Philological Seminary Library.

4.4 : Purchased from

. \G UNIVERSITY FUNDS.

CORNELL UNIVERSITY.

 
"aici ;
THE LATIN LANGUAGE

W. M. LINDSAY
Zondon
HENRY FROWDE

OxrorD UNIVERSITY Press WAREHOUSE

AMEN CorNeER, E.C.

 

Mew York

MACMILLAN & CO., 66 FIFTH AVENUE
THE

LATIN LANGUAGE

AN HISTORICAL ACCOUNT

OF

LATIN SOUNDS, STEMS, AND FLEXIONS

BY

W. M. LINDSAY, M.A.
FELLOW OF JESUS COLLEGE, OXFORD

Orford

AT THE CLARENDON PRESS

TRA
Orford

PRINTED AT THE CLARENDON PRESS

BY HORACE HART, PRINTER TO THE UNIVERSITY
TO

PROFESSOR ROBINSON ELLIS
PREFACE

—

SINCE Corssen’s great work (last edition, Leipzig, 1868-70),
there has been no book devoted to a separate investigation
by Comparative Philological methods of the Latin Language,
its declensions, its conjugations, its formation of the various
parts of speech, and the changes of its pronunciation and
orthography, if we except the short summary (last edition,
Nordlingen, 1889) written by Professor Stolz for the Iwan
Miiller Series of Handbooks of Classical Antiquity. And yet
the additions to our knowledge of the subject since Corssen’s
time have been very great. Not only has the whole Science
of Comparative Philology been, by the help of men like
Johannes Schmidt, Osthoff, and Brugmann}, set on a sounder
basis, but a vast amount has been added to our knowledge of
the Early Latin authors, especially Plautus, of the Umbrian,
Osean, and other dialects of ancient Italy, of Romance, and
above all of the Celtic family of languages, a family closely
united with the Italic group. The time has surely come for
a new treatment of the subject, such as I venture to offer in
the ten chapters of this volume.

I should have liked to have added to them a fuller dis-
cussion of the relation of Latin to the other languages of
Italy. But I had already exceeded the generous limits

 

1] take this opportunity of ac- Grammatik in chaps. iv-viii, and to
knowledging to the fullest extent Seelmann, Aussprache des Lateir in
possible my indebtedness to Brug- chap. ii.
mann, Grundriss der Vergleichenden
vill PREFACE.

allowed by the Delegates of the Press, and it seemed to me
that until more evidence is forthcoming in the shape of
dialectal inscriptions certainty can hardly be attained. It is
much to be desired that some of the money which is being
raised every year for excavations should be devoted to this
field of research. The records of peoples like the Samnites,
who foughi so gallantly with Rome for the rule of Italy, and
whose religion and manners so greatly influenced the ruling
race, should not be allowed to lie neglected. And yet, while
the Latin, Greek, and Etruscan inscriptions of Italy are care-
fully sought after year by year. there has been practically no
organized search for the remains of Osean, Umbrian, Pelignian,
and the rest. I trust that some step may be taken ere long
in this direction.

It remains for me to acknowledge with gratitude the kind
help which I have had from numerous correspondents, both
in this country and abroad, as well as from my Oxford
friends, such as my colleague, Mr. E. R. Wharton. My
special thanks are due to Mr. Sweet for looking through
the proof-sheets of my chapter on Latin Pronunciation, and
to Professors Mommsen, Bormann, Huelsen, and Dressel for
giving me access to the advance-sheets of the Corpus Inscrip-
tionum Latinarum. My friend, Mr. J. A. Smith, Fellow of
Balliol College, has been good enough to go over the whole
book in proof, and to give me many valuable suggestions,
especially on one of the most difficult problems of the
language, the formation of the Perfect Tense.

OxFoRD, August, 1894.
TABLE OF CONTENTS

eggs
List or ABBREVIATIONS
CHAPTER I.
THE ALPHABET.
SEC.
1. The Alphabet
a. The Alphabet of twonty-ons eters
3. The letter F Q 5
4.x
5.4 2 ‘ e ‘
6. The Guttural-symbols
q. Y- and W-sounds
8. Double Consonant
g. Signs for long vowels
to. gg for ng :
11. New Letters for rede Sonntik
12. Influence of Greek Orthography
13. Syllabic Writing 3
CHAPTER II.
PRONUNCIATION.
LA ;
2. Description of re Nain, by Latin Phonotieians
3. Interchange of a and e :
4. Interchange of a and o
5. Anomalies in Romance
6. E
} Deserdptions of the E- er ie Latte Phoneticians ke.
8. i for unaccented é : F
9. i for é in hiatus . 5
to. ‘Rustic’ e foriin hiatus . é .
11. i for é i i
12. i for unaccented é 2 F F ‘ 3 ‘ :
13. & for é 2 ‘ 3 5 :
m4 I i . a . . . qi i
15. Descriptions of the I-sound by Latin phoneticians
16. by Grammarians

PAGE
xxvi

on anunaun Hy

13
17
17
17
18
18
20
aE
21
22
22
23
23
23
25
25
x TABLE OF CONTENTS.

SEC. PAGE
17. Interchange of i and e ; ‘ i é ‘ ‘ r ‘ » 29
18. iin hiatus . 5 5 2 ‘ . . 3 . + go
19. Anomalies in Hamangs Fi ‘ ‘ 3 . a e 3 + 30
20. O rs . : . ~ 339
ar. Deseaeons of the O- sountl by Tati phonsticians 2 ‘ 32
22. Close for open o in accented syllables before certain coriarycttt
groups. ‘ é : “ . ‘ F . ‘ - 32
23. ufor unaccented5 " c * ‘ , c ‘ ‘i + 33
24. u ford = A . i 3 é 5 Z ‘ - 33
25. Other changes of 6 annie 6 s 2 ‘ ‘ : ‘ ° - 84
ob. UY « . . . - 34
27. Descriptions of jhe tieouad by Latin phoneticiana ‘ 5 : » 35
28, Greek v in Latin z ‘ $ ‘ ‘ ‘ ‘ 36
29. o fort “ é % : ‘ * x z ¥ « BF
30. 6 for a F ‘ é : z : 3 . ; - 37
gi. Other changes art ua and a. . . & 3 . « 37
32. Diphthongs if ‘ a i . . - 37
33. Grammarians’ siencu of aipiGnonbe i 2 é ‘ . - 39
34. Ter. Maurus on au. ‘ : " é ‘ s r ‘ » 42
35. auin Romance . ; ‘ ‘ ” 4 ‘ , x a - 40
36. u for accented au Z ‘ 4 i ‘ ‘ a c - 40
37. oandau. i 2 x : A 3 . 7 - 40
38. a for au x ‘ ‘ : ‘ . é . x 42
39. Greek ieimeanigiions of au . ‘ ‘ : i 42
40. aeforau. ‘ dé ‘ i é ‘ . 42
41. e for ae “ é ‘ 2 ‘ é r ° » 42
42. aiforae . z a Fi ‘ : - 43
43. Greek a ‘ n ‘ 5 és . ‘ * ‘ . 43
44. oeande . fi " ‘ i é ‘ c 3 « a4
45. 0c in Romance . ‘ ‘ ‘ ‘ 4 * ‘ Z . 44
46. Greek ev . ‘ ‘ “ 9 ‘ % . ‘ ~ 4A
47. uiofeui. ; ‘ r a é 3 5 : $ - 44
48.J3,V . ; a 7 2 4 F a i - 44
49. Testimony of wamerlans 7 i F 5 ‘ - 47
50. j and v in early Latin 3 F é ‘ , 5 é 5 - 48
51. in late Latin and Romance & a . 7 3 - 49
52. v confused with b in late Latin and Penta 5 . 3 3 - 49
53. Intervocalic v dropped 3 . : ‘ ‘ E ) 3 » §2
54. Postconsonantal v dropped i Ci : Fi ‘ C ‘ - 5§2
55. ai, ei before a vowel . " 3 3 : ‘ z 53
56. H _ 3 i , é “ ‘ 5 i . . ‘ . 53
57. Testimony of grammarians ; : Zi ‘ ‘ a : » 55
58. h between vowels P « , i x é ; , < » 54
59. hin Old Latin . ‘ 7 ‘ 3 é 4 s ; » 57
60. Greek aspirates in Latin . ‘ b « « : é ‘ - 57
61.M,N . a 5 . ‘i ‘ Ss 5 5 60
62. Phonetic descriptions of ripen: m,n i A 4 < 5 . 65
63. The Agma . 2 ‘ : . r ; . ‘ . . 65
64. m, n before consonant : . : : ‘i “ : é - 65

65. Final m ‘ 3 ‘ 2 : . : * ‘ 3 . . 67
SEC.
66,
67.
68.
69.
70.
qT:
72.
73-
14-

76.
77°

79+
80.
81.
82,
83.

85.
86.
87.
88.
89.
go.

gl.
92.

93-
94.
95-
96.
97-
98.

99-

100.
Ior.
102.
103.
Io4.
I05.
106.
107.
108,
109.
IIo.
IQ.
II2,
II.
II4.
115.

TABLE OF CONTENTS.

ns
nx
mn
gn
net
nd
Par: neitie jem in Gat leavisaan ds
Tenues and mediae
Greek tenues in loanwords %
. Confusion of mediae and tenues in Latin wor ds
Mediae and tenues at end of word
Mediae and tenues in the Dialects
. B, P ‘ a
Phonetic descriptions of b p
bs, bt
ps, pt
bm, mb.
b and dialectal f
- bandm
D, T.
Phottetic dspesentious of a, t
dand1
dandr
tl : z . -
Assibilation of ty, dy
K,C,G,QU,GU .
Phonetic descriptions of the Gutturals
qu, gu ‘ é ‘
c, g before narrow vowels .
et, tt .
LR. ~ é
Phonetic descriptions of 1.
ofr
The grammarians on ‘itis pronundlation a 1
of r
Inter disineea af r snd 1
Parasitic vowel with 1, r .
Avoidance of two r’s
rs
rn
In. é ‘
1 before consonant
rl , : * 3
r before consonants .
final r 5 é ‘
Metathesis
ly. : é ‘
ry. . °
F

Descriptions of the sound of f .

Xi

PAGE

Too
xii

SEC,

116,
117.
rz,
11g.
120.
rat.
r22,
129)
124.
125.
126,
127.
128,
129.
130.
I3I.
132.
133.
134.
135.
136.
137.
138.
139.
140.
141,
142.
143.
144.
I45-
146,
147.
148,
149.
150.
I5!.
Pea.
153.
154.

TABLE OF CONTENTS.

mf

Ss, X, z

Phonetic descriptions of a
Latin sin Romance .

Greek ¢, Latin z a
Old Roman z 2
Old Roman s (z), late r
Prosthetic vowel with st, &e.
s before a consonant

><

Final s ‘

Double consonants .

Testimony of the grammarians :
Reduction of 11 to 1, ss to s, after a diphihonig | or fone | r
Confusion of single and double letter in Latin

Double consonants in Italian

Double consonant (not J, s) after long eae

Final double consonant

Final consonants

‘Sandhi’ in Latin

Latin ‘ Doublets’ ‘

Dropping of final consonant in Dain

Dropping of final consonants in Romance

Syllable-Division

Testimony of grammarians

Quantity .

‘Position ’

Shortening of long samme bafore qhother mwa é
Change in quantity of vowel before certain consonant-groups
r with consonant

s with consonant

n with single consonant

1 with consonant

Crasis of vowels, Synizesis, &e.

Vowel-contraction in compounds in the early dranaatiate
Synizesis in Late and Vulgar Latin

Other examples of vowel-contraction

Elision

Parasitic aneearishe

CHAPTER III.

ACCENTUATION,

. Nature of the Latin Accent . . ‘

2. Testimony of the grammarians. (1) On the Nature of the Latin

Accent 3 F 2
(2) On the etréumiiex fitout ay.

3. Accentuation of Greek loanwords

4.

Romance Accentuation . .

PAGE
roo
Io.l
103
103
104
I05
105

‘105
107
IO7
108
108
TIO
IIo
IIg
118
118
IIg
IIg
121
122
122
124
124
725
126
129
131
133
140
141
Iqt
142
142,
143
144
144
144
145

148

154
154
155
156
TABLE OF CONTENTS.

SEC.

5. The Earlier Law of Accentuation

6. Traces of I.-Eur. accentuation in Latin

7. Secondary and main accent

8. The Paenultima Law :

g. Testimony of the grammarians .

io. Exceptions to the Paenultima Law

11. Vulgar-Latin Accentuation

12, Accentuation of the Sentence .

12a. Latin Sentence-Enclitics

13. Syncope z

14. Syncope in the Bueiiedina Gisiect of tan

15. Syncope under the Old Accent Law .

16. Syncope of Final Syllable . : é ‘
17. Syncope under the Paenultima Recent idee (1) Pretonic

(2) Post-tonic .

18. Change of Unaccented Vowels.

19. Other Examples. I. Syllables long ay positon

20. II. Short Syllables (1) in -r

21. (2) in -1 or Labial

The Parasitic Vowel
22, in other short syllables
23. (3) Diphthongs, i ae
au 3 2

24. (4) Dichthenes4 in iietus ,

25. (5) je and ve F

26. (6) Later change of o to u, u to di

27. (7) Greek words with Vowel-change

28. (8) Vowel unchanged. i. in Latin words

29. ii. in Greek loanwords

30. (9) Long vowels
31. (10) Recomposition and Avnileay
32. (11) Pretonic
33. (12) Assimilation, Dissimilailon, ead False Ainatopy
34. (13) Shortening of Syllables long by position

35. Change and Shortening of Vowel in Unaccented Final Syllable =

I, Loss or Syncope of Short Vowel
36. Loss of -e ‘

37. II. Change of Vowel . é

38. Change of final short vowel to é

39. Alternation of final e with internal i

40. III. Shortening of Long Syllable

41. Final long vowel in Hiatus

42. Breves Breviantes

43. Shortening of final -4

44. Shortening of final -é

45. Shortening of final -6

46. Shortening of final -i

47. Shortening of final -a

48. Shortening of final aiphithons

49. Shortening of long vowel before final Conacnatat

xiii

PAGE

‘157

159
159
160
162
163
164
165
166
I7o
177
178
181
183
184
185
I9l
192
192
193
194
195
196
196
196
196
197
198
198
199
199
200
201
201
203
203
204
205
206
206
207
209
210
210
211
212
213
213
213
213
Xiv TABLE OF CONTENTS.

SEC. PAGE
50. Shortening of Final Syllable long by ae é 3 . 214
51. Shortening of Monosyllables . ‘ 215
52. Loss of Final Syllable with -m . $ ‘ . 216

CHAPTER IV.

THE LATIN REPRESENTATIVES OF THE INDO-EUROPEAN SOUNDS,

ILA ‘ . . i ‘ . : . : « BIO.

2. Latin 4 for I.-Eur. a , ‘ . : 221

3. A : : r ‘ P ‘ . . ‘ : F . 2aI

4. L-Eur. i. : i i ‘ ‘ », 223

5. B ; 7 ‘ . * . é 7 ‘ . . 223

6. Lat. éforI.-Eur.é  . ‘ 2 ‘i 4 i . ‘ . 224

q. ifor é é ‘ é r rl 7 ‘ + 225

8. E ‘ A ; é j 5 ‘ 5 ‘ : + 225

g. Latin é for I.-Eur.é . i : fi a 226
10. 6for 8 with wand! . a és . 226
11. i for (accented) e : : 3 - : - 229
12, I s 2 4 : : : ‘ Ny . 230
13. I 3 : : ; APs , : é ‘ Y j . 231
14. Other examples of Lat. i for I.-Eur. i : : 232
15. ié,notii. ‘ r . 2 3 : “ . 232
16. 0 s ‘ ‘ ‘ ‘ : ; - - : & : . 232
17. 0 gs 1s ‘i ’ ‘ : : : . ; ; ‘ 233
18. Latin 6 for I.-Eur.6 . : , : ‘ 234
1g. Latin & for I.-Eur. 6, under influence of v : % » 235
20. ti for 6 a a 3 ‘ , Z fl ‘ 235
2a1.U ‘ ‘ ‘ z ‘ S : is ‘ 237
22, Other examples of Lat. ii, I.-Eur. a ‘ ‘ ‘ + 237
23. U Tay: ~ x 2 : . 237
24. Latin ti for L.-Eur. t : ‘ ‘ . 238
25. Latin itiand Latiné . ‘ : ‘ ‘ ‘ . i . 239
26. The Diphthongs j ‘ ; ® 5 » 239
ay. AL 2 * a < é ¥ * ~ Bae
28. I.-Eur. ai, Latin ae (ai) Sj + 242
29. AI, AE on Inser: soit é 2 é ‘ 3 242
30. AU : ‘ ‘ s . 242
31. Other esaraples 243
32. EI 243

33. Other examples of I. Hint ei

244
34. EI and Lin ee &e. 244
35. EU. . 245
36. Other examples of I. -Enr. eu 246
37. OU, U in Inscriptions 246
38. OT a Fi 246
39. Other examples of I. aks oi 247
40. OI, OF, U on Pie, 247
41. ou 248

42. Other examples of I ait, ou 249
SEC,
. & for older ovi, ové

. The spurious ee ou

. AL . 7 :

. AU

. EL

. BU

. OL

. OU ,
. Variation (Ablaut) of ‘Towels ‘i
. I.-Eur. and Latin éandé6 .

. €ands

. O-d, 6-A

. a-5

. Band &

. &and &

- iandi

. 6and 6

-tiandi

. 6anda

. ou-su

TABLE OF CONTENTS.

Y

. I-Eur. Initial y -

. I.-Eur. y preceded by a eguisenanit

. I.-Eur. y between vowels

. Latinj

. L.-Eur. initial w

. L.-Eur. w (and Latin v) ieiaden oe
. I.-Eur. w after a consonant

I.-Eur. w before a consonant

_N

-M,N.

M 3 3 : ‘ ;
. L.-Eur. m; other examples
. n form
. I.-Eur.ms .
. L.-Eur. mr, ml
. I.-Eur. n; other examples
mn.
. The M- anit N- Souanti

. Other examples of the Nasal onan:
. Other examples of am, an, ma, na

~L

. R

. LR
. I.-Eur.1; other examples .
. L-Eur. r; other examples .

. ss for rs before consonant
. rr for rs before vowel .
.nforr % <

XV
PAGE
250
250
251
252
252
252
252
253
253
258
258
258
259
259
260
260
260
260
261
261
262
264
264
265
265
265
266
267
267
268
268
269
270
270
270
270
271
272
272
273
274
274
275
275
276
276
277
277
277
278
Xvi TABLE OF CONTENTS.

SEC, PaGk
92. The L- and R- Sonants . , : . 278
93. Other examples of the liquid Seuaiits ‘ 3 7 : 279
94. Other examples of al, ar, 14, ra j ‘ ® : . 279
95. Tenues, Mediae, and Aspirates » 279
96. Media or aspirata assimilated to amiielued tatecmante in Takin, » 281
97. Tenuis assimilated to voiced consonant in Latin . * x . 281
98. P é ‘ 4 é é . . ¥ , . 281
99. Other examples of I.-Eur.p . . . ‘i : . 282

Ioo. B : ‘ a ‘ é ‘ i . A a x 282
ror. Other examples of I.-Eur. b : g : 3 : . 282
to2. mn for bn ‘ ‘ ‘ ‘ i : , 282
103. BH . ‘ ‘ . j a . 282
104. I.-Eur. bh ; ater srentales : i a » 283
105. Tt . . . . . 283
106, Other exianriles of Tabiae, t ‘ : , . 284
1o7. L.-Eur. tl , si . i * c : . 284
ro8. I.-Eur. tt . i 7 . ‘ a . 284
tog. D z i , ‘ x i ‘ r ‘ - 284.
110. Other instances of I.-Eur. d, Latin d : 2 . 285
tir. Latinlford . 3 3 ; e ‘ : - 286
112. Latin r for d r é 5 : ‘ . 287
113. trfordr . : : E : - 289
114. DH . ‘ . ‘ « 289
115. Other examnlen of I.-Eur, dh s ; . 289
116. The Gutturals . a i 7 + 290
117. x for Guttural with s ; 3 ‘ ‘ ‘ + 293
118, ct for Guttural with t ; * : F : » 293
11g. gn, gm for en, cm 2 ‘ : 3 : + 293
120. Latin h dropped between eowals : ‘ 2 < , » 204
121. Dialectal f for h 294
122, The Palatal Gutturals: &, G, Ku, 6H 5 : ; ‘ + 295
123. Other examples of I.-Eur. k F ‘ ‘ ‘ ‘ i » 205
124, I.-Eur. kw 296

*125. & : . 296

126. Other aetnglen: of I.-Eur. B , e . . 296"
127.GH . 296
128. Other examples of I. ‘ie: €h 297
129. The Gutturals Proper: K, G, GH, KH 207
K 5 : is 297
1g0. I.-Eur. k ; other dseeanples 298
131.G@ . i‘ : : . 298
132. Other la an of I.-Eur. g 298
133. GH . 298
134. I.-Eur. fis dior saints ‘ é ‘ 298
135. Velar Gutturals with Labialisation: Qu . 299
136. I.-Eur. q3, Latin qu; other examples 300
137. ¢ for qu 5 ‘ ; 300
138. Latin qu of ethos eneta 301
139. R Zor

140. I, Bar, gy Latin v3 ieflige examples 301
SEC.

41.

142. g
143.
144.
145.

746,

147.
148.
149.

I50.

I5I.
152.

153.

154.
I55-
156.

157.
158.
159-
160.

161.

162,
163.

TABLE OF CONTENTS.

Dialectal b

g for I.-Eur, g3

eH» . ‘

I.-Eur. gh? in tating ether examples
The Sibilants: 8, Z

S, Z ‘1 3 .
I.-Eur. s, Latin s ; athens pusieailag ‘
Latin r for isieevoralio sibilant
Initial sibilant before consonant

O. Latin stl, sl, sel :
Sibilant before voiced consonant in aide of etd
Sibilant before r in middle of word

Assimilation of sibilant to preceding r, 1 :
Assimilation of preceding dental to the sibilant
Latin ss for tt

Other groups with a sibilant

Loss of Consonant in Group .

Other examples

Assimilation of Contnnusnts ‘i
Assimilation in Preposition compounded ‘with Verb
Other examples of Assimilation

Lengthening by Compensation

Assimilation of Syllables

CHAPTER V.

FORMATION OF NOUN AND ADJECTIVE STEMS.

1. I, STEM-SUFFIXES

2.

CO MA DUDA W

Hees
BR wON H

15,
16.
17.
18;
19.
20,
21.

Suffixes ending in -6, -a (Nouns anil Rajpatives of the First anid
Second Declension). -O-, -A- 4

. Latin O- and A-suffixes; other examples .
. -16-, -IA-, (-YO-, -YA-)

-U0-, -UA-
I.-Eur. Stems in -wé-

. Latin Verbal Adjectives in -uus, ise -tivus

-NO., -NA-.

. L.-Eur. NO-suffix .

. Latin -nus

. Latin -iInus .

. Latin -Anus

. -MENO-, -MENA-
. -MO-, -MA-

Other examples ‘ , i é :

-RO-, -RA-. : i ‘ ‘ : f i ‘i
Other examples of ihe RO-suffix ; 5 : ‘ : :
Examples of I.-Eur. -téro- and -eré- in Latin é ‘ ‘
I.-Eur. -tro- ‘ ‘ - : : ‘ ‘ , ;
I.-Eur. -dhro- ‘ ‘ ; ‘

-LO-, -LA-. BO

xvii
PAGE
go2
302
302
go2
302
303
305
395
306
307
307
308
308
309
309
309
309
310
Sir
313
314
314
315

316

316
318
318
322
322
323
324
326
326
326
326
327
328
328
328
33°
330
330
33
331
xviii TABLE OF CONTENTS.

SEC.
22.
23.
24.
25.
26.
27.
28.
29.
30,
731.
32.
33-
34.
35-
36.
37-
38.
39-
40.
41.
42.
43.
44-
45-
46.
47-
48.
49.
50.
51.
52.
53-
54:

55:
56.
57:
58.
59:
60.
61.
62,
63.
64.
65.
66,
67,
68.
69.
70.

Adjectives formed by the LO-suffix
Nouns denoting the Agent or the Instrument
Diminutives
Neuters formed with he siti Ale: ‘
The suffix -dhlo-
-TO-, -TA-.
Participles in -tus
Abstract Nouns in -ta (-sa)
Neuters in -mentum
-KO., -KA- ‘
Adjectives with the KO- aiiix
Adjectives in -icius
Suffixes ending in i (Nouns and Adjectives of third Detlonsiany:
Other examples of I-stems
Adjective I-stems from O-stems

-NI-

Other examples of hatte -ni-

-MI-

-RI-, -LI-

Other examples of Latin it -ri-

“PIs 4

Other ssonnwolte of fits: suffix -ti- in Latin

Examples of Latin -tidn-
Adjectival -ti- for -to- in Latin
Other examples of Latin -tat(i)-, -tit(i), -tidin-
Suffixes ending in -Gi (Nouns of fourth Decl.).  -U-
Other examples of U-stems in Latin .
Interchange of U- with O-stems
Other examples of TU-stems ‘
The Suffixes -YH- (Nouns of fifth Deel: and -I-. The Stems in -E.
Other cxamples of Latin Fems. in -7, -ic, &c
-yé- and -i- . j
Suffixes ending in -n oun’ of third Teal; ). -EN-, -YEN-, -WEN.,
-MEN- H
Mase. EN-stems in atin
Suffixes ending in -r (Nouns of third ‘Deal, ). -R-
Neuter R-stems
-ER- and -TER-.
Nouns of relationship .
Latin Nomina Agentis
Suffixes ending in -t (Nouns and RAjoctives of third Deel. Ds
Other examples of Latin T-stems
-NT-. ;
Other examples of farina seine
-WENT-
Other examples of Batite -dsus
Suffixes ending in -d (Nouns of third Decl. :
Other examples
Suffixes ending in a Guttural iiauna ana ‘Aapestiscen of thant Decl; J.
Other examples y

PAGE

332
332
333
333
334
334
335
336
336
336
337
337
338
338
338
339
339
339
339
340
340
341
341
342
342
342

343

343
344
344
347
347

348
349
349
349
350
350
350
350
351
352
352
352
353
353
354
354
355
TABLE OF CONTENTS. XLX

PAGE
71. Suffixes ending in -s (Nouns and Adjectives of third Decl.). -E8- . 355
72. Neuter ES-stems in Latin . ‘

355
73. Adjective ES-stems 356
74. Mase. (and Fem.) ES-stems 356
75. Other 8.stems 357
76. -YES-. : 357
77. Suffixless Forms . ‘ » 357
78. Suffixless stems at end of cheap und in stir. - 358
79. Latin Independent suffixless stems 358
80. II. COMPOSITION ‘ 358
81. Reduplicated Nouns and Atiectives in iistin 363
82. A-stems : f ‘i i A . 363
83. O-stems 364
84. I-stems 364
85. U-stems 364
86. N-stems . 364
87. R-stems 365
88. Dental and Cutiurl Steines 365
89. S-stems » 365
go. Stem-suffixes andl Compesition s in famnee 7 i 365
CHAPTER VI.
DECLENSION OF NOUNS AND ADJECTIVES. COMPARISON OF
ADJECTIVES. NUMERALS.
1. I, DECLENSION OF NOUNS AND ADJECTIVES 366
2. Nom. Sing. I. Masc., Fem. ‘ » BFE
3. Nom. Sing. of A-stems in Latin . . . 373
4. RO-stems 3 374
5. YO-stems 375
6. I-stems 375 _
q. S-stems ‘i ‘ ‘ Z . 376
8. N-stems 3 ‘ , ' . 376
g. Diphthong sionaw- 3 5 2 i é = 397
10. Nom., Ace. Sing. II. Neut. : , ‘ 371
11, O-stems i ‘ “ ‘ i 378
12. I-stems ‘ . ‘i ; . . 378
13. U-stems i é . 2 ‘ i ‘ . 378
14. S-stems é s ‘ Z é : ‘ - 378
15. R-stems: . 379
16, -Sin Nom. Sg. Neut. of Adjectives 379
17. Gen. Sing. . : = ‘ ‘i z ‘ i ‘ . - 379
18, A-stems ‘ a ‘ i F i : A ‘i - . 38t
19. Fifth Decl. Stomi : : : . . : i a » 382
20, O-stems and I0-stems : 5 ; ‘ 4 - 383
21. U-stems. ‘ : a ‘ , 2 : ‘ . » 384
a2, Consonant-stems ~ c : . . a ‘ 5 . 384
23. Dat. Sing. 7 - P ‘ : . : . 385
XX

SEC.

. A-stems ‘

. Fifth Decl. Stems
. O-stems

. U-stems

. Consonant-stems
. Acc. Sing. .

. The endings -im and -em

. Voc. Sing. .

. Other examples

. Abl. Sing.

. O. Latin Abl. wl -d é
. I-stem and Cons.-stem ‘Abl.’ in -i and -e .
. Instr. Sing.

. Locative Sing.

. Locatives in -I and -e in Latin

68.

TABLE OF CONTENTS.

A-stems, &e.

. Nom. Plur. I. Masc., Fem.
. A-stems

. O-stems

. I-stems

. Cons.-stems

. Nom., Acc. Plur. II. Neut.
. Gen. ‘Blew,

-um and -orum in Disianiac,

. Dat., Abl., Loc., Instr. Plural

. A-and O-stems .

. Other stems

« Ase Plur, » ' »

. II. THE COMPARISON OF ADJECTIVES
. The Comparative Suffixes

. The Superlative Suffixes

. Some irregular Comparatives and Supetintives
. ITI. NUMERALS

. One

. Unus

. Two

. Duo

. Three .

. Tres

. Four

- Quattuor

. Five

Quinque

. Six

Seven

. Hight .

. Nine

. Ten ‘

. Eleven to Nineteen

. O. Latin duovicesimus

&

PAGE
386
386
387
387
387
387
388
388
389

391
392
392
395
396
397
397
398
398
399
399
399
401
402
402
403
404
404
404
406
407
407
408
409
410
410
41I

412
413
414
414
414
415
415
415
415
416
416
417
oe GON ON On as 100) 8) oF

et

24,
25.
26.
27.
28.
29.
30.

TABLE OF CONTENTS.

. Twenty to Ninety

. Viginti, &e.

. The Hundreds

. Centum, &e.

. The Thousands

. Mille

. The Numerals in Romans:

CHAPTER VII.

THE PRONOUNS.

I. THE PERSONAL PRONOUNS AND THE REFLEXIVE.

Declension of ego
2 Sing.
Declension of tu.
Reflexive
Declension of sui
rt Plur.
Declension of nos
2 Plur. ‘
Declension of vos

. Il. THE POSSESSIVE PRONOUNS

Their forms P
III. DEMONSTRATIVES .

. O. Latin so-

. The particle -ce .

. Hic

. Iste

. Ile

. Is

. Lpse

. Idem

. The Pronominal Gen. and Dat. Se. :
. IV. RELATIVE, INDEFINITE, AND INTERROGATIVE PRO-

NOUNS .
Stems qi- and qio-
Case-forms .
The stem q3u-
The Possessive cujus .
Other ‘Derivatives ‘i
VY. THE PRONOMINAL ADI ECTIVES
The Pronouns in Romance

CHAPTER VIII.

THE VERB.

. I. THE CONJUGATIONS é
. Traces of the Athematic Conjugation in Tati

1 Sing.

xxi

PAGE
417
418
418
419
419
420
420

421
422
423
423
424
424
424
425
425
426
426
427
429
432
432
433
435
436
437
440
441
442

443
444
445
446
447
447
449
452

453
455
XxXil TABLE OF CONTENTS.

SEC.

00 OI AM FW

Te
iQ.
14.
15.
16.1
19.
18,
19.
20.
ar.
22,
23.
Bais
25.
26.
25:
28
29.
30.
aly
32.
33:
334.
34-
35-

37:
38.
39.
40.
41.
4s
43-
44.
45-

41:
48.
49.
50.
51.

. II, THE TENSE-STEMS (Strong Aorist and S-formations
. ‘Strong Aorist’ forms in Latin . ‘

. Old Latin forms with -ss- (-s-) .

. A. Present. (1) With B-grade of mage an Thematie Vowel

. Other examples . 5 ‘ ;

. Weak grade of root

. (2) With reduplicated root ‘ i ‘ : :
. (3) With root nasalized. i. With sane fats ii. With nasal affix .
. Other examples of nasal infix

Retention of Nasal throughout the Monges:.
Other examples of nasal affix
Other Verb-stems with n
(4) With suffix -YO-, -TYO-
1 in the third Conjugation Presents sil YO. ee :
Other examples of E-grade roots
Of weak grade roots :
Alternative forms in -o and -eo .
Of roots with -a, -6, -d R
Inceptives, and other Verb-stems
(5) Inceptives in -ské- (-sk6-)
Causatives and Intensives in -eyo-
Latin Desideratives in -ttrio
Latin Iteratives or Frequentatives in Atays-
Other Derivative Verbs with the YO-suffix
Other suffixes ,
Other examples of Latin Inceptives
Of Latin Causatives, &c.
Of Latin Desideratives
Of Latin Iteratives :
Of Latin Derivative verbs with ‘YO-sufix .
Of other Verb-suffixes
The Conjugations in Romance
B. Imperfect
Fourth Conj. Impft. in tha

. O. Future.

Fourth Conj. Fut. in oo
Third Conj. Fut. in -ébo

D. Perfect ‘
Other examples of Redupticatad forms
Unreduplicated : . ,

Form of Reduplication

Assimilation of Reduplication- sail to Stee:
Loss of Reduplication

Co-existent Reduplicated and Unrsduptiontal faring

. S-Preterite

Origin of the Perfect in -vi ae aij ,
Shortened forms of the Perfect in -vi
Shortened forms of the Perfect in -si
QO. Latin Perfects in -ai(v)i

Some Irregular Perfects

FAGEK
459
404
465
466
467
467
468
409
471
471
472
472
472
475
475
476
476
476
476
471
477
478
478
478
478
479
481
482
482
483
486
488
489
491
491
493
493
494
501
501
502
503
503
504
504
505
506
508
508
509
SEC.

52.
53-

55-

TABLE OF CONTENTS.

E. Pluperfect
F, Future-Perfect

. G. Tenses formed with Kaattipey Werbee :
III. THE MOODS. A. Subjunctive. (Relics of the I.-Eur.

Optative Mood in Latin.)

. Some O. Latin Subj. and Opt. forms
. B. Imperative .

. Other examples of 2 Sg. iendee with hiaes stem
. Other‘examples of Imper. in -tdd
. Imper. Pass. 2, 3 Sg. in -mind

- 3 Pl. Imperat. .

. IV. THE VOICES .

. Impersonal use of Latin Passive
. Active and Middle

. V. THE PERSON-ENDINGS .
. (1) Active. 1 Sg.

. 28g.

i Athematie: Sg. of féro, ¥8l6

. 38g... :

. The 3 Sg. Pft. in atin

. 2 Plor.

. 2 Plur.

. 3 Plur.

. 3 Pl. Pres. in -nunt

. 3 Pl. Perf.

. (2) Passive Deponent) 1 Sing.
. 2 Sing.

. Use of -re and -ris

. 3 Sing.

. 1 Plur.

. 2 Plur.

. 3 Plur.

. VI. THE INFINITIVE

. Pres. Inf. Act.

. Pres. Inf. Pass.

. Fut. Inf. Act.

. Fut. Inf. Pass.

. VII. THE SUPINES

. VIII. THE PARTICIPLES

Pres. Part. Act. .

. Perf. Part. Act. .

. Perf. Part. Pass.

. ‘Truncated’ Participles .

. IX. THE GERUND AND GEKUNDIVE
. Origin of the suffix -ndo-

. Adjectives in -bundo-, -cundo-, &e.

. Some Irregular Verbs

. Irregular Verbs in Romance

XxXili
PAGE
509
510
510

511
514
516
517
519
519
519
519
521
521
522
524
525
526
526
527
529
529
529
531
531
532
533
533
534
534
534
534
535
537
537
537
538
538
539
54°
54i
541
543
543
544
544
545
547
XXIV TABLE OF CONTENTS.

uw
S
2

an ant wn He

CHAPTER IX,

ADVERBS AND PREPOSITIONS,

. ADVERBS

. Nominative Adverb-forms

. Genitive Adverb-forms

. Accusative Adverb-forms .

. Ablative (Instr.) and Locative Adver < forms
. Adverbs in -tus .

. Adverbial word-groups and dont pouty
. Other Adverbs

. Numeral Adverbs in -ies

. Pronominal Adverbs .

. PREPOSITIONS .
. Ab, ap-, po-, abs, a-, au-, af, absque

. Ab, abs, a

. Af

. Ad

. Ambi-

. An

. Ante

. Apud

. Circum, cirea, circiter

. Clam, clanculum

. Com-, (cum), with, and co-

. Contra (see §§ 1, 4)

. Coram ‘i

. De

. Dis

. Endo

. Erga, ergo

. Ex, ec-, e

Extra

. In

. Infra

. Inter

. Intra, intus
. Juxta

. Ob

. Palam

. Penes

. Per

. Po-

. Post, pone

. Poste, posti-d, pos, po-
. Prae

. Praeter

. Pro, por-

. Prd- and pré-

PAGE
548
553
555
555
559
561
562
565
567
567
572
575
577
577
5717
571

.578

578
579
579
580
581
581
585
581
582
582
583
583
584
584
585
585

585
585
586
586
586
588
588
589
589
589
590
540
TABLE OF CONTENTS.
SEC.
46. Procul
47. Prope
48. Propter
49. Re- ; é
50. Secundum, secus
500. Simul
51. Sine, se
52. Sub, subter, aiid
53. Super, supra, insuper, superne
54. Tenus
55. Trans
56. Uls, ultra
57. Usque
58. Versus, versum, iain renee sendveias ceptlaoroume

CHAPTER X.

CONJUNCTIONS AND INTERJECTIONS.

1. CONJUNCTIONS .
2. (1) Conjunctive.—Que, et, aunt ac, quoque, aif
g. Atque, ac
4. (2) Disjunctive.—Ve, ait way snes seu ‘ ‘
5. (3) Adversative.—At, ast, sed, autem, atqui, re ceterum, verum,
vero
6. (4) Limitative and ornssiives Walden, immo ‘
7. (5) Explanatory —Enim, nam, namque, quippe, nempe, nemut
8. .6) Conclusive.—Ergo, itaque, igitur
9. (7) Optative—Ut, utinam
to. (8) Interrogative. Ne. nonne, num, utrum, an, anne, cur, quare,
quianam
11. (9) Comparative. th uti, ies ceu, quam
12. (10) Temporal.—Quum, quando, dum, donee, ut, ubi
13. (rr) Causal.—Quum, quoniam, quod, quia, quippe
14. (12) Conditional.—Si, nisi, ni, sin, sive, seu, modo, dummodo
15. (13) Concessive.—Etsi, quamquam, quamvis, licet
16. (14) Final.—Ut, quo, quominus, quin, ne, neve, neu, nedum
17. (15) Asseverative Particles.—Ne (nae), -ne
18. (16) Negatives.—In-, ne-, nec, non, haud, ve-
1g. Interjections

INDEX

XXV
PAGE
590
Sgr
sor
sgl
sor
592
592
593
593
593
594
594
595
595

596
598
599
599

616

619
LIST OF ABBREVIATIONS

A. L. L.=Archiv f, lat. Lexikographie u. Grammatik, ed. Wélfflin. Leipz.
1884 sqq.

Amer, Journ. Phil. = American Journal of Philology.

Aneed. Helv. = Anecdota Helvetica, ed. Hagen (a Supplement to the Gram-
matic’ Latini, ed. Keil).

Ann. Epigr. =Cagnat, L’année épigraphique. Paris, 1889 sqq.

Ann. Inst.=Annali dell’ Instituto di corrispondenza archeologica. Rome,
1829 sqq. 8

Arch. Glottol., Arch. Glott. Ital.=Archivio Glottologico Italiano. Rome,
1873 sqq.

*AOny. =AOnvaioy avyypappa meprodixcv. Athens, 1872-82.

B. B. =Beitrige z. Kunde d. Indog. Sprachen, ed. Bezzenberger. Gottingen,
1877 sqq.

B. P. W., Berl. Phil. Woch. =Berliner Philologische Wochenschrift. Berl.
1881 sqq.

Brit. Mus.=The Collection of Ancient Greek Inscriptions in the British
Museum, ed. Sir Ch. Newton. Oxf. 1874 sqq.

Bich. Umbr.=Biicheler, Umbrica. Bonn, 1883.

Bull. + Bullettino dell’ Instituto di corrispondenza archeologica. Rome,
1829 sqq. ie

Burs. Jahresber. = Jahresbericht tiber d. Fortschritte d. Classischen Alter-
thumswissenschaft, ed. Bursian. Berl. 1875 sqq.

Cc. G. L.=Corpus Glossariorum Latinorum, ed. Goetz und Gundermann.
Leipz.

C. I, A.= Corpus Inseriptionum Atticarum. Berl. 1873 sqq.

C. I. G. =Corpus Inscriptionum Graecarum, ed. Boeckh. Berl. 1828 sqq.

C. I. L. =Corpus Inseriptionum Latinarum. Berl. 1863 sqq.

Class. Rev. = Classical Review.

Comm. Lud. Saec.=Commentaria Ludorum Saecularium, ed. Mommsen, in
vol. viii of the Ephemeris Epigraphica (also published in the Monument
Antichi, vol. i, part 3).

Comm. Ribbeck. =Commentationes Philologae . .. Ottoni Ribbeckio. Leipz.
1888. :

Comm. Schweizer-Sidler = Philologische Abhandlungen Heinrich Schweizer-
Sidler ...gewidmet. Ziirich, 1891.

Comm. Woelffl.=Commentationes Woelfflinianae. Leipz. 1891.

Eckinger = Eckinger, Die Orthographic lateinischer Wirter in griechischen Inschriften.
Munich.
LIST OF ABBREVIATIONS. XXvii

Edict. Diocl.= the Edict of Diocletian (contained in the Supplement to vol. iii
of the Corpus Inscr. Lat.).

Eph. Epigr. = Ephemeris Epigraphica. Berl. 1872 sqq. (A Supplement to the
Corpus Inser. Lat.).

Etudes G. Paris = Etudes romanes dédiées & Gaston Paris. Paris, 1891.

Etym. Lat.=Etyma Latina, by E. R. Wharton. Lond. 1890.

Fabr. = Fabretti, Corpus Inser. Italicurum antiquioris aevi. Turin, 1867.

Fleck. Jahrb.=Jahrbticher f. classische Philologie, ed. Fleckeisen. Leipz.
1855 sqq.

G). Cyrill., Gl Philox., Gl. Plac.=the Cyrillus, Philoxenus, and Placidus
Glossaries (contained in vols. ii and v of the Corpus Glossariorum Latinorum,
ed. Goetz und Gundermann).

Harv. Stud. = Harvard Studies in Classical Philology. Boston, 1980 sqq.

Herm.= Hermes. Zeitschrift f. classische Philologie. Borl, 1866 sqq.

I. F, = Indogermanische Forschungen,.ed. Brugmann und Streitberg. Strass-
burg, 1891 sqq.

I. I. S.=Inscriptiones Graecae Siciliae et Italiae, ed. Kaibel. Berl. 1890.

I. N., I. R. N.=Inscriptiones Regni Neapolitani Latinae, ed. Mommsen.
Leipz. 1852.

Journ. Hell. Stud. = Journal of Hellenic Studies.

Journ, Phil. =Journal of Philology.

K. Z, = Zeitschrift f. vergleichende Sprachforschung, ed. Kuhn, Berl. 1872 sqq.

Lex. Agr. = Lex Agraria (No. 200 in vol. i of the Corpus Inscr. Lat.).

Lex Repet. = Lex Repetundarum (No. 198 in the same vol.),

Lib. Gloss. = Liber Glossarum (selections from which are contained in vol. v
of the Corpus Glossariorwm Latinorum).

Mél. Arch.=Mélanges d’Archéologie et d’Histoire Paris, 1884 sqq. (The
publication of the Ecole frangaise de Rome.)

Mem. Ist. Lombard.=Memorie dell’ I. R. istituto Lombardo di scienze,
lettere ed arti. Milan, 1843 sqq.

Mém. Soc. Ling., M. S. L.= Mémoires de la Société de Linguistique de Paris.
Paris, 1868 sqq.

Meyer-Liibke = Meyer-Liibke, Grammatik der romanischen Sprachen. Leipz.
1890 sqq.

Mitth.=Mittheilungen d. kaiserlich deutschen archiologischen Instituts.
Athens, 1876 sqq.

Mitth. (rém.) = ditto (rémische Abtheilung).

Mon. Anc.=Res Gestae Divi Augusti: ex monumentis Ancyrano et Apol-
loniensi, ed. Mommsen. Berl.? 1883.

Mon. Antichi = Monumenti Antichi pubblicati per cura della Reale Accademia
dei Lincei. ‘Milan, 1890 sqq.

Morph. Unt., M. U. = Morphologische Untersuchungen, by Osthoff and Brug-
mann. Leipz. 1878 sqq.

M.S. L. (see Mém. Soc. Ling.). é

Mur.— Muratori, Novus thesaurus veterum inscriptionum. Milan, 1739~42.

Neue = Neue, Formeniehre d. lateinischen Sprache. Berl. 1866 sqq.

Not. Seav.=Notizie degli Scavi di antichita (Atti della R. Accademia dei
Lincei). Rome, 1876 sqq.

Or., Or. Henz. = Orelli, Inscriptionum Latinarum Collectio, vols. i-ii, Ziirich, 1828,
vol. iii (Suppl.), ed. Henzen. Ziirich, 1856.
XXvili LIST OF ABBREVIATIONS.

Osthoff, Dunkles wu. helles 1 (see Transactions of American Philological Associa-
tion 1893, vol. xxiv, pp 50 sqq.).

P. B. Beitr. = Beitrige z. Geschichte d. deutschen Sprache u. Literatur, ed.
Paul und Braune. Halle, 1874 sqq.

Philol. = Philologus: Zeitschrift f.d klassische Alterthum. Gdéttingen, 1846
sqq.

Phil. Soe. Trans. = Transactions of the Philological Society.

Phonet. Stud.=Phonetische Studien: Zeitschrift f. wissenschaftliche u.
praktische Phonetik. Marburg, 1887 sqq.

Probi App. = Probi Appendix (contained in vol. iv of the Grammatici Latini, ed.
Keil).

Rev. Phil.: Revue de Philologie. Paris, 1877 sqq.

Rhein. Mus. -- Rheinisches Museum f. Philologie. Frankf. am Main, 1842 sqq.

Riv. Filolog. = Rivista di Filologia. Rome, 1873 sqq.

Rossi = De Rossi, Inscriptiones Christianae Urbis Romae, 2 vols. Rome, 1861-1888,

S. C. Bacch.=Senatus Consultum de Bacchanalibus (No. 196 in vol. i of the
Corpus Inscr. Lat.).

Stud. Ital. = Studi Italiani di filologia classica. Florence, 1893 sqq.

Studem. Stud. - Studien auf d. Gebiete d. Archaischen Lateins, ed. Stude-
mund. Berl. 1873 sqq.

Suppl. Arch Glott. =Supplementi Periodici all’ Archivio Glottologico Italiano,
vol.i Turin, 1891.

Tab. Bunt. =Tabula Bantina (No. 197 in vol. i of the Corpus Inscr. Lat.).

Versamml. Philolog.= Verhandlungen d. Versammlungen deutscher Phi-
lologen u. Schulminner.

Von Planta = Von Planta, Grammatik da. Oskisch-Umbrischen Dialekte, vol. i. Strass-
burg, 1893.

Wien. Stud.= Wiener Studien: Zeitschrift f. class. Philologie. Vienna,
1879 sqq.

Wilm. = Wilmanns, Exempla Inscriptionum Latinarum, 2 vols. Berl. 1873.

Zy. I. I. 1, Zvet. =Zvetaieff, Inscriptiones Italiae Inferioris Dialecticae. Moscow,
1886.

In the transcription of the various I.-Eur. languages the system of Brug- *
mann, Grundriss d. vergleichenden Grammatik, Strassburg, 1886 sqq. (Engl. trans.;
London, 1888 sqq.) is in the main followed, though in ‘ I.-Eur.’ forms Gutturals
Proper are denoted by k, g. &c. (not as in Brugmann by q, g, &c.), and y, w
often replace Brugmann’s j, yu, while in O. Engl. (Brugmann’s ‘ Anglo-Saxon a)
words the orthography of Sweet, History of English Sounds, is preferred. I follow
Brugmann in distinguishing the Oscan and Umbrian inscriptions written in
the Roman alphabet from those written in the native alphabets by printing
the former in italics, a type reserved in this book for Latin words, stems,
suffixes, and sounds, (On the use of k, g, gh see p. 290.)
THE LATIN LANGUAGE

CHAPTER I.
THE ALPHABET!,

§ 1. Ir an alphabet is to express the sounds of a language
properly, each nation must construct one for itself. But this
ideal was not realized by the ancient languages of Italy. The
Oscan and Umbrian stocks borrowed for the expression of their
language the alphabet used by the Etruscans, who had themselves
borrowed it at an earlier period from the Greeks; and so neither
Oscans nor Umbrians were at first able to express in writing
somé common sounds of their language, such as d and 0, which
were wanting in the Etruscan speech (von Planta, Osk.-Umér.
Dial. i., p. 44). The Latin Alphabet, consisting in the later
Republic of twenty-one letters, ABCDEFGHIKLMNOPQRSTVX,
was borrowed from some Chalcidian colony (e.g. Cumae), to
judge from the form of the letters, which more nearly resemble
those of the Chalcidian inscriptions than of any other Greek
stock. So few Latin inscriptions earlier than the second Punic
War have been preserved, that it is difficult to trace each separate
stage in the process of adapting the Greek alphabet to the
exigencies of the Latin language. The symbols for the Greek
aspirate mutes, © (the ¢/-sound of our ‘ant-/eap’), ® (as in

 

1 Hiibner’s article in Miiller’s Hand- summary of what is known and a
buch d. Klass, Alterthumswissenschaft, list of the authorities.
vol. i, pp. 492 sqq. 1886, gives a

Dp
2 THE LATIN LANGUAGE. (Chap. I.

‘uphill’), & (the symbol for the k/-sound of our ‘in‘dorn’ in
the Chalcidian alphabet, while X was the symbol for the /s-sound,
Attic =), were found superfluous by the Latins, in whose lan-
guage these sounds were unknown, and were retained as symbols
for numbers merely, © for 100 (later modified to C, the initial
of centum), & (later L) for 50, ® for 1000 (later M, the initial
of mile), while the right-hand half of the symbol, viz. D, was
used for the half of 1000, 7.¢. 500, just as V, for 5, seems to have
been the upper half of X (used probably in the Etruscan adapta-
tion of the Greek alphabet for 10) (Ritschl, Opuse. iv. 704 and
722; Mommsen in Hermes xxii. 598). For the f-sound, the ~
bilabial spirant, a sound which in Quintilian’s time was quite
unknown in Greek (Quint. xii. 10. 29), the nations of Italy seem
to have taken the Greek combination of symbols FH (digamma
with aspiration), a combination found in a few of the earliest
Greek inscriptions to express a sound which seems to have been
a development of an original sw- (e.g. Fhexaddwoe, in the proper
name Hecademus, on an inscription of Tanagra (Rohl, Jnser.
Graec. 131), and which may have been at that time some adum-
bration of the fsound. This double letter FH, which we find in
a very old Latin inscription on a brooch found at Praeneste with
FHEFHAKED (=fecit) (C. I. L. xiv. 4123), in the earliest
Etruscan inscriptions, e.g. vhulyenas (the proper name Fulcinins)
(Fabr. Suppl. iii. 306), and in the inscriptions of the Veneti, an
Illyrian tribe of N.E. Italy (Pauli, Altitalische Forschungen iii.
p- 97 sqq.), was in the Etruscan alphabet reduced to a symbol
like the figure 8 (a modification of H, the F being dropt), while
in the Latin alphabet the second element of the compound was
discarded, and F alone was used. The exact course of events
which led to the use of the Greek symbol for the g-sound (in
Chalcidian inscriptions written ¢ not -), to express the Latin
é-sound as well as the Latin g-sound, and in time to the almost
total disuse of the symbol K, cannot, with the evidence at pre-
sent forthcoming, be determined (for a conjecture, see ch. ii. § 75).
On the very old Dvenos inscription, for example (Annali dell’
Inst. 1880), we find FEKED (or FEKED corrected into FECED),
(fecit), PAKARI, COSMIS (cdmis), vIRCO (virgo?) side by side.
The inconvenience of this practice led in time to the use
§ 1.] THE ALPHABET. 3

of a modified form of the symbol C to express the g-sound, the
earliest example of which is found on the as libralis of Luceria
(between 300 and 250 B.c. according to Mommsen), with Ga. /.
(Gai filius) (Rdon, Eeriture e¢ Prononciation, p. 145 sqq.). It
was received into the Roman alphabet at the time possibly of
Appius Claudius Caecus, censor 312 B.c., and took the place of Z,
the symbol apparently for soft or voiced s, a sound which had by
this time passed into the r-sound (see ch. iv. § 148). The symbols of
the Greek vowels . and v were used not only for the Latin vowels
i and w, but also for the y- and w-sounds of words like jam, vos,
a confusion frequently remarked on by the grammarians (e.g.
Quint. i. 4. 10 ‘iam’ sicut ‘etiam’ scribitur, et ‘uos’ ut ‘ tuos’),
which persisted till very late times; though on Inscriptions
from the beginning of the Empire onwards we often find a tall
form of I used for the y-sound (Christiansen, de Apicibus et
L longis, p. 29); and the Emperor Claudius tried without
success to introduce a new symbol, an inverted digamma, for
the ¢-sound.

The third guttural symbol of the Greek Alphabet, Koppa,
was retained for the g-sound of Latin, a sound at first ex-
pressed by Q, e.g. QOI (gui) on the Dvenos inscription, then
by QV.

In the second century B.c. the cultivation of literature at
Rome, in particular possibly the imitation of the quantitative
verse of Greece, led to two usages, perhaps borrowed, the one
from the Greek, the other (if not both) from the Oscan
alphabet, viz. the doubling of a consonant to express the re-
peated or lengthened sound (see ii. 127), the doubling of a vowel
(a, e, #, and o?) to express the long quantity’. The earliest
example of the former is the Decree of Aemilius Paulus, 189 B.c.
(C. [. L. ii. 5041), with posstpErE, &c., beside posEpIsENnT, &c.,
for all the older inscriptions? write the consonant single in such
cases; of the latter, the Miliarium Popillianum, 132 8.c. with
PAASTORES. Ennius is mentioned as the introducer of the double
consonant, while the practice of doubling the vowel is ascribed

 

' Tn Osean this is normally confined 2 As do the oldest Oscan inscrip-
to long vowels in the first syllable. tions and all the Umbrian inscrip-
‘But tristaamentud, ‘testamento’). tions written in the native alphabet.

B2
4 THE LATIN LANGUAGE. (Chap. I.

by the Roman tradition to the poet Accius, another of whose -
spelling reforms was the use of gg for the velar nasal followed
by g (see ch. i. § 63). The practice of doubling the consonant
remained to the latest times, in spite of a temporary resort in the
reign of Augustus to the use of the sécélicus, a sickle-shaped
mark placed above the single consonant, to express its repeated
or lengthened sound; but the double vowel was soon discarded
in favour of the apex, a mark placed above the single vowel, to
express length, originally of a shape like a sickle, or like the
figure 7, later of the form of the acute accent. The apex was
much in fashion till about 130 a.p., when it came to be used at
random over short and long vowels alike, but never attained so
universal use as the doubled consonant. Long 2 was indicated
by the tall form of I1, a form likewise employed to denote the
y-sound, and often also for initial i (Christiansen, de Apicibus et
I longis),

In the last century of the Republic, when Greek Grammar,
and even Greek Phonetics, came to be studied at Rome, the
necessity was felt for the more exact expression of the sound of
Greek loan words, which were more and more entering into
the language especially of the upper classes. For the Greek
aspirates, which had hitherto been represented by the Latin
tenues T, P, C, compound symbols TH, PH, CH were intro-
duced ; and the mispronunciation of these sounds was considered
as great a fault in polite society as the dropping of / is with us
(see ch. 11. § 60). The Greek v (earlier represented by Latin V),
which had by this time the é-sound (see ch. ii. § 14), was now ex-
pressed by the Greek letter itself in its Attic form Y, just as we use
Spanish % in loan words like ‘ cafion,’ while for Greek ¢ (formerly
denoted by s-, -ss-, e.g. setTvs, for Zethus, C. I. 1.1. 1047, patrisso,
&c., Plaut.), the old symbol Z was revived. The reforms proposed
by the Emperor Claudius, the use of the Greek symbol of the
rough breathing for the d-sound (see ch. ii. § 14), of the reversed
C for the ps-sound of seripsi, urbs, &c. (see ch. il. § 78), of the
inverted digamma for the w-sound of vos, &c., did not survive
his own reign (see Biicheler, de 7%. Clandio Caesare grammatico).

 

* Was this too borrowed from an Osean inscription earlier than
Oscan? We have fliet, ‘fient,’ on 211 B.c. (Rhein. Mus. 1888, p. 557).
§§ 2-5.] THE ALPHABET. 5

§ 2. The Alphabet of twenty-one letters. Cicero (Deor. Nat. ii. 37. 93) argues
against the Atomic Theory by showing the improbability of any chance com-
binations of the twenty-one letters of the alphabet ever producing a single
line, much less an entire poem, of Ennius: ‘hoc qui existimet fieri, non
intellego cur non idem putet, siinnumerabiles unius et viginti formae littera-
rum, vel aureae vel quaelibet, aliquo coiciantur, posse ex iis excussis annales
Enni, ut deinceps legi possint, effici ; quod nescio an ne in uno quidem versu
possit tantum valere fortuna.’ This Alphabet, A to X, is often found on
coins of the last century of the Republic (e.g. C. I. L. i. 374, « 100 B.C.) ;
and Quintilian (first cent. a.p.) speaks of x as the last letter of the alphabet
(nostrarum ultima, i.4.9). But Y and Z are added on some coins (¢.g. C. I. L.
i. 393, 454, both with YZ; 417 with Y—all belonging to the last century
of the Republic).

§ 3. The letter F. That early Greek Fh, a development of I.-Eur. sw-, had
some kind of fsound is made not improbable by the analogy of other lan-
guages. In Old Ivish, where I.-Eur. sr between two vowels became (like
sr- in Greek pedua, purds), hr, rh, (e.g. a ‘his’ prefixed to sruth, ‘stream,’
is pronounced @ rhoo), I.-Eur. sw- when preceded by a vowel became /, e. g.
a fiur, ‘his sister’ (I.-Eur. *esyo swesor), which points to a connexion between
hw (wh), and the f-sound. It must however be added that L.-Eur. w- in
Irish regularly becomes jf, e.g. faith, ‘a prophet’ (cf. Lat. vatés). A still
better analogy is furnished by the Aberdeenshire dialect of Scotch, where
the wh- or hw-sound of Scotch ‘what,’ ‘when,’ &¢., appears as f, ‘fat,’ ‘fan.’

§ 4. KX. a, the last letter of the alphabet (Quint. i. 4. 9: x nostrarum
(litterarum) ultima, qua tam carere potuimus quam psi non quaerimus),
was also written xs from early times (e.g. ExsTRAD for extra, on the S. C.
de Bacchanalibus, 186 B.c. C. I. L. i. 196), especially at the period of the
poet and grammarian, Accius (e.g. saxsvm on an epitaph of one of the
Scipios, ¢. 130 B.c., i. 34; PROXSVMEIS for proximis, EXSIGITO, LEXs on the Lex
Bantina, bet. 133 and 118 B.c., i. 197), and is common in the Augustan age
and in plebeian inscriptions of a later epoch (for examples, see Index to
C. I. L. viii. &e.; exsemplo Comm, Lud. Saec. A. 26; and for instances in
Virgil MSS., see Ribbeck, Ind. p. 445). Terentius Scaurus, second cent. a. p.,
condemns the spelling ‘nuxs,’ ‘truxs,’ ‘feroxs’ as an unnecessary repetition
of the sibilant element of the x-sound. The guttural element is repeated
in the spelling cx, e.g. vexor for uxor (a misspelling which has led to the
corruption voxor in MSS. of Plautus, Class. Rev. v. 293), viexit (C. I. L. v. 5735).
(For examples in Virgil MSS., see Ribbeck, Ind. p. 391). We also find xc,
e.g. Ivxera (C. J. L. vi. 14614), and sx, e. g. visxiT (viii. 67), all various ways
of expressing the same sound (a c-sound followed by an s-sound), for which
we also find a more accurate expression, namely cs, e.g. vicsiT (vii. 5723).
This last combination was used to express the sound in the Etruscan
alphabet, the symbol X keing retained only as a numerical symbol, for
the number tro.

§ 5. Z. If we are to believe Velius Longus (7. 51 K), this symbol was
found in the Carmen Saliare; though whether the mysterious jumble of
letters which the MSS. of Varro, L. L. vii. 26, offer as a fragment from this
hymn, cozeulodorieso, &c., can be fairly quoted as an instance of Old Latin
2 is doubtful, for the reading suggests O zew (Greek @ Zed) more than any-

z

thing else; and Varro quotes the passage as exemplifying the old use of s
6 THE LATIN LANGUAGE. (Chap. I.

(not 2) for later r. It is found on coins of Cosa [C. I. L. i. 14 cozano (after
273 B.c.), where the letter should have the ordinary z-form and not the form
printed in the Corpus (see Ritschl. Opusc. iv. 721 )). The dzenoine of the
Dvenos inscription is too doubtful to quote ; for the letters may read not only
dze noine. ‘on the ninth day,’ but also die noine, or even Dvenoi ne. This old Latin
z seems to have expressed the sound of soft or voiced s (but see ch. ii. § 121), the
sound in our verb ‘to use’; while our noun ‘use’ has the hard or unvoiced s,
Between Vowels in Latin s had once this soft sound, and was presumably
written 2; but this sound passed at an early time into the y-sound (c. 350 B.c.,
to judge from the remark of Cicero, Fam. ix. 21. 2, that L. Papirius Crassus,
dictator 415 a. U.c. (=339 B.c.), was the first of his family to change the
name from Papisius to Papirius ; in the Digests (i. 2. 2. 36) Appius Claudius
is mentioned as the author of the change: R litteram invenit ut pro Valesiis
Valerii essent, et pro Fusiis Furii). Martianus Capella tells us that the letter
was removed from the alphabet by Appius Claudius Caecus, the famous censor
of 312 B.c., adding the curious reason that in pronouncing it the teeth
assumed the appearance of the teeth of a grinning skull (Mart. Cap. iii. 261:
z vero idcireco Appius Claudius detestatur, quod dentes mortui, dum expri-
mitur, imitatur). In the Oscan language this soft s-sound was retained
without passing into r. The native Oscan alphabet (derived from the
Etruscan), expresses it by the letters, which is also used for the hard s-sound,
while the z-symbol denotes the ¢s-sound ; but in the later inscriptions, which
are written in Latin characters, 2 is used (e.g. eizazwne egmazwm (in Latin,
earum rerum), on the Bantia tablet, c. 130 B.c.). (On the question whether
the z (Latin character) of Osc. zicolo-, ‘ dieculus,’ represents the soft s-sound or
the ¢s-sound of the letter written in the Oscan alphabet like a capital I with
top and bottom strokes prolonged, and in the Umbrian alphabet with the
same strokes slanting instead of horizontal, and on the occasional use of the
native letter for the s-sound, e.g. Umbr. ze¥ef, ‘sedens,’ see von Planta, Osk.-
Umbr. Dial. p. 71.)

§ 6. The Guttural-symbols. A special symbol for the g-sound, made by
adding a small stroke to the symbol C, is said by Plutarch (Quaest. Rom. 54
and 59; cf. Ter. Scaur. 7. 15 K.) to have been the invention of Sp. Carvilius
Ruga ¢. 293 B.c., presumably because he was the first to write his name Ruga
with the new symbol, as L. Papirius Crassus, dictator 339 B.c., was the first
to conform the spelling of the family-name Papisius to the new pronunciation
Papirius. The remark, however, of Martianus Capella about the action of the
censor of 312 B.c., Appius Claudius Caecus, with regard to the letter Z, whose
position in the Latin alphabet was occupied by the new symbol G, suggests
that the differentiation of the C and G symbols was the work rather of that
many-sided reformer. The exclusive use of the symbol C for the k-sound led
to the disuse of the symbol K, which however, thanks to the conservative
instinct of the Roman nation, was still retained as abbreviation for the proper
name Kaeso, and in a few words before the vowel a, e.g. Kalendae, a common
spelling on inscriptions (see C. I. L. i., Index, p. 583), interkalaris, kaput,
kalumnia. Terentius Scaurus, second cent. a v. (p. 15 K.) tells us that the letter
K was called ka, while the name of C was ce, and that these letters themselves
had been before his time used to indicate the syllables represented by their
names, e.g. krus (for ka-rus), cra (for cera). Velius Longus, first cent, a.v.,
speaks of some sticklers for old usages in his own age, who in their corre-
§§ 6, 7.] THE ALPHABET. 7

spondence always spelt harissime with k not ¢ (p. 53 K.) (see also Quint. i. 7.10;
Prise. i. 12.5 H.; Diom. 424. 29 K. ; Cledonius 28.5 K. ; Maximus Victorinus
195. 19 K.; Probus 10. 23 K.; Serv. in Don. p. 422K ; Donatus, p. 368K.
For spellings with ka in Virgil MSS., seo Ribbeck, Index, p. 429; and for
similar spellings elsewhere, Georges, Lex. Lat. Wort. s, vv. Carthago, caput, carius,
&e., and Brambach, Lat. Orth. p.208.) The symbol C was similarly retained in
its old use for the g-sound in the abbreviations of proper names, C. for Gaius,
Cn. for Gnaeus ; just as an old five-stroked form of the symbol M seems to be
the original of the abbreviation for the name Manius, later written M with
apostrophe. That it persisted in other words also to the beginning of the
literary period, we see from the fact that a large number of archaic words,
quoted by the grammarians from the early literature, are spelt with c not g,
e.g. acetare for agitare (Paul. Fest. 17. 30 Th.). The proper spelling of these
obsolete words was occasionally a subject of discussion, e. g. whether pacvnr
in the XII Tables, nr 1ra PAcvnt, stood for pagunt (cf. pepigi, pango), or for pacunt
(ef. paciscor’, (Quint. i. 6. ro-11; Ter. Scaur. 7.15 K.; ef. Fest. 330 23 Th.);
and probably the perayapaxrypicpds of early C to cand g was almost as fruitful
a source of error as that of E to ¢, 7, «t, of O to o, w, ov in the Homeric text.
Thus frico, not frigo, may be the proper form of the Old Latin verb, used by
Accius frigit saetas (of a boar) Trag. 443 R., &c. (cf. Greek ppicow for ppix-yw) ;
décére (cf. Béopa, mpoodoxdw) of Old Latin degere, ‘expectare’ (Paul. Fest. 51.
32Th.). (On the use of C for the g-sound see also Mar. Victorin. p. 12K.
who quotes Cabino, lece, acna; Fest. 242 and 284 Th., &c.: C is invariably used
for g on the Columna Rostrata (C. J. L. i. 195), an Imperial restoration which
probably followed with some fidelity the spelling of the old inscription.)
The letter Q often takes before wu the place of classical Latin c, especially in
inscriptions of the time of the Gracchi, e.g. PEQVNIA, OQVPARE, QVRA (for a list
of the instances, see Bersu, Die Gutturalen, p. 49) ; though whether Ritschl
(Opuse. iv. 492 , 687), is right in his suggestion that one of the grammatical
reforms of the poet Accius may have been the restriction of k to the c-sound
before a, and of ¢ to the c-sound before u, is quite uncertain. (For Accius’
use of gg for ng in aggulus, &c., ge for nc in agceps, &c., in imitation of the
Greek use of y for the nasal guttural, see below). Marius Victorinus says
(12. 19 K.) : Q et fuisse apud Graecos, et quare desiderat fungi vice litterae,
cognoscere potestis, si pontificum libros legeritis.

§ 7. Y- and W-Sounds :—j and v were not distinguished in Latin MSS. nor
indeed in the earlier printed editions. In Italian some writers keep up the old
Latin habit of using i for j, e.g. Gennaio for Gennajo (Lat. Janudrius) ; others
use j for -i/, ¢. g. vizj,‘vices.’ Even now we generally print the texts of the older
Latin writers, Plautus, Terence, &c., with 7, u, not j, t, partly to give their
language an archaic appearance, but mainly because a large number of words
which in the Classical period, or the Empire, had the y- and w- sounds, had
in earlier times the sound of the vowels (sometimes of the half-vowels) ;
larua, for example, is a trisyllable in Plautus, never a dissyllable. The
minuscule forms v and u are developments of the V, of Capital, and the U
of Uncial writing. The use of the tall I-form on Inscriptions for the y-sound
has already been mentioned, as well as its use for initial 7, and for long 7.
‘How far the I-symbol (in ordinary form or tall form) might be employed for
-yi-, or V for -wu , -wav- is very doubtful. Sittl, in Burs. Jahresber, 1891, p. 250,
quotes abicere for abyic-(?), vEsvivs for Vesuv- (?): cf. Brambach, Orth. p. 94.
8 THE LATIN LANGUAGE. {Chap. I.

On the Monumentum Ancyranum we have Iveytvtis (3. 5 M.), and in
Virgil MSS. iuenis, fluius, exuiae, &e. (Ribbeck, Ind. p. 448). Equally doubtful
is the occasional usage in the earlier history of the Latin alphabet of the
Greek digamma-symbol (whether in the F-form or in the Etruscan form, viz. an
E wanting the middle horizontal line) for the w-sound. Cornutus (ap.
Cassiodor. 148. 8 K.: itaque in prima syllaba digamma et vocalem oportuit
poni, ‘Fotum,’ ‘ Firgo,’ quod et Aeoles fecerunt et antiqui nostri, sicut serip-
tura in quibusdam libellis declarat) implies merely that some of his gram-
matical predecessors made a hobby of writing F for v (ef. Prise. i. 35.17 H.).
The second symbol in the phrase dxenoine on the Dvenos tablet may be
a variety of this symbol in the later form Dvenoi ne, but it may also be 2,
dze noine or (most likely) a form of i. (See above.)

Cicero wrote ii to express the sound of the second element of an 7-diphthong
before a vowel (see ch. ii. § 55), e.g. aito, Muiia, Atiax (Quint. i. 4. 11; Vel.
Long. 7. 54 K.: et in plerisque Cicero videtur auditu emensus scriptionem,
qui et ‘Aiiacem’ et ‘Maiiam’ per duo i scribenda existimavit. He mentions
also Trotia, and with three i’s, coiiicit, Cf. Prise. i. 303 and i. 14 H., who
ascribes the spelling Pompeiii to Julius Caesar).

On inscriptions we find Eli1vs and Erlvs (see Weissbrodt in Philologus, xliii.
pp. 444 sqq.), and in MSS. like the Ambrosian Palimpsest of Plautus, eitus,
aiiunt, &e. (for examples in MSS. of Plautus and Virgil, see Studemund’s
Apograph, Ind. p. 509; Ribbeck, Prol. p. 138).

In the Umbro-Oscan alphabets, which are derived from the Etruscan, the
w-sound is expressed by the digamma, in the form of a capital E wanting
the middle horizontal stroke, while V expresses both the w- and the o-vowels
(the Oscan alphabet came in time to discriminate the o sound by inserting
a dot between the two arms of V). On the question whether Ose. ii and i
correspond to I.- Eur. ty and y in words like Ose. heriiad. and heriam, see
ch. iv. § 63.

§ 8. Double Consonant. Festus in his discussion of the word solitaurilia
(p. 412 Th.), which he derives from taurus, in the sense of xoxévy, and the
Osean sollo- (in Latin totus), declares the doubling of the consonant to have
been a practice introduced by the poet Ennius (239-169 B.c.) into Latin
orthography in imitation of the Greek usage ‘per unum 1 enuntiari non est
mirum, quia nulla tune geminabatur littera in scribendo. quam consuetu-
dinem Ennius mutavisse fertur, utpote Graecus Graeco more usus). The
Roman tradition, which ascribes this spelling reform to Ennius, as well as
the doubling of the long vowel to Accius, is supported by the dates at which
these spellings are first found on inscriptions (double consonant 189 B.c.,
double vowel 132 B.c.); though it is quite possible that Ennius followed, not
the Greeks, but the Oscans, who used double consonants much earlier than
the Romans, and to whose nationality he belonged quite as much as to the
Greek. We do indeed find a double consonant before 189 B.c. in the spelling
HinnaD (the town of Enna in Sicily), 211 B.c. (C. I. L. i. 530), which is a mere
reproduction of the Greek spelling found on coins, e.g. HENNAIon (Head,
Historia Numorum, p. 119); but even after 189 B.c. the double consonant-
sound is often written with the single letter till the time of the Gracchi,
when the double letter became the established spelling (see Ritschl, Opuse.
iv. 165 sqq.).

The sicilicus is only found on a few inscriptions of Augustus’ time : Mumiaes
§§ 8, 9.] THE -ALPHABET. 9

Sabelio C. I. L. v. 1361. osa x. 3743. Marius Victorinus, fourth cent. a.v., states
that it was often to be seen in old MSS. (sicut apparet in multis adhuc
veteribus ita scriptis libris (p.8 K. Of. Isidor. Orig. i. 26. 29).

§ 9. Signs for long vowels. No instance of 00 for 6 is found on the extant
Latin inscriptions, though we have uwootum on an inscription in the Faliscan
dialect, whose orthography was very like the Latin : pretod de zenatuo sententiad
uootum dedet (in Latin, ‘praetor de senatus sententia votum dedit’), (Zvetaieff,
Inscr. Ital. Inf. qo). For t Accius wrote ei (Mar..Victorinus 8. 14 K.), either
because the diphthong ei had by this time become identical with the é-sound,
or in imitation of the Greek orthography (§ r2) ; for Greek e: had taken the
same course as Latin ei, and expressed the same sound as original long # (Blass.
Griech. Aussprache’, p. 51). Lucilius prescribed rules for the use of ei and
‘i longa’; but instead of keeping ei for the original diphthong, and the single
letter for the original long vowel, he used foolish distinctions ', if we are to
believe Velius Longus (56. 7 K.) such as that the double symbol was suitable
for a plural, e.g. puerei Nom. Pl., the single symbol for a singular, e. g. pucri
Gen. Sg. (alii vero, quorum est item Lucilius, varie scriptitaverunt, siquidem
in iis, quae producerentur, alia per i longam, alia per e et i notaverunt, velut
differentia quadam separantes, ut cum diceremus ‘ viri,’ si essent plures, per
e et i scriberemus, si vero esset unius viri, per i notaremus, et Lucilius
in nono :—

‘iam puerei uenere ;’ e postremo facito atque i,
ut puerei plures fiant. i si facis solum,
‘pupilli,’ ‘ pueri,’ ‘ Lucili,’ hoc unius fiet ;

item
‘hoe illi factum est uni;’ tenue hoe facies i:
‘haec illei fecere ;’ adde e ut pinguius fiat.)

The same absurd Peason seems to be assigned for the differentiation of
meille, meillia and miles, militia; of pilum, a mortar (Sing.) and peila, javelins
(Plur.) in another fragment of Lucilius (9. 21-24 M.).

‘meille hominum,’ ‘duo meillia;’ item huc E utroque opus ; ‘ miles,’

‘militiam’; tenues i, ‘pilam,’ qua ludimus, ‘ pilum,’

quo pisunt, tenues. si plura haec feceris pila,
quae iacimus, addes e, ‘peila,’ ut plenius fiat.
Another fragment (or rather two fragments), of more doubtful reading, seems
to prescribe single 7 in the Gen. Sg. of 10-stems, but ei in the Voe. Sg. (9.
17-20 M.) :—
(x) porro hoe ‘ filius Luci ;’
feceris i solum, ut ‘Corneli,’ ‘Cornificique,’
(2) ‘mendaci’ ‘Furique.’ addes e cum dare, ‘Furei,’
iusseris
(unless we read ‘date, Furei,’ and make the et-form Voc. Plur.).

Whether the persistent use of -i in the Gen. Sg. of O-stems on inscriptions
is due to the rule which Lucilius supports, or whether it is to be otherwise
explained, is hard to say (see ch. vi. § 20). Varro, while disapproving of
Lucilius’ arguments, seems to have followed his practice, for Ter. Scaurus (p.

 

‘ Orshould wecallthem mnemonic, rules on the memory of the common
as opposed to scientific, distinctions, -people for whom Lucilius wrote his
meant to impress the orthographic book? (see Lucil. 26. 1 M.).
Io THE LATIN LANGUAGE. (Chap. I.

19 K.), after quoting the passage from Lucilius beginning ‘meille hominum,’
goes on to say : quam inconstantiam Varro arguens in eundem errorem diversa
via delabitur, dicens in plurali quidem numero debere litterae ie (om. MSS.)
praeponi, in singulari vero minime. But in general the spelling ei on
Inscriptions seems to occur for any 7-sound (see the Index to C. I. L. i, and
ef. below, ch. iv. § 34). From the time of Sulla the symbol in use is the tall
I (Christiansen, p. 28), though EI shows itself even later, while from
v. 130 A.D. the tall Iis used at random for the short and long vowel alike
(Christiansen, p. 29). This tall I may be indicated by Lucilius’ phrase ‘i
longa,’ and even by Plautus’ allusion to the ‘littera longa’ in Aut. 77, where
the miser’s old serving-woman in a fit of despondency thinks of hanging

herself :—
neque quicquam meliust mihi,

Ut opinor, quam ex me ut unam faciam litteram Longam.
(Cf. Ausonius ‘iota longum,’ of a hanging body, Epigr. exxviii. 11.) But the
absence of the long form from the Inscriptions till Sulla’s time makes this
doubtful, especially in the case of Plautus. The remark in the Rudms
(v. 1305) that mendicus has ‘one letter more’ than médicus shows that the
long ¢ of the first word was not expressed by ei.
The reason which induced Accius to use EI, and not II, for the long i-sound
was probably the fear of confusion with a common symbol for E, viz. II, in
which a long vertical stroke is substituted for the three horizontal strokes.
There was a similar symbol for F, viz. I, with a short vertical stroke ; both
these by-forms of F and E being probably more used in writing than on
inscriptions, though they are common enough in plebeian inscriptions of
later times, along with a by-form of M with four horizontal strokes | [I.
(See Hiibner, Exempla Scripturae Lat. Epigr.).
In Greek inscriptions a double vowel is found perhaps only in the name
Marcus and its cognates. In the second century B.c. the spelling MaapredAos,
Maapicios, Maapnos is the rule, and it is common till 50 8.c. But the aa is not
found in derivatives where the Greek accent does not fall on this vowel, e.g.
Mapxiavos, MapxedAAevos, &c. (Eckinger, p. 8).
In the first century a.v. the use of wu for @ seems to have been affected for
a time, for the spelling nuuilli occurs on wax tablets found at Pompeii
(Notizie degli Scavi, October, 1887), and uu is often found for a of fourth deel.
nouns in Virgil MSS. (see Ribbeck, Ind. p. 449), €.g. metuus, curruus ; also suus
for sus (cf. Probi Appendix, p. 202. 27 K.). In the Bamberg MS. of the elder
Pliny wus is the regular spelling in the Gen. Sg. and (Nom. and) Ace. Pl. of
fourth decl. nouns (see preface to Sillig’s edition), so that this must have been
Pliny’s own practice (Probus, Inst. Art. 116. 33 K., refers to this spelling).
Lucilius seems to have objected to Accius’ rule of doubling the vowels, at
least in the ease of A, which, he points out, has the same quality when short
and when long (see ch. ii. § 1); hence é and a, he argues, should be written
in the same way, like Greek d and a (9. 4-7 M.) :—
a primum longa, et breuis syllaba. nos tamen unum
hoe faciemus, et uno eodemque ut dicimus pacto
seribemus ‘pacem,’ ‘placide,’ ‘Ianum,’ ‘aridum,’ ‘acetum,’
*Apes, “Apes Graeci ut faciunt.

(On rehemens for rémens, see ch. ii. § 56.)

§ 10. gg for ng. The guttural nasal of English ‘sing’ (ch. ii. § 61) was ex-
§§ 10, 11.] THE ALPHABET. II

pressed before a Guttural by y in Greek, e g. dyyeAos, dyed, and was called by
Greek grammarians the ‘Agma.’ Accius proposed to follow the example of
the Greeks, and express this sound in Latin by g instead of n, e.g. ‘aggulus’
for angiilus, ‘aggens’ for angens, ‘iggerunt’ for ingérunt, ‘agceps’ for anceps.
(Varro ap. Prise. i. p. 30 H.: ut Ion seribit, quinta vicesima est litera, quam
vocant agma, cuius forma nulla est, et vox communis est Graecis et Latinis,
ut his verbis: ‘aggulus,’ ‘aggens,’ ‘agguila,’ ‘iggerunt.’ in eiusmodi Graeci
et Accius noster bina g scribunt, alii u et g, quod in hoc veritatem videre
facile non est. Similiter ‘agceps,’ ‘ageora.’) The Inscriptions offer no
example of this spelling (cf. Eph. Epigr. vii. 928) ; but a trace of its existence
is perhaps found in the spelling ‘ager’ for agger, which the MSS, offer with
singular persistence for a line of Lucilius (26. 81 M.; cf.11.5M.). If Lucilius
and his contemporaries used gg for ng, they would be forced to use the single
letter in words like agger, aggero, &e.

§ 11. New Letters for Greek Sounds: Y, Z,CH, PH, TH, RH. Our
name for y, viz. ‘wy,’ comes from the Latin name for the letter which was
ui’ (Mém. Soc. Ling. vi. 79). Greek v is often represented by Latin wi, and
vice versa, e. g. quinici for xuvicot, and ’AxvaAas for Aquila(ibid. viii. 188 ; Eckinger,
p.123). Before the introduction of the Greek letter, Latin u was used in loan-
words like tumba, &c., while at a later time i was employed, e.g. cignus; and
the Romance forms of these earlier and later-loan words indicate that these
spellings represented the pronunciation of the time (see ch. ii. § 28). Y was
not allowed in native Roman words (Caper vii. 105. 17 K.), though it sometimes
gained a footing through a mistaken idea that a word was borrowed from the
Greek, e.g. sylra supposed to be the Greek #An, lympha identified with Greek
vipgn, &e. (see ch. ii. £ 28). Greek ¢, if we are to believe the grammarians, was
expressed in earlier times by d also (Prise. i. p. 36: y et z in Graecis tantum-
modo ponuntur dictionibus, quamvis in multis veteres haec quoque mutasse
inveniantur, et pro v u, pro ¢ vero...s vel ss vel d posuisse, ut... ‘ Sagun-
tum,’ ‘massa’ pro ZdnvrOos, pata, .. . ‘Sethus’ pro 200s dicentes, et ‘ Meden-
tius’ pro Mezentius) (see ch. ii. § 120).

The earlier expression of Greek 0, ¢, x by t, p, ¢ (e.g. adelpus, Metradati on an
inser. of 81 B c. (?), Not. Scav. 1887, p. 110) remains in words like tus, Greek
Ovos, Poeni for Poivines, culr, Greek xaArg (cf. Quint. i. 5. 20 diu deinde serva-
tum ne consonantibus (veteres) adspirarent, ut in ‘triumpis’). We find b
for ¢ in Old Latin Bruges for &pvyes, and in ballaena for paddAava, the former of
which was used by Ennius, and was still to be found in copies of his poems in
Cicero’s time (Cie. Orat. xlviii. 160 Ennius ... ‘ui patefecerunt Bruges,’ non
Phryges, ipsius antiqui declarant libri), while the latter remained in current
use. (F was not regularly used for ¢ till the middle of the fourth century
A.D. (Hermes xiv. p. 70), though it is often found on plebeian inscriptions from
Severus’ time, and even on Pompeian graffiti we have, e.g. Dafne, (. I. L.
vi. 680). But as early as 146 B.c. we find th, ph, ch in the dedicatory inscrip-
tions! of the Graecizing L. Mummius (¢. I. L. i. 546 corintuo (?), 146 B.C. ;
i. 541 in Saturnians :

ACHAIA CAPTA CORINTO DELE10
ROMAM REDIEIT TRIVMPHANS).

The importance attached in polite society at Rome to the correct pronunciation

 

! They may be later restorations.
12 THE LATIN LANGUAGE. ([Chap. I. §§ 12, 13.

of these aspirated consonants in Greek loan-words led to their wrong use in
native Latin words (ch. ii. § 60), e.g. pulcher, referred to Greek moddypous,
a spelling found as early as 1048 c. ona denarius of Claudius Pulcher (¢. I. L.
i. 380), much in the same way as ‘antem’ (0. Engl. antefn from Gk dvripary
through Low Lat.) has come to be written with th, ‘anthem.’ For Greek
initial 6, and for -5f-, the older spelling was 17, rv, e.g. Regiwn, Burrus (the
invariable form of the name Ilvééos in Ennius, according to Cic. Orat. xlviii.
160°. The use of rh for initial 6- was not approved by Varro, who preferred to
write ‘Rodus,’ ‘retor’ (Varro, LZ L. iii. fr. 57 p. 182 Wilm.).

In Oscan inscriptions similarly Greek aspirates are usually expressed by
tenues, e.g. Arkiia (for "Apxias), Meeilikiieis (for MeaA:xéou Gen.), and so
Pelignian Perseponas, ‘ Proserpinae,’ Gen., but we have also Osc. thesavrei, ‘in
thesauro,’ Loc., &e.

§ 12. Influence of Greek Orthography.—The use of g for the guttural
nasal, advocated without success by Accius (see above), was clearly borrowed
from the Greeks. The spelling ei for the long i-sound, and the employment
of double consonants, may possibly, as we have seen, have come from the
same source. But however natural it may appear for the Romans to have
adopted Greek spelling along with Greek terminology in matters of Grammar
and Phonetics, there is hardly a single instance of the practice that can be
established by proof (see Zarncke’s attempt in Comm. Ribbeck, 1888).

§ 13. Syllabic Writing.—The remark of Ter. Scaurus (p. 15 K.) quoted
above, that k had been employed to denote the syllable ka, c the syllable ce,
suggests (unless indeed he is merely alluding to the common practice of
abbreviating words by writing only the initial letter of each syllable), that
spellings on early inscriptions like LVBS for lub2.n)s on a Marso-Latin inscrip-
tion (C. I. L. i. 183), may be not really evidences of syncopated pronunciation,
but rather traces of an old custom of syllabic writing (see ch. iii. § 14). The
syllabaries found on Etruscan inscriptions (e.g. Fabretti 2403 and 450), as well
as the use of a dot (like the Sanscrit virama), to indicate those consonants
which are not followed by a vowel, in the inscriptions of the Veneti, an
Illyrian tribe of N.E, Italy, are perhaps other indications that syllabic writing
prevailed at an early period in the Italian peninsula. .
CHAPTER IL.
PRONUNCIATION},

$1. A. Iv the words ‘man,’ ‘father,’ the vowels which we are
in the habit of classing roughly as ‘ short a’ and ‘long a,’ are
really very different from each other, and would be phonetically
expressed by two distinct symbols. In Sweet’s Handbook of
Phonetics, while the second is written a, the first is denoted by
a combination of the letters @ and ¢, viz. 2, a symbol which
implies that the vowel has something of the nature of an E-sound.
If we compare our pronunciation of the words ‘ man,’ ‘ hat,’ with
the German of ‘ Mann,’ ‘er hat,’ we see that the German vowel
is the same as the a of English ‘father’ or German ‘ Vater,’
while we might say that our ‘ man,’ ‘hat,’ ‘bat,? have in them
something of the sound of ‘men,’ ‘bet.’ Seelmann, who classifies
the varieties of A as ‘normal a,’ ‘a inclined to an E-sound, and
‘a inclined to an O-sound’ (this last being something not quite
so definitely an O-sound as the vowel of our words ‘all,’ ‘ awe’),
is of opinion that the Latin a had a leaning to e rather than to 0,
and goes so far as to give to Latin « of the Imperial age the
@-sound of English ‘man.’ This however is not the sound of
modern Italian a, e.g. padre, which Sweet now judges to be
identical in quality with the a of English ‘father,’ though, owing

 

1 Seelmann, Aussprache des Latein, Heilbronn, 1885, is the chief book on
Latin Pronunciation.
I4 THE LATIN LANGUAGE. [Chap. IT.

to our smaller use of lip-action in utterance, the vowel has with
us what he terms a more ‘ muffled’ sound. And the evidence at
our disposal is not at all strong enough to allow us to determine
with precision under which class of A-sounds Latin a should be
placed, nor yet how far its quality was altered by the consonants
which accompanied it, nor even whether it had to some extent
a different quality as a long and as a short vowel. On this last
point indeed we have some evidence of weight. We can be sure
that Latin @ and 4, if they differed at all in quality, did not
differ so markedly as Latin @ and é, 7 and @, % and # For
Lucilius (ix. fr. 4 M), in criticizing the proposal of the poet and
grammarian Accius to write a single vowel for a short, a double
for a long vowel (thus a for @, aa for a), says that the vowel a
has the same sound in pronunciation when long as when short,
and should be written in the same way, e.g. pacem, plicidle, &e.,
just as the Greeks write d and a in the same way, e.g. "Apes
and “Apes (the passage is quoted on p. 10).

And his remark is borne out by the evidence of the Romance
languages. In them there are no means of tracing the quantity
of a Latin vowel, unless the long and the short vowel differed in
quality as well as in quantity. This difference did exist in the
case of other vowels, e.g. % and 7; and so in the Romance lan-
guages Latin 7 appears as close ¢, Latin 7 as close z (e.g. Ital.
misi for Lat. mist; Ital. beve for Lat. d%i/), Latin & and a,
however, show no divergence in any Romance language ; and,
when we are in doubt whether a Latin a was long or short, in
a syllable long by position for example, we have to refer to some
other family of languages, which happens to have borrowed the
word at an early period from the Latin. A word like saceus is
shown by its Welsh and Breton forms, sach, not to speak of
Gothic sakkus, O. H. G. sac, O. Engl. seecc, to have had a short
a; but this could not have been told from its Romancé forms,
Ital. sacco, Span. saco. Long « appears in a different guise in
Welsh and Breton (e.g. poc, a kiss, representing Latin pdacem in
the formula of the priest at absolution, pacem do tibi), but not in
Romance, e.g. Ital. pace, Span. paz.

The accounts of the pronunciation of a, given by the Latin
writers on phonetics, do not much help us to determine the shade
§1.] PRONUNCIATION. VOWELS. 15

or shades of the A-sound, which the Latin vowel expressed, nor
are any of their descriptions free from the suspicion of Greek
bias. The evidence to be drawn from the phenomena of the
language itself is equally indecisive. It is true that a becomes ¢
in the unaccented syllable (long by position), as aurifex from
aurum and facis, and in Early Latin in open syllables too, e.g.
abégit, classical abigit, from dé) and dgo, an e retained in classical
Latin before 7, e.g. zmpéro from piéro. But this was the fate of
every short vowel in the unaccented syllable, and not of a alone,
so that ¢ was the natural sound which any short Latin post-tonic
vowel tended to assume, unless attracted by a following Labial
to an O-, U- or U-sound, e.g. occiiyo from ob and cipio, testitms-
nium (C. I. L. i. 197, 3) from stem festi- (see iii. 18). Varieties in
the spelling of foreign names like Sardtca and Serdica, Delmitia
and Dalmétia prove nothing for Latin a2. More important is the
fact that ja-, 7a7- seem to have tended to the pronunciation Jé-,
jej-, with open e. Thus Janudrius became Jénudrius 5 jajunus is
the Plautine form of the classical yeydnus. Here the change of
a to e was due to the influence of the palatal y (our vy) preceding,
just as the # of janipérus was changed to by the same palatal
in Vulgar Latin jinzpirus (Probi Appendix, 199. 8 K.) (Ital.
ginepro, Fr. geniévre, Span. enebro). No such influence is at
work in the mispronunciation stetim for stdtim, a Roman cockney-
ism like London ‘keb’ for ‘cab,’ mentioned by a grammarian of
the fifth (?) century a.p. (Consentius, p. 392, 16 K.: per immuta-
tionem fiunt barbarismi sic: litterae, ut siquis dicat ‘ bobis’ pro
vobis, ‘ peres” pro pedes, ‘stetim’ pro statim, quod vitium plebem
Romanam quadam deliciosa novitatis affectione corrumpit), This
is quoted by Seelmann as a strong argument for his assertion
that-Latin a had in Imperial times the sound of English a in
‘man’; though on the other hand we might argue for an
A-sound more inclining to o from Vulg. Lat. *d¢are, a by-form
of niétare, to swim, which ousted the a-form in Vulgar Latin
about 100 B.c. In Plautus’ time and later vdécare was pronounced
like vocare; the o-sound apparently having been produced by the
influence of the labial v (our w) (cf. Kwdparos, Kodparos for
Quadratus on Gk. inscriptions), as e was by the palatal / (our y)
in Jénudrius. In one of Phaedrus’ fables (App. 21) a man
16 THE LATIN LANGUAGE. (Chap. II.

mistakes the caw of a crow for ave! ave! It is worth
mentioning that Oscan 6, the representative of Ind.-Eur. O, 0,
and, when at the end of a word, of Ind.-Eur. A, which must
have had a sound something like our a in ‘all,’ ‘awe,’ and which
is in those inscriptions which are written in Latin characters
expressed by o (e.g. ¢ovto, ‘state,’ ‘community, Nom. Sing. of
A-stem ; petiro-pert, ‘four times, Acc. Pl. Neut.), rarely by u
(e.g. petiru-pert), as in Greek characters by o (e.g. rwfro), is yet
written by Festus and by Paulus, his epitomator, with a.
Festus, when he mentions the Oscan word for ‘four, writes it
petora (p. 250, 1. 33 Th.) ; and Paulus gives wea, not vero, as the
word for ‘cart’ (p. 560, 1. 17 Th.); though Lucilius, if Festus
(p. 426, l. 7 Th.) quotes him accurately, makes ¢ the Latin equi-
valent of the Oscan Neut. Pl. suffix in so/dé (Lat. tota),

uasa quoque omnino dirimit non sollo dupundi.

The evidence then of the Latin language itself points to
Latin ¢ having had a sound which was liable to influence in the
direction of 0 as well as of e. Into the modifications of Latin
a in each several Romance language, it is hardly necessary to
enter, for they are as likely to be due to the vocal peculiarities
of the nations conquered by the Romans, as to the nuances of
sound in the language of the conquering race. French is the
language where Latin a has been most widely replaced by ¢
(e.g. chef, Lat. capt, while in champ, Lat. campus, though e is
not written, the preceding guttural has been palatalized) ; and in
Portuguese it is something between the a of ‘father’ and the
a of ‘man,’ though before 7 the sound is more guttural. But in
Italy a has what may be called the normal A-sound, that of
English a in ‘father, not that of a in ‘man.’ It is only in two
districts, Emilia (i.e. the Po-valley), and the coast of Apulia,
that it has an E-sound, while in some parts of Italy it tends to
an O-sound (Meyer-Liibke, Jal. Gram. §§ 18-21). Speaking
generally, we may say that the influence of a palatal or often
changes a into an E-sound in the Romance languages (e.g.
Corsican berba), whereas an O-sound is produced under the
influence of such letters as /, v, 4 (e.g. oltro for Latin alter in
some dialects of N. Italy), while before ~ Latin a is in some
§§ 2-4.| PRONUNCIATION. VOWELS. 7

places changed to e, in others to 0. So that the evidence, taken
as a whole, is rather more in favour of attributing to Latin
a a sound which varied to some extent in character, according
to the consonant which accompanied it, than of giving it
definitely the E-character of our a in ‘man.’ And in the
absence of more definite proof, it will be best, for practical
purposes, to use in reading Latin the sound which the vowel bears
in the language of the direct descendants of the Roman people,
the normal A-sound of Italian padre.

§ 2. Descriptions of the A-sound by Latin phoneticians. The formation
of the (Greek or Latin?) sound is described very cleverly by Terentianus
Maurus (second cent. a.p.) (p. 328 of Keil’s edition), in spite of the limitations
of the difficult Sotadean metre (--uu|--uvuj-v-v] -¥%):—

a prima locum littera sic ab ore sumit:
immunia rictu patulo tenere labra,
linguamque necesse est ita pendulam reduci,
ut nisus in illam valeat subire vocis,

nec partibus ullis aliquos ferire dentes.

Marius Victorinus (fourth cent. a.p.) (p. 32 of Keil’s edition) compresses
the same description, in his usual way: a littera rictu patulo suspensa,
neque impressa dentibus lingua enuntiatur. It is still further compressed by
Martianus Capella (fourth or fifth cent. a.p.) (iii, 261, p. 63 of Eyssenhardt’s
edition in the Teubner series): A sub hiatu oris congruo solo spiritu memo-
ramus.

§ 3. Interchange of a and e. Delmétia and Dalmatia: Vel. Longus, p. 73 K.
placet etiam Delmatiam quoque, non ‘ Dalmatiam’ pronuntiemus, quoniam
a Delmino maxima ejusdem provinciae civitate tractum nomen existimatur.
On Inscriptions, we have sometimes a, e.g. Dalmat. (C. I. L. vi. 1607), some-
times e, e. g. Delmatia (C.I.L. iii. p. 280) (see Georges. Lex. Lat. Wortf.s.v.). Jé-,
jej- for ja-, jaj-:—The Vulg. Lat. name of the month was Jenuarius (C. I. L. vi.
1708, of 311-314 A.D., and other inscriptions) [see Schuchardt, Vok.i. 185. So
in Greek inserr. ‘Ievovapiwy C. I. G. 9486 (Catana); “Ievapiov I. I. S. 62
(Syracuse)], which has developed into the Italian Gennajo (cf. Span. Enero),
with open e. Jejianus (with @ according to Ter. Maur. 343 K.), jejentaculum
supplanted the older jajinus, jajentaculum, the Plautine forms (A. L.L. 7. 528).
Jenua, for janua, is indicated by Sardinian enna, genna, and is sometimes
found in MSS. (see Schuchardt, Yok. i. p. 185). Jajunus reappears in late
Latin (in the Itala, e.g. Luc. iii. 20), and in Span. ayunar, while the shortened
forms jantdciilum, jantére are found in MSS. along with the usual jentacilum,
jentire (see Georges, Lex. Wortf. s. vv.) ; and jantare is indicated by Old Span.
yantar, Port. yantar. On Vulg. Lat. jecto for jacto (Ital. gettare, Fr. jeter), see
I, F. ii. Anz. p. 35; and for other examples of a-e, Georges, Lex. Wort/. s. vv.
Sarapis, serracum, Serdica, metaxa, Sabadius, and Dict. s. v. Serranus.

§4. Interchange of a and 0. Ndtdre is the form reflected in the languages of
those countries which were earliest colonized (Sard. nadare, Span. nadar, Port.

a
18 THE LATIN LANGUAGE. [Chap. II.

nadar, Prov. nadar), while *nétdre appears in O. Fr. noer, Raet. nudar, Roum,
innotd, Ital. nuotare, which shows that *notare did not find its way into
Vulgar Latin till about 100 B.c. *Vécittus is the Vulg. Lat. original of Ital.
voto, O. Fr. voit, ‘empty,’ and vicare, vic(u)us (see Georges, Lex. Wortf.) of
Sard. bogare, Span. hueco. Vocatio for vdcdtio is found on the Lex Repetun-
darum of 123-2 B.c. (C. I. L. i. 198. 77: militiaeque eis uocatio esto), and is
the spelling of Julius Caesar in his Lex Municipalis of 45 B.c. (i. 206. 93 and
103: vocatio rei militaris). Plautus puns on vicare, ‘to be empty,’ and vicare
‘to call,’ in Cas. 527: Fac habeant linguam tuae aedes. Quid ita? Quom
ueniam, uocent. Marmor (Greek pappapos) follows the analogy of Nouns in -or.
(For other examples of a-o, see Schuchardt, Vok. i. p. 177 sqq., and Brambach,
Hiilfsbiichlein s. v. Tamyris, and ef. below, ch. iv. § 55.)

§ 5. Anomalies in Romance. Accented Latin a sometimes shows 0, some-
times e in Romance from a variety of causes. Thus Ital. chiovo, from Lat.
clavus, shows o by influence of v; Ital. (dialectal) opre for apre, Fr. ouvrir,
hardly point to Vulg. Lat. *operio for dpério, but are rather influenced by
co(o)\pério, Fr. couvrir, Ital. coprire ; Ital. monco from Lat. mancus is due to
the synonym tronco from Lat. truncus. Vulg. Lat. *grévis (Ital. greve and
grave) may have adapted itself to lévis ; méhum, not mdlwm, the original of the
Romance words for apple (Ital. melo, &c.), seems to be the Greek form prov,
and is indicated by the pun in Petronius, chap. lvi. (p.37. 19 Biich.) con-
tumelia... contus cum malo (Icey. melo). The appearance of e for ain the
unaceented syllable, e.g. Vulg. Lat. alecer (with stem alecro-) for dildcri- (Ital.
allegro, Span. alegre), ceresio- (ceresium and eerasium in Mare. Emp.) from
*ceresus for ctrdsus (Greek xepagdés) in the Romance words for cherry (Ital.
ciriegio, &c.) is due to the same law which produced consecro from saero (ch.
iii). The mispronunciations /etiyo Prob. 212. 4, secratum Mar. Vict. x. 6, are
to be similarly explained. (On the variation of Italian é with I.-Eur. é in
words like Lat. pciteo (Osc. pate-) besides Gk. merdrvuy, see ch. iv. § 61.)

$6. E. The evidence for the pronunciation of Latin e is much
stronger than the evidence at our disposal for Latin a. In the
Romance languages we have clear proof that short and long ¢ had
in the parent-speech a different quality, @ being an open E-sound
like Engl. ‘men,’ é@ a close E-sound like Fr. été. (Our ‘fail,’
‘fate’ have a diphthong of this close e combined with an I-sound.)
These sounds are retained without a change in Italian at the
present day in such words as bello (Lat. Jé/Zvs) with open e, stella
(Lat. s¢é//a) with close e¢, though in open syllables in many
Romance languages open ¢ has developed to ‘e (Ital. criepa, Span.
crieba, from Lat. erépat) (cf. Schuchardt, Voé. ii. p. 328). All
this harmonizes so wonderfully with the evidence we can draw
from the Latin language itself, and from the statements of the
Latin grammarians, as to leave little room for doubt. Accord-
ing to the grammarians long ¢ is ‘an E inclining to an I-sound,
§§ 5, 6.] PRONUNCIATION. VOWELS. 19

precisely what phoneticians nowadays call ‘close e’ (open e would
be ‘an E inclining to an A-sound’), while short ¢ approaches the
sound of the Latin diphthong @, which in the Romance lan-
guages is undistinguishable from Latin 2 (e. g. Ital. cielo, Span.
cielo, from Lat. caelum, exactly as Ital. criepa, Span. crieba, from
Lat. crépat). Latin é and % are merged in the same way in
Romance (Ital. fendo from Lat. findo, like vendo from Lat.
véndo; messo from Lat. missus, like mesa from Lat. mé(n)sa ; all
with close ¢), being distinguished only in the oldest Romance
dialect, viz. Sardinian (Sard. veru from Lat. vérus, but pira from
Lat. *pira, ptrum, for which the Italian words are vero, pera), so
that the two sounds must have become very like one another in
quality at an early period of Vulgar Latin. They were not
however identical, for they are clearly distinguished in Latin
loanwords in other languages (e.g. Lat. ftdes, céra, loanwords of
the second to the fourth cent. a.p., are in Welsh ffydd, ewyr).
And so the probable history of the change of ¢ to 7 in unaccented
syllables (e. g. eligo from Jego), is that the open e first became
close e, and then passed into 7% Italian ¢ in unaccented syllables
is similarly close e, for example, the final e of diece, ‘ten’;
and English speakers of Italian often fail to give this sound
correctly. It is only in the unaccented syllable that we find
2 substituted for e in the mispronunciations censured by the
grammarians, pinaria for penaria, pidato for pedatu, decim for
decem, &c., though on plebeian epitaphs, and the like, we find
instances of i for accented ¢, some of which may be a mere
graver’s mistake of dropping a stroke of II, a common way of
writing E. Before another vowel @ seems to have approached
the sound of 7, to judge from the frequent confusion of suffixes
like -eus and -ius,-eolus and -iolus. Before 2, older ¢i, the 7-sound
was recognized as the correct pronunciation, or at least the correct
spelling, e.g. mies, méiis, contracted to mis; dii, diis, contracted
to di, dis; ii, is, contracted to 7, ig in Plautus, &c. The con-
trary tendency, to substitute ¢ for 7 before a vowel, was a feature
of rustic Latin; and some of these ‘rustic’ forms, especially in
names of agricultural implements and the like, have found their
way into ordinary Latin, e.g. mateola for *matiola (Sanscyr.
matyam). Another dialectal change was to replace @ by ¢ before
C2
20 THE LATIN LANGUAGE. (Chap. IT.

re, e.g. stircus for stercus. The remark of Quintilian (i. 4. 8)
that some new letter was required to express the sound of the
final vowel of Jeri, ere (in ‘here’ neque e plane neque i auditur),
is discussed in § 16. &

In Oscan, as we shall see (§ 14), the short and long E-sounds
seem to have corresponded to Latin ¢, é, in being the one open,
the other close. A short E-vowel, however, when lengthened by
‘compensation’ or any other cause, appears to have retained the
open sound ; for it is expressed by doubling the symbol of the
short vowel (e.g. cestint, Lat. exstané; keenzstur, Lat. censor,
censores), and not by the symbol of the long vowel. For Latin
nevertheless the evidence points to é- for éx, -é(n)s- for -éus-
having had the close sound; for év¢at is the instance given by
a grammarian (see below) of the close E-sound of ordinary
Latin 2, and the Romance and Celtic forms of Lat. mensa,
mé(n)sa (Ital. mesa, Welsh mwys, &c.) point to the ordinary
é-vowel. But Vulg. Latin Jénuarius (for Januarius) is shown
by Italian Gennajo (with open ¢) to have had the E-sound which
is most near a, that is, the open sound. (On ens see § 144.)

§ 7. Descriptions of the E-sound by Latin phoneticians, &c. Teren-
tianus Maurus in his account (329. 116 K.) :—

e quae sequitur vocula dissona est priori,
quia deprimit altum modico tenore rictum,
et lingua remotos premit hine et hine molares,

curiously omits all reference to the difference between short and long e.
There must, however, have been such a reference in some part of his writings,
for Pompeius, as we shall see, quotes him as an authority on this very point.
Marius Victorinus, whose account always closely corresponds with his, after
describing e as follows (33.1 K.): e quae sequitur, depresso modice rictu oris
reductisque introrsum labiis effertur, goes on to say: o, ut e, geminum vocis
sonum pro condicione temporis promit. Martianus Capella (iii. 261) has:
E spiritus facit lingua paululum pressiore. More valuable are the remarks of
those grammarians who give practical hints on the correct pronunciation
of actual Latin words. Servius (fourth cent. a.p.) (in Don. 421. 17 K.) is very
clear: vocales sunt quinque, aeiou. ex his duae, e et o, aliter sonant
productae, aliter correptae . . . e quando producitur vicinum est ad sonum
i litterae, ut ‘meta;’ quando autem correptum, vicinum est ad sonum
diphthongi, ut ‘equus.’ (By the ‘diphthong’ he means ae of aequus, &c.)
Cautions against the confusion of ‘equus’ and ‘aequus’ occur more than
once in the writings of the grammarians. Thus Pompeius (fifth cent. a.p.)
says the one vowel-sound is short, the other long (285. 6 K.): plerumque male
pronuntiamus et facimus vitium, ut brevis syllaba longo tractu sonet . . . siqui
§§ 7-9.] PRONUNCIATION. VOWELS. a1

velit dicere ‘aequus’ pro eo quod est equus, in pronuntiatione hoe fit (ef.
Aleuin 295. 4 K.: ‘aequitas,’ ‘aequus,’ id est justus, ... per ae diphthongon
seribenda sunt ; ‘equus,’ si animal significat, per simplicem e) (see § 41).
Pompeius, in another passage (102. 4 K.), ascribes the comparison of the long @-
to the i-sound to Terentianus Maurus: e aliter longa, aliter brevis gonat...
dicit ita Terentianus ‘ quotienscumque e longam volumus proferri, vicina sit
adilitteram.’ ipse sonus sic debet sonare, quomodo sonat i littera. quando
dicis ‘evitat,’ vicina debet esse, sic pressa, sic angusta, ut vicina sit ad i
litteram. quando vis dicere brevem e, simpliciter sonat. And ‘Sergius’ (in
Don. 520. 27 K.) gives much the same account as Servius: vocales sunt quinque.
hae non omnes varios habent sonos, sed tantum duae,e eto. nam quando e
correptum est, sic sonat, quasi diphthongus, ‘equus ;* quando productum est,
sic sonat, quasi i, ut ‘demens.’

§ 8. i for unaccented 6. Caper (first cent. a.p.) (93. 3 K.): cella penaria,
non ‘pinaria,’ dicendum ; ibid. too. 23 K. primo pedatu, non ‘pidato,’ dicen-
dum ; Velius Longus (first cent. a. p.) (76. 9 K.): ‘comprimo’ quoque per
i malo seribi, quamvis ‘ compressus’ dicatur ; et e contrario ‘decem’ audacius
dixerim, quamvis inde ‘ decies’ trahatur, quoniam, ut supra dixi, sono usita-
tiore gaudet auditus, referring to the form ‘ decim,’ which is found now and
then on inscriptions. (For other examples, see ch. iii. § 22.)

§ 9. ifor 6 in hiatus. The Appendix Probi censures vinia (198. 3 K.), cavia
(198. 5), brattia (198. 6), coclia and cocliarium (198. 6), lancia (198. 8), solia (198. 10),
calcius (198. 10), tinia (198. 19), baltius (198. 23), lintiwm (198. 31), palliarium (198.
9), fassiolus (198. 26). Ariam (e. g. C. I. L. vi. 541, of 88 a. v.), horriorum (e. g. vi.
8680, of 68 a.p.), are frequent spellings on inscriptions. Cf. Greek éppa, dpa
from the end of the first cent. a. D. ; etAcov in the Edict of Diocletian, 301 a. D.;
Tlo7oAo, TlorwwAo. Vulg. Lat. *mia for mea, &e., is shown by Sard. mia,
O. Fr. moie, Roum. mea. Velius Longus (first cent. a. p.) says (77. 16 K.):
nostris auribus’ placet ... ‘miis’ per i, non ‘meis’ per e, ut Terentius :—

at enim istoc nihil est magis, Syre, miis nuptiis aduérsum.

Our MSS. of Terence have not preserved the old spelling in this passage (Heaut.
699); but that it was a spelling current in the older period we see from mieis
on one of the Scipio Epitaphs (C. I. L. i. 38, of 130 B. c.). In the Lex Parieti
Faciendo (C. I. L. i. 577), a copy of an inscription of 105 B.c., we find the Abl.
Pl, abiegnieis, aesculnieis, distinguished in spelling from Acc. Pl. dbiegnea (K. Z.
xxx. 500). Similarly dii, diis represented the pronunciation, although spelt dei,
deis to agree with the other cases, as we learn from Caper (first cent. a. D.)
(109.6 K.): dei non ‘dii’; nam et deabus Cicero dixit : igitur deis ratio, diis
consuetudo ; and the same must hold of ii, iis (eis, C. I. L. vi. 877, time of
Augustus ; ieis, iei, but eos, &e., on the Regulations for the Ludi Saeculares of
17 B.C. (Monumenti Antichi i. iii)); ef. Caper 106. 11 K. eam (MSS. iam)
semper dicendum, quia nihil est ‘iam.’ item non ‘iamus,’ sed eamus. (On
the spelling of the Plur. of is and deus see Georges, Lex. Wortf. s.vv.) By-

 

1 But the use of 1 may have been the long i-sound (i. 9), or -eei-, which
a mere usage of orthography to avoid might suggest the é-sound (ibid.) fol-
the awkward collocation -ei-, which lowed by i.
might be misread as the symbol for
22 THE LATIN LANGUAGE. (Chap. II.

forms in -ea and -ia are sometimes differentiated by the subtlety of gram-
marians, a practice very properly censured by Cornutus (ap. Cassiod. 150. 18
K.) : vineas per e quidam scribendas tradiderunt, si hae significarentur, quas
in agris videmus ; at contra per i, vinias, illas sub quibus latere miles solet,
quod discrimen stultissimum est. nam neque aliunde vineae castrenses
dictae sunt, quam quod vineis illis agrestibus similes sunt. (For other examples
of -ea, -eus varying with -ia, -iws, see Schuchardt, Vok. i. p. 424 ; Brambach
Orth. p. 133 ; and consult Georges, Lex. Wortf. s. vv. glarea, linea, janeus, inustaceus,
virgineus, vitreus, gallinaceus, cavea, urceolus, adored, lanced, oreae, ostrea, pausea, labea,
linteo, phaseolus, nauseo, coprea, cochlea, hordearius ; and Brambach, Hiilfsbiichlein
8. VV. balteus, solea, tinea, bractea.) They have been explained by that tendency
to change i and ¢ in hiatus into consonantal é (y), which turned filium &e. into
“lilyum,’ drea &. into ‘arya,’ so that -eus, -ewm, -ius, ium were merged in the
same sound (see ch. iv. § 63). But they are more easily explained by the
tendency to give a vowel in hiatus the close sound (§ 18).

§ 10. ‘Rustic’ ¢ for i in hiatus. Varro (R. R. i. 2. 14): rustici... viam
‘veham’ appellant. (The Oscan word is via-, the Umbr. vea- and via-'; -eo
for -io is common in inserr. of Etruria, Praeneste, &c. (Sittl, Lok. Versch. p. 10),
e.g. Praenestine fileai (C. I. L. i. 54); the Praenestine form of cicdnia is given
by the MSS. of Plautus, Truc. 690, as conea not ‘conia’: ut Praenestinis conea
est ciconia; in Plaut. Most. 48 the MSS. read dleato ‘on garlic.’ Charisius
(70. 27 K.) censures the pronunciation of ‘alii diserti’ aleum, dolewm, palleum ;
the Appendix Probi rejects alewm (198. 18 K.), lilewm (198. 19), laneo (197. 29),
osteum (198. 5). [For other exx. see Georges, Lex. Wortf. s. vv. alium, ascia, dolium,
Solium, lanio, ostium, pallium, soliwm, spolium ; and Brambach, Hiilfsb. s.vv. feriae,
lilium, sobrius, Both spongia and spongea were used, the latter being, for example,
the normal spelling in the MSS. of Martial (see Georges, s.v. and Friedlinder’s
edition, i. p.118). Schuchardt (Vok. ji. p. 37) gives a number of misspellings
of the kind from MSS. and inscriptions. ]

§1l. ifor é. Schuchardt (Vor. i. p. 227) quotes a large number of spellings
from inserr. of the fourth cent., and later in which an ¢ appears for 2 It
is not easy to decide how many represent a pronunciation of i for Z [thus
Aurilius (third cent.), Cornitius seem to represent that change of 2 to 7 before
a syll. with i (y) in hiatus, which appears in filius, with the ordinary Latin
long i-sound (close 7, § 14), to judge from its Romance descendants, Ital. figlio,
Span. hijo, &c.; on this possible change of z to i, see ch. iv. § 7], how many
a pronunciation of i for é (e. g. filictter?), how many are dialectal (the equivalent
of Lat. @ is an i-sound in several of the Italian dialects, 0. g. Osc. ligud ‘lege’
ch. iv. §5), and how many are mere mistakes. In Greek inscrr. 7 for Lat, 2 is
late, probably dating from the time when Greek 7» came to take the i-sound ;
but Adpidros (-tAAvos) is found beside AdpyAcos in the second cent. (Hckinger,
p. 24). The rare spelling decretuit for décrevit on an inser, of 189 B.c. from
Spain (C. 1. L. ii. 5042) cannot be quoted as an example of the transition of @
toz. The use of ei for ¢ in the unaccented syll. in the word inpeirator (for
impérator) on the same inser. suggests that ef in decretuit may have been meant
to indicate the close e-sound [ef. leigibus xiv. 2892 (Praeneste), pleib. (Eph.
Epigr. i. 3)]. Delirus and delerus are rightly explained by Velius Longus (73. 2 K.},
who follows Varro : delirus is the proper form, derived from lira a furrow, while
the form delérus is due to a fanciful connexion of the word with Gk. Anpety.
“§§ 10-14. | PRONUNCIATION. VOWELS. 23

(On the comparative prevalence of the two spellings see Georges, Lex. Wortf.
s.v., and cf. App. Probi 198. 19: delirus non ‘delerus’) ; torpido beside torpedo
(Caper 106, 8 K.) is merely the substitution of a more familiar for a less familiar
suffix (cf. gréivido for grdvédo ; see Georges, Lex. Wortf.s. v.), and the same is true
of Vulg. Lat. *céninwm (Fr. venin), Bizacinus for Byzactnus (App. Probi,
198. 1), *pergaminum (Ital. pergamino, Fr. parchemin), *pullicinwm (Ital.
pulcino, Fr. poussin) (-énus, -évwm have been in fact supplanted by -%mo- in
Romance languages, e.g. Ital. Saracino, Messina, &c., Fr. serin, a canary, if
from Sirén), *rdcimus (Ital. racimolo, Fr. raisin), *vervicem (Ital. berbice, Fr.
brebis), *mantile (Ital. mantile), &c. (See also Georges, s. vv. criiména, séstimum,
stricus ; ef. App. Probi 199. 6: hermeneumata non ‘erminomata.’)

§ 12. i for accented 6 (see Schuchardt, Vok. i. p. 329 sqq.). Bipinnis for
bipennis, censured in App. Probi 199. 6 K. (cf. Quint. i. 4. 12), is due to confusion
of pinna with penna (ef. Caper 100. 17 K.) ; carictum beside carectum (de Dub. Nom.
v. 573. 2 K. Virgilius in bucolicis ‘tu sub carecta latebas,’ nunc caricta), may
follow the analogy of sdlictum, &c., but it is more likely that the true reading
here is caricea, Plur. of caricewum (Nonius 21. 24 M.?), the original of Span.
earrizo. On vigeo and végeo, filix and félix, fiber and féber, pinna and penna, see
ch. iv. § 11 ; and on bénévilus and bénivilus, bénéficus and bénificus, &e., ch. ili. § 37.
Scida, schida for schtda (see Georges, s.v.) seems to follow the analogy of scindo ;
spicio, sico, &c., of prospicio, prosico, &e. The use of ¢ for t is discussed in § 17.
Before ng every Latin ¢ became by a phonetic law of the language z (see ch. iv.
§ 8), e.g. tingo, confringo, attingo ; and other consonant-combinations may have
influenced é towards the close e-sound or the i-sound (see Georges, Lex. Worty.
s. vv. Vergilius, vergiliae, Verginius, hernia, segmentum, Porsenna, and cf. § 144) ;
dignus has been explained as *dec-nus from décet, and spellings like frumintum
(Schuchardt, Vok. i. 354) may point to a modification of é like that of 6 (ch. iv.
§ 20) before nt.

§13. &foré. Befoie rc we find a for e¢ in the mispronunciation novarca (for
noverca) mentioned in the Appendix Probi (198. 34 K.) ; but forms like ansar
(ib. 198. 22 and 23), passar (ib. 198. 33), carcar in the Acts of the Arval Brothers,
&e. (cf. App. Prob. 197. 32), are better explained as cases of assimilation to
the vowel of the accented syllable (see ch. iii. § 33); and calandae, the Vulg. Lat.
form of célendae [in Greek always xadavéa: (cf. Eckinger), Welsh calan, New
Year’s Day, Mod. Gk. xdAavra}, may be a1 Conj. Gerundive form. (Schuchardt,
Vok, i. p. 206 sqq., has collected a list of examples of doubtful validity.)

§ 14. I. The Romance languages show us that Latin 7, @ differed
in quality, like Latin ¢, ¢. Latin 7, as we have seen, they
merge in Latin @, while Latin 7 remains 7, e.g. Ital. beve, from
Latin 026%2, misi from Latin mist. This i from Latin 7 is, of all
Romance vowels, the least liable to change. In almost every
Romance language it preserves its character unaltered, and
resists every influence of neighbouring consonants, so that there
is great likelihood that it has remained the same on Italian soil
from Roman times till now. Italian si will then exactly repre-
24 THE LATIN LANGUAGE, [Chap. II.

sent the vowel sound of Latin sic, finito of Lat. finttus. This
Italian 7 has the close I-sound, like French si, fini, Germ. sie, our
‘see’ being rather a diphthong, while our ¢ in ‘ bit,’ ‘ fish,’ ‘ kin,’
is an open I-sound, but, according to Sweet, ‘nearer e of “men”
than 7 of German Kind, bitten.” The Latin grammarians
similarly speak of the sound of i as fuller (p/enzor) than that of
%, while the latter is ‘a sound between e and 7. They add
a third I-sound, which in the Romance languages is not dis-
tinguished from ordinary Latin 7%, viz. the 7% of optimus older
optumus, &e., which they style ‘a sound between i and uw.” The
natural inference is that Latin 7 and 7 differed as Latin ¢ and é,
the short vowel being open, the long close, while 7 before a labial,
in words like optimus, had some sound like that of German # in
schiitzen, Hutte. This @ is the same sound as Germ. : of ‘ Kind,’
‘bitten,’ modified by labialization, or as it is usually called, ‘round-
ing,’ i.e. lateral compression of the cheek passage, and narrowing
of the lip-aperture (Sweet, Handbook, p. 13).

The relation of the E- and I-sounds in Oscan seems to offer
a close parallel to that in Latin. I.-Eur. @ is in the Oscan
alphabet e, e. g. edum (Lat. édere, esse ‘to eat’), estud (Lat. esto) ;
2 is i, e.g. bivus (Lat. viz, Nom. Pl.). For é and 7 they have
the same sign (except that for é it is often written double to
indicate length), a modification of this last, which we conven-
tionally write {, e.g. flisnam, a temple, Acc. Sg. (cf. Lat. féstus),
pid (Lat. guid). In Greek characters the three signs are «, 1, €1;
in Latin characters ¢, 7,7. Everything points to their e having
been, like Latin é, an open E-sound, their i the close I-sound of
Latin 7, while like the Romance languages they have merged
close ¢ and open 4 in one sound {. Whether the iu of Oscan
altiumam (Lat. w/timam) indicates the Latin d-sound is, with the
scarcity of material at our disposal, uncertain (see ch. iv. § 23).

In rustic Latin, as we saw (§ 10), i before a vowel was replaced
by ¢, e.g. mateola for *matiola, while in the ordinary language
accented 7 before a vowel, e. g. dies, seems to have had the quality
of long 7. It had possibly the same quality in words like audit,
where the %, originally long (e.g. audit, Plaut.), has been
shortened owing to the difficulty felt by the Romans in pro-
nouncing a long vowel before final ¢ (see ch. iii. § 49). Rustic Latin
§§ 15, 16.] PRONUNCIATION. VOWELS. 25

é for @ in speca, &e., is probably a development of the old
diphthong e, Similarly ¢, the old vowel of the open unaccented
syllable, e.g. O. Lat. dégit for dbigit, was retained in rustic
Latin.

§ 15. Deseriptions of the I-sound by Latin phoneticians. The phone-
ticians describe only the 7-sound. Terent. Maur. (329. 1r9 K.) :—

i porrigit ictum genuinos prope ad ipsos,
minimumque renidet supero tenus labello ;

Mar. Victor. (33. 2 K.): i semicluso ore impressaque sensim lingua dentibus
vocem dabit ; Mart. Cap. (iii. 261) : I spiritus (facit) prope dentibus pressis.

§ 16. by Grammarians. The ii-sound attracted a good deal of attention
from Latin grammarians, and had the honour of being noticed by various
rulers of the Roman Empire. Julius Caesar effected the adoption of the
spelling optimus maximus, and the like, on State inscriptions; his successor
Augustus, we are told, reverted in these forms to the old spelling with u;
the Emperor Claudius took a course different from both of his illustrious
predecessors, and tried to introduce into the Latin alphabet a new letter to
express this particular sound. The statements of the Latin grammarians
about the various sounds of Latin i are not always perfectly clear, and must
be examined in detail. Quintilian (middle of first cent. a.p.), speaking of
the letters wanting to the Latin alphabet says (i, 4. 8) that some special
sign is required for the sound betweeni and ~ in words like optimus, and
similarly for the sound between e and i in words like hére (older hért), sibi, qudsi
(older sibe, guise) (ef. nise, ibe in Virgil MSS., Ribbeck, Index, pp. 436, 451, nise
in Lex Rubria of 49 B.c. (C.1.L.i. 205), and (with ube, sebe, &c.) on late inserr.
(see Georges). In the Appendix Probi (199. 16 K.) we have: nescio ubi non
‘nesciocube’] : medius est quidam u et i litterae sonus ; non enim ‘optumum’
dicimus aut ‘optimum’,’ et in ‘here’ neque e plane neque i auditur. In
another passage he tells us that the change in spelling, optimus, &c., from
optumus, &c., was made by the influence of Julius Caesar, a statement repeated
by Velius Longus in the passage quoted below, and first uttered, according to
Cornutus (ap. Cassiodor. 150. 11 K.) by Varro: Quint. i. 7. 21: iam ‘optimus
maximus,’ ut mediam i litteram, quae veteribus u fuerat, acciperent, Gai
primum Caesaris inscriptione traditur factum. ‘here’ nunc e littera
terminamus ; at veterum comicorum adhuc libris invenio ‘heri ad me uenit,’
quod idem in epistulis Augusti, quas sua manu scripsit, aut emendavit,
deprehenditur . .. ‘sibe’ et ‘ quase’ scriptum in multorum libris est, sed an
hoe voluerint auctores nescio ; T. Livium ita his usum ex Pediano comperi,
qui et ipse eum sequebatur ; haec nos i littera finimus. This example, here
and heri, has given rise to some doubt; for we are accustomed to regard hert
and hert as differing in quantity, like rure and ruri, Tibure and Tiburi (the -e
being the Cons.-stem Locative suffix -i, the -i being the I-stem Abl. suffix -id,

 

17,2. non enim sic ‘optumum’ of one MS., opimum for optimum, is
dicimus ut ‘optimum.’ Perhaps,non certainly wrong. Cf. Quint. i. 7. 21,
enim sincere... aut. Thereading 22.
26 THE LATIN LANGUAGE. {Chap. II.

ch. iv). We have heré, for example, in Martial i. 44 est positum nobis nil
here praeter aprum, but keri in Terence, Eun. 169 heri minas pro ambdébus
uiginti dedi, often shortened by the peculiar metrical law of the comedians
by which dvé was scanned as dvé, cdive as car’, &c. (see ch. iii), as in the line,
Hec. 329 heri némo uoluit Séstratam intro admittere. Some have been led
by this into the mistaken idea that what Quintilian is referring to, is that
interchange of ¢ and 7, which we see on old inscriptions in rendering the
diphthong ei, e. g. ploirume for ploirumei, later plurimi, on an epitaph of one of
the Scipios (C. I. L. i. 32). But, as we shall see from the passages quoted from
other grammarians, the phrase ‘a sound between ¢ and 7’ is the designation
for Latin 7 in such a word as hominem ; and we may be sure that in Quintilian’s
time the word heri, as well as here, would invariably have in the utterance of
everyday speech a short final syllable. He tells us expressly of the word ave
(have) that, although it ought by right to lave a long final vowel, being an
Imperative of a verb of the second conjugation, like splendé, audé, it was never,
except by precisians, pronounced otherwise than haré (i. 6, 21). (A fuller
account of this shortening is given in ch. iii. § 40.) Velius Longus, who like
Quintilian belonged to the first cent. a.p., talks of the ‘exilis sonus’ of
the Latin vowel 7 and, in some cases, of 7, e.g. in 3 Sg. Pres. Ind. of verbs of
the fourth conjugation, audit, ke. (In Plautus and the oldest literature this
iis long, audit, the shortening having been effected by the influence of the
final ¢, just as with us the vowel of ‘note’ is shorter than the vowel of ‘node.’
See ch. iii. § 49.) Ordinary %, asin 3 Sg. Pres. Ind. of the third conjugation,
ponit, &c., he calls the ‘latus sonus,’ while the i of optimus is ‘pinguis.’ Of
this last sound he says that the spelling and pronunciation of « for ¢ in
optimus, manibiae, &c., was regarded in his time as old-fashioned and countrified
(49 K.): i vero littera interdum exilis est, interdum pinguis, ut in eo quod
est ‘ prodit’ ‘ vincit’ ‘condit’ exilius volo sonare in eo vero quod significatur
prodire vincire condire usque pinguescit ut jam in ambiguitatem cadat utrum
per i quaedam debeant dici an per u ut est ‘optumus maxumus.’ in quibus
adnotandum antiquum sermonem plenioris soni fuisse et, ut ait Cicero,
rusticanum, atque illis fere placuisse per u talia seribere et enuntiare.
erravere autem grammatici qui putaverunt superlativa per u enuntiari.
ut enim concedamus illis in ‘optimo,’ in ‘maximo,’ in ‘ pulcherrimo,’ in
‘justiesimo,’ quid facient in his nominibus, in quibus aeque manet eadem
quaestio superlatione sublata, ‘manubiae’ an ‘manibiae,’ ‘libido,’ an
‘Iubido’? nos vero, postquam exilitas sermonis delectare coepit, usque
i littera castigavimus illam pinguitudinem, non tamen ut plene i litteram
enuntiaremus. et concedamus talia nomina per u scribere iis qui antiquorum
voluntates sequuntur, ne tamen sic enuntient, quo modo scribunt; and
again (67 K.): varie etiam scriptitatum est ‘mancupium’ ‘aucupium’
‘manubiae,’ siquidem C. Caesar per i scripsit, ut apparet ex titulis ipsius, at
Augustus per u, ut testes sunt ejus inscriptiones'...  relinquitur igitur
electio, utrumne per antiquum sonum, qui est pinguissimus et u litteram
occupabat, velit quis enuntiare, an per hunce, qui jam videtur eligantior,
exilius, id est per i litteram, has proferat voces; and a little further on
(68. 6 K.): mihi videtur nimis rusticana enuntiatio futura, si per u extu-
lerimus. ita tamen existimo enuntiandum, ut nec nimis i littera exilis

 

1 But the Comm. Lud. Saec. has optimus maximus,
§ 16. ] PRONUNCIATION. VOWELS. 27

sit, nec, u litteram si scripseris, enuntiationis sono nimis plena. Here
his account of the ‘sound between i and w’ is perfectly clear. Optumus had
in old times been written and pronounced with a U-sound; but such
pronunciation was old fashioned and countrified in his own time, the sound
used in polite circles being something between a normal U-sound and
® normal I-sound. But the beginning of the first passage, where he dis-
cusses the other two kinds of I-sound, is not so intelligible. His examples
are evidently carefully chosen synonyms of the third and fourth conjugations;
and one cannot but suppose that he meant to contrast the i of prodit (from
prodo), vincit (from vinco), condit (from condo), with that of the third person
singular of prodire, vincire, condire. Keil supplies a sentence between the
words sonare and in eo, and reads : exilius volo sonare, si dico ab eo quod est
prodere, vincere, condere ; in eo vero quod significat prodire, &c. He refers
the ‘pinguescit’ to the sound of 2, not to the ¢ of optimus, &c., understanding
Velius Longus to distinguish ¢ from i as ‘exilis sonus’ and ‘pinguis sonus.’
These words, ‘ exilis,’ ‘latus,’ ‘pinguis,’ unfortunately lack the precision of
the terminology of modern phoneticians, They remind us of Lucilius’ use
of ‘tenuare’ and ‘plenius facere’ some two centuries earlier, in a passage
not less obscure (9. 14 M.) :—

‘pilam’ qua ludimus, ‘ pilum’
quo pisunt, tenues, si plura haec feceris pila
quae iacimus, addes e, ‘peila,’ ut plenius fiat ;

whereas a later grammarian, Pompeius (fifth cent. a.p.), uses ‘tenuis’ and
‘pinguis’ to distinguish vocalic from consonantal i and w (103 K.): ecce
adverte, quomodo sonat u, ‘unus,’ ecce u vides quam tenuiter sonat. junge
illam ad aliam litteram, et vide quia non sic sonat, sed pinguius sonat,
‘vulnus,’ ‘vanus.’ numquid sie sonat ‘unus’ quando u sola est? non, sed
tenuiter sonat. ‘vanus’ quando dico pinguior sonus est. numquid dicis
‘u-a-enus’? ergo vides quia, si ponantur solae, tenuem sonum habent, si
jungantur ad alias litteras, pingues sonant. similiter et i sic patitur. ‘itur,’
ecce tenuius sonat; si dicas ‘Titius,’ pinguius sonat, et perdit sonum suum,
et accipit sibilum. (¢ palatalized.) This confusion of terms must be borne in
mind in reading the passage we now quote from Consentius (fifth cent. a.p.?),
a passage interesting from its account of the Gaulish and Greek mispronun-
ciations of Latin i (394. 11 K.): iotacismum dicunt vitium quod per i litteram
vel pinguius vel exilius prolatam fit. Galli pinguius hance utuntur, ut cum
dicunt ‘ite,’ non expresse ipsam proferentes, sed inter e et i pinguiorem
sonum nescioquem ponentes. Graeci exilius hane proferunt, adeo expressioni
ejus tenui studentes, ut, si dicant ‘jus,’ aliquantulum de priori littera sic
proferant, ut videas disyllabum esse factum. Romanae linguae in hoc erit
moderatio, ut exilis ejus sonus sit, ubi ab ea verbum incipit, ut ‘ite,’ aut
pinguior, ubi in ea desinit verbum, ut ‘habui,’ ‘tenui’; medium quendam
sonum inter e et i habet, ubi in medio sermone est, ut ‘hominem.’ mihi
tamen videtur, quando producta est, plenior vel acutior esse ; quando autem
brevis est, medium sonum exhibere debet, sicut eadem exempla, quae posita
sunt, possunt declarare. Consentius here uses ‘pinguis’ and ‘tenuis’ or
‘exilis’ like Pompeius, not like Velius Longus, while he distinguishes long ¢
as ‘plenior vel acutior,’ short 7 in hominem as ‘a sound between e and i.’
What he means by saying that in habut, tenui, i had the ‘pinguis sonus,’ must
28 THE LATIN LANGUAGE. (Chap. IT.

be that the words were pronounced in his time ‘ habuyi,’ ‘tenuyi.’ (Seelmann
gives quite a different explanation.)

These three accounts of Latin 7, 7, taken in connexion with the evidence
supplied by the Romance languages, where Latin ? has become a close
E-sound, while Latin i is invariably close I, give us the right to suppose
that the usual Latin i was different in quality from Latin 7, being an open
I, like Engl. ‘ bit’ or Germ. Kind, while 7 was the ordinary close I of Italian
and other languages; though they suggest the further possibility of there
having been a short variety of this latter iin such words as the 3 Sing. Pres.
Ind. Act. of the fourth conjugation, prodit, audit, condit, where the 7%, long in
the time of Plautus, was shortened through the difficulty felt by the Romans
in pronouncing a long vowel before final -t. Seelmann is of opinion that this
short variety of close i came gradually to replace open? in the language of
the educated classes at Rome in the first centuries of the Empire (postquam
exilitas sermonis delectare coepit, Vel. Long.), and so explains the strange
statement of the fourth century commentators on Donatus, quoted in our
discussion of the sound of e (§ 7), that % 7, with %, @, were not distinguished
like & 2, 6, 6; though Consentius a century later enlarges on this very
distinction in the case of i. It is more probable that these commentators,
though they refer to some Latin instances (méta, démens, équus), are really
quoting remarks of Greek phoneticians on the sounds of Greek vowels,
designed to explain the presence of separate signs for iong and short ¢ and o
in the Greek alphabet («, 7, 0, w); and, if this be so, it cannot but suggest
the alarming suspicion that their phrase, ‘¢ is like the diphthong, é like i,’
may really mean that Greek « had the sound of a (as was the case in Attic
Greek by the second cent. a..), 7 of « (the itacism of modern Greek. Blass
ascribes the change of 7 to « to the fifth cent. a. D.).

To pass to the ti-sound of optimus, which in the Romance languages is merged
in ordinary Latin i. The statements of the grammarians we have quoted, par-
ticularly that of Velius Longus (49 K.), show us clearly that in the first century
A.D. the vowel had a sound between wu and i, having had at an earlier period
a U-sound. Still earlier it was an o (see ch. iii. § 18), and we may regard it
as the sound which % (whether originally o or u or a, &c.) took in open syllables
after the accent, when influenced by the presence of a labial; whereas accented
au remained, e.g. ciibo. The passage which we now quote from Velius Longus
extends the same sound to 7 in accented syllables influenced by the labial
sibilant v, as in vir, virtus (75 K.): ‘aurifex’ melius per i sonat quam per u.
at ‘aucupare’ [et aucupium] mihi rursus melius videtur sonare per u quam
peri; et idem tamen ‘aucipis’ malo quam ‘aucupis,’ quia scio sermonem et
decori servire et aurium voluptati. unde fit ut saepe aliud scribamus, aliud
enuntiemus, sicut supra locutus sum de ‘viro’ et ‘ virtute,’ ubi i scribitur et
paene u enuntiatur. unde Ti. Claudius novam quandam litteram excogitavit
similem ei notae quam pro adspiratione Graeci ponunt, per quam scriberentur
eae voces, quae neque secundum exilitatem i litterae, neque secundum pingui-
tudinem u literae sonarent, ut in ‘viro’ et ‘virtute,’ neque rursus secundum
latum litterae sonum enuntiaretur, ut in eo quod est legere, scribere. In the
last words he seems to refer to + of the third conjugation, legit, legimus,
legitis, &c., the sound of which he calls ‘latus’ as opposed to the ‘exilis
sonus’ of 7, and the ‘pinguis sonus’ of i/w; and this confirms our view that
in the passage first quoted from him, the same threefold distinction was
§ 17.] PRONUNCIATION. VOWELS. 29

explained between prodit of third conj. with ‘latus sonus,’ prodit of fourth conj.
with ‘exilis sonus,’ and optimus with ‘pinguis sonus.’ There are a large number
of references by other grammarians to this i/w sound (see Seelmann, p. 205).
Of these we need only quote two; one from Marius Victorinus (fourth
cent.), who points out that this vowel is really the Greek v (Latin y) (see
§ 28); and one from Priscian, who, like Velius Longus, gives this sound of
Greek uv to accented ¢ influenced by a preceding v. Mar. Victor. 19. 22 K. sunt
qui inter u quoque et i litteras supputant deesse nobis vocem, sed pinguius
quam i, exilius quam u. sed et pace eorum dixerim, non vident y litteram
desiderari: sic enim ‘ gylam,’ ‘myserum,’ ‘Syllam’ (MSS. syllabam), ‘ proxy-
mum’ dicebant antiqui. sed nunc consuetudo paucorum hominum ita loquen-
tium evanuit. ideoque voces istas per u (vel per i) scribite. The spelling
myserum, which is found on some inscriptions, may be explained by Greek
puoapés, just as silva was spelt sylva through a fanciful connexion with #An,
and so Sylla for *Syrila (cf. App. Probi 197. 26 crista non ‘crysta’); but it
is difficult to explain gyla (for giila) in the same way. All the Romance
languages point to gitla as the Vulgar Latin form (Ital. Span. gola, Fr. gueule).
The ‘ antiqui’ alluded to are merely former grammarians, whose innovation
in spelling met with little favour, to judge from the instances found on
inscriptions (Schuchardt, Vok. ii. pp. 197 sqq., 218 sqq.). Prisciani. 6ietu
vocales, quando mediae sunt, alternos inter se sonos videntur confundere,
teste Donato, ut ‘ vir,’ ‘optimus,’ ‘ quis’; et i quidem quando post consonan-
tem loco digamma functam Aeolici ponitur brevis, sequente d vel m vel r vel
t vel x, sonum y Graecae videtur habere, ut ‘ video,’ ‘ vim,’ ‘ virtus,’ ‘ vitium,”
‘vix.’ Schuchardt, Vok. ii. p. 221, gives a few examples of vy- for vi- in late
inscriptions ; and in the Appendix Probi (198. 20 K.) we have: vir non ‘vyr,’
virgo non ‘ vyrgo,’ virga non ‘vyrga,’ so that the existence of this tendency
to pronounce accented i as ti after v can hardly be doubted. (The Latin name
for y, Greek v, was ‘ui.’ See below.) But Greek v does not represent Latin i
in this position on Greek inscriptions. Other examples of accented i/u are
simus, written for siimus by some purists of the Augustan age (Mar. Victor.
9. 5 K. Messala, Brutus, Agrippa pro sumus ‘simus’ scripserunt), and by
Augustus himself (Suet. Aug. 87) (cf. C. I. L. ix. 3473. 14); Wibet and libet ; clupeus
and clipeus (see Georges, Lex. Wortf. s. vv.) ; though two of these, simus and libet,
might be explained as enclitic words and so wanting the accent, e. g. amatt-
sumus, lubet-ire, quédlubet, &c. (see iii. 12). (See also Georges, Lex. Wort/. s. vv.
Bruttii, cliens (earlier cluens), linter, seriipulus. The Romance forms point to both
stimus (e.g. Fr. sommes) and simus (e. g. Ital. siamo from *semo, 0. Roum. semo).
Supparum, with byform siparum (see Georges s. v.) seems to be an Oscan word
(Varro, L, L. v. 131), and the mispronunciations ‘ imbilicus’ (Prob. App. 198.
4K.; ef. Ir. imbliu), ‘scoriscus’ (ib. 198. 32 K.), ‘arispex’ (Vel. Long. 73. 9 K.)
have been variously explained.) (On the ti-sound, see Parodi in Studi Italiani,
i. 385.)

§17. Interchange of i and e. The misspellings on inscriptions testify abun-
dantly to the close relation between i and ¢ (close e), e.g. aressemo merentessemo
(C.1.L. if. 2997) (see Schuchardt, Vok. ii. pp. 1-67); but 7 is rarely written
e except in Gaul and Britain, where 6 too appearsasu. In rustic Latin indeed
such words as spica were pronounced spéca (Varro, R. R. i. 48. 2); but it is not
clear whether this was not confined to words which originally had the
diphthong ¢ (cf. vella, Varro, R. R.i. 2 14). Ifso, the ¢ is that dialectal e for
30 THE LATIN LANGUAGE. (Chap. ITI.

I.-Eur. ei which is found in the Umbrian language, e.g. prevo- (Lat. privus,
privatus ; Ose. preivato-), and in various parts of Italy. Demidius for dimidius
(App. Prob. 198. 27 K.) is due to confusion of dz with di- (dis) [ef. demedius,
C. I. L. vii. 140; x. 3428, and in MSS. (see Schuchardt, Vok. ii. 71). Fr.
demi]; Serena for Siréna (App. Prob. 199. 10) to confusion with sérénus (cf. Fr.
serin, canary ?). On délirus and délérus, see ch. iii. The vowel of the open unac-
cented syllable was in Old Latin é, not 7 (see iii. 18; ; and this ancient sound
remained in Rustic Latin. So that ¢ for t of hominem, &c., as well as é for 7 of
spica, &e., and @ for 7 in hiatus of via, &c., characterized the pronunciation of
the country districts. Cicero often alludes to the ‘rustic’ substitution of the
e- for the i-sound in the utterance of his friend L. Aurelius Cotta, the author of
the famous jury-law, the Lex Aurelia Judiciaria of 70 8. c.: quare Cotta noster,
cujus tu illa lata, Sulpici, nonnunquam imitaris, ut iota litteram tollas et
e plenissimum dicas, non mihi oratores antiquos, sed messores videtur imitari
(de Orat. iii. 12. 46. Of. iii. rr. 42; Brut. xxxvi. 137 ; xxiv. 259 ; and Quintilian
xi. 3. 10).

The frequent occurrence on Greek inscriptions of ¢ for Lat. i (e.g. TeBeptos,
Aev7iov) may point to Greek ¢ having had (unlike Latin @) the close E-sound
(see Blass, Aussprache des Griechischen?, p. 23). But it may often be merely
a retention of the early Latin spelling, of the form in which the word was
first borrowed by the Greeks. This ¢ for? is the usual spelling in «opervoy at
all periods, while KauceAros is replaced by Karm:Atos after 50 B. c., Kamerwatrov by
KamitwAcoy in the first cent. a.p.; Aemedos is the form of the Republican,
Aem5os of the Imperial Age. (For other examples see Eckinger, p. 29 sqq.,
and for examples of 7 ¢ in Latin, Georges, Lex. Wortf. s.vv. gillo, hibiscum, minus,
minister, sinus, sine, sinapi, vindico, comissor, solidus. On the late Lat. emitari, see
Schuch. Vok. ii. 20, and ef. Prob. App. 199. 2, and 198. 22 K.)

§18.iin hiatus. The Romance forms of the word for day (Ital. di, Sard. die,
Span. dia, O. Fr. di) have all z, which is the normal representative of Latin
7; and on inscriptions we have sometimes the lengthened form of the letter,
which usually denotes long i, e.g. pis (¢. I. 1. vi. 7527); DIE (10239, also
pRIvsqvam). (On Plvs see § 143.)

§ 19. Anomaliesin Romance. Ital. freddo, Fr. froid, from Lat. frigidus, point
to *friddus, from frig(i)dus (ef. App. Probi 198. 3 K. frigida non frigda), where the
thas been referred to the analogy of rigidus (but see § 127) ; *glérem replaces glirem
in Celtic countries (Fr. loir, 0. Prov. gles), but not elsewhere (Ital. ghiro) ;
Vulg. Lat. *sibilo, *siifilo, beside sibilo, *sifilo, to whistle (e.g. Ital. sufilare,
subillare, sibilare, O. Fr. subler, siffler), have been explained by reference to
sufflare, or to sibulo, the Etruscan word for a fluteplayer ; Greek i in yxpiopa is
treated like Latin 7 in Fr. chréme, Ital. cresma.

§ 20. O. Having discovered that Latin ¢ is open E, Latin
@ close E, we are almost entitled to infer that Latin @ will be
open O, Latin 6 close O. For each language has what phone-
ticians call a ‘basis of articulation, according to which all its
sounds are regulated; and if one set of sounds is treated in
a particular way, any set of corresponding sounds is likely to
§§ 18-20. | PRONUNCIATION. VOWELS. 31

receive a similar treatment. The Teutonic languages, for
example, changed the I.-Eur. Aspirate Mediae to voiced Spirants
(dh to d, the sound of our ¢h in ‘this,’ &c.). They correspond-
ingly moved the Tenues to unvoiced Spirants (é to ¢/ of ‘thin,’
&e.); and similarly the Mediae to Tenues (d to ¢, &c.), a move-
ment or gradation of sounds first discovered by Grimm, and
known as ‘Grimm’s Law.’ From detecting one sound in a lan-
guage, we are thus often able to guess what other sounds will be ;
and we could in the absence of other evidence infer the quality
of the O-sounds in Latin from that of the E-sounds. Evidence,
however, is not wanting. The Romance languages, for example,
show Latin é as open O, Latin 6 (with which Latin @ is merged)
as close O. This open O is in many languages developed in open
syllables to wo (e. g. Ital. ruota, from Lat. réta), as open E to i
($6), while in Spanish wo has further developed to we (e. g. ruede),
a change that reminds us of the substitution of ve- for vo- in
Latin words like verto, older vorto. In Italian we have molle
(with open O) for Latin mdllis, sole (with close O) for Latin sd/,
solem, the open O having the O-sound of German voll, Stock, the
close that of German so, Fr. chaud. Our ‘short 0’ in ‘stock,’
‘folly,’ is a ‘lower’ sound, formed with the tongue lower in the
mouth, than the open O of German (our O-sound in ‘oar’ is
nearer this), while our ‘long 0’ in ‘so’ is a diphthong.

In unaccented syllables in Latin open O, before a Labial or /,
seems, like open E before other consonants, to have become close,
and then to have passed into a U-sound,as ¢ into an I-sound, e. g.
sédilé from sé déld, consiilo (Karly Lat.e:2/-). In Italian, 6, like
é, takes the close sound in syllables after the accent. This uw, as
we saw (§ 14), might sink to i, e.g. gonsilinm ; but as a rule o is
retained in the spelling of compouyds more persistently than e¢,
eg. accdlo,agricbla (agricula, Schucl]f. i1. 133). Even whenaccented,
& seems to have had the closefsound before certain groups of
consonants, such as 7 with angther consonant (not //), m or x with
another consonant, 7, &e.4to judge from such varieties of spell-
ing as Old Latin Culeides for Colchides (Quint. i. 4. 16), Old
Latin and Rustic Latin frundes for frondes, Vulg. Lat. turnus for
tornus; and this is confirmed by the misspellings on plebeian
inscriptions, and the like (collected by Schuchardt, Vos. ii. pp. 114,

 
   
 
 
   
 
 
     
32 THE LATIN LANGUAGE. (Chap. II.

&e.). In Celtic countries ~ is often found on Latin inscriptions
instead of 6, e.g. nepus (for népds). (CLL. xii. 5336.)

The O-sounds of the Oscan language offer the same analogy
to the Latin, as the E-sounds (§ 6). The Oscan alphabet, being
borrowed from the Etruscan, had originally no sign for 0, but only
the sign for w. This w-sign was used for 6 as well as for ~, while for
6 a modification of the sign was used, conventionally written by us
4. In Latin characters w expresses Oscan u, and o Oscan a, though
in the final syllable before a labial w sometimes takes its place [e. g.
Ose. estud, in Latin writing estud (Lat. esté, older esto¢), Ose. pad,
in Latin writing pod (Lat. gudd), Osc. deikum, in Latin writing
deicum (Lat. dicere, older deicere), and dolom, dolum (Lat. détum)].
In Greek characters Oscan u is ov, and sometimes o, Oscan @ is
o. This Oscan 4, as was noticed before (§ 1), represents I.-Eur.
final A of Nom. Sg. of A-stems and Acc. Pl. Neut. of O-stems,
and can hardly have been anything but some form of open O.

21. Descriptions of the O-sound by Latin phoneticians. Terentianus
Maurus distinguishes short from long O (vi. 329. 130-134 K.) :—

igitur sonitum reddere cum voles minori,
retrorsus adactam modice teneto linguam,
rictu neque magno, sat erit patere labra.
at longior alto tragicum sub oris antro
molita rotundis acuit sonum labellis.

This ‘tragic tone in the mouth-cavern’ of 6 is perhaps more applicable to
Greek w, which was open O (Blass, Aussprache des Griechischen®, p. 26), than
Latin 6, and the whole description is possibly, as we have seen, borrowed
from Greek writers on Phonetics. Marius Victorinus (vi. 33. 3-8 K.) summarizes
the older account: 0, ut e, geminum vocis sonum pro condicione temporis
promit ...igitur qui correptum enuntiat, nec magno hiatu labra reserabit,
et retrorsum actam linguam tenebit. longum autem productis labris, rictu
tereti, lingua antro oris pendula sonum tragicum dabit. The commentators
on Donatus (Servius, in Don. p. 421. 17-19 K.) say the same : o productum
quando est, ore sublato vox sonat, ut ‘Roma’; quando correptum, de labris
vox exprimitur, ut ‘rosa’; Sergius, in Don. p. 520. 30-31 0 quando longa est,
intra palatum sonat; ‘Roma,’ ‘orator’; quando brevis est, primis labris
exprimitur : ‘opus,’ ‘rosa.’ Martianus Capella (iii. 261) says merely: O
rotundi oris spiritu comparatur.

§ 22. Close for open o in accented syllables before certain consonant-
groups. (See Schuchardt, Volk. ii. p. 114 sqq.) Before ! and another
consonant 6 became ~ in classical Latin, e.g. consulto (early consolto, C.I. L.
i. 548, latter part of second century B.c.) ; pulcer (but Polc[er], C. I. L. i. 552 of
131 B.c., ef. Prise. i. 27. 12 H.); culpa (Old Lat. colpa Prise. 1.c.). Before m orn
when these nasals are followed by a consonant we see the same tendency.
§§ 21-24.] PRONUNCIATION, VOWELS. 33

The classical spelling is u in umbo, lumbus, wngwis, wncus (see ch. iv. § 20). Before
nd in Vulg. Lat. u replaced classical o (K. Z. xxx. 336), as is shown by the
Romance forms (e.g. Ital. risponde, with close 0, Sard. respundit) ; and in
Italian we have close o in ponte, fronte, fonte, which corresponds with
Priscian’s remark that funtes, frundes, &c., were the older forms retained in
Rustic Latin. (Prisc. i. 26. 35 H. multa praeterea vetustissimi etiam in
principalibus mutabant syllabis; ‘ gungrum’ pro gongrum, ‘cunchin’ pro
conchin, ‘huminem’ pro hominem proferentes, ‘funtes’ pro fontes, unde
Lucretius in libro . . tertio :— :

atque ea nimirum quaecumque Acherunte profundo,

..+ quae tamen a junioribus repudiata sunt quasi rustico more dicta. Of.
Velius Longus p. 49. 15 K. unde in multis etiam nominibus variae sunt
scripturae, ut fontes funtes, frondes frundes; and Charis. p. 130. 29 K. ; sic
ab Ennio est declinatum annalium libro vii; russescunt frundes, non frondes.)
Rumpia is the Latin form of foupaia, the long two-edged sword of the
Thracians, quoted from Ennius by Gell. x. 25. 4, and read in the MSS. of
Livy xxxi. 39. 11. Before rv a close sound of 6, and not the long vowel, is
perhaps indicated by the apex on the o of ornare in some inscriptions (e.g
C. I. L, x. 6104. 1839. 6009) (a fuller discussion of this point in § 145). Greek
«60opvos is cothurnus ; Greek répvos was in Vulgar Latin turnus (so spelt in the
MSS. of Symmachus, Epp. v. 10), e. g. Span. tornar, Ital. torno (with close 0).
The vowel of tornus has been referred to the close sound of Greek o (while w
had the open sound) (K. Z. xxx. 336), and the wu of amurca (Greek duépyy), and
other Greek loanwords in Latin (cf. App. Probi 198. 22 botruus not ‘ butro.’
Cf. Butrio, C. I. L. ii, 668 and Sard. budrone}, might be explained in the same
way. But it is unlikely that the nuances of Greek vowels would be retained
in words naturalized in Latin, and the tendency to give 6 the close sound before
these consonant-groups is visible in genuine Latin words. Perhaps bd is
another group of the kind. Obba was in the time of Nonius (fourth cent.
A.D.) ubba (Non. 146 M. obba, poculi genus, quod nune ubba dicitur).

In Greek Inscriptions we have Movvtavos for Lat. Montinus (C. I. A. iii. 1138,
of 174-8 a.p. ; but usually Movravosy, BovAcaxios, KoupBovAwy (and Kop§-),
Tlovarovjuos (see Eckinger, p. 54). For other examples of o-u before consonant-
groups in the accented syllable, see Georges, Lex. Wortf. s. vv. conchis, dupundius,
formo(n)sus, Corsi, Volscus, proboscis, colostra, bulbus, furnus, fornix, fornax, Fulvius,
triumphus, cochlea; and in the unaccented syllable, s.vv. volsella, to(n)sillae,
promunturium ; also Brambach, Hiilfsb. s. vv. furvus, formica. For classical -wv-
we have O. Lat. -ov- in flovius, elovies (see Georges s.vv.). Curium (for corium)
on the Edict of Diocletian viii. 6 is a strange variety. The Appendix Probi
censures furmica (197. 27 K.). formunsus (198. 9°, detundo (199. 1), purpureticum
marmur (197. 19), a8 well as torma for turma (198. 4 and 28). We have tundunt
on two rustic Calendars (C.I.L. i?. p. 280) for class. tondent. Cf. Sard. tundere).

§ 23. ufor unaccented 6. The mispronunciation pulenta for polenta (Charis.
96. 13 K.; Caper 106. 4 K.) shows this change in the pretonic syllable. Cf.
lulligo for lolligo (Georges s.v), &c. In the post-tonic syllable the change to u
is normal ; see ch. iii. § 18.

§ 24. ufor 6. These two sounds are, as was mentioned above, merged in
the Romance languages. In Late Latin inscriptions the expression of 6 by u
D
34 THE LATIN LANGUAGE. (Chap. IT.

is very common, e.g. patrunus for patrénus. (Cf. App. Probi 197. 28 sobrius
non ‘suber.’) Schuchardt, Yok. ii. p. gt sqq., has collected a large number of
instances from Inscriptions and MSS. (ef. facitud, C.1. L. i. 813’. Forms like
pumilio, pomilio belong to a somewhat different category ; for the original sound
here was ou (cf. pater poumilionom on an old Praonestine cista, Eph. Epigr. i. 20°,
and 6 was a development of ow, in the same way that @ was of ei (see iv. 32).
The same 6 for ou seems to appear in the classical forms robustus, robigo, for
which we have occasional byforms rubustus, rubigo (see Georges s.v., and cf.
Probi Append. 199. 5 K. robigo non rubigo). (This use of 6 and @ for earlier
ou is discussed in ch. iv. § 41.) (Cf. nongentos non ‘ nungentos,’ Bede 281.
26 K.)

§ 25. Other changes of 6 and 6. Curtina, a mispronunciation of cortina
(Dub. Nom. 575 7K.), may follow the analogy of curtus ; faeneris, &c., for faendris,
&e. Vel. Long. 72 and 73 K.) are influenced by généris and the like ; praestélor
and praestilor (Curt. Valerian. ap. Cassiodor. 157. 23 K. ; Alcuin 306. 12K. ;
Bede 286, 19 K.) depend on praestd and praesti ; dstiwn was in Vulg. Lat. astium
(ustei Gen. is found in Mare. Emp. xxviii. 37) (Ital. uscio, O. Span. uzo, Fr.
huis) ; dvum, an egg, was *évum (Ital. uovo, Span. huevo, O. Fr. uef’) ; cérallium
and ciralium are two different forms (Greek «opdAdov and sovpaduov); and
the same must be said of dpilio and tipilio. (The note of Servius on Eel. x. 19
venit et upilio, &c., implies these quantities : propter metrum ait ‘ upilio,’—
nam opilio dicimus—et graeco usus est schemate, sicut illi dicunt otvoya pro
eo quod est dvopa, et otpn pro eo quod est dpy. Cf. Caper 112 K. upilio,
nune opilio.)

§ 26. U, ¥. The Latin grammarians do not speak so much
about the difference of short and long w as they do about % and 3,
perhaps because the first distinction did not so much appeal to the
ear. But in Romance @ and @ take quite different paths, % being
merged in 6, and & preserved, as we found 7 merged in @, and @
preserved, Short # and 6 of Latin are distinguished not only in
Sardinian (the only Romance language which distinguishes Latin
7 and é), but also in Roumanian and in the Latin element of the
Albanian language, though in the two latter % may have first
become close 0, and changed back again to uw. (A. L. L, vii. 61.)
They are distinguished also in Latin loanwords in Welsh.

Latin @ is little altered in Romance, except that in some
countries it has taken a #-sound, in France (lune, for Latin Jina),
a Celtic country, and, perhaps by Greek influence, through the
south-east coast of Italy. Italian #, which seems to retain the sound
of Latin ti, as Italian @ of Latin 7, has the close U-sound of Fr.
sou, Germ. gut, du, while our ‘ two’ is a diphthong ending with
aw-sound, Our short w, e.g. ‘full, ‘put,’ is open U, the German
% of und, Lust, &c. being, according to Sweet (Handb. p. 28),
§§ 25-27.] PRONUNCIATION. VOWELS. 35

rather closer than the English. The ‘ obscure vowel’ of ‘but,’
which is sometimes carelessly spoken of as ‘ short «’ is an entirely
different vowel, not to be called a U-sound at all. In Welsh and
Breton some Latin loanwords show 7 for @, e.g. Bret. dir (Lat.
dirus), but this is probably due to the Celtic tendency to turn
u-sounds into #-sounds (see however K. Z. xxix. 46). Wulg. Lat.
Jinipivus (e.g. Ital. ginepro) for santpérus (Probi Append. 199. 8 K.)
shows the same influence of the palatal spirant j (our y) as Vulg.
Lat. Jenuarius for Januarius (§ 1). But there is little reason to
believe that Latin # had naturally a d-sound. Plautus, Ven. 6 54,
compares the reiterated ¢w tu ‘you! you!’ to the hooting of an
owl :—
Matrona. Tu tu istic inquam. Peniculus. Vin adferri noctuam,
Quae ‘tu tu’ usque dicat tibi? Nam nos iam defessi sumus.

This seems to point to the vo-sound of our ‘ too-whoo,’ though
such comparisons should never have too much stress laid on them.
The palatalization of ¢ before @ is hardly known in the Romance
languages (KX. Z. xxix. 46). The connexion between % and close O
is seen in the numerous misspellings of o for # in plebeian inscrip- °
tions (Schuchardt, Vox. ii. 149, &c.). In unaccented syllables, as
we have seen (§ 14), & tended to the di-sound of optimus, optimus,
which was written ~ to the time of Julius Caesar, afterwards i,
and which in the Romance languages is not distinguished from i.
This was the sound of Greek v, which in older Latin was treated like
Latin u, but afterwards was with more exactness spelt (and pro-
nounced) with the Greek letter Y (y); though in ordinary usage
we often find it, like the #-sound of optimus, represented by ?.
In Oscan #% took after certain letters a yu-sound, e.g. tiurri (Lat.
turrim), Diumpais (Lat. Zuwmpis, lymphis), as in the Boeotian dialect
TUXH Was TLovxa, or in English ‘ tune’ is pronounced ‘tyin ;’ but
there is no trace of this sound in Latin! (see ch. iv. § 7).

§ 27. Descriptions of the U-sound by Latin phoneticians. Ter. Maur.
vi. 329. 142-145 K, :—

hance edere vocem quotiens paramus ore
nitamur ut u dicere, sic citetur ortus:

productius autem coeuntibus labellis
natura soni pressior altius meabit..

 

1 The confusion of Carthaginian miuulec (?) with Lat. mures in Plaut.
Poen. 1009 is no evidence.
D2
36 THE LATIN LANGUAGE. (Chap. II.

Mar. Vict. vi. 33. 8-9 K. u litteram quotiens enuntiamus productis et coeun-
tibus labris efferemus. Martianus Capella iii. 261 U ore constricto labrisque
prominulis exhibetur.

§ 28. Greek v in Latin. Ter. Scaurus says (vii. 25. 13 K.) y litteram
supervacuam latino sermoni putaverunt, quoniam pro illa u cederet. sed cum
quaedam in nostrum sermonem graeca nomina admissa sint, in quibus
evidenter sonus hujus litterae exprimitur, ut ‘hyperbaton’ et ‘hymnus’ et
‘hyacinthus’ et similia, in eisdem hac littera necessario utimur. Y, as a
Greek letter, was not allowed in Roman words (see ch. i.), for the custom
of writing gyla, &c., never gained acceptance (Caper vii. 105. 17 K. y litteram
nulla vox nostra adsciscit. ideo insultabis ‘gylam’ dicentibus. Cf. Bede vii.
273. 33 K.; Ter. Scaur. vii. 22-23 K.; Vel. Longus vii. 81. 5-8 K.; Mar.
Victorin. vi. 33. 11 K.), unless the word was mistaken for a Greek one, e. g.
sylva referred to Greek Ay, lympha to Greek viupn? (Cf. crista non ‘ erysta,
App. Probi 197. 26 K.). The new letter invented by the Emperor Claudius
to express the ii-sound of optimus, optimus is used for Greek v in words like
Nymphius, Bathyllus in the Fasti Antiates written in the reign of Claudius
(C.I. L. i? p. 247). But before the use of the Greek letter ¥, the Greek vowel
was written u (Cassiod. 153. 11 K. Y littera antiqui non semper usi sunt, sed
aliquando loco illius u ponebant: itaque in illorum quidem libris hanc
seripturam observandam censeo, ‘Suriam’ ‘Suracusas’ ‘ sumbola’ ‘sucophan-
tas,’ at in nostris corrumpi non debet ; ef. ibid. 160. 16 K.); and the MSS.
of Plautus indicate such spellings as Hiliria for Illjria, &e. Burrus and
Bruges were the forms used by Ennius for Pyrrhus and Phrijges (Cicero, Orator
xlviii. 160 ipsius antiqui declarant libri). That it was also pronounced like
ordinary Latin w we see from the Romance forms of these earlier Greek loan-
words which make no distinction (e.g. Lat. twmba for Greek rp Bos, is in Ital.
tomba, in Sard. tumba, in Fr. tombe), not to speak of Plautus’ pun on Lydus
and lidus (Bacch. 129), and on chrystilus and criicisdlus (ib. 362). After the
ti-sound of optumus, optimus came to be spelt with 7, the same letter was in
ordinary usage employed for Greek v, e.g. cignus (Greek «v«vos), in Ital.
cecero, being pronounced probably in the same way as the ¢ of optimus, which
in Romance is not distinguishable from ordinary % Tondrus for Tyndareus on
an old Praenestine cista (C. I. L. xiv. 4109) is perhaps to be explained by the
u-sound of o before nd (ch. iv. § 20). Greek xv is often spelt qui, e.g. Vulg.
quiatus for cyathus (see Schuch. Vok. ii. p. 273 sqq. for examples), as Latin qui
ig often expressed by Greek «uv; e. g.’Axvdas for Aquila, Kupevos and Kupwos
for Quirinus on Greek inscriptions (see Eckinger, p. 123). Oe is found for Greek
B in goerus, coloephia, byforms of gyrus, cilyphia, &c. (see Georges 8. vv., and
Schuch. ii. 278). Latin @ is in Greek inscriptions always expressed by o till
the beginning of the Empire, when ov takes its place. We find v especially
in the suffixes -ullus, -ulus, -wrius, &e. (Eckinger, p. 58 sqq.) Sulla, Sylla for
Syrula (§ 16) is always SvAAas.

(For spellings of Greek v and the Latin i-sound with y, u, i, see
Schuchardt’s examples from inscriptions and MSS., Yok. ii. p. 218 sqq.,
and consult Georges, Lex. Wortf. s. vv. cumba, murra, myrtetum, myrtum, lympha,

 

1 Varro Men. 50 B. makes the Gen.  phaticus’ arrepticius G1. Sangall. 912 ;
Plur. lymphon. In Glossarieswe have Ambr. B. 31 supr.
nymphaticus for lymphaticus, e.g. ‘nym-
§§ 28-32. ] PRONUNCIATION. DIPHTHONGS. 37

murmillo, Thynia, Syrus, serpyllum, and Brambach, Hiilfsbiichlein s. vv. thynnus,
syllaba, stilus. The Appendix Probi has: tymum non ‘tumum’ (199. 6);
myrta non ‘murta’ (199. 7); Marsyas non ‘ Marsuas’ (197. 24) ; elamys non
‘elamus’ (198. 20); gyrus non ‘girus’ (197. 27) ; Byzacenus non ‘ Bizacinus’
(198. 1) ; amygdala non ‘amidduda’ (198. 26).)

§ 29. ofort. The coincidence of Latin 6 and % in the Romance languages
makes it natural that we should find o written for % on late inscriptions, and
in plebeian forms. Roman tiles, for example, from the figlina Bucconiana are
in the earlier period marked Bucconiana, but from Diocletian’s time often
Boconiana (C.I.L. xv. p. 386); and Greek azvpag¢ appears in late Latin as
storax (Georges 8. v.). [In addition to the large number of instances of o for %
collected by Schuchardt, Vok. ii. p. 149 &¢., see Georges, Lex. Wortf. s. vv.
columna, urceolus, cunnus, luxurio, verecundus, and ef. App. Probi 198. 23 puella
non ‘poella’; 198. 12 cluaca non ‘cloaca’: 197. 25 columna non ‘colomna’
(on the last example, see § 68 and ch. iii. § 33.)]

§ 30. 6 for a. This interchange, as we have seen (§ 24), is properly
confined to words which had originally the diphthong ou, which became in
Latin a sound expressed variously by 6 and by @ (ch. iv. § 41). Some examples of
the interchange have been mentioned in § 24. To them may be added bocula,
occasionally in MSS. of Virgil for biciila (Ribbeck, Index, p. 391), jocundus for
Jicundus (Georges s.v.), and the examples (many of doubtful worth) collected
by Schuchardt, Vok. ii. p. 181 sqq.

§ 31. Other changes of ti and U. Codliiber was in Vulg. Lat. colober, the u
being assimilated to the accented 0 (see ch. iii. § 33). Hence Vulg. Lat. colobra,
with open o accented before br (cf. ch. iii. § rz), (Sicil. culovria, Span. culebra,
O. Fr. culuevre) ; cf. Append. Probi 199. 2 K. coluber non colober : so colober
on inserr. e.g. Mur. 1144. 3, and in MSS. (Schuch. ii. 149) ; niirus was norus
(see Georges s.v.) or rather *nora (ef. App. Probi 198. 34 nurus non ‘nura”’),
with open o (Ital. nuora, Span. nuera), perhaps by analogy of séror. Liridus
was *hir(i)dus (Ital. lordo, Fr. lourd) ; piimex shows *piim-, in Ital. pomice,
Span. pomez, Fr. ponce) ; niiptiae was *noptia (Ital. nozze, Fr. noces), explained
by analogy of nova nupta (?). Upilio and dpilio, curalium and ciralliwm were
explained in § 25. Auriigo and aurigo are due to interchange of suffixes, not
to transition of vowel-sound. (So gravido and grdvédo, § 19.)

A curious tendency to interchange u-i and i-u appears in Vulg. Lat. stipila
(seen in Ital. stoppia, O. Fr. estoble, Fr. éteule, &c.), unless stup- and stip- are
original byforms. (On mitulus and mytilus, see Brambach, Hiilfsbiichl. s.v., ef.
Gk. Mur:Aqvy and MervAnvy, Lat. Utica and Greek “I7vien.)

§ 82. Diphthongs. We have no reason to doubt that Latin
au, ae were in the classical period, and for some time after,
diphthongal sounds. None of the grammarians who discuss
these diphthongs suggests that they were anything else. But
in various dialects of Italy aw had been early reduced to a single
sound 0, ae to a single sound e¢, a dialectal or ‘rustic’ pronuncia-
tion which shows signs of its presence in the speech of everyday
life. The Romance languages indicate that in Vulgar Latin ae
38 THE LATIN LANGUAGE. (Chap. II.

had become hardly distinguishable from an open E-sound ; and
the reiterated warnings of grammarians, from the fourth cent.
A.D. onwards, against the confusion of words like aeguus and
equus tell the same story. Welsh praidd (Latin praeda, for
*prae-héda or *prae-hida, from prehendo, prae-hendo) must have
been borrowed before this decay set in; but Varro’s use of ae,
instead of 2, to express the sound of Greek 7 (probably open e),
in scaena, seems to show that the process of development had at
least begun before the Imperial Age. On the other hand, au has
been preserved intact by several of the Romance languages; and
in the others (e.g. Italian and French), where it has developed to
o (Ital. cosa, poco, &c., with open 0; Fr. chose), this development
can be proved to be post-Roman. In Latin loanwords in Welsh
we have sometimes az, e.g. aur (Lat. aurum), sometimes close o.
The pronunciation of these diphthongs must have been a com-
bination of the simple sounds of which they are composed; az,
an a rapidly followed by a w (or 0), something like German au ;
ae, an a rapidly followed by an e, something like Welsh ae; but
how modified from century to century, it is impossible to say.
Tn (originally) unaccented syllables in compounds, au was reduced
to @ (through ew ?), e.g. défrido, from fraudo (see ch. il. § 18)
(cf. Ital. udire, from Lat. audire) ; ae, or rather the earlier az, to 7
(through ¢2?), e.g. distisum, from taedeo (ibid.) ; but in later Latin
the reduction was seldom carried out (ch. iii. § 23). The inter-
jection au, only used by women, seems to have been a cry
expressing wonder or indignation, e.g. Ter. Adelph. 336 au, au,
mi homo, sanushe es? while the diphthong ae oceurs in several
exclamations, such as vac (the Lettish wai), Jakae and hahahae, &e.
(Cf. banbari to bark.)

Ae had been in early times ai; and this old spelling was often
used by lovers of antiquity in the Imperial period, though the
pronunciation was of course ae, and not az. A curious feature of
Vulgar Latin, reflected in Romance, was the substitution of a for
au in syllables before the accent, when the next syllable contained
the vowel uw, e.g. Agustus, found on Inscriptions for Augustus.
(Ital. agosto.) The same tendency is shown in the Sardinian
dialect of Italian, where Lat. /aurus is laru, &ec., and in our
‘laughter’ av has an a-sound.
§ 33.] PRONUNCIATION. DIPHTHONGS. 39

Oz, a diphthong used in early times, had been reduced first to
oe, then to a simple sound # (through some é-sound probably)
before the classical period. What was the exact sound of the
later diphthong oe, which we find in words like coetus (from
co(m)-itus), is difficult to determine ; and the small number of
words which possessed this diphthong makes it impossible to
ascertain its treatment in Romance. We have already seen
(§ 28) that it is occasionally found as an expression of Greek
v, e.g. goerus (beside girus), coloephia (beside céljphia). The
interjection oiei was a ery of pain. Thus in Plautus, Mil, 1406,
when the soldier is being thrashed, he shouts: oiei, satis sum
verberatus ; and in Terence, Phorm. 663, the miserly father, hear-
ing of the large sum demanded by the parasite, cries out, as if he
had received a blow: oiei, nimium est.

£u is another diphthong, which arose at a later period through
fortuitous combination, e.g. xéuter, a trisyllable (Consentius p. 389.
28 K.) (from vé and iter, with the accent on the né, § 149), sew (so
neu, ceu, ch. x. § 16 and 11) (by reduction of s?-ve, sez-ve, ch. x.
§ 4); while I.-Eur. ew was, like I.-Eur. ow, in the Italic languages
ow, a diphthong found in early Latin, but reduced to « (as I.-Eur.
ei to 2) by the second cent. B.c. (ch. iv. § 26). Latin ew of the
Interjection “eu (cf. Greek ¢e6) must have been pronounced like
e followed rapidly by « (or 0); for a fifth century grammarian
(Agroecius 122. 11-16 K.) dwells on the distinction between eo,
cho, and heu. Greek ev seems to have been usually pronounced
as a disyllable in Latin. Ui, which can hardly claim to rank as
a Latin diphthong, is seen in the interjection 4w, where it may
express the sound of a whistle, our ‘whew!’ and in the Dative
cui, which does not seem to have much differed in pronunciation
from the Nominative gzé.

§ 33. Grammarians’ account of diphthongs. Nigidius (first cent. B.c.),
ap. Gell, xix. 14.6 a et o semper principes sunt, i et u semper subditae, e et
subit et praeit ; praeit in ‘ Euripo,’ subit in ‘ Aemilio’; Ter. Scaurus (second
cent. a. D.) vii. 16. 5 K.a igitur littera praeposita est...e litter(ae) ... et
apud antiquos i littera pro ea scribebatur, . . ut ‘pictai vestis,’ et ‘aulai
medio’... sed magis in illis e novissima sonat (cf. Quint. i, 7. 18; Marius
Victorinus fourth cent. a. D.) vi. 32. 4-6 K. duae inter se vocales jugatae
ae sub unius vocis enuntiatione prolatae syllabam faciunt natura longam,

quam Graeci diphthongon vocant, veluti geminae vocis unum sonum, ut ae,
oe, au; ef. Ter. Maur. vi. 338. 418-427 K. and 365. 1326-1334; [Probus] de wll,
40 THE LATIN LANGUAGE. (Chap. II.

syll. iv. 219. 25 K.; Servius, in Don. iv. 423. 30 K.; Mallius Theodorus vi. 586.
25-26 K. ; Bede vii. 229. 20-25 K.

§ 34. Ter. Maurus on au. Terentianus Maurus makes a distinction between
Latin dw and du, while Latin eu, he says, is like Greek ev always éu. Au, as in
‘aut Agé,’ ‘aut tbi,’ ‘Aurunci’ of Virgil, he compares to Homer’s avépucay
(presumably dFépucay) and drdp for airdp, in contrast to the (accented) du of
aurum, auspices, Greek atjproy :—

“aut age’ inquit ille vates, saepe dixit ‘aut ubi’
dixit ‘ Aurunci,’ quod aeque barbarum est producere:
pes ubique lege constat, prima cum correpta sit,
consonans et una plenum non queat tempus dare

avépuoay inquit poeta sic et abrdp corripit.
If this means anything, which is doubtful, it ought to mean that in ‘ aut age,
‘aut ubdi’ the diphthong had a more reduced sound than the au of awrum, a reduc:
tion which was similar to that seen in pretonic au followed by a syllable with
u, Aruncus like Agustus (Cf. arvncrto, C. I. L. vi. 13416 ; Arunciin Virgil MSS.,
&e.; Ribbeck, Ind. p. 388.)

§ 35. auin Romance. Had au been an open O-sound in Vulg. Lat. it would
have been merged in Latin 6, has ae as been merged in & But that the o of
Ital. poco, Fr. chose, is a late development we see from the forms of the words,
which would otherwise have been *pogo, *cose (Meyer-Liibke, Rom. Gram. i.
p. 235). We may similarly infer the diphthongal character of au, ae, at the
time of the Empire, from the frequently-repeated statement of the gram-
marians that after a diphthong it was impossible to pronounce a double
consonant, e.g. paulum (not paullum), Paulus (usually spelt Pauilus, but not
so pronounced), while after a long vowel double | was common, e.g. séélla,
Polla, villum (Diminutive of vinum) (see § 127).

§ 36. u for accented au. The change of au to a in the (originally) unaccented
syllable is in conformity with the rule in défriido, the spelling recommended
for Plautus and Terence on the strength of the MSS. by Ritschl (Parerga, i.
540) (see also Georges s.v.). But we find also sed frude in the Lex Repetun-
darum (123-122 B.c.) (C.I.L. i. 198, § 64) (but sed fraude, 5 69), where there
seems no reason for supposing the syllable to have been unaccented [Another
instance of the confusion of au and % in this word is the spelling fraustra, often
found in MSS. of Virgil (see Ribbeck, Index s.v.), which also exhibit frude for
Jraude in A. iv. 675, as the MSS. of Lucretius have frudem ii. 187, frudi vi. 186
see Lachm. p. 85)). Similarly the u for unaccented aw which appears
regularly in the compounds of claudo, seems to have called into life a byform
of the simple verb, cludo, in the first cent. a.p. (Georges, Lex. Lat. Wortf.
s.v. claudo, and p. 750), which remains in the Italian chiudo. (For cludam,
lame, in Plaut. Pseud. 659, read with the palimpsest claudam.)

§ 37. oand au. o for aw is a feature of the Umbrian language (e. g. ote,
Lat. aut) and other dialects, and was preserved in ‘rustic’ Latin, and even
in the Latin of the streets of Rome. Festus tells us of a millionaire who was -
nicknamed Orata (i.e. aurata, goldfish), because he wore two gold ear-rings
(Festus 202. 13 Th. orata, genus piscis, appellatur a colore auri quod rustici
‘orum’ dicebant, ut auriculas ‘oriculas,’ itaque Sergium quoque quendam
§§ 34-38.] PRONUNCIATION. DIPHTHONGS. 41

praedivitem ... Oratam dicunt esse appellatum, &c. (For oricula, ef. App.
Probi 198. 11 auris non ‘oricla.’ Oricla occurs as a eognomen on inscriptions,
C. I. L, xii. 5686, no. 652.)] Cicero’s rival Clodius, was the first of the gens
to change the name Claudius to the plebeian form Clodius, no doubt with the
view of conciliating the mob. Cicero himself in his letters often uses the
more homely forms with 0, e.g. loreolam (Att. v. 20. 4), pollulum (Fam, xii. 12. 2;
oricula (Quint. Fr. ii. 13. 4), like oricilla, Catull. xxv. 2 (see A. L. L. vi. 84), while
plodo is quoted from his ‘ De Gloria’ by Diomede (p. 382. 26 K.), and in Plautus
we seem to find assonance of aurum with ornamentum, ornatus, of auspictum with
omen, of auribus with oculus (Bursian’s Jahresbericht, 1881, p.33). So too Priscian
Gi. 52, p. 39 H.) says: (au) transit in ¢ productam more antiquo, ut ‘lotus’
pro lautus, ‘ plostrum’ pro plaustrum, ‘cotes’ pro cautes: sicut etiam pro 0,
au, ut ‘austrum’ pro ostrum, ‘ausculum’ pro osculum, frequentissime hoe
faciebant antiqui. This usage of au for o [cf. Paul. Fest. 21 (apparently referring
to a passage of Plautus) ausculari dicebant antiqui pro osculari] is found in
Plautus, not merely in aurichalcwm (Greek épefxadxos), where it is due to con-
fusion with aurum, but also in ausciilatur (Bacch. 897, &e.). Aula (or aula as in the
palimpsest) of Plautus became olla, as Paulla, Paula became Polla. It is perhaps
confined to derivatives of ds (see Georges s.vv. oreae, ostium, osculum, osculor), which
seems to have had two parallel stems in early Latin, aus. and ds- (cf. jécwr and
jocur) ; so this gives no evidence on the pronunciation of Lat. 6. In the Lex
Metalli Vipascensis of the first cent. a. p. (Eph. Epigr. iii. p. 180) we have scauria
for the Greek cxwpia, which the Romance languages show us to have been
scdria in Vulgar Latin. Rustic or dialectal o for aw is found in the name
M. Lornti (=M. Laurenti), on a jar in the old Esquiline cemetery (c. 200 B. c.)
(Ann. Inst. 1880, p. 260), while on plebeian inscriptions we have such forms
as Oli (for Auli) on the tombstone of a praeco (Eph, Epigr. iv. p. 297),
Olipor (C. I. L. xi. 1973), &c. In Greek inscriptions we have ‘Acs from the
time of Augustus, but always Tavados (though often MwAda and HoAaa, like
Lat. Pola). (See Eckinger, p. 13.) In cauda (Lith. kéidas) the original vowel
may be 4, and the spelling au be due to the similarity of sound between 6 and
au. (See K. Z. xxviii. 157 for this and other doubtful instances.) [For other
examples of au-o, see Schuchardt, Vok. ii. p. 301 sqq., and Georges, Lex. Wortf.
S.VV. caupo, auspicor (Diom. 383. 10 K. Claudius octavo Historiarum ‘Flacco
ospicatur ’), caulis, cauwliculus, caurus, raudus (also rudus), paused, lawretwm, plaudo,
claudus, claustrum, sorix (cf. Mar. Vict. 26. 7 K. sorix vel saurix, C. GL.
Vv. 242. 33), codex.] (See Diomedes, pp. 382-3 K., Probus Inst. 118-9 K.)
Suetonius (Vesp. viii. 22) tells us an anecdote of the homely Vespasian :
Mestrium Florum consularem, admonitus ab eo plaustra potius quam plostra
dicenda, postero die ‘ Flaurum’ salutavit [cf. the glosses: plostrum dicimus
magis quam ‘ plaustrum’ (C.@.L. v. 93. 13), and: ‘odit’ audit (ibid. 89. 7
and 125. 26).]

§ 38. afor au. (Schuchardt, Yok. ii. p. 305 sqq.) Agustus for Augustus, e. g.
C. 1. L, ix. 1365 (411 a. D.) (ef. Greek ’Ayouaradios, Mitt. Inst. xiii. p. 236 7. 5;
Eckinger, p. 12). The Romance name of the month points to Vulg. Lat.
Agustus, e.g. Ital. agosto, Span. agosto, Fr. aout, and shows that the a was
not merely a conventional symbol for 0. Asculio was the Vulg. Lat. form
of qusculto (cf, Caper 108. 6 ausculta non ‘asculta’), as we see from the
Romance forms, e. g. Ital. ascoltare, Span. ascuchar ; *agurium of augurium
(Raet. far agur, to consider, Ital. sciagurato, from *exaguratus, unlucky,
42 THE LATIN LANGUAGE. (Chap. II.

Span. jauro). Cladius often occurs for Claudius on inscriptions (eg. C. I. L.
li. 4638, of 275 a.v.) (ef. Greek @acros, C.I.A. iii. 10, of 209-210 A.D. ;
Bull. viii. p. 247, of 11 a. D., from Eumenia). For similar spellings in Glossaries
(e. g. ‘agustae’ sanctae ; fastus for faustus, &e.), see Léwe, Prodr. p. 421. In
MSS. of Virgil, &¢., we find Arunci for Aurunci (Ribbeck, Ind. p. 388, ef.
Arunceio, C. I. L. vi. 13416); and modern Italian place-names like Metaro,
Pesaro show a similar change.

§ 39. Greek transcriptions of au. In Greek inscriptions we find usually
av for Latin au ; but also ao, e. g. BaogTw, C.I.L. ix. 6229 and 6230 ; @acarwes
6209 (the form Haodos does not occur till the fourth or fifth cent. a. D.) ;
also aov, e.g. TlaovAdwva, C.1.G. 6665; AovAov (2656 b add.) (see Eckinger,
p. 13).

§ 40. ae for au. de is found now and then on inscriptions for ay, e. g.
maeso(leum), C. I. L. i. Fast. min. ix of 1 a.p.; Paelinus, &e.

§ 41. efor ae. (Brambach, Orthogr. p. 205; Schuchardt, Vok. i. p. 224 sqq.)
E for ae (ai) is a feature of the Umbrian language, e. g. pre (Lat. prae), and is
found on Latin inscriptions in the Umbrian territory, e.g. Cesula, 0. I. L.
i, 168 (Pisaurum), and elsewhere (see Sittl, Lok. Verschied. p. 4). It was
a feature too of rustic Latin, as we see from Varro, L. L. vii. 96 rustici pappum
‘Mesium,’ non Maesium ; v. 97 in Latio rure ‘edus’; qui in urbe, ut in
multis, a addito aedus; and from Lucilius’ ridicule of a praetor who called
himself Cecilius instead of Caecilius (ix. 10 M. Cecilius pretor ne rusticus fiat.
Cf. Diom. 452. 17 K.). The same variation of ¢ and ae found its way into
ordinary pronunciation in the case of country-terms, e. g. faenisicia and fenisicia,
the hay-harvest. The Romance forms point to sépes, not saepes ; séptum, not
sdeptum (e.g. Port. sebe, Span. seto) (Gréber, A. L. L. v. 465). From Varro’s
remark that scaena (and scaeptrum ?) represented the pronunciation of Greek
oxnv;, oxqmTpov at his time, we should infer that this ae had a sound
approaching to long open e, for Greek 7 probably still had at this period the
open sound (Varro, L. L. vii. 96 obscaenum dictum ab scaena ; eam ut Graeci
Accius scribit ‘scena.’ In pluribus verbis a ante e alii ponunt, alii non, ut
quod partim dicunt ‘scaeptrum,’ partim ‘sceptrum,’ alii Plauti ‘Faenera-
tricem,’ alii ‘Feneratricem’; sic ‘faenisicia’ ac ‘fenisicia’). This spelling
of the title of a play of Plautus, Feneratrix, for Faenérdtrix, the Usuress, agrees

| with another remark of Varro that fenus, not faenus, was the pronunciation of
Old Latin, used by Cato and others (Non. 54 M.; Varro lib. iii de sermone
Latino: ‘faenus autem dictum a fetu, et quasi fetura quadam pecuniae.
Nam et Catonem et ceteros antiquiores sine a littera ‘fenus’ pronuntiasse
contendit, ut fetus et fecunditas). How thoroughly ae (through we?) became
identified with the long sound of open e¢ at a later time, we see from the
remark of a fifth century grammarian, that équus, when the first syllable,
through being accented, was unduly lengthened in pronunciation, became
aequus (Pompeius 285. 6K. plerumque male pronuntiamus, et facimus vitium
ut brevis syllaba longo tractu sonet. .. si (quis) velit dicere ‘aequus’ pro eo
quod est equus). (Cf. prehendo, with shortening of prae before a vowel, as
deamo of dé, and Marius Victorinus’ use of -aeus to express the disyllabic
pronunciation of Gk. -evs (67K...) Another grammarian of the same century
gives a caution against the confusion of vae and vé (Agroecius 114. 21 K.>, of
quaeritur and quéritur (id. 116. 18 K.); while he speaks of the first syllable of
§§ 39-43. ] PRONUNCIATION. DIPHTHONGS. 43

praemium, pritium, précor, as if they were distinguished in writing only, not in
pronunciation (id. 115 K. praemium cum diphthongo scribendum ; pretium,
precor sine diphthongo. Veteres enim majoris rei sermones cum diphthongo,
et quadam dignitate scribi voluerunt). Even in the fourth cent. Servius, in
a note on Virgil, Aen. i. 344:—

huic conjux Sychaeus erat, ditissimus agri
Phoenicum, et magno miserae dilectus amore,

thinks it necessary to point out that miserae is the Adjective, not the Adverb
miseré. The ‘Orthographies’ of Bede and Alcuin (the latter served as
a text-book for Carlovingian scribes of MSS.) abound in similar distinctions
(e.g. quaeritur and queritur, Alcuin 308. 16 K.; Bede 287. 8 K.; quaestus and questus,
Alcuin 308. 17 K.; saevitand sevit, Aleuin 310. 5 K.; Bede 289. 30 K.; caelo and celo,
Alcuin 299. 6K. ; Bede 268. 27 K.), some of which may have been taken from
earlier grammarians [ct. Charisius (fourth cent.), p. 98 K. on the spelling erwmna
for aerumna ; Marius Victorinus (fourth cent.), p. 25 K. on the spelling cesaries
for caesaries]. Philargyrius, the Virgil Scholiast, on Ecl. iii. 39, defends the ae
of haedera (for hédéra) by connecting the word with haereo (ef. Paul. Fest. 71. 26
Th. hedera dicta, quod haereat, sive quod edita petat, vel quia id, cui
adhaeserit, edit.). With all this it is no wonder that it is often difficult to
decide whether the proper spelling of a word is with ae ore. [For the rival
claims of e and ae in some words, see Georges, Lex. Wortf. s. vv. meles, nenia,
gleba, magna, muraena, paelex, feles, cetra, ne (the Interjection), gaeswm. |

On Greek inscriptions we find ¢ for Latin ae from the middle of the secorid
cent. a.D., e.g. KexAros, but never 7. (Eckinger, p. 78.) Instances of Latin
ae for Greek 7 in inscriptions and MSS. are given by Schuchardt, Vok. i. p. 227
sqq., a very frequent case being that Genitive ending of female names in -aes
(Greek -7s) from the last century of the Republic, e. g. Laudicaes (C. I. L.i, 1212,
which is discussed in ch. vi. § 18.

§ 42. ai for ae. The old spelling a is found on Imperial inscriptions,
especially in the reign of the grammarian-emperor Claudius (e. g. C. I. L. vi.
353, of 51 A.D., Caisare}; but we have the express testimony of Terentius
Scaurus (second cent.) (16. 7 K. sed magis in illis e novissima sonat), not
to speak of Quintilian (first cent.) (i. 7. 18 cujus secundam nunc e litteram
ponimus), that the second element, as pronounced, was e, not 7. The change
of the earlier ai to the classical form of the diphthong, ae, took place in the
second cent. B.c. (e.g. aedem, beside aiquom, tabelai, datai, &e., on the 8. C.
Bacch. of 186 B.c., 0. L.L. i. 196). The spelling aei, found once or twice
towards the end of the second cent. B.c., e.g. conquacisivet, Caeicilius, Caeician[us],
may mark the transition (see ch. iv. § 29). In ain for aisne, aibat (disyll.) the
diphthong must have had the sound of O. Lat. ai.

§ 43. Greek «. Before a consonant Greek « is always 7 in Latin, e. g.
Atrides. Before a vowel it is @ till the first cent. a.p., then 7 Thus
Alexandréa, Daréus, &c., are the earlier spellings ; Alexandria, Durius the later.
(See Brambach, Hiilfsbiichlein, p. 4.) The -é-, -7- was often shortened (cf. § 143).
e.g. balntum (Gk. Badavetov) (cf. Prise. i. p. 71 H. and p. 73 H. on Alphéus,
Hectortus, &c.). The Greek diphthong which probably passed into the 7-sound
about roo Bc. is a common expression of Latin i, e.g. “Avrwveivos, but of
Latin 7 only in hiatus, e.g. drpetov (for atrium), MoumAecos (for Publius) (see
44 THE LATIN LANGUAGE. [Chap. IT.

Eckinger, p. 42). Latin e in words like Pompeius is in Greek me; but from
the first century a. p. we find also «, e. g. Wovmeos (Eckinger, p. 81).

§ 44. oe and e. Alcuin and Bede give almost as many rules for the
distinction of oe and e, as for the distinction of ae and e [e. g. cepit and coepit,
coepta and incepta, Alcuin vii. 299. 18 K. ; Bede vii. 269. 14 K. ; fedus (quod est
deformis) and foedus, Alcuin vii. 301-302. 2 K.; Bede vii. 273. 4 K.; cf.
Orthogr. Bern. 293. 9 K. ; pene, penes, and poena, Aleuin vii. 306. 35 K. ; Bede
vii. 286. 1 K.], some of which may come from earlier grammarians. [For
variations in spelling between oe, e, ae see Schuchardt, Vok. ii. p. 288 sqq.,
and consult Georges and Brambach s. vv. cena, caenum, faeteo (ef. Span. hiede),
amoenus, fenus, maereo, paene, proeliwm, caelebs, caelum, caecus, oboedio, JSoedus,
fecundus, obscenus, pomoerium, femina, fetus, &c.] Greek is in classical Latin 6,
e. g. melodia, but earlier oe, e.g. comoeds (ef. Thraex and Thrax, Blass, p. 43).

§ 45. oe in Romance. *péna for poena is indicated by the Romance forms
(e.g. Ital. pena, Span. pena, Fr. peine), and probably */edus for foedus, foul
(e. g. Span. hedo, feo). Céna (e. g. Ital. cena, Span. cena) is thought to have
been the correct spelling (cf. Osc. kersna-), though the spelling with oe (due
to confusion with Greek xo.wds, as coelum, for caelum, confused with xoidos) is
very old (cozn- on » Praenestine cista. Mél. Arch. 1890, p. 303).

§ 46. Greek ev. Marius Victorinus vi. 66-67 K. consimili ratione quaeritur,
Orpheus in metro, ut

non me carminibus vineat nec Thracius Orpheus,

utrum trisyllabum an disyllabum sit, an idem nomen duplici enuntiatione
promatur, aut sine a littera, ut Peleus Pentheus, aut cum a, ut ita declinetur
Orphaeus, ut Aristaeus. visum est tamen hoc posse discerni, ut illa sine a
littera graeca sit enuntiatio, haec latina, quae per diphthongon effertur. The
proper spelling is ewhoe, Euhius, euhan, not evoe, Evius, evan (see Brambach,
Hiilfsbiichlein s. vy.). For a corruption of Greek ev in vulgar pronunciation,
see App. Probi r99. 6 hermeneumata non ‘erminomata.’ On an old mirror
of Praeneste we have Taseos (Tasei, Gen.) for @dcebs (Eph. Epigr. i. 23).

§ 47. ui of cui. Quint. i. 7. 27 illud nune melius, quod ‘ cui’ tribus quas
posui litteris enotamus, in quo pueris nobis ad pinguem sane sonum qu et oi
utebantur, tantum ut ab illo ‘qui’ distingueretur. Ter. Scaur. 28 1 K. c
autem in dativo ponimus, ut sit differentia cui et qui, id est dativi [et
vocativi] singularis et nominativi et vocativi pluralis. Annaeus Cornutus
ap. Cassiod. 149. 8 K. ‘qui’ syllaba per q u i scribitur ; si dividitur, ut sit
cui et huic, per ¢.

§ 48. J,V. That Latin 7 and v had some sound like our y, w,
and not like our j, v, there can be no doubt whatever. We see
this from the close relation that exists between 7 and j, wu and v
in different forms of the same word in Latin, e.g. jam and
nunciam (3 syll.), tenuis and tenvia (3 syll.), as well as from the
express testimony of grammarians. The signs j and v, which
suggest to us a difference between the sound of these letters and
§§ 44-48.] PRONUNCIATION. 45

of the corresponding vowels, are, as we saw (ch. i. § 7), of quite a
late date. In the Roman period jus and vos were written with i
and w, us, wos, exactly like pius and twos (ch. i. § 1). The only
question is whether j and v were actual consonants (y, #) or half-
vowels (2, #). Our y, for example, in ‘ you’ is a spirant consonant,
but is often in pronunciation weakened into a half-vowel (Sweet,
Handh. p. 37). The distinction is so slight a one that it seems im-
possible to determine the exact pronunciation of j and v in a dead
language like Latin; and probably the pronunciation varied at
different times; but we certainly have one or two clear testi-
monies to the consonantal character of these sounds. Thus a
fifth (?) century grammarian remarks on the difficulty experienced
by the Greeks of his day (as by the Greeks of our own time) in
pronouncing this y-sound in such a word as jus. They make the
word, he says, almost a disyllable (Consentius v. 394 K.). And
the same writer in another passage mentions a corresponding
mispronunciation of the w-sound in veni (v. 395.15 K. u quoque
litteram aliqui pinguius ecferunt, ut, cum dicunt ‘veni,’ putes
trisyllabum incipere). And much earlier, in the first cent. a.D.,
we have a distinction drawn between v of valente, primitivo,
&e., and the ~ of guis. The former is said to sound ‘cum aliqua
aspiratione’ (Velius Longus vii. 58. 17 K.), much as Varro, the
contemporary of Cicero, says that v- had a strong thick sound
(crassum et quasi validum) in vafer, velum, vinum, vomis, vulnus,
&e, (LZ. L. iii. fr. p. 148 Wilm.). That this consonantal character
of j and » intensified and developed itself as the centuries went
on, we see from the Romance languages, e.g. Italian, where
Latin v has become our v-, Latin j our j-sound (e.g. vostro, Lat.
voster, giurare, Lat. jardre); and it is possible that the further
back we go in the history of the Latin language the less con-
sonantal was the sound of j and v!. But there is no evidence of
this, unless we count as such the scansion dicio in Plautus and
Terence, where the 7 appears to be so entirely sunk in the follow-
ing 7 (e), as not to make the first syllable long by position,

 

1 [,-Eur. swé- became Latin sd-, indicate that the v of sve- was more
e.g. soror (through *svo-?), but not consonantal than ordinary r.
L.-Eur. wé-, e.g. vetus, which may
46 THE LATIN LANGUAGE. (Chap. II.

whereas in dlicio of the classical poets the first syllable is so
lengthened (4. L. L. iv. 560) (but see ch. il. § 25). This merging
of j in a following 7 has been compared with the merging of the
uw of gu, which Velius Longus tells us was more vocalic than the
v of valente, &e., in a following wu, e.g. cum (earlier quom) for
quum, locuntur (earlier loguontur) for loguuntur (cf. § 93). A similar
unconsonantal character for v in early Latin has been inferred
from the reduction of gvé to a in such Plautine forms as odléscor
for obliviscor, dinus for divinus; but in the absence of express
testimony, such as we have for the consonantal character of j
and v at a later time, it is impossible to decide positively so
minute a point. An untrained ear can hardly distinguish between
the spirants y, vw, and the half-vowels /, y, nor yet between the
various nuances of the z-sound, such as our wf, e.g. ‘which,’ the
unvoiced w, differing from the voiced w of ‘witch’ as p from 4,
¢ from d, ec from g, or such as in French ‘oui,’ the consonant of
the vowel of French ‘sou,’ while our w is the consonant of the
vowel of English ‘ full,’ ‘put’ (Sweet, Handbh. p. 42). So much
we can say, that the pronunciation of j and v certainly became
more and more removed from the half-vowels in the centuries of
the Empire; and it is natural to infer a movement in the same
direction in the Republican period. But when exactly 7 and v
ceased to be half-vowels and became consonants, or how far their
character varied according to their position in the word it is
impossible to determine with precision.

The same tendency to syncopate a short unaccented syllable
that produced caida out of ciélida affected the vowels 7, « when
they preceded other vowels. The word /drua is a trisyllable in
Plautus ; it has become a dissyllable in classical Latin, just as
livtdum of Plautus became ddrdum. The only forms known to
Plautus are miluos, rélicuos, gratiis, which in classical Latin
are milvos, reliquos (by the middle of the first cent. a.D. relicus),
gratis. In the first cent. a.p. ¢énuis wavered between a dissyl-
lable and a trisyllable (Caesellius ap. Cassiod. vii. 205 K.).

This reduction of the vowel 2 after ¢, ¢ led, as we shall see
($§ 90, 94), to the assibilation of these consonants. Titius
became * 7%¢yus and then something like *7itsus (cf. our ‘ orchard’
for ‘ort-yard’), as we learn from the remark of a fifth cent.
§ 49.] PRONUNCIATION. J, V. 47

grammarian, quoted below: si dicas ‘ Titius’ (i) pinguius sonat
et perdit sonum suum et accipit sibilum.

As to the pronunciation of words like A/aia, Pompeius, ejus,
where the diphthong is followed by a vowel, we have very clear
information from the grammarians that the 2-sound was shared
both by the first and the second syllable, Mai-ia or Mai-ya, not
Ma-ia, Ma-ya. To express this sound Cicero proposed to write
Hatia, diaz with two 7’s (Velius Longus vii. 54.16 K.; Quint.
1, 4. 11); and on inscriptions we find spellings like martlorEm
(C.L. L. it. 1964, col. iii. 10) (see ch. i. § 7), where the long form
of I may express the consonantal or half-vocalic sound j, as in
conIvwx (C.J. L. vii. 8, &c.) (ch. 1. § 1). Whether it was this
already existing practice of writing long I for 7, which made
Claudius abstain from proposing a new letter for j, when he
introduced the inverted F-sign for v, or whether he followed the
Greek alphabet which had a sign for w (the digamma), but none
for y, we cannot say. Possibly the reason is to be found in the
more rapid development of the w-sound (Latin v) than of the
y-sound (Latin j).

F and 6 (which had by this time become between vowels the
bilabial spirant) were, as early as the third cent. a.p., hardly
distinguishable, as is seen from the frequent warnings given by
the grammarians against confusion of Jabat and lavat (Probi
Appendix 199. 22 K.), dibedo and “ivido (ib. 201. 4 K.), &e., &e.
Indeed one grammatical treatise (of the fifth century) is
devoted to this very subject: Adamantii sive Martyrii de
B vocali et V vocali, It was summarized by Cassiodorus for
the book on Orthography which he compiled for the use of
Benedictine copyists of MSS. (Keil, Grammatici Latiui, vol. vii).

At some time before the fifth cent. a.D., when precisely we
cannot say, initial v, and possibly 7 in other positions too, seems
to have passed from the bilabial spirant (Spanish 4) to the
labio-dental spirant (our +). After /, 7 it assumed in time the
sound of the voiced mute 0.

§ 49. .Testimony of grammarians. Quintilian i. 4. 10, tr, after saying that
a letter is wanting to the Latin alphabet to express the sound of v in servus,
vulgus, the Aeolic digamma, goes on to speak of the consonantal (pro consonan-

tibus) character of the vowels i, u, e.g. iam but etiam, wos but tuos. Of conjicit
he says, littera é sibt insidet, and so with u in vulgus, servus. In another passage
48 THE LATIN LANGUAGE. (Chap. II.

(i. 7. 26) he tells us that serwos was the spelling of his teachers, serwus that of
his own time, but that neither spelling quite expressed the sound, so that
the emperor Claudius had good reason to introduce a new letter like the
Aeolic digamma (cf. xii. ro. 29). The usual expression of the grammarians
for j and v is ‘(i, u) transeunt in consonantium potestatem’ (e.g. Mar.
Victorin. vi. 5. 18 K.; Donat. iv. 367. 12 K.; Charisiusi. 8.1: ef. Diom. i.
422.14 K.; Ter. Maur. vi. 341. 536 K.). Later they talk of the ‘ pinguis sonus’
as opposed to the ‘exilis’ or ‘tenuis’ (vocalic), the first to use this term being
Servius (fourth cent.) (iv. 422. 1 K.), e.g. Pompeius (fifth cent.) (v. 103 K.
‘vanus’ quando dico pinguior sonus est. numquid dicis ua nus? ergo vides
quia, si ponantur solae, tenuem scnum habent, si jungantur ad alias litteras,
pingues sonant. similiter et i sic patitur. ‘itur,’ ecce tenuius sonat ; si dicas
‘Titius,’ pinguius sonat et perdit sonum suum et accipit sibilum). Finally
Priscian (sixth cent.) speaks of the ‘diversus sonus’ of j and v from 7 and u,
and questions the soundness of Censorinus’ (third cent.) contrary opinion
(i. p. 13 H. non sunt in eisdem, meo judicio, elementis accipiendae : quamvis
et Censorino, doctissimo artis grammaticae, idem placuit) (cf. Nigidius
ap. Gell. xix. 14. 6). In another passage Priscian talks of v and b as quite
similar in sound (i. 18. ro H.), where he says that caelebs should be written
*caclevs, the word being derived from caelum and vita, and meaning literally
caelestium vitam ducens (!), were it not that v is never allowed to stand before
aconsonant. He goes on to say that b had this sound in very early Latin,
because Quintilian quotes Belena for Héléna (Fedéva) (ef. Serv. in Don. 422. 2 K.,
and ¢.1I. LZ. i. 1501) from early literature. This remark is interesting as
showing how early MS. corruptions showed themselves. When we turn to
the passage in Quintilian (i. 4. 15), we find that he is discussing the use of b
for Greek 7 and ¢ in early Latin. His examples are Burrus (for Iuppés), Bruges
(for Spiyes) and balaena (for pdéddAawva). The whole passage is taken from
Verrius Flaccus, who used these same examples. In our MSS. of Quintilian
there is the corruption Belena for balaena, a corruption which must have also
existed in the MS. used by Priscian, and which led him to make this mistake
(Fleck. Jahrb. 1889, p. 394). We notice that Consentius (fifth cent. ?) happens
to use pinguis in precisely the opposite sense when he speaks of that mis-
pronunciation of vent which made the word almost like a trisyllable (v. 395.
15 K.), unless indeed he is referring to the bilabial (w) as opposed to the
labiodental spirant sound (v) (see below). Other barbarisms which he
mentions as ‘in usu cotidie loquentium’ are so-lu-ii for disyllabic solvit, wam
for uvam, induruit (a trisyllable) (v. 392. 35 K.).

§ 50. j andvin early Latin. Priscian (i. p. 17. 3 K.) is certainly wrong in
explaining the sine invidia of Terence (Andr. 66) by the vocalic character of v
(see ch. ili. § 34) ; Accius’ augira (Tray. 624 R.): pré certo arbitrébor sortis,
6racla, aditus, aigura, may bea byform, and not a case of suppression of ¢ (y);
progenie mi genui on a hexameter line of a Scipio epitaph of ¢. 130 B.c. (C. ZL.
i. 38) is perhaps a graver’s error for progeniem genui; the use of -i, not -d, in
the Gen. Sing. of 10-stems in the older writers has nothing to do with the
sound of j (see ch. vi. § 20), nor have the Plautine forms ain (always), aibat
(occasional) (see ch. viii. § 35) ; peitro, where the r of the preposition has been
dropped, owing to the consonantal nature of the 7, seems to be a later spelling
than periero (see Georges, Lex. Wortf. s. v.); and the true account of puletwm
fleabane (also pulegiwm, see Georges) is a matter of doubt (see ch. iv. § 116).

t
§§ 50—-52.] PRONUNCIATION. J,V. 49

Between 7 and another i, v disappears at a very early time, e. g. obliscor, dinus in
Plautus (see Rhein, Mus. xxxv. 627) ; and Plautus’ trisyllabie avonculus (aunculus
or aonculus) seems to be a suppression of pretonic + like the later Noembris
for Novembris (see below). The vocalic nature of v in cave (pronounced with
é, ch. iii. § 44) is seen from Cicero’s story (Dir. ii. 84) of the confusion of Cauneas
(se. ficus vendo) with cave ne eas, as well as from the spelling causis for cave sis
in Juvenal ix. 120; of v in ave (pronounced with -2, Quint. i. 6. 21) from
Phaedrus’ fable (App. 21) of the man who mistook the caw of a crow for
this word (famila for familia on an inser. of Ameria in Umbria, 6. ZL. xi.
4488, may be a dialectal variety, like the Oscan famelo ‘familia’ of Bantia, Zv.
I1.I, 231). Our w is similarly suppressed in ‘Ha(w)arden,’ ‘Main(w)aring,’ &c.

§ 51. in late Latin and Romance. With Latin j (our y) were merged in
Vulgar Latin g before e, i, and d@ before i followed by a vowel (see below), for
these three Latin sounds are indistinguishable in the Romance languages.
Spellings therefore on late inscriptions like Diuliali (Rossi 1118, of 568 a. D.°,
Madias (Rossi 172), Giove (I.R.N. 695), Gianuaria (Fabry. x. 632, Interamna,
of 503 a.D.) do not indicate that j had passed from the y-sound (see A. L. L.
i. 220), but that -diu-, -dia-, gio-, gia- were pronounced like -yu, -ya, yo-, &e.
The occasional spelling with Lat. z, Greek ¢, e.g. Zanuario (C.1.L. x. 2466),
(ovaea (I. I. S. 826. 22, Naples), «oous (Lat. co(n)jux, C. I. L. x. 719, Surrentum)
is probably nothing but an attempt to indicate the spirant sound of j (our y)
as opposed to the vocalic sound of 7; for Lat. z, Greek ¢ had at this time the
soft or voiced s-sound of our verb ‘to use,’ and not our j-sound, nor the sound
of -dz- in ‘adze’ (see § 120). (For other examples see Schuchardt, Vok. i.
pp. 66 sqq.) This Vulgar Latin y-sound of triple origin is y in Spanish (in
most situations’, in Sardinian, and (by Greek influence ?) in South Italian,
but in ordinary Italian (except when pretonic, e.g. rione from Lat. regidnem,
ajuta, pronounced ayuta, from Lat. adjatat) it has become the sound of our j;
while in French (in most situations) it has assumed the sound which we
write sin ‘pleasure,’ z in ‘azure.’ Thus Latin jugwm is Span. yugo, Ital.
giogo, Fr. joug ; Lat. majus is Span. mayo, Sicilian mayu, Ital. maggio ; Vulg.
Lat. Jenuarius is Sicil. yennaru, Ital. gennajo, Fr. janvier. In loanwords in
Welsh Latin j has the y-sound, e.g. Ionawr (Lat. Janudrius), dydd Iau (Lat.
dies Jovis). In Greek inscriptions, besides the usual z, e.g. IovAcos, Movmecos,
we have sometimes n and e, e.g. "HovdAcos, ElovAcos, Tacos and Tanos (see,
Eckinger, p. 80).

The barred @ of the Pelignian dialect (Petiedu, uidadu, Uibdu, afded in the
same inscription, Zvetaieff, Inscr. Ital. Inf. 13) expresses some sound into which
consonantal i(y)and di in hiatus had developed (Latin Pettiedia, *viam-do, Vibidia,
abiit) (Rhein. Mus, xliii. 348 ; Class. Rev. vii. 104), and seems to be a dialectal
anticipation of the coincidence of di in hiatus and j in Vulgar and late Latin.
After a short accented vowel y suffered some similar change in Teutonic,
witness Goth. iddja I went, from the root EI to go.

§ 52. v confused with b in late Latin and Romance, From the beginning
of the second century a.p. we begin to find b and v interchanged on inscrip-
tions (see Schuchardt, Vok. i. 131 and iii. 67 ; Brambach, Orth. p. 238), and by
the third century the confusion is complete. The b-symbol is, as is natural,
used for the v-sound more frequently than the vowel symbol (capital V, uncial
U, see i. 7) for b. Latin b had probably by this time become, when between

E
50 THE LATIN LANGUAGE. (Chap, II.

vowels, a spirant (see § 78), so that the tendency is to restrict V, U to the
vowel- (u), B to the spirant-sounds (0, v). (For examples of the interchange
see the Indices to the Corpus.) In Greek inscriptions ov is the earliest spelling
for Lat. v, and continues to be the usual spelling throughout the Imperial
period, e.g. dpovadis Mon. Ancyr., Oceonactavos (never Beon-), overepavos and
overpavos, &c, But we find 8 occasionally even in the first cent, a.p., the
earliest examples being Aafios, AccBios (the usual spelling), %:ABavos, Badepios.
This use of 8 may have been stimulated by the preference of a single to
adouble symbol. Ae:Sios is more’pleasing to the eye than Aeiowos ; and in this
way we may explain why Latin ov is more often 08 than oov, (It is often ov,
e.g. Novos) (see Eckinger, pp. 82 sqq.) Little light however is thrown on the
pronunciation of Latin v by this Greek use of 8; for in the first place, the
pronunciation of 8 itself in the Imperial age is uncertain (Blass supposes it
to have become a spirant, as in modern Greek, in the second cent. a.p.
Aussprache d. Griech.? p. 91), and in the second, the use of 8 followed in all
probability the use of bin the Latin spelling. (Thus on the Edict of Diocletian
vulva is spelt bulba in the Latin inscription, BovABy in the Greek.) We some-
times find ov and B on the same inscription, e.g. Nepova and Nepfa (second
cent.), PAaouavos and SAaBiavos on an inser. of Cyrene, 117-125 A.D. (see
Eckinger). The remarks however of the Grammarians point, as we have
seen, to v having retained its connexion with the vowel wu till a later time in
correct pronunciation ; and the same thing is indicated by the loss of inter-
vocalic v in paimentum, &e., for pdvimentum (see below). At what time the
bilabial spirant v (our w) became the labiodental spirant v (our v) is not easy to
say. It would be rash to conclude from spellings like convivium, convivio
(where the m of com is changed to ”) on the Lex Municipalis of Julius Caesar
(C.I.L. i. 206), and still more from invitei, inviteis (where the n of in is retained)
on the Sententia Minuciorum of 117 B.c. (i. 199), that the change had taken
place in the Republican period; for as early as 189 B.c. we have inpeirator
(Wilm. 2837), and in the Sen. Cons. de Bacchanalibus of 186 B.c. conpromesise
(i, 196), clear instances of m before an undoubted bilabial. The facts certainly
point to com-, im- being the oldest spellings before v- (and /-, see § 64), e.g.
comvovise (and coventionid) i. 196 ; comvalem, comfluont (but also conflowont) i. 199 ;
and the im wita of the Palimpsest of Plautus (Merc. 471), comuiuas (Men. 224),
may rest upon old tradition ; but the substitution of -n for -m of a preposition
‘before a consonant in a compound is no certain evidence for the nature of
the consonant (see § 65). More weight may be attached to Cicero’s deliberate
preference of the spelling com before v, mentioned by Marius Victorinus
(fourth cent. a.p.) (18. 14 K.): item consonantes inter se, sed proprie sunt
cognatae, quae simili figuratione oris dicuntur, ut est b, f, m, p, quibus Cicero
adicit u, non eam quae accipitur pro vocali, sed eam quae consonantis obtinet
vicem, et anteposita vocali fit, ut aliae quoque consonantes. quotiens igitur
praepositionem sequetur vox cujus prima syllaba incipit a supradictis litteris,
id est b, f, m, p, v, quae vox conjuncta praepositioni significationem ejus
confundat, vos quoque praepositionis litteram mutate, ut est ‘combibit,’
* comburit,’ ‘ comfert,’ ‘ comfundit,’ ‘commemorat,’ ‘comminuit,’ ‘comparat,’
‘compellit,’ ‘comvalescit,’ ‘ comvocat,’ non ‘conbibit,’ ‘conburit,’ et similia.
sic etiam praepositio juncta vocibus quae incipiunt a supradictis litteris n com-
mutat in m, ut ‘imbibit,’ ‘imbuit,’ ‘imfert,’ ‘imficit,’ ‘immemor,’ ‘ immitis,’
‘impius,’ ‘impotens.’ The ordinary rule that com-, im- are used before », b,
§ 52.) PRONUNCIATION. J,v. 51

m is quoted by Priscian (i. p. 31. 2 H.) from Pliny, Papirian, and Probus (ef.
Papir. ap. Cassiod. 162. 6 K.; Prob. 150. 6 K.) with no mention either of
Ff (which Mar. Vict. must have taken from some older grammarian), or of v.
It is true that Cicero’s spelling, comuocat, &c., might equally well be taken as
a proof of the more vocalic nature of v in his time; for before a vowel com
is often the form in use, e. g. cimédo, cimitor, &. [Caesellius Vindex (end of
first cent. a.p.) (ap. Cassiod. 206. 17 K.) recommends com- before a vowel, con-
before a consonant or v: tunc pro m littera n litterae sonum decentius effe-
remus]. Butthe Latin and Teutonic loanwords give a similar indication of
a change in the pronunciation of v (at any rate of initial v), during the period
of the Western Empire. The early Latin loanwords in Teutonic languages
show invariably w for Latin v-, e.g. Goth. wins, our ‘wine,’ ‘wall,’ ‘-wick ’
(Latin vinum, vallum, vicus). But Teutonic loanwords in Italian &c., which
date from the Gothic occupation in the fifth cent. a. p., show gu- for Gothic w-
(e.g. guarire from Gothic warjan ; guisa, our ‘ -wise’), an indication that the
initial w- sound had passed out of use in Latin. An examination of the
Romance languages does indeed suggest that the change from the bilabial to
the labiodental spirant was not completed in the Vulgar Latin of all the
provinces; but on the other hand the close connexion of the w- and the
v-sounds, and the frequent passage of a language from either sound to the other,
weaken the force of the evidence. In Vulgar Latin intervocalic b had been
merged inv. This v, of double origin, has the labiodental sound in Italian
and French ; but is bilabial in Spain, and (possibly through Greek influence)
in South Italy. Spanish and South Italian also merge initialband v. Thus,
while initial and intervocalic 6 of Latin bibo receive a different treatment in
Italian bevere, they have the same spirant sound in Spanish beber, Sicilian
viviri, Calabrian vivere. The identification of Latin v and intervocalic b in all
the Romance languages, and therefore in Vulgar Latin, shows that it was in this
position, in the middle of a word between vowels, that b first became a spirant
sound (see below). Confusions of spelling between b and » are usually of
this sort, e.g. Daniivius, the spelling of the classical period, later Danubius (see
Georges, Lex. Wortf.s.v., and for other examples, s. vv. abellana, gabata, viduvium,
Suebi, sebum, Vesuvius, sevir). That the development also of v differed accord-
ing to its position in a word is a natural inference, and is confirmed by the
evidence. Initial accented v would, owing to the stress with which a con-
sonant was pronounced in this position, develop its consonantal character
more rapidly than intervocalic v, especially than pretonic intervocalic v (see
below). A good instance of a confusion of spelling due to this is the word
vénéficus, which so often assumed the form beneficus, that it produced in late
Latin a new word for a sorcerer, maleficus (A.L.L. i. 79) (cf. Probi App.
200. 9 K. inter beneficum et veneficum hoc interest, quod beneficum bene
facientem significat, veneficum autem veneni datorem esse demonstrat).
Vétillum is the correct spelling, not bédtilum (Nettleship, Contributions to Lut.
Lexic. s.v.).

After r and 1 the same thing seems to have happened; cf. late Lat. albeus
(Agrim. 82. 24), arba (75. 19), Vulg. Lat. corbus, curbus (Fr. corbeau, courbe,
&e.). Pliny’s example of preconsonantal J is the word silva (§ 99); and the
classical spelling of the Perfect of ferveo, where rv is followed by wu is ferbui not
fervui (feruui) (ef. Georges, Lex. Wortf. s.vv. vulva, ervum, gilvus; Probi App.
198.7 alveus non ‘albeus.’ Albeus occurs often on inscrr., e.g. C.1.L. x. 1.
52 THE LATIN LANGUAGE. (Chap. IL.

1695, 1696, 4752, 6850, Eph. Epigr. iii. 48). The only (?) early example of rb, ib
becoming 7», lv is acervissimam (I. N. 1951, of 155 A.D.), a misspelling due to con-
fusion of two similar words acerbus and acervus, and not to be taken as evidence
of a change of the sound7d to the sound rv. But 72, 2 for rv, lv is common on
inserr., e.g. coserba, Helbius, salbus, serbat, serbus, balbis (see index to C.I.L. xiv).
Assimilation also often played a part in the development of » and b; e.g.
verve is in Vulg. Lat. *berbix (Fr. brebis, Ital. berbice) ; vervactum is *barbactum
(Span. barbecho, Sard. barvatu, Port. barbeito), and the only change of rb to rv
that is common to all the Romance languages, viz. morvus for morbus, seems to
show the influence of the initial m (Span. muermo, Port. mormo, Prov. vorma,
Fr. morve, Sicil. morvu. See A. L. L. iv. 121). Primilegium for privilégium (Caper,
111, 2K.) is due to confusion with primus. We have / for v in the spellings
Judicafid (C.I.L. vi. 6592), Mafortio (le Blant, 7.G. 612 A, of 527 a.p. from
Narbonne).

§ 53. Intervocalic v dropped. Between vowels v seems to have retained
a voealic character much longer. It was dropped before w of the Nom. Sing., thus
divus (older deiv-) became *deius, deus, Gnaevus became Gnaeus, &e. (ch. iv. § 70), but
was usually restored from the other cases, e.g. rivus from rive, &e., but Vulg.
Lat. had rius, &e. (Ital. rio, Prov. rius, O.Fr. riu} ; between similar vowels it
is very prone to disappear, e. g. 7-2, sis, for si vis, obliscor, dinus (Plaut.), just as
in Mod. Tuscan between e-e, bee for beve ; late spellings like noicius, for novicius,
Noe(m)bris for Novembris, &e., are very frequent, especially when v stands before
the accent. (For examples see Schuchardt, Vok. ii. pp. 471 sqq., @ g. Flaus
C.I.L. 1. 277, viii. 9422, ao EB. B. v. 777; ef. the remarks of grammarians
like Probus, Inst. 113. 17 K. hoe ovum et non hoe ‘oum’; Probi App. 198
5. K. flavus non ‘flaus’; ib. 199. 2 K. rivus non ‘rius’; ib. 198. 8 K. favilla
non ‘failla’; ib. 199. 2K. pavor non ‘paor’; ib. 197. 28 avus non ‘aus’;
similarly on Greek inscriptions NoeuBpios is the usual form (as early as 73
B.c. in S. C. of Orupus) ; cf. ’Oxraios (time of Augustus) ; ’Afavos(C. I. I. 4750) ;
Zenpos, &c. (Eckinger, p. 92) (see also Georges, Ler. Worf. s. vv. longao, boa,
boo, Ribbeck, Index, p. 448 for spellings in Virgil MSS. like fluius, exuiae, iwenis,
beside which we find fluventa, bovum, fluridus, fluvitantem, ingruvit, tenuvia).

§ 54, Postconsonantalv dropped. Vulg. Lat. v(in classical Latin the vowel 1)
is also dropped after consonants not only before u (for examples see Schuchardt,
Vok. ii. pp. 464 sqq.), e.g. mortus for mortuwus, cardus for carduus (ef. cardelis
Petron 46. 4) (Ital. morto, cardo ; Span. muerto, cardo ; Fr. mort, chardon
from *cardo, -dnis), just as -guu- became -cu- in the beginning of the first cent.
A.D. e.g. ecus, locuntur, locutus (see § 93), but also when pretonic in words like
Jén(ujirius, Febr(wrarius, batt w)ére, cons(u)ére, contin(u)ari (see Georges and
Brambach s. vv. and for contin,w)ari, A. L. L. viii. 129, 136. Examples of this
spelling in Inscriptions and MSS. have been collected by Schuchardt, Vok, ii.
pp. 467 sqq. Compare the Romance forms, e.g. Ital. gennajo, febbrajo,
battére, cucire, &c.) (see ch. iii. § 15). In the App. Probi we have: 199. 12
Februarius non ‘ Febrarius’ ; 197. 23 vacua non ‘ vaqua,’ vacui non ‘ vaqui’
(ef. Febrarius in various Latin inscriptions, such as C. I. L. ix. 3160; xiv.
58. 2795). Pituita must have had in ordinary speech the trisyllabic pronuncia-
tion which Horace gives it (Epp. i. 1. 108 nisi cum pituita molesta est), and not
the quadrisyllabic of Catullus (xxiii. 17 mucusque et mala pituita nasi). For
Aelius Stilo’s derivation of the word was ‘quia petit vitam’ (ap. Quint. i. 6.
§§ 53-56. ] PRONUNCIATION. H. 53

37), and the Vulgar Latin form was *pipita or *pippita (Ital. pipita, Span.
pepita, Fr. pépie; cf. Mid. Engl. pippe, Swiss pfiffis). On the other hand
suavis seems to have been a trisyllable in Vulg. Lat. (as in Sedulius, e. g.
i. 274, and later poets), e. g. Ital. soave, O.Fr. so-6f, Prov. sodu. Servius (ad
Aen, i. 357) tells us that many persons in his day considered suddet to be
a trisyllable.

§ 55. ai, ei before a vowel. Velius Longus says that Cicero wrote Matia, &e.,
because he thought these words should be written as they were pronounced
(auditu emensus scriptionem) ; so cojicit might be written cotdicit to express the
sound of the first syllable coi and the second and third syllables ticit (Vel.
Long. vii. 54.16 K.) : in plerisque Cicero videtur auditu emensus scriptionem,
qui et ‘ Aiiacem’ et ‘ Maiiam’ per duo iscribenda existimavit : quidam unum
esse animadvertunt, siquidem potest et per unum i enuntiari, ut scriptum
est. unde illud quod pressius et plenius sonet per duo i scribi oportere exis-
timat, sic et ‘Troiiam,’ et siqua talia sunt. inde crescit ista geminatio, et
incipit per tria i scribi ‘ coiiicit,’ ut prima syllaba sit coi, sequentes duae iicit. ..
at qui Troiam et Maiam per unum i scribunt, negant onerandam pluribus
litteris scriptionem, cum sonus ipse sufficiat. hance enim naturam esse qua-
rundam litterarum, ut morentur et enuntiatione sonum detineant, quod accidit
et in eo quod dicimus ‘ hoc est’ [pronounced ‘ hoccest’ p. 54. 12], cum ipsa
vastitas litterae in enuntiatione pinguescat. atque ipsa natura i litterae est
ut interjecta’ vocalibus latinis enuntietur, dum et prior illam adserit et
sequens sibi vindicat. So Priscian (x. 1. 494) says that aio was spelt aiio
in former times, and is still pronounced ‘ayyo’ (iloco consonantis habet duplicis).
Our ordinary pronunciation Tré-ja, éjus is wrong. The first vowel of the
diphthong ‘retained its natural quantity, dero, Gaius, but dus, dio, major (see
Arch. Glott. Ital. x), as we see from Romance forms like Ital. peggio (with open
e) for Latin péor, and from the remark of Terentianus Maurus (p. 343 K.), that
in Troja, Maia, pejor, jejunium the vowel preceding j is short in each of these
words, though the syllable is long. Similarly ejiilo, to utter the cry e (Plaut.
Aul. 796 ei mihi!.. Cur eiulas?) must have been pronounced ei-iulo. In
unaccented syllables j, 4 seem to have been dropped after a short vowel in
Latin, e.g. the Adj. suffix -eus for *-eyos (Riv. Filolog. 1891 p. 18) (ch. v.).
Spellings like aiio are sometimes found in MSS. of classical authors, e. g. aioin
the archetype of Hor. Epp. i. 15. 45 was written aiio, whence the corruption
alio in several MSS (Class. Rev. v. 296); eiius in the Ambrosian Palimpsest of
Plautus, Most. 981 &c. ; ptiaculum (in the Vetus Codex pilaculum) Truc. 223.

§ 56. H. Latin 4, the representative of Indo-European GH (e.g.
hostis, our ‘ guest’) must in prehistoric times have had some
sound like German ch in ‘ach,’ Scotch ch in ‘loch,’ but by the
literary period had been reduced to the mere spiritus fortis, our
h. We have no reason to doubt that the sound was dropped in
Vulgar Latin as early as the middle of the third cent. B.c., for
we have not a trace of initial or medial 2 in any of the Romance
languages, not even the oldest ; and one of the earliest tasks of
grammarians at Rome was to draw up rules for the correct use of
54 THE LATIN LANGUAGE. (Chap. II.

initial /, their usual practice being to appeal to the Sabine dialect
where J.-Eur, g/- had become f (e. g. fostis), as gh in our ‘ enough’
(ch. iv. § 121). The Greek aspirated consonants 4, x, ¢ (t-h, k-h,
p-h, as in ‘ an¢-/eap,’ ‘ink-horn, ‘ up-Aill’) were expressed by the
simple tenues ¢, c, » in the Latin of the Republic, until at the
time of Cicero it was felt necessary to express them more
accurately by ¢h, ch, ph (ch. i. § 11); and this pronunciation was
carefully followed in polite circles. The struggle to attain the
new shibboleth of fashion led to ludicrous misapplications of the
h-sound by the uneducated classes, which have been satirized by
Catullus in his famous epigram on Arrius (84) :—
‘Chommoda’ dicebat siquando commoda vellet
dicere et insidias Arrius ‘ hinsidias’ ;

and the dropping of 4 seems to have been even in the time of
St, Augustine an unpardonable breach of manners. (On rf, rrh
for Greek p-, -6- see ch. i. § 11.)

Between vowels the omission of 4 was sanctioned by current
usage in a number of words such as némo (for *xe-hemo), débeo
(dehibeo), praebeo (praehibeo), praeda (for *prae-heda, *praehida),
By the first cent. prendo and ui? had established themselves in
pronunciation, also deprendo, through reprehensus was heard as
well as reprensus.

In the Umbrian language the length of a vowel was often
indicated by writing it before and after an 4, e.g. comohota (Lat.
commota) ; and it has been suggested that this usage may have
been adopted in Latin in a few words like vehemens, just as the
Oscan habit of doubling a vowel to express its length (e. g. tristaa-
mentud, Latin ¢estamento abl.) was adopted by Accius (ch. i. § 9).
Vehemens, according to this theory, is derived from vé and mens,
like résanus from vé and sanus (Ktym, Lat. p. 113). We have
NaHARTIS (C./.L. xi. 4213, time of Augustus), as well as
NART(is) (ib. 4201, 240 A.D.), &e., in Latin inscriptions from the
Umbrian territory, and Cicero (Orat. xlv. 153) speaks of the
name Ala (dhala) as representing Avilla (but cf. Diom. p. 424,
Dositheus, p. 382 K.). We find 4 put to the same use
in modern German, through analogy of words like stahel
‘steel’ (with 4 for I.-Kur. 4; cf. O. Pruss. stakla) which became
stal.
§ 57-] PRONUNCIATION. H. 55

§ 57. Testimony of grammarians: Quint. i. 5.19 quamquam per adspira-
tionem, sive adicitur vitiose sive detrahitur, apud nos potest quaeri, an in
seripto sit vitium, si h littera est, non nota. cujus quidem ratio mutata cum
temporibus est saepius. parcissime ea veteres usi etiam in vocalibus, cum
‘aedos ircosque’ dicebant. diudeindeservatum, ne consonantibus adspirarent,
ut in ‘Graccis’ et ‘triumpis.’ erupit brevi tempore nimius usus, ut ‘choronae
chenturiones praechones’ adhuc quibusdam inscriptionibus maneant, qua de
re Catulli nobile epigramma est. inde durat ad nos usque ‘ vehementer ’ et
‘comprehendere’ et ‘mihi’: nam ‘mehe’ quoque pro ‘me’ [leg. mi?] apud
antiquos tragoediarum praecipue scriptores in veteribus libris invenimus.
Similarly Gellius ii. 3. 1-4: h litteram sive illam spiritum magis quam
litteram dici oportet, inserebant eam veteres nostri plerisque vocibus ver-
borum firmandis roborandisque, ut sonus earum esset viridior vegetiorque ;
atque id videntur fecisse studio et exemplo linguae Atticae. satis notum
est, Atticos ixéuv et i pronomen et multa itidem alia, contra morem gentium
Graeciae ceterarum, inspirantis primae litterae dixisse. sic ‘lachrumas,’ sic
‘sepulchrum,’ sic ‘ahenum,’ sic ‘vehemens,’ sic ‘incohare,’ sic ‘helluari,’
sic ‘halucinari,’ sic ‘honera,’ sic ‘honustum’ dixerunt. In his enim verbis
omnibus litterae seu spiritus istius nulla ratio visa est, nisi ut firmitas et
vigor vocis, quasi quibusdam nervis additis, intenderetur. Then he goes on
to tell of a bookhunting friend of his who had bought for twenty gold
‘sigillarii’ a MS. of the second Aeneid, ‘mirandae vetustatis,’ which was
reputed to have belonged to Virgil himself. In v. 469 telis et luce coruscus
aena, the last word had been corrected to ahena, just as aheni, not aeni, was
the reading of the ‘optimi libri’ in Georg. i. 296. This account of h as (like
the Greek spiritus asper), a mere ‘nota adspirationis,’ not properly called
a ‘littera’ is a commonplace of the grammarians, e.g. Mar. Victor. vi. 5.
27 K.; ib. vi. 3; Charisius i. 265. 20 K.; Priscian i. 47, &¢. The only con-
tradiction is the absurd remark of Pompeius (v. 117. 14 K.), that in Virgil’s
line (den. ix. 610) terga fatigamus hasta, the h causes length by position,
a remark often repeated by the later writers on metre and followed in practice
by the Christian poets.

Terentianus Maurus in his description of the sound of h discusses its claims
to stand in the alphabet (vi. 331. 213) (Cf. Quint. i. 4. 9):

nulli dubium est faucibus emicet quod ipsis

h littera, sive est nota, quae spiret anhelum.

quin hance etiam grammatici volunt vacare,

quia non adicit litterulis novum sonorem,

sed graecula quaedam scholicae nitela vocis

vocalibus apte sedet ante posta cunctis,

‘hastas’ ‘hederas’ cum loquor ‘ Hister’ ‘hospes’ ‘hujus.’

Marius Victorinus says (vi. 34. 7 K.) profundo spiritu, anhelis faucibus,
exploso ore fundetur; and Martianus Capella (iii. 261) H contractis [con-
rasis Eyss.] paululum faucibus ventus exhalat. Cf. Priscian i. 24 ; Alcuin vii.
303. 18 K.

Rules for the use and omission of initial h are very frequent in the gram-
marians. Nigidius (first cent. B.c.) emphasized the importance of correctness
in the use of this letter: rusticus fit sermo si adspires perperam, a dictum
quoted by Gellius (xiii. 6. 3), who explains that by ‘rusticism’ Nigidius
meant what grammarians of a later date called barbarismus, Velius Longus
56 THE LATIN LANGUAGE. (Chap. II.

quotes Varro’s argument for the pronunciation hdréna, viz. that the Sabine
form of the word is fusena. Similarly haedus is supported by faedus, hircus by
Jircus (Vel, Long. vii. 69. 4-10 K.). Quite a number of dialectal forms have
been preserved for us through the grammarians’ practice of using dialectal /
asa criterion for Latin h, e.g. fordewm (with fasena, firci, faedi) (Vel. Long.
vii. 81 K.): the doubtful fariolus (Ter. Scaur, 11 K.) (with faedus, fordeum, and
p. 13: fircus) : Falisean haba (id. 13 K.): fibra (=herba) (Nigidius ap. ‘Serv.’
ad Georg. i. 120): forda bos, a cow in calf, Fordicidia (Paul. Fest. 59; 73 Th.
folus, fostis, fostia (id. 59): horctus, good (id. 73) : hanulum, a shrine (id. 73):
Suma (=humus’, Haunti (= Faunii) (glosses ap. Liwe, Prodr. 426) ; and a large
number of etymologies were made on the strength of this relation between
fand h, such as Formiae ‘velut Hormiae’ from Greek 8ppos (Paul Fest. 59) :
horreum from far (id. 73): firmus from Greek épya (id. 64). So Servius (ad
Aen. vii. 695): Faliscos Halesus condidit. hiautem, inmutatoh in f, Falisci
dicti sunt, sicut febris dicitur quae ante ‘hebris’ dicebatur, Formiae quae
‘Hormiae’ fuerunt, dd rijs épyys: nam posteritas in multis nominibus f pro
h posuit. These dialectal words are often loosely called ‘old Latin’: haba,
for example, which Terentius Scaurus expressly declares to have been
a Faliscan word (13 K.), is referred by Velius Longus (69 K.) to the
‘antiqui’; and Quintilian (i. 4. 13), amongst other genuine instances of old
Latin, such as Valesii, Fusii, mertare, says: quin ‘fordeum’ ‘faedosque’
{dicebant], pro adspiratione f ut simili littera utentes. There is however no
reason to believe that in Latin itself these forms were used, though they may
have been heard in the country districts about Rome, where dialectal
influence often strongly asserted itself. At other times grammarians defend
the use or omission of # by more or less ingenious etymologies, e.g. Servius
in Don. iv. 444. 28, 29K. dicta est enim [harena] quod harida sit terra ;
Charisius i. 103. 21, 22 K. harena dicitur quod haereat, et arena quod areat ;
gratius tamen cum adspiratione sonat. Velius Longus (vii. 68. 18, 19 K.)
defends dlica : cum ab alendo possit alica dici, et aliculam existiment dictam,
quod alas nobis injecta contineat, and ortus: quod ibi herbae oriantur.
Charisius says of this word dlica that Verrius Flaccus approved of the form
without h, whereas a line of Lucilius ran : nemo est halicarius posterior te
(i. 96. 9 K.). Caper’s dictum is: alica non halica (vii. 107. 12 K.). Another
doubtful case was the salutation dive. Quintilian (ij. 6. 21) tells us that though
the proper form was dvé, the verb being avére and not havére, yet no one, except
a precisian, thought of saying anything else than havé: multum enim lit-
teratus, qui sine adspiratione et producta secunda syllaba salutarit (‘avere’
est enim), ... recta est haec via: quis negat? sed adjacet alia et mollior et
magis trita, &e. (For examples of uncertainty in the use of h-, see Georges
and Brambach, s. vv. Hiberus, harena, haurio, exaurio, harundo, haruspex, hebenus,
hedera, helluor, Henna, heia, eiulo, Hilotae, Aedui, atica, allec, halucinor, Hadria,
Halaesa, Halicarnassus, Hamilcar, Hammon, Hannibal, Hanno, elleborwm, ercisco,
erctum, erus, Hadrumetum, haedus, hamus, hariolus, hibiscum, hinnueus, hircus,
hostia, holus, holitor, onustus, wmeo, umerus, ulcus, Hister, Hirpini, onero, Ilerda,
Tlyria.) Cf. Probi App. 199. 17 K. adhuc non ‘aduc’ (aduc in C. I. L. v. 6244).

The right employment of h is a leading subject in Alcuin’s handbook
of Orthography (vii. 300. 27 K.; 303. 11, 13 and 19; 306. 2); and St.
Augustine (Confess. i. 18) playfully remarks that the dropping of an h was
generally regarded as a more heinous sin than an offence against the law of
tak 60.) PRONUNCIATION. H. 57

ristian charity: si contra disciplinam grammaticam sine adspiratione
primae syllabae ‘ominem’ dixerit, displiceat magis hominibus, quam si
contra tua praecepta hominem oderit, quum sit ‘homo.’

§ 58. h between vowels. Quintilian (ix. 4. 59) says that depreidere, not depre-
hendere, was the form in use in his time. Gellius (second cent.) (ii. 3) speaks of
ahenun cf. uheneam, Comm, Lud. Saec. A 60, &e.), vehemens, incohare (along
with lachrumae, sepulchrun, helluari, halucinari, honeraw and honustus) as old-
fashioned forms now obsolete. A fourth century grammarian, called Probus,
says that frdho retains the h in spelling merely to indicate that the «and oare
prenounced separately, the word being spoken ‘trao’ (iv. 185.5 K.). On the
other hand in the second century Terentius Scaurus while declaring that
prendo, never prekendo, was the form in use, says that rho ‘sine dubio
aspiratur,’ and speaks of remens and vehemens, reprensus, and reprehensus as
optional (vii. 19. 14 K.\ [ef. Velius Longus (second cent.), vii. 68. 15 K.,
who gives remens and reprendo as the usage of the ‘elegantiores,’ prendo as
universal, and Annaeus Cornutus (first cent.), the friend of Persius, who
mentions prendo, remens, nil as the pronunciation of his day (ap. Cassiodor. vii.
153. 7 K.) (see also Aleuin vii. 311. 26, 27 K. ; Papirian vii. 159. 18-21 K.;
Eutyches vii. 200. 8 K.; Caper vii. 98. 12 K.)]. (For examples of confusion
in spelling, see Georges and Brambach s. vv. cohors, incoho, aeneus, Ahenobarbus,
Dahae, Phraates, coerceo, euhan, prooemium, periodus. For Greek compounds with
aspirate initial of second member following a consonant, see exedra, exodus,
synodus, Panhormus, &e. On the interjections aha, ehem ef. Richter in Stude-
mund’s Studien, i. il.)

§59. hin Old Latin. H was dropped earliest between vowels (e.g. “2m0);
and the disuse of initial 2 would no doubt begin with words which were
preceded in the sentence by a word ending in a vowel. Teutonic loanwords
with h- in Romance lost their & rapidly in Italian, Spanish, and Portuguese,
but retained it for some time in French, which in the Middle Ages abounded
in consonantal terminations. This was doubtless the principle of elision of
a final vowel before initial 2 in Latin poetry, whether the vowel was actually
final, or was followed by the vague nasal ‘after-sound,’ m (see §§ 153, 61).
There is no reason to suppose that initial A was in Early Roman poetry more
resistive of elision, than in the classical period. The Plautine flagitiim hominis
formed really a single word (ch. iii. § 12), and the hiatus is to be compared to
hiatus in compounds like circitit from circewm it, The weak nature of early h
is seen in compounds like cohonesto | co- before a vowel as in coeo, coorior, &e.)
which in Accius, Trag. 445 R. appears as cinesto (see § 149). Nihil is always
a monosyllable in Plautus apparently. But the dropping of initial ” on the
older inscriptions is hardly known. (See Sittl. Lok. Verschied. p. 39.)

§ 60. Greek aspirates in Latin. The Greek aspirates lost their aspiration
in loanwords used by the early writers, c.g. Plautus, as we gather from the
MSS. (see the statistics given in Fleck. Jahrb. 1891, p. 658 n.), from puns like
those on Chrysalus (Crusalus) and erucisalus, on Charinus (Carinus) and careo :—
Pseud. 736, non Charinus mihi quidemst sed copia, on Thales and talentum Capt.
274, and from the statements of later grammarians (ef. the pun on excalciaverat
‘eum adspiratione secundae syllabae’ (robbed of one’s money, yaAxés), and
excalceaverat, i. e. taken off one’s boots, calcei) in Porphyr. ad Hor. 8. i. 8. 39).
58 THE LATIN LANGUAGE. (Chap. II.

Quintilian for example (i. 5. 20) says: diu deinde servatum ne consonanti-
bus [veteres] adspirarent, ut in ‘Graccis’ et in ‘triumpis.’ There are not
wanting in Plautus indications that the vulgar Greek pronunciation of x as
‘k-kh (see Blass, p. 86) influenced some loanwords in popular use so as to
lengthen (by position) the previous vowel.  <Acc‘h)eruns, Acc(h)il(Des, like
bracc(h)ium seem to be required by the metre (Baier, Philologische Abhandlungen
eu Hertz. 1888°. Similarly the word tricae, whose origin has been traced to
8. Italy, where the word was applied to hair-shackles put on the legs of fowls
to prevent their straying, seems to be nothing but the Greek zpixes in
a Latinized form (bicae and *triccae, like brachium and bracchium, &e.). The
proverb ‘dpinae tricaeque,’ used of trifles (Mart. xiv. 1. 7) :—

sunt apinae tricaeque et siquid vilius istis,

has likewise been referred to S. Italy, where dpava: (Latinized apinae), ‘the
unseen realms,’ was in popular story the name of an imaginary country
of bliss, like Aristophanes’ ‘Cloud-cuckoo-land’ (Ribbeck, Leipziger Studien,
1887). The Vulg. Latin muttus, a word, from which Fr. mot is derived (mitus,
Non. 9. 16 M.?) seems to be similarly Greek piOos (muttus for mittus), as tricae
for *triccae; also stiippus a rope is Gk. orpégos (Festus, 452. 21 Th., says that at
Tusculum the word had the sense of a wreath, and that a Faliscan ‘Garland-
festival’ was called Struppearia ; the Romance forms point to stréppus, a form
found in this passage of Festus)... Gk. ¢ was in early Latin transcription
p, e.g. Pilipus on a denarius of the time of the Gracchi (C.I.L. i. 354),
though sometimes (like Gk. 7 in Burrus, Tuppés) b, e.g. Briiges for Spu-yes,
balaena for pdAAava (Quint. 1. 4. 15, from Verrius Flaccus) (see § 49). F was
not regularly used till the middle of the fourth cent. a.p. (see ch.i. § 11). Blass,
Griech. Aussprache®, p. 85, dates the change of Greek ¢ from the p-h to the
f-sound at about 400 a.p.; and the language of Diomede (fourth cent.) seems
to imply that the difference between Lat. fand Gk. ¢ was in his time very
slight (423. 28 K. ‘et hoc scire debemus quod f littera tum scribitur, cum
latina dictio seribitur, ut ‘felix,’ nam si peregrina fuerit, p et h scribimus,
ut ‘Phoebus,’ ‘ Phaeton’). The difference between the two sounds in Cicero’s
time is seen from Quintilian’s story of Cicero ridiculing a Greek witness who
could not pronounce the first letter of Fundanius (Quint. i. 4. 14). ©® is
however, as is natural, the Greek transcription of Latin f (Eckinger, p. 97) ;
e.g. Sovdavios is the Greek transcription of this very name on an inscription
of 81 B.c. (Bull. ix. p. 457, from Lagina in Caria). The old spelling persisted
in a few words like tis (Gk. 6vos), &c. (see ch.i. §11). A curious interchange
of s and th is occasionally seen in the Notae Tironianae, e. g. agatho for
agadso, Apollopisius for Pythius (Schmitz, Beitr. 109). With the introduction
however of Greek grammatical studies at Rome a more exact transcription
came into fashion (see ch. i. § rr), and it is to this tendency that we must refer
the aspiration of some consonants even in Latin words about this time, not
merely in words which were supposed to be borrowed from Greek, e.g. pulcher
(ch. i.§ 11) (referred to roAdxpous, Ter. Scaur. vii. 20. 4-8), and sépulchrum (sé and
pulcher! Charis. i. 73 17; ef. C.I.L. i. 1007 heic est sepulerum hau pulerum
pulcrai feminae), lachryma (to Gk. ddxpipa), &e., but to others which could
hardly be so misunderstood, .e.g. praccho, lurcho. In the Orator xlviii. 160,
Cicero tells ug that he was forced in spite of his convictions to yield so far to
popular usage as to pronounce pulcher, Cethégus, trimmphus, Karthago, though he
§ 60.) PRONUNCIATION. H. 59

still adhered to Orcivius, Mato, Oto, Caepio, sipulerum, corona, lacrima: quin ego ipse,
cum scirem ita majores locutos esse, ut nusquam nisi in voeali aspiratione ute-
rentur, loquebar sie, ut ‘ pulcros,’ ‘Cetegos,’ ‘triumpos,’ ‘Kartaginem’ dicerem.
aliquando idque sero convicio aurium cum extorta mihi veritas esset, usum °
loquendi populo concessi, scientiam mihi reservavi. ‘Orcivios’ tamen et
‘Matones,’ ‘ Otones,’ ‘Caepiones,’ ‘sepulera,’ ‘coronas,’ ‘lacrimas,’ dicimus,
quia per aurium judicium licet. Similarly Quintilian, in a passage already
mentioned (1. 5. 20), says: diu deinde servatum, ne consonantibus [veteres]
adspirarent, ut in ‘Graccis’ et in ‘triumpis.’ erupit brevi tempore nimius
usus, ut ‘choronae,’ ‘chenturiones,’ ‘praechones,’ adhuec quibusdam in inscrip-
tionibus maneant. qua de re Catulli nobile epigramma est. In the second
cent. a.D. pulcher was the current pronunciation (Ter. Scaur. vii. 20. 4-8 K. ;
Vel. Long. vii. 69. 13-17 K.), also Carthago, Gracchus, Otho, Bocchus ; unaspirated
were ctlo, coclea, cocledre (Vel. Long. l.c.). In the fourth cent. Orcus, Vulcdnus,
corona, ancdra, sépulcrum (Mar. Victorinus vi. 21. 20 K. ; Serv. ad Georg. iii. 223,
but for ancora cf. Serv. ad Aen. vi. 4), but Gracchus (Charis. i. 82. 11 K.), pulcher
(Serv. I. c.) were the forms in use. Pulcher held its ground most persistently
in spite of the rule, first apparently stated by Varro (Charis. i. 73. 17 K.),
and often repeated by the Grammarians (Ter. Scaur. vii. 20. 4-8; Probus
Cath. iv. ro.19 K.; Ter. Maur. vi. 932. 2ig-22rt K.; Mar. Vict. vi. 34. 5-6 K. ;
ef. Vel. Long. vii. 69. 13-17 K.) that no consonant should be aspirated in
a native Latin word. On the other hand this (Gk. Ovos), chéréna (from Gk.
xopés, Etym. Lat. p. 23), with lurcho, stpulchrum, Orchus, &c., were only adopted
by imitators of the Ciceronian age (Probus Cath. iv. 10. 19 K., Serv. ad
Aen. vi. 4; Mar. Victor. vi. 21. 20 K.; ef. Ter. Scaur. vii. 14 K.). Of ciréna
Festus (26 Th.), quoting probably from Verrius Flaccus (time of Augustus),
says: corona cum videatur a choro dici, caret tamen aspiratione. For
examples of these varieties of spelling on inscriptions, see Brandis, De conso-
nantium aspiratione apud Romanos (in Curtius, Studien, ii. 1869). Consentius
(v. 392. 19, 27) censures the mispronunciations Tracia, Trachia, Chartago. For
79 we have pth in pthoibus in the Comm. Lud. Saec. The use of rh for Gk. initial
6- was not approved by Varro (L.L. iii. fr. 58, p. 182 Wilm.) (see ch i. § rz).
{For examples of this confusion of spelling, see Georges and Brambach, s.vv.
ancora, area, tropaeum, baccar, Cethegus, Gracchus, murra, Orcus, Otho, pulcher, Regium,
rhombus, talasio, letum, simulacrum, charta, Bosporus (Bosphorus not till third or
fourth cent. a.p.), chlamys, chorda, clatri, cochlea, concha, cothurnus, cyathus, lurco,
lumpa, schema, schola, sepulcrum, raeda, Raetia, Ramnes, Rhodope, Rhodus, rhiis,
Riphaeus, romphaea, theatrum, Viriathus, triumphus, racana, ciniphes («vimes), triclinium
(Abl. Plur. trichilinis, C.I.L. ix. 4971; xiv. 375, 17). On the spellings
Calphurnius and Calfurnius, see Schuch. Vok. i. p. 18, and for the confusion of
ph, p, f, see ibid. on the spelling phidelis, and Georges on phaseolus, sifo, sulfur
Late Latin culfus ‘see A.L.L. vii. 443) is the precursor of the Romance forms
of Gk. xéAmos (Ital. Span. golfo, &.). In the Probi App. we have (199. 7 K.)
strofa non ‘stropa’; (199. 17) amfora non ‘ampora’; (197. 19) porphyreticum
marmor, non ‘ purpureticum marmur,’ and perhaps (199. 8) zizifus [zizibus
MS.] non ‘zizupus.’

The Romance forms show that Vulgar Latin retained the old equivalence
of the Latin tenues to the Greek aspirates ; e.g. Gk. «dAagok is Ital. colpo, O.Fr.
colp, Fr. coup ; xaA4v is Ital. calare ; 04AAos is Ital. tallo, Span. tallo, Fr. talle,
[Cf. the cautions given in Prob. Appendix against stropa, ampora (see above).]
60 THE LATIN LANGUAGE. [Chap. II.

So that misspellings like ch for c, th for t, ph for p on plebeian inscriptions
cannot have implied a different pronunciation. The aspirated forms were to
the uneducated Romans mere equivalents of the tenues. Ch, the equivalent
of c, was in Italian utilized to distinguish the guttural from the palatalized
sound, e.g. chi (Lat. qui), chiave (Lat. clavis) (see Schuch. Vok. i. p. 74).
Similarly 2 is written, but not pronounced, in Italian to distinguish a few
synonyms like ho (Lat. hdbeo), and o (Lat. aut), &e.

§ 61. M,N. The pronunciation of the nasals varied according
to their position. At the beginning of a word or a syllable m
and x had their normal sound. What this was for m there can
be little doubt. M, the lip-nasal, has in all Romance languages
at the beginning of a word the same sound, that namely of
our m. The N-sounds on the other hand vary considerably. There
is ‘dental ,’ as Sweet calls it, the point-nasal, with many
varieties according as the tongue touches the teeth (the true
‘dental’? 2 as in French, Italian, &e.), or the gums a little
behind the teeth, as in English, and so on. There is ‘ palatal’
n, Sweet’s front-nasal, as in Fr. Boulogne, vigne, Italian ogni,
Spanish sefior, cafion (something like our ‘ vineyard’), There is
‘velar’ or ‘guttural’ n, Sweet’s back-nasal, of English ‘sing,’
German singen. And 2 (like m), may be unvoiced, as in Ice-
landic knif, hnut, a sound common in England 200 years ago
in words beginning with kn like ‘know,’ ‘ knife,’ which have now
lost all trace of the initial 4. Voiceless m is heard in the inter-
jection ‘hm!’ Normal Latin ~ was not the true ‘dental’ (as
Italian ~ is) according to the Latin phoneticians. The tongue
touched not the teeth but the palate; what precise part of the
palate, we are not told (Nigidius ap. Gell. xix. 14. 7). Before
a guttural, » was ‘velar’ or ‘guttural’ , like Greek y in
dyyedos, ayxdAn, a sound called the Agma by Greek phoneticians
and by their Latin imitators (Nigidius 1. ¢.; Priscian 1. 39 ;
Mart. Vict. vi. 19. 11 K.), to express which Accius proposed to
follow the Greeks and write gy, e.g. aggiilus, aggens, agguila,
iggérunt. At the end of a syllable, before a consonant, Latin
m, x had again a parallel in Greek, and accordingly have received
attention at the hands of the Latin grammarians. The sound is
described as ‘something that is neither m nor x,’ as in Greek
odpBvé (Mar. Vict. vi. 16. 4 K.), a description that would apply
to the sound of in our own ‘unpractical, ‘unmerciful’ in
§ 61.) PRONUNCIATION. NASALS. 61

careless utterance, as well as to the preposition in Ital. impero,
intacco. There was however one sound of Latin m which had no
parallel in Greek ; and here the Latin grammarians do not give
us so much information as we could wish. The curious usage
of Latin poetry, by which a word ending with -m elides its
final syllable before an initial vowel or /, just as though it ended
with a vowel, has nothing like it in Greek. Quintilian (ix. 4. 40)
tells us that final m before an initial vowel was hardly pro-
nounced, and had a sound not represented in the alphabet:
neque enim eximitur, sed obscuratur, et tantum in hoe aliqua
inter duas vocales velut nota est, ne ipsae coeant. Cato, he adds,
wrote ‘dicae’ ‘faciae’ for dicam, faciam (cf. Quint. i. 7. 23;
Paul. Fest. 20. 6 and 51. 10), although this spelling was often
changed by ignorant persons: quae in veteribus libris reperta
mutare imperiti solent, et dum librariorum insectari volunt
inscientiam, suam confitentur. And according to Velius Longus
(80. 20 K.), Verrius Flaccus, in the time of Augustus, proposed
a new letter, the first half of the ordinary letter M, to express
final m before an initial vowel: ut appareret exprimi non debere.
Priscian (i. p. 29. 15 H.) says: m obscurum in extremitate
dictionum sonat, ut ‘templum’; apertum in principio, ut
‘magnus’; mediocre in mediis, ut ‘umbra.’ What the exact
sound of -m was, is not easy to determine. From Latin poetry
we see that a word ending in -m, e.g. finem, is, when the next
word begins with a vowel, treated like a word ending in a vowel,
e.g. fine. In both cases the final syllable suffers what is called
‘elision,’ fin(em) onerat and jfin(e) onerat (see § 153). Final -m
therefore lacks the weight of an ordinary consonant, the power
to prevent two vowels from coalescing, and in this respect is on
a par with initial 4-. Before e.g. honorat the final syllable of
fine, finem suffers ‘elision’ in exactly the same way as before
onerat. But are we to say that in fimem the em became a nasal
vowel, an e spoken ‘through the nose, or in stricter terms,
spoken with the passage into the nose not covered by the uvula?
In this case em would have a sound like that of our exclamation
‘eh!’, spoken with something of a nasal twang ; and to give this
sound to the Latin interjection hem (expressing surprise, sorrow,
indignation, &e., e.g. Ter. And. 435: quid Davos narrat? .. .
62 THE LATIN LANGUAGE. [Chap. II.

nilne hem? Nil prorsus), would not be unnatural. Or should
we say that e had its ordinary sound, and that this sound was
followed by some reduced form of m, probably some adumbration
of unvoiced or whispered m, at any rate something of as slight
a consonantal character as 4? A very probable account is that
-m was reduced through the lips not being closed to pronounce
it. If instead of closing the lips, all that were done were to drop
the uvula, a nasal sound would be given to the following initial
vowel, so that finem onerat- would be pronounced finewonerat with
a nasalized 0 (Gréber, Commentationes Woelfflinianae, pp. 171 sqq.).
‘When the next word began with a consonant, final m seems to
have had more weight in ordinary Latin, if not in Vulgar Latin,
for it never fails to make. its vowel long by ‘ position’ in poetry
of all periods, whereas final s in the earlier poetry usually does so
fail. Plautus, for example, could not end an iambic line with
nullum fert, though he does with nudlus fert (which we often
write nullw’ fert), But that it had not the definite s-character
of initial +, that m of viam continet did not sound like m of mira
continet we see from the remark of a first century grammarian,
that in the phrase etiam nunc, although m was written, something
else (like e¢zannunc), was pronounced (Vel. Longus vii. 78. 19 K.
cum dico ‘etiam nune,’ quamvis per m scribam, nescioquomodo
tamen exprimere non possum). How far this differed from that
sound of the nasal before a consonant in the middle of the word,
which is described as ‘something that is neither m nor x,’ it is
difficult to say. Compounds with prepositions at any rate, like
co(n)necto, comprimit, continet seem quite on a level with etiam
nunc (or etramnune?). In both these cases Latin m is treated like
Greek v in cvyxaddo, cvvteivw, cvpBdddrw, and the spellings found
on inscriptions ey kixd@, éu putavel@, ty yuvatka, ri Bovdny,
which has led in the Rhodian dialect of the present day to
a complete assimilation of the nasal to the consonant, e. g. tix xdri
(rHv yap) (G. Meyer, Griechische Grammatik,? § 274). As close
a parallel is offered by Sanscrit final which is described as ‘a nasal
of a servile character always to be assimilated to a following
consonant of whatever nature that may be’ (Whitney, Sanscrit
Grammar, § 71). This Sanserit -m before initial y, » becomes
‘a nasal semivowel, the counterpart of each respectively.’ If
§ 61.) PRONUNCIATION. NASALS. 63

the same happened in Latin, if conivnctus, coniux were pronounced
cotiunctus, coiiux, with the first 2 nasalized, and conuentio were
*eouuentio with the first w nasalized, it would explain why it is
that, in spelling, the nasal is often omitted, coventionid (C. 1. L.
1. 196), covugi (id. 1064) (cf. Sweet Primer, p. 104).

Final 2 had not this weak sound. Priscian tells us (1. p. 30)
n quoque plenior in primis sonat et in ultimis partibus sylla-
barum, ut ‘nomen,’ ‘stamen’: exilior in mediis ut ‘amnis,’
‘damnum.’ In the Umbrian language however it seems to
have been on the same footing as Latin final m, for it is some-
times omitted, e. g. nome (Lat. nomen), and sometimes written m,
e.g. numem. Both in Umbrian and Oscan final m shows the
same character as in Latin, e.g. Umbr. ku and kum (Lat. ca),
Ose. via and viam (Lat. viam), con preivatud (Lat. cum privato)
and cum atrud; and in both languages a nasal is often omitted
before a consonant in the middle of a word, e.g. Umbr. iveka
and ivenga (Lat. juvenca), uzo- and onso- (Lat. humero-); Ose.
aragetud (Lat. argento) and praesentid (Lat. praesenti): Umbr.
apentu and ampentu (Lat. impendito); Osc. Aanovs (Lat. Lam-
ponius). This omission occurs on Latin inscriptions too, e.g.
pEcEBRIS (C./.L. i. 930), MERETI (iii, 2702, &c.), but was a ten-
dency not allowed to develop, as we see from the fact that in
Romance » and m are always retained, e.g. Ital. Dicembre, Span.
Diciembre, Fr. Décembre. The Plautine scansion of wempe in
such a line as Cas. 599 :—

quin ta suspendis té? Nempe tu te dixeras,

has often been quoted as an instance of such a suppression of the
nasal (wZpe) in ordinary speech. But the true scansion is xemp(e)
with the same syncope of final -¢ as in fw and tune, ac (for *atc)
and atque, nec and neque. A nasal is only dropped before certain
consonants in Latin according to fixed laws and never without
doubling the consonant or lengthening the vowel (Skutsch,
Forschungen, i. § 2). Before s, for example, this was the case ;
so mensa was pronounced mésa, with the ordinary close sound of
the long vowel, to judge from the Romance forms, e.g. Ital.
mesa (with close ¢); in consitles the x, though written, was not
sounded (Quint. i. 7. 28, 29). Whether this ¢, 0 were at any
64 THE LATIN LANGUAGE. (Chap. II.

period of the language nasal vowels we are not told. In O.Engl.
n was dropped before th (of ‘ thin,’ &c.), s, £ with nasalization and
lengthening of the preceding vowel, e.g. map, ‘mouth’ (Germ.
Mund), gos, ‘ goose’ (Germ. Gans), fif, ‘five’ (Germ. fiinf).

Another internal group that calls for notice is mz. In most
Romance languages this has become wz, e.g. Ital. danno (Lat.
damnum), donna (Lat. domna and domina), colonna (Lat. colwmna),
but in French the ~ has yielded to the m, e. g. dame (Lat. damnum
and dém(2)na) (colonne is a bookword). That in the Latin pro-
nunciation the ~ had here a weak sound seems to follow from
Priscian’s remark (i. p. 30) : n exilior in mediis [sonat] ut ‘ amnis,’
‘damnum’ (where the syllable begins with m, a-mmnis, da-mnum,
while in éiamnune the syllable begins with ~), as well as from
Quintilian’s many centuries earlier (i. 7. 28, 29): quid quae
scribuntur aliter quam enuntiantur?...‘columnam’ et ‘con-
sules’? exempta n littera legimus. gw likewise takes different
paths in the Romance languages. Usually it becomes a palatal
n-sound, e.g. Ital. legno (Lat. ignwm), Span. lefio, but in Sar-
dinian 2m, e.g. linna, There is no evidence to show that the
g in this position in Latin took the velar guttural sound 7; and
probably the pronunciation was merely ordinary g followed by x.
net was pronounced at least in Vulg. Latin xt (probably not with
velar z), with suppression of the ¢, e.g. guintus (Ital. quinto,
with close ‘) (see § 144). md shows a tendency to ww (as in
Oscan and Umbrian) in forms like Plautus’ dspennite, distennite ;
but this assimilation was not carried out in literary or Vulgar
Latin (e. g. Ital. risponde, Lat. respondet).

M, v are ‘ liquids’ (dypat) in Greek, because they readily com-
bine (like A, p) in one syllable with a preceding mute. But in
Latin, though this was the case with 7, r, it was not with
m,n. Greek loanwords in Latin which presented this com-
bination, tended to be pronounced with a parasitic vowel, e. g.
Téctimessa for Greek Téxunooa: and this is the form in which
they appeared in the earlier literature. At the end of the
Republic, when it was considered a requisite of polite speech
meater exactness the Greek sound of these loan-
me was generally abandoned; though even in
Ptace it persisted in some forms, such as mina

  
  
§§ 62-64.] PRONUNCIATION. NASALS. 65

(Greek wvG), and in Vulgar Latin, as the Romance languages
testify, it never was given up (e.g. ctctnus, Ital. cecero, for

ktxvos) (see § 154).

§ 62. Phonetic descriptions of normal m,n. Ter. Maur. vi. 332. 235:

at tertia (sc. littera m] clauso quasi mugit intus ore ;
quartae [n] sonitus figitur usque sub palato,
quo spiritus anceps coeat naris et oris.

Mar. Vict. vi. 34. 12, 13 K. m impressis invicem labiis mugitum quendam
intra oris specum attractis naribus dabit; n vero sub convexo palati lingua
inhaerente gemino naris et oris spiritu explicabitur. Martianus Capella
(ii. 261), M labris imprimitur ; N lingua dentibus appulsa collidit. Priscian
i. 29, 30 H. m obscurum in extremitate dictionum sonat ut ‘ templum,’ aper-
tum in principio ut ‘magnus,’ mediocre in mediis ut ‘umbra.’... n quoque
plenior in primis sonat et in ultimis partibus syllabarum ut ‘nomen,’
‘stamen,’ exilior in mediis ut ‘amnis,’ ‘damnum.’ Nigidius (ap. Gell. xix.
14. 7), speaking of the Agma, says: si ea littera {n] esset, lingua palatum
tangeret. Ter. Scaurus mentions the exertion necessary to produce the
labial m (as also 6, p), vii. 14. 3 K. non sine labore conjuncto ore.

§ 63. The Agma. Nigidius (i. c.), in Augustus’ time, speaks of the n in
words like anguis, increpat, ingenuus as a sound ‘between n and g,’ & spurious
n (adulterinym), in which the tongue does not touch the palate, as in normal n.
Varro (ap. Priscian i. p. 30) says it is a sound common to Greek and Latin,
written g in Greek, and by the poet Accius in Latin, e. g. aggulus, agceps (ef.
Mar. Vict. vi. 19. 11). Marius Victorinus (fourth cent.), vi. 16. 4. K., declares
that this was the sound of the nasal before qu in nunquam, numquam, quanquam,
quamquam, which is a sound between n and g, though, he says, it is usually
spoken of by grammarians as if it were the sound between m and n of Greek
odpBvé. Spellings on late inscriptions like nuncqvam (C.I.L. v. 154) NVNC-QVAM
(iv. 1837), vNcqvam (x. 8192) may indicate this pronunciation, though in
IVNCXI (viii. 8692), &e., cx may be merely the common symbol for XY as in
vexor (ii. 3330), a spelling which has led to the corruption voxor in MSS. of
Plautus (Class, Rev. v. 293).

§ 64, m,n before consonant. Marius Victorinus, speaking of nunquam, num-
quam, &c. (vi. 16. 4 K.) says: clari in studiis viri, qui aliquid de orthographia
scripserunt, omnes fere aiunt inter m et n litteras mediam vocem, quae non
abhorreat ab utraque littera, sed neutram proprie exprimat, tam nobis deesse
quam Graecis : nam cum illi capBvé scribant, nec m exprimere nec n. sed
haec ambiguitas in his fortasse vocabulis sit, ut in ‘Ampelo,’ ‘ Lycambe.’
nam in nostris supra dictis non est. The word odyBvg, apparently the stock
example of Greek phoneticians for this ‘sound which is neither m nor n,’ is
an unfortunate instance. It is not a native Greek word, but a loanword from
another language, and in its original form seems not to have had a nasal
[Aramaic sabb’ka (Daniel iii. 5), Greek capBien]. [Is the word connected with
the Latin sambicus, elderwood, where we have a similar variety of spelling
between sambiicus and sdbucus (sab- in Romance)? On Lat. labrusca, in Vulg. Lat.
lambrusca (a MS, reading in Virg. Ecl. v. 7}, Ital. lambrusca, Fr. lambruche,
Span. lambrusca, see below.] We might be inclined from this to believe that

F
66 THE LATIN LANGUAGE. [Chap. IT.

the Greek phoneticians were speaking of a sound that is not a native Greek
sound, were it not for spellings on Greek inscriptions like roymfjs, ’OAvvri~, Cret.
avpdrapos, &e. This Greek sound was appealed to by Latin grammarians to
explain varieties of spelling like eorwmdem and eorundem, q tus and q
where m was required hy the etymology, true or false (eorwm, quam) (Cor: autns
ap. Cassiod. vii. 152. 3), but by the pronunciation, and even, as we have seen,
to explain numquam and nunquam, tamquam and tanquam, &e. Here again a
doubt suggests itself whether the reference of this sound ‘between m and n’
to the Latin language is not based on mere varieties of spelling which were
not varieties of pronunciation ; but an appeal to the inscriptions tends to
remove it. SENTEMTIAM (C.J. L. i. 206), DECENBER (ii. 4587, &c.) (see Indices
to C.J. Z.) can most easily be explained on the supposition that this sound
really existed in Latin, unless indeed they are due to the general confusion
between mi and nt, nb and mb caused by the co-existence of etymological and
phonetic spellings like comtéro and contéro, inbiitus and imbitus. (The spelling
sentemtiam is that of the Lex Julia Municipalis and may be due to some ortho-
caney eboony of Julius Caesar; for on the same inscription we have damdum,
Zama i lei, ty dam, ti cenwee la, ¥ tum, tamtae, tamt ) The
eueltiag of these verbs compounded with the prepositions in, cum (com) is
frequently discussed by the grammarians (see the passages collected by
Seelmann, p. 279) (cf. jandidum for jamdidum in Virgil MSS., Ribbeck, Index
s.v.). To the ordinary consonants before which » becomes m, viz. b, p, m
(cited by Priscian, i. p. 31. 2 H., from the elder Pliny), Marius Victorinus
(fourth cent.) adds f (probably from some earlier grammarian‘, and on Cicero’s
authority v, e.g. comfert, comvocat. Whether the change in the fashion of
spelling compounds of com, in with verbs beginning with /, v, is a proof that
these spirants passed from a bilabial to a labiodental pronunciation is dis-
cussed in § 52. It is possible that the nasal was not sounded before jf, v or
only slightly sounded (as before h, or a vowel initial), so that the variation of
m, nin the spelling would not indicate a change in pronunciation (cf. the
suppression of Latin n before f, v in Provencal : see below). The form co- often
appears on inscriptions before j, v in compounds as before a vowel or h (codrior,
<0e0, cohortor, cohaereo, but also comest, and in the older spelling comauditum, &c.).
Before v only on early inscriptions, e. g. covenvmis (C. I. L. i. 532), COVENTIONID
(i. 196) ; before j also on later, e.g. corcrro (C.J. L. ii. 1964, col. ii. 51), corver
(C. I. L. i, 1064, 1413, vi. 2516, &c.) (see Indices to C.I.Z.). The nasal is also
often dropped in simple words before mutes, m before Labials, e.g. Novebris, n
before Dentals and Gutturals, e.g. eudem, provicia (Mon. Ancyr.) (see Schuchardt,
Vok.i. p. 105). Greek inscriptions treat the Latin nasals in the same way, e. g.
Noe{p., Womwvtos, on the one hand ; Noev pros, Movrwvios, on the other (Eckinger,
pp. 109 sqq.). Occasionally a nasal is wrongly inserted in plebeian and late
inscriptions, m before Labials, » before Dentals and Gutturals, e.g. semptem,
singnifer. (Examples from inscriptions and MSS. are given by Schuchardt, Vok.
i. p. 113 sqq.) (Co may be an original byform of com in cogo, &c. ; see ch. ix.
§ 22,)

In Romance, however, there is no trace of any variety of sound in these
vases. The classical spelling is invariably reproduced, e.g. Ital. immobile
‘Lat. immdbilis’, tanto .Lat. tantus); and this fact strengthens the doubt
expressed above regarding the existence of this ‘sound between m and n’ in
Latin. Vulg. Lat. *rendo (Ital. rendo, with open e, Span. rendir, Fr. rendre, &.)
follows the analogy of prendo; and lambrusca, just quoted, for labrusca, should

 
§ 65.} PRONUNCIATION. NASALS. 67

probably be explained in the same way. But the dropping of the nasal in
such a form as infatibus, quoted (perhaps from some older grammarian) as a
‘barbarismus’ by Julian, Bishop of Toledo (end of seventh cent.) (Exc. in Don.
v. 324. 9 K.) is not reflected in the Romance forms (Ital. infante, Span.
infante, Fr. enfant, &c.). Before s, where we know that in Latin the nasal
was dropped in pronunciation with lengthening of the preceding vowel, e. g.
mensa, the pronounced form, mésa, is the form reflected in the Romance
languages, e. g. Ital. mesa, with close e. Before f, the same thing seems to have
happened in Latin, but almost the only instances of Latin words with nf in
Romance are compounds with the preposition or particle in, e. g. infans. These
show n, except in Provencal, where n is dropped, e. g. effas, eferms, efranher
(Lat. infringere), efern, afra (Lat. infra), cofes, cofondre [as » before v in evers,
eveja (Lat. invidia), covens (Lat. conventus), covertir]; but this x of Vulgar
Latin may be due to what is called ‘Recomposition’ (see ch. iii. § 18°.

§ 65. Final m. In the only other Indo-European language which has not
changed final -m into n we are confronted with a curiously similar difficulty
to that in Latin. The native grammarians of India, who at an early time
devoted themselves to a minute and exhaustive analysis of the phonetics
and accidence of Sanscrit, the sacred language of India, have left conflicting
accounts of the sound of m at the end of a word. Some hold that in a word
like Sanser. tam (Greek rér), the sound was that of a nasalized (anundsika)
vowel ; others teach the ‘intervention after the vowel of a distinct nasal
element called the anusvara, or after-tone’ (Whitney, Sanscrit Grammar ®, § 71).
The Romance languages do not help us to decide whether one or any of these
processes took place in Latin, for they offer no indication that in Vulgar
Latin jinem differed at all from jine. The only cases where Latin final m is
preserved are a few monosyllables; and in these m has been changed to n
(rem, French rien, quem, Span. quien). The Latin nasals in the middle of
a word have passed into nasal vowels in countries under Celtic influence, viz.
France and North Italy, and also in Portugal, where Latin n between vowels
has passed into a nasal vowel, e. g. Romio, Lat. Rémdnus, mao, Lat. ménus,
laa (contracted to 14), Lat. lana. The Portuguese suppression of intervocalic
n is not a complete parallel to the Latin usage; for it is in the middle of
a word that the nasal is so treated, and m is never suppressed like 2, e. g.
fumo, Lat. famus, fama, Lat. fama. The n is described as having first
nasalized the previous vowel, *mano, (with nasal a), and then having been
dropped (Meyer-Liibke, Rom. Gram. i. p. 314). Nor is the suppression of inter-
vocalic m in Irish in such a word as fearail, manly. This word is a compound
of fear, man (cf. Lat. vir), and, amhail, like (cf. Lat. similis). Between two
vowels in Irish m passed into a v-sound; and this v often combines with
a preceding vowel into a nasal diphthong like German aw nasalized. In an
unaceented syllable, as in fedramhail, this diphthong is so far reduced, as to
allow the last two syllables to coalesce into one. In Latin, m never had this
tendency to become v; and so the elision in such a phrase as feram illud cannot
be explained by the Irish reduction of fearamhail tojfearail. A better parallel
to the latter would be the reduction of comuentio, couentio to contio. (But
see ch, ix. § 22). (On Port. tam with m sounded like nasal w, see Sweet
Phil, Soc, xvii. 203.)

Final -m is omitted very frequently on the oldest inscriptions till 130 B. c.,
or thereabouts, and again on late plebeian inscriptions (see § 137). Still
68 THE LATIN LANGUAGE. (Chap. II.

there is no evidence that -m was more easily dropped in early Latin poetry than
in the classical age. Indeed Priscian G. p. 30 K.) speaking of final -m says :
vetustissimi tamen non semper eam subtrahebant, and quotes a hexameter
of Ennius (4. 354 M.) ending with milia militum octo (cf. Enn. A. 322 M.
beginning dum quidem unus); but this remark need not drive us to the
opposite extreme, of believing that -m was more sounded in early than in
classical times. To the Roman ear at all periods a syllable ending in -m seems
to have been the equivalent in hiatus of a syllable ending in a long vowel.
This appears to be the rule in Saturnian versification (see Amer. Jour. Phil. xiv.
309); and Plautus, with the older poets, Horace, with the classical poets,
allow prosodical hiatus in the one case as much as in the other. Lucilius,
for example, scans as a short syllable without eliding, quam (i. 32 M. irritata
canes quim homo quam planius dicat. He is speaking of the letter r), exactly
as he shortens quo (xxx. 24 M. quid seruas qué eam, quid agam? quid id
attinet ad te?). Horace’s niim adest is on a par with his si m? amas ; and the
prosodical hiatus quoted from Ennius by Priscian may be equated with the
instance quoted by Cicero (Or. xlv. 152) from the same poet Scipid inuicte (A.
345 M.). Nor was this a mere usage of poetry. The same thing is seen in
compounds, such as cémest, cdire (the spelling with com probably was the older
usa,e before a vowel, cf. comauditwm, and comangustatum, Paul. Fest. 46 Th. :
comegit Gl. Plac. xiv. 39 G.; on cigo, see above), which have the first syllable
short, like pracire, praeeunt. The compound of circum and i is a trisyllable
ciretit ; of anté and it a dissyllable, anteit. (On flagitium-hdominis in Plautus, see
above, § 59.) But the disregard of -m in metre before an initial consonant
(except under the law of Breves Breviantes, like Plautus’ almost invariable
énim)} is unknown until late plebeian verse, e. g. wmbré(m) levem (along with
talés amici) on the epitaph of a praeco (C. I. L. vi. 1951), moriente(m) viderent
(vi. 7578), &c. Even in the careless hexameters of the dedicatory inscription
of Mummius (i. 542; of 146 B.c.), which contain facilia occupying the place of
a dactyl, pacé need not represent pacem (as in i. 1290: pacem petit), but may
be the Abl. (as in Plaut. Rud. 698: tua pace) :

tua pace rogans te
cogendei dissoluendei tu ut facilia faxseis.

The dropping of final -m in vulgar pronunciation is attested by the remarks
in Probi App. (198. 27) triclinium non ‘triclinu’; (199. 14) passim non
‘passi’... numquam non ‘numqua’... pridem non ‘pride,’ olim non
‘oli’; (199. 17) idem non ‘ide,’ and by the spellings on late and plebeian
inscriptions (see § 137’. Consentius (fifth cent.?), p. 394 K., alludes under
the name of ‘Mytacismus’ to a practice of joining -m to the initial vowel of
the next word : sicut plerumque passim loquuntur ‘ dixeram illis.’ Similarly
Pompeius (fifth cent.), p. 287. 7 K., quotes from Melissus (second cent.) the
rule for the correct pronunciation of a phrase like hominem amicwm as a mean
between the two extremes, ‘homine mamicum’ and ‘homine amicum.’
Velius Longus (54. K.) says: cum dicitur ‘illum ego’ et ‘ omnium optimum,’
‘illum’ et ‘omnium’ aeque m terminat, nec tamen in enuntiatione apparet ;
with Quintilian (ix. 4. 39), quoted above, cf. Diom. 453. 9 K. ; Serv. in Don.
445. 14 K.

 

1 Cf. Enn. A. 287 non enim rumores ponebat ante salutem (with non enim
in all MSS.).
§§ 66-68. ] PRONUNCIATION. NASALS 69

Before the pronunciation alluded to by Velius Longus is found expressed
in writing, in MSS. of Virgil (see Ribbeck, Index, p. 430); so ‘ etiannunc’ on
the Herc. Papyri (Class. Rev. iv. 443).

Tanne for tumne is quoted by Festus (p. 542 Th.) from Afranius: tanne
arcula tua plena est aranearum? Cf. Quint. viii. 3. 45 on the sound of cum
before a word beginning with n- (cf. Cic. Orat.xlv. 154; Fam, ix.22.2). Final
-n of the preposition in is sometimes changed to -m before an initial labial
consonant (see Ribbeck, Ind. p. 433 for instances in Virgil MSS. like im burim,
im flammam, im mare, im puppibus). So forsum and forsitam in MSS. (see Ribbeck,
Ind. p. 420, and Georges, Lex. Wortf., and for other examples of the confusion
of -m and -n, Schuchardt, Vok. i. pp. 117 sqq.).

§ 66. ns. (See § 144.) In Greek inscriptions we find dropped before sin Latin
words frequently, at all dates and in all localities. The nasal is usually dropped
in the terminations -ans and -ens, also in -ensis, e. g. KAnuns, xactpnoa. But
Latin census and its derivatives usually retain 1, e.g. xijvoos, Knvowpivos (Mon.
Anc.), also the combination -nst- (Eckinger, pp. 114, 115). [For the variation of
s with ns in Latin spelling, see Georges, Lex. Wortf. s. vv. centies, decies, Consentia
(modern Cosenza), pinso, mensis, mensor, &e. Cf. Probi App. 198. 9 ansa non
‘asa’ ; 198. 2 Capsesis non ‘Capsessis.’] The pronunciation of ns as s, with
lengthening of the preceding vowel, led to the use of ns for s after a long
vowel, e.g. thensaurus for thesaurus [see Georges s.v. Other examples, e. g.
occansio (ef. Rhein. Mus. xvi. 160), in Schuchardt, Vok. i. p. 112]. The Appendix
Probi gives cautions against the use of ‘ occansio’ for occasio (198. 21 K.), of
‘Herculens’ for Hercules (197. 25 K.). Velius Longus (p. 79. 1 K.) says that
Cicero ‘libenter dicebat foresia, Megalesia, hortesia, and Papirian (ap. Cassiod.
160. 14 K.) says that tosus, tusus, prasus were the older spellings, but that the
rule of his time was to retain n in the P.P.P., not in Adjectives, e. g. formosus
(ef. Probi App. 198. 14 K. formosus non ‘formunsus’; Caper 95. 18 K.:
Ter. Scaur. 21. 10 K.; we have formonsae, C. I. L. vi. 2738); in the P.P.P. the
n seems to have been restored from the other part of the verb. Charisius (58.
17K.) says: mensam sine n littera dictam Varro ait quod media poneretur ;
sed auctores cum n littera protulerunt, Vergilius saepe, &c. (ef. Varro L. L.
v. 118). On quotiens (the better spelling) and on vicenswmus, &c., see Georges,
Lex. Wortf. s. vv., and Brambach, Lat. Orth. p. 269.

§ 67. nx. Of the spellings conjux and conjunx (statistics in Georges, s.v.),
Velius Longus (first cent. a. D.) says (p. 78 K.) that the spelling without n is
due to the analogy of the other cases. conjugis, conjugi, &e. In actual pronun-
ciation, he declares, the 7 is heard ; for ‘subtracta n littera, et difficilius
enuntiabitur et asperius auribus accidet.’ The comparison of other I.-Eur.
languages, e.g. Greek ov-(ug, Sanser. sam-yuj-, shows that the form without
n must have been the original form of the Nom. too, and that the x has been
introduced by the analogy of jungo (cf. Ter. Scaur. p. 20. 10 K.).

§ 68. mn. In the fifth cent. Pompeius (p. 283. 11 K.) mentions as a barbarism
columa (ef. the Diminutive cilitmella) for célumna, which looks very like the
pronunciation mentioned by Quintilian as normal in his day (columnam
exempta n legimus). This colum(nja seems to have become *coloma, as céltiber
became colober, by assimilation of unaccented u to accented 0 (see ch. iii. § 33),
whence colomna (Probi App. 197. 25) with open accented 0, the origin of the
Romance words for pillar, as *coldbra (open 0) of the Romance words for snake.
7° THE LATIN LANGUAGE. [Chap. IT.

But the analogy of célitimen may have had something to do with the pronuncia-
tion colum‘n)a (ef. scdémellum and scamnum. See Georges s.v.). Sollennis is
explained as a byform of sollemnis (from *amno-, around, Ose. amno-), due to
a supposed connexion with annus (Etym. Lat. p. 97). (On confusions of mn
and nn, n, see Schuchardt, Vok. i. p. 147, and Georges, Lex. Wortf. s. vv. antenna,
Portunus, lamna.) The insertion of p between m and » occurs in late plebeian
spellings like calumpnia, dampnum, sollempnis (see Schuchardt, Vok. i. p. 149, and
Georges, s. v. damnum). On mptsee below. Hiemps (cf. consumpsi, &c.), is the
regular spelling of the Codex Mediceus of Virgil, and is accepted by Ribbeck ;
though this form is condemned by the grammarians, e. g. Caesellius ap. Cassiod.
161. 17 K. ; Ter. Scaur. 21. 6 and 27. 3 K. ; Alcuin 303. 8 K.

§ 69. gn. Spellings on inscriptions like ingnominiae (C. I. L. i. 206, 45 B.C.),
congnato (x. 1220) seem to be mere etymological spellings like inpello, &c.
Singnifer, on a soldier’s grave (C.J. L. vi. 3637’, has been explained above in
§ 64 (cf. § 144). Nor are we entitled to conclude that g passed into a nasal
sound before m from spellings like subtemen and subtegmen, examen, and perhaps
exagmen (Class. Rev. 1891, p. 294) (see ch. iv. § 116). (For instances of the
spelling of gn-, consult Georges and Brambach s.vv. coniveo, conitor, dinosco,
cognosco, navus, natus, narus, aprugnus, and see Schuchardt, Vok.i. p. 115. On
conitor, &e., but cognosco, &e., see ch. iv. § 119.)

§ 70. net. The suppression of the guttural in quintus is something like the
dropping of -g of ‘going’ in the mispronunciation ‘goin’ to.’ Quinctus is
the spelling of the Republic, Quintus of the Empire, according to Brambach.
So on Greek inscriptions of the beginning of the second cent. B.c. Kovyxros,
but also Kowvros (Eckinger p. 122). For the byforms nanctus and nactus (see
Brambach) we have a parallel in sactus (reflected in Welsh saith) beside
sanctus, while Vulgar santus (see Georges) is Welsh sant. The -ct- is not
a development of -nct- but a byform (see ch. viii. § 10).

§ 71. nd. In all 8. and Central Italy nd has followed the course taken in
Umbrian and Oscan and become nn; and similarly Latin mb is mm. In ordinary
Italian, Latin nd has become nn (n) in the pretonic syllable, as is shown by ne
for Lat. inde, manucare for Lat. mandicdre. Similarly we have mm for mb in
amendue beside ambidue (Lat. ambo duo), the same assimilation as we have
in English, e.g. ‘lamb’ (nn for nd is seen in ‘Lunnon town’ for ‘London
town,’ &¢.). (For Latin confusions of nd and nn, see Schuchardt, Vok, i. p. 146,
e.g. Secunnus, and consult Georges s. v. grundio). Nt competes with mpt in
lanterna (better laterna) apparently from Greek Aazmrjp, and tempto (not tento),
Pomptinus (not Pontinus), pédétentim, &c. (see Georges and Brambach). Thus
voluntas and viluptas are sometimes confused in MSS. through the intermediary
form volumptas (Schuchardt, Vok. i. 5). But emptum, redemptum, consumptum are
the established spellings, though Marius Victorinus (21. r2 K.) demands on
etymological grounds emtum, redemtum, consumtum, as also redemsi, consumsi,
&c. (ch. iv. § 76).

§ 72. Parasitic vowelin Greek loanwords. The Early Latin instances have
been discussed by Ritschl, Opusc. ii. 469 sqq., who on the strength of MSS.
spelling, and the requirements of prosody, restored to Plautus such forms as
drdchiima, téchina, Alciména, Alet , Clicinus, gimindsium, Procina. Marius

 
§§ 69—73.] PRONUNCIATION. MUTES. 71

Victorinus (8. 6 K.) tells us that the form Tecmessa was first used by Julius
Caesar Vopiscus (an older contemporary of Cicero) in the title of his tragedy
of that name, and was so pronounced, at his orders, by the actors: juxta
autem non ponebant em: inde nec Alemenam dicebant nec Tecmessam, sed
‘Alcumenam’; inde ‘Alcumeo’ et ‘Alcumena’ tragoediae, donec Julius
Caesar, qui Vopiscus et Strabo et Sesquiculus dictus est, primus ‘Tecmessam ’
inscripsit illam, et in scena pronuntiari jussit (cf. Prise. i. p. 29 H. [u]
saepe interponitur inter cl vel cm in Graecis nominibus, ut “HpasdAjs
‘Hercules,’ ’AoxAnmés ‘Aesculapius,’ et antiqui “AAxpyyy ‘ Alcumena,’
"Adxpaioy ‘ Alecumaeon’). Similarly the Sardinian mouflon was called in
Latin miisimo (a name applied to a much-prized breed of diminutive horses ;
ef. Lucilius vi. 15 M. praedium emit, qui vendit equum musimonem), but in
Gk. povopor (Strabo). (On the parasitic vowel between a mute and/in Latin
words, like péric.w)lum, see ch. iii. § 13.) The difference of Greek and Latin in
this respect is seen in the fact that Aespos is the earliest and most usual form of
the name Décimus, older Decumus, on Gk. inscriptions, just as awkward com-
binations with 1, e.g. AevrAos, “ApBovoxda, MacxAos, Tareptdos, Movptaa. (Lat,
Porcula) are commoner on Greek inscriptions than on Latin (Eckinger, pp. 47,
75). The parasitic vowel is really the visible expression of a ‘ voice-glide’
(Sweet, Handb. p. 84), asin Germ. Knie pronounced ‘ kenie’; ef. Fr. canif from
Low Germ. knif. (For examples on inscriptions, e. g. Himinis, C.I.L. i. 982,
see Seelmann, p. 251, and ef. below § 154.)

§ 73. Tenues and Mediae. In pronouncing s, ¢, ¢ the vocal
organs are in the same position as in pronouncing 4, d, g, but the
breath comes through the open glottis, as the space between the
two vocal chords which stretch across the larynx is called.
With 4, d, g we close the glottis, by drawing these vocal chords
together, and produce what phoneticians call ‘ voice” 3B, d, g are
now therefore usually termed ‘ voiced’ mutes as opposed to g, ¢, ¢,
the unvoiced or ‘breath’-mutes. An older designation was
tenues and mediae. In some languages what are called tenues
and mediae do not really differ by the absence and presence of
‘voice,’ but merely by energy and weakness of articulation. In
one German-Swiss dialect, for example, German £ and g are really
the same consonant pronounced strongly and pronounced weakly.
For such languages the terms ‘fortes’ and ‘lenes’ are more
suitable than ‘ breath-mutes’ and ‘voice-mutes.’ In investigating
the sound of the Latin mutes we have accordingly to consider
whether the tenues differed from the mediae in being uttered
with the glottis open, or merely in being articulated with greater
energy. Another point to be taken into consideration is that
mutes, especially voiceless mutes, have in many languages
a ‘ breath-glide, what we roughly call 4, after them. In Danish
7% THE LATIN LANGUAGE. (Chap. ITI.

every initial ¢ is pronounced with this 4 following ; and the same
peculiarity in Irish-English is well known. That Latin a, ¢, ¢
were not so pronounced we can infer from the fact that for the
more exact expression of the Greek aspirates, pi. th, ch were
brought into use in the last century of the Republic, which
shows that p, ¢, c had not, at least at that period, the sound of
¢, 6, x (like our ‘upAill,” ‘antheap, ‘inkdorn’). The other
question, whether the Latin tenues and mediae are more properly
distinguished as breath- and voice-mutes, or as fortes and lenes,
is more difficult to settle. The Latin phoneticians, who, as we
have seen (p. 28), are not very safe guides on any point of Latin
pronunciation, are especially at fault here; for neither they nor
their Greek masters seem to have carried their analysis of sounds
as far as the phonetivians of India, who had at an early time
discovered the distinction between unvoiced (aghdsa) and voiced
(ghé§avant) consonants, and its dependence on the opening (vivdra)
or closure (samvara) of the glottis. The Latin phoneticians talk of
p and b,of ¢andd, of ¢ and g as entirely different types of sounds,
produced by different positions of the vocal organs. Seelmann
professes to find in their descriptions evidence that p, ¢, ¢ had
a more energetic articulation than 4, d,g. This is certainly true
of their account of ¢ and g; but it is doubtful how far it is true
of the others, and even if it were, how much authority should be
allowed to these descriptions. A better reason for believing that
the Latin tenues were pronounced with more energy of articula-
tion than the mediae has been found in the fact that the Greek
tenues, which must have lacked this energy, are often represented
in Latin as mediae, e.g. Greek xwGids, Lat. gdbius. This is, as
is natural, especially the case in the initial accented syllable,
which seems in Latin to have been uttered with a strong stress.
It must be added, however, that an examination of the instances
shows that they are almost wholly confined to Greek x, especially
when preceding certain sounds; and that the same tendency is
shown by Latin ¢ to be weakened in the same position to g, e. g.
Vulgar Latin *gavia for cdévea (Ital. gabbia).

On the other hand, when we consider the Latin loanwords in
Welsh and the Teutonic languages, we are led to believe that this
energy of articulation was not the only thing which distinguished
§ 73.] PRONUNCIATION. MUTES. 73

the tenues from the mediae in Latin. Had it been, we should
probably have found the two classes of mute confused in their
Welsh and Teutonic forms. But this is not the case; cf, Welsh
poe, Lat. pacem ; Welsh bendith, Lat. bén(é)dictio; Lat. cdgquina,
cocina is our ‘ kitchen,’ Lat. gemma our ‘ gem.’

And in Italian of the present day p, ¢, ¢ are unvoiced, 4, d, g
voiced. So we have grounds for believing the Latin tenues to
have been unvoiced, the Latin mediae to have been voiced; and
the guttural mutes, if not all three classes, to have been also
distinguishable as fortes and lenes.

In native Latin words the tenues and mediae are not confused
to any great extent. The same tendency that turned I.-Hur. @
into ¢ before 7 in atro-, &c. (ch. iv. § 113) is seen in the old spellings
mentioned by Quintilian (i. 4. 16) ddewanter and Cassantra (cf.
C.I.L. i. 59, ALIXENTROM; I50I, ALIXENTE(r) CASENTER(a),
both inscriptions from Praeneste, and in Ital. Otranto for Greek
‘YSpobs -obvTos, Lat. Hydruntwm). In very early times the single
letter c (Greek y) was used for the c-sound and for the g-sound ;
but, as we saw before, the two sounds must have been throughout
this period distinguished in pronunciation, though not in spelling.
It is perhaps only at the end of a word that we find a real
variation between tenuis and media. Final syllables were pro-
nounced as weakly in Latin as initial syllables were pronounced
strongly ; and we might expect to find the tenuis fortis at the
end of a word replaced by the media lenis.

This is apparently the explanation of the Roman preference of
the spelling ad, 0b, sub to ap (as in ap-erio), op (as in op-erio,
Oscan op), *s-up ; though in actual utterance these words were no
doubt sounded with -p when followed by a word beginning with
a tenuis, e.g. ab templo, ob templum (like obtineo, proncunced
op-tineo). The spelling was not so established in the case of
similar subordinate or proclitic words ending in a dental, e.g.
at, often written ad; but the change on plebeian and late
inscriptions of final -¢ of verbs to -d, e.g. religuid, is probably
due to this weakening. On the other hand, a final is often
reduced to a whispered sound in languages, and a voiced
consonant, if whispered, sounds more like an unvoiced.

In the Romance languages the Latin tenues and mediae, when
74 THE LATIN LANGUAGE. [Chap. II.

initial, and when the initial syllable, to which they belong, has
the accent, retain their identity with wonderful persistence, e. g.
Ital. puro (Lat. purus), bene (Lat. béné), tale (Lat. tals), duro
(Lat. darus); but in the middle of a word, and when in the
unaccented syllable (though not after the Latin diphthong az,
e.g. Span. poco), the tendency is almost universal to turn the
tenuis into a media, the media into a spirant. In one language,
however, Roumanian, the tenuis is usually preserved, e. g. mica
(Lat. mica), liptuct (Lat. Jactdca), muta (Lat. miitdre), and
in Italian the reduction of the tenuis is of limited extent, e. g.
‘amico (Lat. amicus, Span. amigo), uopo (Lat. jpws, Span. huebos),
vite (Lat. vitis, Span. vide), fuoco (Lat. fdcus, Span. fuego), &e.,
though before a we have the media in miga (Lat. mica), strada
(Lat. strata), lattuga (Lat. Jactica), &c., and when the vowel
following has the accent, e.g. siguro (Lat. séciirus), mudare
(Lat. mitare). Misspellings on plebeian inscriptions like Amada
(le Blant, LZ. G. 576), iradam (Or. 2541, of 142 a.D.), Segundae
(Mur. 2076. 10) are precursors of these changes. That they
obtruded themselves into the recognized Latin pronunciation is
more than doubtful. One tendency indeed of the Romance
languages, to turn pr, tr, er into br, dr, gr, e. g. Span. padre (Lat.
puter), sobra (Lat. stpra), magro (Lat. médcer), lagrima (Lat.
licrima, cf. C. I. L, ix.648 LAGREMAS), is directly contrary to that
treatment of d before 7 in early Latin (dtro- for *ddro-, &c.),
which we have just mentioned. In Italian / remains after any
vowel except a, e.g. vetro (Lat. vitrum, Span. vedro), but padre
(Lat. piter), and pr is retained when it follows the accented
vowel, e.g. sopra (Lat. sépra), capra (Lat. céipra), but cavriuolo
(Lat. capreotus), obbrobrio (Lat. opprobrimm; cf. Or. Henz.
6086 ii).

(For the phonetic descriptions of the Latin tenues and mediae by Roman
grammarians see §§ 79, 86, 92.)

§ 74. Greek tenues in loanwords. The most frequent instance of the
change of a Gk. tenuis to a Lat. media, is the change of Gk. « to Lat. g,
especially before the vowels a, 0, u. Thus gummi for xéupe, gdbius for cwids,
giiberndre for kuBepyay, Stguntum for ZdxvvOos, &c., have always, or usually,
gin Lat. (see Georges, Lex. Wortf. s.vv.). The spelling varies in conger and
gonger, gorytus and corytus, commdrus and gammarus (see Georges). Ter. Scaurus
(xiv. 9, 10)says that some pronounced gawndce, some cwunace ; so gamellus and
§§ 74, 75.] PRONUNCIATION. MUTES. 75

céimellus (cf. Probi App. 198. 9 calathus non ‘galatus’; Gloss. ap. Mai, Cl.
Aue. vi. 578 corax per ¢ non per g; and see Georges and Brambach x. vv.
Caieta (now Gaéta), Agrigentum, grabatum). In the modern Milanese dialect, the
French cabriolet has similarly become gabriolé. Before n,c became g in Latin
(ch. iv. § 116), so that the spellings cygnus, Gnisus, Gnidus are only natural (see
Georges and Brambach). Of the final -ca of dmurca (Gk. éuédpyn), Servius (ad
G. i. 194) says that it was written with c, but pronounced with gy. (A similar
interchange of -ca, -ga, is seen in leuga and leuca, raca, and raga, &e.) For Gk.
» we find Latin d in the initial accented syllable before the vowel u in the
word buxus, and in the Old Latin name of King Pyrrhus, Burrus (the form
used by Ennius in his Annals, according to Cicero Or. xlviii. 160 Burrum
semper Ennius, numquam Pyrrhum; ipsius antiqui declarant libri; ef.
Quint. i. 4. 15, and Ter. Scaur. 14 K., who adds Byrria as the equivalent of
Gk. Tvppias). The form burrus was retained in rustic and colloquial Latin ;
burra was a name for a cow, burrus for a red-faced man, as we learn from
Paul. Fest. (p. 22. 32 Th. burrum dicebant antiqui quod nunc dicimus
rufum ; unde rustici ‘burram’ appellant buculam, quae rostrum habet
rufum. pari modo rubens cibo ac potione ex prandio ‘burrus’ appellatur),
who also mentions burranica potio; lacte mixtum sapa, a rufo colore (p. 26. 19
Th.), and burranicum ; genus vasis (p. 26.7’. Quintilian (i. 5. 13) quotes Cicero’s
phrase Canopitarum exercitum with the remark, ipsi Canobon dicunt. (On Latin
Canopus, Gk. KaywBos see Brambach s.v.) In’ Probi Append. (199. 5 K.)
we have, plastanon ‘blasta.’ Gk. Bardnov for raraviov is quoted hy Hesychius
as belonging to the Sicilian dialect (wardma.. mornpia. tives 5& Sid Tod B
Batdvia Aéyovaw, and Bardma: 7a Aondbia, 7H 58 Aegis Stxedrey). So carpiitinae
crépidae of Catull. xeviii: 4 are in Gk. xapRariwa:and kapratwa. Old Lat. Telis
for @eris (C. I. L. xiv. 4102, on a Praenestine mirror) (cf. Varro, L. L. vii. 87
lymphata dicta a lympha; lympha a Nympha, ut quod apud Graecos Oeris,
apud Ennium: Thelis illi mater, and &. R. iii. 9. 19 antiqui ut Thetim
‘Thelim’ dicebant, sic Medicam ‘Melicam’ vocabant) perhaps implies an
intermediary form with d for Gk. 7. The relation of cotonea, the origin of the
Romance words for quince (Ital. cotogna, Fr. coing) to Gk. «viwvia is not clear
(cf. Macrob. vii. 6. 13 mala cydonia quae cotonia Cato vocat; Pliny, N. H. xv.
to). In Vulg. Lat. we find additional examples of g for Gk. «, such as *grupla
(Ital. grotto), *garofulum (from xapudpuardor), *gontus : not to mention *gattus for
(Teutonic?) cattus, and probably gamba. Of spellings on Gk. inscriptions may be
instanced xad:nwy for ciligarum, caparyapov (from serrdcwm), oyakyn allon the Edict
of Diocletian ; also T'avzios for Cantius, and Kavé:ros for Candidus (Eckinger, pp. 98,
100, 102). Thesame interchange of tenues and mediae is seen in Gk., e. g. rhxw
and riyavov, especially in loanwords, e. g. Tamys, Tamis and bdéms, ’Apmpaximrns
and ’ApBpamerns : in dialects we have, e.g. xAd-yos' yada, Kpfjres, Hesych., and in
later Vulg. Gk. 7 often becomes 5 between two vowels, Itis thus often possible
that the Roman word represents a Greek byform. [Other examples of the
variety in Latin loanwords are carbdsus, Gk. xdpmacos, criimina, Gk. ypipéa
a bag, galbinum, Gk. yxaABdvn, spelunca, Gk. onnjdvyya, not to mention incitega,
Gk. éyyvdhun. Cf. also latices with Gk. Adrayes, pldga with Gk. wAdé. On citrus
(also cedrus), and Gk. #é5pos, see ch. iv. § 113. Both Creisita and Crisida occur
on old Praenestine cistae for ‘Chryseis’ (C. I. L, xiv. 4109; i. 1 501) ]}.

§ 75. Confusion of mediae and tenues in Latin words. A large number
of seeming instances are not due to any Latin, or even Italic, law of sound,
76 THE LATIN LANGUAGE. (Chap. II.

but are survivals of that interchange of media and tenuis, which shows itself
occasionally in I.-Eur. roots, e. g. siicus and sigo. Many are dialectal varia-
tions ; for in parts of Italy the tenuis was used where the Latin form had
the media and vice versa, just as in modern Italian the Neapolitan dialect
has ¢ corresponding to the @ of lapidi (Lat. ldpides), &c. ; the Roman dialect
affects grosta, gautela, &c., with g- for c-. Thus Quint. i. 5. 12 tells us that
a certain Tinea of Placentia used ‘ precula’ for pergiila (see K. Z. xxx. 345). Of
the misspellings of this kind in inscriptions and MSS. (collected by Schuchardt,
Vok. i. pp. 124 sqq.), not a few are due to the similar appearance of the letters
G, C, B, P. But there is a residuum of undoubted instances of variation
between the tenuis and the media, at least forc and g. For p and b we have
the (dialectal?) word ropio, ared mullet, given as nickname to Pompey, who had
a florid complexion (Mar. Sacerd. 462 K. quotes a lampvon, perhaps a Fescen-
nine line sung by soldiers at his triumph, quém non ptdet ét rabet, nén est
h6émé sed répio). The word, which should probably be read in Catull. xxxvii. 10
(see Sacerd. 1. c.), is evidently connected with robus and rufus, for Sacerdos adds,
ropio autem est minium aut piscis robeus aut penis (ef. Ter. Scaur. 14 K. on
the doubtful examples of Paldtiwn and ‘ Balatium’ (by analogy of balo) ; Publicola
and O. Lat. Poplicola (by analogy of populus) ; propom on early coins for probum,
C.I.L. i. 19; aduocapit in the Carmen Arvale}]. For ¢ and d, probably not
petiolus, ‘a little foot’ (?) (Afranius ap. Non. 160 M. atque ddeo nolo nido
petiolo ésse plus (MSS. es pus!, for this is better explained as peciolo- (Ital.
picciuolo, O. Fr. peguel, &c.). But the most examples are of g for c, as we
found to be the case with Greek loanwords ; and this perhaps throws some
light on the early use in the Latin alphabet of the Greek Gamma-symbol as
the symbol for Latin ¢ as well as for g. In Vulgar Latin -cit- and -cer- in the
proparoxytone syllable seem to have become -git-, -ger-, e. g. *plagitum, *fager,
to judge from the Romance forms (see Arch. Glottol. ix. 104). (Is digitus
a similar transformation of dicitus? Dicitus is censured in Probi App. 198. 10,
and occursin MSS. See Schuchardt, Vok. ii. 413) ; initial cra- may have become
gra, e.g. Ital. grasso, Span. graso, Fr. gras from Latin crassus; Ital. grata, Span.
grada from Lat. cratis (Meyer Liibke, Rom. Gram. i. p. 353); Lat. gavia for cavea
is reflected in Ital. gabbia, Span. gavia, Prov. gabia; so *gonfldre in Ital. gonfiare,
&c. [For other examples of g-c, see Georges and Brambach s. vv. vicesimus, viceni,
triceni, tricies, nongenti, cremia, neglego, graculus, gurgulio, &e. ; on the change of
d before r to t, e. g. dtrox for *adrox (cf. idium), see ch. iv. § 113, of c before n to g,
e. g. dignus, see ch. iv. § 119; bibo (I.-Eur. *pibo) is due to the Latin tendency to
assimilate adjacent syllables (ch.iv. § 163); cf. also Quint. i. 6. 30 nonnumquam
etiam barbara ab emendatis conatur discernere, ut cum Triquetram dici
Siciliam an ‘ Triquedram,’ meridiem an ‘medidiem ’ oporteat, quaeritur.]

§ 76. Mediae and Tenues at end of word. Quintilian, who includes
among the points of inferiority of Latin to Greek the use of -b, -d at the end
of syllables (xii. 10. 32 quid quod syllabae nostrae in b litteram et d innituntur
adeo aspere, ut plerique, non antiquissimorum quidem, sed tamen veterum,
mollire temptaverint, non solum ‘aversa’ pro ‘abversis’ dicendo, sed et in
praepositione b litterae absonam et ipsam f [s edd.] subiciendo), mentions (i. 7.
5) with disapproval the practice of distinguishing ad, the preposition, from at,
the conjunction. The right use of -¢and -din words like at and dd, séd, quit (from
queo’, and quid (from quis), quit and quéd, &c., is a subject of frequent remark in tho
grammarians, e. g. Ter. Scaurus (12. 8 K.) approves sed on the ground that the
§§ 76, 77.] PRONUNCIATION. MUTES. a7

old form was sedum {cf. ib. 11. 8 K.; Vel. Long. 69-70 K. ; Probi App. 202,
37 K.; Cassiod. 212. 5 K.; Aleuin 308. 8 K., and (on haut and haud) 303. 3 K.; on
caput and apud Bede 264. 35 K., &e.]. Charisius (229 K.) quotes ad for at from
a speech of Licinius Calvus; and Vel. Long. 70 K. says that sed in sed enim
‘dlitteram sonat.’ Instances of the confusion or suppression of final -d and -¢
in inseriptions and MSS. have been collected by Schuchardt, Vok. i. pp. 118 sqq.
and Seelmann, pp. 366 sqq. They include, beside the cases just mentioned,
the use of -d for -t, or the suppression of -t, in verbal forms, like rogad, C. I. L.
iv. 2388 (but on O. Lat. feced, &c., for fécit, &c., see ch. viii. § 69), peria (for péreat)
iv. 1173; also the loss of -t in the combination -nt, fecerun vi. 3251, just as -t
is lost in the combination -ct, lac from lact from older lacte, an I-stem). The
preposition is spelt at in the Lex Col. Jul. Genetivae Urbanorum of 44 B.c.,
except when the next word begins with d-, e. g. at it judicium atsint, i. 2. rg ;
ateo .. ad decuriones, iii. 8. 7 (Eph. Epigr. ii. p. 122), but always ad (even in
adtributionem, &c.), beside aput in the Lex Julia Municipalis of 45 B c. (CLL,
i. 206). This uncertainty of usage has been taken as evidence of the final
dental having been uttered faintly, or having been a sound intermediate
between d and ¢, like the final dental of German, written dt, in Stadt, &.
The instances, however, mentioned by grammarians are all words which
would be closely joined in utterance with a following word, quid tibi?, quid
dicis, ad templum, ad deos, &c.; so that it is most natural to believe, as was
suggested of the confusion between -m and -n, that the sounds adapted them-
selves to the initial of the following word. Quid tibi would be pronounced as
quit tibi, and ad templum as at templum (like at-tineo, &c.). The spellings quid tibi,
ad templum would be historical (like ad-tineo), not phonetic. The weakness of
final d is better shown by its suppression after a long vowel. Thus the
Abl. Sg. ceased to be pronounced with -d about the end of the third cent. 8. o.
(see § 137), though d remained till later in the monosyllables med, ted,
while naud was retained before words beginning with vowels (Ritschl, Opuse. ii.
591, &c., Vv. 352) ; and the affection of -t by the spellings on plebeian and late
inscriptions of verb-forms like reliquid, &e. In Vulgar Latin -/ cannot have
been dropped till after the conquest of Gaul (4.L. L. i. 212°.

P is not found at the end of any Latin word, if we except volup, for viliipe,
the Neuter of an adj. *volupis; but it is common in Oscan, e. g. op (Lat. 0b),
ip ‘there.’, (On the spellings optineo, obtineo, &., see § 80'. On the other
hand final g never appears in Latin, though we have ¢ in ac (for atqu[e] ), nec
(for néqufe]), lac (for lact[e]), ite for ili-c[e!, &e. ; but nec is written neg-
invariably in negotium, and usually in neglego (see Georges, s.v.). (Of. ch. x.

§ 18.)

§ 77. Mediae and tenues in the Dialects. There is a good deal of inter-
change of tenues and mediae in Umbrian and Oscan (e. g. Osc. deketasiti and
degetasits), which has led some to the theory that the Umbro-Oscan mediae
were not voiced (Conway, Amer. Journ, Phil. xi. 306), while others refer the
variation to the defects of the Umbrian and Oscan alphabets, which being
derived from the Etruscan had not originally the means of distinguishing
fully the tenues: from the mediae (the Umbrian alphabet, for example, uses
the tsign for both ¢ and 4d, the #-sign for both k hnd g, &e., see von Planta,
Gramm. od. Onibr, Dial. i. p. 547). In a Falisco-Latin inser. (Zvet. I. I. I. 72)
we have gondecorant, gonlegium, beside communia,...

i
78 THE LATIN LANGUAGE. (Chap. II,

§ 78. B, P. Latin 4, p were labial mutes, apparently with the
same sound as J, p in Ital., e.g. bene (Lat. déxé), pino (Lat.
pinus), and English 6, p, Between vowels 4 became in course
of time a labial spirant, and by the third cent. a p. became
identified with Latin v (w) (see § 52). In Spanish, 4 has
a w-sound, which differs from our w, in that the back of the
tongue is not raised, nor the cheeks narrowed. It is the same as
the w of South German wie, wein.

B was often written, though » was pronounced, before s, ¢, in
such words as wrbs, obtineo. It was a frequent subject of discus-
sion among Latin grammarians whether these words should not
be spelt with py, so that the spelling might agree with the
pronunciation. The d was defended in urds, &c., on the ground
that it would be unreasonable to spell a nominative case with
p and the other cases with 4, urbis, wrbi, &c., and in obtineo, &e.,
because the form of the preposition when alone, and often in
composition, had the 4, 0b, obdiico, &e. The -é of the preposition
similarly assimilated itself in pronunciation to m in compounds
like submitto, summitto. In Gmitto all traces of this 0 have
disappeared (see ch. iii.).

Latin 6 represents an I.-Eur. aspirate in words like rider,
stem rubro-, .-Eur. *rudh-r0, Greek é-pv@pés. In these cases
J in various parts of Italy corresponded to Latin 4, e.g. Umbrian
rufro-. This dialectal f for 4 is seen in forms like s#/tdus, beside
genuine Latin sid7/us, forms which do not prove anything about
the pronunciation of Latin 4, but are merely corresponding words
to the Latin, which have come from some dialect or other. A good
many of these dialectal #forms have found their way into the
Romance languages.

§ 79. Phonetic descriptions of b, p. To the usual phoneticians, Ter.
Maur. vi. 331. 186-193 K. :

b littera vel p quasi syllabae videntur
junguntque sonos de gemina sede profectos :
nam muta jubet portio comprimi labella,
vocalis at intus locus exitum ministrat.
compressio porro est in utvaque dissonora ;
nam prima per oras etiam labella figit,
velut intus agatur sonus; ast altera contra
pellit sonitum de mediis foras labellis ;
§§ 78-81.] PRONUNCIATION. LARBIALS. 79

Mar. Vict. vi. 33. 15 K. (whose b seems to be the » of Ter. Maur.) b et p
litterae conjunctione vocalium quasi syllabae (nam muta portio penitus latet:
neque enim labiis hiscere ullumve meatum vocis exprimere nisus valet, nisi
vocales exitum dederint atque ora reserarint) dispari inter se oris officio
exprimuntur, nam prima exploso e mediis labiis sono, sequens compresso ore
velut introrsum attracto vocis ictu explicatur; Martianus Capella iii, 261
B labris per spiritus impetum reclusis edicimus . . P labris spiritus [spiritu
Eyss.) erumpit ; we may add the remark of Terentius Scaurus vii. 14. 3 K.
b cum p et m consentit, quoniam origo earum non sine labore conjuncto ore
respondet.

80. bs, bt. Latin bs had the sound of Greek y (Vel. Long. vii. 61 K.), and
was one of the sounds for which the Emperor Claudius proposed a new letter,
on the ground that a separate sign for-cs (x) justified a separate sign for ps.
The general opinion however pronounced this new letter unnecessary. Some
even went so far as to question the necessity of x (Quint. i. 4. 9 nostrarum
ultima [sc. x], qua tam carere potuimus, quam psi non quaerimus). The
spelling abs is defended on the strength of ab by Velius Longus (vii. 61 K.),
who also tells us that some authorities always spelt opstitit, absorpsi, wrps, nupsi,
pleps (id. vii. 64 and 73-4 K.) (cf. Mar. Vict. vi. 21. ro K.: Ter. Scaur. vii. 14.7;
21.8K.). It was Varro who laid down the rule that nouns with -p- in the
Genitive should have -ps in the Nominative, nouns with -b- should have -bs,
e.g. Pelops, Pélipis, but plebs, plébis, urbs, urbis (Ter. Scaur. vii. 27. 11 K.; ef.
Varro, L.L. x. 56). [So in the Appendix Probi: (198. 4 K. and 199. 4) celebs
non ‘celeps’; (199. 3) plebs non ‘pleps’; (199. 11) labsus non ‘lapsus.’]
That obtinuit was pronounced optinuit, we are told by Quintilian (i. 7 7) :
secundam enim hb litteram ratio poscit, aures magis audiunt p. The
spellings ps, pt are common enough in MSS. and inscriptions (see Indices to
C.I.L.) (ef. Obseus, Opscus and Opicus, old forms of Oscus (as supscribo, subscribo
of suscribo), Fest. 212 and 234 Th. ; and see Georges, Lex. Wortf., s.v. cambsi,
campsi]. Curtius Valerianus (ap. Cassiod. 157 K.) says ps belong to the same,
bs to different syllables. In the Lex Col. Jul. Genetivae Urbanorum of 448. c.
op- is used in opsaepire, optemperare, opturare, optinere, but always ab-, never ap-
(Eph. Epigr. ii. pp. 122 and 221). Absinthium (vulgar absentium), absida late Lat.
for apsis follow the analogy of abs.

§ 81. ps, pt. In Romance the sounds of Latin pt and ps have been retained in
Roumanian, but in other languages have passed into #, ss, e. g. Ital. sette (Lat.
septem>, cassa (Lat. capsa), esso (Lat. ipst), medesimo (Vulg. Lat. *met ipsimus,
O. Fr. medesme, Fr. méme}. Isse for ipse found its way into colloquial Latin
(see Georges, Lex. Wortf.s.v.), though, if the story mentioned, but discredited,
by Suetonius (Aug. 88), be true, the use of issi (or iat ?)} for ipsi by a ‘ legatus
consularis ’ led to his being cashiered by Augustus as ‘rudis et indoctus.’ Cf.
sussilio for supsilio, subsilio ; and for some examples of ss for ps in MSS. and late
plebeian inscriptions, see Schuchardt, Vok. i. 148 ; for uy t for pt, ib. i. 143, and
see Georges 8. VV. scratia, septimus. The lap-dog, the subject of one of Martial’s
prettiest epigrams (i. 10g) was called Issa (i.e. ¢psa in the sense of domina),
‘M’lady.’ Its master had made a painting of it : inquatam similem videbis
Issam, Ut sit tam similis sibi nec ipsa.

1 Plautus puns on opsecro and mox seco, Mil. 1406.
80 THE LATIN LANGUAGE. [Chap. II.

§ 82. bm, mb. Ommentans was the spelling in a line of Livius Andronicus’
translation of the Odyssey (ap. Fest. 218. 14 Th. aut in Pylum deuenies aut
ibi ommentans. Cf. Gl. Plac. ommentat: expectat). Amnégo occurs on
inscriptions (C. I. L. vi. 14672) ; amnuo in Glosses (Léwe, Prodromus, p. 421).
(On dmitto, submitto, pronounced summiitto, &e., see Brambach, Hiilfsb.’ pp. 16-18.
On mb becoming dialectally mm, see § 71.)

§ 83. b and dialectal f. Alfius was the dialectal, Albius the Latin form of
the name. The two forms are found, for example, on Interamna inscriptions
(Albius, C. I. L. xi. 4240, Alfia, 4242). So with other proper names like Orbilius
and Orfilius. Sifilus, a mispronunciation of stbilus, censured in the Appendix
Probi (199. 3 K.; ef. Non. 531. 2), was a dialectal variety ; similarly scrofa
a sow has been connected with scrébis. In glosses we find crefrare with cribrare,
bufus with bibo (Lowe, Prodr. p. 421), and in modern Italian sufilare (cf. Fr.
siffler) beside sibilare (Lat. sibilare), tafano (Lat. tdébdnus), &c. (other examples
in Arch. Glott. Ital. x. 1).

§ 84. band m. B became m in Latin before m, n (ef. swmmitto, amnego, above).
But glimus and glibus (ef. Probi App. 198. 8 globus non ‘glomus’) are two
different stems, globus, -i and glomus, -eris (see Rom. Forsch. vii. 217).

§ 85. D, T. We have clear evidence that Latin ~, the dental
nasal, was, like our 7, not a pure dental (see § 61). The dental
mutes, d@ the voiced dental, ¢ the unvoiced, cannot then have been
pure dentals either. The Latin phoneticians speak of Latin d and
tas differing in more respects than the mere presence or absence
of what is technically called ‘ voice’; though the suspicion under
which they stand of being unduly influenced by their Greek
authorities makes them uncertain guides, In Italian. ¢, e.g. tu
(Lat. ta), d, e. g. dono (Lat. dézo) axe both pure dentals, differing
like any other unvoiced and voiced mute. But there is on
Italian soil a curious sound, a cacuminal /, exemplified by Sicilian
cavaddu (Lat. céballus), on which see Meyer-Libke, Ital. Gram.
§ 264.

Both / and + are sounds closely related to d, the position of the
tongue, &c., being very similar in the formation of all three
sounds. In d there is a complete closure of the mouth passage ;
in 2 the middle of the passage is closed, but the sides are left
open ; in 7 there is an opening in front at the tip of the tongue.
Through neglect of the side closure 7 has passed into 7 in words
like dicriima (older dacruma ) (see ch. iv. § 111); through neglect
of the front closure it passed into * in Old Latin in words like
arfuise (later adfuisse) on the Senatus Consultum de Bacchanalibus
(C.1. L. i. 196, of 186 B.c.) before the bilabial spirants f and r
§§ 82-85.] PRONUNCIATION. DENTALS. 81

(see ch. iv. § 112), A fifth (?) century grammarian speaks of the
mispronunciation peres for pedés as one specially affected by the
poorer classes at Rome in his time (Consentius v. 392. 15 K.) ;
and the same change of sound is still found in dialects of Italy.
In Naples, for example, pere is the word used for ‘ foot’ to this
very day. Before 7, d seems to have been changed to #, e.g.
ater, stem atro- for *ddro (see ch. iv. § 113), a tendency seen in
spellings like Alewanter, Cassantra, which Quintilian tells us he
had noticed on old inscriptions at Rome (i. 4. 16), and which is
found on Praenestine cistae (see § 73). Of dr, d/, &c. we are told
‘nullo modo sonare d littera potest’ (Cassiod. 151 K.; 207 K.).
Before /, ¢ could not be pronounced, but passed into the sound of
ec, just as the phrase ‘at least’ often takes with us the sound
‘acleast.’ The J.-Eur. suffix -¢/o had on this account become -c/o in
Latin words like périclum (ch. v. § 25); and when at a late period
the Latin suffix -¢wdus became contracted by the syncope of the
penult, it was changed to -clus, vétilus, for example, becoming
veclus (Ital. vecchio). Another change of /, namely its tendency
to be assimilated by a preceding ” in words like distenno for
distendo is discussed in § 71, and its assimilation in compounds
like adtineo, pronounced attineo, adsum, pronounced assum, in ch. iv,
§ 160. The most important changes of 7, ¢, however, are those
which these letters experienced when they were followed by 7
before another vowel. The same syncope that reduced vetulus to
veclus, cilida to calda, made ‘ Tityus’ out of Titius, < hodye’ out
of Addie. Through this combination of y with a preceding
consonant in unaccented syllables, a new series of sounds,
unknown in Latin, has arisen in Romance. Latin simia has
become French singe (through *simya), Latin apiwm Fr. ache
(*apyum), Lat. rébies Fr. rage (*rabyes), Lat. cambiare Fr. changer
(*cambyare). Dy, as we saw before (§ 51), became identified with
gt, ge, and Latin 7 (our y), and has assumed in Italian the sound
of our j, e.g. Ital. giorno from Lat. diurnus; while ty has
developed in Italian into the sound of ¢s, a sound reduced in
French to an s-sound, in Spanish to a sound like our // in ‘thin’
(written in Spanish z), e.g. Ital. piazza, Fr. place, Span. plaza,
all from Latin plitea, *platya. The grammarians of the later
Empire have fortunately left us a good many remarks on the
G
82 THE LATIN LANGUAGE. (Chap. II.

palatalization of ¢, so that we can trace pretty clearly the course
of its development in Latin. It seems from their accounts to
have begun in the fourth cent. a.p., and to have been fairly
established by the fifth ; and this is confirmed by other evidence,
such as the fact that in the Latin loanwords in Welsh (borrowed
during the Roman occupation of Britain which ceased in the fifth
cent.), ¢y has not become an s-sound. About the same time cy
became assibilated; and so confusions of -ci- and -ti- before
a vowel are common in late inscriptions and in MSS.

§ 86. Phonetic descriptions of d, t. Terentianus Maurus makes the back
of the tongue come into play in the formation of d, which would make Latin
d to have been what phoneticians call ‘dorsal’ d (from Lat. dorsum, the back),
like the d of Central and 8S. Germany. Seelmann understands ¢,1,7, and
also to have been dorsal sounds. Ter. Maur. vi. 331. 199-203 K. :

at portio dentes quotiens suprema linguae
pulsaverit imos modiceque curva summos,
tune d sonitum perficit explicatque vocem ;
t, qua superis dentibus intima est origo,
summa satis est ad sonitum ferire lingua.

Similarly Marius Victorinus speaks of the two sounds as having marked
difference in their formation (vi. 33. 24 K.): d autem et t, quibus, ut ita
dixerim, vocis vicinitas quaedam est, linguae sublatione ac positione distin-
guuntur. nam cum summos atque imos conjunctim dentes suprema sui parte
pulsaverit, d litteram exprimit. quotiens autem sublimata partem, qua
superis dentibus est origo contigerit, t sonore vocis explicabit. They repre-
sent the formation of ¢, in conformity with what we have already learned about
n, as the contact of the tongue with the alveolars, or gums of the upper teeth,
whereas in uttering @ both the lower and the upper teeth are touched by
the tongue, which is so bent down as to touch the lower teeth with its tip,
and the upper with its blade. Martianus Capella (iii. 261) : D appulsu linguae
circa superiores dentes innascitur. . . T appulsu linguae dentibusque impulsis
extunditur [extruditur Zyss., extuditur MSS. ].

§ 87. dand1. In some Italian dialects @ in the Latin suffix -idus becomes 1 if
the stem ends in a labial. Thus Lat. pidus is in the Neapolitan dialect
tiepolo. Some examples of / for din MSS. and late inscriptions are collected
by Schuchardt, Vok. i. 142.

§ 88. dand r. In the Abruzzi (the ancient country of Oscan and Sabellian
tribes) we find d (English tz in ‘there’) andr for Latin d, e.g. dicere and
ricere (Lat. dicere), da and ra (Lat. dat). The close connexion of é with r, as
phases of d, we see from Spanish, where in the literary language d has assumed
the @-sound in words like ‘ Madrid,’ while in the Andalusian dialect this ¢ has
sometimes developed into 7,'e. g. soleares, sometimes been dropped e. g. naa,
for *nada. Final @ is weakly pronounced in Spanish, and often dropped ; and
the same is true of the Galician dialect of Portuguese, e.g. bondd (Lat.
§§ 86-90.] PRONUNCIATION. DENTALS. 83

bénitatem). In Provengal too Latin d became @ and was dropped when final.
All this throws light on the Umbrian treatment of I.-Eur. d, which in the
middle of a word is expressed by a peculiar sign in the Umbrian alphabet
(conventionally written @ or *), a sign rendered in Latin characters by rs, e.g.
kapide, capirse (Lat. ciipidi, Dat. of capis, a bowl), and which seems to inter-
change with 7, e. g. tertu and tedtu (Lat. dédto, or rather *dédato), but which at
the end of a word is often dropped, e. g.asam-a and asam-ad (Lat. ad dram, or
rather *aram ad), always when a long vowel precedes, e. g. pihaclu (Lat.
pidctlo, older prdciléd, Abl.). Not unlike is the Latin treatment of d, with the
occasional change to ron the one hand, and the loss of final d after a long
vowel [e. g. piacold(d), but quéd] on the other. In Italian d is always dropped in
words like fe (Lat. fides), and in the other Romance languages d between any
two vowels is liable to the same thing; e.g. Latin médulla is in Spanish
meollo, in French moelle, though in Italian it is midolla; Italian preda (Lat.
praeda) is in Sardinian prea, &c.

§ 89. tl. Veclus for vétilus, viclus for vitilus, capiclum for cdpitilwn, were
mispronunciations in vulgar speech (Probi App. p. 197. 20 and 198. 34 K.).
Stlis the old form of lis (Quint. i. 4. 16) is spelled sclis on inscriptions (e. g. C. I. L.
X. 211 and 1249). Caper censures the use of sclataris for sildtaris, a pirate ship,
marculus for martilus, a priest of Mars (vii. 107. 1; 105.21 K.). (For examples
of cl for td in MSS. and late inscriptions see Schuchardt, Vok. i. 160.)

§ 90. Assibilation of ty, dy. In the fourth cent. we have an indication that
ti, di before a vowel were in process of change. Servius (in Don. iv. 445.
8-12 K.) tells us that they often pass into a sibilus (which need not imply
an s-sound) when in the middle of a word, though often they retain
a pronunciation in accord with their spelling (etiam sic positae sicut di-
cuntur ita etiam sonandae sunt, ut ‘dies’ ‘tiaras’). The same grammarian,
in a note on Virgil, Georg. ii. 126, remarks that the Greek word Média must be
pronounced in Greek fashion sine sibilo, that is to say without that consonantal
y-sound which Latin médius, media had in the time of Servius, that ‘ pinguis
sonus’ of i which the grammarians, as we saw before (§ 14), declared to be
particularly alien to Greek pronunciation. In the early part of the fifth cent.
Papirian (MS. ‘Papirius’) is more explicit. The letters ti before a vowel, in
words like Tatius, dtia, justitia, have, he says, a sound as if 2 (i.e. Greek ¢,
which had at this time the soft or voiced s-sound) were inserted between
‘them (ap. Cassiodor. vii. 216. 8 K.): ‘justitia’ cum scribitur, tertia syllaba
sic sonat quasi constet ex tribus litteris t, z, i. This, he points out, is the
case only when ti is followed by 2 vowel, and not always even then, not, for
example, in Genitives like ofii nor when s precedes ti, e. g. justius, castius. In
the same century Pompeius censures as a fault the very pronunciation allowed
in the preceding century by Servius, whereby ti, di were pronounced as spelled.
He lays down the rule (vy. 286. 10 K.): quotienscumque post ti vel di
syllabam sequitur vocalis, illud ti vel di in sibilum vertendum est... ergo si
volueris dicere ti vel di, noli, quem ad modum scribitur, sic proferre, sed sibilo
profer. He goes on to say that this pronunciation is not found with initial
#, dé; nor with the combination sti; for here ‘ipsa syllaba a litteris accepit
sibilum,’ a remark which shows pretty clearly that Pompeius understands
by sibilus an s-sound, cf. v. 104.6 K. si dicas ‘Titius,’ pinguius sonat [i] et
perdit sonum suum et accipit Sibilum. A grammarian of the fifth cent.(?),

ce
84 THE LATIN LANGUAGE. [Chap. IT.

Consentius (395. 3), describes the new sound of ti in étiam as ‘breaking some-
thing off the middle syllable’ (de media syllaba infringant). He, like
Pompeius, declares the old pronunciation to be a ‘vitium,’ and tells us that the
Greeks in their anxiety to correct this fault were apt to go to the extreme of
giving the new sound to ti even when not followed by a vowel, e. g. in optimus
Gnediam syllabam ita sonent quasi post t, z graecum ammisceant). Finally
Isidore in the seventh cent. tells us (Orig. i. 26. 28) that justitia ‘sonum 2
litterae exprimit,’ and (xx. 9. 4) that the Italians of his time pronounced
hodie as ozie. The spellings on inscriptions confirm this account of the
grammarians, though, as was to be expected, the assibilation shows itself
on plebeian inscriptions even earlier than the fourth cent., e. g. Crescentsian(us)
(Gruter, p. 127, vii. 1, of r40 a. D.), and even in the case of accented ti, di;
Isidore’s statement about the pronunciation of hodie (now oggi) is perhaps
supported by oz (C. I. L. viii. 8424): z (=zes, for dies) (C. I. L. v. 1667), &e.,
this z being pronounced like our z in ‘amaze.’ But dy- first passed through
the stage of y, unlike ty- (see ch. iv. § 62), and this z may be merely an attempt
to express the y-sound. Seelmann, p. 323, gives a list of these spellings. Some
may be dialectal, for in Osean we have on the Bantia tablet (Zv. 231) Bansa-
(Lat. Bantia), zicolo- (Lat. diecula) (pronounce 2 as above); and in Etruria the
assibilation of ty seems also to have been known (see Sittl, Lok. Verschiedenheiten,
p. 11); Marsus (cf. Martses Abl. Pl. on « Marsic inscr.) was the native name
for Martius. The rationale of the change of sound is easy. While forming
the ¢t-sound the tongue unconsciously adapted itself to the position for the
y-sound, so that the interval between the two letters was bridged over by
a glide-sound which the Latin grammarians compare to Greek ¢, like the
connecting p in the group mpt, from original mt, in words like emptus. (For
a full account of the process see K. Z. xxix. 1 sqq., especially p. 48. On the
interchange of ti- and ci-, see § 94.)

§ 91. K, C, G, QU, GU. What we call Guttural Consonants
are more properly divided into (1) Gutturals proper, or Velar
Gutturals, or simply ‘ Velars,’ formed by the back of the convex
surface of the tongue against the soft palate or velum, and
(2) Palatals, formed by the middle of the convex surface of the
tongue against the hard palate; and these two classes, which are
also called back gutturals and front gutturals, might be still
further subdivided according as the sound is made more to ‘the
back, or more to the front of the mouth. The Velars and Palatals
may be found side by side in a language. German ch, for
example, with a broad vowel like 0, a, is a velar, e.g. ‘ach,’ but
with a narrow vowel like 2, in such a word as ‘ich,’ it is
a palatal, being spoken more in the front of the mouth, so that
it often sounds like English s4. Italian cf of chi, chiesa, is
spoken more in the front of the mouth than ¢ of casa, and the
same is true of Engl. & of ‘ key” compared with ¢ of ‘caw.’ The
§ 91.] PRONUNCIATION. GUTTURALS. 85

distinct lines of development which the Latin gutturals, ¢ (A), g,
took before broad and before narrow vowels, makes it possible, or
even probable, that in Latin, as in Italian, ¢ in centwm had a more
palatal sound than ¢ in cantus, contus, &c., although this distine-
tion is not mentioned by any of the Roman grammarians. The
only guttural of which they give us a clear account is qu, in
which the w-element seems to have been more of a vowel than
Latin v (our w). A first century grammarian (Vel. Long.
vi. 58. 17 K.) makes the difference to consist in the latter being
sounded ‘cum aliqua aspiratione,’ i.e. as a consonantal spirant,
not as a half-vowel, like ~ of quis, and in the fourth century w of
guoniam, quidem is said to be ‘ nec vocalis nec consonans’ (Donat.
iv. 367. 16 K.). Priscian (seventh cent.) says the same of the
u of sanguis, lingua (i. 37), so that Latin gu, gu must have had
a sound very like their sound in Italian quattro, &. The
palatalization of ¢, g before a narrow vowel is found in all
Romance languages, with the exception of a dialect in the island

of Sardinia. It was also a feature of the Umbrian language, so ~

that we should expect it to have appeared early in Vulgar Latin
at least. But all the evidence points to as late a period as the
sixth and seventh centuries A.D. as the time when the change of
sound took place. No grammarian hints at a difference of sound
in ¢, g before a broad and before a narrow vowel, although the
assibilation of ¢ before a vowel is mentioned again and again.
Greek transcriptions of Latin words with ¢ invariably reproduce
it by «, in cases like KHNZON for censwm, KPHEKHN® for
crescens ; Latin loanwords in Welsh (first to fifth centuries) show
that Latin ¢ was hard in all positions, e.g. Welsh cwyr (Lat.
céra), ciwdawd (Lat. civitdtem), and similarly German Keller
(Lat. celldrium), Kiste (Lat. cista); it is not till the seventh
century that spellings like paze for pace (Muratori, 1915. 3) assert
themselves on inscriptions. At an earlier period, it is true, cz (cy)
before a vowel in unaccented syllables, and ¢ (¢y) in the same
position, had been confused, so that ci was written ¢ in words
like sd/dtiwm, and expressed like genuine ¢i a sibilant sound.
But this proves nothing for ¢ in words like centum, citra. G
before ¢, 2 became (like dy) the y-sound, and is in the Romance
languages indistinguishable from Latin 7 (our y) (see § 51). The
86 THE LATIN LANGUAGE. (Chap. IT.

group ct has become /#é in Italian, e.g. Ottobre, Lat. Octodris,
and had assumed the sound in late Latin, to judge from spellings
on inscriptions like Jattuca in the Edict of Diocletian, Otodris
(Rossi 288, of 380 a.p.) and autor.

§ 92. Phonetic descriptions of the Gutturals. Ter. Maur. vi. 331. 194-205 K. :
utrumque latus dentibus applicare linguam
¢ pressius urget: dein hine et hine remittit,
quo vocis adhaerens sonus explicetur ore.
g porro retrorsum coit et sonum prioris
obtusius ipsi prope sufficit palato.
. .. k perspicuum est littera quod vacare possit
et q similis; namque eadem vis in utraque est.

Mar. Vict. vi. 33. 20 K. ¢ etiam et g, ut supra scriptae, sono proximae oris
molimine nisuque dissentiunt. nam ¢ reducta introrsum lingua hine atque
hine molares urgens haerentem intra os sonum vocis excludit: g vim prioris
pari linguae habitu palato suggerens lenius reddit... [q, k] quarum utram-
que exprimi faucibus, alteram distento, alteram producto rictu manifestum est.
Mart. Cap. iii. 261 ; G spiritus [facit] cum palato...K faucibus palatoque
formatur ... Q appulsu palati ore restricto. Ter. Scaur. vii. 14. 1 K. x littera
cognata est cum ¢ et g, quod lingua sublata paulum hae dicuntur. Bede
(228. ar), in criticizing Donatus’ remark, quoted above, on the pronunciation
of win qu, explains him to mean that ‘tam leviter tum effertur ut vix sentiri
queat.’ Pompeius (v. 104. 25 K.) calls the ~ a ‘pars litterae praecedentis.’
Priscian (i. 6) seems to speak of it as the ‘contractus sonus’ of normal u;
but the passage is corrupt and the meaning uncertain. The statement of the
phoneticians that Latin ¢ was uttered with more energy of articulation than
g, is confirmed by certain phenomena of the language, as has been shown
in § 73. ;

§ 93. qu, gu. In Oscan and Umbrian, where I.-Eur. qu had become p, Latin
qu is expressed in loanwords by kv (Osc. kvaisstur, Umbr. kvestretie (Lat.
quaestiivae]). In Faliscan the qu-sound is written cv (cu or cv) e.g. cuando.
Greek transcriptions have normally xov-, e.g. Kovadparos: but «o- is the
earliest expression of qui-, e.g. Kowsxtios (C.I.G. ii. 770, of 196-4 B.c., see
Eckinger, p. 120 sqq.). In the Augustan age when o before a final consonant
was weakened to u even after v, u (ch. iv. § 20), qu, gu became before this u
reduced to ¢, g, which points to their being more like cu, gu than cw, gu, relicus
from reliquos (in the time of Plautus rélicuos, of four syllables), lécuntur, sécuntur,
extingunt. The grammarians of the first cent. a. D. were puzzled by the want
of correspondence between Nom. Sing. écus, Nom. Pl. equi, and reconstituted
the Nom. Sing. as equus (in the time of Trajan) (Vel. Long. 59. 3 K. auribus
quidem sufficiebat ut equus per unum u scriberetur, ratio tamen duo exigit) ;
guu in extinguunt, &e., followed somewhat later. In the fifth cent. we find
co definitely ousting quo, and go, gud, though qué seems to have been pro-
nounced cd as early as the beginning of the second cent. B.c. (see ch. iv.
§ 137). Grammarians find great difficulty in deciding which verbs should be
written with -guo and which with -go. The rule they usually follow is
to write -go when the Perfect has -xi, ungo, tingo (see Bersu Die Gutturalen’.
(A Vulg. Lat. *laceus for ldqueus is the original of Romance words for ‘noose’
§§ 92-94.] PRONUNCIATION. GUTTURALS. 87

like Ital. laccio, Fr. lacs; cf. Probi App. 197. 27 K. exequiae non ‘execiae’;
Cassiod. 158. 15 K. on ‘reliciae’). The spelling of the Pronoun gui in its
various forms was also matter of discussion as early as the time of Quintilian.
He tells us (i. 7. 27) that in his younger days the Dative Singular used to
be written quoi to distinguish it from the Nom. qui, but that the fashion had
since come in of spelling it cui: illud nunc melius, quod ‘cui’ tribus quam
posui litteris enotamus, in quo pueris nobis ad pinguem sane sonum qu et
oi utebantur, tantum ut ab illo ‘qui’ distingueretur. Annaeus Cornutus,
Persius’ teacher, tells us that at a much earlier period Lucilius laid down
the rule that qu should be used when a vowel followed in the same syllable,
otherwise cu, and this rule he himself accepts. His comments on it seem to
show that there was not much difference in the sound. ‘Some,’ he goes on
to say, ‘think we should spell as we pronounce, but I do not go so far as that’
(ego non omnia auribus dederim). Then he adds, ‘ qui’ syllaba per q u iscri-
bitur ; si dividitur, ut sit cui ut huic, per ¢ (ap. Cassiod. 149. 1 K.). So Ter.
Seaurus (first cent.), 27. 18 K. quis quidem per ‘cuis’ scribunt, quoniam
supervacuam esse q litteram putant. sed nos cum illa u litteram, si quando
tertia ab ea vocalis ponitur, consentire jam demonstravimus. ec autem in
dativo ponimus, ut sit differentia cui et qui. Velius Longus (first cent.)
(75. 10 K.) thinks it necessary to point out the distinction between dquam
Noun and dcuam Verb. Another tendency that appears in late Latin is to
make a short vowel before gu long by position, as indeed any consonant
followed by u'(w), e.g. Ital. Gennaio with double n from *Jenwarius; aqua
is scanned with the first syllable long by the Christian poets, and appears in
Ital. as acqua (cf. Probi App. 198. 18 K. aqua non ‘acqua’). A sixth cent.
grammarian ventures to give this quantity to the word in a line of Lucretius,
vi. 868 quae calidum faciunt aquae tactum atque vaporem, where, however,
the MSS. read Jaticis. Lachmann proposed to read agiiae of three syllables, but
was not able to prove that this form (like Horace’s siliiae) existed in Old Latin
(see Schroeder in Studemund, Studien, ii. 20). In Plautus and the older
dramatists, where the short syllable of a word like péti, léci has a shortening
influence on the following long syllable, so that the words may be occasionally
scanned péiti, lici (see ch. iii. § 42). a short vowel before qu seems hardly to have
had this shortening power, e. g. rarely (if ever) léqui. So to the ear of Plautus
qu almost made a preceding vowel long by position, unless we say that loqui,
&c., sounded to Plautus something like a trisyllable. At any rate qu can
hardly have had merely the ‘rounded ’ k-sound of Russian.

§ 94. c, g before narrow vowels. That ¢, g remained hard before e, i, &c.
(when a vowel did not follow), down to the sixth and seventh centuries 4. D. we
have a superabundance of proof. For the earlier period we may point to the.
fact that in Umbrian, where ¢ («) before a narrow vowel became a sibilant,
expressed by a peculiar sign in the native alphabet, the Latin ¢ was not used
for this sound in inscriptions (from the time of the Gracchi) written in Latin
characters, but a modification of s, namely s with a stroke like a grave accent
above it, e.g. deken (Lat. décem), sesna (Lat. cena). That Plautus (who by the
way was an Umbrian) makes a play on the words Sdsia and sicius, proves

nothing (Amph. 383) :

Amphitruonis te ésse aiebas Sésiam.—Pecedueram :
nam ‘Amphitruonis sécium’ dudum me ésse volui dicere.
88 THE LATIN LANGUAGE. [Chap. II.

He makes a play on arcem and arcam in Bacch. 943: atque hie equos non in
arcem verum in drcam faciet impetum.

At Cicero’s time the spelling pulcher, Gracchi with ch for supposed Greek x is
evidence that in declension of nouns and adjectives (acer, acris, &c.) the c did
not change toa sibilant when it came to stand before an e or an i, as it
does in Italian (amico with hard ¢, amici with sibilant c). Varro (ap. Prise.
i. 39) quotes agceps (another spelling of anceps) as one of the words where the
Agma-sound (the ng of ‘thing’) was found in Latin before c (therefore pre-
sumably hard c), In the first cent. a.p. Plutarch and Strabo render Latin ¢

before a narrow vowel by Greek «, Kixépwy, &. None of the grammarians of
the Empire hint at a variety of pronunciation for c, g, not even Priscian in
the sixth cent. ; and all through this period we have Greek « for Latin c in
all positions (on documents of the sixth cent. Sexip, dwvarpim, &e.), and on
Latin inscriptions an interchange of ¢, k, ¢ (e. g. pake, C. I. L. x. 7173: cesquet
for quiescit, viii. 1091) (see Seelmann, pp. 342 sqq.). This interchange is not
regulated by any principle. We do not find & used for ‘hard ¢,’ c for ‘ soft c’
&e., as would have been the case had there been a real difference of pro-
nunciation. All the examples quoted for interchange ofc before a narrow
vowel (not in hiatus) with a sibilant earlier than the sixth cent. in S. Italy,
the seventh cent. in Gaul, are illusory (see G. Paris in Acad. Inscr, 1893, Comptes
Rendus, xxi. p. 81). ;

The evidence that Latin c was what we call hard ¢ before e, i down to
a late period is thus overwhelmingly strong. But while holding to this fact
we may make two concessions. First, that c before e, i was probably more of
a palatal (like Italian ch in chiesa) than a velar (like Italian ¢ in casa). This
palatal character was more and more developed in the Romance languages till
c became a sibilant. Since however this assibilation is not known in the
Sardinian dialect of Logudoru, it may be that at the time of the occupation
of Sardinia (c. 250 B.c.) Latin c had still a velar character before narrow as
before broad vowels. Second, that ci (ce) before a vowel underwent the samc
process of assibilation, as ti before a vowel did in the fifth cent. a.p., although
interchange of spelling between prevocalic ci and ti before that time means
merely that cy, ty were confused, as cl, 41 were confused (cf. Quint. i. 11. 6),
not that both cy and ty expressed a sibilant sound. (For instances of the con-
fusion see Schuchardt, Yok. i. pp. 154 sq., and consult Georges and Brambach
s. vv. Mucius, mundities, negotiwm, otium, nuntius, Porcius, propitius, provincia,
spatium, Sulpicius, indutiae, infitiae, condicio, contio, convicium, diciv, fetialis, solacium,
suspicio, uncia, &e. The earliest examples date from the second cent. a.p.) On
Greek inscriptions Latin ci and ti are similarly confused, the earliest example
being ’Apovi.avos (Any. iv. p. 104) of 131 4.p. In a Pisidian inscription
(Journ. Heil, Stud. iv. p. 26), of 225 a. D., with Mapaiavos, the o (written C) is no
doubt merely a confusion with the Latin letter c. :

G before e, i may have been a palatal, rather than a velar, even earlier than
c; for in Sardinian hard g is not preserved as hard c isin this position. At
what precise period it was developed to y we do not know. The Appendix

Probi mentions as a mispronunciation ‘ calcosteis,’ for calcostégis, though this

may be a case of that spirant pronunciation of Greek y, like Tarentine éAios
for ddtyos, Boeotian idy for éyw. In the Romance languages it is treated exactly
like Latin j (y), e.g. Ital. genero, Span. yerno (Lat. géner), like Ital. giace,

Span. yace (Lat. jacet). It is dropped between two vowels in spellings like
§§ 95, 96.] PRONUNCIATION. LIQUIDS. 89

vinti for viginti (C. I. L. viii. 8573), the precursor of Ital. venti, &c. ; so trienta,
(il. 5399), &e. (A.L.L. vii. 69. Seethe list in Schuchardt, Vok. ii. 461). Vulg.
Lat. *mais is seen in Fr, mais, Ital. mai, &. But this dropping of intervocalic
g is found also before other vowels in late inscriptionsand MSS., e. g. frualitas
for frugalitas (see Schuchardt’s list, Vok. i, 129), as in Vulg. Lat. eo (Ital. io,
&e.) for ego, just as in Italian and other Romance languages, every intervocalic
g when pretonic is dropped, e. g. Ital. reale from Lat. regalis. So g in this posi-
tion may have become a spirant (like g of German Tage), just as intervocalic
b became a w-sound in the third cent. (§ 78).

§ 95. ct, tt. For examples see Schuchardt, Vok. i. 134; Rhein. Mus. xlv. p. 493,
and consult Georges and Brambach s. wv. cottana, coturnix, setius (?), vettonica,
pittacium, brattea, salapitia, virecta. So nictio, to ‘give tongue,’ of a dog who has
picked up the scent, is spelled nittio in the lemma of Festus (p. 188. 1. 16 Th.),
where he quotes the spirited line of Ennius, Ann. 374 M. :

nare sagaci
Sensit ; voce sua nictit ululatque ibi acuta.
Autor, with autoritas, is censured in the Appendix Probi (198. 30 K.), and is
found on late inscriptions (C. I. L. viii. 1423; cf. xii. 2058, of q91 a.D.). For
net, which became nt by loss of the guttural, see § 70. So mulcta became multa
(Georges s. v.).

gm or at least Greek yz seem t6 have tended, like im, to the sound um. In
Vulg. Lat. sagma was *sauma (Prov. sauma, Fr. somme), Isid. Orig. xx. 16. 5
sagma quae corrupte vulgo ‘sauma’ dicitur (v./. salma; ef. Span. salma, Ital.
salma and soma). Cf. Probi App. 198. 11 pegma non ‘peuma.’ (For other
examples, see Schuchardt, Vok. ii. p. 499.) For gn, see § 144.

§96. L,R. The liquids ¢, 7 are, as we have seen (§ 85), closely
connected with the voiced dental mute d. The tongue has
a similar position with each of the three sounds; but while with
d the mouth passage is completely closed by pressure of the
point of the tongue against the front, and of the edges of the
tongue against the sides of the mouth, with / the sides are left
open, and with 7 the front. The connexion of the three sounds
in Latin is seen, as was before remarked, in the interchange of
a with Zin lingua, older dingua, &c., of d with 7 in arfuisse for
adfuisse, &c., and, as we may now add, of 7 with / in words like
caerileus for *caeluleus, not to mention occasional occurrences of
the dental nasal for 7 or /, like Vulgar Latin menetris for mérétria,
cuntellum for cultediwm.

The grammarians give us a good deal of information about
the pronunciation of 7 in different parts of the word. It had
a ‘pinguis sonus,’ or ‘plenus sonus,’ in two cases, (1) when it
ended a word, or syllable followed by another consonant, e.g.
sol, silva, albus; (2) in combinations like jl, cé, e. g. flavus, clarus.
go THE LATIN LANGUAGE. (Chap. II.

In contrast with this ‘pinguis sonus’ it has what is called an
‘exilis’ (or ‘tenuis’) ‘sonus’ (presumably its normal sound) in other
two cases, viz. (1) at the beginniny of a word, e.g. lectus, léna,
“ipus, and especially (2) when it ends one syllable and begins the
next, e.g. 2-/é, Métel-dus, al-a [pronounced with two ?’s as in
our ‘ mill-lade,’ ‘ hotel-landlord ’ (see § 127) ]. When we examine
the development of Latin 7 in the Romance languages we find
the explanation of this distinction. In Italian, for example,
Latin 7, when initial, or when repeated, has the normal /-sound,
e.g. lana, valle (Lat. va//is), pelle (Lat. pelZis), but after a con-
sonant 2 has become an /y-sound, now reduced to ¢ (the half-
vowel), e.g. chiaro (Lat. clarus), pieno (Lat. plénus), fiume (Lat.
jiimen). At the end-of a syllable before a consonant, it has in
most Romance languages been reduced to a w-sound, eg. Fr.
autre, Prov. autre, Span. otro (Lat. alter), and so in parts of Italy,
e.g. Sicil, autru, while in other parts it is represented by an
z-sound, e.g. aitro in the Florentine dialect. All this points to
Zin clérus, &c., and / in alter, &., having been pronounced with
what phoneticians call an ‘off-glide’ and an ‘on-glide’ clarus,
alter, which glides have been more and more developed in the
Romance languages, till they reduced, or even completely extin-
guished, the /-sound. In Sardinian, which reflects the oldest
type of Vulgar Latin, 7 remains unaffected after a consonant to
a large extent, e.g. klaru, plenu, flumen, so that this affection
of / may not have beer: begun in Vulgar Latin till about 200 B.¢.

With regard to 7, we should expect from the analogy of the
Romance laiguawes that Latin + was trilled (i.e. formed with
the tip of the tongue vibrating), like the German and Scotch 7,
not like English r in ‘red.’ This is confirmed by the Roman
name for 7, ‘littera canina,’ the growling letter, Pers. i. 109
(Latin /irrio must have expressed the sound better than English
‘growl’), and by Lucilius’ description of it as like the growl of
a lazy dog, or as he puts it, like ‘what care 1?’ in dogs’ lan-
guage (ix. 29, 30 M.):

r non multum abest hoe cacosyntheton atque canina
si lingua dico ‘nihil ad me.’

This rough sound of Latin 7 explains the reluctance of the

Romans to begin two successive syllables with a consonant
§§ 97, 98.] PRONUNCIATION. LIQUIDS. gi

followed by 7, a reluctance seen in forms like praestigiae for
praestrigiae, incrébui for increbrui, and in spellings on inscriptions
like propius for proprius. Before s, 7 was assimilated, e.g. riissus,
risus for riérsus, as we see from Plautus’ pun on Persa and
pessum (Pers. 740 Persa me pessum dedit). It was assimilated
too before 7, e.g. perdicio, pronounced, and often spelled, ped/icio.
Metathesis of 7 (and 7) was as common in bad Latin as in bad
English, as interpertor (? interpétror) for interpretor, coacla for
cloaca testify, and other mispronunciations censured by the
grammarians. Dr became 7 in Latin, e.g. drow (cf. dium) (ch. iv.
$113). Neither 7 nor / remained unaffected by the palatalizing
influence of y, that later sound of 7 in words like Janudrius, facio,
hédie, which worked so great a transformation of the language
in the later period of Roman history. The palatalization of
* led to its disappearance in Italian, e.g. Gennajo (Vulg. Lat.
*Jen(u)aryus), a process exemplified in earlier times by the form
peiuro for peritiro, while dy has become the 7 mouillée, written g/in
figlia, miglia (cf. our ‘ million’), bigliardo (our ‘ billiards’), and in
some dialects, e.g. the patois of Rome and the neighbourhood, has
sunk to y. Some spellings on late Latin inscriptions seem to be
precursors of these changes of ry, ly. (On Umbrian /-, see ch. iv.
§ 85.)
§ 97. Phonetic descriptions of 1. Ter. Maur. vi. 332. 230-234 K. :

adversa palati supera premendo parte
obstansque sono quem ciet ips: lingua nitens
validum penitus nescio quid sonare cogit,
quo littera ad aures veniat secunda nostras,
ex ordine fulgens cui dat locum synopsis ;

Mar. Vict. vi. 34. 10 K.: sequetur 1 quae validum nescioquid partem palati,
qua primordium dentibus superis est, lingua trudente, diducto ore personabit ;
Mart. Cap. iii. 261 1 lingua palatoque dulcescit.

§ 98. of r. Ter. Maur. vi. 332. 238, 239 K. :

vibrat tremulis ictibus aridum sonorem
has quae sequitur littera ;

Mar. Vict. vi. 34. 15 K. sequetur r quae vibrato +vocis palatum linguae fastigio
fragorem tremulis ictibus reddit; Mart. Cap. iii. 261 R spiritum lingua cris-
pante corraditur. Ter. Scaurus (13. 10 K.) mentions the connexion ofr and?
with d: item 1 et det r ets [inter se mutuis vicibus funguntur], cujus rei
maximum argumentum est, quod balbi, qui rexprimere non possunt, aut 1
dicunt aut s, nec minus quod capra per diminutionem capella dicitur et frater
fratellus. :
Q2 THE LATIN LANGUAGE. (Chap. II.

§ 99. The grammarians on the pronunciation of1. The earlies taccount
(ap. Prise. i. p. 29 H.) is that of Pliny the Elder, who gives / three varieties of
sound : (1) ewilis ; quando geminatur secundo loco posita ut ‘ ille’ ‘ Metellus’;
(2) plenus : quando finit nomina vel syllabas et quando aliquam habet ante se
in eadem syllaba consonantem ut ‘sol’ ‘silva’ ‘flavus’ ‘clarus’; (3) medius, in
other positions: ut ‘lectus’ ‘lectum.’ Similarly in the fifth(?) cent. Consentius
(v. 394 K.) makes only two divisions (1) pingwis: cum vel b sequitur, ut in
‘albo,’ vel ¢ ut in ‘pulehro,’ vel f ut in ‘adelfis,’ vel g ut in ‘alga,’ vel m ut
in ‘pulmone.’ vel p ut in ‘scalpro’ (2) exilis: ubicumque ab ea verbum
incipit, ut in ‘lepore’ ‘lana’ ‘lupo,’ vel ubi in eodem verbo et prior syllaba
in hac finitur, et sequens ab ea incipit, ut ‘il-le’ et ‘ Al-lia.” Not so clear is
his account of the two mispronunciations to which ille was liable. The Greeks,
he says, pronounce ‘ille mihi dixit’ swbtilius as if tlle had only one?; others
pronounce ‘ille meum comitatus est iter,’ or ‘illum ego per flammas eripui’
pinguius, ‘ut aliquid illic soni etiam consonantis ammiscere videantur.’
Possibly this means that the Greeks made the double | into one as we
do in pronouncing Italian (see § 127), and that others (e. g. Spaniards) gave
it the ly-sound that it now has in Spain, e. g. villa (Ital. villa), which is pro-
nounced like Italian viglia. Diomede (i. 453. 3 K.) remarks on the fault of
pronouncing J in léicem or almam ‘nimium plene.’ Servius in Don. iv. 445.
12-13 K. calls it a ‘labdacismus’ (mispronunciation of 1) to make a single J,
e. g. Liicius, too ‘tenuis,’ or a double 1, e.g. Métellus, too ‘pinguis.” Pompeius
(v. 286-287 K.) makes the same remark, and explains it thus: debemus
dicere ‘largus* ut pingue sonet ; et si dicas ‘lex,’ non ‘lex,’ vitiosa sunt per
labdacismum, item in gemino 1], si volueris pinguius sonare, si dicamus
‘Metellus’ ‘ Catullus,’ in his etiam agnoscimus gentium vitia; labdacismis
scatent Afri, raro est ut aliquis dicat 1: per geminum 1 sic locuntur Romani,
omnes Latini sic locuntur : ‘Catullus’ ‘Metellus.’ His explanation would
doubtless be intelligible to his auditors when accompanied by his oral examples
of the different sounds ; to us, who have to infer these, it is not so clear. All
that seems certain is that initial 7 had some distinction of sound from thel of
Métellus, Cdtullus, &c., but whether this distinction consisted merely in the
more emphatic articulation which every initial consonant received in Latin
or in some other modification, such as the slight on-glide, which initial J has
in the Gaelic language, and which makes a word like long, a ship (Lat. longa
se. navis), sound almost like ‘along,’ it is impossible to say (cf. § 117; ch. iv.
§ 149). The sound of the initial is described as ‘pinguis,’ in comparison with
the 2 of Metellus, but as ‘exilis’ in comparison with the 2 of cérus, alter, &e.
(See also Isid. Orig. i. 31. 8). In O. Engl. also there were three different
kinds of J, (1) deep gutteral 7, as in ‘chalk,’ (2) ordinary J, as in ‘field,’ (3)
palatal 7, as in ‘whi(l)ch’ (Paul’s Grund. i. p. 860). The fact that e could
become o before 7, but not before UJ, in Latin (ch. iv. § 10) suggests that normal
Latin 7 was deeper, or less palatal, than Ul.

§ 100. of r. Varro (L. L. iii. fr. p. 146 Wilm.) mentions the rough sound
(¢asperum’) of crua, dere, vépres (also crura), beside viluptas, mel, léna.

§ 101. Interchange of rand1. On the confusion of fldgro and frdgro, and
the misspelling of both as fraglo, see .A.L. LZ. iv.8. In Probi App. 201. 19
the distinction between the two words is carefully pointed out (cf. ib. 198. 9
flagellum non ‘fragellum’) (Ital. fragello). Pliny gave the rule for the
§§ 99-102.] PRONUNCIATION. LIQUIDS. 93

employment of the suffixes -lés and -ris, that -lis should be used when the
stem contained an 7, -ris when it contained an J, e.g. augitrdle, midldre (ap.
Charis. 135. 13; cf. Prise. i. p. 132 H.). So too the I.-Eur. suffix -tlo-, Latin
-clo-, became -cro- after a stem with 1, e.g. lavacrum, fulcrum, simulacrum. The
same tendency to dissimilation ig seen in Vulgar Latin forms like veltrahus,
beside veriridgus (see Georges s.v., and cf. Prov. veltres, O. Fr. viautre), and
pelegrinus (C.I.L. iii. 4222, &e.), from which come the Romance words, Ital. pelle-
grino, Fr. pélerin, our ‘pilgrim,’ &c.; in the mispronunciationstelebra (see Georges
s.v.), censured in Probi App. 198. 21 K.; in the spellings of MSS. and late
Latin inscriptions, collected by Schuchardt, Vok. i. 136 sqq. (cf. meletrix, Non.
202. 13; 318. 6); inspellings on Greek inscriptions like BapSAAea, BapBidAdos,
beside BaABiAAeia and BadBiAdAos, SeBAapios, MeAxoupravos, &c. (see Eckinger,
p. 107); in Romance forms like Ital. albero (from Lat. arbor), reclutare (our
‘recruit’), Mercoledi (from Mercitrii dies), urlare (from iliilare), which show
that the same process is going on in modern, as in ancient, times on Italian
soil. The town Cagliari in Sardinia was in Latin called Carales plur. or Caralis
sing.; but we find byforms Calaris and Cararis. From Crustuwmeria, or
Crustumium, we have the adjective-forms Crustiiminus and Clustuminus (Greek
Kpogropeva and KAovoroupeva, Eckinger, p. 107).

§ 102. Parasitic Vowel with 1, r. The sounds J, r were called ‘liquidae’
by Latin phoneticians, A, , v, p, bypai by Greek), because they united easily
with a preceding consonant. Cf. Mar. Vict. vi. 20 eaedem autem ‘liquidae’
dicuntur, quando hae solae [he includes m, n] inter consonantem et vocalem
immissae non asperum sonum faciunt, ut ‘clamor’ ‘ Tmolus’ ‘Cnosus’ (MSS.
consul] ‘ Africa.’ But in Latin, especially after the time of Plautus, there
was a tendency to facilitate the pronunciation of a mute followed by J, particu-
larly when post-tonic by the insertion of a vowel, written on early inscrip-
tions 0, later 7. Thus pd-clum, which is formed with the I.-Eur. suffix -dlo-
(ch. v. § 25), became po-colom, po-culum. These forms with the parasitic vowel
underwent at a later time the same process of syncope that reduced célida to
calda, Titius to Tityus, porcilus (from the stem porco- with the suffix -lo-) to porelus,
and so resumed their earlier appearance poclum, &c. Traces are not wanting
of the same parasitic vowel-sound showing itself between a consonant and 1.
The development in Romance of a word like patrem, suggests that it must in
Vulgar Latin have sounded almost like a trisyllable, *pat*’rem (Meyer-Liibke,
Rom. Gram. i. p. 251) ; and the same is indicated perhaps by Varro’s derivation
of Gracchus, or, as he spelt it, Graccus (quasi *Geraccus), from gero, ‘quod mater
ejus duodecim mensibus utero eum gestaverit’ (ap. Charis. 82. 7 K.), certainly
by spellings on inscriptions like Terebuni for Trébon- (Eph. Epigr. i. 116), though
a good many of these may be dialectal. For this parasitic vowel was
a marked feature of the Oscan language, and its kindred dialects, e.g. Oscan
aragetud (Lat. argento Abl.), Pelignian Alafis (Lat. Albius). Bede cannot be
right in explaining spondaic hexameters, like

illi continuo statuunt ter dena argenti,

as ending really with a dactyl and spondee, the last two words being pro-
nounced ‘denarigenti’; for it is not the case that all, or most, spondaic lines
end in a word in which r is combined with a consonant. But his description
of the sound is interesting, though of course he is not to be regarded as an
authority on Latin pronunciation, except when he is quoting from some older
94 THE LATIN LANGUAGE. (Chap. II.

grammarian. He says (p. 250. 11 K.), after instancing some spondaic
hexameters ending with argenti, incrémenta(!), respergebat, interfectae, intercepto,
neque enim in quinta regione versus heroici spondeum ponere moris
erat, sed ita tamen versus hujus modi scandere voluisse reor, ut addita in
sono vocali, quam non seribebant, dactylus potius quam spondeus existeret,
verbi gratia, ‘intericepto’ ‘incerementa’ ‘ interefectae ’ ‘resperigebat’ et per
synalipham ‘denarigenti.’ quod ideo magis r littera quam ceterae con-
sonantes patitur, quia quae durius naturaliter sonat durior efficitur, cum ab
aliis consonantibus excipitur ; atque ideo sonus ei vocalis apponitur, cujus
temperamento ejus levigetur asperitas ; and he goes on to say that this use
of a parasitic vowel between 7 and a consonant was much affected by monks,
when chanting the responses at divine service. At the same time it is
possible that the existence of this parasitic vowel may explain another feature
of Latin poetry, namely the optional treatment of a short vowel before a mute
followed by r as long by position.‘ Plautus, who, as we have seen, avoided
the parasitic vowel with J, as alien to the conversational Latin of his time,
also refuses to allow the first syllable of a word like pdtri, dgri to be scanned
long, though (see ch. ili. § 42) such a syllable is not so short as the first syllable
of pati, dgi, &e., for it cannot exercise a shortening influence (by the law of
Breves Breviantes) on the following syllable. Plautus scans pdti, digi, but
only pati, dgri. In the same way he always scans vehiclum (the invariable
form), cubiclum (though this last word is an exception to the rule in being
usually quadrisyllabic, cubiculum), and never vehiclum, cubiclum. And this is
the usage in all the dramatic poetry of the Republic. But Ennius in his
Epic, Lucilius in his Satires, allow themselves such scansions as nigrum, litrat,
Sibras, triclini ; and this suggests that in poetry, where the words were sounded
with more deliberation than the rapid conversational utterance of the drama
allowed, the presence of this parasitic vowel was felt to add another unit of
time, another ‘mora’ to these syllables, so that they might on occasion be
treated as long. Lucilius’ scansion ‘triclini (Inc. 145 M.), for example, reminds
us of the form used by Varro, tricilinium (R. R. iii. 13. 2. So the MSS.), and
the forms found on inscriptions trichilinis (C. I. L. ix. 4971 ; xiv. 375, 17, &e.),
should perhaps be replaced by the quadrisyllabic form. In the time of
‘ Servius the accent rested on the second syllable of maniplus, so that the word
was regarded in ordinary conversation, either as being almost a quadri-
syllable, or as having a penult equivalent to a long syllable (Serv. ad Aen.
xi. 463 maniplis: in hoe sermone ut secunda a fine habeat accentum usus
obtinuit) ; and the Vulgar Latin shifting of the accent from the first to the
second syllable of words like tenebrae, &c. (see ch. iii. § 11), is no doubt to be
justified in the same way; though in Servius’ time it was not allowed in
correct pronunciation (Serv. ad Aen. i. 384 peragro: ‘per-’ habet accentum
... muta enim et liquida, quotiens ponuntur, metrum juvant, non accentum).
This explanation of the optional scansion pdtri, fabrum competes with another
(§ 142), according to which the consonant before the r was doubled in pro-
nunciation, as it is in modern Italian fabbro, &c. (beside fabro), just as
a consonant was doubled in later Latin before consonantal u (w) in acqua
(Ital. acqua),and before consonantal i (y), the development of J, in Ital. occhio
(Vulg. Lat. oc(u)lus), &e. Itis quite possible that the shifting of the accent to
the second syllable of words like tenéb’rae, may have had the effect of
strengthening the sound of the mute. The doubling of the consonant in the
§ 103.] PRONUNCIATION, LIQUIDS. 95

proparoxytone syllable is a feature of Italian, e.g. femmina, collera, legittimo
(§ 131).

The parasitic vowel between a mute and J is generally wanting on
Greek inscriptions, perhaps because the Greeks were more conversant with
combinations like yA, 7A, &c. than the Romans, e.g. AevrAos, BryAevria (Lat.
Vigilantia), and the syncopated forms of -ulus (as in porcu-lus, &e.) are usual in the
instances quoted by Eckinger (p. 75), "ApBovoda, Macros, Tarepxdos, Toupxaa,
though it must be added that they mostly date from a time when Syncope

. had taken a strong hold of the Latin language itself. The Appendix Probi
condemns several of these syncopated words (as he condemns calda, &e.),
including with them some whose vowel in the classical spelling is not
original but parasitic: speclum, masclus, veclus, viclus (for vitulus), vernaclus,
articlus, baclus, juglus, oclus, tabla, stablum, tribla, vaplo, capiclum. He also
mentions mascel, figel, which may be South Italian, for the Oscan equivalent
of famulus was famel (Paul. Fest. 62. 1 Th.). The early date of the parasitic
vowel with / is seen in the old Latin form piacolom, quoted by Mar. Vict.
p. 12 K., and pocolom, the usual spelling on the Praenestine vases (C. I. L.
i. 43 sqq.). Plautus seems to regard the use of these lengthened forms as
a licence, only to be resorted to in cases of metrical necessity ; for they are
found, especially when a long vowel precedes the syllable with J, only at the
end of a line or hemistich, e.g. Capt. 740,

periclum vitae meaé tuo stat periculo (see ch. iii. § 13).

(For the parasitic vowel in Greek loanwords with m, n preceded by a con-
sonant, see § 154.) Calicare, from calx, lime (Paul. Fest. 33. 8 Th. calicata
aedificia, calce polita; ib. 41. 21 calicatis, calce politis; ib. 53. 16 decalica-
tum, calce litum ; Gl. Cyrill. oviw, decalico, calce albo; C.I.L. i. 1166 basili-
cam calecandam), apparently the normal spelling, and to be read probably
in the Placidus Glossary (60. 19 G.), (where the MSS. have decaleatis, de calce
albatis), is not a case of parasitic vowel between / and a following consonant
(like Gk. Kaadrropyios), (Dittenb. Syll. 240 of 138 B.c.), and (on the edict of
Diocletian) xadunos. It merely retains the Greek loanword x4aA¢ in its
dissyllabic form, instead of syncopating it to its usual form in Latin, calz.
(For examples of the parasitic vowel in misspellings on inscriptions, see
Seelmann, p. 251.)

§ 103. Avoidance of two r’s. (See 4. L. L. iv. 1 sqq.) The changes of r to
lin Vulg. Lat. pulégrinus, &c., and the doubtful cases of n for r, e. g. low Latin
menetrix, may be considered as examples of this avoidance. The use of gndritior
for gnarior (not before Augustine), férocior, for *ferior, mdgis verus beside verior,
magis miris médis (Plaut. Mil. 539) for mirioribus modis, décentior for décdrior in
Quintilian and Tacitus, sanctior for sdécrior, all exhibit the same tendency.
Pompeius (283. 13 K.) mentions mamor as a mispronunciation of marmor ;
and the form Mamers, Mamertini beside Marmar of the Carmen Saliare, also
Fabaris, the Latin name of the Sabine river Farfarus, perhaps show the same
dropping of r in the reduplication-syllable. So too the form porrigo was
preferred to *prorigo. On praestigiue, from praestringo, to dazzle, crébresco and its
compounds (riibesco, is like pittesco, from a stem without r), see Georges, Lex.
Wortf. s.vv. On inscriptions, de propio (Ital. propio and proprio) for de priprio
occurs (Not. Scav. 1890, p. 170), propietas (C. I. L, ix. 2827 of 19 a.v.), &e. (see
Schuchardt, i. p. 21, for other instances). Vulg. Lat. *trono, to thunder, for tino
96 THE LATIN LANGUAGE. (Chap. II.

(Ital. tronare and tonare, Span. O. Port. Prov. tronar), has inserted 2 for
onomatopoeic effect ; friustrum (Probus 199. 3 K. frustum non ‘frustrum’),
found, with crustrum and pristris, in MSS. of Virgil (see Ribbeck’s Index) and
aplustrum (aplustre) for Greek dpAacroy : crétarias (Caper. 108. 13 K. cetariae
tabernae, quae nune ‘cretariae’ non recte dicuntur) by false analogy (see
Schuchardt, i. 21 for other examples: and cf. Fr. trésor from Lat. thésaurus,
beside Ital. tesoro ; Span. estrella from Lat. stélla beside Ital. stella and (dial.)
strella. In Italian we find the same tendency; e.g. Federico, Certosa (Fr.
Chartreuse), arato (Lat. dratrum), frate (Lat. fratrem), deretano (Vulg. Lat.
*deretranus from retro), &c.; gomitolo, a ball of thread, from Lat. glimus, shows
the same suppression of one of two /’s.

§ 104. rs. Velius Longus 79. 4 says: sic et dossum per duo s quam per r
dorsum quidam ut lenius enuntiaverunt, ac tota littera r sublata est in eo
quod est rusum et retrosum. Cf. Probi App. 198. 29 persica non ‘ pessica’ (a
peach). [For other examples, see Georges, Lex. Wortf. s. vv. controversia (and
other compounds of -versus, e. g. prosa), Marspiter, Sassina, Thyrsagetes, assa, &e.]
In the Sententia Minuciorum of 117 B.c. (C. I. LZ. i. 199) we have controvorsieis,
controvosias, suso vorsum, sursuorsum, and sursumuorsum, deorsum and dorsum side
by side. Since double consonants are usually written single on this inscription
( posidebunt, posedeit, &e., beside possiderent) controuosias probably represents the
pronunciation controuossias. (On these spellings in Plautine MSS., see Ritschl,
Prolegg. p. civ.)

§ 105. r-n. Menetris, a byform of mérétrix (see A. L. L. iii. 539 and cf. Probi
App. 198. 28 K. meretrix non ‘menetris’) may have been influenced by
mdnére or Greek pévw (cf. Non. 423. 11 M. menetrices a manendo dictae sunt).
Nor is cancer a clear case of ‘ Dissimilation’ for *carcer (Greek xapx-ivos), seeing
that a nasal in the reduplication-syllable is not unknown in other I.-Eur.
languages, e. g. Greek -yoy-ydAAw, Tov-Bopi(w, O. Ind. cafi-curyate, cf. gin-grire.
In the Gaelic language n when following c, gy becomes 7, such a word as enu,
a nut, being pronounced cru (with nasal uv). Some see this change in gréma.
a land-measuring instrument, which they consider to be the Greek yvwpwr.
But other instances are wanting ; and gn- in Latin became n, not gr, e.g. nosco,
natus. (Cf. Probi App. 197. 32 pancarpus non ‘ parcarpus’) (see ch. iv. § 80).

§ 106. 1-n. Nuscitiosus, ‘qui plus videret vesperi quam meridie’ (Fest. 180.
21 Th.), and nusciosus, ‘qui plus vespere videt’ (Liwe, Prodromus, p. 17), are
byforms of luscitiosus and lusciosus, which may be due to the analogy of nox.
Leptis, ‘filia fratris’ (ib. p. 340) seems to be a byform of neptis. The Diminutive
of colus, a spindle, was in Vulg. Lat. *conuc(u)la (Ital. conocchia, Fr. que-
nouille). A more certain example of for J is the mispronunciation censured
in Probi App. 197. 24 K. cultellum non ‘cuntellum’ (see below) (see also
Seelmann, p. 327; Schuchardt, Vok. i. p. 143). When Latin n follows n in
successive syllables we find I-n in Ital. veleno and veneno (0. Fr. velin) from
Latin vénénum, Bologna from Bénénia, Palestrina from Praeneste, Praenestinus,
calonaco and canonico (cf. Ital. gonfalone, a banner, Fr. gonfalon, Span.
confalon, O. Fr. gonfanon, Prov. gonfanons from O. H. Germ. grand-fano).

§ 107. 1 before consonant. For the u-affection of 1 before a consonant, see
some instances collected by Schuchardt, Vok. ii. p. 493 sqq., e.g. cauculus for
calculus in MSS. (ef. Georges). In the Edict of Diocletian (301 4.D.) we have
§§ 104—-111.] PRONUNCIATION. LIQUIDS. 97

xavroviarop: for calcilatiri. The letters Land I are so similar that spellings like
sairem (le Blant. J. G. i.) may be nothing but a graver’s error. On the other
hand the Umbrian form of the Latin Volsignus was certainly Voisienus; and in
C. I. L. xi. 5389 and 5390 (=i. 1412) we have epitaphs of a father in Umbrian,
and a son in Latin, with the father’s name Voisieno- and the son’s Volsieno-.
In Probi Appendix 197. 24 K. the mispronunciation cuntellum for cultellum is
mentioned. This treatment of i before a consonant is found in Central Italy
(Latium, Sabina, la Marche, and Umbria) in modern Italian, e.g. untimo for
ultimo in a fourteenth cent. text (see Wien. Stud. xiv. 315 n.). Cf. muntu from
Pompeii (C. I. L. iv. 1593).

§ 108. rl. Velius Longus 65. 11 K. per vero praepositio omnibus integra
praeponitur, nisi cum incidit in 1 litteram, adfinem consonantem, quam
elegantioris sermonis viri geminare malunt quam r litteram exprimere, ut
cum ‘pellabor’ malunt dicere quam perlabor. nec aliter apud Lucilium legitur

in praeposito per,
‘pelliciendo,’ hoe est inducendo, geminato 1 (Lucil. ix. 32 M.};

‘pellicere’ malunt quam perlicere, unde et apud Virgilium non aliter legimus
‘pellacis Ulixi’ (see ch. iv. § 160). Inthe Probi Appendix 198. 14 K. we have :
supellex non ‘superlex,’ with the (marginal?) note utrumque dicitur. (On
the late spelling swperlex, see Georges, Lex. Wortf. s.v.) The only example of
rl in the Indices of the Corpus is perlegere (vol. i).

§ 109. r before consonants. R is sometimes dropped before a consonant on
Greek inscerr., e. g. Koara (Lat. Quarta) (C. I. G. add. 43151), Sarwvdos (Brit. Mus.
ii, 341, from Cos); and in Latin plebeian ‘inserr., we have misspellings like
Fotunate (C. I. L. vi. 2236) for Fortindtae (sometimes Fort-, i.e. Fort- with close
o, see § 145). ([Cf. the rude Faliscan inser. Zvetaieff, Inscr. Ital. Inf. 63 with
Maci Acacelini (as Votilia for Voltilia) for the usual Marci Acarcelini of n° 62, &c.}
But it would be unsafe to rely on these as evidence that Latin r ever became
the mere voice-glide which English r always becomes when not followed by
a vowel, e.g. in ‘here,’ ‘hark’ as opposed to ‘herein,’ ‘ harass.’

§ 110. final r. All final consonants were, as we have seen, weakly pro-
nounced in Latin. Some instances of the omission of -r in spellings of
inscriptions and MSS, have been collected by Schuchardt, Vok. ii. p. 390.

111, Metathesis. Quint. i. 5. 13 ‘Trasumennum’ pro Tarsumenno, multi
auctores; i. 5. 12 duos in uno nomine faciebat barbarismos Tinga Placen-
tinus, si reprehendenti Hortensio credimus, ‘preculam’ pro pergula dicens.
Examples from Plautus are Phyrgio, Aul. 508, corcotarii, Aul. 521. Consentius
(392. 23 K.) censures perlum for prelum, reilquum for reliquum, interpertor for
interpretor, coacla for cloaca, displicina (a schoolboy’s joke surely) for disciplina :
Diomedes (452. 30 K.), leriquiae for reliquiae, lerigio for religio (and tanpister for
tantisper) ; Julian, in Don. v. p. 324. 18 K. intrepella for interpella ; Probi Appendix
has (199. 12 K.) glatri (leg. clatri?] non ‘cracli.’ This late Latin cracli (ef.
Probi App. 195. 23 K.) comes from *eratli as veclus (ib. 197. 20) from *vetlus for
vetulus. Clustrum for crustlum is found on inscriptions of the Empire (e. g. clustrum
et mulsum, Not, Scav. 1877, p. 246 of second cent. a.D., ef. «AoverporAaxois, Athen.
xiv. p. 647 ¢, d). Colurnus is the adjective from corulus. (On the confusion of
fragro, flagro and the form fraglo, see A.L. L. iv. 8.) In Italian nothing is

H
98 THE LATIN LANGUAGE. [Chap. II.

commoner than this Metathesis with r. Thus in S. Italy crapa is used for
capra (Lat. cdépra), which reminds us of the statement of Paul. Fest. that the
old Latin word for capra was crepa (33. 36 Th. caprae dictae, quod omne
virgultum carpant, sive a ecrepitu crurum. Unde et ‘crepas’ eas prisci
dixerunt. The Luperci, who wore goatskins and ran about striking people
with goatskin thongs, were called crépi, ib. 39. 34 K. crepos, id est lupercos,
dicebant a crepitu pellicularum, quem faciunt verberantes) ; so interpetre for
interprete (should we read interpetror in Consentius 392. 23 K.?); and preta
for petra is used in various parts of Italy ; formento is Lat. frimentum, farnetico
Lat. phrénéticus, &c., &e. For 1 we have padule, a marsh (Lat. pdlidem) ; falliva
beside favilla, fiaba, a fable, for Lat. fabla, fabiila. Paduan requilia for reliquia
(should we read requilum in Consentius 392. 23 K. ?) is in Venetian leriquia
(ef. leriquias of Diomedes 452. 30 K.); in many parts of Italy, grolia is used
for gloria. [For some instances of Metathesis in late inscriptions and in MSS.
spellings, see Schuchardt, Vok. i. p. 29 on Prancatius for Pancratius, padules for
paludes (cf. Ital. padule), and Seelmann, p. 330 on Procobvera for Porcobera, &c. ;
and for examples in Latin byforms, see Georges, Lex. Wortf. s.vv. pristis,
crocodilus, Trasumenus, trapezita, and Wolfflin, A. L. L. viii. 279 on accerso and
arcesso.} The Latin Proserpina (C.J. L. i. 57 PRosEpnar dat. case) for Greek
Tlepcepévy (Pelignian Perseponas gen. case) may be due to the analogy of
proserpo; but Vulg. Lat. *alenare for dnkélare (Ital. alenare, Fr. haleiner),
*plopus for poplus, popiilus, poplar-tree (Ital. pioppo, Roum. plop, Catal. clop) are
clear cases of metathesis of 1.

§ 112. ly. For misspellings on late inscriptions and in MSS. like fius for filius,
see Schuchardt, Vok. ii. pp. 486 sqq. Some of them may be due to the confusion
of the L and I (see above). Ital. giglio, a lily, seems to come from a form
*lytlyum,

§ 118. ry. Servius ad Aen, ii. 195 approves pejuro for the Verb, but perjurus
for the Adjective : in verbo r non habet : nam pejuro dicimus, corrupta natura
praepositionis : quae res facit errorem, ut aliqui male dicant ‘pejurus’ ut
pejuro.

§114. F, The Latin phoneticians cannot be suspected of any
influence from Greek sources in their account of f, a sound
unknown to the Greek alphabet; so their description may be
taken as a true account of the pronunciation of f at their time,
or possibly even at an earlier, the date, namely, of the treatise on
Latin phonetics from which they seem to have borrowed. That
time was probably the Augustan age. Their words leave no
doubt whatever that / was a labiodental spirant, as it is in
Italian and most languages, formed by the upper teeth pressed
against the lower lip, not a bilabial spirant, formed by the upper
and lower lips pressed against each other :

imum superis dentibus adprimens labellum
spiramine leni (Terentianus Maurus, second cent. A. D.).
§§ 112-114.) PRONUNCIATION. F. 99

The spzramen lene was more a feature of the normal f-sound
when a vowel followed, than in combinations with consonants
like fr, jl, as we learn from Quintilian. who, when discussing the
more musical nature of the Greek language than the Latin, speaks
of Latin /, especially in words like frangit (to a less extent when
followed by a vowel), as rough and harsh compared with the
softer sounds of Greek. This more vehement articulation of
J before a consonant explains the different treatment of the Latin
spirant in Spanish, in words like haba (Lat. faa), humo (Lat.
fiimus), but fraga (Lat. frdgum, *fraga). Quintilian’s account
does not mention the labiodental character of the sound (xii. 10.
29 paene non humana voce vel omnino non voce potius inter
discrimina dentium efflanda est), but is quite consistent with it.
It is, in fact, very like the account given by phoneticians of our
J, as ‘formed with a strong hiss, by pressing the lower lip firmly
against the upper teeth, and thus driving the breath between
the teeth’ (Sweet, Handb. p. 41).

But it is highly probable that Latin f was at some time
bilabial, as it is to this day in Spanish, where v (4) is bilabial
too. Bilabial f naturally tends to become labiodental, because
by bringing the teeth into play it is possible to give a stronger
and more distinct sound than can be produced by the lips alone.
The voiced bilabial spirant v has, as we saw before (§ 48), become
labiodental v in Italian and other Romance languages. And we
have some evidence of f being still bilabial in the last centuries
of the Republic from spellings like im fronte (C. I. L. i. 1104),
not to speak of comfluont beside conflouont on the Sententia
Minuciorum of 117 B.c. (C.J. LZ. i. 199), and possibly from the
fact that ad, in composition with a word beginning with f (or
v, b, &e.) became in Republican Latin ar, e.g. azfuise on the 8. C.
de Bacchanalibus of 186 B.c. (C. I. Z. i. 196). Another passage of
Quintilian tells us of the difficulty felt by Greeks in pronouncing
this thoroughly Roman letter (i. 4. 14 Graeci adspirare f ut ¢
solent). He illustrates it by the story of Cicero’s ridicule of
a Greek witness who could not pronounce the first letter of
Fundanius. By the fifth cent. a.p., however, the Greek aspirate
had become a spirant, differing from Latin f only in being
bilabial.
100 THE LATIN LANGUAGE. (Chap. IT.

§ 115. Descriptions of the sound of f. Quintilian (xii. ro. 29) : nam illa, quae
est sexta nostrarum, paene non humana voce, vel omnino non voce potius,
inter discrimina dentium efflanda est: quae, etiam cum vocalem proximo
accipit, quassa quodammodo, utique quotiens aliquam consonantem frangit,
ut in hoc ipso ‘ frangit,’ multo fit horridior ; Terent. Maur. 332. 227 K. :

imum superis dentibus adprimens labellum,
spiramine leni, velut hirta Graia [i.e. ¢, p-h} vites,
hane ore sonabis, modo quae locata prima est ;

Marius Victorin. 34. 9 K. f litteram imum labium superis imprimentes
dentibus, reflexa ad palati fastigium lingua, leni spiramine proferemus; Mart.
Cap. iii. 261 F dentes [faciunt] labrum inferius deprimentes. In the sixth
cent. a.D., Priscian mentions as the only difference between Latin fand Greek
¢, that the former was not pronounced /ixis labris (i. p. 11.27 H. hoe tamen
scire debemus. quod non fixis labris est pronuntianda f, quomode p eth ;
atque hoe solum interest) [Blass, Griech. Aussprache®, p. 85 dates the change
of Greek ¢ (written in Latin ph, or as Priscian puts it ‘p et h’), from the
aspirate to the spirant sound at about 400 A.p.]. Two centuries earlier than
Priscian, the difference between Latin f and Greek ¢ seems to have been very
slight, for Diomedes (fourth cent. a.p.), from whom, or from whose original
authority, Priscian may be quoting, says (423. 28 K.): et hoe scire debemus
quod f littera tum scribitur, cum Latina dictio scribitur, ut ‘felix.’ nam si
peregrina fuerit, p et h scribimus, ut ‘Phoebus,’ ‘Phaeton.’ Fis the normal
equivalent of Greek @ in Greek loanwords from the middle or end of the
fourth century onwards, e. g. strofa, Greek orpoph (see ch. i. § 11).

The remark of Priscian (i. p. 35. 17 H.) that F, the Aeolic digamma, used
to have the sound of consonantal v (w), probably refers to a fashion of some
early grammarians of writing ‘ Fotum,’ ‘ Firgo,’ &c., alluded to by Cornutus
ap. Cassiodor. 148. 8 K. and by Donatus ad Ter. Andr. i. 2. 2 (see ch. i. § 7).

§ 116. mf. Mar. Victorinus (18. 14 K.) : item consonantes inter se [invicem
sibi succedunt], sed proprie sunt cognatae, quae simili figuratione oris dicuntur,
ut est b, f, (s!, m, p, quibus Cicero adicit v, non eam quae accipitur pro vocali,
sed eam quae consonantis obtinet vicem, et anteposita vocali fit ut aliae
quoque consonantes. quotiens igitur praepositionem sequetur vox cujus
prima syllaba incipit a supradictis litteris, id est b, f, [s], m, p, Vv, quae vox
conjuncta praepositioni significationem ejus confundat, vos quoque prae-
positionis litteram mutate, ut est ‘combibit’ ‘comburit’ ‘comfert’ ‘com-
fundit’ ‘commemorat’ ‘comminuit’ ‘comparat’ ‘compellit ’ ‘comvalescit’
‘eomvocat’ non ‘conbibit’ ‘ conburit’ et similia. sic etiam praepositio
juncta vocibus quae incipiunt a supradictis litteris n commutat in m, ut
‘imbibit’ ‘imbuit’ ‘imfert’ ‘imficit’ ‘immemor’ ‘immitis’ ‘impius’
‘impotens.’ He must be quoting from some grammarian of the Republic in
his rule about /; for the usual teaching of-the grammarians of the Empire is
that the consonants before which m is used are b, p,m; and Priscian, i. p. 31.
2 H. quotes as early an authority as the elder Pliny to this effect (cf. Prise.
i.p. 29. 18 H. ‘am’ praepositio f vel ¢ vel q sequentibus in n mutat m:
‘anfractus’ ‘ancisus’ ‘anquiro’). But the spelling with m before f (and v)
was undoubtedly an old usage, of which such MS. spellings as comferre, Poen.
1048, comfragosas, Men. 591 in the Plautus Palimpsest, im flammam, Aen. xii. 214,
comfieri, Aen. iv. 116 in Virgil MSS. may be relics. On the other hand since
§§ 116-117. ] PRONUNCIATION. 5, X, Z. IoL

inpérdtor is found both in early and later times beside imperator, and since the
nasal may have been in pronunciation dropped before f (ef. cofisse, covenere in
Virgil MSS., Ribbeck, Ind. p. 393), the evidence of these early spellings is not
conclusive.

$117. S, X,Z In the noun ‘use’ and the verb ‘ to use’ the
letter s has two different sounds, which we often call ‘hard s’
and ‘soft s.’ Hard s is more scientifically termed ‘ unvoiced,’
soft ‘voiced’ s, the two sounds differing exactly as the unvoiced
and voiced mutes, p and 0, ¢ and d, c and gy. The Latin s in
a word like urls was hard or unvoiced s, we know from the
frequently repeated statements of the grammarians, that the
spelling wrys expressed the pronunciation; while the spelling
with 6 was justified only by the analogy of other cases, urdis,
urbi, urbem, &e. (see § 80); and p in sumpsi, hiemps tells the same
tale. Initial s, whether in the accented or unaccented syllable,
may also be put down as unvoiced s, since the Romance languages
agree in giving it this sound (e.g. [tal. si, Fr. si, Span. si for Lat.
si; Ital. sudare, Fr. suer, Span. sudare for Lat. sudare), and
similarly when s is the initial of the second member of a com-
pound, in words like Ital. risalire, Fr. résilier, Span. resalir from
Lat. résilire, later resalire. None of the Latin grammarians
ever suggest that Latin s had anything but one and the same
sound ; and their silence is evidence of some weight that the soft or
unvoiced variety of s was unknown in Latin. This voiced s-sound
seems to have been the sound of Greek ¢ in and after the Macedo-
nian period (cf. (udpaydos for cudpaydos, Zuvpva for Spipva),
which explains why Oscan voiced s, which corresponds to
Umbrian and Latin 7, is in those inscriptions which are written
in Latin characters, expressed by z, e.g. eizac (Umbr. erak),
eizazune egmazum (in Lat. earum rerum) on the Bantia Tablet
ce. 130 B.c. Between vowels s had at an early time in Latin
(c. 350 B.c. according to Cicero, Fam, ix. 21. 2) become 7, as it
did in Umbrian ; and this earlier s may very well have had in
this position the voiced sound (ch. iv. § 146). Intervocalic s in the
classical and Imperial period is only found as a rule where there
had been formerly some consonant combined with s, e.g. formosus
earlier formonsus, causa earlier caussa, dsus earlier ussus (see ch. iv.
§ 148). In such words s has become voiced in many Romance
102 THE LATIN LANGUAGE. (Chap. IT.

languages, e.g. Fr. épouse (Lat. sponsa), but only in those in
which every Latin unvoiced consonant becomes voiced in this
position. In Italian intervocalic s is unvoiced, except in the
few cases where Latin unvoiced mutes also become voiced, e.g.
sposare, like mudare (§ 73). There is hardly any evidence, there-
fore, that s in classical Latin was in any circumstances pronounced
like our s in ‘to use’; and the opinion, a widely spread one,
that the change of spelling from caussa to causa, &c., indicated
a change from hard to soft s is utterly wrong.

The pronunciation of double -ss- is discussed in § 131. Here
we need only mention the curious practice that grew up in the
later Empire of prefixing / to initial s¢, sy, sc, seen in spellings on
inscriptions like istatuam (Orelli 1120, of 375 A.D), spose (i.e.
sponsae C.I.L, viii. 3485),-and in Romance forms like Fr.
épouse. These last show that this spelling does not indicate an
sh-sound of s before a mute like German stehen (pronounced
‘shtehen’), but that there was an actual i-sound before the
s-sound, an i-sound which developed from a vowel-glide, due to
beginning the word before the vocal-organs were properly in
position for the initial consonant. These ‘initial on-glides,’ as
phoneticians would term them, are a feature of Romance lan-
guages (cf. Greek épvdpds), but not of Teutonic. We have
already seen that there is some indication of initial / having had
an ‘on-glide’ in Latin (§ 99); and spellings on inscriptions
suggest the same for other initial consonants. Can this have
been the ‘ circa s litteram deliciae’ which elocution teachers had to
correct in their pupils (Quint. i, 11. 6), just as singers are
taught nowadays to avoid the ‘breathy’: gradual beginning of
an initial vowel ?

X had, as the grammarians repeatedly tell us, the sound of
e followed by the sound of s. The ¢ (as the ¢ of ct, § 95),
tended to be dropped after a consonant, whence the spelling
mers in Plautus for mera; and in careless pronunciation # in
any position tended to ss (so ct became ¢¢, § 95), as we see
from forms like cossim for coxim, used in the farces of Pomponius
(ap. Non. 40 M.). There are traces, too, of the substitution of se
for cs, e.g. ascella is the Late-Latin form of awil/a. In Italian
we have ss [as in ancient Pelignian, e.g. wsur (Lat. wxdres), and
+

§ 119.) PRONUNCIATION. S X, Z. 103

other dialects], for example, sasso (Lat. sawum), but before
a consonant (as probably in Vulgar Latin), s, e.g. destro (Lat.
dexter) (as in ancient Umbrian destra, &c.). Latin loanwords in
“Welsh indicate cs, e.g, O.W. Saes for Latin Saxo, croes for Latin
créx, but s before a consonant, e.g. estron (Lat. extrdneus), estynn
(Lat. eatendo).

Z of Old Latin had perhaps the soft or voiced sound of s, which
passed into the 7-sound about the time of Appius Claudius, the
famous censor, when z was discarded from the alphabet (see ch. i.
§ 5). Greek ¢ differed from it in causing length by ‘ position.’

Final -s after a short vowel was weakly pronounced at all
periods of the Latin language, and in the early poetry often did
not constitute ‘ position’ before an initial consonant, though by
Cicero’s time it was regarded as an essential of correct pro-
nunciation to give s at the end of a word its full sound.

§ 118. Phonetic descriptions of s, x : Tor. Maur. vi. 332. 239-243 K. :

mox duae supremae
vicina quidem sibila dentibus repressis
miscere videntur: tamen ictus ut priori
et promptus in ore est, agiturque pone dentes,
sic levis et unum ciet auribus susurrum.

Mar. Vict. vi. 34. 16 K. dehinc duae supremae, s et x, jure jungentur. nam
vicino inter se sonore attracto sibilant rictu, ita tamen, si prioris ictus pone
dentes excitatus ad medium lenis agitetur. Mart. Cap. iii. 261 S sibilum facit
dentibus verberatis. ... X quicquid C atque S formavit exsibilat. Cledonius,
v, 28.1 K. s... sibilus magis est quam consonans.

§ 119. Latin sin Romance. Initial s- becomes our sh (cf. Ital. scimmia) in
Venice and some other parts (similarly intervocalic s becomes the voiced
form of this sound, as in our ‘ pleasure’), and was possibly voiced s in ancient
Italic dialects (e.g, Faliscan Zexto- for Sextus (?)). On a late inscription of
Tibur we have zaBrna (C. I. L. vi. 12236). Intervocalic -s- between the accented
and unaccented vowels is unvoiced in Spanish (where however all sibilants
are unvoiced), Roumanian and Italian, e.g. Italian mese (Latin ménsis)
(muzex on a Naples’ inscription (C. I. L. x. 719), if it represent actual pro-
nunciation, must have been a dialectal variety]. Italian sposa (with voiced
s and open 0) is influenced by sposare (Latin spo(n)sare), where the o and the
s precede the accent ; the voiced s of rosa is anomalous, but may represent
Greek ¢; or rosa may be a bookword and nota direct descendant of the Latin,

‘for its French and Spanish forms too are irregular (Gréber’s Grundr. p. 522).
It is voiced in the other Romance languages, in which also (as in Spanish)
unvoiced mutes between vowels become voiced, e. g. Fr. épouse (Lat. spo(n)sa),
chose (Lat. causa) with voiced s, like O.Fr. ruede (Lat. rita, Span. ruede), vide
(Lat. vita, Span. vida). It is voiced also in North Italian. Intervocalic -s-
104 THE LATIN LANGUAGE. (Chap. II.

before the accented vowel is voiced in Italian, e. g. sposare (Latin spo(n)sare),
precisely as any Latin unvoiced mute becomes voiced in this position ; e. g.
mudare (Latin miatare), pagare (Lat. pdcare). So does any s which by Syncope,
&c. has come to stand before a voiced consonant, e.g. sdegno (Vulg. Lat.
disdigno), while in Spanish it has developed to @, » (though written s), e. g-
desden. ;

§ 120. Greek £, Latin z. The letter z, the Greek letter ¢, was, as we saw
(ch. i. § 1), brought into use at Rome in the transcription of Greek words (and
of those only) about the close of the Republic. Previously to that time ss had
been used, e. g. massa (Greek waa), which at the beginning of a word was s,
e.g. Setus (C.I.L. i. 1047, 1299, Greek 2900s) (Plautus makes this s- alliterate
with ordinary s-, e.g. sonam sustuli Merc. 925, solve sonam Tuc. 954), and,
if we are to believe the grammarians, d, e.g. Medentius for Mezentius. [But
Septidonium a mispronunciation of Septizonium, a building at Rome (Probi App.
197. 23), seems to be a popular etymology from donum.] (Prise. i. 4g y etz
in graecis tantummodo ponuntur dictionibus, quamvis in multis veteres
haec quoque mutasse inveniantur, et prov u, pro ¢ vero...s vel ss vel d
posuisse, ut.. ‘Saguntum,’ ‘massa’ pro Zdxvvdos, pa(a,... ‘Sethus’ pro
2700s dicentes, et ‘Medentius’ pro Mezentius.) Blass, in his book on Greek
Pronunciation, gives the history of the Greek sound as follows, The com-
bination 2d in words like dos (Germ. Ast), i(w (Lat. sido for *sisdo), “A@nvace
(‘A@qvas-de) was expressed by the letter ¢ (the Semitic letter Sain, a symbol in
the Semitic alphabet of voiced s‘, as the combination ks by ¢ (the Semitic
Samech). This was the original use of ¢. It came to be applied to the
combination dz (from dy), e.g. meds (for medyds), (amAouros (for bidmAouTos) ;
and at this period came the transference of the Greek alphabet to Italy, with
the result that in the Italic alphabets, Umbrian, Oscan, &c., the z-symbol
had the sound of dz or ts. In course of time dz came round to the sound of
zd, so that me(és and i{w had now the same sound of ¢. This 2zd-sound further
developed into the sound of zz, or z, apparently in the Macedonian period ;
and so we find the town Gaza, whose Semitic name has voiced s or Sain,
written in Greek characters Ta¢a. This then was the sound which the
Romans had to express in Greek loanwords, voiced s, not the earlier sound
zd. Voiced s, as we have found reason to believe, was a sound unknown in
Latin words since 350 B.c., which explains Quintilian’s remark (xii. 10. 28 ;
ef, Maxim. Victorinus, vi. 196. 3 K.) about the beauty of the sound of ¢, and
its absence from the Latin alphabet. To express it, double or single s (the
unvoiced s-symbu]) was used by the early Republican writers and occasionally
by later authors (e.g. saplutus, Petron. 37, for (awAovros), perhaps even d (with
the sound of th in ‘this’ ?), until a later age felt the necessity of employing,
for the sake of exactness, the Greek letter itself, as they did also in the case
of Greek v, ¢, x, 9. The history of (is a common point of discussion among
Greek grammarians who remark on its origin from the combinations o6 and
50, and their remarks are repeated by their Latin imitators, but need not be
taken to imply that ¢ had at the time of the Empire any other sound than
that of voiced s (Mar. Vict. vi. 6.6 K.: Maxim. Vict. vi. 196. 3 K. : Audacis’
exc. vii. 327 K.). Thus Velius Longus (vii. 50.9 K.), in criticizing the remark
of Verrius Flaccus : ‘sciant z litteram per sd scribi ab iis qui putant illam
ex s et d constare,’ states positively that ¢ had not the sound of a double
letter, unlike y and ¢: denique siquis secundum naturam vult excutere hanc
§§ 120-128.) PRONUNCIATION. |S, X, Z. 105

litteram, inveniet duplicem non esse, si modo illam aure sinceriore explo-
raverit... et plane siquid supervenerit, me dicente sonum hujus litterae,
invenies eundem tenorem, a quo coeperit. The interchange of dy and z on
late inscriptions, e. g. baptidiata, Rossi i. 805, of 459 a.D., and in spellings like
zabulus for didbélus (see Georges, Lex. Wortf.s.v.) implies merely that the spirant
y-sound which dy had come to take in Latin (§ 51, ef. Madia for Maia) was
felt to resemble the sibilant sound of voiced s. Z has however the ts-sound
in alphabets derived from Latin, e. g. O. Engl. Bezabe ‘ Bathsheba.’

§ 121. Old Roman z, found in the Carmen Saliare (Velius Longus, vii. 51. 5
K.), and according to tradition discarded through the influence of Appius
Claudius, one would naturally suppose to have had the same sound as that of z
in the Umbrian, Oscan, and other Italic alphabets, viz. dz or ts'; so that Claudius
might, like Papirius, Ruga, and other traditional reformers of spelling, have
exemplified the new orthography in his own family-name by writing Claudius
for an earlier *Clauzus. This would harmonize well with the fantastic
remark of Martianus Capella, that Claudius objected to the letter because it
gave the teeth the appearance of a death’s-head (iii. 261 Z vero idcirco
Appius Claudius detestatur, quod dentes mortui, dum exprimitur, imitatur),
which happily describes the appearance of the mouth in uttering the dz-
sound of our ‘adze.’ But this description will also suit for the voiced
s-sound ; and the contemporary change of intervocalic s (voiced s} to 7,
exemplified in the new spelling of Papisius as Pdpirius, as well as the use of
z for voiced s on a very early coin of Cosa, suggests that this rather was
the sound of early Roman z (see ch. i. § 5’, though the matter is uncertain.

§ 122. Old Roman s (2), later r. The change of intervocalic s to r is
a common one in various languages, and is generally taken to imply that the
s first became voiced s, then passed into r (cf. Span. desden, see above). That
Old Roman s of Fisius, &e. took this course is indicated by the fact that the
Oscan sibilant, corresponding to Latin and Umbrian 1, is in the inscriptions
in Latin characters written z, and not s.

§ 123. Prosthetic vowel with st, &c. For instances of the prosthetic
vowel in MSS. and late inscriptions, see the list given by Schuchardt in Vok.
ii. pp. 338 sqq., who refers the earliest traces of its use to the second cent. a. D.
It is written 7 or e, e.g. istudium, estudiwm, sometimes hi-, he-, e.g. histudiis
(often misread in MSS. as hisstudiis), rarely y, ae, and is often confused with
the prepositions in, ex. Thus iscribere, escribere may represent either scribére or
inscribere or exscribere. In late Latin where ab is used before an initial vowel,
@ before an initial consonant, we find ab normally before an initial sp-, sc-, st-,
even though the prosthetic vowel is not expressed in writing (A. TL. L, iii. 149).
Along with the dropping of this prosthetic vowel, and the restoration of the
original form scribere, &c., went the dropping of the initial ¢-, e- of i(n)scribere,
e(x)scribere, so that we get spellings like splorator for eaplordtor (Cagnat, Ann.
Epigr. 1889, no. 55), Spania for Hispania (see Schuchardt’s list, Vok. ii. pp. 365
sqq,). In Italian, a language in which almost every word ends in a vowel,
the prosthetic vowel has been dropped, e.g. studio, and with it the genuine

 

} Initial Z, however, in Umbro-Ose. may have been a variety of s, e.g. Umbr.
ze¥ef ‘sedens’ (von Planta, i. p. 71).
106 THE LATIN LANGUAGE. (Chap. II.

initial vowels of words like Lat. historia, Hispania, instriimentum, Ital. storia,
Spagna, stromento or strumento. But after words like con, in, non (all ending
in a consonant), both vowels are restored in pronunciation, so that the
spelling con estudio, non estoria represents the actual sound. These forms
studio and estudio (istudio), storia and estoria (istoria) are what are called
‘ doublets,’ the one being used after a final vowel, the other after a (rare) final
consonant; and that is, no doubt, the explanation of these double forms
splorator, esplorator, &c. on late inscriptions. In French, where consonant
endings were far more preserved than in Italian, the prosthetic vowel
remains, e.g. étude, écrire. It is before st-, sc-, sp-, &c., for the most part, that
the prosthetic vowel asserted itself sufficiently to require expression in spell-
ing ; but its presence before other consonantal initials may be inferred from
occasional spellings like ilocus, ireddere, imerito (misread in MSS. as immerito), &e.
(see the list of examples in Schuchardt, Yok. ii. pp. 360 sqq. ; some in MSS. are
merely misreadings of critical signs). In Italian, where, as we have said,
almost every word ends in a yowel, we could hardly expect this Latin pros-
thetic vowel to show signs of itself. Indeed the tendency is rather for a genuine
initial vowel to be suppressed under the influence of a preceding final vowel.
Thus la apecchia, the bee (Lat. dpiciila) has become la pecchia, Lat. ecclésia has
become chiesa, inimicus, nemico, and so on. Whether the same Procope is the
explanation of the mispronunciation rabo for arrdbo, which Plautus puts into
the mouth of the slave in the Truculentus, for the sake of poking fun at the
Praenestines, it is impossible to say :

STR. tene tibi
rabonem habeto .
AST. Perii, ‘rabonem.’ quam esse dicam hance beluam ?
Quin tu arrabonem dicis?
STR. ‘a’ facio lucri,
Ut Praenestinis ‘conea’ est ciconia.

We do not find mention of the prosthetic vowel by Latin grammarians till
quite late times, which shows that however far it had developed in Vulgar
Latin, it did not threaten to encroach on the speech of the educated classes.
Thus Isidore (seventh cent.) derives escarus (i.e. scarus) from esca (Orig. xii. 6. 30
escarus dictus eo, quod escam solus ruminare perhibetur), and iscurra (i.e.
scurra), somewhat comically, from the same word (ib. x. 152 [under I not E]
iscurra vocatur, quia causa escae quempiam consectetur; cf. ib. xx. 4. 9
discus antea ‘iscus’ vocabatur a specie scuti). He warns his readers against
the mispronunciations yspissa, yscena, ystimulus, (4. 509, App. 3. 40. Ar. spissa,
scena, stimulus et cetera similia y carent). Similar warnings are given in the
Glosses ap. Mai, Cl. Auct. against iscena, iscandalum, iscapha, iscribtura (vi. 580),
and directions to write ‘per solam s’ sceda (vii. 578 b), stimulus, spissa, and
splendor (vi. 581). (Theophilus non ‘izofilus,’ Probi App. 198. 1, should
perhaps read ‘T. non ziofilus,’ and in 199. 10, stabilitus non ‘istabilitus,’ is
a mere conjecture.) It is not found in the early Latin loanwords in Teutonic,
e.g. O. H. Germ. scriban (Lat. scribo), or Celtic languages, e.g. O.Ir. scol,
Bret. skol (Lat. sc(h)ola). But Welsh, which has the same tendency as late
Latin to use a prosthetic vowel (written y, pronounced like u of our ‘ but’),
before initial s followed by a consonant has subsequently added this y- to these
Latin loanwards, e.g. ysgol, ysgrifo, as it has done to other words of a similar
§§ 124, 125.] PRONUNCIATION. 5, X, Z. 107

form, e.g. ysgub, a sheaf. That Procope had shown itself in Vulgar Latin
we see from the Romance forms, Ital. bottega, Span. botica, Fr. boutique,
which point to Vulg. Lat. *poteca for dépéthéca, and Ital. morchia, Span. morga
from Vulg. Lat. *murca for dmurca, both Greek words. (On ste for istd, see
ch. vii. § 17.)

A further result of the confusion of a word like scrivo with a compound,
exsoribo, inscribo, was that some words beginning with sc-, sp-, st-, &e. were
regarded as compounds with the prepositions ex, im, and were deprived of
their initial s. This, at any rate, seems to be the explanation of forms like
Vulg. Lat. *pasmus for spasmus (Span. pasmo, Port. pasmo, and the French verb
pamer, to swoon) (see the list of examples in Schuchardt, Vok. ii. pp. 354 sqq.).
Another result possibly was that such a form as sponere for exponere being
regarded as the equivalent of ponere, the letter s- might occasionally be
prefixed at random to words beginning in c-, t-, p-, &c., e.g. spictus for pictus
(Schuchardt, 1.c., mentions a few doubtful examples; but includes cases
where the s- was original, e.g. O. Lat. stritavus, later tritavus. See ch. iv. § 146.)

§ 124. s before a consonant. I.-Eur. s before m, n, &c. was dropped in Latin
or rather assimilated (iv. 159), e. g. primus (Pelignian Prismu for Lat. Prima),
cémis (on the very ancient Dvenos inscription cosmis), diimus older dummus
(cf. Dusmus). So tra(n)s became tra- before j, d, and optionally before m, p,
according to Velius Longus, 66. 9 K., e. g. transtulit, but trajecit, tradusxit trans-
mistt or trdmisit, transposuit or traposuit, Spellings on late plebeian inscriptions
and in MSS., such as prebeteri for presbyteri (Rossi, i. 731, of 445 4. D.), have
been collected by Schuchardt, Yok. ii. pp. 355sq. But though s is suppressed in
this position in some Romance languages, notably in French, e. g. chateau
(Ital. castello, Span. castillo) from Lat. castellum ; blamer (Ital. biasmare,
Catal. blasmar, Prov. blasmar) from Vulg. Lat. *blas(i)mare for blasphemare,
this suppression is by no means universal, and was unknown in French itself
at an earlier stage, e.g. blasmer; so that these spellings cannot convince us
that Latin s in the middle of a word had at all the same weak sound that it
had at the end of a word. (Schuchardt’s examples of the assimilation of ¢ or
t to s, with ss for sc and st, e.g. Crissana, Vok. i. pp. 145 8q., are perhaps better
explained as cases of palatalization of ¢, ¢.)

A vowel before si, &c. is not shortened under the influence of a preceding
short syllable in Plautine versification any more readily than a vowel before
any other consonant group, e. g. voliiptdtem, beside potéstatem, ministéerium. (See
ch, iii. § 34.)

§125. x. On the spellings cs, cx, xs, &c. see ch. i. § 4, and for the interchange of
«x with ss and (with consonant) s, see Georges and Brambach s. vv. mixtus,
Ulixes, sesoenti, Esquiliae, Xerues, Sestius, &e. Schuchardt, Vok. ii. p. 351, and i. 133,
gives some instances of es- for ex- before c, t, pin late Latin inscriptions and in
MSS. (Cf. Placidus’ Glossary, 67. 18 G. exspes, sine spe. . . ‘espes’ vero sine
x nihil est), and of -ss-, -s- for -x-, e. g. vissit for viait, Alesander for Alexander.
Vissit for vixit is common on late Christian inscriptions (e. g. C. I. L. x. 4546),
but the earliest instance of ss for x is probably on an epitaph of a cavalry
soldier at Cologne, which cannot be later than Nero’s reign [ve]ssillo (A. L. L.
viii. 589.) On mers for merx (mer(c)s, like pars for par(t)s) in MSS. of Plautus,
see Ritschl, Opuse. ii. p. 656. Caper 98. ro K. allows both cals and calx: cals
dicendum, ubi materia est, per s; at cum pedis est, calx per x. In late Latin
108 THE LATIN LANGUAGE. [Chap. II.

final -s and -x are often interchanged. Thus x is written for the final s
(originally ss) of miles, cries, piples, léctiples. All these are forms censured in
the Appendix Probi (197. 28 K. ; 198. 29; 199. 4,5); and on inscriptions we
have milex, milix, pregnax, &e. see Seolmann, p. 353). The similarity of the
sound of this -s with -x may be inferred from Probus, Inst. 126, 36 K. quae-
ritur qua de causa miles per s et non per x litteram scribatur, &c. Similarly
obstetriz was made opstitris (Probi App. 198. 34 K., cf. 198. 28 meretrix non
‘menetris’) ; and on inscriptions we have conrvs and corvs (Greek Kozoyc,
C.I.L. X. 719), subornatris, &c. (see Seelmann, p. 353). The felatris (iv. 1388
and 2292) on inscriptions of Pompeii reminds us of -s, -ss for x, cs, on Oscan
inscriptions of the same town, e. g. meddiss (for *meddicés Nom. Pl.) Zv. LL.I.
140 (cf. Osc. Santia for Bavéias).

§ 126. Final s. (See Havet on ‘I’S latin caduc’ in Etudes dédiées a G. Paris.
1891 ; he shows that it is the rule, and not the exception, that -s does not
constitute ‘ position’ in the older poetry ; cf. Plautine endings of lines like
estis vos), Cicero (Orator, xlviii. 161) : quin etiam quod jam subrusticum videtur,
olim autem politius, eorum verborum, quorum eaedem erant postremae duae
litterae, quae sunt in ‘optimus,’ postremam litteram detrahebant, nisi vocalis
insequebatur ; ita non erat ea offensio in versibus, quam nunc fugiunt poetae
novi; ita enim loquebamur:

qui est omnibu’ princeps,
non ‘omnibus princeps,’ et
uita illa dignu’ locoque,

non dignus. quod si indocta consuetudo tam est artifex suavitatis, quid ab
ipsa tandem arte et doctrina postulari putamus?; Quint. ix. 4. 38 quae fuit
causa et Servio, ut dixit, subtrahendae s litterae, quotiens ultima esset aliaque
consonante susciperetur, quod reprehendit Luranius, Messala defendit. nam
neque Lucilium putat uti eadem ultima, cum dicit ‘Aeserninus fuit’ et
‘dignus locoque,’ et Cicero in Oratore plures antiquorum tradit sic locutos.
(On the dropping of -s on inscrr. see § 137.)

$127. Double Consonants. No point of Latin pronunciation
is more certain than that a double consonant in such a word as
bueca was really pronounced as a double, and not as a single
consonant, with ‘the first syllable ending in one c, and the
second syllable beginning with another c,’ as the Latin gram-
marians put it, or in more scientific language, with a new force-
impulse beginning in the second half of the consonant. The
word would be uttered, not, as we are accustomed to pronounce
it, with one c-sound, but with the double c-sound of our ‘ book-
case.” The statements of the grammarians are so clear on this
matter as to leave no room for doubt; and even without their
help, we might have inferred the Latin usage from the evidence
of the Romance languages. For although it is only the Italian
which has entirely preserved to this day the double pronunciation
§§ 126, 127.] PRONUNCIATION. DOUBLE CONSONANTS. 109

(e.g. Ital. boc-ca, but Span. boca, Fr. bouche), there are traces in
the others of its previous existence. Latin ss is hard s, where
Latin s has become soft, or voiced s. Latin rz, nn, UW have
developed into different sounds in Spanish from Latin 7, x, J;
and in French a Latin vowel before a double consonant has been
differently treated from one before a single consonant: tals
becomes tel, but vallis, val ; manus becomes main, but axnus, an.
The only thing open to question is whether the spelling with two
consonants did not sometimes indicate a lengthened rather than
a doubled consonant, a consonant on which the voice dwelt for
a time, without dividing it between two syllables. This distinc-
tion between a long and a double consonant is more clearly
marked in the case of a mute (e.g. long ¢ and double c), than of
a liquid, nasal, or sibilant (e.g. long 7 and double /, long » and
double x, long s and double s). The greater force and abruptness.
of the mute as compared with the liquid would make the syllable-
division in ducca more readily caught by the ear than in mille.
This lengthened pronunciation may have been given to m, &c.,
in dmitto for ammitto (cf. ammissam in the Medicean MS. of Virgil,
A, ii. 741); and it was probably a stage in the development of
words like mzlia older millia, causa older caussa, casus older cassus.
From the statements of the grammarians, and from the
spelling of Inscriptions and the oldest MSS., we see that the
orthography, and presumably the pronunciation, of the Empire
did not allow ss after a diphthong, nor (with possible exceptions)
after a long vowel, nor yet // between a long ¢ and another 7.
The caussa, cassus, glissa, missi, militia of an earlier time were
reduced to causa, cdsus, glisa, misi, milia, and show in Italian
to-day the single letter in spelling and pronunciation (cosa,
chiosa, misi, &c.). Seelmann’s explanation is that the length of
the diphthong would detract from the length of the consonant
in cau-ssa, and make it no longer than a single consonant causa,
while in mda the similarity of the articulation of 7 and 7 was the
reason why the vowel organs passed so quickly over the inter-
vening /-position back to the 7-position, as to prevent the voice
from dwelling for the due period of time on the / itself. How-
ever that may be, we can at least be positive that the spelling ss
did not, as Corssen suggests, merely indicate the hard or unvoiced
L1O THE LATIN LANGUAGE, (Chap. IT.

quality of the s-sound (e.g. Engl. ‘ass’ with hard s, ‘as’ with
soft s).

The practice of writing the consonants double was not adopted,
as we saw (ch, i, § 8), until the time of Ennius. But thereis no
evidence, apart from this fact, to show that the pronunciation of
bucca, penna, &c., in earlier times was not the same as the later
pronunciation (like our ‘bookcase,’ ‘ penknife’). Plautus may
have written these words with a single letter; still he always
treats the first syllable as long by position ; so that it would be
as rash to infer that the older spelling was anything more than
a mere usage of orthography, as to regard the temporary use of
the sicilicus in the Augustan age (ch. i. § 8), e.g. osa (C.L.L. x.
3743), a8 an indication that the consonant had at that time
a lengthened rather than a doubled pronunciation.

§ 128. Testimony of the grammarians, The grammarians’ rule is ‘ Write
two consonants, when two consonants are pronounced’: ubi duarum con-
sonantum sonus percutiet aures, Mar. Victorinus vi. 9-10 K.; who quotes
scb-batis, suc-cis, ef-fert, ef-fugit, fal-lit, gal-lus, val-lus, macel-lum, nadl-lus, pal-lium
Pal-las, an-num, Cin-nam, ap-paratum, lap-pam, Ar-runtium, bar-rum, cur-rit,
fer-rum, as-siduum, Cas-sium, fes-sum, At-tius, Vet-tius, and adds: nam ut color
oculorum judicio, sapor palati, odor narium dinoscitur, ita sonus aurium
arbitrio subjectus est. Similarly Papirian (ap. Cassiodor. vii. 162. 10 K.) says:
sono internoscemus, quoting ac-cedo, at-tuli, as-sidwus, ap-pareo, an-nuo, al-ligo.
So Vel. Longus vii. 61-62 K. : ac-cipio, ac-currere, ag-gerat, Pliny (ap. Priscian,
i. p. 29. 8): d-le, Metel-lus. They speak also of one syllable ending with the
consonant, and the next syllable beginning with the same consonant (prior
syllaba in hac finitur, et sequens ab ea incipit, Consentius, v. 394. 35 K.,
who quotes il-le, Al-lia). Similarly Priscian, i. p. 45. 5 of il-le, p. 46. 8 of
Sab-burra, sab-bata, gib-bus, gib-berosus, gib-ber, ob-ba, . . . sub-bibo, p. 47. 5 of vacca
(MSS. bacca), buc-ca, soc-cus, ec-quis, quic-quam, p. 47. 9 of abad-dir, abud-dier, ad-do,
red-do, red-duco (‘ quod etiam reduco dicitur’,) p. 48. 5 of of-sicio, suf-jicio, af-fectus,
ef-ficio, dif-ficilis, dif-fundo, p. 49. 29 of lip-pus, ap-paret, p. 50. 25 of mit-to, Cot-ta,
at-tinet. Velius Longus’ remarks on the pronunciation of reduco and reddo must
be understood in the same way (vii. 66. 3 K.).

§ 129. Reduction of 11 to 1, 8s to 8, after a diphthong or long vowel.
Some grammarians ascribe this reduction to a diphthong, others to any
long vowel. Quintilian (i. 7. 20-21) tells us that caussae, cassus, divissiones was
the spelling of Cicero’s time, and that the double s was found in autograph
MSS. both of Cicero and of Virgil, and adds that in still earlier times (i.e.
before the introduction of double letters), jussi was spelt with a single s.
Velius Longus (vii. 79. 20 K.) censures the proposal of Nisus (first cent. a. D.)
to write comese, consuese, and his argument ‘quia juxta productam vocalem
geminata consonans progredi non soleat,’ and declares positively that
‘geminari consonantes productis vocalibus junctas usus ostendit,’ quoting
§§ 128,129.) PRONUNCIATION. DOUBLE CONSONANTS. III

as examples errasse, saltasse, abisse, calcasse. He inclines however to the spelling
paulum on the ground that paullum ‘repetito eodem elemento [sc. 1)... enun-
tiari nullo modo potest,’ and declares the true rule to be that the presence of
a diphthong, not of any long vowel, forbids the doubling of a consonant (cf.
Prise. i. p. 109. 22 H.). Still he contrasts dosswm (for dirswm), with rasum,
retrdsum ; and in another passage (72. 11 K.) he approves of the spelling and
pronunciation accisdtor, as of cdmisdtor, Annaeus Cornutus (ap. Cassiodor.
149. 12-15 K.), speaking apparently of the old spelling caussa, says: in qua
enuntiatione quomodo duarum consonantium sonus exaudiatur, non invenio.
Terentius Scaurus (21-22 K.) declares that neither s nor 7 are doubled, unless
the preceding vowel is short; when it is long, the syllable ends with the
vowel, and the consonant begins the next syllable, e. g. plau-sus, lii-sus. The
spelling caussa he makes etymological (due to cavissa), not phonetic: apparet
‘eausam’ geminatum s non recipere, quoniam neque in fine praecedentis
alterum potest poni, neque a gemino sequens incipere. The remarks of
Velius Longus (72. 19 K. s vero geminata vocis sonum exasperat), and of
Marius Victorinus (viii. 5 6 K. iidem (sc. antiqui] voces quae pressiore sono
eduntur, ‘ausus,’ ‘causa,’ ‘ fusus,’ ‘odiosus,’ per duo s scribebant ‘ aussus’),
must be regarded in the light of the previously quoted statements; though
the latter may imply that the sound of an -s-, which represented a former -ss-
was not quite the same as the sound of ordinary s. That this was probably
the case with final -s (e. g. miles for *miless) we shall see below (§ 133). Else-
where Terentius Scaurus defends the spelling pawlum on etymological grounds,
comparing pullum, pusillum (20.15 K.) ; and Annaeus Cornutus (first cent. a.p.,
ap. Cassiodor. 149. 19 K.) speaks of some grammarians who wrote mailo (the
older spelling, as we shall see) for malo, because they connected the word with
Greek paAAov. Another reason apparently alleged for this spelling mallo, nollo,
was the analogy of the Infinitive maile, nolle, to judge from Papirian’s dictum ap,
Cassiodor. 159. 1 K.: malo per unum 1, quod est magis volo; malle per duo.
1, quod est magis velle ; nolo per unum 1, est enim non volo, nolle per duo ],
quod est non velle (cf. Probi App. 201. 33 K. inter velit et vellit hoc interest
quod, &c.) (cf. vellint, C.I.L. v. 2090; vii. 80; nollis, vii. 140), A further
instance of the influence of an etymological theory on spelling is furnished
by Alcuin (310. 32 K.), who defends the spelling solemnis by referring the
word to stleo. But if we overlook spellings warped by etymological theories,
and here and there a traditional spelling retained, we may lay down the rule
that J after a diphthong, and s after a long vowel or diphthong, were not
written, or pronounced double in the period of the Empire, so that it is
unlikely, for example, that cessi (though from cédo), jussi [though the spelling
jous- occurs on old inscriptions (see ¢. I. L. it. Index p. 583), and cf. jussus
(along with Anniis!), vi. 77], ussi (though from iro; cf. A.L.L. ii. 607), had
a long vowel in the Imperial age. ((C¥ssi, wissi, according to Priscian, i.
p. 466. 6, 7H.) The use of il after a long vowel, but not after a diphthong,
shows that the diphthongs still retained their diphthongal sound.

The statements of the grammarians about the older spelling are borne out
by a reference to the Republican inscriptions. On the Lex Rubria of 49 B.c
we have promeisserit, remeisserit, repromeisserit, and on other inscriptions

, missit, paullum, millia, milliarium (see the Index to C. I. L. vol. i’.
pp. 6or- 2); on the Comm. Lud. Saec. both caussa and causa, but always quaeso ;
on the Mon. Ane. millia, clausum and cl um, . ; io.

 
112 THE LATIN LANGUAGE. [Chap. II.

Paullus is the usual form even on later inscriptions, also Pollio and Polio (see
Georges, Lex. Wortf. s.vv.; ef. Polla with apex on 0, C.I.L. xi. 4572, &e.).
Aulla, the oldest spelling, preserved in the Ambrosian MS. of Plautus (see
below), similarly became aula and olla (see Georges) ; crisso (with 7, ef. crispus)
became criso (ib.); glossa and glossema became glosa, glosema (Liwe, Prodromus,
Pp. 1 sqq.) ; nassiterna is the old spelling of this old word, like nasswm (later
nasus) (see Georges). Thus abscisio (from caedo), and abscissio (from seindo)
were not distinguished in spelling till Tiberius’ reign.

In the best and oldest MSS. of Republican writers, aud (archaistic) writers
of the Augustan age, such as Plautus and Virgil, we havea good many spellings
with U, ss, where the later orthography used the single letter. Thus in
Virgil MSS. we have examples of the old spelling, with double s, of the Perfect
and Perf. Part. Pass. of verbs like edo (adessus, ambessus, exessus, obessus, peressus,
semessus, i.e, adéssus, &c.), video (provissa, &c.; also the verb invisso) Seigieo
(aussa’, and other verbs in -do, -deo; haereo (haessit) and haurio (haussere,
hausserat) : mitto (missi, missere); we have -nss- and -ss- for later -ns-, -s- in
conprenssa and compressa, emenssi, &c., lapidossa, undossi, &c.; similarly caussa,
incusso, &c. ; and in foreign words cassia (Gk. xaccia, a misspelling of. xécia),
Crinisso (A. v. 38), gessa (A. viii. 662), Passiphae, Rhessus ; though some of these
last may be a wrong spelling (e.g. gessa should be gaesa, for O.Ir. gae, O. H.
Germ. gaizon- point to gaiso-, with g for I.-Eur. gh-, cf. Sanscr. hé¥as). And
in the Ambrosian Palimpsest of Plautus we have essum, essurire, essitabunt,
exscissus (Most. 826), ussus, vib incusses, uisso, ucisse, dimissero, quaesso, caussa,
-ossus (laboriossi, odiossae, n , radiossus) ; also , wassd, and even
sesse (perhaps rightly, for *séd-sz(d)), Merc. 249, Stich. 365, and noss, Stich. 536 ;
and in Greek words paussam, Alcéssimarche, and (perversely) bédssilice, Poen.
577 (ef. bassim, C.I.L. i. 1181). For i we have in Virgil MSS. some words
where s, x, &c. have been dropped before / with lengthening of the vowel, or
rather have been assimilated, like quallus, G. ii. 241 (cf. qudsillus), anhellitus and
anhellus, tellum, vellum, along with some of doubtful origin, collwm, a strainer,
G, ii. 242 (probably first *cavillum, then caulum or collum, then cdlum, like
Paullus, Paulus, Poltio and Polio), illex, A. vi. 180, mallim, G@. iii. 69 and A. iv. 108,
paullatim, millia, opillio; as well as -ella, -ellus in loquella, querella (the normal
spelling), Philomelia, fasellus. (On olli Dat., dlim Adv., see ch. vii.): And in the
Plautus Palimpsest, aulla, a.jar (later ada and olla), millia, paullum (see
Ribbeck’s Index, and the Index to Studemund’s Apograph of the Codex
Ambrosianus). In the Palatine MSS. of Plautus we have also noillo (see
Goetz, preface to the Stichus, p. xiv).

Of these, we know that millia, &., where long i precedes and i follows the
double 1, became milia, &c. in the Imperial age. Pompeius (185. 16 K.)
quotes the rule of the elder Pliny : Plinius Secundus in libris dubii sermonis
ita expressit, ‘mille non debemus aliter dicere nisi per geminum 1, in numero
plurali unum | ponere debemus et dicere milia’ (ef. ibid. 172.13 K.). Mitia,
vilicus are the normal spellings on inscriptions, beside mille, villa, from the
reign of Tiberius ; but in earlier inscriptions (excepting in very early ones
where no consonant is written double) we have ll. In the Monumentum
Ancyranum, that valuable evidence of the orthography of the Augustan age,
we have millia, milliens. (On stilicidium from stilla, Diminutive of *stira, stiria,
see Lachmann ad Luer. i. 313.) In other circumstances double il was retained
after 1 long vowel. e.g. villa, stélla, which show the long vowel and double J,

 

 
§ 130.] PRONUNCIATION. DOUBLE CONSONANTS. 113

or its traces, in the Romance languages, e.g. Ital. villa, stella (with close ¢),
as in the Welsh loanword ystwyll, Epiphany. So mille, rallum, stilla, villum
(Dim. of vinwm), trilla, cordlla (Dim. of cordna), alus (from wnus), nillus, &e.
(Priscian i. p. 109. a1 H. attests villum, and dllus; and on inscrr. we have wlla
C.IL, ii. 1473; dilli, vi. 10230; nillum, x. 4787 ; villani, ix. 348, &c.) But in
a group of words, as before remarked, we have Ui in the older, 7 in the Imperial
spelling, viz. words where there has been what is called ‘compensatory
lengthening,’ e.g. qudlus for *quas-lus (cf. quédsillus), vélum for rex-lum (ef.
vewillum). Of these words Cicero says (Orat. xlv. 153), quin etiam verba saepe
contrahuntur non usus causa, sed aurium ; quo modo enim vester ‘Axilla’”
Ala factus est nisi fuga litterae vastioris ? quam litteram etiam e ‘ maxillis’
et ‘taxillis’ et ‘paxillo’ et ‘vexillo’ et ‘pauxillo’ consuetudo elegans Latini
sermonis evellit.

That the suppression of the s-sound was in the earlier period expressed by
doubling the letter, we may infer from these spellingsin Virgil MSS. and auila
(for aux-la, ef. auxilla) in the Palimpsest of Plautus. The change to the single
7 seems to have been made after the reign of Augustus, simultaneously with
the adoption of J for UW in paulum, milia, &c., of s for ss in causa, fusus, &e., and,
as we shall see, of m for mm in a similar case of ‘compensatory lengthening,’
dumus from *dus-mus. Anhellus, if for *anhenslus from stem *an-anslo-, must
have had long e. Thus léquella for *loques-la, quérella for *queres-la may have
been the older forms, which were banished for a time, and were restored in
later Latin. (For statistics, see Brambach, Orth. p. 259.) Mallo, nollo are
attacked by the grammarians of the Empire, along with millia, caussa, fussus,
&c., and represent with these the older fashion of spelling and pronunciation.
Diomedes (p. 386. 13 K.) blames those ‘ qui geminant 1 litteram et enuntiant’
in these two verbs. (Does Velius Longus allude to the verb in p. 80. 5 K..
where he says: quis autem nescit ‘malum’ una | littera scriptam multum
distare a ‘mallo’ eodem elemento geminato?) How far other consonants
were doubled after a long vowel is discussed below.

§ 130. Confusion of single and double letter in Latin. In the misspellings
of inscriptions and MSS. we find a double written for a single consonant,
especially in the case of (1) mute before 7, e. g. frattre (C. I. L. viii. 111), suppra
and suppremus in Virgil MSS. (Ribbeck, Ind.) ; (2) before consonantal u (w),
e. g. tennuis in Virgil MSS., strennwior in MSS. of Lucil. xvi. 19 M. (cf. Probi App.
198. 18 K. aqua non ‘acqua’); (3) s before mute, e. g. disscente (C. I. L. iv. 1278).
We find the same doubling of a consonant in the first two cases in Italian in
fabbro, acqua, &c. (see below) ; and in classical Latin we have perhaps traces
of them in the normal spellings, quattuor, battuo. The third type of misspelling
probably reflects the attraction of s to the first syllable (see § 139). The
opposite error, of writing a single consonant for a double, appears especially
(1) after a long vowel, e.g. nula, Eph. Epigr. iv. no. 557 (Consentius warns
against mispronunciations like mile, vila, 392. 7 K.); (2) in syllables before
the accent, usually in words compounded with prepositions, where the final
consonant of the preposition has been assimilated to the initial of the verb or
noun, e. g. acepi, comunis, but also in other cases, such as Diminutives, vw. g.
sacellus (see Georges, Lex. Wortf. s.v.). There are in the classical language
a few traces of the reduction of a long syllable to a short in the pretonic
syllable, possibly in Diminutives like dfella (from offa), mdmilla (from mamma)
(cf. Ital. vanello), but certainly in prepositional compounds. Thus ommitto

T
I14 THE LATIN LANGUAGE. [Chap. II.

from ob and mitto has become émitto (Priscian i. p. 46. 18 H. omitto dicimus pro
‘ommitto’) ; in reddico the preposition was changed, perhaps by the analogy
of other compounds, to re, but not in reddo, where it is accented (unless the
true explanation here is that reddo represents *re-dido with a reduplicated form
of the verb found in Umbro-Osean, ch. viii. §9). (Velius Longus 66, 3 interdum
haec d littera geminatur, quotiens ab eadem littera sequens vox incipit; nec
tamen semper, siquidem ‘reddere’ dicimus geminata d,.. . unde adnotanda
imperitia eorum qui sic ‘redducere’ geminata d littera volunt enuntiare,
quasi ‘reddere,’ tamquam necesse sit totiens eam duplicem esse, quotiens
sequens vox ab eadem littera incipit.) The versification of Plautus shows us
that after a short syllable the preposition in the pretoniec syllable of a com-
pound was especially liable to be so slurred in pronunciation, that it might
optionally be scanned as a short syllable, e. g. quid decépit? (see ch. iii. § 34). How
far the weakness incident to the first syllable of accepit after a short syllable
adhered to it in other circumstances, and tended to reduce the acec- to the
sound of ac-, it is difficult to say. We have ore cirupto in Lucil. ix. 1. M.,
expressly attested by Consentius (400. 8 K.) ; and there seems no reason for
setting aside the reading of the MSS. in Luer. vi. 1135 an caelum nobis ultro
natura cérumptum Deferat, a reading confirmed by Isidore, Nat. Rer. 39.
The usual practice, where the preposition is assimilated, is to write a double
letter in some cases, e.g. corruptus, ommentans (Liv. Andron.); in others to write
a single letter and lengthen the vowel, e. g. dmitto. It is not always easy to
draw a hard and fast line between these two practices. A scansion like
Plautus’ quid dmittis suggests that the second word was pronounced rather
ammiittis (ef. ammissam in Virgil MSS., Ribbeck, Ind.) than dmittis ; for Plautus
does not shorten the first syllable of the compound in quid insanis, &c., where
we know the ito have been a long vowel before the group ns (see ch. iii. § 34).
The a of amittis could then hardly be on precisely the same footing as a vowel
long by nature. Adm-, though written amm- was not written dm-. (Amentum,
beside ammentim and admentum, is probably non-existent. See Nettleship, Con-
tributions s.v.) Again, Gellius says of the preposition com, compounded with
ligo, and necto (ii. 17. 8) coligatus et conexus producte dicitur. (Com before n-
is always cén- in the best spelling, e.g. cdnubiwm, but not before gn, e.g. cognatus,
unless gn has previously become x, e.g. cinitor.) The late spelling oportunus (see
Georges) may be due to the analogy of dportet; but on Greek inscriptions we
often have w, instead of the usual 0, in compounds with com, the assimilated
m being sometimes omitted, e.g. Kwyzodos, sometimes expressed, e.g. nwppextwp
(Eckinger pp. 51-2). It is therefore an open question how far these late
spellings, such as corigia in the Edict of Diocletian, indicate a real reduction
of the double consonant to a single, cérigia, or a transference to the vowel of
the extra length of the consonant, cérigia. Greek spellings of Latin words are
very uncertain guides ; for Consentius mentions as a fault of the Greeks their
inability to pronounce the double consonant in words like jussit, ille, 395. 13 K.
s litteram Graeci exiliter ecferunt adeo, ut cum dicunt ‘jussit,’ per unum s
dicere existimes ; 394. 25 K. ubi enim [Graeci] dicunt ‘ille mihi dixit,’ sic
sonant duae 1] primae syllabae, quasi per unum | sermo ipse consistat!; and
on Greek inscriptions we find double confused with single, single with double
consonant in Latin words to 9 very great extent, especially 7 and U, but not

 

' In Martial ii. 60 puer Hylle has assonance with puerile.
§ 130.] PRONUNCIATION. DOUBLE CONSONANTS. 115

often s and ss (see Eckinger). Similarly in Latin inscriptions we find Greek
words misspelt in this particular, e. g. fesera, eclesia (So in Vulg. Lat. ; ef. Ital.
chiesa, &c.), bassilica (e.g. C.I.L. iv. 1779), &c. So Probi App. (199. 9 K.)
censures ‘bassilica’; (198. 11) ‘cammera’ (cf. Sicil. Neap. cammara) ; (198.
17) ‘dracco’ ; (198. 27) ‘fassiolus.’ (In each case the vowel is a, the quality
of which was the same, whether short or long.) On a Republican inscription
(C.I. L. i, 1181) we have bassim, and in the Palimpsest of Plautus bassilice.
The Plautine form of the name ’IAAupia is Hiliiria. In other foreign words we
have a like confusion, e. g. Britanni and (later) Brittani (see Georges s.v.), as
we have a confusion in the quantity of the vowel of Batdci, &e. Very often
a wrong etymology, or wrong association, is the cause of a misspelling ; e. g.
pellex, a late spelling of paelex (see Georges), was due to connexion with pellicio ;
and the established spelling accipiter for *actipeter [probably with d, weak grade
of 6 of Greek wev-7érns (epithet of hawk in Hesiod, Op. 210), Sanscr. Aagupdtvan-,
ch. iv. § 54] to connexion with accipio; cf. the vulgar form acceptor (Caper 107.
8K. accipiter non ‘acceptor’) used by Lucilius (inc. 123 M.) exta acceptoris
et unguis. The misspelling cominus for comminus (see Georges) is due to the
analogy of aminus; and the analogy of Diminutive terminations -ellus, -illus, -ullus
is generally believed to be responsible for the later spellings camellus, anguilla,
cucullus, &. The corrupt form cémellus instead of camélus (Greek #épundos)
is indicated by the spelling in the Itala (see Rénsch, Itala, p. 460), and in
the Edict of Diocletian (11. 6, &.), as by the Italian cammello (with open ¢),
Span. camello, Fr. chameau (cf. phasellus for phasélus in Virgil MSS.) ; anguila,
the spelling of good MSS. of Latin authors, is reflected by Span. anguila
(A, L. L. viii. 442) 5 on cucilus and cucullus, see Brambach, Hiilfsbiichlein, s.v. The
same explanation is generally given of -ella for -da of loguela, querela, suadela,
lutela, medela, &c. (on which see Brambach, Orthographie, p. 258 sq.). The
grammarians approve of the single 7 in these words (Ter. Scaur. 11. i K. on
querela ; Mar. Vict. 17. 9 K. on loquela, querela, suadela, tutela, also camelus ; Caper
96. 6 K. on querela, loquela) ; but by the time of Papirian, the latter part of the
fourth cent. a.p., querella was the usual spelling (see Papir. ap. Cassiod. 159. 4 K.
Cf, Bede 287. 6 K.; Alcuin 299.6 K.; Quaest. Gram. Cod. Bern. 83. Suppl.
175.7 K.). At the same time we have seen that querella (for *queres-la), loquella,
&c. were probably the older spellings, and stand beside quallus for *quas-lus in
Virgil MSS., so that their use in Vulgar Latin may be really a case of
adherence to the older form, just as we find vulgar spellings like ussus, vissus,
messor, fressus, allium (on these see Georges), and Vulg. Lat. *vessica, attested
by Ital. vescica, &c. (ef. Capsesis non ‘Capsessis,’ Probi App. 198. 2). Pila
was in late Latin *pilla, pillula (see Georges s.v. pilula, and ef. the Romance
forms). Péno, for pé-s(i)no, a compound with the preposition po, for *apo, a by-
form of ab, was treated in vulgar speech as if *por-s(i)no, a compound with
por-. Hence the double s (for 7s, as dosswm from dorsum) in possitus (quasi
‘por-situs’), possut on plebeian inscriptions (e.g. possud/, 0.1. L. v. 5623; vil.
47. 137. 246; dipossitus, Rossi, i. 103, of 348.4.D.). In the Appendix Probi 202.
12 K. one is warned against the confusion of sera, a bolt, with serra, a saw,
a confusion seen in Ital. serrare, Span. cerrar (ef. 201. 33 on velit and
velit). Other misspellings depend merely on the substitution of a single
consonant with long vowel for a double consonant with short vowel, and
vice versa, e.g. Probi App. 199. 4 K. garrulus, non ‘garulus’ (probably
garulus) ; 198. 21 K. caligo, non ‘calligo.’ This substitution was allowed, as

12
116 THE LATIN LANGUAGE. [Chap. II.

we have seen, in prepositional compounds, like ditto, to which we may add
stipendium for stip[i]pendium (slupendiorum, C. I.L. vi. 2496, 2787, 2795 3 stependiorum
3069, of 221 A.D. point to i in vulgar pronunciation), fricae (see § 60) for
*tricc[h]ae (cf. Ital. treccare, beside Neapolitan tricare). It appears to be
consistently carried out in a prow of words, ciépa (ef. Sanser. kipas) and
clippa, stipa and stippa (Greek orémn and orinmn), miicus and miiccus, pupa
(cf. pvveAE, C.I.L. X. 4315 } PVPIVS, PVPIA, Vi. 6021) and prippa, giitus and giittus,
mittus (Greek puOos, see Class. Rev. v. 10) and miittus, *biitis (Greek Bodris) and
*piitis (cf. Ital. botte, bottiglia). The forms with short vowel and double
consonant seem to be those of late Latin and Romance (e. g. cuppa, Ulp. Dig.
xxxiii. 6, 3 § 1 and xxxiii. 7, 8 M.; Augustine, Conf. ix. 8.18; Not. Tir. 156:
puppa, Acron, in Hor. 8. i. 5. 65; Ital. coppa, Span. copa, &c., while capa is re-
flected in Ital. cupola, Span. cuba and in the Welsh cib), so that the variety in
form has arisen through the consonant being allowed to assert itself before the
articulation of the vowel had been completed, and thus to take away from
the vowel some of its force. (Similarly Middle High German muoter (I.-Eur.
*mater), with long vowel and single consonant has become n modern German
mitter.] Stréna seems in the same way to have become strénna at the end of
the Republican period ; for stréna is indicated by Span. estrena, &c., strénna
by Ital. strenna (with open e), Fr. étrenne, though the byform strenua (see
Georges) may indicate confusion with strénuus, later strennuus (see above). In
Italian the same thing is very common, e.g. venni for véni, leggi for légi,
brutto for brite ; and this may be the explanation of the puzzling form tutto
for Latin totus (see Kirting, Lat.-Roman, Worterb. s.v.; and ef. Consent. 392. 1 K.
‘tottum’ pro toto, ‘cottidie’ pro cotidie), as of Latin Juppiter (the usual spelling,
see Georges) for Jipiter! (ch. vi. § 32). Sticus however retains this form in
Vulgar Latin, and similarly brdca ; while both classical and Vulgar Latin show
baca (see Georges, and cf. Ital. bag-ola, Fr. baie), which, if the ordinary deriva-
tion be correct (see Liyma Latina s.v.), should be bacca for *bat-ca. Latin ciccus
seems to have been in Vulg. Latin *cicus, to judge from Ital. cica, cigolo, &e.

Scribes of Irish nationality were specially liable to miswrite a double for
a single consonant in a Latin word; for in the orthography of their own
language the double letter often indicated merely that the consonant had not
degenerated into a spirant. Thus a repeated substitution of double for
single consonants in a Latin MS. is frequently an indication that the MS.
has been written in an Irish monastery (see Zimmer, Glossae Hibernicae, proll.
xi). Again the confusion of single with double consonants in inscriptions
may often be due to local influence. Thus the Greeks, as we have seen, had
a difficulty in pronouncing the Latin double consonants; the Osean dialect
often shows a double letter, where a single is etymologically correct, especially
hefore a y-sound, e.g. Vitellia (Lat. I/alia), before a w-sound, e.g. dekkviarim
(cf. Lat. decem), before an 7, e.g. alttrei (Lat. alter’), and sinianty ss before t,
e.g. kvaisstur (Lat. quaestor), which remind us of Latin misspellings like acqua.
Srattre, disscente; in the Umbrian inscriptions a double consonant is never
found in those written in the native alphabet, and very seldom (sometimes
perversely, e.g. ennom, cf. Lat. enim; avvei, ef. Lat. vis) in those written in
Latin characters.

 

1 Fuppiter, quippe (ch. ix. § 7), ipsippe (ch. vii. § 20) suggest that this doubling
of p was a usage in the literary language.
§ 180.) PRONUNCIATION. DOUBLE CONSONANTS. 117

But in spite of all these facts, it still remains true that there is usually
a sufficient consensus between inscriptions and the best MSS. to enable us to
decide with certainty on the single, or on the double letter, as the classical
spelling of the word, a spelling with which the Romance forms, as well as
the loanwords in Celtic and Teutonic languages, show a remarkable agreement.
The classical form generally agrees with the form postulated by the etymo-
logy of the word, though there are some exceptions, e.g. bdca (see above)
instead of bacca, damma (but in the proper name, Dima; see Georges) instead
of dama, This implies that a sharp line was drawn in Latin between the
single and double consonant! (e.g, milus and miillus), a fact which should
make us suspicious of etymologies which ignore this distinction, such as the
identification of annus, a year (with d, Terent. Maurus v. 1239), cf. perennis),
with dnuus, a ring. For examples of this kind of uncertainty in Latin
spellings, see Georges, Lex. Wortf. and Brambach, Hiilfsbiichlein s.vv. lammina
and lamina (syncopated lamna), vacillo and vaccillo, mamilla rarely mammilila,
buccella and bicella, disicio and dissicio (ef. porricio), cotidie and cottidie (see Etyma
Latina s.v.), mutonium and muttonium, muttio and mutio, glutto and gluto, murgiso
and murgisso, mantisa and mantissa, favisue and Favissae, favisor and favissor,
comi: and isor, Masinissa and Massinissa, ph lus and passeolus, T'issaphernes
and Tisaphernes, Porsenna and Porstna (ef. Pompeius, p. 284 K., who also
censures (a suppositious?) Catilinna], meddix and medix (Oscan meddis Nom.,
medikeis, Gen.), Apuleius and Appuleius (but Apulia better than App-), Marcomani
and (later) Marcomanni. (Panus and pannus may be different words. See
Berl. Phil. Woch. 1887, p. 214.) The double consonant is declared by them to be
the better spelling of bracchium (with cch for x of Gk. Bpdyioy: see § 60), littera
(cf. Rumanee, e.g. Fr. lettre, and Welsh llythyr. In the Lex Repet. (ALL. i.
198) of 123-122 B.c., once LEITERAS, but the spelling of a single for a double
consonant is usual on this inseription, and ei is used for i in seine], futiilis,
caccdbus, ciilleus, tralleum, cuppes, cuppedo (for citp-), lolligo (Fritsche, ad Hor. S. i.
4. 100), fello (for fé-lo), helluor, sollers, sollemnis, sollicito, pappare (Plaut. Epid. Goetz,
pref. p. xxx), cippus, liypus, cdiperro (see Nettleship in Class. Rev. 1892, p. 168),
Messalla (cf. Hispallus with U on C.I.L. i. 39), Sallustius, barritus, Arruns, Arretiun
(now Arezzo), alléc, Allecto, Allifae, Sardiénidpallus, ballaena (Gk: padd- rather than
par-), ballista, Sallentini, civinnus, pétorritum, Trdstimennus, Appenninus (and Ap-),
nummus, immo, dissextum ; the single of balbutio, litus (so Vulg. Lat., e.g. Ital.
lito and lido), bucina, alucinor, besalis, belua, stivio, muriola (cf. Paul. Fest. 125,
13 Th. murrina, genus potionis, quae Graece dicitur nectar. _Hanc mulieres
vocabant muriolam), Erinys, Apulia, Sufes, tépete, Larisa, sarisa, Gnosus, Parnasus,
talasio, pedisequus, ilico ; the double consonant appears to be the older spelling,
the single the later in mantellum (Plaut.) and mantéle, stéllio and. stelio, pilleus
(so, for example, in MSS. of Martial; see Friedlinder’s edition, i. p. 117), and
pieus (pill- in Romance, K. Z. xxxiii. 308), marsuppium and marsiipium, Marpessos
and Marpésius, and possibly the legal parret and pdret (Fest. 292. 25 Th., parret,
quod est in formulis, debuit et producta priore syllaba pronuntiari et non
gemino r scribi, ut fieret ‘paret,’ quod est invenitur, ut comparet, apparet).
The spelling paricida for parricida belongs to a period before the doubling of

 

1 Plautus, however, puns on mittis Vérberon etiam, dn iam mittis ?
and mitis in Mil. 1424, when the Mitis sum equidem fustibus.
soldier is getting a thrashing:
118 THE LATIN LANGUAGE, (Chap. IT.

consonants was practised. Of Greek loanwords we have O. Lat. creterra (Gk.
«pnTnp), perhaps grammosus (from Gk. yAnpn) in Caecil. Comm. 268 R. (but gramae
Plaut. Curc. 318, Biich. Rh. Mus. xxxv. 72), grabattus (Gk. «paBaros\, &e.
A.L.L. viii. 367)1. (See also Ellis Catullus p. 338 on Varus and Varro,)

§ 131. Double consonants in Italian. These are not known in the dialects
of Umbria and of North Italy ; and their use varies a good deal in different
parts. They form one of the greatest difficulties to English learners ; for
a double consonant is unknown in our language, except in compound words
like ‘ bookcase,’ ‘ penknife’ ; as they proved a stumbling-block in old times
to Greeks (ef. §§ 99 and 117 on the Greek mispronunciation of I, ss). Double
consonants have replaced Latin single consonants before y, e.g. occhio (Lat.
be(uylus), vendemmia (Lat. vindémia) ; before the w-sound of Latin aqua (Ital.
acqua); before 7, e.g. fabbro (Lat. faber). (Compare the doubling of a con-
sonant in these positions in Oscan orthography, and similar misspellings in
late Latin inscriptions.) Doubling is very common under the accent of
a paroxytone word, e.g. femmina (Lat. fémina), legittimo (Lat. legitimus), and
on the first syllable, when it hasa secondary accent, e.g. pellegrino, tollerare
(ef. late Latin suppellectilis, see Georges). Forms like allodola (Lat. alauda),
commedia (Lat. cémoedia) seem to be due to the analogy of words compounded.
with prepositions, like Latin «lliido, commédus (ef. Ose. Appelluneis
* Apollinis’ ?).

§ 182. Double consonant (not 1, s) after long vowel. We know that
the e of fressus, for frensus, later fresus, was long ; but we cannot tell whether
the true explanation of the change from dwmmetum to diimétum (see Georges)
is to make it similarly a reduction of a double to a single consonant after
a long vowel, or merely a substitution of a vowel length (dm) for consonant
length (timm), like dmissam for ammissam (in MSS. of Virg. 4. ii. 741), as the
-amm- of flammen, a blast, in Virg. MSS. (see Ribbeck) seems to represent the
usual -dm- of fla-men. A spelling like ruppes for ripes in Virgil. MSS. (see
Ribbeck) suggests rather the alternation of ctippa with capa, ptippa with
piipa (see above); and the quantity of the vowel before the double con-
sonant of lammina (see Georges), vaccinnia (see Ribbeck, Index) is quite uncer-
tain. Clear cases of a long vowel before nn are mercennarius (the correct
spelling, according to Brambach, Hii/fsbiichi. s.v., later mercenarius), tinnire (for
7 and nn are attested by Port. tinir, Sard. tinnire), and perhaps hinmuleus (also
inuuileus, inuleus, see Georges) (ef. Agroecius 115. 14 K. hinnuleus, ut 7 acutum
sit, quia nomen a sono vocis accipit), and Vinnius (also Vinius; see C.I.L.
vi. 28978 sqq. Long i is indicated for Vinnia, 28986) ; but before other double
consonants they are difficult to establish. NArrem on the carefully written
inscription of the Emperor Claudius (48 a.p.} at Lyons (Allmer et Dissard
vol. i. p. 70"; Boissieu p. 136) may be due to the analogy of gndrus, narus.
which made the spelling naro (proposed by Varro, if we are to believe Papirian

 

' Sometimes the interchange of Lat. Acca Larentia, Gk. "Axed (a name
single and double consonant isI-Eur. of Demeter), O. Ind. akki; Gk. Zevvd,
In pet-names we often finda double &c. The double consonant of Eppius,
consonant : e.g. Gk. vavva beside O.  Seppius, &c. has been so explained,
Ind. nana; Lat. mammas atque titas (ch. though others refer it to a dialectal
y. § 81) beside Gk. drra, O. Ind. atta; doubling before y like Osc. Vitellia.
§§ 1381-1384.) PRONUNCIATION, FINAL CONSONANTS. 11g

ap. Cassiodor. 159. 8; ef. Varro. L.L. vi. 51) approved by some grammarians
(e.g. Velius Longus 80. 9 K.), though never accepted in popular usage (see
Georges) (cf. varus and Varro). (On *trippa, the original of Ital. trippa, our
‘tripe,’ &c., see Kérting’s Lexicon s.v.) A certain instance of the reduction
of t to ¢ after a diphthong is the late form autor (censured, with autoritas, in
Probi Appendix 198. 30 K., and found on late inscriptions, e.g. C. I. L. viii.
1423 ; ef. xii. 2058, of 491 a. D.), where the t represents ¢t for original ct (see § 95).
Ital. freddo, Fr. froid point to *friddus, from frigdus, a vulgar form of frigidus
(Probi App. 198. 3 frigida non ‘frigda’) ; O. Span. frido to *fridus or *friddus.

§ 1383. Final double consonant. A finaldouble consonant was not allowed
in Latin orthography, but was written single, e.g. miles for *miless, from
*milit-s. But that it differed in pronunciation from an ordinary final single
consonant, we may infer from the forms censured in the Appendix Probi,
where -x is wrongly substituted for this -s (originally -ss) (197. 28 K. miles
non ‘milex’; 198. 29 aries non ‘ariex’; 199. 4-5 poples non ‘ poplex,’
locuples non ‘locuplex’), forms which are found on inscriptions (e. g. milex,
C. I. L. vi. 37, 2457, 2549, &c.). We may infer also that there was a change in
its pronunciation in course of time ; for in Plautus miles has the last syllable
long (Aul. 528), while in Ennius, Lucilius, &c., its last syllable is short (Ann.
277 M.; so milés, Lucil. xi. 8 M.), though never shortened before an initial
consonant like ordinary -tis,-is. Plautus also scans ter for *terr (cf. terruncius,
the true spelling ; see Rhein. Mus. xlvi. p. 236) from *ters, *tris (Greek rpis), as
a long syllable (Bacch. 1127), as he scans es (2 Sg. Pres. Ind. of sum), prodes, &e.
like *ess, *prodess (contrast cir, Lucil. xv. 9 M. ; prodés, id. ine. 128). A relic
of this usage remains in the scansion of hoc for *hoce from *hod-(c)e as
a long syllable by the classical poets; and the remarks of the grammarians
on this scansion explain the reason of the change and uncertainty in the
quantity of these final syllables. Thus Velius Longus (54. 6 K.), commenting
on Virgil’s ‘ hoe erat, alma parens’ says: ergo scribendum per duo ¢, ‘ hoc-c-
erat alma parens,’ aut confitendum quaedam aliter scribi, aliter enuntiari ;
Pompeius (119. 13): item clittera aliquando pro duabus consonantibus est...
ut...‘ hocerat alma parens’ : ‘hoc,’ collide ¢, ut sit pro duabus consonantibus.
in illo alio exemplo brevis est, ‘solus hicinflexit sensus’ : sic lubrice et leniter
currit, (Velius Longus also fails to make this proper distinction between hocc
for *hod-c and hi-c.) They show us that in pronunciation *hocc (and presumably
*corr, *ess) were actually sounded with double consonant when the next
word began with a vowel, at least if the accent fell on them, while before
a consonant initial, and probably when unaccented, the double consonant
would be reduced to a single, hoc fuit but hoce erat. The unaccented nature of
the Substantive Verb *ess, and of the final syllable of *miless, prodess, &c.
explains their speedy reduction in Latin prosody (see ch. iii).

§ 184, Final consonants. A final consonant is always more
liable to weakening than an initial, because of the general
tendency of languages to pronounce with diminishing stress.
It is especially so in English after a long vowel. Thus in the
word ‘ cat’ the ¢ is uttered with less force than the c, while the
same final after a long vowel, as in ‘cart,’ is still weaker. It is
120 THE LATIN LANGUAGE. (Chap. IT.

a rule of our language that a final consonant is always short
after a long vowel, as we may see, if we contrast a word like
‘heel’ with a word like ‘hill’? The weakness of Latin final
consonants has been already mentioned, Final -d was dropped
in pronunciation after a long vowel about the end of the third
cent. B.c.; final s does not constitute position before an initial
consonant in that species of poetry which most closely imitated
ordinary pronunciation, Dramatic Poetry; final -m offers but
slight resistance to the elision of the vowel which precedes it ;
the tenues fortes seem to have been replaced when final in pro-
nunciation by the mediae lenes, e.g. ab, sub, religuid, &e, (cf.
negotium ?) (see § 73).

In considering the pronunciation of final consonants it is
necessary to regard not isolated words, but words as they stand
in the sentence. The accent of a word, when standing alone, is
something different from its accentuation in the sentence; e.g.
Greek apés, but mpds méAtvy 7AOe. And the same is true of its
pronunciation. The Greek orthography indicates the first dis-
tinction, but rarely the second ; though we find it to some extent
on inscriptions rnu moAw, éy duxys, &e., especially in Cretan inscrip-
tions (see ch. ii. § 41). But in writing Sanscrit the principles
of ‘Sandhi’ (i.e. putting together, synthesis), to use the native
term, were carefully followed by the grammarians of India. The
neuter Demonstrative, for example, tad (Lat. zs-d#d) had its final
d changed according to the following initial consonant in tat
tapas, that heat (Lat. is-¢wd *¢epus), tal lihati (Lat. is-tud Lingit),
tan nahyati (Lat. is-tud nectit), &c. There was something like
this in Latin. Traces of it appear occasionally in inscriptions
and MSS8., e.g. im burim in MSS. of Virgil, G. i. 170; and we
have ‘doublets’ like negue, atgue before vowels, nec, ac before
consonants ; but for the most part it is not indicated in spelling.

Final consonants lingered longest in monosyllables, especially
accented monosyllables, and before being entirely discarded in
spelling, passed through the ‘doublet’ stage; that is to say,
they were retained in pronunciation in certain positions in the
sentence, before an initial vowel usually, and dropped in others ;
e.g. hau scio but haud habeo, just as the r-sound is found in
English only before an initial vowel following without any pause,
§ 135.] PRONUNCIATION. FINAL CONSONANTS. 121

e.g. ‘here he is.’ The same process went on in the Romance
languages, of which French was the most retentive of final
consonants till comparatively modern times; though now, for
example, final -¢, -s, -r exist only in pre-vocalic ‘ doublets, e.g.
vient-il ? with ¢ sounded, but il vien(t). An English example of
‘Sandhi’ is the different vowel-sound of the article ‘the’ before
a vowel and before a consonant, and an example of the abandon-
ment of one ‘doublet’ and the exclusive use of another is the
preposition ‘ with,’ which now ends only in the ¢4-sound of thin,
but which in early modern English had in certain collocations
the ¢/-sound of ‘this.’ Both ‘doublet’ forms remain in.‘ my’
and ‘ mine,’ ‘ one’ and ‘a,’ ‘an,’ ‘naught’ and ‘not,’ &e.

The treatment of final vowels in Latin is most naturally
considered in connexion with the changes produced by the accent ;
for they are affected much in the same way as the vowels in
post-tonic syllables (see ch. iii. § 40).

§ 1385. ‘Sandhi’ in Latin :—Verrius Flaceus proposed a new symbol for
final m, when the next word in the sentence began with a vowel, a symbol
like the half of the ordinary letter M ; while Cato the Elder wrote dicae for
dicam, faciae for faciam (see § 61). The tendency of final -m, -n to adapt
themselves to a following consonant-initial, is seen in spellings on inscrip-
tions, like im balneum, C.1.L. iv. 2410, imbello, iii. 4835, im pace, viii. 10542
(for examples see Indices to C.J. Z.) and in MS. spellings like im mare, im
medio, im pace, im puppim in Virgil MSS. (see Ribbeck, Ind. p. 433), im praeda,
im uita in the Plautus Palimpsest (see Ind. to Studemund’s Apograph.). Caper
(106. 17 K.) says: in Siciliam dicendum, non ‘is Siciliam,’ card 7d v, non
xara 76 o, quia nunquam sine n pronuntiatur (leg. insicia .. non ‘is.’?). We
have eliannunc in the Herculanean papyri (Class. Rev. iv. 443), and etiannum,
jandudum, &c. in MSS. of Virgil (see Ribbeck), spellings which agree with the
statement of Velius Longus (78. 19 K. cum dico ‘ etiam nune,’ ‘ quamvis per m
scribam, nescio quomodo tamen exprimere non possum), and Cicero’s remarks
on the sound of cum followed by n- (Or. xlv. 154 ; Fam. ix. 22. 2; ef. Quint. viii.
3.45; Diom. 450. 34 K. ; Pompeius 293. 17 K. ; Prise. i. 372. 8and 594. 21 H.
&c.). Est was curtailed in writing, as in pronunciation, like our ‘is,’ in ‘it’s,’
‘he’s,’ &., audiendust, audiendast, audiendumst, &c., a spelling recommended by
Mar. Victorinus (22. 14 K.), and found in MSS., e.g. in Virgil MSS. acerbist,
locutast, ventumst, amantemst, cupidost, suprast, &c. (see Ribbeck’s Index, p. 419), in
the Plautus Palimpsest copiast, aegrest, homost, olimst, palamst, meliust, &c., and
similarly with es, iratas, dignus, itwrus (generally printed by editors irata’s, &c. ;
once with es Imperat., viz. ‘molestus, Most. 955) (see Studemund’s Index,
p. 505). One may perhaps see the beginnings of the suppression of final
consonants in the tendency of pronunciation mentioned by Consentius
(fifth cent. ? a. D.) (395. 7 K.), the tendency to detach a final consonant from its
word, and join it to a following initial, ‘si cludit’ for sic ludit, ‘si(c) custodit’
122 THE LATIN LANGUAGE. [Chap. II.

for sic custodit: item litteram ¢ quidam in quibusdam dictionibus non latine
ecferunt, sed ita crasse, ut non discernas, quid dicant : ut puta siquis dicat
‘ sic ludit,’ ita hoe loquitur, ut putes eum in secunda parte orationis cludere
dixisse, non ludere ; et item si contra dicat illud, contrarium putabis. alii
contra ita subtiliter hoc ecferunt, ut cum duo ¢ habeant, quasi uno ¢ utrum-
que explicent, ut dicunt multi ‘sic custodit.’ (Cf. his remarks (394. 7 K.)
on the pronunciation ‘ dixera millis’ for dixeram illis.]

§ 136. Latin ‘Doublets.’ By pretonic Syncope (see ch. iii. § 13) ac (for *atc),
nec, neu, seu, replaced atque, neque, nere, sive before a word beginning with a con-
sonant. In Dramatie poetry the final -¢ is always suppressed, in similar
circumstances, of nempe, and often of unde, inde, quippe, ile, and perhaps iste.
Similarly proin, dein seem to have developed from proinde, deinde, when
a consonantal initial followed. Final -d, after it had been dropped after
a long vowel in the pronunciation of most words, remained in monosyllables
like haud, med, ted ; haud being the form used before a vowel, haw before a con-
sonant (Caper. 96. 4 K. ‘hau dolo’ [leg. haud uolo?] per d recte scribitur,
etenim d inter duas vocales esse debet. quod si consonans sequitur, d addi
non debet, ut ‘hauscio’; Mar. Vict. 15.21 K. So in Plautus, Ritschl Opuse.
ii. sor 2. and v. 352) ; the same probably being true of med, ted. (On qui(n)e,
quandoc and quandoque, see ch. x. § 15, ch. ix. §10) Preposition ‘doublets’
@, db, abs ; @, ec, ev, &c., on which see ch. ix. §§ 12 and 29.

§ 137. Dropping of final consonant in Latin. It was a rule of Latin,
pointed out by Julius Caesar, in criticizing Varro’s spelling Jact, that no
word could end in two mutes (Pompeius 199 K.; Caper 95 K. On lacte, tact,
lac, see Georges, Lex. Wortf. s.v.). Nor was a double consonant allowed to end
aword. Plautus gives to miles, es, &c. the scansion of miless (for *milit-s), ess,
&e. (see ch. viii. § 2); but almost the only trace (a doubtful one) of spelling with
-ss is noss in the Ambrosian Palimpsest in Stich. 536; though Velius Longus
(54. 6 K.), commenting on Virgil’s ‘hoc erat, alma parens,’ half proposes to
write hoccerat : ergo scribendum per duo ¢, ‘hoccerat alma parens’ aut con-
fitendum quaedam aliter scribi, aliter enuntiari (ef. Pompeius 119. 13 K. ;
Prise. ii. p.6. 1 H. So hoccine for *hocce-ne, *hod-ce-ne, Prise. i. p. 592. 22 H.).
Mel (for *mell, *meld), cor (for *corr, *cord), ter (for terr, ef. terr-uncius, *ters, *tris)
are short in Ovid, &., though long (neither mdi nor mé are found) in Plautus ;
but the difference between -s (from original -ss) and ordinary -s is shown even
at a late period by spellings like milex, praegnum (see § 125).

Final -d after a long vowel is written throughout the S. C. de Bacchanalibus
of 186 B.c. (C. I. L. i. 196, sententiad, exstrad, facilumed, &e.), though it is not
found in the decree of Aemilius Paulus Macedonicus of 189 B.c. (ib. ii. 5041,
in turri Lascutana, ea tempestate), and probably does not appear in Plautus,
excepting in the Pronouns (Abl. and Acc.) med, ted, sed. Even these Pronoun
forms are out of use by Terence’s time. The retention of haud along with
haw shows the course which this final @ must have taken. Before vowels it
would remain pronounced until the preconsonantal form had driven the full
form from the field (so in post-Augustan poetry we find nec more and more
supplanting neque) ; before consonants it would probably be first assimilated,
e.g. haud ligo, pronounced haulligo, like alligo, haud scio, pronounced hausscio
like a(s)scisco, then dropped. After a short vowel, it is often written -t on
late inscriptions (and indeed from the end of the Republican period), but is
not dropped (see Seclmann’s list, p. 366).
§§ 136,137.] PRONUNCIATION. FINAI, CONSONANTS. 123

Final -m is dropped in early inscriptions before a consonant or a vowel-
initial with equal frequency, in the earliest inscriptions more after o of the
Gen. Plur. (perhaps not yet shortened before -m)}, than after o of the Acc. Sg.
Mase. and Nom. Ace. Sg. Neut. On the older Scipio epitaphs it is usually
dropped, e. g. oino (Acc. Sg.), dutonoro (Gen. Pl.) (C. I. L. i. 32°. But from c. 130 it
is regularly retained in spelling [as also on State inscriptions like the S. C.
Bacch. of 186 8.c., the (restored) Columna Rostrata, &c.], until the plebeian
inscriptions of a later date (see § 65). On these it is not merely dropped
but also is written -n, as final -n is occasionally written -m (see Seelmann’s
lists, p. 364). It never fails, as final -s may fail, to constitute ‘ position ’
before an initial consonant in early poetry ; though the frequency of the
scansion enim before a consonant in Plautus (where the final syllable is
shortened by the law of Breves Breviantes) suggests that this represents the
usual pronunciation of the word. Final m before an initial vowel seems to
have been equally adapted with a final long vowel or diphthong for what is
called ‘Prosodical Hiatus,’ i. e. for being scanned as a short syllable, instead
of being elided. Ennius, for example, ends a line with millia militiim octo, as
he begins another with Scipid invicte. (Cf. circu(myire, septu(m)ennis, septu(m)-
aginta, but septumus, &c.) (On the treatment of -m in poetry, see § 65; and
on its weak pronunciation in ordinary speech, § 61.) The course it took is
perhaps indicated by Consentius (394. 7 K.), who says that the common way
of pronouncing a phrase like ‘dixeram illis’ was to detach the -m from the
first word, and join it to the initial of the second. (Cf. Pompeius 287. 7 K.)
(see §. 61.)

Final -s is dropped on early inscriptions especially in the Nom. Sing. of
I0-stems, written -io, or -i (see Index to C. I. L. i'. p. 602). That both io and -7
represent the same sound -i(s) (ch. vi. § 2) is quite possible. (Cf. Cornelio on
one Scipio epitaph, C. I. L. i. 31, c. 250 B. c., Corneli on another, i. 35, v. 160
B.c.) But asarule -s is dropped only after a short vowel, except in the
dialect of Pisaurum, e. g. matrona(s) (ib. i. 167 sqq.), and is more often retained
than dropped.

Cicero speaks of its failuré to prevent elision of a preceding 7 in the phrase
vas’ argenteis ; and some have thought that it is occasionally elided before an
initial vowel in Plautus, e. g. com(is) incommodus, Bacch. 401, amatu(s) es written
in the MSS. amatus. But all the instances admit of other explanations ;
wmatu’s is rather a case of prodelision, like our ‘it’s’ for ‘it is’ (the length of
the wis due to the double s with which es ended in Plautus’ time, ch. viii. § 2) ;
plur(is) existumo of Plaut. Pers. 353 may easily be a mistake for plure, which
Charisius tell us was used in O. Lat., and soon. Whether aequcinimitas implies
a pronunciation aequ(us) animus is doubtful ; it seems rather to come from the
Compound (ch. v. § 80) aequ-animus (cf. the gloss ‘ Animus aequus’ duae partes
orationis ; ‘animaequus’ ipse homo, ¢C. G. L. v. 266. 11-12). (On final s not
constituting ‘position’ before an initial consonant in the older poetry, see
§ 126.) The Latin loanwords in Teutonic seem to have still possessed -us
‘e.g. Goth. sakkus, a U-stem, Germ. kurz, from Lat. saccus, curtus\, but to
have lost the final consonant of -um (Zeitschr. Roman. Philologie, xvii. 559).

Final -t is often written -d on late inscriptions, e.g. reliquid, fecid (see
Seelmann’s list, p. 366), which probably indicates change to the media lenis
in pronunciation. It is dropped with great frequency in the graffiti of Pompeii,
e. g. valia, ama (see Index to C. 1. L. iv.).
124 THE LATIN LANGUAGE. (Chap. II.

Final -nt loses the dental, and is written -n, or -m, on late inscriptions, e. g.
Jecerun (see the Indices to the Corpus), though, no doubt, the -t was heard
before an initial vowel, like the -t of Fr. vient in vient-il? Dedro (C.J. L. i.
177 Matre Matuta dono dedro matrona ‘Matri Matutae donum déderunt
matronae’) is a form belonging to the dialect of Pisaurum in Picenum. The
dropping of final consonants (-m, -d, -r, -f; -t, -n, -s) in this order of frequency
is a feature of Umbrian (see von Planta i. 568).

§ 188. Dropping of final consonants in Romance. Lat. final consonants
are better preserved in monosyllables than in other words. tis in Italian
e, before vowels ed ; in O. Fr. and Prov. eand ed; in Span. y and e; aut*is in
Italian 0, od; Fr. ou, Span. o; dd in Ital., Prov. and O. Fr. is a before con-
sonants, ad before vowels, &c. -M remains in the monosyllables, Fr. rien
(Lat. rem), Span. quien (Lat. quem), &c.; -1 in a monosyllable like mel, Fr.
miel, Span. miel, Ital. miele; -« in the monosyllable cor, Fr. cueur, O. Span.
cuer, Ital. cuore ; -» remains in non (Ital. no and non). In longer words,
-m is dropped, e. g. Ital. dieci from Lat. décem, amava from Lat. dmabam ; -t is
retained in Fr., e.g. O. Fr. aimet, but Ital.ama, Span. ama. In Sardinian the
form used ‘in pausa’ (at the end of a sentence, &c.) is amat, before a vowel
amad, e.g. amad issu, before a consonant ama, e.g. ama su padre; -l and -r
are lost in Italian, e. g. frate, tribuna, insieme (cf. Span. ensieme, but Fr.
ensemble) ; -s is lost in Ital. (though in monosyllables it leaves an i, e.g. noi.
crai, which is absorbed in a preceding e, e. g. tre), but it is retained in Fr.
and Span., e. g. Ital. tempo, Fr. temps from Lat. tempus, and from Lat. cantas,
légis, Fr. chantes, lis, Span cantas, lees ; -d is lost in Ital. ché, Span. que from
Lat. quid, but remains in 0. Fr. qued ; -c has disappeared in Ital. di (Lat.
dic), si (Lat. sic) ; -nt is -n in Ital., Span. e.g. Ital. aman-o, Span. aman, but
remains in Fr., e. g. aiment ; -z remains in Fr. six, Span. seis, but notin Ital.
sei. It thus appears that French has been far more retentive of final con-
sonants than Italian or Spanish. In the Sardinian dialect of Italian (Sardinia
was the earliest province, and its dialect is a descendant of the earliest stage
of Vulgar Latin), all final consonants remain, except -m, e.g. tempus, amas,
amat, ses, amant, nomen, but adapt themselves to the following initial, e. g.
est bennidu (pronounced ‘es b-’) (see Meyer-Liibke Ital. Gramm. p. 156). But
in standard Italian there are still traces of these lost final consonants of
monosyllables, e.g. ebbene for e bene (Lat. et béné), ovvero for o vero (Lat. aut
véro), dimmi for di mi (Lat. dic mihi), checcosa for ché cosa (Lat. quid causa),
where the double consonant is due to the final having assimilated itself to
the following initial, as Latin a arose from a collocation like ab-binis, *am-me
(for ab me), *ap-patre (for ab pdtre). In French we see ‘Sandhi’ carried to far
greater lengths than Italian, where almost every word ends in a vowel.
Before an initial vowel, French -s, -t,-r are heard in pronunciation, and
a nasal vowel resolves itself into an oral vowel followed by». And, more
curious still, -1 of words closely joined to a following word beginning with
a consonant suffers the same change as | before a consonant in the middle of
a word and becomes w; e.g. du pére, au pere, beau, like autre, &c. InS.
Spain -s becomes h, or is dropped, e.g. ‘Cadi(h)’ (Storm. Engl. Phil? i. p. 71).

§ 189. Syllable - Division. The Romance languages show
a remarkable agreement in their division of the word into
§§ 188-140.] PRONUNCIATION. SYLLABLE-DIVISION. 125

syllables, their principle of division being to make the syllable
end with a vowel, and begin with a consonant, or combination of
consonants. Any combination of consonants, that is pronounce-
able at the beginning of a word is made to begin the syllable,
with the one occasional exception of combinations beginning
with s, where the s is in some languages allowed to end the
preceding syllable. An Italian says o-bli-quo, te-cni-co, e-ni-gma,
a-tle-ta, no-stro, be-ne, a pronunciation which often offers con-
siderable difficulty to Englishmen, who would, for example, more
naturally pronounce the last word as ben-e, like ‘any’ A
Spaniard says ha-blar, bu-llir, but nues-tro, attaching the s to
the first syllable. The Roman division of syllables was that of
the Romance languages, not of the English, as is proved to
certainty by the very precise and unmistakable statements of
the grammarians on the subject. Their rule is ‘ Never let a syl-
lable end in a consonant if the consonant can possibly be
pronounced at the beginning of the next syllable’; and they
give examples like pdte-stus, no-ster, a-mnis, ma-gno, a-gmen.
The same method is followed in those inscriptions which indicate
the syllables by dots, e.g. C./. Z. vi. 77 T-AN-NI-VS‘-HE-DY-PNVS,
11682 VI-XIT-AN-NIS, as well as by contractions, where the initial
letters of the syllables are used, like ma (magnus), omB (omnibus),
pp (propler); though on inscriptions we often find s taken with
the preceding syllable in words like caE-LES‘TI (vi. 77), SES
TV-LE-IVS (ix. 4028), with which we may compare misspellings
like disscente (vide § 130). Occasionally a grammarian urges
the advisability of regarding the etymological formation of com-
pounds like abs-témius, 0b-liviscor; but such remarks only show
that the natural pronunciation of these words was ab-stemius, o-
bliviscor, just as we in natural utterance disregard the formation
of phrases like ‘at all,’ ‘ at home,’ and pronounce ‘a-tall,’ ‘a-tome.’

§ 140. Testimony of grammarians. Servius, in Don. iv. 427. 20 K., states
the rule as follows: quotienscumque quaerimus, quae consonantes in
seribendo sibi cohaereant vel cui syllabae imputentur, utrum priori an
sequenti, similitudo aliorum nominum hunce solvit errorem. ut puta ‘ aspice’
... intelligimus...setp...consonantes sequenti tantummodo dare nos
debere, eo quod invenitur sermo qui a duabus istis consonantibus inchoetur,
ut ‘spica.’ similiter ‘amnis’: debemus m et n sequenti syllabae dare in
scribendo, quoniam invenitur sermo qui ab his consonantibus inchoetur, ut
126 THE LATIN LANGUAGE. (Chap. II.

‘ Mnestheus,’ * attulit’: non possumus duo t sequenti syllabae dare, quia
nullus sermo invenitur, qui a duabus t consonantibus inchoetur, et hoc
in ceteris consonantibus observabimus. plane scire debemus, conexiones
quod dico consonantium non eas quae latinis syllabis congruunt, sed
etiam quae graecis, excepta scilicet ea syllaba quae constat de b et
d, quae in latinum sermonen numquam ita transit, ut cohaereat, ut est
B5édAa. quando enim scribimus ‘ abditur,’ non possumus a in una syllaba
ponere et b et d in sequenti. He thus testifies to a-spicé, a-mmnis, at-tulit, ab-
ditur. Similarly Caesellius (ap. Cassiod. vii. 205. 1 K.) to pote-stas, no-ster, ca-pto,
plo-strum, lu-strant, capi-strum, clau-strum, ra-strum, campe-stre, a-stla (for astula),
pe-stlum (for pestilum), car-po, dor-sum, Por-cius, Pa-ris, la-pis, tu-tus, sol-vo, ner-vus,
rol-ro, lar-va, pul-vis, te-nu-is. but dissyllabic ten-vis, be-lu-a and bel-va, ma-lu-a and
mal-va ; Terentianus Maurus (vi. 351, v. 879 K.) to o-mmis, a-mmnis, and (v.
904 K.) ma-gnus, di-gnus, a-gnus, si-gna, pu-gna (v. 941 K.), fa-xo, a-ais, ne-cus, u-xor,
no-xia ; Marius Victorinus (vi. 29. 20 K.) to a-mnis, ar-ma, a-xis (ef. Charisius,
i. rr. 19 K.; Dositheus vii. 387. 4 K.); Caper (vii. 96. 9 K.) to no-strum,
ve-sirinn, maje-stas; Dositheus (vii. 385. 5 K.) to a-gmine, ma-gno ; Priscian (i. p. 42
H.) to a-bdomen, My-gdonides, Abo-dias, A-tlas, Ae-tna, i-pse, nu-psi, scrixpsi, scri-ptum.
dra-chma, a-gmen, vi-ctrix, sce-ptrum, thus admitting, unlike Servius, bd into the list
of pronounceable combinations, and (p. 50 H.) pa-scuc, lu-scus, Co-smus, pro-spera,
te-stis, &. Bede and Alcuin insist on copyists of MSS. breaking up words at
the end of a line according to these rules, ma-gnus, pro-pter, colu-mna, &e,
Etymological division is recommended by Quintilian (i. 7. 9) with the
instances hari-spex, abs-temius (quia ex abstinentia temeti composita vox est) ;
by Caesellius (ap. Cassiodor. vii. 206. 1 K.), ob-liviscor, and (205. 18 K.), di-spicio.
abs-tulit, trans-tulit, abs-condit) ; by Aleuin (vii. 306. 4 K.), ob-stipui, ob-sum, ob-strepo,
obs-olevit (ef. Cassiodor. vii. 204. 19 K.). Priscian similarly says (i. p. 45 H.)
si antecedens syllaba terminat in consonantem, necesse est etiam sequentem
a consonante incipere, ut ‘ar-tus’ ‘il-le’ ‘ ar-duus,’ nisi sit compositum, ut
‘ab-eo’ ‘ad-eo’ ‘ per-eo,’ but adds that Herodian in his treatise on Ortho-
graphy declared it to be ‘rationabilius sonoriusque’ to follow the ordinary
syllable-division in the case of Compounds too ; and in another passage (i. p.
42) he hesitates between a-bnito and ab-nuo, Terentius Scaurus (vii. r2. 1 K.)
censures ‘nes-cio’ for ne-scio, a mispronunciation which shows the tendency
already mentioned (§ 139) to detach s from a following consonant or conso-
nant group, or perhaps rather to divide it between the two syllables, ‘nes-scio.”

The law of Breves Breviantes in Plautine prosody, it may be mentioned,
takes no account of syllable-division. Shortening is allowed (after a short
syllable) of a pretonic syllable long by position in words like gubérndbunt.
cavilldtor, volintitis, where the consonant group is divided between two
syllables neither more nor less readily than in words like egéstdti, venistiti,
where the consonant group is confined to one syllable.

§ 141. Quantity. he quantity and the quality of a vowel are
two different things. We are apt to distinguish in our minds
a long and a short vowel (say ¢ and é) by quality, not by
quantity, thinking of ¢ as an open E-sound, of é as a close
E-sound, whereas the terms ‘long’ and ‘short’ should be
§ 141.) PRONUNCIATION. VOWEL-QUANTITY. 127

applied only to the amount of time taken in pronouncing the
vowel, so that there is, properly speaking, a long and a short
open E and a long and a short close E. It is true that differ-
ence in quantity and in quality often go together; thus Latin ¢
was, like our @, open E, Latin é was close HE, though the long
sound of open E was also known in Latin, and was written
ae (§ 6). The Romance languages, which have lost all other
distinction of the Latin long and short vowels, distinguish them
according to quality (e.g. Lat. 4é/us is Ital. bello with open E,
Lat. sté/a is Ital. stella with.close E), though this distinction
of quality does not always correspond to distinction of quantity
(e.g. the 7 of Lat. video and the @ of Lat. credo are similarly
represented in Fr, vois, crois) ($6). Consonants, too, may differ in
their quantity like vowels. For example, English final consonants
are long after short, short after long vowels, e. g. ‘hill,’ ‘ heel.’

We may distinguish at least three degrees of quantity or
length,—long, short, and half-long,—an example of the last
being the vowel of our ‘note,’ while ‘node’ and German ‘ Noth’
have a long vowel. Latin half-longs may be detected by the
metrical scansion of a syllable as either long or short, e.g. in
Plautus’ time the final syllables of dmat, ténet, abit, dolor
(ch. iii. § 40)?.

The marked distinction between a long and a short vowel in
Latin made it possible for the Romans to imitate the quantita-
tive metre of the Greeks. Their own native metre, the Satur-
nian, which is represented in literature by the ‘Odyssea’ of
Livius Andronicus and the ‘ Bellum Poenicum’ of Naevius, but
which was banished from the domain of poetry by Ennius, was,

 

vowel being written with ei, or with
the tall form of i (to indicate the long

1 A poetical scansion may of course
be traditional. Thus Martial (iii.g5. 1)

has havé, although Quintilian (i. 6. 21)
tells us that in the ordinary pronun-
ciation of his time the final vowel
was short. Nor is variation in the
scansion of proper names proof of
half-long quantity; it is rather to
be referred to ignorance or careless-
ness, The first syllable of Fidenae,
for example, was certainly long, the

sound) on inscriptions, and being

-usually so scanned by poets; but

Virgil (A. vi. 773) has urbemque Fide-
nam. Scansions like Italia (an imita-
tion of a Greek prosodical usage) are
mere metrical licences, and prove
nothing about actual pronunciation
(‘Italiam’... extra carmen non de-
prendas. Quint. i. 5. 18).
128 THE LATIN LANGUAGE. (Chap. II.

like the metre of the Teutonic and other I.-Eur. stocks,
accentual, not quantitative’. But the Romance languages do
not possess this distinction, We are in the habit of calling an
accented vowel, such as the second vowel of the Italian word
‘Toscana,’ long; but in reality it is pronounced with no more
length than the unaccented vowels of the same word. One of
the chief differences of such a language as Italian from Teutonic
languages is the equal length which it assigns to each vowel,
even a final unaccented vowel. Contrast, for instance, the final
e of Ital. notte with that of Germ. Gabe, or the final i of Tivoli
in the Italian and in the usual English pronunciation of the
word. Almost the only really long syllables in Italian are
syllables long ‘by position,’ e.g. the first syllables of ‘ tanto,’
‘tempo,’ which, as we shall see (ch. iii. § 4), have in fact a cireum-
flex accentuation, ‘tanto,’ ‘témpo.’ Similarly the Spanish accent
does not impair the quantity as the English accent does ; and
in French the usual quantity of every vowel is the half-long,
e.g. jeune, (See Storm on Romance Quantity in the Ponet.
Stud. 1888.) All this points to a period of ‘Vulgar Latin ’
when all vowels were equally short or half-long, and when the
only predominance of one vowel over another would be that con-
ferred by the stress of accentuation. And we detect traces of
this process of ‘levelling’ in the evident uncertainty of the gram-
marians of the fourth and fifth centuries a.p. about the quanti-
ties of words for which they have not one of the classical poets
to appeal to, and above all in the errors in scansion of those

 

The Saturnian line had three %x(,) XX, Xxx || XXXX, XX

accents (main or secondary, ch. iii. § 7)
in the first hemistich (one always
on the first syllable of the line), and
two in the second, and like Romance
poetry reckoned (with permissible
variations) a definite number of syl-
lables to the line, seven to the first
hemistich, six to the second. Its
two chief types were :

A-type—

Xx(,) Xx, XXX || XXX, XXX
dabunt mélum Metélli Naéuio poétae,
B-type (less usual) —

prim(a) incédit Céreris Prosérpina puter,
a variety of the second hemistich of
the A-type being || xxxx, xx adloct-
tus sammi,and of the B-type i! xXx, Xx
fuisse uirum. (See Amer. Journ. Phil.
vol. xiv.)

2 The passages quoted from the
grammarians by Seelmann, p. 75,
are not conclusive ; e.g. Ter. Scaurus’
distinction of facilis Sg. from facileis
Pl. is a matter of orthography, not
of pronunciation, and is suggested
by Lucilius’ proposed distinction of
the symbols ¢ and ei (see ch. i. § 9).
§ 142. | PRONUNCIATION. VOWEL-QUANTITY. 129

Christian poets who imitate the quantitative verse of the
Augustan poets. Grammarians often censure mispronunciations
due to the overmastering of quantity by accent, e.g. Céres (Mar.
Sacerd. 451. 13 K.), picés (Consent. 392. 18 K.), piper and drator
(2b. 392. 3, 11 quod vitium Afrorum speciale est) ; they frequently
caution against the confusion of eguus (with accented short
open e) and aeguxs (with accented long open e) (Pompeius,
285. 8 K., &.). Cf. the haphazard use of the apex and tall I on
late inscriptions (ch. i. § 1). For a discussion of the influence of
accentuation on the quantity of the Latin vowels (e.g. late Lat.
ddélum for eidmdov), see ch. iii. § 1; for variations like péciscor
and pdcem, (I.-Eur. pik- and pak-), see ch. iv. § 51.

$142. ‘Position.’ In Latin poetry a syllable is scanned long,
even though it have a short vowel, if the vowel precedes any
consonant-group requiring a certain period of time for pronun-
ciation. A long syllable of this kind is said to be long by
‘ position ’ (positio, e.g. Quint. ix. 4. 86; i. 5. 28); and the way
in which a Roman apprehended this length by ‘ position’ may
be seen from a passage of a fifth cent. grammarian (Pompeius,
112. 26 K.): ut puta si dicas ‘ et,’ unum semis habet. e vocalis
est brevis, unum habet tempus. t consonans est, et omnis con-
sonans dimidium habet tempus: ecce ‘et’ unum semis habet
tempus. adhuc non est nec longa nec brevis; plus tamen habet
a brevi, minus quidem habet a longa. adde ad ‘et’ s, etiam fit
longa. quare? e brevis unum tempus habet, t dimidium tempus
habet, s dimidium tempus habet: ecce duo tempora sunt, fece-
runt duo tempora longam syllabam. With a naturally long
vowel there would be really extra-length, but there is no account
taken in Roman poetry of the different length of, say, the second
syllables of calesco (with é) and modestus (with ¢), both being
treated as long syllables. Plautus, however, seems not to shorten
by the Brevis Brevians Law (ch. iii. § 42) a syllable with naturally
long vowel, scanning: guts tincédit ? but not quis tmststit ? quis
infértur ? (before s-, f- the vowel of 7 was long, § 144).

Consonant groups which admitted of more rapid pronuncia-
tion were not necessarily scanned long, viz. groups composed of
a mute and a liquid (7,7). Thus in Virgil agrum (with @) may

K
130 THE LATIN LANGUAGE. (Chap. II.

be scanned with the first syllable long or short as the poet
chooses. Plautus and the older dramatists, who follow more
the actual pronunciation of everyday life, never scan such a
syllable long, though in other than drarhatic poetry this scansion
is found; e.g. Ennius in his epic has wégrum (Ann. 187 M.), sacru-
ficare (ib. 233), &e., with first syllable long, whence we may
infer that in rapid unconventional utterance such a syllable was
short, but in measured ceremonious speech the longer dwelling
of the voice on the mute and liquid justified a long scansion.
That such a syllable differed from an ordinary short syllable is
seen in the avoidance by the dramatists of the shortening of
a vowel after a mute and a liquid by the Brevis Brevians Law ;
e.g. Plautus scans a? readily, but avoids a scansion like dgrv.
The same pronunciation of a mute with 7 seems to have pre-
vailed in Imperial times, to judge by Servius’ note on Virg. A.
i. 384 Libyae deserta peragro] ‘per’ habet accentum ; nam ‘a’
longa quidem est, sed non solida positione ; muta enim et liquida
quotiens ponuntur, metrum juvant, non accentum (cf. Quint.
1.5. 28; ix. 4. 86), but not of a mute with Zin the word maniplis,
according to the same authority (Serv. ad 4. x1. 463 in hoc sermone,
ut secunda a fine habeat accentum usus obtinuit). The establish-
ment of the pronunciation manx¢plus with long second syllable (but
short 7) has been plausibly referred to the longer form manipulus
(on the presence and absence of the parasitic or svarabhaktic vowel
in Latin between a mute and /, see § 102); but it may be
objected that Plautus scans yopdus with first syllable short, as well
as trisyllabic populus ; and makes the suffix ¢/o- (ch. v. § 25) one
syllable, with preceding vowel scanned short, in véAiclum, though
(normally) two syllables in cibictilum. In the Romance lan-
guages the accent hus been shifted to all penultimate vowels
followed by a mute with 7, e.g. Ital. allegro (with accent on
second syllable) from Lat. ddacris (Vulg. Lat. *adecro-) (see
ch. iii. § 11), which shows that in Vulgar Latin the combination of
a mute with 7 came universally to constitute length by position.
This, too, has been explained by the supposition of a parasitic
vowel, *aléc*ro-, like the occasional spelling arbéterium (in
poetry, however, arditrium is invariably scanned with short,
second syllable), but may also be referred to the practice which
§ 148.] PRONUNCIATION. VOWEL-QUANTITY. 131

we see most clearly in Italian of lengthening a mute before 7,
e.g. fabbro (Lat. fadro-), febbre (Lat. fe6ris), and Ital. occhio
from Lat. dewlus (Vulg. Lat. oclus, *occlus ?), suggests a similar
account of the Imperial Latin pronunciation of maniplus. We
see the same doubling in English ‘fodder’ (from ‘food ’), ‘ bitter’
(from ‘ bite’), ‘apple.’ Grober ascribes this consonant lengthen-
ing (Comm. Woelfl. p. 171) to what is called the ‘legato,’ as
opposed to the ‘staccato’ pronunciation ; that is to say, *fabro-
was pronounced with linking of the two syllables fa- and bro-,
not with that marked break of one syllable from another that
we see in Ital. be-ne (§ 139). The same ‘legato’ pronunciation
of the two syllables of a word like factum, omnis (but cf. § 139),
he makes the scientific explanation of the scansion of the first
syllable as long (similarly with fae tumulwm, &c.), and by the
tendency to attach an s in a group like st, sc, sp to the pre-
ceding syllable (seen in misspellings like disscente, § 130), he
explains the scansion esto, nescio, &e., with first syllable long.
On the other hand, when a word ended in a vowel and the next
began with st, sc, sp, the ‘legato’ pronunciation did not equally
assert itself, so that Lucretius allows a scansion like Ziberd sponte
(v. 79) and the like, though Virgil does not!. The Italian pro-
nunciation of festa, pescare, aspro, &c., lengthens the s, similarly
the 7 in alto, the 7 in morte, the ~ in mondo, the m in campo,
and so on.

In early Latin poetry final -s as a rule does not before an
initial consonant constitute length by position, a fact due to the
weak pronunciation of -s at that period (§ 126). Similarly initial
A-, both in early and classical poetry, has not the weight of an
ordinary consonant. But final -m always has this weight (§ 6 5).

$143. Shortening of long vowel before another vowel. In
the word pius the 7 was originally long (cf. Osc. Pifhioi ‘ Pio’
Dat. Sg.), and the scansion pia (MSS. diu, dia) has been ascribed
to Ennius in his Epic (ap. Cic. Rep. i. 41. 64:
pectora pia tenet desiderium, simul inter
sese sic memorant: O Romule, Romule die),

 

1 In A. xi. 308— ponite. spes sibi quisque, &c.,
spem siquam adscitis Aetolum a pause in the sentence intervenes
habuistis in armis between the two words.

K 2
132 THE LATIN LANGUAGE. {Chap. IT.

though Plautus in his homelier dramatic poetry recognizes only
pus. The shortening of the i is due to the difficulty found by the
Romans in maintaining the long quantity of a vowel before
another vowel. The same is true of diphthongs; we find, for
instance, the compound of prae and *hendo assuming the form
pré-hendo (and even prendo, § 58). The amount of length
assigned to a long vowel or diphthong in such a position would
differ at different periods, and even in the pronunciation of the
same period. Plautus is no doubt using the colloquial pro-
nunciation of his own day when he scans Chius (Adj.) (Poen.
699; Cure. 78); whereas wa%us was not in vogue at Quintilian’s
time (extra carmen non deprendas, sed nec in carmine vitia
ducenda sunt, Quint. i. 5. 18), and Servius (ad Virg. A. i. 451)
says that audiit, and not audiit, lentit and not lentit, was the
ordinary pronunciation, the forms with the short penult being
a usage of poetry. Here the retention of long 7 in Servius’
pronunciation may be ascribed to the presence of the forms with
», audovit, lenivit, and similarly we find in Ter. Phorm. 573
audieras. The fii of Ennius, e.g. Ann. 431 M.:

nos sumus Romani qui fuimus ante Rudini,

is sometimes used (especially at the end of a line, i.e. through
metrical necessity) by Plautus, who makes similar use of /2ers,
Sierem. (On Ennius’ aduiit Perf. &c. see ch. viii. § 50; we have
FvveIT, C.J. £. 1. 1051.)

The same shortening must have appeared in the pronunciation
of the sentence, when a word ending in a long vowel or diph-
thong preceded a word beginning with a vowel, so that the
‘prosodical hiatus’ of Latin poetry, e.g. Plaut. tii amas, Enn.
Scipid inuicte (cf. Enni imaginis), Virg. qui amant, &c, was
a native Latin usage and not an imitation of Greek versifica-
tion ?. Final long vowels would, therefore, have a short variety
or ‘doublet,’ which occurred as often as a vowel-initial followed,
and this fact, coupled with the tendency of the accent to weaken
a long final, especially in iambic words (ch. 111. § 40), explains the
early shortening of final @, e.g. ¢errd, and the later shortening of
final -6, e.g. pond in Imperial Latin (ch. iii. § 45).

 

! It seems to be the rule in Saturnian metre (Amer. Journ. Phil. xiv. 310).
§ 144. | PRONUNCIATION. VOWEL-QUANTITY. 133

On inscriptions we not unfrequently find a short i before
another vowel in the middle of a word written with the tall
form of the letter, the usual sign of 7 (ch. i. § 1), e.g. DIz, with
privsgvam (C.J, L. vi. 10239); and in the Romance languages
the 2 of dies is represented by the usual representative of Latin #,
e.g, Ital. di, Prov. dia, Fr. di, Span. dia. This at first sight
seems to be in direct opposition to the usual law of shortening
a long vowel before another vowel. But it is unlikely that
a short vowel was lengthened in this position; all that the
Romance forms and the spelling with tall J need imply is that
the i had the quality (not necessarily the quantity) of long Z, in
other words, had the close and not the open sound ($14). This
is certainly the explanation of Romance *pio, with close ¢ (Ital.
pio, &c.), and of the plvs of inscriptions, e.g. C. I. D. vi. 1058,
for we have seen reason to believe that a Jong 7 shortened in
Latin retained the quality of long 2, audit, &c., of the classical
and later period being pronounced with the close, not the open
i-sound ($14). But the z of dies must have been originally
short (ch. iv. § 63). (The examples from Romance are discussed
in K.Z. xxx. 337; additional examples of tall I in dies, pins on
inscriptions are given in Christiansen, De apicibus et I longis,
p. 32.) (Cf. the sound of Engl. ‘ the’ before a vowel.)

$144. Change in quantity of vowel before certain consonant-
groups. The quantity of a vowel which stands before a group
of consonants or a double consonant in Latin is not so easily
determined as the quantity of a vowel followed by a single con-
sonant. In a word like tnélactabilé the scansion of the word by
the Latin poets will fix the quantity of the vowel of every
syllable except the third. The w of the third syllable is long by
‘position,’ as it is called, because it stands before the consonants
ct, but we cannot tell from a line like—

venit gumma dies et ineluctabile tempus,

whether it is also long by nature or not. To ascertain the
natural quantity of these vowels which are long by position},
we can refer to two main sources of information; first, the

 

1 For a list of them. see Marx, Hiilfsbiichlein?, Berlin, 1889 (a book to be
used with caution).
134 THE LATIN LANGUAGE. (Chap. ITI.

inscriptions which denote a long a, ¢, 0, « by an apex, a mark
like the symbol of the acute accent (and from ¢. 130 B.c. by
doubling the vowel), a long z by the tall form of that letter ;
second, the Romance languages which, as we have seen, dis-
tinguish a long from a short e, 7, 0, u, when, as seems usually
to have been the case, the long and the short vowel differed in
Vulgar Latin in quality as well as in quantity. Neither of these
sources are wholly satisfactory. The apex and tall z seem to be often
used at haphazard, especially on inscriptions later than 150 A.D.,
and the latter has other uses than to express long 7, such as for
initial 2, consonantal 7 (our y), and so on?. The Romance lan-
guages, and the Teutonic and Celtic loanwords, often indicate
a quantity different from that which can be inferred for a word
in Classical Latin, a very natural thing if we consider how much
the pronunciation of a vowel is liable to be influenced by the
consonant-group next which it stands, and by the analogy of
other words of a similar form. Some help is occasionally
afforded by the statements of grammarians on the quantity of
this or that vowel, though even they sometimes show by their
hesitation that the pronunciation of such vowels was in their
time not always definitely established. Aulus Gellius (second
cent. A.D.), for example, discusses the proper quantity of e in
quiesco (vii. 15), and decides for é on the analogy of calésco,
nitésco, stupésco and other Inceptives, as well as of the noun guiés
(cf. quiésco C.J. L. vi. 25531), though he adds that a friend of his,
an educated man, invariably pronounced the word with short
e, guiésco, In another passage (ix. 6) he recommends the pro-
nunciation detito, against a common pronunciation of his time,
dctito, which was defended by the analogy of the short vowel of
the simple verb ggo. The grammarians of a later date, when the
distinction between long and short quantity was beginning to
disappear, are still more at a loss about those quantities for which
they have not the authority of the classical poets to fall back
upon. It is difficult to believe Priscian (ix. 28) (sixth cent.)
when he posits a naturally long penult for all perfects with e,

 

* See Christiansen, De apicibus et I longis inscriptionum latinarum, 1889 (a Kiel
dissertation).
§ 144.] PRONUNCIATION. VOWEL-QUANTITY. 135

e.g. 2léat, and for no others, e.g. dizi, nor yet when he makes
the a of mansi long by position only (ix. 27). Greek transcrip-
tions, too, are often dangerous guides ; for the quality of Greek
« and 7, oand w, differed, as we have seen (§ 32, § 21), from that of
Latin ¢, é, 6, 6, the Greek short vowels being, at least in the
Attic period, close and the long vowels open, while the Latin short
vowels had the open, the long the close sound. Greek ov is no
indication of the long @ of Latin, but merely of the w-sound of
Latin «, as opposed to the #-sound of Greek v. Greek «, how-
ever, almost always indicates Latin 7 (see Eckinger).

Etymology indeed will often help us. Thus we can infer
a long vowel in the first syllable of Zuctus, grief (the w is marked
long on inscriptions), from the analogy of Jigeo; though we
should never have guessed that Aésternus, unlike Aéri, had a long
vowel, if we had not been informed of the fact by a Latin
grammarian (Mar. Victorin. vi. 15. 15 K. ‘ hesternum ’” producte
dici debet: nemo enim est, qui latine modo sciat loqui, qui aliter
quam producta syllaba ‘ hesternum ’ dixerit). The metrical treat-
ment of words by the early dramatists may also be appealed to,
if it be granted that a vowel long by nature is seldom or never
shortened by the influence of a preceding short syllable, in words
like véléptdtem, or phrases like quid tgndras (see ch. iii. § 34). With
all these aids it is possible to gain a good deal of information
about the quantity of vowels long by ‘ position’ in Latin, quite
enough to prove the irrationableness of our usual method of
pronunciation which ignores all distinction of quantity in their
case1, though hardly enough to settle satisfactorily the question
with which this paragraph proposes to deal, namely the extent
to which the influence of one consonant-group tended to shorten
a vowel naturally long, of another to lengthen a vowel naturally
short. “To ascertain the limits of our knowledge and of our
ignorance on this subject it will be necessary to make a more
minute examination of the several words involved than is
generally wanted.

 

' Our ordinary propunciation of  illéx, ‘lawless.’ We pronounce both
Latin makes no distinction, for ex- ‘illéx’!
ample, between illéz, ‘alluring,’ and
136 THE LATIN LANGUAGE. [Chap. II.

In one case at least we seem to have safe ground under our
feet. Cicero (Orator, xlvili. § 159), tells us that im- and con-
lengthened their vowel when compounded with a word begin-
ning with s or 7: quid vero hoc elegantius, quod non fit natura
sed quodam instituto, ‘indoctus’ dicimus brevi prima littera,
‘insanus’ producta, ‘inhumanus’ brevi, ‘infelix’ longa.  et,. ne
multis, quibus in verbis eae primae litterae sunt, quae in
‘sapiente’ atque ‘felice,’ producte dicitur, in ceteris omnibus
breviter. itemque ‘ cOmposuit ’ ‘cdnsuevit ’ ‘ céncrepuit ’ ‘con-
fecit’: consule veritatem, reprehendet; refer ad aures, probabunt.
That this rule should be extended to all vowels before xs, (xf),
we see from such statements of grammarians as that Present
Participles in -exs, -ans had in the Nominative a long vowel
(Probus iv. 245. 13 K.; Pompeius, v. 113. 23 K.), while the
original shortness of this ¢ is indicated by the Romance languages
for the other cases (e.g. Ital. -ente with open ¢ in the penult) ;
that ¢ was long in the termination of Numeral Adverbs in -ies,
-ies (Probus iv. 247. 9. K.) (ef. O. Ind. kiyunt, &c.), and in the
Nominative Singular of déns, géns, méns, &e., (Bede vii. 230.
15 K.), while ¢ in the other cases of these nouns is indicated by
the Romance forms (e.g. Ital. dente, gente, with open e, Span.
diente, miente). Probus, however, seems to inculcate insdus,
insintis as opposed to fans, fontis (iv. 6.12 and 28. 26 K.: ef. Prise.
vii. 39). Inscriptions, too, show the apex in words like cLEMENS
(C. I. L. iit. 4550), PROCKDENS (vi. 1527 d 28), and a host of
other examples with xs (see a list of them in Christiansen, De
apicibus, &c. p. 41); while Greek inscriptions have -yvs, e. g.
Tpovinvs (Eckinger, p. 115). Finally Romance forms like Ital.
teso (with close ¢) from Latin /é(x)svs, the participle of tendo
(with open ¢) from Latin ¢éndo, not only indicate a long vowel
before zs in Latin, but also seem to show that this long é had
the same quality as the usual Latin ¢ (close ¢), and was not
a mere protraction of the open e-sound of short ¢1. Quintilian

 

1 The spellings t(h)ensawus for -no- (with long open E, § 41), but
Onoavpds, Scaptensula for Sxant? tAq are to be compared with the mis-
or SxanryovAn, Chersonensus for Xep- spelling censured in Probi App. 198.
advncos (see Georges) do not then 21 K., occansio for occasio (see § 66).
offer -ens- as the equivalent of Gk. Long open E was written ae (§ 41).
§ 144. | PRONUNCIATION. VOWEL-QUANTITY. 137

(i. 7. 29) tells us that in the word consu/es the nasal was not
sounded, a fact possibly expressed by the usual abbreviation of
the word on inscriptions, cds. The dropping of the nasal is also
indicated by spellings like novies beside noviens, and by the
Romance forms, e.g. Ital. teso (Lat. /é(n)sus), and Celtic and
Teutonic loanwords, e.g. Welsh dwys (Lat. dé(x)sus), O.H.G.
isila (Lat. a()suda, cf. Diom. i. 409. 3 K.; Serv. in Don. iv. 442.
30 K.) (cf. Irzros, C.L L. vi. 19873).

The grammarians who repeat the rule of Cicero with regard
to - and con- (Gellius, ii. 17; Probus, iv. 149. 33 K. and 253.
22; Diomedes, i. 433. 15 K.; Serv. ad den. i. 187 ; Max. Vict.
vi. 204. 16 K.; Audacis exc. vii. 354. 21 K.), often add the
remark that the rule was not strictly followed in the pronun-
ciation of their time. Thus Diomedes (i. 409. 3 K.) says of
im- and con- before s, f, ‘ plerumque producuntur’ (cf. Cledonius, v.
76.9 K.); and Servius (iz Don. iv. 442. 28 K.) intimates that
the rule was often violated in practice, plerumque enim non
observantes in barbarismos incurrimus. This probably indicates
a tendency of later Latin to give z- and con- in these compounds
the same short vowel-sound that they had in other compounds
like tucedo, céncedo, and in the simple forms %m, ciim; and this
will explain why it is, that on all but the earliest inscriptions of
the Empire, the instances of apexed vowels before xs are not so
frequent in these compounds, as in other words (see the lists
given by Christiansen), and also why, both on inscriptions and in
Romance forms, instances of a long vowel before xf are rare'.
For the combination #/ hardly occurs except in the case of verbs
beginning with f compounded with in- and con-. The word
consul was perhaps not regarded as a compound; for the o is
marked with the apex on inscriptions with great persistency.
In Welsh, too, the Old Welsh form cusil points to a Latin
co(n)silinm (cf. Diom. i. 409. 3 K.: Serv. im Don. iv. 442. 30 K.).
But, for compounds, which were realized in popular usage to be
compounds, the pronunciation of Cicero’s time must have gone

 

1 The tall IJ of Inrerr (C.J.L. inscription has the tall form, e.g. Ira,
vi. 7579), quoted by Christiansen Imprrra, It is always difficult to be
and Seelmann, is no certain indica- sure whether Ins-, Inr- on inscriptions
tion of long 4. Every initial iofthis indicate long i, or merely initial 7.
138 THE LATIN LANGUAGE. [Chap. II.

more and more out of fashion under the Empire. Its prevalence
in the time of Plautus may be inferred from the fact that
Plautus is averse to éns-, inf- being shortened by the influence
of a preceding short syllable (see § 142). In Umbro-Oscan we
see lengthening of a vowel before us, wf in Osc. keenzstur,
‘censor,’ (with long open E, § 6), Umbr. aanfehtaf’, ‘ infectas.’
The remark of Priscian (ii. 63), that the terminations -guus,
-gna, -gnum are always preceded by a long vowel, has been
extended by Marx and others into a rule that the combination
-gn- always lengthened a preceding vowel. Of Priscian’s
examples (régnum, stdgnum, benignus, malignus, abiégnus, privignus,
Paelignus), abiégnus, régnum, stagnum had probably originally
a long vowel (cf. abiés, régem, stare); privignus gets a certain
amount of confirmation from the spelling prIvIano on a soldier’s
epitaph (C. Z. Z. vi. 3541); but dentgnus, malignus receive none from
Romance forms like Ital. benigno, maligno, which are probably
‘bookwords,’ acquired by recent borrowing from Latin, not
naturally transmitted by continuous usage from Roman times. —
The Romance forms (eg. Ital. degno, Span, des-den, ‘dis-
dain’; possibly ‘ bookwords, A. L. D. viii. 324), point to Vulg.
Lat. dignus; but the word has the tall 7 in C.L.Z. vi. 6314
pIenz, and elsewhere. They point also to signum (cf. sigillum),
(Ital. segno, Span. sefia, &e.); but on inscriptions we have
sIenum (C.2. LZ. vi. 10234, a carefully written inscription of
153 4.D., and elsewhere), sIGNIFICABO (vi. 16664). The gram-
marian Diomedes (fourth cent.), speaking of the rhythmic
arrangement of some of Cicero’s clauses (i. 470. 9 K.), seems
to speak of dignitas as an anapaest, just as he calls jistam
a trochee; and if this be the right construction of his words,
it suggests that he pronounced dignitas. Welsh swyn, a charm,
Old Irish sén, blessing, sénaim, to bless, to sain, Old High Ger-
man ségan, charm against evil, blessing, are all from a late
Latin signum in its Christian sense of ‘the sign of the cross’;
and their form indicates a form segnum, with close e, a develop-
ment of an earlier signum, not signum (cf. § 14). The Romance
forms indicate, too, a short vowel in lignum, pignus, pugnus, and
show us that if the lengthening of a vowel before gn was
a tendency of Latin pronunciation at all, it was not one so
§ 144. | PRONUNCIATION. VOWEL-QUANTITY. 139

marked, and so persistent, as the lengthening before us. [For
further discussion of this question, see Bezz. Beitr. xvi. 189 sqq. ;
Mém. Soc. Ling. vi. 34 note; K. Z, xxx. 337, where it is suggested
that the change in the vowel was one of quantity merely, not of
quality, so that dignus, signum, would have the long open i-sound.
Before gu, by a phonetic law of Latin, 2 became i (ch, iv. § 8)]. The
spelling PlamEn(tum) on an African inscription (C. J. L. viii. 1344)
is not evidence enough for a lengthening of the vowel before gm;
nor is the exact relation clear between sudtegmen and subtémen,
exagmen (?) and ewdmen (Class. Rev. vol. v. p. 294: Btym. Lat.
p. £26) (see ch. iv. § 116).

In the Perfect Participle Passive and kindred formations of
verbs whose Present ends in -go, preceded by a short syllable,
e.g. /égo, dgo, we find a long vowel. Thus /éctor, léctum, dctum,
fictor (from a third-conjugation form */igere?) are attested by
Aul. Gellius (xii. 3 and ix. 6), eto by Porphyrio (ad Hor. 8.
1. 6. 122), while on inscriptions we have /éctor (C. I. L. vi. 9447,
the epitaph of a grammaticus, and so presumably correct in
spelling ; vi. 27140), adlécté, xiv. 376 (second cent. a.D.), &e., detis
(vi. 1527 d 59, B.C. 8-2), &e., infrdcta (ix. 60, c. 100 A.D.) (while
a for frango is proved by effringo, confringo (+ from @)), récte}
(xii. 2494, beginning of first cent. a.p.), ¢éctor (vi. 5205), and
the like (see Christiansen, p. 47, and cf. Ankros, mporynxro[pos| on
Gk. inserr., K. Z. xxxili. 402.). The long vowel is also found in
the Perfect (properly S.-Aorist, see ch. viii. § 39) of these verbs,
réwi, téwi (Prise. ix. 28, who adds iddéwi*), rémit (C. IL. v. 875,
105 A.D.), dé@it (x. 1793) (see Christiansen, p. 49). Whether
it is due to the consonant-groups g and #, g and s, or is a
lengthening peculiar to the Perfect and kindred forms of the verb
(see ch. viii. § 39), it is difficult to say, but the latter supposition
is certainly the more probable. The single instance on inscrip-
tions of a form that is not Verbal, viz. mda(imo) (vi. 2080, the
Acts of the Arval Brotherhood, ¢. 120 a.D.) is not sufficient

 

1 So that there wascompleteasson- alluded to by Hor. (Epp. i. 1. 59) rex
ance between réx and récté in the  eris aiunt Si recte facies.
children’s verse: 2? Plautus puns on tllectus, the Verbal
réx eris, si récte facies; si non Noun from illicio, and lectus (from lego),
facies, nén eris, a bed, Bacch. 55.
140 THE LATIN LANGUAGE. {Chap. IT.

evidence for the former. Diomedes (431.17 K.) says the ¢ of
nia is short. A similar difficulty presents itself in connexion
with the parallel formations from Verbs ending in -xgo. Gellius
(ix. 6) attests actus beside wingo; and on inscriptions we have
sejunctum (C. I. L, vi. 1527 € 38, B.c. 8-2), &e., definctis (v. 1326),
&e., conjunwit (xi. 4333, time of Antonines), extinctos (vi. 25617,
A.D. 10), ceuctus (x. 4104) (see Christiansen, pp. 44 sqq.). Here
the spelling sdnctus (cf. Osc. saahtim, Umbr. sahata), frequent
on inscriptions, seems to show that the lengthening occurs
before original xe (sacer 1), as well as before xe from original ng ;
but the few instances of non-verbal forms, viz. conjiéna (vi. 6592,
6593), the numeral guinctus, frequent on inscriptions (Christian-
sen, p. 46) (cf. gutugue, quini), are again insufficient evidence to
separate the lengthening from the ordinary lengthening of the
Latin Perfect. The absence of the nasal in xactus (beside
nanctus), fictus (but sinctus Ter. Hun. 104), plebeian defuctus
(il. 4173), and sactissimae (vi. 15511; v. 6580) (cf. Welsh saith
beside sant, § 70), is explained in ch, viil.§ 10. In gutntus it is the
guttural that disappears. All the Romance forms point to guintus
(cf. Greek Kovewros and Koewros, though the earliest forms are
Kovyxrios and Kowros: see Eckinger, pp. 122 sqq.), but declare
for the short vowel in piinctus, unctus, cinclus, tinctus, cinxi, fina.
In the absence of express testimony, such as we have for ws,
uf, it is impossible to be sure that the combination of g with
a consonant, like , #, ¢, s, whether preceded or not by a nasal,
had a lengthening effect on a preceding vowel; though there
certainly are a good many apparent indications of this. It is
equally impossible to decide whether the supposed influence
exerted on the vowel may have been a change of quality merely,
and not of quantity, just as the combination xgw had the effect —
of changing an o toa w, e.g. wnguis for *onguis (ch. iv. § 20).
145. r with consonant. Spellings on inscriptions like Fortun(a), Fortunata
(C. I. L. vi. 7527) (ef. Fotunate vi. 2236) suggest that the 0, which was certainly
originally short (Lat. dr for I.-Eur. 7, see ch. iv. § 92), has been lengthened by

the influence of the following rt. If this be so, it could only have been a local
pronunciation, or at least one that never gained a secure footing in the

 

1 A fresh complication arises from a stem sacri- (sacres, Plautus), point-
the fact that beside sdcro-, we have ing to a root sic- beside stic-.
§§ 145-147.]_ PRONUNCIATION. VOWEL-QUANTITY. 141

language ; for the Romance languages testify abundantly to short vowels in
words like porcus, cornu, certus, &e. Marius Sacerdos (vi. 451. 5 K.) quotes
pernix as a barbarism ; and Pompeius (v. 126, 5) censures the mispronun-
ciation drma, So that the initial o of drno, if long, as attested by inscriptions,
[e.g. Ornav(it) C.I.L. x. 6104 (time of Augustus), ornare xii. 4333 (time of
Antonines ; (for other instances on inscriptions, see Christiansen, p. 53], and
by Celtic loanwords (e.g. Welsh addurn, ‘ ornament,’ addurno, ‘to ornament,’
Lat. adorno) must have been originally long and cannot owe its length to the
influence of the following rn. (Similarly forma, drdo, orca.) But it is more
likely that the o was not really long, but merely had the quality of long
Latin 6, in other words was close 0. Plautus seems to scan érn- after a short
syllable ; though the instances are so few as to leave a slight doubt (Trin. 840

‘might possibly be novo cum drnatu, Awl. 721 eo drnatus). There are not wanting
indications that r with a nasal tended to modify the quality of a vowel, e.g.
fornus and furnus, formica’ and furmica, turnus (Greek répvos), just as in Italian
to-day (though not in the Toscana) close e becomes open before + with a con-
sonant, e. g. verde, erpice (Meyer-Liibke, Ital. Gram. § 54). How far this may
explain the discrepancy between the firmus of inscriptions (Christiansen,
p- 53), and the Vulg. Lat. firmus, postulated by Romance forms like Ital.
fermo, and the Welsh loanword fferf, it is difficult to say. Arvum with short
a expressly attested by Audacis exc. p. 328. 8K., originally the neuter of
the adjective druus, e.g. Plaut. Truc. 149 non aruos hic sed pascuos ager est,
appears with long a in an inscription of Tiberius’ time (drvdli, C. I. L. vi. 913).
(For other cases of long vowel before r with consonant on inscriptions, see
Christiansen, pp. 51 sqq.) (Arma, attested by Serv. in Don. 426.11 and 36K.
Prise. Acc. 521. 15, Audacis exc. 328. 6, is proved by ivermis, for @ is not
weakened to e (ch. iii.) ; dra Pomp. 130. 7 is proved by coerceo. |

§ 146. s with consonant. The short vowel before sp, sc, st, &c., attested by
the Romance languages for words like vésper, piscis, crisia, is quite strong
enough evidence to disprove the theory that an originally short vowel was
lengthened before these combinations, and to show that the long vowel
indicated by inscriptions for pastor, priscus, tristis, jtistus, &c. (see Christiansen,
PP. 54 8qq.) must have been originally long. Diomedes (p. 431, 31; 432, 16.)
attests fenéstra, dsper; Quintilian (ix. 4. 85) agréstis, Audax (359. 15 K.) campéstris,
&e. Heésternus (Mar. Victorin. vi. 15. 15 K.) beside héri is puzzling; also the
discrepancy between cristum of C.J. L. i. 1199 and Vulg. Lat. criista indicated
by the Romance forms, e.g. Ital. crosta. Festus (86. 8 Th.) distinguishes
histra, wallowing-places, from listra, purifications.

§ 147. n with single consonant. Véndo (Ital. vendo with close e), fontem
(Probus 6. 12 K.; but Romance *font- by analogy of other nt-stems), princeps
(Ital. principe, &c.) are enough to disprove the theory that this combination
shortened a preceding long vowel. There are, however, points of difficulty.
Servius (in Don. 426. 34 K.) attests princeps ; Diomedes 433. 18 céntio (by anal.
of cém-?) (for coventio, and so originally contio, or perhaps *ciintio ; French nonce,
annoneer points to a Latin form niintius (for noventius, and so originally nintius,
Mar. Victorin. vi. 12,18 K.) ; Romance words for ‘eleven,’ like Span. once, Fr.
onze, point to a Vulg, Lat. tindecim (properly tindecim, ftom tinus and decem)
(sinciput is usually explained as *sém(i)-caput). There are some indications
that the quality of a vowel was liable to change before this combination.
Thus -ond- appears as -und- in frundes, a form ascribed to Ennius (see K. Z.
142 THE LATIN LANGUAGE. (Chap. II.

XXX. 336); unguis seems to be for *onguis (ch. iv. § 20). The remark of a late
grammarian (Anon. Bern. Suppl. 111 H.), that hirundo, arundo have i, is a
mistake based on a misunderstanding of Priscian i. p. 123. 7 H.

§ 148. l with consonant. Vulg. Lat. remdlcuwm (Ital. rimorchio, Span. remol-
que, Fr. remorque), properly remilewm (from Greek fipovAKéw) is not evidence
sufficient to justify us in supposing that this combination had the effect of
shortening a preceding long vowel. Ultra, the form attested by the Romance
languages, was in all probability also the classical form; in Varro, L. L. v. 50
read uLs, miswritten in the Archetype wis, as in v. 83, and then changed by
scribes to ouis: ultra, not dlira, is the true reading of the Claudius Tablet at
Lyons, col. 1. 1. 40 (Allmer et Dissard, Inscriptions antiques, Musée de Lyon, vol. i,
pp. 70 sqq.). The influence of this combination, however, in changing the
quality of a vowel is seen in culmen beside céliimen, rult beside volo, &c. (see
ch, iv. § 20).

The shortening of a long vowel before certain single final consonants,
-r, -t, &c. is, like the shortening and change of final vowels, dependent on
the Accent, and so is discussed in the next chapter. (On the Assimilation of
Consonants, e. g. summitto for submitto, see ch. iv. § 159.)

$149. Crasis of vowels, Synizesis, &c. Two neighbouring
vowels in the middle of a word became a Diphthong if the
second was 7 or uv, e.g. coetus for co-itus (used literally in Plaut.
Amph, 657 primo coetu uicimus), suffered Crasis if they were
suitable vowels, e.g. cdmo from co-emo (cf. Engl. ‘doff’ for
do-off, ‘don’ for do-on), while if the first was ¢ or w (or in
certain cases ¢ or 0) Synizesis was a common result, e.g. larva
from /drua (a trisyllable in Plautus). Compounds of a Preposi-
tion ending in a vowel, and a Verb, &c., beginning with a vowel
or 4, show vowel-contraction more regularly in the early drama-
tists than in the Augustan poets, e.g. coercé (a dissyllable) Pacuv.,
conestat (for cohonestat) Accius, whether it be that these con-
tracted forms are a relic of the earlier accentuation of the first
syllable of every word, cderce, and the uncontracted the result of
the shifting of the accent, cd-érce, or that the contracted belong
to the conversational language of everyday life, the uncontracted
to the artificial diction of the higher poetry. Possibly a tri-
syllabic coerce is a ‘ re-composition’ like é-néco beside older enico
(with weakening of unaccented vowel), or adcurro beside aceurro
(with assimilation of consonants) (see ch. iv. § 159). Synizesis went
hand in hand with Syncope [/drua became /arva at the same
time that /éridum became lardum (see ch. iii.)], and asserted itself
more and more under the Empire ; e.g. guetus (*quyétus) for quiétus
is a common spelling on late inscriptions (cf. Ital. cheto, Span.
§§ 149, 150.] PRONUNCIATION. CRASIS, ETC. 143

quedo, Prov. quetz). The palatalization of a consonant under
the influence of a following 7 (become y) before a vowel has
played a great part in the Romance languages, e.g. Fr. bras
from bracchium, *braccyum (see § 48). A final vowel before an
initial vowel suffered elision (see the next section), and the same
thing may have happéned to the e of ne- in neutiquam, &e.,
which is scanned with the first syllable short (or should we pro-
nounce wytitiquam, nyullus, nyusquam?), while néiter, in which
the accent by the Penultima law fell on the e-, was pronounced
as a trisyllable.

§ 150. Vowel-contraction in compounds in the early dramatists. Coerce
(dissyll.) is found in Pacuvius, Trag. 47 R. :
gradere dtque atrocem céerce confidéntiam

(ef. 1. 345); and in Plautus deartuare (Capt. 640, 672), deasciare (Mil. 884) ;
deosculari (Cas. 136, 453, 454, 467) are quadrisyllables. But dehortari (four
syllables) stands in contrast to hortart in Poen. 674—

neque vés hortari néque dehortari decet.

(Ennius, Ann. 401 has the same verb in Tmesis: de me hortatur.) The
phrase coemptionalis senex, used of old, and therefore valueless, slaves who were
bought not singly but in numbers (from coemere, to buy in a lump) is irreve-
rently applied to his master by the cunning slave in the Bacchides, 976—

nune Priamo nostro si ést quis emptor, coémptionalém senem
uendam ego,

where the word coemptionalem is scanned with five syllables, as cohonestat appears
in the form conestat in a line of Accius (Trag. 445 R.)—

pré se quisque cum corona clérum conestat caput. (MSS. conectat, constat.)

Of the compounds of hédbeo, débeo always has the contracted form in Plautus ;
cohibeo may in all cases scan as cébeo, and prohibeo as prabeo ; praebeo is sometimes
spelt praehibeo in the MSS., but the scansion may always be trisyllabic, and
must be so in Merc. 1023 ; and the same holds true of all the older poets. In
Terence we find only prendo and reprendo, not prekendo (unless possibly Andr.
353), nor reprehendo (from prae and *hendo). (For other instances, see Klotz,
Altrom. Metrik, p. 139.) In the classical literature the full forms of these verbs
are generally restored [but e. g. cdgo from *co-ago, cdgito from *co-agito (ch. viii.
§ 31), débeo, praebeo, &c J, though Derivative words often retain the shorter form,
e.g. cOpula from *co-apula, from dpere, to fasten ; praeda for *prae-heda (cf. pre-
hendo) ; praemium from *prae-emo (cf. eximius from ex-émo), &e. (see ch. v. § 4).

Coepi Perf., with its O.Lat. Present coepio, comes from an old verb dpere
(ef. dpiscor) meaning ‘to fasten’ (Paul. Fest. 14. 2. Th. comprehendere
antiqui vinculo ‘apere’ dicebant), derivatives of which are aptus, aptare, as
well as copula just mentioned. In the Perfect we should expect céépi like céegi
from cégo (co-ago), and this scansion is occasionally found, as in this hexameter
line (usually referred to Ennius’ Annals) (536 M.)—

rex ambas intra fossam retinere coepit
144 THE LATIN LANGUAGE. (Chap. II.

(ef. Lucr. iv. 619). Like dissyllabie coepi (the usual scansion both in the early
and in the classical poetry) is Terence’s trisyll. coemisse (Ad. 225).

§151. Synizesis in Late and Vulgar Latin. For a list of spellings from
late inscriptions and from MSS. like quesco (for quiesco), Febrarius (Ital. Febbrajo,
Span. Febrero, &c.), see Schuchardt, Vok. ii. pp. 444 sqq., and ef. Georges, Lea.
Wortf. s. vv. vird(i)arium, sesqu(i)alter, vac(u)efacio, ventr(i)osus, sem{i)ermis, sem(i)-
ustus, sem(i)juncia, vitr(e)arius, alv(earium. All these show suppression of %
(w), t (y) before an accented vowel, long by nature or by position. Similarly
Lat. coactus has become Ital. quatto, Prov. quait, Span. cacho, through Vulg.
Lat. *quattus (from *cwactus), Lat. codgulum, Ital. quaglio, Span. cuajo, and so
on. Forms with Synizesis occasionally appear in the Latin Poets (classical
as well as ante-classical), e.g. praemjatores, Naevius, Com. 17 R., injurjatum,
Lucilius ii. 9. M., matvisti, id. Inc. fr. ix M., genva, tenvia, arjete (with the first
syllables of these three words scanned long by position) (see Luc. Miller, De
Re Metrica, pp. 249 sqq.). (On ‘scyo’ or ‘scié,’ see Charisius, p. 16. 9 K.)

§152. Other examples of vowel-contraction. Nil from nihil (always
monosyllabic in Plautus); némo from *ne-hemo; dissyllabic deinde, proinde
(according to the grammarians these were accented on the first syllable ; see
next chapter, and cf. the Plautine scansion périnde, Stich. 520) ; cdmburo for
*co-amb-uro 5 bimus for *bi-himus, ‘of two winters’ (cf. Engl.‘ twinter,’ meaning
a two-year old beast) from hiems: cpia and O. Lat. cdpi- Adj. for *co-opia,
*co-dpis- (ef. in-dpi-) ; antehac (see § 58 for other examples of the loss of inter-
vocalic h with contraction). The loss of intervocalic w, y led to contraction
in words like ditior for divitior, dinus a form of divinus, std for *stayo (Umbr. stahu,
ch. viii. § 2) (but ¢ and o do not coalesce in méneo, pleo, &c., nor a and accented
@ in dhénus) ; on these see ch. iv. §§66and 7o. So did the loss of intervocalic
m in cogo for *cdmdgo, &., (unless co was a by-form of com, ch. ix. §22). The
tendency to contraction of vowels appears at all stages of the history of Latin,
and asserted itself in colloquial Latin even more than in the literary language.
(For a fuller list of examples see Stolz in Miiller’s Handbuch, ii, p. 275). (On
the merging of i in a following i, u in a following w, see § 48.)

§ 158. Elision. Elision of a final vowel, or vowel preceding
final -m, before the initial vowel (or 2 with vowel) of a follow-
ing word is a feature of Latin poetry. That it was also
practised in speaking. we see from passages like Cicero, Orator,
xliv. §150; xlv. § 152; Quint. ix. 4. 33; x1. 3. 33-34; Seneca,
Epp. 40 (cf. Cicero’s story of Crassus mistaking Cawneas (sc. ficus
vendo) for cave ne eas; Div. ii. 40]. Marius Sacerdos (448.6 K.)
says that in reading a line like Virgil’s.monstrum horrendum, &c.,
the final -wm of monstrum was entirely suppressed, but this does
not quite agree with the statement of Probus (ap. Gell. xiii. 21. 6)
that ¢urrim had a more melodious sound than turrem in the line
lurrim in praecipits stantem, &e. In the Saturnian Poetry a final
syllable ending in -m seems to have been not elided but left in
§§ 151-154. | PRONUNCIATION. PARASITIC VOWELS. 145

prosodical hiatus (see § 65), like the -wm of cireum in the com-
pounds circti(m)ago, circii(m)eo, or of sublatum, &e., in the forms
sublatuiri, &e. (ch. viii. § 87), and the same treatment is found occa-
sionally in the older poets, e.g. Ennius (dun, 354 M.) (quoted
by Prisician i. p. 30 H.) millia militttm octo, and even in the
Augustan poets with monosyllables, e.g. nttm abest Hor. ;
a final long vowel was also shortened, not elided, like any long
vowel before another vowel in the middle of a Latin word,
eg. pré-hendo, illius, and so in the older poets frequently,
occasionally in Augustan poetry, e.g. qui amant, Virg., Esqui-
linaé alites, Hor. (see § 143). This must be a native usage, and not
an imitation of Greek poetry. Cicero’s remarks on this subject
may be quoted (Orat. xlv. 152): nobis, ne si cupiamus quidem
distrahere voces conceditur: indicant orationes illae ipsae horri-
dulae Catonis, indicant omnes poetae praeter eos, qui, ut versum
facerent, saepe hiabant, ut Naevius :

uos, qui accolitis Histrum fluuium atque algidam,

et ibidem :

quam numquam uobis Grai atque barbari ;
at Ennius semel :

Scipio inuicte ;
et quidem nos:

hoe motu radiantis Etesiae in uada ponti :

hoc idem nostri saepius non tulissent, quod Graeci laudare etiam
solent. How far they indicate a change in actual pronunciation
or in the mere technique of verse-making is doubtful (cf. ch. iii.
§ 41).

§ 154. Parasitic vowels (cf. §§ 72, 102, and ch. iii. $13). When
two adjoining consonants are not easily pronounced together a
vowel is often inserted to facilitate pronunciation. This is called
Anaptyxis. The inserted or ‘ parasitic ’ vowel (sometimes styled
in the terminology of the Sanscrit grammarians ‘ svarabhaktic ’
vowel, from Sanse. svara-bhakti-‘ partial vowel’) is often seen
in the older Latin loanwords from Greek, when the Greek word
contained a combination of consonants which was not easily
pronounced by Roman lips. A Roman did not begin a word

L
146 THE LATIN LANGUAGE. (Chap. II.

with the letters mn-, as the Greeks often did, so the Greek pva
took in Latin the form mina, just as in French the Low German
knif became canif; similarly we find in Plautus ¢échina (e.g. Poen.
517), drachtima and the like. Marius Victorinus (8. 6 K.) says that
the un-Latin character of the combination em produced the forms
Aletimeo, Alctiména (so on an old Praenestine mirror, C. J. L.-xiv.
4102), 7éctimessa, and adds that the tragedian Julius Caesar
Vopiscus (an older contemporary of Cicero) was the first to
conform the third word to the Greek Téxunooa, writing the title
of his tragedy Zeemessa, and ordering the actors to pronounce the
name in this way on the stage (cf. Prise. i. 29. 5 H.). (For
a list of Greek words so treated, see Ritschl, Opuse. ii. pp. 469-
523.) The same thing is found in native Latin words. The
suffix ¢do- (ch. v. § 25), for example, which indicates the instrument
with which an action is performed, or the place of its perform-
ance, is in Lat. -cudo-, as well as -clo-, e.g. véhiciilum, ‘that by
which one is carried’ (in Plautus always véhiclum), ciibtedlum,
‘the place where one lies down.’ In Plautus the -clo- form is
the more usual, especially after a long vowel; e.g. périciilum is
a quadrisyllable only at the end of a line (i.e. through metrical
necessity) in his plays, so that in his time the parasitic vowel
between ¢ and 7 had not quite asserted its claim to rank as
a separate syllable (see § 102), though between consonants of
less affinity, e.g. 2 and Z in the ending -d¢é/uwm, it is normal; he
uses both pdpulus and poplus (the latter only at the end of a line)
(cf. pilumnoe poploe, quoted from the Carmen Saliare by Festus,
244. 24 Th., a phrase for the javelin-bearing Romans). On the
oldest inscriptions we have poplo- (e.g. poplus, C.D. ii. 5041, of
189 B.c.; poplom on the (restored) Columna Rostrata, pro poplo
Ariminesi, Not. Scav. 1887, p. 120); piaclum (C.I. L. xi. 4766)
may be like cedve ‘caedere’ due to Umbrian influence; but
pocolom on the early Praenestine vases (C. I. L. i. 43 sqq.), tabola
(i. 197, 198), &c. (Marius Victorinus, if we can trust the read-
ing, quotes from the ‘libri antiqui foederum et regum’ piacolom
with populot Romanoi.) At a later time the spelling -cwé-
established itself so firmly in the language that it became
impossible to discriminate an original co-lo, eg. cor-cu-lum
(formed with the diminutive suffixes 4o- and Jo-, ch. v.), por-
§ 154. | PRONUNCIATION. PARASITIC VOWELS. 147

culus for *porco-lo-, from an original -clo-; and still later the
wave of Syncope which swept over the language reduced all these
forms to the same type, porclus, stablum, cubiclum, &e. (see
ch. lil. § 13).

Anaptyxis played a great part in the Oscan language, and its
kindred dialect, the Pelignian. We have in Oscan aragetud for
Lat. argento (Abl.), Helevis for Lat. Helvius, teremennid for
‘*terminia’ (Lat. termini), with liquid preceding ; and with
liquid following, paterei (Lat. patri), Sadiriis (Lat. Satrius), in
Pelignian sacaracirix (Lat. siécratrices), pristafalacirix (Lat.
praestabildtrices), and so on. The inserted vowel takes the
quality of the vowel in the syllable containing the liquid, e.g.
aragetud for *ar-getud, paterei for *pa-trei (so Lat. stdébiilum for
*sta-blum, stébilis for *sta-blis). The long preceding syllable is
the reason of its absence in Osc. maatreis (Lat. mdtris) (cf.
O. H. G. hlitres beside fogales and Plautus’ preference of periclum,
&e.). It is not found in the initial syllable, so that Zerebonio
(C. I. L. i. 190), if a dialectal form, does not belong to the Oscan
dialect (cf. Terebuni, Eph. Epigr.i. 116; Ital. calabrone from Lat.
crabro, a ‘hornet’ 1), Accentuation often seems to influence its
presence or absence in Latin; for example, Plautus has usually
miiniiplaris (once at least manipularis), as in classical Latin
disctpulus stands beside disciplina. But the cross-working of
Anaptyxis and Syncope, and the difficulty of ascertaining in
which words a suffix has been directly added to a stem ending
in a consonant, and in which there was originally an intervening
vowel (ch. v. § 21) has hitherto prevented the drawing up of exact
rules for its use in the language. [A full list of Latin examples
will be found in Stolz, Lat. Gram. (in Miller's Hands. Klass.
Alterthumswissenchaft) p. 277 (and ed.) (see also above, § 102)].

 

1 Varro’s derivation of the name  p. 184, Wilm.), suggests a pronuncia-
Gracchus, ‘a gerendo’ (Lib. Gramm. tion like G*racchus.
CHAPTER III.

ACCENTUATION 1.

$1. Nature of the Latin Accent. Was the Latin accent one
of pitch or stress? Did the accented syllable in a Latin word
differ from the other syllables in being uttered at a higher note
than they were, or with a greater force? The two things are
obviously quite distinct. For a syllable to be sounded at a high
or low note is one thing, with energy or with gentleness is
another, just as a musical note may be sounded strongly or
gently (forte or piano), a thing quite different from its being
a note high or low on the musical scale. Most languages do
indeed combine in a greater or less degree pitch-accent with stress-
accent. The accented syllable, if pronounced with more energy
than the unaccented, is generally at the same time pronounced at
a slightly higher (or lower) pitch. But, for all that, it is
usually possible to say decidedly of one language: this language
has a stress-accent ; of another : this language has a pitch-accent.
Our own language for example is clearly a language of stress-
accent. It distinguishes its accented syllables by giving them
greater energy of articulation than the unaccented ; and it shows
the usual characteristics of a language with stress-accentuation,
namely, a slurring or Syncope of short syllables immediately
following the accented syllable (e. g. ‘ méd(i)cine’ ; cf. dém(o)sel,’
‘fént(a)sy ’ and ‘fancy,’) and an obscuring or reduction of un-
accented vowels (e. g. ‘father,’ where the ¢ has the sound of the

 

1 Seelmann, Aussprache des Latein,  centuation Latine, Paris, 1855, of the
Heilbronn, 1885, is the chief expo-  pitch-theory. For sentence-accentua-
nent of the stress-theory; Weil et tion, see the Class. Rev. v. pp. 373,
Benloew, Théorie générale de VAc- 402.
ACCENTUATION OF WORD. 149

‘obscure’ vowel of the word ‘ but’ ; ‘savage,’ ‘minute,’ ‘ orange ’).
In a long word, say the adjective ‘ characteristical,’ we might
number each syllable according to the amount of force with
which it is uttered, the strongest (with the main accent) being
the fourth syllable of the word, the next strongest (with the
secondary accent) the first. The weakest syllables are, as is
usually the case with stress-accentuation, those following im-
mediately on the most strongly accented, thus ‘ chirkctéristicdl.’
In the Romance languages the accent is, like ours, an accent of
stress, but this stress is much weaker than ours, corresponding
to our secondary stress rather than to our main accent. This is
notably the case in French, where the stress is weaker than it is,
for example, in Italian. But the Romance languages show the
same tendency to syncope of short unaccented syllables, and to the
reduction of unaccented vowels, as our language does, though in
a much less marked degree (e.g. Italian gridare from Latin
quiritare, balsimo from Latin dalsémum). Pitch-accentuation is
seen in English more in the accentuation of the sentence than of
single words. A question like ‘Are you ready?’ differs by its
rising tone from a statement of fact like ‘He is ready.” In
some languages however, such as Swedish, Lithuanian, Servian,
these tone-distinctions are cleary marked in single words, a word
of the same spelling as another being often distinguished from
it by the tone alone. In English we have no example of this,
unless it be such a word as ‘rather,’ which by a difference of
tone can imply two different meanings, in answer to a question
like ‘Is it raining?’ If we are asked, ‘Is it raining?’ and
reply ‘Rather, the word, if we give it one tone, will imply
‘slightly,’ ‘not much,’ with another tone will convey the notion
of ‘heavily, ‘violently.’ But in the main the distinctions of
tone are unknown in our language ; and it is this that makes it
difficult for us to understand the nature of a language which
uses entirely or predominantly a pitch-accentuation, such as in
ancient times the Greek language, and of living languages,
Chinese. In the case of a dead language, we have two means
of ascertaining whether its accent was one of pitch or of stress.
We have the phenomena of the language itself on the one hand,
and we have the statements of native grammarians, if they are
150 THE LATIN LANGUAGE. (Chap. III.

trustworthy, on the other. Both these means of evidence point
to the pitch-character of the ancient Greek accent. The words
of the language do not show that Syncope and Reduction of
unaccented vowels, which we have seen to be characteristic
effects of a stress-accent. The Greek grammarians’ accounts of
the accent of their own language point in the same direction.
Modern Greek has, however, a stress-accentuation, so that the
Greek accent must have changed its nature in the course of time,
though at what precise period the change took place it is difficult
to say. No doubt the nature of the accent differed more or less
in different parts of Greece; and the accent in one dialect may
have allowed stress to predominate over tone at an earlier period
than in another (in the N. Greek dialects, for example, as in the
N. Greek dialects of modern times ; see Hatzidakis, K. Z. xxx.
388). The accent is taken into account in Greek metre in the
verse of Babrius, a contemporary probably of Augustus, and
author of a verse translation of Aesop’s fables.

When we turn our attention to Latin, we are confronted with
the difficulty that, while the Latin grammarians often speak of
their accent in terms properly applicable only to a pitch-accent,
all the features of their language point to its having been
a stress-accent. The reduction of the unaccented vowel (e.g.
dbigo, &c., but Greek andyw, &e.), the Syncope of syllables
following the accent (e.g. oljwrgo from oljirigo, caldus from
ciiltdus), all indicate unmistakably the presence of a stress-
accent. And the difference of its accentuation from Greek,
though not a single grammarian definitely informs us of this
difference, comes out clearly in the treatment of Greek loan-
words, especially in the language of the less educated Romans.
Greek Yoqia (with short 1) became So/ta, a stress-accent replacing
the pitch-accent with the result of lengthening the accented
vowel; Greek eidwAov became édélum. Instances like these show
that the Romans had much the same difficulty as we have, in
pronouncing Greek words with a short accented paenultima, or
with an accented antepaenultima and long penult. The difficulty
would not be so great for a Roman as for us, if his stress-accent,
like that of his modern descendant, the Italian, was not so strong
as ours; nor would it be so much felt at an earlier period, when
§ 1.] ACCENTUATION OF WORD. 151

the distinctions of quantity were more vividly marked (see ch. il.
§ 141) than in the later Empire. The Hungarian language, -
where the sense of quantity is equally vivid, accentuates the first
syllable of every word without detracting from the quantity of
vowels in the following syllables. No doubt too the nature of
the stress-accent would differ in various parts of Italy in ancient
times, as it does to-day (see Meyer-Liibke, Ital. Gram. § 122,
p. 71). In Praeneste, if we are to believe such indications as the
spellmg meornnra for the name Magolnia on inscriptions, and
perhaps the form conea for ciconia, ‘a stork, Syncope was carried
to greater lengths than in Latin, and the stress of the accent
must have been stronger (see § 14 below). But that the Latin
language of all periods, at which we have definite knowledge of
it, was a language of stress-accentuation, is proved by all the
evidence at our disposal, and disproved by nothing except the
silence of the grammarians. The same tendency to Syncope,
which before the literary period produced undecim out of *#nd-
décim, is seen working in the Early Literary time in words like
oljurigo (Plaut.), oljurgo (Plaut. and Ter.), and in the Augustan
age in calidus and caldus (the form preferred by the Emperor
Augustus, Quint. i. 6. 19), while virdis for viridis asserted itself
still later, and the same tendency, as we have seen, still shows
itself in modern Italian. And hand in hand with Syncope
goes the reduction and change of unaccented vowels. How
then are we to explain the absence of comment on the part of
the grammarians? We must, I think, take three things into
consideration. First, that the study of Accentuation, and
all the terminology used, came to the Romans from Greece.
It was Tyrannio who in the first cent. B.c. brought this new
lore to Rome, including among his earliest pupils possibly Varro
and certainly Cicero’s friend Atticus. Cicero, in a letter which
has been preserved (ad Att, xii. 6. 2), banters his friend on his
enthusiasm for so trivial a subject (te istam tam tenuem Gewpiav
tam valde admiratum esse gaudeo . . . sed quaeso quid ex ista
acuta et gravi refertur ad rédos ?), and in his own treatise, the
‘Orator,’ published at this time, makes mention of the wonderful
“law of nature’ which prescribes that the accent shall never be
further from the end of a word than the third syllable (Or. xviii.
152 THE LATIN LANGUAGE. (Chap. III.

58). The word accentus itself was nothing but the Greek word
mpotpdia in a Latin dress; and not only the terms employed,
but the description of the phenomena of accentuation are taken
directly from Greek authorities.

In the second place, the contrast between their accent and the
Greek would not be felt so markedly by Roman grammarians as
it would by us, whose accent has so much stronger a stress than
the Latin, or modern Italian, a consideration which makes it less
surprising that they did not remark on the essential difference
between the two systems of accentuation. And thirdly, the
Greek accent itself had probably at the time of these gram-
marians already entered that process of change which ended in
the stress-accentuation of modern Greek. The Greek writers
on accentuation would no doubt go on using the terminology of
the earlier phoneticians, without perceiving that their terms and
descriptions were no longer so applicable to the actual phenomena
as they had once been; and if the Greek contemporary theorists
on accent misused the terminology in this way, a Roman imitator
might be excused for carrying the misuse a little further, in
applying the same terminology to Latin accentuation. Indeed,
the writers on Latin Grammar were seldom Roman by birth;
they were usually Greeks, and would have the same difficulties
in describing the Latin accent as a Frenchman in describing the
strong stress-accent of English. These considerations may
explain how it is that only a few statements of the writers on
Latin Grammar are rid of the terms ‘high’ and ‘low’ (instead
of ‘strong’ and ‘ weak’) accent, such as the remark of a fifth-
century grammarian, that the accented syllable in a Latin word is
the syllable which would be heard at a distance, when the others
were inaudible (quoted in § 2). We may then believe the Latin
accent to have been in the main an accent of stress, like that of
modern Italian, though like it (and the accent of the Romance
languages generally), the stress-accent may have been accom-
panied by a higher tone than the tone of the unstressed
syllables.

The discredit, which we have found it necessary to attach to
the language the grammarians use in describing the nature of
accentuation, makes us hesitate about accepting their distinction,
§1.] ACCENTUATION OF WORD. 153

evidently borrowed from the Greek, of three kinds of accent,—
grave, acute, and circumflex. They postulate a circumflex for
those syllables with naturally long vowels which would have
one in Greek, namely, long paenultimas followed by a short final
syllable, e.g. Rémié but Rémaé. The quantity of the final syl-
lable is the chief factor in Greek accentuation, but not in Latin,
where the quantity of the paenultima takes its place, so that one
would not expect the accent of the first syllable of Romdé to differ
from that of Roma@é. Other circumflex words are, according to
the grammarians, long monosyllables like rés, flds, més, and final
long syllables of words whose last vowel has been dropped by
Apocope or Syncope, e.g. adie from *illice, nostrds from nostratis.
In modern Italian these apocopated words have a strong acute
accent on the final syllable, e.g. bonta, citta, virti, for donitatem,
civitatem, virtutem, but in Spanish a word like amé (Latin dmavit,
Vulg. Lat. *amaut) has a tone-circumflex in the last syllable, the
voice rising first and then falling slightly. There is hardly
evidence enough to enable us to test this theory of a circumflex
in Latin, nor to show, supposing such an accent did exist,
whether it was a tone-circumflex, formed of a rising and falling
tone, like Spanish amé or our ‘Oh!’ when used sarcastically, or
a stress-‘ circumflex,’ formed by two impulses of the voice, some-
thing like our diphthongal pronunciation of a long vowel,
e.g. ‘foe, ‘two. In modern Italian a paenultima long by
position has a very long pronunciation, and we might write
a circumflex accent over the first syllable of words like tanto
(pronounce ‘tanto’), tempo (pronounce ‘témpo’), while the
different treatment of a Latin penultimate from a Latin ante-
penultimate vowel in words like popolo (Lat. pdpulus), uopo
(Lat. dpus), suggest the possibility of a Latin Roma, &e. as
distinct from Rémulus, &c.

On the whole, then, we may say that a circumflex accent may
have existed in Latin in words like /lés, id/éc, and even in Réma
(though the grammarians’ distinction of Rdéma, Rémae is doubt-
ful), but that certainty on this matter is not to be had. The
Latin accent was an accent of stress, a stress which was not so
strong as ours, and which may have been accompanied, as in
Romance, by a high tone.
154 THE LATIN LANGUAGE. (Chap. ITI.

§ 2. Testimony of the grammarians. (1) On the Nature of the Latin
Accent. The remarks of the Latin grammarians on accentuation have been
collected by Schoell, De Accentu lingnae Latinae (in the Acta Soc. Philology. Lipsiensis,
vol. vi. 1876). They include remarks not only on accentuation, but on
proper pronunciation generally ; for accentus is often defined as ‘vitio carens vocis
artificiosa pronuntiatio’ (ib. p. 78). Here are one or two of the more important
descriptions of the Latin accent. Varro, who probably learnt the theory of
accentuation from Tyrannio (Schoell, ib. p. 6), speaks of the aititudo of a word
as opposed to its longitudo (i.e. the quantity of the syllable). By altitudo he
means its accentuation, cum pars verbi aut in grave deprimitur aut
sublimatur in acutum (ap. Serg. de Acc. p. 525. 28 K.) (ef. ib. p. 533. 4 cum
verbum enuntietur aliqua in eo syllaba necesse est summum illud vocis
fastigium possideat). His imitator, Martianus Capella (fourth and fifth
cent. A.D.) (iii. p. 65. 19 Eyss.), prettily describes accentuation as ‘ anima vocis
et seminarium musices,’ adding, quod omnis modulatio ex fastigiis vocum
gravitateque componitur ; and Nigidius, a contemporary of Varro (ap. Gell.
xili. 26. 1-3 H.), describes the accentuation Valeri by the words summo tono est
prima, deinde gradatim descendunt (cf. Audac, exc. 7. 357. 14-358. 1 K.). On
the other hand Pompeius (fifth century a.p.) (5. 126-7 K.) uses language
suitable to stress-accentuation (plus sonat), when he says, illa syllaba, quae
accentum habet, plus sonat, quasi ipsa habet majorem potestatem, and goes
on to use the illustration of the accented syllable of the word optimus being
the only syllable heard at a distance, finge tibi quasi vocem clamantis ad
Jonge aliquem positum, ut puta finge tibi aliquem illo loco contra stare et
clama ad ipsum. cum coeperis clamare, naturalis ratio exigit ut unam
sylabam plus dicas a reliquis illius verbi; et quam videris plus sonare
a ceteris, ipsa habet accentum. ‘optimus,’ quae plus sonat? illa quae prior
est. numquid hic sonat ‘ti’ et ‘mus’ quemadmodum ‘op’? Ergo necesse
est, ut illa syllaba habeat accentum, quae plus sonat a reliquis, quando
clamorem fingimus. [The same language is used by Servius (fourth cent.), in
Don, iv. 426. 10-20 K.]. Some would explain this difference of language by
supposing the Latin accent in the time of Varro to have been more of a pitch-
accent than it was in the time of Pompeius, while others try to make out
that it is those grammarians who were themselves Greeks, or who follow
implicitly Greek authorities, who speak of ‘high’ and ‘low’ accent, while
the native grammarians of a more independent turn of mind use the more
correct terms, ‘strong’ und ‘weak.’ I cannot see much ground for discri-
minating between the accent of Varro’s time and of a later age. The same
processes of syncope and vowel-reduction are at work at both periods and the
cause of these processes must have been the same stress-accentuation. But
there may well have been a change in the Greek accentuation which became
more and more apparent in each successive century.

(2) On the circumflex accent. Servius (in Don. 426. 10 K.) distinguishes
the acute accent of drma from the circumflex of Misa, acutus dicitur accentus
quotiens cursim syllabam proferimus, ut ‘drma’; cireumflexus vero, quotiens
tractim, ut ‘Musa’ (cf. Pompeius, 126. 4 K. non possumus dicere ‘Arma,’ non
possumus dicere ‘Musa’; Cledonius, p. 31. 30 K. ‘drma’ excusso sono
dicendum est, while ‘Réma’ is pronounced tractim). Similarly Priscian
(i. p. 7. 11H.) speaks of three different sounds of a, with the acute, the grave,
and the circumflex accent, as in hdmis, hamorum, hdmus, or drae, ararum, dra,
§§ 2, 3.] ACCENTUATION OF WORD. 155

and Vitruvius, in a passage borrowed apparently from Aristoxenus, says of the
words sol, lux, flos, vox, nee unde incipit nec ubi desinit (sc. vox] intelligitur, sed
quod [v. l. nec quae] ex acuta facta est gravis, ex gravi acuta. (Archit. v. 4. 2).

§ 8. Accentuation of Greek loanwords. (See the passages
quoted by Schoell, pp. 201 sqq.) An educated Roman would of
course pronounce a Greek word correctly with the same quantity
and accentuation that the Greeks themselves gave to it. The
grammarians of the Empire prescribe the Greek accentuation for
such Greek words in a Latin author as retained their Greek
form and declension. Thus in Virg. Georg. i. 59 Eliadum palmas
Epiros equarum ; Servius, in his note on the passage, says that
the word Epiros, since it has its Greek form, must be pronounced
with the accent on the first syllable, Hyiros, unlike the Latinized
form Epérus: sane ‘Epiros’ graece profertur, unde etiam ‘ Ey
habet accentum ; nam si Jatinum esset, ‘Epirus,’ ‘pi’ haberet,
quia longa est. But Greek loanwords which became naturalized
at Rome were adapted to the Latin accentuation and declension.
This was specially the case in the Republican period, according
to Quintilian (i. 5. 60), who tells us that Julius Caesar followed
the old habit of using Calypsonem, a form which Quintilian him-
self does not approve, though he accepts Castdrem, Olympus,
tyrdunus. In the plays of Plautus, who uses the language of the
educated society of his day, Greek words appear as a rule in
a Latinized form with a Latin accent, but in some words of use
among the common people the Greek accent is retained with the
effect of altering the quantity. Thus the gold coin known as
a ‘Philip, Greek @idirzos, is always Philippus with the second
syllable shortened after an accented short syllable. The shorten-
ing of the second syllable, long by position not by nature, is,
like the reduction of the vowel in Zurentum (Greek Tapavra,
Accus., now T4ranto), an indication that the Roman accent in
early times, as well as late, produced an effect on the word that
the Greek accent did not produce, an effect always traceable to
a stress-accentuation. But whether a naturally long vowel was
at this period ever shortened by the accent in a Greek loanword.
is doubtful. Ancora (Greek dyxipa), with o instead of x, is not
a certain example, and still less the hypothetical cunt/a (Greek
koviAn, not xdvida) in Plaut. Trin. 935 (Journ. Phil, xxi. 205). It
156 THE LATIN LANGUAGE. [Chap. III.

was frequently done at a later age when the long and short
quantity had more approximated to each other (see ch. 11. § 141),
e. g. butirum (Greek Bovripov), biésfémus (Greek BAdrpnpos), édd-
lum (Greek cidwAov), érémus (Greek pnjos beside épfiuos), selinum
(Greek céAzvov), our ‘ celery.’ In all these popular words which
passed into the Romance languages (Ital. ermo, sedano, &c.) the
rule seems to be that the Greek accent was always retained, even at
the cost of the quantity, except in oxytone words, which followed
rather the Latin accentuation, e.g. tap/nus (Greek tamewvds) [see
Meyer-Liibke, Gram. Rom. Sprach. i, p. 34, and cf. Anecd. Helv.
177.4 H, on ‘abyssus’ (48vocos) : paenultima positione longa sed
acuitur antepaenultima. So adijssus Paulinus of Nola (19. 651;
35. 228); Cyprian, Gall. gen. 288 P.]. This inability of the Romans
to reproduce the Greek accentuation of a final syllable is a subject
of frequent remark in the grammarians. A Greek writer of the
sixth cent. (Olympiodorus in Aristot. Meteor. p. 27) makes the
curious remark that the Roman paroxytone pronunciation of
words like Tpatkoi, &c., was due to their haughtiness (da rév
xéurrov), and had earned for them the epithet of the ‘ overween-
ing” Romans (dev brepnvopéovtes éxdAnOnoay bd trav TomTar).
I have heard a Frenchman ascribe the English mispronunciation
of words like ‘ Francais’ to the same cause.

§ 4. Romance Accentuation. The clearness with which each
part of the word is pronounced in Italian and Spanish always
strikes an English traveller. Every syllable has due effect given
to it. There is nothing like the swallowing of parts of words,
that is seen in our own and in other Teutonic languages. The
unaccented vowel in Italian notte, Spanish noche, is clearer and
more definite than in German Gabe; but on the other hand
there is less difference in quantity between a long and a short
vowel, the accented long vowel in the word ‘Toscana,’ for
example, being hardly, if at all, longer than the unaccented
vowels of the word. The accent is one of stress, but is accom-
panied by a high tone, the drop of the voice in a Spanish word.

like mano (Lat. manus), being about one-fifth, =Es—: though

mano,
sometimes in emphatic utterance the word is pronounced with
§§ 4, 5.| ACCENTUATION OF WORD. 157

a lengthening of final vowel and a slight rise of the voice on
the second, the unaccented, syllable, SE. The French accent

mano.
must have at one time been of very powerful stress, so great is
the reduction which French vowels and syllables have undergone,
but it is now much weaker than in any other Romance language,
so weak that it is usually difficult to say on which syllable the
accent rests. The difference of pitch, say between the two. syl-
lables of the word ‘jamais!’ is often very considerable, especially

in excited utterance, = or = (see Storm in Phon.
Stud. 1888). jaraaid jamais

§ 5. The Earlier Law of Accentuation. The Indo-European
accentuation, which we can generally ascertain from the Sanscrit
and Greek (e.g. Sansc. pita, Greek wari, father), has not left in
Latin the traces which it has left in the Teutonic languages.
According to a law discovered by Verner, and known as ‘ Verner’s
Law, a Teutonic spirant, developed from an Indo-European
unvoiced mute, remained unvoiced when the immediately pre-
ceding vowel bore the accent in the Indo-European, but became
voiced when that vowel did not. The Indo-European verb
*wérto, to turn or become (Sanscr. vartami, Lat. véto), is in
Goth, vairpa (our ‘worth’ in ‘ Woe worth the day!’) with the
unvoiced-spirant sound (our // in ‘thin’); while Gothic fadar
(Engl. ‘ father’) had the voiced-spirant sound of our ¢/ in ‘ then.’
By the same law Indo-European s appears in Teutonic words as
unvoiced or voiced under the same conditions, e.g. unvoiced in
Old High Germ. mis, our ‘mouse’ from Indo-Eur. *miis (Sanscr.
miu’, Gk. pis, Lat. més), where the immediately preceding vowel
had the accent, voiced in Goth. diza-, our ‘ ore’ from Indo-Eur.
*4yos (Sanser. dyas, Lat. aes), where the Indo-European accent fell
on another vowel. (On the change of voiced s to 7 in Latin, see
ch. iv. § 148). The Lithuanian accentuation, too, often enables us,
when we compare it with the Greek, to determine in what cases
the Indo-European accent was ‘circumflex, and in what cases
‘acute,’ e.g. circumflex in Gen. Sg. of A-stems (Greek rips, Lith.
rafikés), acute in Nom. Sg. of the same stems (Greek ripy, Lith.
ranka from *rankii) (see Hirt’s articles -in Indogerm. Forsch. i. &c.).

We have, however, in Latin, as Corssen proved, traces of an
158 THE LATIN LANGUAGE. (Chap. III,

older accentuation than the system which prevailed in the
classical period, which show us that at some early time the
Indo-European accent-law had been replaced by a new law,
namely, that the accent should fall on the first syllable of every
word. A change of the same kind seems to have taken place
in the Teutonic languages (see Paul’s Grundriss Germ. Phitol.
i. p. 339), and probably also in Celtic (Thurneysen in Revue
celtique, vol. vi); and in some languages of the present day,
such as Lettish, this uniform accentuation of the first syllable
prevails. The traces it has left in Latin are these :—

(1) Syncope of the second syllable of a word when that syllable
was short, e.g. wxdecim, which under the ordinary Latin accentua-
tion must have been *uuddecim, a compound of wuus and décem
(see § 13).

(2) Reduction of vowels, which. would by the accentuation of
the classical period bear the accent, e.g. infringo from in and
frango; concido from cum and caedo; triennium from tri- (tres)
and annus; which point unmistakably to an earlier *cducaido,
*trianniom, &c. (see § 18). At what precise period the change, no
doubt a gradual one, from this earlier system to the Pxnultima
Law of Cicero’s time began and completed itself, it is difficult to
ascertain. But there is some evidence that it was still incomplete
in one particular in the period of the Early Drama, for the
metrical treatment of words like factlius, mulierem (KUL), in
the plays of Plautus and Terence, indicate that the pronunciation
of such words in their time laid the accent on the first, and not
on the second syllable. A line in which the metrical ictus falls
on the second syllable occurs very rarely in their plays (Pizlologus,
li. 364 sqq.). At the same time the incidence of the metrical
ictus in all other types of words points to the prevalence of the
Paenultima Law for all words, except these quadrisyllables with
the first three syllables short. But though we cannot fix the
time when Latin words passed from the old to the new accentua-
tion, when, for example, sépientia became sapiéntia, témpestatibus
became tempestdtibus, we can guess, partly from the analogy of
other languages, partly from the inherent probabilities of the
case, what the nature of that change was. A long word like
stipientia, tempestatibus must have had at all periods a secondary
8§ 6, 7.] ACCENTUATION OF WORD. 159

as well as a main accent; it could hardly be pronounced other-
wise, as we can see from our own pronunciation of such words as
‘characteristical’ (with secondary accent on first, main accent
on fourth syllable). So that séientia would be more accurately
written sdpiéntia. The change from the old accentuation to the
new would be, in reality, nothing but a usurpation by the
secondary accent of the prominence of the main accent; sdépiéntia
would become sdpiéntia, témpestatibus would become tempestétibus.
Dimidius, unless it takes its -mi- by analogy of dimidiatus, &e.
(which is unlikely), must have been accentuated on the first
syllable about 250 3.c., for the change of unaccented @ to 7 is not
found on the oldest inscriptions ($ 22).

The Umbro-Oscan dialects seem to have passed through the
same stages as Latin. Traces of the first stage, the accentuation
of the first syllable, are e.g. Osc. Maakdiis, Vezkei, ‘ Vetusco ’
(with syncope of the second syllable), of the second stage (the
Paenultima Law), Osc. teremenniu, ‘*terminia, with doubling of
consonant before the y-sound and after the accented vowel (see
von Planta, Gramm. Osk-Umbr, Dial. i. p. 589).

§ 6. Traces of I.-Eur. accentuation in Latin. The occasional appearance
of @ for L.-Eur. @in Latin has been explained by the I.-Eur. accentuation by
Wharton (Ztyma Latina, p. 119), who thinks that é (and 6) became @ when
they preceded the syllable which bore the I.-Eur. accent, e.g. magnus from
*meg-nos, Gk. péyas). For other theories of the kind, see Bugge in Bezz. Beitr.
xiv. 60. 67. 70; Froehde, ib. xvi. 182. 191. 215 ; Stolz in Wien. Stud. viii. 149 ;
Conway, Verner’s Law in Italy, &c. (on a trace of the I.-Eur. accent in Umbro-
Osean, von Planta, i. p. 491).

§ 7. Secondary and main accent. The Saturnian verse recognizes this
secondary accent, if we are right in regarding it as accentual and not quanti-
tative verse, with three accents in the first hemistich and two in the second
(see ch. il. § 141),

e.g. dibunt mdlum Metélli | Naéuio poétae.
For a five-syllabled word always counts for two accents in Saturnian verse,
e.g. magna sapiéntia || multisque uirtutes,
e.g. dédet Tempestatibus || aide méretod,
e.g. onerdriae onustae || stébant in flustris ;
and a four-syllabled word (at any rate of the forms — = -* and ¥—  *) does the
same at the beginning of the line,
e.g. immoldbat duream || uictimam pulchram,
e.g. supérbiter contémptim || cénterit legidnes,
e.g. Cornélius Lucius || Scipio Barbatus.
A Latin secondary accent in long words such as drmatiira is indicated by the
Romance forms, which treat the vowel of the first syllable in the same way
160 THE LATIN LANGUAGE. (Chap. III.

as they treat accented a. Italian Fiorentino beside Firenze may point to the
secondary accent having been stronger in the first syllable of Lat. Flérentinus
than of Lat. Flérentia (cf. Ital. tollerare, scellerato, &c., with doubling of the
consonant which follows the vowel with secondary accent) (see Meyer-Liibke,
Gram. Rom. Sprach., i. p. 501). There are some indications (e. g. Alliteration)
that the first syllable was even in the classical period pronounced with
a certain amount of stress.

§ 8. The Paenultima Law. The law of accentuation which
prevailed in the classical and subsequent periods is that known
as the ‘Paenultima Law, a very simple one, namely, that the
accent falls on the antepenultimate syllable, if the paenultima be
short, on the paenultima itself, if long, e.g. dédres, decores,
The earliest notice of Latin accentuation, the remark of Cicero
(Or. xviii. 58) mentioned above (§ 1), speaks of it as a law of
nature that the accent should never go further back in a word
than the third syllable from the end. We have seen reason to
believe that at an earlier period this ‘law of nature ’ was broken
in the case of four-syllabled words, like factlius, beginning with
three short syllables. These were at the time of Plautus accented
on the fourth syllable from the end, facilius, &. But in all
other words the evidence to be obtained from the versification
of Plautus, and such processes of language as syncope and
reduction of unaccented vowels, points to the operation of the
Paenultima Law in the earliest literary period.

The Latin grammarians agree in pointing out the difference
between the Greek and Roman systems of accentuation, and the
greater simplicity of the Roman, which (like the Aeolic, and
unlike the Attic, &c.) never lets the accent fall on the last
syllable of a word; though they are strangely silent on the
difference, which one would have thought would have been quite
as striking, between the pitch-accent of the Greek, and the
stress-accent of their own language. They posit for Latin the
three kinds of accent used by the Greeks, the acute, the circum-
flex, and the grave, understanding by the last term rather the
absence of accent than any particular form of accent, and assign-
ing the circumflex, as we have seen, to vowels long by nature in
the penultimate syllable of words whose final syllable is short,
and in monosyllabic words. Thus the name Céthégiis, with
a naturally long paenultima and a short final, takes the cireum-
§ 8.] ACCENTUATION OF WORD. 161

flex on the paenultima, Cethéyus, and the grave accent on the
antepaenultima, Céthégus; the name Cétullus, with a natur-
ally short vowel in the paenultima, takes the acute accent instead
of the circumflex, Catdé//us, with the grave on the antepaenul-
tima, as before, Catddlus.

Monosyllables like Zua, spes, flos, sol, mons, mos, fons, lis, whose
vowel is naturally long, have the circumflex, /éx, spés, &c., while
ars, pars, pix, nia, fax, with vowel naturally short and lengthened
only by ‘position,’ take the acute accent, drs, pars, &e. Apoco-
pated words like illic, nostras retain their old circumflex accent,
illic(e), nostrd(ti)s.

A compound word, or word-group, like réspitblica, jusjitran-
dum, malésanus, intérealoct, has only one accent, respiblica, male-
sdnus, interedloci, &c., though, if resolved into two independent
words, each takes its separate accent, résque piblica, mdle sdnus,
intérea léct. Thus Argilétum, which the etymological fancy of
the Romans explained as Argi letum, ‘the death of Argus,’ is
mentioned as a unique example of a word combining all three
accents, the acute on the first, the grave on the second, and the
circumflex on the third syllable, Argidétum. (On these word-
groups, see below, § 12.)

The secondary accent, which, as we have seen ($ 7), must
have existed in longer words like arddrétum, témpérdtus, intémé-
rdtus, existimdtus, is ignored by the Roman grammarians, unless
we are so to understand the media prosodia, mentioned by Varro,
in imitation of the péon tpoowdia of Greek Accentual Theorists,
an accent which he describes as something between the grave
(i.e. entire absence of accent) and the acute accent. The
secondary accent shows traces of itself in Italian in the doubling
of the consonant in words like pellegrino (Lat. pérégrinus),
scellerato (Lat. scé/érdtus), tollerare (Lat. tolerare), &c. (see
above, § 7).

The simplicity of the Latin accentuation made it unnecessary
to indicate by written signs the accent with which a word was
to be pronounced. The mark of the Greek acute accent, a line
sloping up from left to right, and placed above the vowel of the
syllable, was used in Latin inscriptions to indicate a long vowel,
and was called the ‘apex’ (see ch. 1. § 1). It was employed

M
162 THE LATIN LANGUAGE. (Chap. III.

especially where two words of similar spelling differed in quantity
alone, e.g. médélus and mddus (Quint. i. 7. 2), and has remained
with a quantitative use in several alphabets derived from the
Latin, e.g. the Irish alphabet.

§ 9. Testimony of the grammarians. uintilian’s remarks on Latin pro-
nunciation are generally so free from Greek bias, that his account of the
Paenultima Law is worth quoting in full (i. 5. 29-31) : difficilior apud Graecos
observatio est [sc. legis sermonis], quia plura illis loquendi genera, quas
duadéxrous vocant et quod alias vitiosum, interim alias rectum est ; apud nos
vero brevissima ratio. namque in omni voce acuta intra numerum trium
syllabarum continetur, sive eae sunt in verbo solae, sive ultimae, et in iis aut
proxima extremae, aut ab ea tertia. Trium porro de quibus loquor, media
longa aut acuta aut flexa erit, eodem loco brevis utique gravem habebit sonum
ideoque positam ante se, id est ab ultima tertiam, acuet. Est autem in omni
voce utique acuta, sed nunquam plus una, nec unquam ultima, ideoque in
disyllabis prior ; praeterea nunquam in eadem flexa et acuta: itaque neutra
cludet vocem Latinam. Ea vero quae sunt syllabae unius, erunt acuta aut
flexa, ne sit aliqua vox sine acuta. In another passage (xii. 10. 33) he praises
the variety of the Greek accent in contrast to the monotonous accent of the
Romans, who never give the acute or the circumflex to the last syllable of
a word, so that every word has its last syllable, sometimes both its last
syllable and its paenultima, grave (i. e. unaccented) ; and adds that Roman
poets like to give a charm to their lines by the use of Greek names pronounced
with the Greek accent. (For other similar statements of the Paenultima
Law, see the passages quoted by Schoell pp. 100 sqq.: e.g. Diomedes 431. 6
K. ; Donatus 371.2 K.; Servius in Don. 426. 15 K. &c.) Examples are sil,
dés, Catullus, Céthégus (Sergius, De Acc. p. 483. 11 K.) indoctissimus (Pomp. 127.
15 K.), Caélius, Sallistius, Curidtius, caélum, Cicero, Galénus, Galéni, Camilli (Mart.
Cap. iii. p. 65. 22 Eyss.), Romdnus, Hispdnus (‘ Priscian,’ De Acc. p. 520. 17 K.),
ab, mél, fel, drs, pdrs, pix, nix, fds, Via, spés, flés, sil, mons, més, fons, lis (Diom.
431. 15 K.), mix, rés (Don. 371. 8 K.), néc, now (Serv. in Don. 426. 27 K.), rés
(Serg. De Acc. 524. at K.), aés, ét, qué (Pomp. 128. 15 K.), déus, citus, ddtur, drat,
pontus, cohors, lima, Réma (Diom. 431. 18 K.), hora, léges, sdlus, homo (Dositheus
378. 1 K.), méta, Créta, népos, bonus, mdlus (Don. 371. 11 K.), marinus, Crispinus,
amicus, Sabinus, Quirinus, lectica, Metéllus, Marcéllus, ldtebrae, ténebrae, Fidénae,
Athénae, Thébae, Ciimae, tabdllae, fenéstrae, Sérgius, Mdllius, dscia, fiscina, Jrilius,
Claudius, Romdni, legdti, praetéres, praedénes (Diom. 431. 23 K.), &e.

Of compound words and word-groups with one accent (like our ‘sé6n-in-
law,’ ‘man-of-war,’ ‘ pockethindkerchief,’) we have examples such as male-
sdmus, interedloct (Don. 371. 22 K.; Diom. 433. 30 K.; Pomp. 130. 18 K. ;
Cledonius 33. 12 K.). Argiletum (Prise. ii. p. 113. 10 H. ; but with all three
accents, Mart. Cap. iii. p. 68. 15 Eyss.), propediem (Don. ad Ter. Ad. v. 5. 7 (888) ],
respublica, jusjurandum (Prise. i. p.177. 10 H.; i. p. 180. 12 H.), jurisperitus,
legislator, praefectusurbis and praefectusurbi, tribunusplebis, tribunusplebi, mentecaptus,
orbisterrae, orbisterrarum, paterfamilias, paterfamiliarum, armipotens, armorumpotens,
magistermilitum, asecretis, acalculis, aresponsis, abactis (Prise. i. p. 183. 5 H.), istius-
modi, hujusmodi, cujusmodi (Prise. i. p. 440. 2 H.). Of hujuscémodi, &e., Priscian
(i. p. 205. 16 H.) says that some regard them as two separate words, but the
accent, resting as it does on the last syllable of the pronoun, shows that they

 
§§ 9, 10.] ACCENTUATION OF WORD. 163

are compounds. He distinguishes the separate accentuation of decimus et
septimus, &e. from the single accent of septimus-decimus, &e. (de Fig. Num.
XX. p. 413. 11 K.).

§ 10. Exceptions to the Paenultima Law. The rule of Latin Accentuation,
that final syllables are always unaccented, is, according to the grammarians,
violated, or apparently violated, by certain classes of words. They are words
which have dropped or contracted their last syllable, so that the accent,
which in the uncurtailed form fell on the paenultima, remains in the curtailed
form on the same syllable, which has now become the ultima. Under this
category come :—

(1) Nouns, or rather Adjectives, in -as, Gen. -atis, indicating the country
of one’s birth, e. g. cujds, nostras, Arpinds, with primds, optimds (Caper ap. Prise.
i, p. 128. 23 H.). These words, which in Early Latin (e. g. Plautus) have the
full form cujatis, nostratis, when ata later time they became contracted, retained
their old accentuation ; and so nostrds, ‘a countryman of ours,’ was dis-
tinguished by its accent from nostras, Acc. Pl. Fem. of the Possessive Pronoun
(Priscian i. p. 454. 11 K.).

(2) Some Verbal Forms ; addic, addic, &e., fumat (for Siimévit), audit (for
audivit), and the like (Servius ad Aen. iii. 3). Another remark of Servius (ad
Aen, i. 451), and other grammarians, throws some light on the last example.
They tell us that audit and not audtit, lentit and not lenvit, was the ordinary
pronunciation, the forms with the short penult being an artificial usage of
poetry, much as in English the word ‘ wind’ is allowed a different pronuncia-
tion in poetry from its ordinary one, From -vit to -it is so short a step that it
is difficult to justify a disbelief of the grammarians’ statements about audit.
The third Sing. Perf. Act. in Romance languages (e.g. Span. amé, Latin dmavit,
Ital. dormi, Latin dormivit) points to Vulgar Latin forms in accented -aut, -i
(ef. -aut on graffiti of Pompeii, C. I. L. iv. 1391, 2048).

(3) Words ending in -c (the Enclitic -c?), whose last syllable is long by
nature or by position, e.g. adhiic, posthdc, antehdc, istic, illic, istitc, illéic, istine,
illine, istéc, illdc, istéc, illée (Caper ap. Prisc. i. p. 130. 2 H.). Vulgar Latin
accentuation of the final vowel of illic, iliac, &c. is indicated by the Romance
adverbs, e. g. Ital. li, 14, Span. alli, alld.

(4) Words ending in -n (the Enclitic -né), whose last syllable is long by
nature or by position, e. g. tantén, Pyrrhin (Servius ad Aen. x. 668, &c.). This
rule cannot however have been absolute, for forms like vidén ut in Old Latin
Poetry, and even in Augustan poets (Virg. Aen. vi. 779; Tib. ii. 1. 25) show
that when the final vowel of the particle was elided, the verb might retain
the ordinary accentuation, viden, like vides. Servius (fourth cent.) tell us that
vidén was the usage of his time (ad Aen. vi. 779 viden ut geminae stant vertice
cristae] ‘den’ naturaliter longa est, brevem eam posuit, secutus Ennium : et
adeo ejus est inmutata natura, ut jam ubique brevis.inveniatur), and Plautus
seems, when -qut, -né is elided, to let the metrical ictus fall normally on the
syllable which would have the accent in the absence of the particle, e. g.
prospéerequ(e), surruplasqu(e) (Amer. Journ. Phil. xiv. 313).

An accent originally on the antepaenultima remains in the curtailed form
on the paenultima, according to the grammarians, in contracted vocatives and
genitives of 10-stems, e.g. Vergili, Valéri, tugurt (Serv. ad Aen. i. 451; Prise.
i. p. gor. 21 H.). Gellius (second cent. 4.p.) tells us that Nigidius Figulus
(first cent. B. c.) wished to distinguish Valeri Voc. from Valéri Gen., but adds

M 2
164 THE LATIN LANGUAGE. (Chap. IIT.

that in his own time such an accentuation as Vdleri Voc. would sound very
strange : siquis nunc Valerium appellans in casu vocandi secundum id prae-
ceptum Nigidii acuerit primum, non aberit quin rideatur (N. A. xiii. 26),

Lastly, Interjections are excluded by the grammarians from the ordinary
rule. They are said to have no ‘certi accentus,’ whatever that may mean ;
and the statement of a late grammarian (Audacis exc. 361. 11 K, that papaé
and attdt (also ehem, MSS. hoehem?) were accented on the last syllable, is
confirmed, in the case of at least the former, by the incidence of the metrical
ictus in Plautus (always papaé, never pdpae). The Greek edye appears in
Plautus’ dramas with the last syllable lengthened! (cf. the MSS. spelling
eugae), and the phrase eugae-eugae always has the metrical ictus eugae-eugae.

Greek words, as was mentioned before (§ 3), when they were used by
a Latin author with their Greek form and declension, retained also their
Greek accent, e. g. Eptros in Virg. Georg. i. 59 (Serv. ad loc.).

§ 11. Vulgar-Latin Accentuation. The Latin accentuation is retained
with wonderful tenacity by the Romance languages. Where they agree in
deviating from the classical Latin accent, the accentuation which they repro-
duce is that of Vulgar Latin. There are four important cases of deviation :—

(1) First of all, in words ending in -idrem, -iéhun, e. g. miilitrem, filislum, the
accent in Vulgar Latin was shifted from the ¢ to the e and 0, mutiérem, filidlum.
The precept of an unknown grammarian (Anecd. Helv. p. ciii. K.) sanctions
this usage (mulierem in antepenultimo nemo debet acuere, sed in penultimo
potius), and in Christian poets of the third and fourth centuries we find
scansions like insuper et Salomon, eadem muliére creatus, Drac. Satisf. 161 ;
ef. Ital. figliuolo, Span. hijuelo, Fr. filleul. Nouns in -té, Gen. -ittis followed
a somewhat different course. Their Nominative became -és, and this form
was extended to the other cases, e.g. paRETES (C.J. LZ. vi. 3714), Ace. Sg.
*paretem is attested by the Latin loanword in Welsh, parwyd, and by the
Romance forms, e.g. Ital. parete (with close e), Span. paréd, &c., while
*muliérem is the original of Ital. mogliére (with open e in the penult).

(2) Again the occurrence of a mute with the liquid r at the beginning of
the last syllable seems to have attracted the accent to the penult. Thus
Vulg. Lat. *tenébrae is attested by Span. tinieblas, and other Romance
forms. We cannot be wrong in connecting this with the practice of Latin
poets of treating a short syllable before a mute with » as a long syllable,
when it suits their convenience (see ch. ii. § 142).

Servius (fourth cent.) (ad Aen. i. 384) seems to say that the accent was not
in his time attracted to the penult in correct pronunciation ; for he remarks
with regard to peragro in this line of Virgil ; ‘ per’ habet accentum... muta
enim et liquida quotiens ponuntur metrum juvant, non accentum (ef. Diom.
431. 28 K.).

(3) In Compound Verbs the accent seems to have shifted to the stem-vowel
of the verb in Vulgar or Late Latin, e. g. recipit is indicated by Ital. riceve, Fr.
regoit ; renégat by Ital. riniega, O. Fr. renie. With this we may connect the
tendency in the spelling of post-classical inscriptions, and of our earliest MSS.
to restore the vowels in compound verbs to their undecayed form, e. g. con-
sacro, compremo (§ 18).

 

1 Like our ‘ bravo!’ ‘hillo!’ often pronounced with the voice dwelling on
the final vowel.
§§ 11, 12.] ACCENTUATION OF SENTENCE. 165

(4) Lastly, the Romance forms of the Numerals give indications that the
Vulgar Latin accentuation was viginti, quadrdginta, &e. (see Meyer-Liibke,
Gram. Rom. Sprach. i. p. 494). Triginta is one of the barbarisms ‘ quae in usu
cotidie loquentium animadvertere possumus,’ enumerated by a fifth-century (?)
grammarian (Consentius p. 392.4 K.). Ona fifth-cent. inser. (vid. A. L. L.
v. 106) we have quarranta for quadraginta (Ital. quarénta), and an epitaph in
hexameters has vinti for vigintt (Ital. venti) [Wilm. 569 (ef. C.F. L. viii. 8573) :

et menses septem diebus cum vinti duobus].

$12. Accentuation of the Sentence. Hitherto we have been
considering only the accentuation of words by themselves. But
there is also such a thing as the accentuation of the sentence;
and the ‘accent which a word would bear, if uttered separately,
may be different from the accent assigned to it when standing
with other words in a sentence. The Greek preposition zpds,
for example, had, if mentioned by itself, an acute accent. But
in the sentence its accent was obscured by the accent of the
noun which it governed, e.g. mpds méAuv, and this by the Greek
system was expressed by replacing its acute by a grave accent.
So that the Greek system of marking the accents recognized
both the word-accent and the sentence-accent.

By the Indo-European sentence-accentuation the verb in
a main sentence was treated as a subordinate word, and ap-
parently occupied the position proper to enclitic words, namely,
the second place in the sentence, while in dependent clauses it
received the accent like any other word, and stood at the end of
the clause, an arrangement which has been, curiously enough,
preserved in German to the present day (see Wackernagel in
Indog. Forsch. i. pp. 333 sqq.). Indefinite pronouns were enclitic
or subordinate words, while interrogatives were accented (cf.
Greek dup rus and ris avjp;). Other enclitics were the copula
*que (Greek re, Lat. qué), the personal pronouns (unless specially
emphasized), &c.

We can determine with a fair amount of accuracy the accen-
tuation of the Latin sentence, partly by the help of the remarks
of Latin grammarians, partly from observing the phonetic
changes of Latin words in the Romance languages, where an
accented word or syllable is not subject to the same laws of
development as an unaccented, partly from the analogy of other
languages, and to a large extent from the versification of the
166 THE LATIN LANGUAGE. (Chap. III.

early dramatists. For Plautus and Terence leave, as a rule, the
subordinate words of the sentence, the words, in fact, which we
omit in writing telegrams, in the theses of the line, where no
metrical ictus falls on them:

e.g. Plaut. Trin, 21 [rogat] ut liceat possidére hanc nomen faibulam,

where wt and hanc, words which might be omitted without
obscuring the meaning, are relegated to the theses of the first
and fourth feet. The Latin grammarians, with their usual
adherence to Greek terminology, speak of the ‘attraction’ of the
accent by enclitics or subordinate words. But this statement of
the facts is corrected by Quintilian (i. 5. 25, 26), who shows
that a subordinate relative, like gwadis in the sentence, talis est
qualis Cicero fuit, or a subordinate preposition, like czrewm in
Virg. Aen. iv. 254 quae circum litora, cireum Piscosos scopulos,
&e., is really in Latin united with the following word into
a word-group, which takes the ordinary accent of a single word:
cum dico ‘circum litora,’ tamquam unum enuntio dissimulata
distinctione, itaque tamquam in una voce una est acuta. The
Latin pronunciation would thus be qualis-Céicero, circum-litora.
Priscian (i. p. 183 H.) objects similarly to a statement of Helleniz-
ing grammarians, that guts, the indefinite pronoun, in siguis,
numguis, &e., is an enclitic like ts in efris, and prefers to call
siguis a compound or word-group with the natural accent of
a single word. With this correction, the rules of the Latin
grammarians about enclitics and subordinate words are in the
main probable enough in themselves, and are confirmed by the
evidence of the Romance languages, and the early dramatists’
versification, though some of their distinctions between the
accentuation of words of the same spelling, such as né, ‘ verily,’
né, prohibitive, ze ‘lest’; ddeo, the verb, and adéo, the adverb ;
ut, ‘how, ut, ‘in order that’; érgo, ‘ therefore,’ ergé, ‘on account
of,’ require additional evidence before we can accept them.

§12a, Latin Sentence-Enclitics. Among Latin Sentence-Enclitics we may
class :—

(1) Enclitic Particles like qué (I.-Eur. *q3é, O. Ind. ca, Gk.7e, all unaccented),
vé (I. Eur. *we, O. Ind. va (unacecented), Gk. *Fe in 9-2 from *9-Fe], and so on.
Their enclitic nature is shown by their being always joined in writing with
the preceding word, e.g. atque, sive. In the rapid utterance of ordinary con-
§ 12°.) ACCENTUATION OF SENTENCE. 167

versation these words often lost their final vowel, and so appear also in the
forms ac (for *atq), seu, &c. (see § 35).

(2) The various parts of the substantive verb. The mode in which es, est
are written in the best MSS. of Plautus, for example, amatus (amatu’s), amatust,
amatumst for amatus es, amata est, amatum est (ef. ventumst, &c. in Virgil MSS.,
Ribbeck, p. 419), shows that they were treated as mere appendages of the
past participle passive. For the unaccented nature of érat, rit, &e. we have
proof, if proof be needed, in Romance forms like Ital. era and Span era (Lat.
erat), O. Fr. ert (Lat. erit), for an accented é would have taken another form,
such as Ital, *iera, Span. *yera (cf. Ital. niega, Lat. négat). It need hardly be
said that the extent to which these words, and indeed all ‘Sentence-Enclitics,’
were suppressed, would depend on the caprice of the speaker, on the nuance
of thought, on the style of composition, &. No hard and fast rule can be
laid down about them, just as no rule could be made for the use of ‘’s’ for
‘is,’ ‘’re’ for ‘are’ in English. A sentence, for example, of Cicero, ending
with the words licitum est, is quoted by a grammarian as an instance of
a sentence ending with a monosyllable (Mar. Sacerd. 493. 14 K.).

(3) The personal and possessive pronouns, when unemphatic. In the
Romance languages two distinct series have been developed for the personal
pronouns : (a) the enclitic, e.g. Ital. mi, ti; Fr. me, te: (b) the accented, e. g.
Ital. me, te; Fr. moi, toi. Similarly a Vulgar-Latin possessive *mus, *mum,
*ma, beside the regular meus, mewm, mea, is indicated by French mon, ma, and
Vulgar-Italian ma-donna, padre-mo, &¢., which resemble O. Lat. sis for suis, &c.
(see ch. vii. § rr), in such a line as Ennius, Ann. 151 M., postquam lumina sis
oculis bonus Ancus reliquit. It is true that Priscian (ii. p. 141.15 H.) expressly
says that there is no distinction in Latin corresponding to the Greek distinc-
tion between «fdév pe and «ldev eye, ode éxetvov, apud nos autem pronomina
eadem et discretiva sunt ut ‘vidit me’ vel, ‘ vidit me, illum autem non’; but
he seems to refer rather to the identity of the written form of the emphatic
and unemphatic pronoun, than to that of their intonation in discourse. In
Plautus and Terence the ictus always falls on the preposition in phrases like
in me, dd me, intér se, unless the pronoun is emphatic (or elided), just as in
Greek we have mpés pe, mpés ce, &c., or as in English we lay the stress on ‘ for,’
‘with’ in ‘for me,’ ‘with him,’ &c. In O. Irish this tendency of pronuncia-
tion reduced the pronouns to mere suffixes, e. g. for-m, ‘on me,’ for-t, ‘on thee,”
&e. (but cf. Censorin. ap. Prise. ii. p. 51. 11 H.).

(4) The demonstrative pronouns, when unemphatic. The Romance forms
point to (i)lum pdtrem, (il)la mater, &e. as the origin of the definite article in
all the Romance languages, while ille pater seems to have been pronounced
sometimes #(le) pater, e. g. Ital. il padre, Span. el padre, Prov. el paire, some-
times (il)le pater, e. g. Fr. le pére. Similarly (4)ste, indicated by Ital. stasera,
‘this evening,’ is actually found in old MSS. (see Neue, Formenlehre 3 ii.
pp. 402 sq.; on sta in the Itala, see Georges, Lez. Wortf. s. v.), and isté is men-
tioned as a barbarism by a third-century grammarian (Mar. Sac. p. 451. 10 K.).

(5) The relative and indefinite pronouns, while the interrogative and
exclamatory were accented, e. g. tdlis est qualis Cicero fuit, but qualis fuit Cicero !
The grammarians often distinguish between the accentuation of quis, qudlis,
quantus, quit, qudtus, cujus, abi, unde, &. when used interrogatively, and when
they are merely relative or indefinite pronouns (e.g. Prise. i. p. 67. 5 H.,
interrogativum est quod cum interrogatione profertur, ut quis, qudlis, qudantus,
168 THE LATIN LANGUAGE. [Chap. III.

quot, qudtus, cum suos servant accentus. Infinitum est interrogativorum
contrarium, ut quis, qualis, quantus, quot, quotus, cum in lectione gravi accentu
pronunciantur, ef. ii. p. 127. 2 H. Partit. p. 501.14 H. So for qui interrog. and
rel. ib, ii. p. 9. 20 H.; cujus, interrog. and rel. ib. ii. p. 179. 3 H. ; quo, ubi, unde,
qua, ib. ii. p. 132. 3 H. ; ii. p. 83. 12 H. ‘qua’ quando relativum est gravatur ;
quando, Charis. p. 111. 27 K.; Prise. ii. p. 82. 24 H.; quorswm, id. ii. p. 83. 11
H. &e.). Their usual expression for the subordination of the relative and
indefinite forms of these words is that they ‘have the grave accent’ (gravi
accentu pronuntiantur, gravantur), though sometimes they use language more
applicable to Greek (e. g. motos interrog., words rel.), and say that ‘they take the
acute accent on the last syllable’; and Charisius (p. 111. 27 H.) gives an actual
Latin example of this accentuation of the last syllable in the sentence quando
tot stragis acervos Vidimus. Quintilian, as we saw, corrects this statement
of the case, and shows that the true account for the Latin language is to say
that the relative is joined with the noun or important word beside it, this
word-group taking the accent of any ordinary word. The line from Virgil
would thus be really pronounced quandd-tot stragis acervos, &c., so that the
accentuation of the final syllable of quando is due to the accident that it
stands next to a monosyllabic word. The relative would have no accent in
a sentence like quem testem te adducturum dixeras, for it would be joined
with the noun into a word-group, quem-téstem, with the accent on the first
syllable of the noun. When in a line of poetry the relative followed the
noun, as in Plautus (Amph. 919) Testém quem dudum te ddducturum dixeray,
the accentuation would presumably be by the same reasoning testém-quem; and
this presumption seems to be confirmed by the versification of the dramatists.

(6) Prepositions, while adverbs, are accented. Thus we should say sipru
habitat, but supra moénia est ; ante vénit, but ante Caésarem vénit. This distinction
is often inculcated by the grammarians ; e.g. Palaemon (first cent. a. D.) (ap.
Charis. p. 189. ro K.= Diom. p. 407. 19 K.) insists on the different accentua-
tion of infra, supra, extra, intra, ultra, citra, circa, juxta, contra, subtus, coram, ante,
post, prope, usque, super, when adverbs, and when prepositions; cf. Charis.
p. 231. 24 K.; Audax, p. 353. 22 K.; Probus, Inst. p. 149. 27 K.; Mart. Cap.
lil. p. 67. 21 Eyss.; Prise. ii. p. 28.24 H.; ii. p. 30. 25 H.; ii. p. 33.1 H.;
ii. p. 42. 7 H.; ii. p. 45. 25 H.; ii. p. 51. 11 H. &.). Priscian (ii. p. 27. 4 H.)
says that Latin prepositions, like Greek, had, by themselves the acute accent
on the last syllable (swpér, bép), but in the sentence lost this accent ; (accen-
tum habent praepositiones acutum in fine, tam apud Graecos quam apud nos,
qui tamen cum aliis legendo, in gravem convertitur) (ef. Don. p. 391. 11 K.) ;
and Quintilian, as before mentioned, says that what really happened in
Latin was that the preposition was fused with its noun into a word-group,
which was then accented like any ordinary word, e. g. cirewm-litora (accented
like circumsistite, circumlitio). Indeed the words are often written together in
old MSS. and inscriptions ; cf. Mar. Vict. 23. r2 K. and Indices to C. I. L. (so
Umbr. preveres ‘ante portas’). This suggests that in collocations like in via,
per dolum, in manus the preposition itself may have received the accent of the
word-group, unless it were desirable for some special reason to give pro-
minence to the noun. This view is supported by some word-groups,
which established themselves in Latin usage, such as obviam, sédtilo (st,
siné dolo), comminus, déniio (dé novo), admidum, afitim (but ef. Gell. vi. 7), as
well as by the versification of the dramatists, which also points to in-rem,
§ 12°.) ACCENTUATION OF SENTENCE. 169

in-spem, in-jis, &e. (cf. quamobrem.) A preposition placed after its noun
received an accent, as in Greek, according to the grammarians (e.g. Prise.
ii. p. 27. 4 H. cum praepostere ponuntur, monosyllabae acuto, disyllabae
paenultimo acuto proferuntur) ; but remained unaccented in phrases like
virtutem propter imperatoris or justitia in legum, where the preposition is followed
by a genitive dependent on the noun (Censorinus ap. Prise. ii. p. 33. 20 H.).

(7) That conjunctions, like prepositions, had a different pronunciation
according as they came dest or second in the sentence is asserted by Priscian
(ii. p. 24. 21 H.) praepositae gravantur omnibus syllabis, postpositae acuuntur
in principio. He refers expressly to igitur, quiniam, saltem, so that he would
have us pronounce, e.g. igitwr Cicero rénit, but vénit igitur Cicero. That the
monosyllabic conjunctions ¢f, séd, and the like, were enclitic words in the
sentence, may be proved, if proof be wanted, from the versification of the
early dramatists, in whose lines these conjunctions are relegated as a rule to
the theses, and do not receive the metrical ictus, and also from the Romance
languages, where the Latin monosyllabic conjunctions have suffered the same
phonetic changes as the unaccented syllables of Latin words. Et, for
example, if the word had been accented, would have become *iet, or some
similar form, instead of Ital. e, Fr. et, Span. y.

(8) Auxiliary verbs in Latin must, like those in other languages, have
been enclitic, or rather, according to the Latin practice, must have been
joined with their verb into a word-group, e.g. volo-scire, coctiim-dabo, missiim-
Sacit, cave-facias. In Plautus volo-scive is always scanned vold-sctre, never vold-
scire ; and the metrical ictus in his verses of phrases like factum-volo, facids-
volo, missdém-face, cavé-parsis, cavé-faxis supports our rule. Cicero’s story about
Crassus at his departure for Parthia mistaking the cry of a fig-seller, Cawneas !
Cauneas! (se. ficus vendo) for cive né eas (Div. ii. 40, 84) seems to show that in
ordinary talk this verbal phrase was treated as a word-complex with a single
accent cau(e)-n(e)-eas. Similarly a verbal phrase like épéram-dére, fidem-dire,
dono-ddve would probably have ordinarily only one accent, just as we throw
the stress on the noun ‘noise’ and not on the verb ‘make’ in the phrase
‘to make a noise.’ The dramatists let the metrical ictus fall on these phrases
thus : fidém-dans, operdm-dat, operdm-dabam, dond-data. The reduction of other
unemphatic verbs to mere members of a compound word is indicated by the
traditional way of writing quilibet, quamvis, quantumcis, and the like. Svs, ‘if
you please,’ from st vis (cf. suitis, Plur.), is an enclitie appendage of the
imperative, e. g. priptra-sis as much as dum in exctité-dum, aspict-dum. (Plautus
gives to these phrases the ictus properd-sis, excitedum, aspicedum, &c.)

(9) Some nouns too of subordinate meaning must have become members
of word-groups. In English ‘thing,’ ‘kind,’ ‘ state,’ ‘part’ are used in this
way without stress, in such sentences as ‘something (nothing) of that kind,’
“some parts of England.’ That médus, rés were so used in Latin we see from
the traditional spelling quiémédo? qudre? So diesin propediem (Don. ad Ter.
Ad. 888), quotidie, postridie. Gellius (x. 24) says that in the time of Cicero
and the earlier period, the phrase diequinte or diequinti was in vogue, ‘ pro
adverbio copulate dictum, secunda in eo syllaba correpta,’ and we may guess
that dies formed a compound with triginta (viginti) from the fact that these two
numerals are perhaps never found in Plautus and Terence with the ictus on
the last syllable, except when dies (or minae) follows, e.g. Men. 951

at ego te pendéntem fodiam stimulis triginta dies,
170 THE LATIN LANGUAGE. [Chap. III.

where the last two words seem to make a compound noun, like our ‘ fort-
night,’ ‘twelvemonth.’ Lécus too might be subordinated in a phrase like
intérea loci (Don. ad Ter, Eun. 255), tébi loci, &e. The versification of the early
dramatists, and the compound words in Romance suggest as similar word-
groups phrases like vaé-mihi, vaé-miserd-mihi, bene-rém-gérit, male-rém-gerit (with
metrical ictus normally on these accented syllables in Plautus), ad-illam-héram.
(Ital. allora, Fr. alors), ad-mentem-habére (Prov. amentaver, O. Fr. amentevoir),
avis-struthio (Fr. autruche, Span. avestruz), avis-térda (Ital. ottarda, Fr. outarde,
Port. abetarda), foris-fictre (O. Ital. forfare, Fr. forfaire), male-hdbitus (O. Sp.
malato, Prov. malapte, Fr.malade), &c., (ef. Engl. ‘goodbye’ for ‘God be wi’ ye’).

$18. Syncope. The syncope or suppression of an unaccented
vowel is a common feature of languages. which have a stress-
accent, and is carried to the greatest length by the language
whose stress-accent is most powerful. The Celtic languages had
a stronger stress-accent than Latin, and so we find in Old Insh
some words borrowed from Latin (which the Romans had them-
selves borrowed from the Greeks), reduced by syncope to a much
greater extent than they were in Latin, eg. felsub, Latin
philoséphus, apstal, Latin dpostdlus. And in countries under
Celtic influence, such as France or the northern parts of Italy
(e.g. in the Romagna dmeng for ddminica, Sunday), Latin words
have been curtailed much more than in other parts of the
Romance-speaking world. In ancient Italy, too, we see syncope
more developed in some districts than in others, e.g. Myolnia for
Magolnia on a Praenestine inscription (C. J. L. i. 118), though it
is not always easy to say when such forms are merely graphic,
and indicate the use not of a contracted pronunciation, but only
of a contracted system of writing. The conditions under which
vowel-syncope was carried out differed at different periods. In
Latin a vowel between ~ and m was not syncopated, because the
consonant-group 2m was difficult to pronounce, e.g. dima not
*anma. But in the Romance languages syncope has been pushed
a stage further, e.g. Prov. anma, alma, arma, Old Fr. anme,
alme, arme, Fr. Ame, Span. alma, Sicil. arma, Ital. alma (in
poetry), the unmanageable group um being often changed to
rm, just as original xm in Latin *can-men from céino, *gen-men
from géno, gigno became rm in carmen, germen (ch.iv.§ 78). Simi-
larly pertica, which resisted syncope in Latin, in Italian (pertica),
and in other languages, has succumbed in French (perche, our
‘perch’) and Provencal (perga). Frigidus appears in all the
§ 13.] ACCENTUATION. SYNCOPE. 171

Romance languages in a syncopated shape (Ital. freddo, Fr. froid,
&e.) (ch. ti. § 132). Viridis, too, was in Vulg. Lat. vérdis (Ital.
verde, Span. verde, Fr. vert), and céiidus appears early as caldus.
Analogy also may often prevent syncope, or, after words have
been syncopated, may restore them to their original form. Thus
porgo, for example, was restored to porrigo by the analogy of the
perfect porrexi; and the analogy of other adjectives in -7dus
where this termination was preceded by some uncombinable con-
sonant, e.g. frigidus, may account for the existence of unsyncopated
adjectives like cadidus ; for the consonants in calidus, 7 and d, are
of a kind that would be easily combined. A Nom, Sing. like
hortus, if syncopated to *horts, *hors, would soon be restored
to its old form by analogy of the other cases orti, horto, &e, In
the compound cofors we do indeed find this monosyllabic form ;
and one might be tempted to think that the syncope of ortus to
*hors had led to the word becoming an I-stem for an O-stem,
*hors, *hortis, from hortus, hortt. But the I-stem of cohors is
more easily accounted for by the Latin predilection for I-stem
compounds of QO-stem nouns, e.g. exanimis, unanimis from
dnimus (ch. v. § 34.). It is doubtful whether any clear case of
a change of stem through syncope of the final syllable of a Nom.
Sg. is to be found in Latin.

It appears, then, that vowels resisted syncope when they stood
between consonants which did not easily combine, and that the
analogy of unsyncopated forms might prevent or efface syncope
in whole classes of words. With these exceptions, it seems to
have been the law of Early Latin that 2, 7 in the syllable after
the accent always suffered syncope, unless they were long by
‘ position.” This 2, 7 might be original 2, 7, or the reduced (post-
tonic) form of original @, 3 ($ 18). The Early Latin accent fell,
as we have seen above (§ 5), on the first syllable of each word,
so that every @,% in a second syllable not long by position must
have suffered syncope.

The syllable -7%-, preceded by a consonant, followed laws of
its own. It appears in the posttonic syllable as é, e.g. paternus
for *patri-nus. Similarly -/7- appears in the posttonic syllable
as wl (5/) in fécultas, &e. (see ch. iv. § 13). The syllable -vé- also
stands apart from others owing to the vowel-nature of its con-
172 THE LATIN LANGUAGE. [Chap. III.

sonant, which facilitated syncope. Thus we have syncope of
a vowel preceded by v even in a syllable long by position, e.g.
aunculus (Plautus), the ordinary conversational form of dvonctlus
(cf. anculus, C. I. L. viii. 3936, ix. 998); or rather the semivocalic
w has dropped out between the two vowels. Similarly détior
for divitior, déorsum (dissyllabic deorsum in Plaut.) for dévorswm,
and the like, are different from the ordinary cases of syncope.
The tendency to syncope continued to assert itself at all
periods of the language. Words which resisted it at an earlier
period often, as we have seen, succumbed at a later. The new
law of accentuation, the Paenultima Law (§ 8), brought with it
the possibility of a new variety, namely, suppression of the
syllable preceding the accent. Pretonic syncope is a feature of
Indo-European, where indeed it was much commoner than post-
tonic syncope, but it could play no part in Latin so long as the
accent remained invariably on the first syllable. It is often
difficult to say whether a case of syncope is pretonic or post-
tonic. In words like drdére, drdérem, for example, we say
that the syncope of 7 of *artdére, *aridérem is due to the new
accent on the penult; *aridére, *ariddrem, but it might possibly
be referred to the influence of the old accent on the first syllable,
*évidere, *éridorem. Words like arténa (Greek dpérawa), perstroma
(Greek repiorpwpa) Lucil. (i. 41 M. and Lowe, Prodr. p. 347),
both borrowed no doubt after the old accent law had ceased to
operate, are clearer cases of pretonic syncope. So are enclitic or
subordinate words which drop final @ before an initial consonant,
e.g. nempe, proimie, deinde, which before a consonant often took
the forms *xemp (so scanned by Plautus and Terence, ch. x. § 7),
prow, dein, as atque, néque became ac (for *atg, ate), nec. So
benficium, malficium, calficio, &e. And the influence of the
following accent, rather than the mere addition of extra syllables,
seems to be the real factor in the syncope in the literary period
of such words as frigddéria (Lucil. viii. 12 M.) beside frigidus,
caldérius beside célidus, portérium beside pértitor, postridie beside
postéri, altrinsécus beside ditéri. The weakening effect on an
unaccented syllable of a following accent is shown by Plautine
scansions like sénéctitem, volintdlem, potéstatem, pertstréma, where
an additional weakening element is supplied by the short syllable
§ 13.] ACCENTUATION. SYNCOPE. 173

preceding. These scansions must reflect the pronunciation of
these words in ordinary conversation, Similarly calé-facere
became calé-fécere and cal-facere, and ministérium passed into min-
stérium oy mistérium (cf. Plaut. Pseud. 772), the consonant-group
st being specially adapted to combination with a preceding liquid
or nasal (cf. per()stroma above). The unaccented -v7- of dvtdus,
which resisted syncope in the simple adjective-form, succumbs to
the influence of the following accent in the lengthened derivative
*avidére, audére, to have a mind for, to dare [e. g. si audes (Plaut.),
if you please, in the classical period sd/es].

Arid- of dridus becomes ard- in ardére, ardérem, and by their
analogy sometimes appears in the simple adjective (ardus, Lucil.
xxvil. 40 M.). Similarly aet- for aevit- in aetés may have come into
use first in the lengthened cases aetdtis, aetdti, aetdtem, or in deri-
vatives like aetérnus, though here the syncopated form of the
trisyllable established itself in ordinary usage, unlike ardus. Forms
like caldérius may have had some influence in introducing caldus
into the colloquial usage (it hardly came into the literary) of the
Augustan period. Quintilian (i. 6. 19) tells us that Augustus
stigmatized as a piece of affectation the use of calidus for caldus
(non quia id non sit latinum, sed quia sit odiosum, et, ut ipse
Graeco verbo significavit, weplepyov), and yet in the Appendix Probi
we find caéda under the same condemnation as frigda, virdis
(198. 3 K.).

Post-tonic syncope, under the new accent law, seems, during
the Republic and Early Empire, to occur only when the accented
vowel is long}, e.g. jargo (in Plautus still jurigo), usiinpo for
*usiripo, -ds for (Plautine) -dtis in nostrés, Arpinds, summds,
&c., though we find it in the period of the Early Literature after
a shortened vowel in words of four or more syllables where three
short syllables followed each other before the final syllable, e.g.
bilinéum [so Plaut. and Ter.,and dalineator, Rud., 527 (A.)], a spell-
ing which did not yield for some time to later la/newm [balinearium,
C.I_L.i. 1166 (¢.130B.¢.); Caper (first cent. a.D.)(108. 7 K.) prefers
balneum ; cf. Gloss. Plac. 9. 29, and see Georges, Lew. Wortf.s.v. ;

 

1 Syncope after a long vowel is due to the tendency to make a long vowel
extra long.
174 THE LATIN LANGUAGE. [Chap. IIT.

balineum occurs frequently on the Lex Metalli Vipascensis of the
first century a.v. (Eph. Epigr. iii. p. 166)], dpttimus [the spelling
opitva of an archaizing inscription, C. 1, L. i. 1016 (=vi. 1958)
of the late Republic or early Empire shows that this form had not
been so long obsolete as to be forgotten, though as early as the
Scipio Epitaph, C. J. Z. i. 32, e. 200 B.c., we have optvmo, ‘ opti-
mum’]. These words, as we saw above (§ 5), had in the time of
Plautus and Terence the accent on the first syllable, balinéum
producing talneum, *épitumus optumus. Similarly dpificina,
opificina (Plaut. Mil. 880), produced officina, puéritia (with the
ictus puéritia in all the instances, not many, of its occurrence in
the early dramatists) produced the puertia of Horace [C. i. 36. 8
actae non alio rege puertiae ; Charisius (fourth century a.D.) still
recognized pueritia as the correct form, 266. 7 K.]. But forms like
caldus, virdis, domnus belong to colloquial or to Vulgar Latin,
and were not as a rule established in the language till the later
Empire, though valde, older viilide, Plaut. Pseud. 364, and a few
other words, were current at a much earlier time. The same
wave of syncope that reduced viridis, diminus, &c., to dissyllabic
form attacked w, ¢in hiatus (cf. ch. ti. § 48). As early as the latter
half of the first century a.D. ¢énuis varied between a dissyllable
and a trisyllable (Caesellius ap. Cassiod. vii. 205. 16 K.); cardus (for
cardwus) (ch. ii. § 54), mortus (for mortuus), &c., are the precursors of
the Romance forms (Ital., Span. cardo, Ital. morto, Span. muerto,
Fr. mort); while the similar reduction of 7 (e), led to that
palatalization of consonants which has so transformed the whole
appearance of the Romance languages, e.g. Ital. piazza, Span.
plaza, Fr. place from Vulg. Lat. *platya, Lat. plitéa, &e. (see
ch. i. § 48).

Forms like saeclum beside saeculum are not to be classed with
forms like jurgo beside jurigo, for saeclum is the older form, while
in saeculum a vowel has been inserted between the ¢ and the /
to facilitate pronunciation, a vowel which is generally called
a ‘parasitic? vowel, or, in the terminology of the Sanscrit
grammarians, a ‘svarabhaktic’ vowel (from Sanscr. svarabhakti-,
‘partial vowel’); see chap. ii. §154. The termination -célus, -ciilum
in Latin sometimes represents the I.-Eur. suffix -d/o, which
indicates the instrument with which an action is performed, or
§ 18.) ACCENTUATION. SYNCOPE. 175

the place of its performance, sometimes the suffix -colo, a com-
pound of two I.-Eur. diminutive suffixes, -co and -/o. To the
first class belong words like wzh%ewlum, ‘that by which one is
carried’ ; pdculum, ‘that out of which one drinks’; cibiculum,
‘the place where one lies down’; périculum from *perior, peritus,
experior: draculum from orare. To the second, diminutives like
corculum, uxorcula, sucula. These two classes of terminations
are not distinguished by us in our ordinary practice of writing
Latin; they show, however, in the hands of Plautus a notable
difference of metrical treatment. For in his verses the first
suffix appears normally as one syllable, -clus, -clum, reflecting
without doubt the current pronunciation of his time; the second
as two syllables, -cu/us, -culwm. Thus Plautus has always
véhiclum, with that monosyllabic form of the suffix which we
find invariably when by dissimilation the c/ is changed into cr,
e.g. ambiilacrum, ‘a place for walking in,’ for *ambulaclum,
while diminutives like aurteula, pulvisculus retain the dissyllabic
suffix in his plays as persistently as dyricdla, incdla, or any other
compound of the verb cd/o. So do nouns formed by the
addition of the suffix -do to -co-stems, in distinction to those
formed by its addition to -c-stems, e.g. porcu-lus, céloc-la, ‘a
yacht.’ Probably not a single instance occurs in his verses of
-col- reduced to -cl- by syncope, even after a long vowel, e.g. never
*eorclum from cor-cu-lum, *porclus from poreu-lus (Class. Rev. vi.
87). (But privicloes, ‘ priviculis,’ Carm, Sal., ch. vi. § 49.)

Forms with the parasitic or svarabhaktic vowel, with -colo-
for -clo-, are indeed not infrequent with him. Cudiculum, for
instance, always or almost always, appears as a quadrisyllable.
But he uses these expanded forms as a rule (especially when
a long vowel precedes), only at the end of a line or half-line ;
that is to say he regards the equivalence of the parasitic vowel
to an actual short syllable as a licence only to be resorted to
in cases of metrical necessity. For example, periclum is the
normal form of the word, while pericuum occurs only at the end
of a line or hemistich. Capt. 740 is a good example of this
distinction :

periclum vitae méae tuo stat periculo ;

and it is only at the end of a line that the phrase nu/dumst peri-
176 THE LATIN LANGUAGE. (Chap. ITI.

clum (e.g. Pseud. 1076) becomes nullum periculumst (e.g. Capt.
91). Poculum, too, shows this full form usually in a similar
position, while saeclum is never allowed trisyllabic scansion
at all. The Romance languages show that a later wave of
syncope not only reduced saec(u)/um, &c. to their original form
saeclum, but also words like porcudus to porclus, e.g. Ital. cerchio,
(Lat. cire(u)/us), teschio (Lat. test(a)la), spillo (Lat. sptn(u)la).
(CE. oclus, C.I. L. x. 7756, &c., crustlum, xi. 3303, of 18 a.v.,
Proclaes, xv. 1157, of 123 a.v., Vitlus, vill. 9432, &e., Aevtaos,
&e. on Greek inscriptions, Eckinger, pp. 73-5, Masclus (Gk.
Mackdos, &c.); so in Probi App. 197. 20-22 K. speculum non
‘speclum,’ masculus non ‘ masclus,’ vetulus non ‘veclus,’ vitulus
non ‘viclus,’ vernaculus non ‘vernaclus,’ articulus non ‘arti-
clus, baculus non ‘baclus,’ angulus non ‘anglus,’ jugulus non
‘juglus, and 198. 18 oculus non ‘oclus,? 198. 23 tabula non
‘tabla,’ 198. 27 stabulum non ‘stablum, 198. 34 capitulum
non ‘capiclum,’ 199. 9 tribula non ‘tribla,” 199. 14 vapulo
non ‘vaplo’ (MS. baplo); and see George’s Lex. Wortf. s. vv.
coag(u)lo, aedic(u\la, assec(u)la, bub(u\lus, conch(u)la, Vist(u)la,
cop(u)la, cop(u)lo, eubie(u)lum, vit(u)lus, discip(u)lina, eatemp(u)lo,
fil(uyla, fig(u\linus, laterc(n\lus, Here(u)le, jug(ujlans, manic(u)la,
ment(ujla, masc(u)lus, orac(u)lum, poc(u)lum, peric(u)lum, pedi-
c(u)lus, saec(ujlum, scrup(u)lus, seatula (sescla), surc(u)lus, spec(u)-
dum, sub(ulla, temp(ullum, trich(i)la, vine(u\lum, vet(w)lus, ver-
nac(u)lus, Asc(u)lum, Vist(u)la, &e.; cf. Schuch. ok, i. 402 sqq.).
Fr. roule, &c. from unsyncopated 7df#/a points to a re-formed
diminutive; so in Roumanian, e.g. teule or tiule (Lat, tégu/a)
(Taverney in Etudes... G. Paris, p. 267). (See also § 21.)

Another case of syllable-suppression which differs from
ordinary syncope is that of words like areudii (‘qui excubabant
in aree,’ Paul. Fest. 19. 10 Th.) for arct-cithiz, where the
proximity of two almost identical syllables c? and c# has caused
or facilitated the dropping of the former. [Similarly sémi-
modius and semodius, débilitare for *debilitatare, fastidium for
*fastitidium, wldlatria for tddlila-, limitrophus for limitotro-,
Restitus for Restitaitus (A. L. L. viii. 368), &c., as in Greek du(d.)-
hopeds, Wn(ho)popia, Nei(r0)mvpia, &e.].

And a large number of words, like diem (beside diwissei),
§ 14.] ‘ACCENTUATION. SYNCOPE. 177

aucissem (beside audivissem), audi-trie (beside auditor), gi-gno
(beside gé-ws), are often wrongly included with genuine cases
of Latin syncope like /ardum, jurgo. In some of these there is
either no syncope at all (so some explain dixem for *dic-sem,
like es-sem; but see ch. viii. § 3), or, if there is, it took place in
the I-Eur. period, e.g. I.-Eur. *83-6né (ch. iv. § 51), -tri- (d.) ;
others, e.g. divem, if a reduction of diwissem (ch. viii. § 3), are
rather to be explained like ar(ci)eubii above; audissem, if
a reduction of audivissem (ch. viii. § 3), like sis for si vis.

§ 14. Syncope in the Praenestine Dialect of Latin. On the Latin
inscriptions found at Praeneste there are a large number of omissions of
vowels, e.g. Deumius (for Déciimius), C.I.L. i. 1133, Gminia (for Géminia), Eph.
Epigr. i. 72, Diesptr (for Diespiter), C.I.L. i. 1500; Ptronio (for Pétronio), Eph.
Epigr. i. 92; a full list in Sittl, Lokal. Verschied. Lat. Sprache, p. 22. This
tallies so remarkably with a reference by Plautus to a peculiar pronunciation
of the Praenestines, conea for cicénia, ‘a stork,’ that it is likely that these
spellings represent the actual sound of the words. In the Truculentus of
Plautus the surly, taciturn slave Truculentus, whose mispronunciations are
more than once a subject of jest (cf. line 683), turns the word arréibo, ‘a prepay-
ment,’ ‘earnest-money,’ into rabo, ‘a raver’ (cf. rabere, ‘to rave,’ rabula,
‘a bawling pettifogging lawyer’); and on being taken to task, says that he
has pocketed a part of his arrato as the Praenestines do with ciconia (line 690) :

‘ar’ facio lucri,
ut Praenestinis ‘conea’ est ciconia.

(Cf. misisia for Ital. amicizia in the modern Parmese dialect ; Ital. nemico
for Lat. inimicus, &.) The omitted vowel is e or i, once a (Mgolnia, C.I.L.
i. 118) (besides Acmemeno for Agdmemno, on an old Praenestine cista, Eph.
Epigr. i. 19), and the omission is easily explained by the influence of the
accent in words like Diésp(<)t(e)r, P(e)tronio, and, if we admit that the ante-
paenultima was accented in these words in the Praenestine dialect at this
period, in D(e)cumius, G(e)minia (ef. Cem(i)na, C.I.L. i. 99). But an accented
vowel is omitted in Trtia (for Tertia), Eph. i. 108, Pol(t)dia, Eph. i. 95, and
even a long accented vowel in Atlia (for Atilia), Eph. i. 33.

A grammarian of the second cent. a.p. (Terentius Scaurus, pp. 14, 15 K.) tells
us of a practice of an earlier date of substituting a letter for the name of the
letter. The name of c was ‘ce,’ of d was ‘ de,’ of k was ‘ka’; and so cra was
written for céra, kra for kara (cara). His example for dis almost exactly our first
example of the Praenestine contraction, viz. Deimus for Decimus. These words
of Terentius Scaurus suggest that spellings like Albsi for Albési, Albensi on an
inscription of Alba Fucentia, a town not far removed from the Praenestine
district (Zvetaieff, Inscr. Ital. Inf. 46), lubs mereto (for lubés, lubens merito) on
an inscription found near Avezzano, in the same neighbourhood (C.I. L.
i. 183), are indications of a syllabic system of writing in partial use in this
region of Italy (ch. i. § 13), and leave us in doubt about the real nature of
Praenestine pronunciation.

N
178 THE LATIN LANGUAGE. [Chap. III.

§ 15. Syncope under the Old Accent Law. (1) The preposition ambi-
(Gk. dui) loses its second syllable in compounds like dnculus, ‘a servant,’ for
*dmbi-cilus (Gk. dyugi-rodos, Sanser. abhi-caras), an old Latin word from
which came anculare, ‘to serve’ (Paul. Fest. 15. 7 Th.) and ancilla ; ancipes
(Plaut. Rud. 1158), later anceps (from ambi- and cdéiput) ; am-plector, &e. The
same shortening may have caused that confusion of the old preposition indo
(endo) (ch. ix. § 27) with the preposition in (en) (ind(o)grédior, &c. becoming by
syncope identical with ingredior) which led to the disuse of indo and the
adoption of im in its place. Thus indaudio (Plaut.) was completely ousted
by inaudio in the time of Terence, and in the classical period compounds
with indo are only found as archaisms in poets, e.g. indtipérator Juv.

(2) The second syllable of the first member of a compound is syncopated in
hospes for *hosti-pes, princeps for *primi-ceps (cf. primi-génia, an epithet of the god-
dess Fortuna, unsyncopated because mand g do not easily combine) ; forceps
for formi-ceps (from formus, ‘hot,’ connected with Gk. Oepyds. Paul. Fest. 59. 18
gives us this etymology, forcipes dicuntur, quod his forma, id est calida,
capiuntur; cf. Vel. Long. 71.15 K.) 3 quindecim from quinqué and décem ; wndecim
from tinus and decem ; universus is a re-formation from an older syncopated
form which we find on the Decree of the Senate against the Bacchanalian
orgies, an inscription of the time of Plautus (186 3B.c.) (though Plautus
himself uses the four-syllabled universus) (C.I.L. i. 196. 19 homines plous V
oinuorsei uirei atque mulieres sacra ne quisquam fecise uelet) (this may be
a mistake for *oinwuorsei, like swrsuorsum on the Sentent. Minue. 1. 15) ; vindemia
for *vini-démia ; Marpor (C. I. L. i. 1076) for Marci-por, &c.

(3) The first syllable of a verb compounded with a preposition is suppressed in
pergo for *per-rigo (cf. perrext, perrectum); porgo, the old form of porrigo, *por-régo (cf.
Fest. 274. 15 Th. antiqui etiam ‘ porgam’ dixerunt pro porrigam), e. g. exporgere
lumbos, ‘to stretch one’s legs,’ Plaut. Pseud. prol. 1, ef. Epid. 733 ; the word, sanc-
tioned by the usage of Virgil (A. viii.274 pocula porgite dextris), appears now
and then in the Silver Age poets (Val. Flace. ii. 656; Stat. Theb. viii. 755, &c.),
but the classical form is porrigo; surgo for surrigo, *sub-régo ; a deponent perf.
participle sortus for *swrctus, formed on the analogy of the syncopated pres. ind.,
was often used by Livius Andronicus (Paul. Fest. 423. 1 Th.) ; beside surripui
(classical surripui) we have the syncopated form surpui in Plautus (e.g. Capt.
760), and even a perf. part. pass. formed after its type, surptus (Rud. 1105).
Lucretius also uses the contracted form of the present (ii. 314 motus quoque
surpere debent), and Horace (8. ii. 3. 283 unum me surpite morti; ef. C.
iv. 13. 20 quae me surpuerat mihi) ; pono for *pé-sino (the preposition is pé-
a byform of *épd, db), postus, depostus, compostus, very common for pi-situs, &e. ;
like postus is prae-std, earlier prae-sti (Cassiod. 157. 22 K.), ready, at hand, for
*prae-situ ; cette for *cedite, *cé-ddte, where the particle c?, ‘here,’ ‘hither,’ is
prefixed to the verb, asin Oscan ce-bnust, ‘hue venerit’ (Zvetaieff, Inscr. Ital.
Inf. 231. 20). These syncopated forms were probably far more frequent in
the early period, than at a later time, when the same tendency to re-
composition which produced con-sacro out of consecro, ad-sum out of assum
(ch. iv. § 160) restored por-rigo, sur-ripui, &e. The older forms might remain
undisturbed in derivatives whose connexion with the verb was unnoticed,
e. g. refriva faba (referiva, Plin. xviii. 119), the bean brought back by the farmer
from the field for luck (in Fest. 380. 17 Th. we are told that the word was
also associated with refrigo, ‘to roast,’ ‘parch’) ; apricus, Aprilis may be similar
§ 15.) ACCENTUATION. SYNCOPE. 179

traces of an ap-(e)rio (cf. dp-ério), unless the syncope in all three words was pre-
tonie syncope under the Paenultima Law of Accentuation, ref(e)riva, ap(e)ricus,
Ap(e)rilis, Verbs beginning with a vowel unite it into one sound with the
final vowel of the preposition, though to what period of the language this
crasis should in each case be referred is uncertain, for it might be caused by
« following as well as by a preceding accent. Crasis was the rule in such
compounds in the early period, to judge from the usage of the older poets.
Coerce, for example, is dissyllabic in Pacuvius (Yrag. 47 R.),

gradere Atque atrocem coerce confidéntiam (see ch. ii. § 150).

(4) The first syllable of a reduplicated perfect of a compound verb is
dropped in reppéri, rethili, recctdi, &e., where the double consonant seems to
preserve a trace of the syncope (see ch. viii. § 44). The syncope would in these
perfects be facilitated by the Latin tendency to drop one of two neighbouring
syllables of like sound (see on arcubii for arci-ctibii, above § 13); and it is
natural to suppose that the perfects without reduplication, like ex-scidi (O. Lat.
scicidi), con-curri (and con-ciicurrt, older *con-cécurri), &c., originated in this way
(ch. viii. § 44), just as in modern Greek B.fd(w has become Baw in compounds
like b:aBagw, éuBatw, &e. Another syncopated verb-form is cante (2 Plur.
Imperat. of ccino), quoted from the Carmen Saliare by Varro (L. L. vii. 27).
But as a rule all traces of syncope in the declension of the verb have been
obliterated, the full forms (e.g. cdnite) having been restored through the
influence of forms where the consonants were not adapted for combination
(e. g. sistite), or where the syllable in question did not immediately follow
the (early) accent (e.g. céncinite), as well as from the analogy of other con-
jugations, v.g. améd-te, moné-te, audi-te. In Umbrian and Osean these
imperatives are syncopated, ¢.g. Umbr. sistu (Lat. sistito), Osc. actud (Lat.
dgito). The same is true of derivative adjectives, e.g. in -idus, himidus,
frigidus, ctilidus, silidus, rigidus, aridus, &e. (but nudus for *novidus from a root
nogw-, Sanscr. nag-nas, our ‘naked,’ is syncopated; on a#dus see below), in
-icus, e. g. dnicus, ctvicus, médicus, but the nouns Plancus, lurco, ‘a glutton’ (from
litra, ‘the mouth of a sack,’ according to Paul. Fest. 86. 23 Th. lura, os cullei, vel
etiam utris ; unde lurcones capacis gulae homines), juncus, &ce. are syncopated
(see below on raucus). Similarly vividus is saved from syncope by the influence
of other adjectives in -idus, while the noun vita for *vivita (Lith. gywata) is
not. The Umbrian adj. in -co-, titco- (Latin piblicus) from tota-, *touta-, ‘the
community,’ ‘people,’ shows the contraction which Latin adjectives of this
formation escape (cf. Ose. toutico-).

(5) Diminutives in -lo-, on the other hand, extended the syncope from
dissyllabic to other forms, e. g. allus from *ano-lus, villum from *vino-lum, Ter.
Adelph. 786, and from their analogy, céralla for *cordnula, persilla for *persdnula,
ampulla from amp(h)ira (Greek dppopG, Ace. of dupopeds), &e.

(6) Greek words borrowed at an early period probably owe their syncope to
the early accent, e. g. Herctiles (Greek ‘HpaxAjjs), Polliices (the early form of Polliix),
Plaut. Bacch. 894 (Greek MWoavdevens), cala (Greek xéAr¢, with &) ; cf. Praenes-
tine Acmemeno (above, § 14).

(7) Other examples are alter from diliter ; postulo from *posci-tilo ; ulna from
*ulina (Greek wAévn) ; original -In- becomes -U- in Latin, ev. g. collis from *colnis
(Lith. kétnas, Greek xodwvds) ; propter from *propi-tér.

(8) For the syllable -ri-, examples are: sdcerdds from *sacré-dés, *sdcri-disy

N 2
180 THE LATIN LANGUAGE. (Chap. III.

from sdcer and root dd, ‘to give’; dcerbus for *dcri-dho-, with the same termina-
tion (belonging to root dhz, ‘to make’) which becomes -dus when not preceded
by 7, e.g. frigidus, calidus (see ch. iv. § 114) ; sticellum for *sacer-lum, *sacré-lom,
sdcri-lum, &e. Nouns like dger (Greek d-ypés), cdiper (Greek xampos), and adjectives
like dcer show this treatment of the unaccented final syllable of the Nom.
case, whereas hortus, &c. are saved from syncope by the analogy of other cases,
horti, horto, hortum, &c., and of other Nominatives, whose final syllable did not
immediately follow the (early) accent, like dnimus, aitumnus. In the Italian
dialects we find this -er- in similar positions, e. g. Osc. Aderla- from *Adrola-
(Lat. Atella), Abella- probably from *Abrola- (Lat. *Apelia), Umbr. ager, pacer,
‘propitious,’ from pacri-, often spelt -r-, e.g. Sabine Atrno- (Lat. Aternus), Ose.
Tantrnnaiim (Gen. Pl.) (ch. iv. § 92). In Latin we seem to have -er- for -ri,
in the accented syllable too, e. g. ter (Greek pis), in Plautus (e. g. Bacch. 1127)
scanned as a long syllable, and so pronounced terr (cf. terruncius, the right
spelling ; see Biicheler in Rhein. Mus. xlvi. 236), from *ters. The substitution
of é for ri would then be due rather to metathesis (cf. N. Ital. fardor, &c. for
fredor, &c., Meyer-Liibke, Rom. Gramm. i. p. 291, and see above, ch. ii. § 111).
Ter would however be unaccented in phrases like ter-mille (cf. terdeciens written
as one word on Mon. Ane. i. 29). On ter, see ch. vi. § 61, on er for ri ch. iv. § 13.
(9) For the syllable -vi-, e. g. autwmo for divi-timo (Greek ofw for 6f-1w) ; claudo
*clavi-do, from clavis ; curia for *civiria (cf. Volscian covehriu, Zv. Inser. Ital. Inf.
47) 5 gaudeo for *gavideo, part. gavisus (Greek ynOéw for yaFe-Séw) ; naufragus for
*navi-fragus; niper, ef. adj. nuperum Acc. Sg., Plaut. Capt. 718 recéns captum
hominem ntiperum nouicium), for *nivi-pérus from névus and péro 3 praeco for
*prae-vico from prae and vico; praedes, earlier praevides (praevides Plur., praes
Sing. on the Lex Agraria of Sp., Thorius, 111 B.c., C.I.L. i, 200, but only
praedes on the older Lex Repetundarum, 123-2 B.c., i. 198), from prae and vas,
perhaps suffered syncope after the new accent law, as did idus (as early as
Lucil. inc. 172 M.), in Plautus only avidus ; raucus for rdvicus from révis, ‘ hoarse-
ness’ ; vita for *vivita (see above) ; auceps for *dvi-ceps ; Opiter, a name given to
a child ‘who had a grandfather for a father,’ cujus pater avo vivo mortuus
est (Paul. Fest. 207. 15 Th.), Plur. Opiteres (Liwe, Prodr. p. 396), seems to be
colloquial Latin for *Aupater for *dvi-pdter from dvus and pdter. The form
méavolo is found (with mdlo) in Plautus, but in the classical period only malo.
The syllable -vi-, -vé- is syncopated, even when long by ‘ position,’ in auspex
for *dvi-spex ; niintius, older néventius [the older form occurs in a prophecy of
the famous Marcius, published 213 B.c. (ap. Fest. 164. 28 Th.) quamvis
noventium duonum negumate, ‘quamvis bonum nuntium negate’) ; nundinae
for *nédven-dinae from novem and din- a bystem of dies (Sanscr. dinam, O. Slav.
dini). And -v- is dropped even before a long vowel when a vowel of the
same quality precedes, e.g. labrum, earlier ldvabrum, Lucr. vi. 799. (Marius
Victorinus, ix. 20 K. quotes lavaibrum for labrum among other instances of Old
Latin forms such as hacetenus for hdcténus, hocedie for hédit, semol for simul) ;
latrina for ldvatrina, the old word for a bath, supplanted by the Greek loan-
words bal(i)neum (BaAdvetov) (cf. Non. 212. 7M. latrina ...est lavatrina, quod
nune balneum dicitur), as balnewm itself was succeeded by lavacrum : divinus
was early contracted to dinus (Leo in Rhein. Mus. xxxviii. 2), e.g. reidinai and

 

1 Or for *ndu-fragus, which became néa-fragus (ch. iv. § 45), and so with
claudo, gaudeo (?).
§ 16.] ACCENTUATION. SYNCOPE. 181

res deina on an old inscription (C.I.L. xi. 4766), so obliscor for obliviscor in the
early dramatists (see Georges, Lex. Wortf. s. v.), and in universal usage, sis, for
si vis, from which a plural sultis was formed. (On loss of intervocalic -v-
see ch. ii. § 53 and ch. iv. § 70.

This liability of every short second syllable to syncope under the Early
Accent Law makes it dangerous to infer from Latin forms the presence or
absence of a short vowel in the corresponding Indo-European forms, e. g. to
infer from the distinction between Lat. titra, citra, infra, siipra, extra, contra on
the one hand, and wiltérior, citérior, inferior, siipérior, extérior on the other, that
the original stem-suffix was -tr-, -r-in these adverbs and -ter-, -er- in these
adjectives. Priscian (ii. p. 30. 1 H.) tells us that the older forms were supera,
infera, extera, &c. quaedam etiam syncopam passa sunt, ut ‘supra’ pro
‘supera,’ et ‘infra’ pro ‘infera,’ et ‘extra’ pro ‘ extera,’ nam antiqui trisyl-
laba ea proferebant, ut Cicero in Arato :

Torvus Draco serpit supter superaque retorquens
Sese,

tenuit tamen, ut disyllaba magis ea proferantur (ef. ii. p. 55. 23 H.). Stipéra is
found in an elegiac epitaph, of the time of the poet Accius (to judge from its
use ofa double letter to indicate a long vowel), C.J L.i. 1011 Ree fuit ee
vero plus superaque parens, and on another inscription, with the same
indication of date (cf. Ritschl, P. L. MU. p. 46) (-ce- for @ in seedes) we have infera,
i. 1166 quae infera scripta sont, but on the earlier 8. C. de Bacch. (i. 196 of
186 B.c.) we have suprad and exstrad, so that supera (which is used also by
Lucretius) may be a form that is not genuinely antique. On the spelling
arbiterium for arbitrium, see Georges, Lex. Wortf. s. v., and cf. magistero- C. I. L.
i. 73, malgi|steratus, Eph. Epigr. ii. 298 ; on dealéra and dextra, see Brambach,
Lat. Orth.; on sinistera for sinistra (e. g. Ter. Eun. 835), on Tibéris and Tibris, see
Georges, Lex. Wortf. s. vv., and for other examples of syncopated byforms, s. vv.
sol(i)dus, Vir(é)domarus, frig(i)dus, ful(i)ca, Temese (Gk.) and Tempsa (Lat.) ; but
audac-ter, later auddci-ter, privi-gnus beside privi-genus (ef. Paul. Fest. 225. 2 Th.
‘oenigenos’ unigenitos), teg-men beside tégi-men, &c. admit of other explana-
tions. (On the use of a Parasitic Vowel with r see ch. ii. § 102.)

Syncope is carried even further in Umbro-Oscan than in Latin, but in
Umbrian the Perf. Part. Pass. is not syncopated, while the 3 Sg. Imper. is (but
not -net-, e.g. kanetu, ‘let him sing’) ; thus sektu is Imper., segetu is P. P. P.
(see von Planta, i. p. 214). Contrast Osc. toutico- with Umbbr. totco-, ‘ publicus,’
Ose. minstro- with Lat. ministro-, Osc. Vezka- with Lat. Vetusco-.

§16. Syncope of Final Syllable. In Oscan and Umbrian, as in Gothic, &, 6,
%, but not %(2), in a final syllable are syncopated, e.g. Osc. hurz, Lat. hortus,
Bantins, Lat. Bantinus, tavtiks for *touticos from touta., ‘ community,’ ‘ people,’
Umbr. emps, Lat. emptus, pihaz, Lat. pidtus, all with syncope of -is; Ose.
peddeg for *med-dik-és Nom. Pl. of meddix, the title of the Oscan chief
magistrate, censtur for *censtor-ds, Lat. censdrés, with syncope of -és of Nom. Pi.
(see ch. vi. § 40). Umbr. pacer for *pac-ri-s, ‘propitious,’ with syncope of -ris.
In Latin we have this syncope, unless it should rather be called metathesis
(§ 15. 8), in Nom. Sg. of -ro-, and 7i- stems, e.g. dger for *ag-ro-s, in-tég-er for
*en-tag-ro-s, dicer for ac-ri-s, viliicer, &c., but perhaps in no others. Quaituor
might be for *quetwor-ts, Mase. (Doric Gk. réropes, Sanser. catvaras), but may
182 THE LATIN LANGUAGE. (Chap. III.

also be the Neuter form (Sanser. catviri) (see ch. vi. § 63). The contraction of
-atis to -ds in the Nom. 8g. of adjectives or nouns denoting the place of one’s
birth, e. g. Arpinas, *Casilas (Umbr. Casilos), is later than Plautus, who always
uses the full form -atis: while Campans (Mase. not Neut.) in his cruel sneer
at the conquered Campanians, Trin. 545:

Campans genus
Multo Surorum iam antidit patientia,

seems intended to mimic an Oscan *Campans for Campdnus, like Osc. Bantins
for Bantinus. Like Arpindtis, later Arpinas, are Sammnitis, later Samnis, Laurentis,
later Laurens, Tiburtis, later Tiburs, &c. (Prisc. i. p. 134 H.). Nominatives Sg. of
i-stems like sors (in Plaut. Cas. 380 sortis), quiés (if an I-stem like O. Pers.
Siyatis, Av. Syéitis) may have dropped i in their final syllable, not by
syncope, but by analogy of consonant-stems ; cf. nubs used by Liv. Andr. for
nibes (Serv. ad Aen. x. 636), plebs and plebés (for other examples, see Ritschl.
Opusc. ii. 652). Or these may be instances of parallel stem-formations, like
pent-, peno-, penos- of pénus, Gen. peniis, penum, Gen. peni, penus, Gen. pendris,
So vidlens and vidilentus, fluens and fluentum, &c., epulonus (Paul. Fest. 55,
15 Th. ‘epolonos’ dicebant antiqui, quos nunc epulones dicimus), and
Epiilo, centurionus and centiirio, curionus, decurionus and citric, dectirio (Paul. Fest.
34. 36 Th. ‘centurionus’ antea, qui nune centurio, et ‘curionus’ et
‘decurionus’ dicebantur), infans, once used by Accius (Zag. 189 R. infans
facinus) in the sense of infandus. Compounds like in-dex, tii(s)-dex, vin-dex
differ from causi-dicus, iwrt-dicus, fati-dicus in being formed directly from the
weak verb-stem dic-, like Sanser. a-di8-. So conjux (cf. Sanscr. sayuj-, Gk.
aug) beside bijigis and bijtigus, bigae, &e. So for(m)ceps, au-ceps, prin-ceps beside
urbi-cdpus (Plaut.), hosti-capas (Paul. Fest. 73. 10 Th. ‘hosticapas’ hostium
captor) ; dpi-fex, arti-fex, carni-fex beside mini-ficus, magni-ficus ; rém-ex beside
prod-igus. Man-sués (Acc. mansuem and mansuttem) beside man-suétus, in-quizs
beside in-quiétus are like Gk. diyjs and dbufros, dxpns and dkpyyros. Praeceps,
anceps from prae-, ambi- and cdéput, in Plautus praecipes (Rud. 671), ancipes
(Rud. 1158) (cf. procapis, Paul, Fest. 281. 22 Th. ‘procapis’ progenies, quae
ab uno capite procedit ; and concapit (?) of the XII Tables ap. Fest. 556.
27 Th. tignum iunctum aedibus uineaue et concapit ne soluito) were after-
wards assimilated to compounds of cépio, e.g. prin-ceps (ef. Prisc. i. p. 280.
15 H. antiqui tamen ‘ancipes’ et ‘praecipes’ et ‘bicipes’ proferebant in
nominativo... idem tamen vetustissimi etiam ‘praecipis’ genetivum.. .
secundum analogiam nominativi protulerunt). Old Latin Polliicés (Plaut.
Bacch. 894, cf. Gk. MloAvbedens) was shortened to Pollix, probably by analogy
of lax, Gen. licis. Priscian (i. p. 282. 12 H.) tells us that the old forms of
concors, discors, &e. were concordis, discordis (cf. i. 354. 13 H.) (ef. late Lat. orbs,
e.g. Ven. Fort. ix. 3.14; orbis non ‘orbs’ Probi App. 198. 8 K.).

O-stem adjectives often have their Nom. 8g. shortened through theirtendency,
especially when compound (ch. v. § 34), to become I-stems. Thus hildrus (Gk.
idapds, one of those loanwords from Greek to express subtlenuancesof feeling, for
which the Romans had no word of their own, like our loanwords from French,
such as ‘ triste’) became, after the time of Plautus, hilaris (ch. v. § 34) ; fortis
was perhaps originally forctus (Paul. Fest. 73. 9 Th. ‘ foretum’ pro bono dice-

- bant), though as early as the XII Tables we find /orctes for loyal allies, (ap. Fest,
524. 15 Th. in XII cautum est, ut idem juris esset ‘Sanatibus’ quod
§ 17.] ACCENTUATION. SYNCOPE, 183

‘ Forctibus,’ id est bonis, et qui numquam defecerant a Populo Romano ; ef.
Paul. Fest. 59. 26 ‘foretes,’ frugi et bonus, sive validus, where Paulus may
have put the Nom. Pl. forctes by mistake for the Nom. Sg. forctis) ; sdécro- and
sdcri- are parallel stems in O. Latin (sdcres porci, ‘ pigs for sacrifice,’ Plaut. Men.
289, Rud. 1208, cf. Fest. 464. 7 Th.), and similarly mdno- and mani-, good,’
though in classical Latin the only survival of this group was im-manis ‘bad,’
‘hurtful,’ and the di Manes. In the Carmen Saliare occurred the phrase Cerus
manus, explained by Paul. Fest. 87. 29 Th. as creator bonus, and at Lanuvium
the old word manis was in use even in the time of Macrobius (fourth cent. a..)
(Macr. i. 3, 1g nam et Lanuini ‘mane’ pro bono dicunt; sicut apud nos
quoque contrarium est ‘immane’). Varro (L. L. vi. 4) connects with O. Latin
manus, ‘good’ the adverb mane, ‘early,’ and in support of his etymology
mentions a curious Greek custom of uttering the words pds dyaddy as a good
omen when a light was brought into the room, diei principium ‘mane,’ quod
tum manat dies ab oriente, nisi potius quod bonum antiqui dicebant
‘manum,’ ad cujusmodi religionem Graeci quoque, cum lumen affertur,
solent dicere gas dyadéy (see ch. vi. § 38). The O-stem hortus became in the
compound *co-horto- an I-stem *co-hortis, which (like sors, &e. above) changed
its Nom. Sg. to co-hors. Even the Perf. Part. Pass. sdéndtus appears in the
form sanati- in the expression quoted from the XII Tables by Festus 524. 10 Th.
for the repentant allies, who had first revolted and then returned to their
allegiance, ‘Sanates’ dicti sunt, qui supra infraque Romam habitaverunt.
quod nomen his fuit, quia, cum defecissent a Romanis, brevi post redierunt
in amicitiam quasi sanata mente. And the Old Latin legal phrase dare
damnas esto, tantum damnas esto (Cato ap. Gell. vi. 3. 37; Quint. vii. 9. 12, &c.)
may be a case of substitution, for the usual O-stem damndto-, of an I-stem
damnati-, which has taken a cons.-stem Nom. Sg. damnas, like aetas, tempestas.

IO-stems had at all periods a tendency to pass into I-stems, The older
adj. termination -drius (frequent in Plautus, vid. Lorenz ad Pseud. 952,
e.g. singularius, virginarius) may have been often replaced by -aris in MSS. of
Plautus (ch. v. § 4). Cf. Caper 112. 2 K. vates olim ‘vatios’ dicebant ; so Verres
and Verrius. In Vulg. Lat. -ius (-eus) became -is in actuaris, abstemis, sobris,
caerulis, consanguinis, &e. (Léwe, Prodr. p. 420), thus repeating the early
confusion between -io- and -i- in the declension of names like Cuecilius, Ace.
Caecilium, Caecilis, Acc. Caecilim (see ch. vi. § 5). But none of these are clear
cases of the change of stem of a Latin word owing to the syncope of its final
syllable in the Nom. Sg. Perhaps the most likely instance is Lucipor, from
Lucius and puer (stem piéro-), of which the Plur. is given by Pliny, H. NV.
xxxiii. 26 as Luctpores (cf. Dat. Sg. Naepori on an inscription of the end of the
Republican period, C.J.L. i. 1539 ¢), but even this might be otherwise
explained. The weakening of final vowels in Latin (see below) gives an
& priori probability to the syncope of final short syllables like -is, -ds, -és as
in Osean and Umbrian, but it has not yet been satisfactorily proved that
syncope did actually occur in any syllables except those immediately
preceding or following the accent. (Schuchardt, Vok. ii. 394 sqq. has collected
a number of instances on late plebeian inscriptions of the omission of
a short vowel of the final syllable, e. g. fect for fécit.)

§ 17. Syncope under the Paenultima Accent Law. (1) Pretonic. Compounds
of jacio like céilé-facio, which shortened their @ by the law of Brevis Brevians
(see below), took the further step of suppressing the vowel altogether before
184 THE LATIN LANGUAGE. (Chap. III.

the accent of the next syllable, cal-facere, cal-factus. Quintilian (i. 6. 21) tells
us that in his time the full form caléfacere was never used in ordinary talk.
Olfacere, not *olefacere, is the regular form. Ritschl proposed to help the
metre occasionally in Plautus by reading benjictum, malficium for bénéficium,
miléficium, benfacta, malfacta (e.g. Trin. 185) for benefacta, malefacta of the MSS.
He supported his proposal by the old spelling BENVENTOD on a coin of
Beneventum (C.I. L. i. 19), c. 250 B.¢. On later inscriptions spellings like
BENMERENTI are frequent, also maLpictvm (see Ritschl, Opuse. ii. 716). So
firmly established was the syncopated form of compounds of facio like
olfacio that even Gréfacio, whose 2 could not be shortened by ordinary phonetic
change, since it is preceded by a long syllable, seems after their analogy to
have been made a quadrisyllable by Cato, for the MSS. of the Res Rustica
agree wonderfully in presenting the word in this form (c. 69 ; 125; 157. 12).
To pretonic syncope we must refer the currency of the forms disciplina, figlina
beside discipulus, figulus, and on later inscriptions vetranus (ef. C.I.L. iii. Ind.
p. 1159 for vittranus, &c. (on Greek inscriptions almost always ovezpavos or Be-
tpavos). Festus 466. 16 Th. tells us that scéna, an old word for the priest’s knife
(used by Liv. Andr. Com. 2 R. corruit quasi ictus scena) had a byform sacena ; and
another obsolete term sculna, discussed by Gell. xx. 11, a synonym of séquester,
was explained by a grammarian, who compiled a sort of ‘Slang Dictionary ’
(Lavinius ‘De Verbis Sordidis’) as a contraction of *secuina. Vulg. Lat,
mat(a)tinus (Ital. mattino, &c.) may be explained either as a case of the
suppression of one of two similiar neighbouring syllables, like Res(ti)titus
above (§ 13, p. 176) or of pretonic syncope, such as is seen in Ital. cervello (Lat.
cerebellum), vergogna (Lat. vérécundia), bonta (Lat. bénitdt-), gridare (Lat.
quiritare), dritto (Lat. directus), &c. Procope is common in Italian, owing to
the frequency of final vowels, e.g. vescovo (Lat. giscipus\, nemico (Lat.
inimicus), cagione (Lat. occdsion-), &c. Synizesis of the pretonic short vowel
is seen in Vulg. Lat. qu(i)éus, *dyurnus (Ital. giorno) ; coactus became *quattus
(Ital. quatto), &c. (cf. Georges, Lex. Wortf. on Num(i)torius, Lug(u)dunum).

(2) Post-tonic. A good example of syncope after a long accented syllable
under the new Accent-law is the word barca (our ‘barque’), a word which
seems to have been introduced at the time of the naval displays given by
Caesar for the amusement of the people, and which is clearly a contraction of
*parica from the Egyptian baris (Prop. iii. 11. 44) (see Rhein. Mus, xlii. 589).
Another is lamna (Hor. ¢. ii. 2. 2 inimice lamnae), in Vulg. Lat. lanna (Arnob. ii.
41), the older form of which was lammina (e. g. Plaut. Asin. 549). And we have
many words which appear in Plautus in their full form, but in later writers
are reduced by syncope, such as obiirigo, by Terence’s time always obiiirgo,
nouns or adjectives in -dtis denoting the country of one’s birth, &., e.g.
infimatis (Stich. 493). The same shortening tendency attacked u, i in hiatus,
e.g. larua is a trisyllable in Plautus, a dissyllable later, so gratis, later
gratis, while it has left traces of itself in spellings on old inscriptions like
iugra (for jagéra) on the Lex Agraria of Sp. Thorius, 111 B.c. (C.. L. i. 200.
14, 25), not to mention others which may be dialectal, such as PROSEPNAI
(Dative) on a very old mirror of Cosa (¢. I. L. i. 57. -ar, not -ars, is what is
written ; see Rhein. Mus. xlii. 486), and cepre for caedere on an early inscription
of Spoletium in Umbria (C.J. L. xi, 4766). Ardus for dridus appears occasionally,
e.g. Plaut. Aul. 297; Pers. 266; Lucil. 27. 40 M., and on an inscription
copied in the Empire from an original of 105 B.c. (C.1L.i1.577. 2. a1=
§ 18.] ACCENTUATION. VOWEL-WEAKENING. I 85

x. 1781), which also contains wda (2. 18) for ivida (but dridus, Plaut. Rud.
574, 726, 764, &c.) ; so Raude for Ravide, Catull. xl. x; aspris for aspéris, Virg.
Aen. ii. 379 (cf. aspritido, asprétum, asprédo, and other derivatives, as well as
Ital. aspro), aspriter, Sueius ap. Non. 513 M. Syncope after a short accented
syllable is seen in soldus, used even in the Lex Municipalis of Julius Caesar,
458. ¢.(C. I. L. i. 206. 114, 115), and admitted by Horace into his Satires (S. ii. 5.
65 metuentis reddere soldum, and 5S. i. 2. 113), in possum for pité-sum (ch. viii.
§ 97), and in ferme, for férime, Superl. of féré, if the corruption fert me of the
Palatine MSS. in Plaut. Trin. 319 be evidence of the spelling ferime in Plautus’
time. Plautus has never the form culmen, which appears to be a form proper
to the oblique cases, so that the declension was : Nom. céldimen, Gen. cdl(u)minis
(cf. Georges, Lex. Wortf. on later(é)culus, and possibly fer(4)culum). In Vulgar Latin
we have slave-names like Marpor (C. I. L. i. 1076), Naepori (Dat. Sg.) (i. 1539 @),
of which full forms like Quintipor, Marcipor, Gaipor, are given by Festus (340.
17 Th.), mattus for médditus, ‘drunk’ (Petron.), virdis (cf. Probi App. 199. 9 viridis
non ‘virdis’) ; on vir(i)desco, vir(t)darium, see Georges, Lex. Wortf. s.vv.; dictus
for digitus (see Georges); fridam for frigidam on an inscription of Pompeii
(¢. I. L. iv. 1291) (ef. Probi App. 198. 3 K. calida non ‘calda’; frigida non
‘frigda’; infrigdo for infrigido, Oribas. fragm. Bern. iv. 34. p. 1. 6 and 10 Hag.);
cf. frigdor (see Georges, Lex. Wortf. s.v.) ; calda is readin Cato, R. R. vi. 1 and 75,
Varro, R. R. i. 13, &e., and the proper name Cald(us) is found on coins as
early as 109 B.c. (C.J. L. i. 382); on domnus for déminus, see Georges 8. v., and
cf. the proper name Domnus, Domna, Gk. Aopvos (C. I. G. i. 6505, end of second
cent. a.D.), and Vulg.-Lat. nit(i)dus, horr(i)dus, rig(i)dus, col(a)phus, &c. are
indicated by the Romance forms, e. g. Ital. netto, ordo (but with close initial
0), reddo, colpo, &. (For a list of syncopated fornis in late inscriptions and
MSS., see Schuchardt, Vok. ii. pp. 394 sqq.)

§ 18. Change of Unaccented Vowels. In a language with
a stress-accent the unaccented vowels are liable to be obscured.
We see this in our own language, where the unaccented vowels
in words like ‘father,’ ‘ sister,’ have become what we call par
excellence ‘the obscure vowel,’ the vowel-sound of w in ‘but.’
We notice too a difference in this respect between Italian pro-
nunciation and our own; for an Italian pronounces the vowels
of the unaccented syllables more clearly, and does not slur them
to the same extent as we do. But in Italian also the same
tendency to weaken an unaccented vowel is present, though not
in so marked a degree. The unaccented vowel often fails to
preserve its individuality, and is open to influence from
a neighbouring consonant, 7, for example, changing a preceding
short vowel to e, 2 changing one to o. Thus Latin arbor, or
rather its oblique case-form arbdrem, &c. has become in Italian
albero ; Latin déilis has become debole. And in the pretonic
syllable of signore (Lat. sézzérem), midolla (Lat. médudia), the
186 THE LATIN LANGUAGE. (Chap. III.

unaccented vowel has become «7. Exactly the same thing
happened in Latin. In the last chapter we saw that a short
vowel in the syllable following the accented syllable remained
unsyncopated only when its syllable was long by position, or
when some other cause prevented syncope. But though un-
syncopated, it did not remain unaffected. Its quality was
changed. In a syllable long by position we see a short
unaccented vowel becoming ¢, e.g. remex, from rémus and dgo,
in other syllables i, e.g. remigis, jérigo (Plaut.), later jirgo.
Under the influence of a following labial consonant or / it
assumes a v- or d-sound (see ch. ii. § 16), e.g. occtipo, in-cipio, from
eipio; a following 7 makes it e, e.g. pépért, from pdrio. Some
vowels retained their individuality better than others. Short
0 in compounds of verbs like voco, rogo, &c. remains unchanged,
€.g. convoco, invoco, irrogo, arrogo; short win tu-tudi, &e.

Final syllables too cannot have been so liable to affection as
others, or the difference between Nominatives Singular of differ-
ent stems, such as cémis, dpus (Old Lat. opos), manus, &e. could
not have been so well maintained. Perhaps they were saved by
the analogy of trisyllables, and longer words, where the final
syllable was not in the weakest of all positions, viz. immediately
following the accent.

Even diphthongs were changed, their first element being
affected, ac becoming 7 (through *e7), aw becoming # (through
*eu), Just as single a was originally weakened to ¢ (see below).
Thus the compound of 0d and caedo’ became, under the influence
of the early accent, occido, of ob and claudo, occlido. But long
vowels were more resistive of change, e.g. invado, from vddo,
wrépo, from répo.

The regularity with which these changes of short vowels and
diphthongs are carried out in the second syllables of Latin words
is a strong proof of the fact discovered by Corssen, that the
Latin accent at some early time rested invariably on the first
syllable; for it is the syllable immediately following the
accented syllable, which in a language with stress-accent is most
liable to be affected. A syllable with a secondary accent, like
the paenultima of *pdrri-catda- (under the old accent law) would
not be liable, just as in the Romance languages the vowel of the
§ 18.] ACCENTUATION. VOWEL-WEAKENING. 187

first syllable of words like classical Latin drmatdra, &c. shows
the same treatment as the vowel of the syllable with the main
accent (cf. Ital. Fiorentino from Florentinus, like fiore from
fidrem, but Firenze from Floréntia). It might, however, change
its vowel after the analogy of kindred words where the same
vowel followed immediately on the accent, e.g. *éc-caido, and so
we get the Old Latin form paricidas (Paul. Fest. 278. 10 Th.).
On the other hand the analogy of the simple word with accented
root-vowel would often save the vowel of the compound from
being changed, e.g. vades et subvades, XII Tab., where the a of
vades is not weakened as it is in praevides (C. I. L. i. 200), later
praedes. And at any period in the language the sense of the
relation of a compound to a simple word might lead to the
restoration of the vowel in the compound to its accented quality,
e.g. *prévicare might become provdcare, though the noun praeco
(for *prae-vico) was left unchanged ; énico might become e-neco ;
consecro, con-sucro, This restoration of compounds to their
unweakened form, ‘Recomposition’ as it is sometimes called,
is a feature of the late Republican and the Imperial period, and
possibly had some connexion with the grammatical studies
imported from Greece towards the close of the Republic, and
prosecuted with great zest for many centuries.

In the period of the earlier literature the change of unaccented
vowels is more the rule than it is later, e.g. always enico in
Plautus, &c., in spite of the old practice of separating the pre-
position from its verb by tmesis, 0d vos sacro, for obsecro vos, sub
vos placo, for supplico vos.

Analogy, however, was at work in all periods, and exerted its
influence now in one way, now in another. The analogy of the
Nominative preserved from change the vowel in the oblique cases
of arborem, fulguris, &c., at least in the literary lauguage
(cf. Ital. albero ; fulgerator, Gruter. Jnscr. xxi. 3); the analogy
of the Oblique Cases, integri, integro, &e. has substituted e for
2 in the Nominative integer, Compounds, too, which were made
for the occasion, or were rarely used, like O. Lat. hosti-capas,
hostium captor (Paul. Fest. 73. 10 Th.), wrbi-capus (Plaut.), would
escape the change which befel a word established in use, like
prin-ceps, muni-ceps. But with thes: exceptions the change of
188 THE LATIN LANGUAGE. (Chap. III,

short vowels of the second syllable is very regular in Latin,
though the oldest inscription extant, Manios med fefaked Numa-
stot, ona brooch perhaps of the sixth cent. B. c. found at Praeneste,
is suggestive of an epoch when this law was not in operation.
The exact rules of change seem to be these, The older
representative of 2, the modification of a short vowel in an
ordinary short unaccented syllable was e (Gk. «); while the
older representative of «dé, the modification of an unaccented
short vowel before a labial or 7, was 0 (Gk. 0). # was replaced
by i, 0 by w about 230 B.c. Up to that time the process of
change might be so described. An unaccented short vowel was
changed before a labial 7 to o, in all other circumstances to e.
Thus on old Praenestine jewelcases, &c. we find spellings like
Belolai (C.L.L. i. 44) for Bellulae, Salutes (i. 49) for Salutis,
Aecetiat for Aequitiae (al. Angitiae), (i. 43); and these older
spellings often persist to a much later period. The MSS. of
Plautus, for example, preserve traces of adegit for abigit, Capt.
814; easolatum for eaulatum in Mere. 593 (B), Most. 597 (A),
&c., and the Lex Repetundarum of 121 B.c. (CL. ZL. i. 198) has,
with the conservativeness of legal orthography, forms like
detolerit, oppedeis side by side with detulerit, ediderit, &e. H was
especially long retained after the vowel i, e.g. ébrietas, pirietem.
And after consonantal i (y) we find conieciant on the Lex
Repetundarum, proiecitad (for projicito) on the Titulus Lucerinus
(Eph. Epigr. ii. 298); while the spelling zieciates, Plaut.
Truc. 298 has led to the corruption ¢//eciatis in the Ambrosian
Palimpsest (so in Lucretius MSS. traiécére, iii. 513. For other
examples, see Lachmann ad Luer. ii. 951); 0 was similarly re-
tained after 7, ¢, e.g. ftliolus, Piiteol, lit. ‘little wells, and after
vocalie or consonantal wu (w, w), e.g. parwolus. (See ch, iv. § 70.)
The 7é of compounds of jacio, &c. became 7, e.g. conicio,
through loss of accent (ch. iv. § 51), and similarly «é of compounds
of guatio, &e. became %, e.g. concutio. This older ¢ remained
in short syllables before 7, e.g. pepert. Also in syllables
long by position, except where the first of the two consonants
was a labial or 7; and even into these it found its way in
time with the exception of the combination of 7 with another
consonant (not //), e.g. con7emno, older condumno (both forms are
§ 18.] ACCENTUATION. VOWEL-WEAKENING. 189

found on the Lex Bantina of 130 B.c., 0.2. L.i. 197); surreptum
(surruptum Plaut.), but always insulto, insulsus, inculco (ch. iv. § 10).
The 0, proper to syllables whether short or long by position in
which the vowel was followed by a labial or /, became w, which
might pass into the #-sound (ch. ii. § 16), written at first w, later i.
The spelling of MSS. of Plautus, ¢estumonium, &c. became in
time testimonium, &c. In Superlatives i for earlier ~ was adopted
for State Inscriptions through the influence of Julius Caesar
(Quint. i. 7. 21; Varro ap. Cassiod. p. 150. 11 K.), sé on the Lex
Julia Municipalis of 45 3.c. (C.D. i. 206) maximam and maau-
mam, though we find it occasionally used long before his time,
eg. proxsimum (i. 1291, an inscription which Ritschl dates
‘not after 130 B.c.’). J came in earliest probably in syllables
which were followed by a syllable with 2 in hiatus, e.g. recipio
(recipit on a Scipio epitaph of c. 180 B.c., i. 33).

The same vowel appears in confringo, infringo, &c. in accordance
with the phonetic law of Latin which gives us 7 for e in the
accented syllables of words like tingo (Gk. réyyw), ch. iv. § 11.

The succession of 0, u, 7 in words like maxomos, maxumus,
maximus is also seen in the parasitic or ‘Svarabhaktic’ vowel
(ch. ii. § 154) of poculum (Plant. poclum), &e. The earliest spelling
is 0, e.g. on the Praenestine vases of third cent. B.c. belolai pocolom
(C. I. L.i. 44), Salutes pocolom (i. 49), Aisclapi pococolom (for poco-
lom), (Eph. Epigr.i. 5). The classical Latin spelling is ~, poculum,
stibulum, &c. The 7 in J-stem Adjectives, &c., e.g. stabilis, dgilis,

facilis, where i follows in the next syllable, is in O. Lat. ¢, e.g.
fameliai (C. I. L. i. 166), on Greek inscriptions KauxeAuos, &e.

An o which had escaped the reduction to e became at the end of
the third cent. B.c. w, e.g. dpus, earlier opos (2b. 1. 52), Licius,
earlier Luciom (ib. i. 32) (cf. ch. iv. § 17); imdustrius (older endos-
truo-, if we may believe Paul. Fest. 75.28 Th.) ; -w¢ in 3 Pl. for
older -ont, e. g. praedopiont of Carm. Sal. (Fest. 244.13 Th. MS.
-oti-) in the sense of pracoptant, Au became w, 7, e.g. sdtura, satira.

The weakening of the diphthong ai (later ae) to 7 was fre-
quently abandoned in the late Republican and Imperial time,
a number of forms which exhibit this weakening, e.g. consiptum,
obsiptum, from saepio, being recognized as Old Latin forms.
The same weakening may have occurred when a: stood in hiatus,
190 THE LATIN LANGUAGE. (Chap. ITT.

but here by the Latin law of shortening a long vowel in hiatus
(ch. ii. § 143), 7 sank further to %, e.g. Bovianum for Bovianum
(Oscan Bivaiantd Abl.), Marius (cf. Osean Maraiio-). Similarly
unaccented aw in hiatus sank to % in éluo, eluderum (Cato) from
liwere (Old Lat.), Zavderum. (For other examples, see Parodi in
Stud. Ital. i. 385.) (For reduction of final vowels, see § 37.)

Greek loanwords in Latin show the same changes of the post-
tonic vowels, though a vowel may be retained unchanged in
words which were borrowed after the operation of the law
affecting that particular vowel, or which never became part and
parcel of the common language. The change is seen in bélinewm
(Plaut. &c.), classical alnewm (Badavetov), triitina (rputdvn),
tilentum (tédavrov), phiilerae (pddapa), &e., but not in pldtanus
(mAdravos), barbarus (BdpBapos), &e. Vulgar Latin citera (x.Odpa),
Probi App. 197. 26 K.), Ital. cetera and cetra, but Span. gui-
tarra from citdra (xOdpa) (as from xdypapos Ital. gambero, but
Span. gambaro), carry out the vowel-reduction which was
omitted in the classical forms of these words. The analogy too
of native words may often have interfered with the normal
development of these unaccented vowels ; the 6 of ancora
(dyxipa) and the e of plicenta (mraxobvta, Acc. Sg.), for example,
may have arisen in this way, just as zpdé0upov became protulum
(Liwe, Prodr. p. 376) by the analogy of diminutives, or Iepce-
govn, Proserpina (Prosepnai, Dative, on an old mirror of Cosa,
C.L. L. 1. 57) by the analogy of proserpo.

Under the early law of accentuation, when the accent fell on
the first syllable of every word, pretonic change could take place
only in proclitic or subordinate words like prepositions preceding
their nouns. Whether Menerua of early inscriptions (e.g. C. LL.
1 191 Meneruai; cf. Quint. i. 4. 17), a quadrisyllable in Plautus
(ch. iv. § 148), became Mixerva through loss of accent in the first
syllable or by analogy of minor is uncertain. But the pretonic
change of au to wu in Ital. udire (Lat. audire), uccello (Vulg.
Lat. *aucellus from avis), and of ae to z in Ital, cimento (Lat.
caementum), cisello, our ‘chisel’ (Lat. caese//wm), may have
already occurred in Vulgar Latin. The pretonic syllable is
often assimilated to the accented, e.g. momordi for earlier
memordi, and the same tendency in the post-tonic syllable
§ 19.] ACCENTUATION. VOWEL-WEAKENING. 191

is seen in mispronunciations like ¢onotru (Probi Append. 198.
32 K.), preventing reduction in dlicer, hébétem, &e.

A long vowel in an unaccented syllable was not shortened
(except in the final syllable, see §§ 40-50 infr.) until a late period,
when the length of all long syllables had been reduced to some-
thing not far removed from a short syllable (see ch. ii. § 141).

But a syllable long by position, when preceded by a short
syllable and followed immediately by the accented syllable, was
so reduced as to be often scanned as a short syllable by the
early dramatists, e.g. voliiptatem, senéctutem (Plaut. Ter.).

In Oscan and Umbrian, though syncope is of frequent occur-
rence, the quality of an unsyncopated unaccented vowel is
retained in the spelling. The name, for instance, of the Latin
poet, Propertius, who was a native of Umbria, is in Umbrian
form Propartio-, not Propertio- (Vois. Ner. Propartie on an
Umbrian inscription, C.L Z. xi. 5389, would be in Latin Vols.
Propertii, Neronis f.; cf. xi. 5518 sqq.).

§ 19. Other Examples. I. Syllables long by position. Anteclassical
exercirent from sarcio, Ter. Hewut. 143 (e in all the MSS.; cf. Paul. Fest. 57. 12 Th.
exercirent : sarcirent) ; ommentans, from manto, Frequentative of mdneo, quoted
by Fest. (218. 14 Th.) from Livius Andronicus (cf. Gl. Plac. ommentat:
expectat, &c.); inpetritum: inpetratum (Paul. Fest. 77. 3 Th.); inermat:
armis spoliat (id. 78. 28 Th.); iniéx: inductor, ab inliciendo (id. 80. 29 Th.,
with quotation of Plaut. Asin. 221) from O. Lat. licio (id. 83. 36 Th., lacit:
inducit in fraudem. Inde est ‘allicere’ et ‘lacessere’; inde ‘lactat,’ ‘illectat,’
‘ oblectat,’ ‘delectat.’ Cf. 83. 14 Th. lacit: decipiendo inducit. ‘Lax’ etenim
fraus est) ; procestria (id. 282, 6 Th.), apparently from castra, seems to be the
word equated with Gk. mpodorea in the ‘Philoxenus’ and ‘Cyrillus’ Glossaries ;
compectus is in Plautus the Participle of the compound of pédciscor, compactus
(ch. ii. § 144) of compingo. We have ¢ before a labial with a consonant in incepsit,
the old ‘Perf. Subj.’ of incipio (Paul. Fest. 76. 23 Th.); peremne dicitur
auspicari, qui amnem, aut aquam, quae ex sacro oritur, auspicato transit
(Fest. 316. 32 Th.) ; indeptare: consequi (Paul. Fest. 75. 27 Th.). The gloss
indepisci: adsequi, adipisci, on the same page, J. 31 (ef. Gloss. Plac.) is
perhaps given more correctly in the ‘Philoxenus’ Glossary, indepti: dvicav-
Tes; praeceplat: saepe praecipit Carm. Sal. (Fest. 244. 10 Th.); inebrae aves :
quae in auguriis aliquid fieri prohibent (id. 78. 7 Th.). But enubro: inhi-
benti (id. 54. 7 Th.) Cf. the questionable spellings in the ‘Philoxenus’
Glossary, eniber, enibra, enibrum (for enub-?). On the Falisco-Latin inscription
of the Falisean ‘collegium cocorum’ in Sardinia (Zv. I. I. I. 72), an inscription
with bad spelling and worse metre, we have aciptum for acceptum in the first
line: Gonlegium quod est aciptum aetatei agedai.

Classical examples are génetrix beside genitus ; obstetrix beside stator, constituo ;
(but praditrix, &e., influenced by proditor) ; fulgétrum (all these Neuters in -trum
192 THE LATIN LANGUAGE. [Chap. ITI.

and Fems, in -tra have 2, except a few with d, e.g. verttrum, mulctira, ardtrum. But
tonitru: see A.L.L. i. 111); td-ent-idem from ante; expers from pars; ptrennis
from annus, and imberbis from barba, with the usual I-stem of Compound
Adjectives ; incestus from castus ; forceps from formus, ‘warm,’ and cdpio ; compesco
from *péic-sco (ef. péciscor). So in Reduplicated Perfects, e.g. peperci from
pareo ; fefelli from fallo, And in Final Syllables like miles for *milets, *milit-s (in
Plaut. the last syllable of such words is long by position, ch. ii. § 137) ; cornicen
for *cornicens, *cornu-can-s. An original o becomes e(¢) in triginta for *trigenta (Gk.
Tpidxovra) (on ¢ for e before nt, ef. ch. ii. § 147), ile from unaccented olle (ch. vii.
§ 13), and perhaps piren-die (cf. Osc. perum) (on -undo- and -endo in the Gerund,
-wnt- and -ent- in the Pres. Part., see chap. viii.). An original « becomes e in
con-sternari (cf. Gk. mripoua, O. H. G. stornem); an original i perhaps in
O. Lat. magester Quint. i. 4.17). Other examples of the variation of weakened
and unweakened forms are: comperco and comparco, contrecto and contracto, aspergo
and aspargo, dmando and amendo, dispertio better than dispartio, bipartitus and
bipertitus, quinquepertitus and quinquepartitus, retracto better than retrecto, conspergo
and conspargo, céiliandrum and caliendrum, attrecto and attracto (so perhaps Sarepta
and Sarapta), on which see Georges, Lex. Wortf. s.vv.; ef. abarcet Paul. Fest.
11. 36, abercet id. 19. 26 Th. On Greek inscriptions we have mpiveuw, pavenp,
BigtAdaptos, ovrtpavos, &c., from the end of the first cent. a.p. ; see Eckinger ;
prae-fiscini is usually derived from fascinum, but neither exintero beside exentero,
nor bipinnis beside bipennis are certain cases of the change of e to i, nor yet
Antistius beside Antestius (§ 39). (On i for e in infringo, triginta, see ch. iv. § 11.)

§ 20. II. Short Syllables (1) in-r. The compound of léx and rumpo has in
Plautus the spelling legeriipa (e.g. Pers. 68, corrupted to lege rumpam), cf.
viveradix, Cato, R. R. xxxiii. 3), though at a later time the usual ‘ Composition-
Vowel’ i was used, e. g. pinnirdpus, Juvenal (see Rev. Phil. 1892, p. 109); from
paro come aequipero, impero, pauper, but dpi-parus ; jiniperus (and junipirus) (see
Brambach, Lat. Orth. p. 142), derived by Verrius Flaccus from jitvénis and pirus
(‘Serv.’ ad Eel. vii. 53); sdcer, soceri may be the direct development of *swékiiros,
Gk. éxupés, Skt. Svaéuras, but see § 15, K. Z. xxxii. 564) ; cineris, cineri, but cinis,
cinisculus (cf. Georges, Lex. Wortf. s.vv. Silerus, mataris, Samiramis, and for
plebeian spellings like Caeserem, see Schuchardt, Vok. i. 195, ii. 214). [The
late spelling facinerosus is capable of being explained, like temperi Adv. beside
temport Dat., by the variation of the suffix -os- and -es- in the Declension of
these Neuter stems (ch. v. § 71); ef. pignera for pignora (see Georges) ].

§ 21. (2) in -1 or Labial. Anteclassical : consoluerunt and cosoleretur on the
8. C. de Bacchanalibus of 186 B.c. (C.L.L. i. 196. But consuluere i. 185
beside consoltu i. 186 on two old inscriptions of Venusia); the MSS. of Plautus
show exsolatum, Merc. 593 (B), exolatum, Most. 597 (A), &c. (see Brix ad Trin. 535) ;
consol on two inscriptions of 211 B.c. (i. 530-1) on another of 200 B.c. (Not.
Scav. 1887, p. 195), and so normally till the third Punic War, even in one of
71 B.0., consolibus beside consulibus (C. I. L. i. 204); exsoles is the Old Latin form
(Cornutus ap. Cassiod. p. 152. 7 K.; Caesellius ap. eund. p. 204. 2 K.), while
Velius Longus says, ‘consol’ seribebatur per 0, cum legeretur per u (p. 49.
14 K.); incolomis is the spelling of the best MSS. (B, C) in Plaut. Truc. 168 (ef.
colomnas C. I. L. i, 1307). Oremains in the classical period in vinolentus (perhaps by
analogy of vind lentus), somnolentus, and sangutnolentus. For the Superlative suffix
we have the oldest spelling o in the proper name Maxomo in an inscription in
the Faliscan dialect (Zv. I. I. I.60 Maxomo Iuneo he cupat, ‘Maximus Junius hice
§§ 20,21.] ACCENTUATION. VOWEL-WEAKENING. 193

cubat’) (cf. Gk. Aexo[pos C,I. A. iii.61. A (3). 18, end of first cent. a. D. (?) ; maxu-
mus, &., as was said above (§ 18), is the usual spelling on inscriptions till the
time of Julius Caesar, though maximus, &e. is occasionally found much earlier.
The spelling of Plautus has u in words like magnujice, Pseud. Jo2 (A.); pultufagis,
Most. 828 (A.); sociufraude, Pseud. 362 (A.) ; sacruficem, Pseud. 327 (both A, the
Ambrosian Palimpsest, and the Palatine family of MSS.); carnufex, &c. (see
Index to Studemund’s Apograph of A, p. 522). So Oinumama for Unimamma, an
Amazon, on an old Praenestine cista (C.I.L. i, 1501) ; testumonium on the Lex
Bantina of 133-118 B.c. (i. 197); Cornusicia on an inscription (i. 1087), which
Ritschl dates ‘not long after Caesar’ (cf. Gk. Kopvogutos, e.g. C. I. G. 6948),
tubulustrium (Varro), but aedisicandam 108 B.o. (C. I. L. i. 565 and Eph. Epigr. viii.
460), opiparum on the old Falisco-Latin inscription with aciptum (Zv. I. I. I. 72),
vadimonium and aedificium on the Lex Agraria of 111 B.c. (C. I. L. i. 200) ; testi
monium on the Lex Repetundarum of 121 B.¢. (i. 198), &c. Manufestus is the
anteclass., mantfestus the classical spelling (Georges, Lex. Wortf.s.v.). [For other
examples see Georges s.vv. Hadrumetum, quadrupes, septu(m)ennis, septu(m)aginta,
crassupes (Gk. Kpascorns, Bull. vi. p. 280, of the Republican period), manupretium,
maritumus, incolumis, coluber, marsuppium, monumentum, cornupeta, aurufex, existumo,
lacruma mucculentus, recupero, ustulo, acupenser, sterculinum, intubus, sescuplex,
victuma, pontufex]. The influence of a following syllable with i (especially in
hiatus) is shown in familia (O. Lat. famelia, § 18) beside famulus; subrimti
haedi, from rumis, mamma (Paul. Fest. 369. 8 Th.) beside subrumari (Fest.
442. 32 Th.); moinicipieis beside mancup(um) on the Lex Agraria of 111 B.C.
(C. I. L. i, 200); manibieis, Eph. Epigr, i. p. 215 (but manubies, ib. viii. 476, on a
Capua inscription ¢. 135 B.c.); surripias is the spelling of both families of MSS.
of Plautus in Pseud. 876, surripere in Pseud. 290, 675, surripitur in Mil. 602, but
with u in the next syllable surrupui, surrupuisse seem to be the Plautine forms
(also surruptus) ; we have recipit on a Scipio Epitaph of c. 180 B. co. (C.I.L.i. 33),
accipito and conciliwum on i, 197 of 133-118 B.C., accipito, conciliatum, conciliaboleis
on i. 198 of 123-122 B.c.; acipiant on i. 199 of 117 B.C.; only the é-form is
quoted of inipitus: implicatus vel inretitus (Gl. Plac.), from root ap- (cf. aptus).
[For inépite : inpetum facite (Paul. Fest. 78. 5 Th.), see below]. In Gk. inscrip-
tions Serroupios is very rare; we find almost always Zerripuos. (See also Georges
on Lanivium).

Classical: u remains in contubernium (but adtibernalis, Paul. Fest. 9. 9 Th.)
from taberna; nuncupo, occupo from capio, occulo, &c., and was retained in the
spelling of Dat. and Abl, Plur. of some U-stems to distinguish them from
similar I- or Cons.-stems, e.g. artubus, but according to the second-cent,
grammarian in the spelling only (Ter. Scaur. p. 25. 11 K. nemo autem tam
insulse per u ‘artubus’ dixerit) (trebibos on an old inscription in the British
Museum, Eph. Epigr. ii. 299) ; dissupo is the anteclassical, dissipo the classical
spelling (Georges, Lex. Wortf. s.v.); so victuma and victima (ib, s.v.) 3 monu-
mentum and monimentum were both used, e.g. monimentu (C.I.L. i. 1258, ‘not
after 130 38.0.’ Ritschl), while monementum and monomentum are incorrect
spellings (Georg. s.v.). Dicumentum, &c. but specimen, &c. by Assimilation,

The Parasitic Vowel. Anteclassical: piacolom, the old spelling according
to Mar. Victorinus (p. 11. 14 K. ut apparet ex libris antiquis foederum et
legum, qui etiamsi frequenti transcriptione aliquid mutarunt, tamen retinent
antiquitatem... pro ‘piaculum’ ibi ‘ piacolom.’), is on a law of 58 B.0. (C. 2. L. i.
603), viacul- (piaclum on the Spoletium inscription, xi. 4766), but the ancient o

0
194 THE LATIN LANGUAGE. [Chap. III.

remains in Plautine spellings like aemolos, Acc. Pl., Pseud. 196 (A.); epolonos
dicebant antiqui quos nunc epulones dicimus (Paul. Fest. 55. 15 Th.) ; agolum :
pastorale baculum, quo pecudes aguntur (Paul. Fest. a1. 37 Th.); Tuscolana,
C.I.L. i, 1200; tabolam on S. C. Bacch. of 186 B.c. (i. 196); taboleis, popolum
(beside popul(o)) on Lex Bantina of 133-118 B.. (i. 197); singolos, taboleis (and
tabula), conciliaboleis on the Lex Repetundarum of 123-122 8B. c. (i. 198); singolos
(but vinculeis) on the Sententia Minuciorum of 117 8B.c. (i. 199) ; tabolam, singolis
oni. 208, an inscription referred by Ritschl to about the time of the Lex Agraria
(i. 200, which however has only tabula, tableis, singula, trientabule(is)), Viz. 111 B.C. ;
angolaria (but opercula), on the (restored) Lex Parieti Faciendo of 105 8.c. (i.
577), 80 that the old spelling does not seem to have died out till the end of the
second cent. B.c. (On Greek inscriptions we have Aevrodos (first cent. B.c.),
Aevredos (c. 140 B.c.), Aevrudos (first cent. A. D.), but usually AevrAos (cf. ch. ii.
§ 102); the Gk. loanword drachma is in the earlier writers drac(h)uma; for
spellings like vigulum, vigulo, vigelia, titelus, sibelo, sepulivit, see Georges. )

§ 22. in other short syllables. Anteclassical: accédo (for accido) is pre-
served by the MSS. in Enn. Trag. 77. 206 R.; Luer. ii. 1025, v. 609 and
elsewhere (see Ribbeck, Prolegom. Verg. p..416) ; so timedus in Naev. Com. 35 R.;
acetare dicebant, quod nune dicimus agere (Paul. Fest. 17. 30 Th.). Similarly
e is retained without weakening in spellings of the oldest MSS. of Plautus like
detenet, Pers. 505, contenuum, Stich. 214, contenuo 623, &c. (so the corruption ad te
alienent, Pers. 497, points to attenent not aitinent) ; in the MSS. of Poen. 266 proseda
(cf. Paul. Fest. 282. 16 Th. prosedas meretrices Plautus appellat), optenui on
a Scipio Epitaph of ¢. 130 B. c. (C. I. L. i. 38) ; conregione in the augur’s formula
(Varro, L. L. vii. 8; Paul. 46. 24 Th.) ; promeneruat, promonet, Carm. Sal. (Fest.
244. 12 Th.) ; cf. mereto(d) on a Scipio Epitaph of c. 215 B.c. (C. 0. L. i. 32) (but
meritod i. 190, ‘early part of the sixth cent. a.v.c.’), and even on a recent
inseription (i. ro12). This ein Old Latin spellings often appears for I.-Eur. jin
syllables unaccented under the Old or the Paenultima Law, e. g. aidiles Nom.
Sg. on a Scipio Epitaph of c. 250 B.c. (i. 31); Fabrecio (i. 106) ; Tempestatebus
on a Scipio Epitaph of ¢. 215 B. c. (i. 32) ; Lepareses for Liparenses (Gk. Auwépaz),
(quoted probably from Ennius by Paul. Fest. 87. 6 Th.), and Greek inscrip-
tions often retain the older orthography, e. g. coueriov (usually), KaieAtos (till
c. 50 B. c., then KaietAros), KarerwaAcov (usually, Kamer- not till first cent. a. v.),
Aeredos (in Rep., but Aemdos in Empire), Aoperros and Aoptios. But the
weakening to 7 is old, as is seen from dimidius, which must have changed etoi
at a time when the accent rested on the first syllable ; conjice on an old
Praenestine cista of third cent. B.c. (Mél. Arch. 1890, p. 303); subigit and
opsides on a Scipio Epitaph of ¢. 200 8B. c. (i. 30) ; habitarent oppidum, possidere
on the Decree of L. Aemilius Paulus, 189 B.c. (ii. 5041) ; obstinet, dicebant
antiqui, quod nunc est ostendit, ut in veteribus carminibus, &c. (Fest. 228,
6 Th.); prospices, prospice, Carm. Sal. (Fest. 244. 13 Th.) ; énico is the old
spelling, later eneco (Georges, Lex. Wortf. s.v.); prosicium, quod praesecatum
projicitur (Paul. Fest. 282. 13 Th., ef. prosiciae, Gl. Philox.); exsicas from ex
and seco, Plaut. Rud, 122; obigitat antiqui dicebant pro ante agitat (Fest. 214.
2Th.); jurigo, later jirgo (cf. jargium) ; gallicintum from cano, by analogy of
which was formed conticinium (cf. Gl. Plac. p. 58. 24 G. conticinio: tempore noctis
post galli cantum quando cecinit et conticuit) (see Goetz, praef. in Plaut. Asin.
xxv). Classical : Juppiter from pater ; sistite (cf. Gk. torare); compitum, explained
by Varro, ‘ubi viae competunt’ (Z. L. 6. 25) (cf. propitius) ; dimico from maco
§§ 22, 23.] ACCENTUATION. VOWEL-WEAKENING. 195

(cf. macto) ; of the rare weakening of o we have examples in Compounds like
himicida, armiger, &c. for the Composition-Vowel, which is 4 in other languages,
is tin Latin (see ch. v. § 83) ; inquilinus beside incola? Of %, examples are cornicen
(Gk. xopyoxdapios, kopvoverapios and xopvikAapios) 3 stipercilium (ef. Gk. xddAa,
Plur.) ; both inclutus (incluto in all the MSS. of Plaut. Pers. 251) and inclitus are
attested spellings ; (cf. arbita, not arbuta, in the MSS. of Lucretius, v. 941 and
965). For other examples of e-d, see Georges, Lex. Wortf., s. vv. eligo, compitum,
tremebundus, caeremonia, fenisictum, cervesia, ploxenum, subsicivus, quatenus, internecio,
brotinus, seneca, querimonia, intellegentia, neglego, interimo, also for late and plebeian
spellings like segitis, patena, tredecem, decim. (On late adjecentia see Schuchardt,
Vok. i, 193.) The change of e to i in syllables long by position is claimed for
braefiscini génista, &e, (on these see ch. ii. § 12), certainly with right in im-
fringo, &e. (see ch. iv. § 11), before a consonant-group like ng (so tingo for
*tengo, Gk. réyyw). Assimilation saves the vowel in segetem, teretem, &c.

§ 23. (3) Diphthongs, ai, ae. Anteclassical : distisum et pertiswm dicebant,
quod nunc ‘distaesum ’ et ‘pertaesum’ (Paul. Fest. 51. 25 Th., ef. 271. 2 Th.).
Festus, 372. 7 Th. tells us that Scipio Africanus Minor was twitted for his
use of pertisus by Lucilius :

Quo facetior videare, et scire plus quam caeteri
‘Pertisum’ hominem, non pertaesum, dices.

Lucilius was right, for compounds with intensive per are Separable Compounds
like bene-facio, sat-ago (see below) ; pertaesum is the spelling on the Claudius tablet
at Lyons ; consiptus was used by Ennius, according to Paul. Fest. 43. 37 Th. (ef.
45. 15), and an example is quoted by Non. 183. 14 M. s. v. venor: teneor consipta,
undique uenor (Enn. Trag. 254 R.) ; adsipere et praesipere dicebant antiqui, sicut
nos quoque modo dicimus ab aequo ‘iniquum,’ ab quaerendo ‘inquirere’ (Paul.
Fest. 16. 9 Th.) ; obsipiam, quoted from Caecilius by Diomedes (p. 383. 10 K.
quod vulgo ‘obsepio’ dicimus veteres ‘obsipio’ dixerunt. Caecilius, &c.)
(Com. 65 R.); praecidaneam poream dicebant, quam immolare erant soliti
antequam novam frugem praeciderent (Paul. Fest. 273. 5 Th.). (Gellius, iv. 6
discusses this word and its cognate succidaneae, which, he says, was sometimes
mispronounced in his time succidaneae: succidaneae nominatae, littera i scilicet
tractim pronuntiata; audio enim quosdam eam litteram in hac voce corripere) ;
occisit is quoted from the Laws of Numa by Festus (194. 21 Th.) ; so decidito in
XII Tab., inceideretis on 8. C. Bacch. of 186 8. c. (C.L I. i. 196, 27). But exquaereis
quoted by Priscian (i. p. 38 H.) from Plaut. Aul. 800, and the MSS. of Plautus
often show this spelling of the word (see Ritschl, Opuse. iv. p. 141) (so defaecato,
Aul. 79, but deficatam, Most. 158 are the likely spellings) ; conquaeri, conquae-
siuerit, exaestumaverit occur on the Lex Repetundarum of 123-122 B.c. (C.I.L.
i. 198), while on the Edictum Popillianum of 132 8B. c. (i. 551) we have the
curious spelling conguaeisiuei (cf. i. 547, an inscription of 141 or 116 B.c., with
Caeicilius (and consulto), while a similar inscription, i. 548, has the older spelling
Cajcilius (and consolto, § 26)]. Later, the retention of ae became the rule, e. g.
opsaeptum on the Lex Col. Jul. Urbanorum of 44 8. ¢. (Eph. Epigr. ii. p. 105) ;
lapicaedinis on the Lex Metalli Vipascensis of the first cent. a. D. (Eph. Epigr.
iii. p. 166) ; we have usually fabri subaediani on inscriptions (C. I. L. x. 6699. 5 ;
vi. 9559. 8, &c.), or subediani (Vi. 9558. 7; Vili. 10523. 5) (of which last, subidiani
on ii, 2211. 7, seems to be a misspelling). The weakened forms are used in
the classical period in the compounds of quaero, in existimo (ef. Mar. Victor.

O02
196 THE LATIN LANGUAGE. [Chap. III.

p. 22. 6 K. quid enim facietis in his quae, velitis nolitis, et scribenda sunt et
legenda ut scripta sunt, ut exempli gratia ‘existimo’ non ‘exaestimo’), in
Sastidium for *fasti-tidium (§ 13, p. 176), &e.

au; offucare aquam: in fauces obsorbendam dare (Paul. Fest. 223. 8 Th.) ;
defrudo seems to be the spelling of Plautus and Terence (Ritschl, Parerg. Plaut.
P. 540) ; accuso, incuso, &e. from causa are classical forms. The & of the
compounds of claudo was in time adopted in the simple verb too, cludo by
analogy of recludo, &e. (of Ital. chiudo) (see Seelmann in Git. Gel. Anz. Aug.
15, 1890) (cf. sed frude § 64, beside sed fraude § 69, on the Lex Repetundarum,
C.I. L.i, 198). The 3d of explodo, &c. is not due to the loss of accent, but is a by-
form of au found in the simple verb. (Diom. p. 382. 26 K. plaudo frequens
est, apud veteres plodo; then after quoting the form ploderent from Cicero,
he adds, secundum eam consuetudinem qua ‘au’ syllaba cum ‘o’ littera com-
mercium habet, ut cum dicimus ‘claustra’ et ‘clostra,’ item ‘caudam’ et
‘eodam’ et similia), just as oe (older oi) and % are byforms, e.g. commiinis,
imminis, comornem in §.C. Bacch. (¢.I.L. i. 196), immoenis (Plaut.). Oboedio
from audio is difficult to explain. (See also Georges, Lex. Worlf. s. vv. dissaeptum,
exquiro, existimo, and Brambach, Orth. on pertaesus, lapicidinae.)

§ 24, (4) Diphthongs in Hiatus. (On these see also ch. iv.) The u-diphthong
is retained in db-avus, dt-avus, trit-avus (O. Lat. strit-avus) but becomes % in aluo
beside lavo, eruum (ervum) from *ereg}- (Gk. épéBivO0s), dénuo for de novo, &e. The
Greek’ Ayar(F)of became Lat. Achivi (through *Acheiv-); Gk. €dat(F Joy, olivo-, older
oleivo-, which became when -om was weakened to -um *olei(v)um Nom., *oleivi
Gen. &c. (see ch. ii. § 53), whence olewm (for ey before a vowel loses y, e. g. eo, ‘I
go,’ for *éy6, ch. iv. § 63), and olivum, olet and olivi, &c. (like dei(v)us, deiviy whence
deus and divus, det and divi, ch. iv. § 70); Gk. fouydala, a Thracian claymore,
became rumpia (Enn. Ann. xiv. fr. 8 M.; Liv. xxxi. 39. rr). In cloaca for
clovaca, the » has been dropped, as usual, before the accented vowel (ch. ii.
§ 53), while nécivus and nocuus are different formations (ch. v. § 7).

§ 25. (5) je andve. Oni, t%as a weak or unaccented form of yé, wé in Indo-
European see ch. iv. § 51. Whether the z of dbicio, &. should be explained
as a similar Latin weakening, or asa modification of -ji-, is an open question.
The % of concutio may also be compared with the use of Greek «vu for Lat. -qui-
(especially unaccented), e.g. Greek ’Axvaas for Lat. Aquila, Greek Kupevos for
Lat. Quirinus, see ch. ii. § 28. Cf. ancunulentae ‘unclean,’ (Paul. Fest. 8. 29 Th.),
and inguinare ; bigae is the reduction of bi-jugae (see Georges, Lex. Wort. s. v.
bijugus), quadrigae of quadri-jugae ; abiviohas the first syllable short in the old
dramatic poets (cf. ch. ii. § 48, p. 45).

§ 26. (6) Later change of o to u, u toi, i. In syllables long by position
this is the usual development of original 0, for cases like tri-gint&é with e(i)
for unaccented o (cf. Greek zpidxovra) are rare, e.g. vétustus from I.-Eur.
*wetos (Greek éros, ‘a year’) and similar derivative TO-stems from Neuter
S-stems, férundus and similar Gerundial DO-stems from third Conjugation
Verbs, viluntas and the like formations [that finestus, ferendus, ferentarius (cf.
Osc. Herentat-, the Oscan Venus), &. show a weakening of o to e, and not
rather a bystem funes-, ferend-, ferent- cannot be proved ; cf. ch. viii. §§ 89, 94, and
see above, § 20; ef. lugubris (-os) and funebris (-es)]. Similarly in final syllables
long by position we have -unt in g Pl. of Verbs for O. Lat. -ont, e.g. nequinont
(Liv. Andron.), cosentiont (Scipio Epitaph) (ch. viii. § 73). The change of 0 to u
§§ 24-27.] ACCENTUATION. VOWEL-WEAKENING. 197

in unaccented syllables is further discussed in ch. iv. § 20. Diipundius (and
dupondius ; see Georges, Lex. Wortf. s.v.), promuntirium, are not good examples,
for before nd, nt we find even accented o becoming a u-sound, e.g. O. Lat. frundes,
Acheruntem (ch. ii. § 22). For examples of the change in syllables not long by
position, see Georges, Lex. Wortf. s. vv. formidolosus, adulescens, lemures, fulgurio,
bajulus, lautumiae, and cf. Brambach, Orth. on the misspellings pulenta, amulum,
Aequiculi and Georges on subules, eburis Gen., rigura Plur., vinulentus, sanguinu-
lentus, somnulentus, tripudo. But coralium (Greek xwp-) and curalium (Greek xoup-)
are not examples, nor mamphur (leg. mam/far) the ‘thong’ round a turner’s wheel
(Paul. Fest. ror. 1 Th.) (see Meyer-Liibke, Comm. Schweizer-Sidler, p. 24), and
O. Lat. colina is a doubtful form. Examples of u-i are inclutus, later inclitus, de-
Srutum and defritum (see Georges), arbutum and arbitum (Luer.), sdtura and satira.

§ 27. (7) Greek words with Vowel-change. a. Aleria ( AAadia in Herodo-
tus), a town in Corsica (cf. the Scipio Epitaph, c. 215 B.c. C.I.L. i. 32 hee cepit
Corsica Aleriaque urbe) ; tessera (réocapa); Agrigentum (’Axpdyavra Acc.), now
Girgenti; Tarentum (Tdpayta Acc.), now Taranto or Tardnto; Alixentrom
(AAééavSpor) on a Praenestine cista of third cent. B.o. (i. 59), and on another
(i. 1501) Alixente(r) (AAéEavSpos), Casenter(a) (Kaccdvbpa), Ateleta (‘Araddyrn) ;
Hecuba, O. Lat. Hecoba (Quint. i. 4. 16) (ExaBy) ; camera («apapa), also camara,
(the spelling approved by Verrius Flaccus, Charis. 58. 23 K.), which was
specially used in the sense of a decked boat (see Georges, Lex. Wortf. s. v.) 5
Camerina and Camarina (Kaudpiva), crdpula (xparmddn) (see Meyer, Rom. Gram.
i, pp. 32, 36); machina (unxavy, paxava) ; Cutina (Kardyy); scutula (oxvrddn) ;
strangulo (orpayyadaw).

e. catapulta (xaraméArys) ; scopulus (oxdmedos) ; tarpessita Plaut. (rpame(irns) ; pha-
retra (papérpa); Acheruns Plaut. (Axépmv) ; enocilis (Liwe, Prodr. p. 376) (€yxeAus).

t. dapsilis (BaysAnjs) ; cupressus (kumdpiocos).

o. amurca (audpyn) ; cothurnus («dopvos); epistula (émaroAn) ; also epistola (see
Georges, Lex. Wortf. s.v.); paenula (pawdrns) ; tribulus (rpiBodos) ; Patricoles
(Tlarpoxdos), the old form, used by Ennius (Trag. 314 R.; aline of Livius
Andronicus is quoted by Gellius, vi. 7. 11, with this name in the form Patroclus,
without any divergence in the MSS.) ; lautumiae. Avernus, popularly connected
with dopvos, and late Lat. averta (Greek dopr7) admit of other explanations.
(See Solmsen, Stud, Lat. Lautgesch. p. 23). On the spelling numisma (Gk. vépiopa)
see Keller ad Hor.Epp. ii. 1. 234, and on late Lat. zabulus for diabolus, Georges,
Lex. Wortf. 3. Vv.

v. arytaena, but artaena (artena) in Lucilius (dpirawva) ; incitega (eyyvOqen)
(Paul. Fest. 76. 3 Th. incitega: machinula, in qua constituebatur in convivio
vini amphora, de qua subinde deferrentur vina) ; mattea, ‘mincemeat’ (Varro,
L.L. v. 112) (parrtn). (The word appears in a curious military term
mattiobarbulus, used by Vegetius for a leaden bullet, or a soldier armed with
these, apparently for parrvo-ndpBodos, lit. ‘mincemeat-scattering.’ See
A. L. L, ¥. 135) 3 serpillum, (if trom Greek épmvddos), with s by analogy of serpo.

a, av. Achivi (Axatol) ; olivum and olewm (dato) ; oliva and olea (éAaia) ;
Centurum, Centaurum (G]. Plac. p. 54. 7 G.) (Kévravpor).

Parasitic Vowel. ‘HpaxAjjs is on Praenestine cistae and mirrors Hercle.. .
(GIL. xiv. 4105), Hercles (? Fercles) (C.I.L. i, 1500), Hercele Acc. (i. 56),
and on old Praenestine inscriptions (xiv. 2891-2) Hercole Dat. On a Roman
inseription of 217 B.c. (i. 1503) Hercolei Dat. So Hercolei (i. 1175), Hercoli
(i, 815), but Herculis Gen. on an inser. of 146 B.¢. (i. 541), classical Hercules,
198 THE LATIN LANGUAGE. (Chap. LIT.

mehercle (ef. Prise. i. p. 27. 13.H. Romanorum vetustissimi in multis dictioni-
bus loco ejus (u) o posuisse inveniuntur... ‘Hercolem’ pro ‘ Herculem’) ;
*AoxAnmés is Aiscolapio Dat. on an old inser. (Ann. Epigr. 1890, no. 85, but
Aisclapi, Eph. Epigr.i. 5), classical Aesculapius; ’AXcpyqvy is in Plautus Alcwmena ;
on techina, &c., musimo, see ch. ii. § 72.

§ 28. (8) Vowel unchanged. i. in Latin words. Anteclassical : incantassit
and excantassit of XII Tab. (ap. Plin. xxviii. 18), but ‘occentassint’ antiqui dice-
bant, quod nune convicium fecerint (Fest. 196. 12 Th.) ; ancaesa, dicta sunt
ab antiquis vasa, quae caelata appellamus (Paul. Fest. 15. 10 Th.), but Prise.
i. p. 29. 20 H. cites as instances of am-, ‘anfractus,’ ‘ancisus,’ ‘anquiro,’ and
Varro, L.L. vii. 43 explains ‘ancilia’: quod eaarma ab utraque parte,ut Thracum,
incisa ; perfacul antiqui, et per se ‘ facul’ dicebant, quod nune facile diximus
(Fest. 266. 20 Th.) is normal, for compounds with per- ‘very’ seem not to
change the vowel, e.g. persalsus (beside insulsus), persapiens (beside insipiens) :
perfacilis (beside dificilis), being what are called ‘Separable’ Compounds,
cf. per pol saepe peccas, Plaut. Cas. 370, per dpus est, Ter. Andr. 265 (so that
Lucilius was right in his objection to pertisum, see above) ; procapis progenies :
quae ab uno capite procedit (Paul. Fest. 281. 22 Th.) ; concapit tignum
XII Tab. (ap. Fest. 556. 27 Th. tignum iunctum aedibus uineaue et concapit
ne soluito); resparsum vinum (Paul. Fest. 353. 6 Th.) ; concapsit, conprehen-
derit (C. G.L. v. 182, 22). occanuere (3 Pl. Pft.) is quoted from Sallust’s
Histories by Priscian, i. p. 529. 5 K.

Classical : rédarguo, but ‘rederguo,’ was used by Scipio Africanus Minor
(Fest. 372. 7 Th. redarguisse per e litteram Scipio Africanus Pauli filius dieitur
enuntiasse, ut idem etiam ‘pertisum’); dlacris, but Vulg. Lat. alecer (so
in a glossary in MS. Vind. 482) (Ital. allegro, &c.) ; augurditus, augur were
formerly ‘augeratus,’ ‘auger’ according to Priscian, i. p. 27. 17 H.; impetus,
but ‘inipite,’ inpetum facite (Paul. Fest. 78.5 Th., apparently a corruption
for ‘impite,’ impetum fac), ‘compitum’; undecim, duodecim weaken the e¢ of
the final syllable but not of the paenultima ; incola, but O. Lat ‘inquilinus’
(eh. vi. § 10) ; inaequalis, but ‘iniquus,’ &. U remains in tiitudi (see ch. viii.
§ 39), pécudem, contumax, &e. (See also Georges, Lex. Wortf. s. vv. instauro, con-
quaestor, comparo, sepelio, &c.)

§ 29. ii. in Greek loanwords: amygdala (duuyddAn) (but Vulg. amiddula, Probi
Appendix 198.26 K.), artemo Lucil. (dprépwr) ; astraba, the title ofa play ascribed
to Plautus (dorpé8y) ; ballaena or balena, Plaut. &c. (pdAdawa) ; balanus, Plaut.
&c. (BéAavos) ; barathrum, Plaut. &c, (Bdpadpov); calamus, Plaut. &e. («ddapos),
apparently Vulg. Lat. *calmus (Ital. calmo and calamo, Fr, chaume) ; cantharus,
Plaut. &c. («dvOapos); cinaedus, Plaut. &c. (xivatdos); cottabus, Plaut. &c.
(xé7TTaBos) ; cymbalum, Luer. &e. (eduBadrov) ; dacdalus, Enn. &c. (Saidados) ; drapete,
Plaut. (Spanérns) ; gaunacum, Varro (xavvdcn); gausape, Lucil. &c. (yavodmns,
yatoaros) ; Hecata, Plaut. &e. (‘Exdrn) ; hilarus, Plaut. &c., later hilaris (fAapés) ;
Hiluria, Plaut., later Illyria (IAAvupia) ; lapathus, Lucil, (Adma@os) ; machaera.
Plaut. &c. (udxatpa) ; malacus, Naev., Plaut. &c. (udAaxos) ; margarita, Varro,
&e. (uapyapirns) ; metallum, Varro, &c. (uéradaAov) ; murena, Plaut. &c. (udpatva);
narcissus (vapkiacos) ; obrussa, ‘touchstone,’ Cic. (d8pd¢n, dBpugov) ; onager
(Gvarypos) ; palaestra, Plaut. &c. (madatorpa) ; petasus, Plaut. &c. (aéragos) ;
phalanga, Varro, &c.; pittacium (mrrdmov) ; ptisana, Varro, &c. (rriodv7) ;
raphanus, Cato, &c. (pdpavos) ; sesamum, Plaut. (ojcapor), but seswma, Plaut.
§§ 28-31.] ACCENTUATION. VOWEL-WEAKENING. 199

Poen. 326, sesima (see Georges s.v.) ; slomachus, stomachor, Ter. &e. (ordpaxos) ;
Tartarus, Tartarinus, Enn. &. (Téprapos), sometimes mispronounced ‘Tarterus’
(Consent. g92. 17 K.); thalamus (@dAapos) ; thesaurus, t(h)ensaurus, Plaut. &e.
(Snoavpés) ; tropaeum, Accius, &c. (rpdmaov) ; tympanum, Plaut, &e. (rUpmavor) ;
tyrannus, Plaut. &. (rdpavvos) ; paedagogus, Plaut. &. (wadaywyés).

§ 30. (9) Long vowels. None of the examples adduced to prove that long
unaccented vowels were sometimes changed are conclusive: déliro from lira,
‘afurrow,’ root leis- (0.81. léha, Lith. lysé, ‘a garden-bed,’0. H. G. wagan-leisa,
&c.), is the correct form, while deléro, as Varro (ap. Vel. Long. 73.2 K.) pointed
out, is due to confusion with Greek Anpeiv. Délinio (so spelt in all the MSS.
apparently of Plaut. Stich. 457), beside delenio, subtilis (but protlum, &e.),
suspicio, convicium, all with i in the following syllable, show the change to
which even accented @ is liable, e. g. Plintus (ch. iv. § 7). Occidamus, attributed
to Plautus, as an example of ob in composition, by the MSS. of Festus (196. 10
Th. occidamus Plautus ponit pro contra cedamus, cum plurimae aliae prae-
positiones familiariores huic verbo sint; cf. Paul. 197. 1 Th.) is clearly
a corruption for occédémus. For not only does Placidus’ Glossary of Plautus
(p. 89. 4 G.) give occedere: occurrere vel obviam cedere, but the MSS. (the
Palatine family) of Plautus read in the passage referred to by Festus, viz. Pseud.
250, Accedamus hac obviam, where the corruption accedamus points to an
original occedamus. Consiva, an epithet of the goddess Ops (Fest. 210. 26 Th.,
Varro, L. L. vi. 21) has been connected with constro, consévi. The examples of
unchanged é are numerous, such as the compounds of cédo, repo, célo, credo, crétus,
spéro, irrétio from réte, &e. For the change of @ to@ through want of accent
(for a similar change through influence of palatal j (y) in Vulg. Lat. Jénuartus,
&e., see ch. ii. § 3) the examples usually adduced are anhélus (cf. halo), and subtél
(cf. talus). But anhélus (spelt anellus in MSS. of Virgil; see Ribbeck’s Index)
has probably come from *an-énsios, the « of halo, from *édnslo (root an augmented
by s), having been changed to ¢ while its quantity was still short. The word
subtel quoted by Priscian (i. p. 147. 9 H.) as an instance of -¢l, and explained as
76 KotAov Tod Todds (what does he mean by hostis hostilis, subtel subtilis,
i, p. 131. 21 H.?) may similarly be due to a change of the short vowel in the
original form *sub-tax-lus (ef. taxillus) (or from tellus ?). None of the Compounds
of claémo, fama, fatus, clérus, pdreo, pax, placo, pravus, rado, vado, gnarus, gratus, labor,
mano, nitus,gndvus, &e. ever change the vowel. Profestus is a compound of féstus
(cf. feriae for *fesiae), not of fastus, fas (cf. nefastus). Nor dod, % change; witness
the Compounds of piéro, dono, flds, &e. Praestélor and praestilor come, the one from
praesto, the other from praestit (§ 15. 3). Pejéro and aéro (ef. conierat, coniurat,
C. G, L. iv. 322. 33) have not yet been thoroughly explained.

§ 31. (10) Recomposition and Analogy. In Vulgar Latin, as was mentioned
before (§ 11), the accent seems to have rested on the first syllable of the
verb in Compound Verbs, e. g. renégat, Ital. riniega, O. Fr. renie ; dimérat, Ital.
dimora (with close 0), Fr.demeure. The vowel of the simple verb usually
appears unchanged in the Compound, e.g. reddédit, Ital. rendiede, O. Fr.
rendiet. From the inscriptions of the Empire and the remarks of gram-
marians we see that the same ‘etymological’ treatment of Compound Verbs
was a feature of Imperial Latin. On the Latin Papyri of Herculaneum (first
cent, a. D.) the preposition of a Compound Verb, &c. is usually retained in
its simple form and not assimilated to the initial of the verb, or noun, e. g.
200 THE LATIN LANGUAGE. (Chap. ITI.

ad-siduo, ad-fini (Class. Rev. iv. 443), by a similar ‘re-composition’; and Velius
Longus (p. 62. 16 K.) mentions adluo, adliquor, adldbor as the forms in use at
his time, though Assimilation was the custom with other verbs, e. g. alligo (see
ch. iv. § 159). The same grammarian, in another passage, while he approves
of the pronunciation commendo, adds that the popular pronunciation was
commando (73. 10 K, quamvis ‘commendo’ dicamus, tamen ‘ commando’ in
consuetudine est.) (So amendo and amando. See Georges, Lex. Wortf. s. v.). And
his remark on the word comprimo shows the tendency of his time (first cent. A.D.)
to follow in these Compound Verbs the Analogy of the Simple Verb, or of
the Perfect Participle Passive (76. 9 K. ‘comprimo’ quoque per i malo scribi,
quamvis ‘compressus’ dicatur). (Cf. Mar. Vict. 10. 6 K. sacratum autem
in compositione ‘consecratum ’ facit per s et e, non per s et a, sic et castus
facit ‘incestum’ non ‘incastum’; Caper 110. 7K. ‘insipiens’ non ‘insapiens’ ;
Diom. 378. 30 K.; Prise. i. p. 437. 25 H.) The analogy of the Perf. Part.
Pass. (or was it Assimilation ?) brought e instead of i into the second syllable
of perpeti, depecisct, &¢., while the analogy of the simple verb is seen in spellings
on Imperial inscriptions like consacravit (C. I. L. vi. 3716, of 182 A. D.), consacravt
on the Mon. Ancyr. ii. 30; iv. 25) (for other examples see Seelmann, Ausspr.
p. 60). Often the two forms, the old with changed vowel and the new
popular form, are retained side by side, and are used by the grammarians to
express different shades of meaning. Thus Velius Longus (75. 6 K.) differen-
tiates aspergo the Verb, from aspargo the Noun ; Caper (100. 5 K.) protinus the
Adverb of time, from protenus the local Adverb. The i of the Oblique Cases
of levir, *laevir, ‘ brother-in-law’ (cf. Greek 8anp, I.-Eur. *daiwer-) and indeed of
the Nom. Sg. too, is due to the analogy of vir (cf. Non. 557. 6 M. levir dicitur
frater mariti, quasi laevus vir) ; of the inferior spelling génitrix, for genetrix
(see Georges, Lex. Wortf. s.v.) to the analogy of genitor. Sometimes the
Analogy of the Compound affects the Simple Verb, when the Compound is
more frequently in use than the other. The Analogy of conspicio, aspicio,
despicio, &c. changed the spelling of the little used simple verb from specio (e. g.
Varr. L. L. vi. 82, Plaut. Cas. 516) to spicio ; complico, explico, &c. have effected
the change of *pleco (Gk. mAé«rw) to plico. (For other ex. of ‘ Re-composition,’
see Seelmann, Ausspr. p. 60, and Georges, Lex. Wortf. s. vv. dispando, infacetus, prac-
canto, infarcio, peremo, indamnis beside indemnis ; see also above, § 28).

§ 32. (11) Pretonic. Miniscitur pro reminiscitur antiquitus dicebatur (Paul.
Fest. 88. 12 Th.) (or by Analogy of Compound ?) ; the Preposition en of O. Lat.
became in from its position before the accented syllable in phrases like
in-aéde esse, in-téemplum ire, &. Caper (p. 93. 3 K.) corrects the mispronunciation
pinaria cella for ‘penaria,’ and (p. 106. 4 K.) pulenta for ‘polenta,’ and (p. 100,
23 K.), pidato for ‘pedatu’ in the phrase ‘primo pedatu’; Probi Append. 198.
5 K. sinatus (C.I. L. i. 206, 1. 135 ; viii. 10525, &c.). We may similarly explain the
u of Ulixes, from ’Odvaceds, the ‘ Aeolic’ form of ’Odvaceds (Quint. i. 4. 16, who
also quotes the spelling Pulixena for Polyxena) ; cf. IovBwos on a Gk. inscrip-
tion of Syracuse, I. I. S. 125; BovAovpmos on a Cyzicus inscription of the
Republican period, Mitth. vi. 124. (See also Georges, Lex. Wortf. for the
spellings rutundus, lulligo, ciminum, Sigambri, Lundinium, &c.) In Italian the
influence of the following labial is shown in somiglia (Lat. *simitiat), domanda
(Lat, demandat), dovere (Lat. debere), of a following r in smeraldo (Lat.
smaragdus). But examples in Republican Latin of the weakening of initial
syllables are doubtful (cf. § 7).
§§ 82-34.] ACCENTUATION. VOWEL-WEAKENING. 201

" § 33. (12) Assimilation, Dissimilation, and False Analogy. In Italian
the unaccented vowel is often assimilated to the vowel of the neighbouring
syllable. Thus Latin aequalis has become uguale ; cronaca (Lat. chronica) owes
its penultimate a to Assimilation. The same tendency is seen in Vulgar
Latin *aramen for aeraémen (Span. arambre, Port. arame, Prov. aram, &c.),
¥salvaticus for silvaticus (Fr. sauvage, our ‘savage,’ Span. salvaje), &c., and in
classical Latin in Perfects like momordi, poposci, cucurri, of which the older
forms were memordi, peposci, cecurri (Gell. vi. 9). So strong is the tendency in
Latin to assimilate completely an initial syllable which bas some resemblance
to a following syllable that we find this Assimilation even in the accented
syllable of Perfects like pupiigi, older pepugi (Gell. ib.) [cecini reflects the older
spelling *ce-cen-ei, but when the Stem-syllable had originally i, we haveiin the
Reduplication-syllable, e. g. di-dic-i (see ch. viii. § 22) from *dic-sco]. (On the
Assimilation of Syllables in Latin, see ch. iv. § 163). Mispronunciations of
this kind censured in the Appendix Probi (197-9 K.) are: toloneum, tonotru,
passar, ansar, parantalia, butumen, and on late inscriptions we have misspellings
like monomentum (C. I. L. vi. 2888, 11131, 24481, xiv. 416 and 523 and 864; Bull.
Comm. Rom. 1880, p. 137, 1887, p. 43), optomo (C. I. L. ii. 4291) (cf. oppodum as early
as the Lex Agraria of 111 B.c., i. 200, 81), passar (I. R. N. 7160; C. I. L. vi. 2698),
ansare (v. 7906), pataris (vi. 2060, 12, the Act. Arval. of 81 a.».), carcares (vi.
2065, 2066, 2067, the Act. Arval. of 87-90 a. D.), cubuc(u)larius (C. I. L. vi. 6262,
8766), figilinae (xv. praef. p. 8). See also Georges, Lex. Wortf. on the spellings
lucuna, lucusta, tiburis, Berenice, carcar, passar [e. g. Itala (Ash.) Lev. xi. 5, (Taur.)
Math. x. 29 and 31, (Cantabr.) Luc. xi. 150], Ptolomais, Dolobella, tugurium, and
ef. Romance forms like Span. pajaro, Ital. passaretta (from Vulg. Lat. passar).
The opposite tendency, viz. Dissimilation, perhaps appears in Vulgar Latin
in a word like vicinus, where the first 7 (close 7) has been changed to open i
(Span. vecino, Prov. vezins, &c.). To the false analogy of léicus has been
referred the a of Vulg. Lat. *lacusta (Roum. lacusta), while forms like Prov.
langosta, O. Fr. langoste point to an original l’angusta (illa angusta). The
tendency of plant-, bird-, and beast-names to be changed by all sorts of false
analogies is well seen in the dialectal Italian descendants of Lat. vespertilio,
‘a bat’ (Tose. pipistrello, and vipistrello Caserta sportiglione, Pisa pilistrello,
Parma palpdstrel, &.). (For exx. of vowel retained by Assimilation, see
§§ 22, 29). :

§ 34. (13) Shortening of Syllables long by Position. In the dramatists of
the Republic a syllable long by nature or by position is occasionally scanned
asa short syllable when a short syllable precedes, a law of Prosody which is
usually called the Law of Breves Breviantes. Of final syllables, syllables
whether long by nature or by ‘ position’ are shortened by this law especially
in iambic words like cavé, putd, feriint, legiint, the liability of a final unaccented
syllable to be shortened being increased by the precedence of a short accented
syllable (see next section), Putting final syllables aside for the present, the
usual case of syllable-shortening is in a word of four or more syllables, where
a syllable long by position is preceded by a short syllable, and followed by
the accented syllable. Thus voliiptatis, voliiptatem, voliptarius, &c. are common
scansions in the early dramatists, and voliintatis, juvintutis, gubérnare and
gubérnator, egéstatis, ventistatis, supéllectilis come next in order of frequency. The
normal scansion of all these second syllables is that of Classical poetry; but
the position of the syllable between a short syllable on the one hand and the
202 THE LATIN LANGUAGE. (Chap. III.

accented syllable on the other, made it especially liable to be slurred in
pronunciation, so that the dramatic poets, who followed more closely the
pronunciation of everyday life than others, felt themselves at liberty, when
exigencies of metre demanded, to treat it as a short syllable. In the word
ministerium this pronunciation was carried so far as to syncopate the second
syllable, minsterium, misterium (Ital. mestiero, Fr. métier, Chaucer’s ‘mistery,’
ed. Morris, iii. 348); and this form seems to occur as early as Plautus,
Pseud. 772:
paruis magnisque misteriis praefulcior,

where the MSS. offer miseriis. Less frequently we find the preposition shortened
in a Compound when preceded by a short monosyllable (or elided dissyllable),
e. g. Capt. 83 in dcculto, Most. 896 tibi dptemperem, phrases which may be considered
as word-groups in-occulto, tib(é)-optémperem, and so fall under the same category
as the polysyllables voluptatis, voluptarius just mentioned, but also, e. g. Trin. 318
quid éxprobras? Capt. 70 quia inuocatus, where the accent does not fall on the
syllable immediately following the preposition. The tendency of a preposition
in a Compound to be weakened (ch. ii. § 130) (cf. 5-mitto for *om-mitto, *obmitto ; ré-
cido, ré-latus, ré-duco, earlier reccido, rellatus, redduco, but see ch. ix. § 49), is here in-
creased by the precedence of a short syllable ; or perhaps the truer explanation
is that the Preposition was regarded as separable from the other member of the
Compound, and quid é&:-, qui(a) in- show the same shortening as in the final
syllable of iambic words. Similarly in Greek and other loanwords a syllable
long by position may be shortened when the preceding short syllable has the
accent, as in Plautus always Philippus (@/A:nmos) in the sense of a ‘Philip,’ a gold
coin, and in the Christian poets abjssus (dBvacos) (Paul. Nol. 19. 651; 35. 2285
Cypr. Gall. Gen. 288 P.). In Vulgar and Late Latin we have syllables long by
nature shortened in this way, e. g. ertmus (pnyos) in the Christian poets (e. g.
Prud. Psych. 372; Cath. v. 89), whence the Romance forms, Ital. erémo and
ermo, O. Fr. erme, Span. yermo, &e. ; mertbatur, a mispronunciation censured
by Consentius 393. 23 K. (also drator 392. 11 K.); verécundus in the Christian
poets (e. g. Fort. vii. 6. 10) (cf. vericundus @, I. L. x. 1870), whence the Romance
syncopated forms of ver(e)cundia, Ital. vergogna, Fr. vergogne, Span. vergu-
enza, &c. ; but the instances which can be quoted from the early dramatists
are so few and so uncertain as not to warrant us in ascribing this pronunciation
to an earlier time (see Journ, Phil. xxi. 198; xxii. 1). In Ter. Phorm. go2 an
uerébamini, some MSS. have an ueremini; and Clutémestra or Clutaémestra (KAvra-
pnorpa, a better spelling than KAvra:pyqjorpa), in Livius Andronicus, Trag. 11 R.,
may be a case of false analogy, like orichalcum (épelxadxos), which is in Plautus
aurichalcum, by confusion with aurum, and owes its short 7 to this earlier form
(ef. aquaeductus non ‘aquiductus’ Prob. App. 197. 26 K., like terrae motus
non ‘terrimotium’ ib. 198. 32). Ancira (dyxupa), where the shortened vowel
follows a syllable which is not short but long by position, seems, with its o for
uv before r, not to be a direct development of the Greek word. The early
dramatists do not shorten by the law of Breves Breviantes the prepositions
in, con in Compounds when the letter following the preposition is s or f (see
Journ. Phil. 11. ce.); and we know from Cicero (Or. xlviii. § 159) that the 7, 0
were long in these cases. Caléfacio, &c. (in Quintilian’s time apparently calfacio,
i. 6. 21), are really separable compounds, cale facio (ef. facit dré, Imer. vi. 962),
so that the e¢ is properly regarded as a final vowel; and the same is true of
diéquinte (cf. Gell. x. 24. 1).
§ 85.) ACCENTUATION. REDUCTION OF FINAL SYLL. 203

§ 85. Change and Shortening of Vowel in Unaccented Final
Syllable. The final syllable in. Latin requires a separate treat-
ment, for besides the want of accent, there are other weakening
influences to which a final syllable is always liable. Phoneti-
cians tell us (Sweet, Primer, § 105) that ‘the general tendency
of language is to pronounce with diminishing force,’ so that in
English, for example, the ¢ of ‘cat’ is pronounced with more
force than the ¢, and the final consonants of ‘obliged’ are
‘whispered’; and in Portuguese the final o of a word like campo
(Lat. campus) is similarly uttered with what is known as
‘whisper,’ not with ‘voice. When a vowel actually ended
a word, it would also be liable to elision, more or less complete,
before a word which began with a vowel or the letter 2.

I. Loss or Syncope or SHorr Vower. i. Final vowel.
The weakness of a final short vowel in Latin is seen in Plautine
versification. Plautus (according to Langen, in Philologus, xvi.
p- 419) shows a preference to elide a final short vowel rather
than allow it to constitute by itself a thesis, so that endings of
iambic lines like eapectare vis, where the final & of expectare
forms the thesis of the last iambus, are not common. The
weakness of final @ in particular, the vowel to which, as we
shall see, every short final vowel was changed, is shown still
more by its occasional suppression in words like quippe, unde,
inde, and perhaps ile, iste, before an initial consonant in Plautus
and the early dramatists. Nempe is always scanned nemp in
this position by Plautus and Terence, while protnde, deinde, have
developed the byforms proim and dem, and négue, atque, the
monosyllables xec, ac (for *atc). All these are words which
would naturally be closely joined in utterance with a following
word, so that we may compare the Italian suppression of -e, -o,
after , /, r in word-groups, such as of the final vowel of bello,
buono, signore, &c., in phrases like bel tempo, buon giorno,
signor padre, tal cosa, &c. Similarly the subordinate or
auxiliary verbs facio, dico, dico, lose their -e in the 2 Sg. Imper.
Jac, dic, duc (see ch, viii. § 28). The same loss of -2, whether due
to syncope in a word-group, or to elision before an initial vowel,
or to both causes, has produced -/ from -/e, -r from -re, in forms
like bacchdnai for earlier *bacchandle, calcar for *calcare (Neut. of
204 THE LATIN LANGUAGE. (Chap. III.

cabcaris, for calcare ferrwm, ‘the iron attached to the heel’), &c.,
and has reduced the particles -ce, -ne, to -c, -2, in hic, hunc, viden,
audin, &e. The loss of final -wm in nihil for nihilum, séd for
sedum (Ter. Scaur. 12, 8 K.), &., can have been due to elision,
but not to syncope (see ch. x. § 18).

ii, In final syllable. The syncope of a short vowel in a final
syllable ending in a consonant has been already discussed in
§ 16. We there saw that this syncope, a prominent feature
of the Oscan and Umbrian languages, e.g. Osc. hirz (Lat.
hortus), Umbr. emps (Lat. emptus), is difficult to establish with
certainty for Latin, since vdlens beside violentus, mansués beside
mansuetus, rémex beside prodigus, &c., may be instances of parallel
stem-formations like pénu-, peno-, penos-, of penus, Gen. penis,
penum, Gen. peni, penus, Gen. pendris; and even stronger examples,
such as Arpinas, older Arpindtis, praeceps, older praectpes, may
have arisen otherwise than by syncope.

§ 36. Loss of -e. For other examples in Plautus such as Pseud. 239 mitt(e) mé sis,
and for a list of instances of quipp(e), nemp(e),&c. see Skutsch, Forsch. i. Plautus’
use of -ne and -n seems to depend, not on whether the initial of the following
word isa vowel or a consonant, but on whether the preceding syllable is short or
long (Schrader, De part, ‘ne’. . . apud Plautwm) (for Terence’s use of -ne, -n, see
Dziatzko ad Phorm. 210 Anh.) ; while he employs the forms hisce, illisce; &e.
before an initial vowel, hi, di, before an initial consonant (Studemund in
Fleckeisen’s Jahrb. 1876, p. 73). Parallel forms like atque, and ac (for *atc),
Ital. tale and tal, which have arisen from the same original form according to
its position in the sentence, are called ‘doublets.’ (German ‘Satzdoubletten’)
(ch, ii. § 136.) The Latin mdgis has thus become in Italian mai, when used
independently as an Adverb; but ma, with loss of the final vowel, when
used as a Conjunction, and so joined to a following word. In Oscan, avt, in
the sense of Latin autem or at, and avti, in the sense of Lat. aut, may be
similar doublets. The syncopated form of the I.-Eur. preposition *Apé
(Greek do, Sanser. dpa) has become universal in Latin, e.g. ap-trio, ab-diaco
(cf. sub, Greek to), almost the only trace of the final vowel being yo-situs,
pono for *po-s(i)no ; I.-Eur. *péri (Greek ép:, Sanser. pari) is Lat. per- in perma-
gnus, persaepe, &c,; I.-Eur. *éti (Greek é7:, Sanscr. Ati) is Lat. et (Umbr. et) ;
I.-Eur. *5pi (Greek ém-oGev) is Lat. ob (Oscan op); I.-Eur. *ambhi (Greek
dui) is Lat. amb-ustus, an-cisus), whether the Syncope of these words took
place in the Latin period (ab from *ape, earlier *apo), or at a much more
remote period (cf. Goth. af, English ‘of,’ ‘ off’). Weuw, sew, ceu, which are not
used in Latin poets before a vowel, are cases of Syncope in the Latin period
o sive &e.; also quin (see ch. x. § 16) for qui-ne [cf. Ter. Andr. 334, if née) :

efficite qui detur tibi;
égo id agam mihi qui ne detur) ;

sin for si-ne ; quot, tot (cf. toti-dem, Sanscr. kati, téti). (On fer and vel, see ch. viii.
§§ 36, 37.) ACCENTUATION. REDUCTION OF FINAL SYLL. 205

§ 58, and on em, originally *eme, the Imper. of dmo, ‘to take,’ ch.x.§ 19). Aslate
as the time of Terence we find abduce used before a vowel, abduc before a con-
sonant, while face is the form employed at the end of a line (Engelbrecht,
Studia Terentiana, p. 63) ; but in the classical period, owing to the prevalent
use of these imperatives dic, duc, fac, in word-groups, i.e. in close connection
with a following word, the syncopated ‘doublet’ has ousted the other form,
just as in post-Augustan poetry we find nec more and more supplanting
neque, and usurping the position before vowel- as well as before consonant-
initials. Other Imperatives occasionally appear without final -é, e.g. inger mi
Catull. xxvii. 2 (see ch. viii. § 58). So with -¢ of the Infinitive. Biber dari is
quoted by Charisius (124. 1 K.) from Fannius (cf. Caper 108. 10 K. bibere non
‘biber’) ; and a plausible etymology of instar, a word first used in Cie. Verr. ii. 5.
§ 44, and literally meaning ‘ weight’ (cf. Cic. Off. iii. 3. 11 ut omnia ex altera
parte collocata vix minimi momenti instar habeant), makes it the Infinitive,
used, like biber, as a Substantive, of insto, ‘to be of equal weight,’ ‘to show equi--
poise of the balance,’ like Swiss-German ‘die Stimmen stehen ein,’ ‘ the votes
are equal.’ (Wolfflin in A. L. L. ii. 581.) Bustar or bostar, glossed by Bovordo.ov
in the ‘Cyrillus’ and ‘Philoxenus’ Glossaries, may be for -stare, as instar for
instare. We find -al, -ar for -ale, -are in trisyllabic or longer Nouns like
animal, but from sedile, &. we do not find *sedil, &c., nor from tle, &e. ul,
though subtél (7d Kotdov rod modés, Prise. i. p. 147. 11 H.) is said to stand for
¥subtéle, Neut. of *subtélis from talus. Strempse, an old legal word, found in
the phrase sirempse lex esto, ‘let the same law apply,’ e.g. Plaut. Amph.
prol. 73:
sirempse legem iussit esse Iuppiter,

is found without the final -e in the Tabula Bantina of 133-118 B.c. (C. 0. L.
i. I97. 13 siremps lexs esto), and other early laws. Luacte, the Plautine form
(though lac ig the reading of the MSS. in Amph. 601, perhaps a corruption
of iact) is lact in Varro, L. L. v. 104 (lacte Cato ap. Char. 102.9 K.), and in
classical Latin lac (cf. Charisius, 102. 4 K. lactis nominativum alii volunt lac,
alii lact, alii lacte ‘e’ postrema). Véliip, ‘pleasurably,’ seems to be for
*volupe, Adverbial Neut. of an Adj. *volupis ; and Ritschl (Opusc. ii. 450) would
analyze the volupest of Plautus, Mil. 277, &c. into volupe est, a form which the
phrase seems actually to bear in late Latin writers, like Arnobius,
Prudentius and others (see Georges, Lex. Wortf. s.v.). We have similarly
Focul for facilz, e.g. Lucilius vi. 3 M. nobilitate facul propellere iniquos, and
dificul for difficile (see Nonius, p. r11. 21 M.; Paul. Fest. 61.32 Th.; Fest. 266,
20 Th. ‘ perfacul’ antiqui et per se facul dicebant, quod nune facile dicimus).
The O.Lat. Adverb poste (e.g. Enn. 4. 244 M. poste recumbite, uestraque
pectora pellite tonsis) is in classical Latin post ; ante does not appear without
the final -e in Latin, but we have in Oscan ant, as well as plist, post, Umbr.
post. The suppression of a final short vowel was a common feature of Oscan
and Umbrian, e.g. Oscan nep, Umbrian nep (Lat. neque), &e. (See also ch. x.
§§ 9 and 12, on wt and titi-nam, dinec and donique, and cf. Georges, Lex Worif. s. vv.
altar(e), animal(e), autumnal(e), boletar(e), cervical(e), cochlear(e), laquear(e),
‘pulvinar(e), virginal(e), lucar, specular, toral, torcular, vectigal, &e. Quint. i. 6. 17
speaks of tribunale as out of use in his time).

§ 37. II. Change of Vowel. i. Short Vowel. We have already
seen (§ 18) that, in the syllable immediately following the
2.06 THE LATIN LANGUAGE. (Chap. IIT.

early accent, every short vowel was changed to é@, unless diverted
by a following labial to 6. It is probable that short final vowels
took the same course, and were one and all changed to & This
é might be dropped (§ 36) or retained, but did not become 7, as
unaccented 2 in the middle of a word did (§ 18), so that @ is
preeminently the final vowel of the Latin language. A final
7 becomes @ in médre for *mari, dnimale (later animal) for *animali,
&c., while in the middle of the word it remains, e.g. maria,
animalia. Similarly final -2, when, by the addition of a particle,
it ceases to be a final vowel, becomes 7%, e.g. bénificus beside
bene, quippint beside guippe, sicine beside sic(e), hoccime beside
hoe-ce from *hod-ce (ch. vii. § 16).

ii. Diphthong. A diphthong in the final syllable was treated
like a diphthong in the posttonic syllable. As we have ez,
class. 2, for posttonie ai in inceido (S. C. Bacch.), class. inctdo
from O. Lat. caido, class. caedo, so we find final ¢7, class. 7 repre-
senting I.-Eur. ai (or ai? ch. viii. § 66) in the 1 Sing. of the Perfect
Active, &e., e.g. tétid-i (older -e), And while an example of the
weakening of oi to ei, 7 in the posttonic syllable is difficult to
find (§ 18), it is regular when final, e. g. foideratei (S.C. Bacch.),
class. foederati, from an original ending -o7, On the treatment
of the final long diphthongs -ai, -éi, &., see ch. iv. §§ 45 sqq.

iii. Long Vowel. In the post-tonic syllable, as we saw (§ 30),
a long vowel was not changed through the influence of the
preceding accent. Nor was it changed in quality in the final
syllable, though its quantity suffered. Long final a became
-@ in terrd, arvd, &c., but did not pass into another vowel, such
as é. The shortening of long final vowels is discussed below

(§ 40).

§ 38. Change of final short vowel to 6. An example of é- for an original -6
is the ending of the 2 Sg. Imperat. Pass. and Depon., e.g. sequere for *sequesd
(Gk. éme(c)o, ch. viii. § 77), of -¢ for -%, perhaps sat from an older *sat(é), if this
was a u-stem *satu- (cf. satu-r) (but see ch. ix, § 4). An -d which has escaped this
weakening (e. g. endo, on which see ch. ix. § 27) became -% (as in the posttonic
syllable, § 26), e.g. indi. (On noenti, a byform of noenum, see ch. x. § 18).

§ 39. Alternation of final e with internal i. Other examples are istic, illic
from iste, ille, increased by -ce ; isticine, illicine, further increased by -ne ; hicine,
nuncine, tuncine ; tutin, for tute with ne, is the spelling of the MSS. in Plaut. Mil.
290 ; undique from unde, indidem from inde; ante is anti-in compounds like antict-
§§ 88-40.] ACCENTUATION. REDUCTION OF FINAL SYLL. 207

pare, antistes, antistita, antigerio (O. Lat. for valde), and antisto (a better spelling
than antesto: see Georges, Lex. Wortf. s.v.) ; facilin for facile with ne, servirin
for servire with ne, is the spelling of the MSS. in Plaut. Men. 928 and 795;
vbenivolus, benificus, malivolus, malificus (beside benevolus, malevolus, &e., a spelling
much discussed by the grammarians, e. g. Vel. Long. 76-77 K.; Alcuin 298.
14 K.; Probus, 119.2K. See Brambach, Lat. Orth. and Georges, Lex. Wortf.
8. vv.) (For additional examples see Ritschl, Opusc. ii. 556).

§ 40. III. Shortening of Long Syllable. i. Final long vowel
or diphthong. A long vowel or diphthong in the middle of
a Latin word may be shortened in’ hiatus, e.g. prus (cf. Oscan
pithio-), dalnéum (Badaveiov), déamo, préhendo (see ch. ii. § 143).
The same thing happened apparently to a final long vowel or
diphthong when the next word began with a vowel or 4, so that
scansions like Plautus, Au/. 463 mé% hondris, Asin. 706 dé hérdeo
(cf. class. méherele), Ennius, Ann. 45M. Seipid inuicte, need not have
been imitations of Greek poetry, but rather expressed the actual
Latin pronunciation. How far the shortening proper to this
position may have attached itself to the vowel, even when a con-
sonant initial followed, is not easy to say. From the earliest period
of Latin literature we find a tendency to shorten every final long
vowel. Some offer more resistance than others; -7 and -# than
-@ and -6. The final @ of Nom. Sg. of A-stems and of Nom.
Acc. Pl. of Neuter O-stems seems never to occur even in the
earliest poetry in any but a shortened form, while in the Oscan
and Umbrian dialects it has been reduced to some sound which
is written 0, and which is treated by Lucilius as a short vowel
(Lucil. ize. 106 M.), if we may trust the quotation by Festus
(426. 7 Th.) Lucilius: ‘uasa quoque omnino dirimit non sollo
dupundi,’ id est, non tota). Final 6 of verbs and nouns is, on the
other hand, always long in the earlier poetry, except when the
precedence of a short vowel, especially an accented short vowel,
allows it to be scanned as a short syllable, e. g. legd, modo, less
frequently pellegs, dicitd. But by the time of grammarians like
Charisius and Diomedes (fourth century a.p.), this -o was
universally shortened in pronunciation, so that a fifth-century
grammarian (Pompeius, p. 232 K.), cannot explain Virgilian
scansions like canté, except on the theory that they are imitations
of the Greek -w of 0.6, &e.!_ The course of development taken
by -d in the literary period, viz. its shortening first in iambic
208 THE LATIN LANGUAGE. Chap. III.

words like /egé, then in cretic words like ped/egé, finally in all
words, e.g. cant, we may suppose to have been taken in the
pre-literary age by final -d. From ferd, &c. the shortening
would spread to efferd, &c., and would in time be extended over
every Nom. Sg. Fem. of A-stems and Nom. Ace. Pl. Neut. of
O-stems. That the shortening was mainly the work of analogy
we see from ¢trigintd, quadragintd, &c., which, though really
Neuters Plural, were regarded as mere numerals and so escaped
the shortening which was enforced on every Neuter Plural Noun.
But it must have been aided, partly by the inherent weakness of
every final syllable, partly by the shortening of a final long
vowel in pronunciation when the next word began with a vowel.
The former presence of a final consonant does not seem to have
made much difference. Ovid scans estd (earlier estéd) as he
scans Sudmd; and Plautus allows the shortening by the Brevis
Brevians law of daté, dicitd, probé, maxumé, mant and other
Ablatives (earlier datéd, &c.).

ul. Long vowel followed by consonant. The quantity of a vowel
in a final syllable is often influenced by a following consonant,
In English the long vowel-sound of ‘node’ becomes a half-long
sound before the dental tenuis, ‘note. Similarly in Latin
a long vowel tended to be shortened by a following final ¢, 7,
&c. Under the shortening influence of a preceding short
accented syllable, the final syllable (with naturally long vowel)
is readily shortened by Plautus in words like tenet, amat, and,
to a less extent, soror, moror, but seldom in words like fenes,
amas, moras (for the statistics, see Leppermann, De correptione,
&ec. p. 78); and in classical poetry every originally long vowel
is scanned as a short vowel before final -¢, -r, &c., but not before
final -s. Final -2 also shortens a preceding long vowel; thus
bacchanal (for bacchanale) became, when the accent shifted to the
second syllable, dacchandl, as caledr (for caledre) became, under
similar circumstances, calcir. And, though we cannot trace the
effect of final -m in poetry, seeing that a syllable so ending is
elided before an initial vowel, we are told by Priscian that it had
the same power of shortening a long vowel (even in monosyl-
lables), e. g. spém, rém, diém, meridiém (Prise. 1. 23. 13; 366.21 H.).

ui. Pinal syllable long by position. In Plautus legiint, dixeriint
§ 41.] ACCENTUATION. FINAL SYLLABLES. 209

are admitted as well as Jegd, diver. But in the hexameters of
Ennius, Lucilius, &¢., these shortenings of final syllables long by
position are avoided, as they were in the poetry of the classical
period. They were apparently regarded as vulgarisms, much as
the change of final -ng to -n is with us.

§ 41. Final long vowel in Hiatus. In Greek poetry (dactylic, anapaestic, &c.)
besides the shortening of final diphthongs like ai, o: before an initial vowel
(a seansion due to the consonantal character of 1, dvdpa por tvvere being pro-
nounced dvipa poyévvere, G. Meyer, Griech. Gram.? § 154), we sometimes find
shortening of final @, w, 7 in similar circumstances. This shortening seems
to have reflected the ordinary pronunciation, as we can see from inscriptions
in the Cretan dialect, a dialect in which the nuances of sound taken by a word
in its various positions in the sentence were more regularly expressed in the
orthography than in other dialects. On the Tablet of Gortyn, for example,
v7 is written we when the next word begins with a vowel (K. Z. xxxiii. 133)
In the native metre of the Romans, the Saturnian, a final long vowel or
diphthong (or syllable in -m), seems similarly to be left in Prosodical Hiatus, i.e.
shortened, not wholly elided, Before an initial vowel or h- (see ch. ii. § 143);
and this Prosodical Hiatus, as well as Hiatus proper, such as the non-elision of
a final short vowel, is much more common in Plautus than in Terence, as it
was in Naevius, according to Cic. Or. xlv. § 152, than in Ennius. Plautus
employs it in dialogue metres with (1) (accented ?) monosyllables preceding
a shortinitial syllable, e.g. quéeam? (2) iambic words with verse ictus on the first
syllable, e.g. méi honoris ; (3) monosyllables following a short final syllable which
has the verse ictus, e.g. omnid quaé isti dedi ; (4) iambic words or word-endings,
when the final syllable has the verse ictus and the following initial syllable
is short and has the natural accent, e.g. vir habitat, una operés ebur, obsequt
animo. In Anapaestic Metres also with (5) Cretic words, and in other cases.
(For a list of examples, see Klotz, Altrém. Metrik, p. 119. They include not
merely instances of dissimilar vowels, but also of similar, e.g. i-¢ in ét
imdgine, Pseud. 1202). Terence, and apparently Lucilius, restrict it to the
first of these cases; but Virgil has not only examples like qui amant (Ec.
viii. 108), but also like vale valé inquit (Ecl. iii. 79), and sub [ltd alto (A. v. 261).
Virgil thus employs it (1) to prevent the entire suppression by elision of
a monosyllable (accented ?), ending in a long vowel or -m ; (2) in cases where
a long final vowel would be shortened by the law of Breves Breviantes in the
dramatists. ‘That a vowel shortened in Hiatus was not so long as an ordinary
short vowel, we may infer not only fromthe fact that it is normally elided,
but also from the disinclination shown by Plautus to allow a vowel so
shortened to constitute by itself the thesis of a metrical foot. It is allowed
to go with another short syllable to form a resolved thesis, e.g. qui éém, but
is never allowed to dispense with the proximity of another short syllable,
except in a few phrases which may be considered as word-groups or com-
pound words, e. g. dé hérdeo, Asin. 706, which might be written de-hordeo like
déhortari, Poen. 674. The difference in this respect between Plautine and
Saturnian versification (see ch. ii. § 143) need not imply a change in the pro-
nunciation of these final long vowels in Hiatus ; but may be due merely to
the different character of the verse. The Saturnian poetry was of a more

P
210 THE LATIN LANGUAGE. (Chap. III.

solemn and dignified tone than the conversational verse of the dramatists,
and would naturally be uttered with a greater pause between the words,
Imitation of the Greek dactylic and anapaestic prosody is inconceivable in
the Saturnian poetry, and unlikely in the trochaic and iambic verse of
Plautus ; so that we can hardly be wrong in supposing this Prosodical Hiatus
to reflect the ordinary pronunciation in Latin, as it did in Greek.

§ 42. Breves Breviantes. The syllables most affected by this law are those
ending in a long vowel in words which were in ordinary talk closely joined
with a following word. Forms like mihi, tibi, sibi, modi, cits, &c. have forced their
way even into classical poetry ; and in Plautus we find this shortening chiefly
in verb-forms, which go closely with a following word, e. g. voli-sctre, abi-rus,
cavt-dicas, while the examples of nouns are mostly confined to adverbial forms,
e.g. domi-restat, domb-prodit, or subordinate words like homé (see statistics in
Leppermann, De correptione, p. 78). Ennius in his Epic restricts this usage
to words ending in a vowel, and subsequent Hexameter poets follow him,
e.g. puts but not legiint, dixerd (Hor. S. i. 4. 104) but not dixeriint. That this
shortening was not a mere metrical licence, but reflected the actual pronuncia-
tion, we see from Quintilian’s remark (i. 6, 21) that havd, not avé, was the
normal form in his time, as well as from Phaedrus’ fable of the man who
mistook this word for the caw of a crow (App. 21), and Cicero’s story (Div. ii.
40) of Crassus mistaking a figseller’s cry, Cawneas (sc. ficus vendo), for cave ne eas.
The spelling causts for cave sis in Juvenal ix. 120 points to the same thing ;
and Servius (ad Aen. vi. 780) says that vid?n was the pronunciation of his
day. (Should we read rogdn for rogiés in Pers. v. 134?) Plautus in his
dialogue metres allows the scansion of a cretic word as a dactyl in the first
foot only of the line or hemistich ; Terence not at all; but Horace in his
Satires and Epistles has Pollid, dixerd, mentid, &c. ; commodd Catull, x. 26 is probably
Neut. Plur. (see Owen ad loc.). (On the operation of the Breves Breviantes
Law in Plautus, see Journ. Phil. xxi. 198 and xxii. 1.) Plautus requires that
the preceding short syllable shall be perfectly short; he does not allow
a short vowel preceding a mute and liquid to act as a Brevis Brevians, e.g. not
pitri like pati, nor even a short vowel preceding qu, except under particular
circumstances. But in classical poetry we find putréfacta, liquéfiunt, &e.

§ 43. Shortening of final -4. We have -4 in Greek words in the early
poets (Enn. A. 567 M. agoed longa repletur is very uncertain ; cf. Gk. dyud),
just as we have in later poetry, e.g. Stat. Theb. vi. 515 Nemed (cf. Prise. i.
p. 202.16 H.). But the instances quoted of -4 in Nom. Sg. of A-stems or
Nom. Acc. Pl. of O-stems seem to be illusory. They are really cases of
(1) metrical lengthening, e.g. Enn. A. 149 M. et densis aquilé pinnis obnixa
volabat, a lengthening of a short syllable before the penthemimeral Caesura,
like the lengthenings before the hephthemimeral in A. 85 M. sic expectabat
populis atque ora tenebat, Virg. A. iii. 464 dona dehine auro gravid sectoque
elephanto, where we have an originally short syllable (e.g. popults, I.-Eur.
-6s) lengthened, by a metrical licence borrowed from Greek poetry, before
the two chief caesuras of the hexameter ; (2) syllaba anceps, e.g. Plaut. Mil.
1226 namque édepol uix fuit cépid | adetndi atque impetrandi (at the end of
the first hemistich of an Iambic Septenarius, like -is in Truc. 149 non druos
hic sed pdscués | ager ést : si aratiénes) ; Plaut. Rud. 1086 TR. Tit crepundia (a).
GR. Quid, si ea sunt aurea? TR, Quid istic tua? (at change of speaker,
§§ 42-44. ] ACCENTUATION. FINAL SYLLABLES. 211

like -& of Voc. Sg., I-Eur. -é, in Pers. 482 TO. Quid agis? DO. Credo. TO.
Unde agis te, Dérdalé. DO. Credé tibi). Or they are cases of wrong scansion,
e.g. Trin, 251 nox datur: ducitur familia tota (where the metre is Anapaestic
with familid, not Cretic with familia), Mil. 1314 Quid uis? Quin tu itbes ecferri
é6mnia quae isti dedi [where we should scan émnid quaé isti, not 6mnia qu(ae),
isti], or of wrong reading, e.g. Asin. 762 Ne epistula quidem Ulla sit in aédibus
(where we might insert wsquam before ulla, as in Rud. 529, and scan epistuld,
not epistuld). A few apparent instances of din Plautus have not yet been
explained, viz. Bacch. 1128 ; Epid. 498; Men. 9744. (For a list of examplesin
Plautus, see C. F. Miiller, Plaut. Prosodie, p. 1; in Ennius, see Reichardt in
Fleck. Jahrb. 1889, p. 777.) In the Saturnian fragments there is no reason for
scanning -@ in Nom. Sg. or Neut. Pl. (see ch. ii § 141). Final -é for -dd is long in
Early Latin, as in Classical poetry, e.g. Abl. mensd, era, Adv. extra, supra
(exstrad, suprad on S.C. Bacch. of 186 B.c., 0. I. L. i. 196), so that Early Latin
contra, frustra (e.g. Plaut. Rud. 1255 ne tu frustra sis, at the end of an Iambic
line; Naev. praet. 6 R. contra redhostis, at the beginning of the second
hemistich of a Trochaic Septenarius ; Enn. ap. Varr. L. L. vii. 12 quis pater
aut cognatu’ uolet nos contra tueri?) cannot have been originally *contrad,
*frustrad. But it may be ‘shortened by the influence of a preceding short
syllable, like any other long vowel, e. g. venustissumd, Poen. 1177, gratia, Stich.
327, rusticd, Pers. 169. Similarly with -d of Imperatives of the first Conjuga-
tion. We have in the dramatists amd, puti beside amd, puté, and so even
e.g. Persius iv. 9. hoc puté non justum est. But this shortening was not
extended by analogy to all Imperatives in -@ We never find *plantad for
planta, *mandé for mandé in Early or in Classical poetry. This is perhaps due
to the influence of the other Persons of the imperative plantito, plantate, while
for nouns like mensa the length of the final -a would not be impressed on the
memory by other cases like mensae, mensam, &c. The -é of Numerals like
quadraginia, &c. is not scanned as a short syllable till late times, e. g. C. I. L.
vi. 28047 (=Meyer, Anth. 1326) quadraginté per annos ; vi. 29426 (= Mey. 1389)
septvaginti, when Abl. -@ is similarly treated, e. g. C. I. L. xiv. 3723 hic situs
Amphion ereptus prima juventa (see ch. ii. § 141).

§ 44. Shortening of final -é6. Final -2 of the Imperatives of the second
Conjugation is scanned short by Plautus under the same conditions as final
-@ of first Conjugation Imperatives, e.g. moné and moné, cavé (almost always
short). That this scansion corresponded with the pronunciation we see from
the. remark of Quintilian (i. 6. 21) that havé, not avé, was the universal pro-
nunciation of his time (multum enim litteratus, qui sine adspiratione et
producta secunda syllaba salutarit—‘avere’ est enim—et ‘calefacere’
dixerit potius quam quod dicimus, et ‘ conservavisse,’ his adiciat ‘ face’ et
‘dice’ et similia. recta est haec via: quis negat? sed adjacet et mollior et
magis trita) (cf. § 42). This shortening was not extended to Imperatives
with long penullt, e. g. splendz, never*splendé, Similarly the -é of calé, frige, &c. in
the compounds calefacio, frigefacio is in all Latin poetry scanned short only
when the first syllable is short, cdléfacio (but never *frigéfacio), which in
Quintilian’s time was apparently pronounced caffacio (Quint, i. 6. 21 quoted

above ; for this spelling see Georges, Lex. Wortf.s.v.). Cato’s arfacit, so spelt
in MSS. of the Res Rustica 69; 125; 157. 12, seems to follow the analogy of
calfacit. These Compounds were Separable Compounds (cf. facit are for arefacit,
Luer. vi. 962), so that their -2 is properly regarded as final -2. This -¢ already

P2
212 THE LATIN LANGUAGE. {[Chap. III.

shortened to some extent under the influence of the preceding accented short
syllable in calé, madé, &c., is in the Compounds cale-fiicio, made-facio (cf. Prise.
i. p. 402. 10 H.) subjected to the additional weakening influence of a follow-
ing accented syllable, and so is scanned by the dramatists invariably as
a short syllable, although other writers sometimes make it long, e. g. Enn. Ann.
573 M. patéfecit; Catull. lxiv. 360 tepéfaciet beside tepéfacsit of xviii. 29 (see
Ritschl, Opusc. ii. p. 618). A short vowel before a mute and liquid (and to
some extent before gu) (see ch. ii. § 93) was not so short as a short vowel before
a single consonant, and was not so capable of acting as a Brevis Brevians in
the dramatists’ versification. Hence Ritschl was wrong in scanning piitré-
Jacit, Plaut. Most. 112, though Ovid has putréfactus, liquéfiunt. Similarly in the
compound of dies and quintus, &c. the @ of the second syllable, properly
regarded as a final -2, was shortened in the Republican forms diéquinte, &e.
(Gell. x. 24. 1 ‘die quarto’ et ‘die quinto’. . .ab eruditis nune quoque
dici audio, et qui aliter dicit pro rudi atque indocto despicitur. Sed Marci
Tullii aetas ac supra eam non, opinor, ita dixerunt ; ‘diequinte’ enim et
‘diequinti’ pro adverbio copulate dictum est, secunda in eo syllaba correpta.
Divus etiam Augustus, linguae Latinae non nescius, munditiarumque patris
sui in sermonibus sectator, in epistulis plurifariam significatione ista dierum
non aliter ususest.) Final -2in the Abl. of the fifth Declension is treated by
Plautus exactly as final -@ of first-Declension Ablatives, that is to say, it
is occasionally scanned short when preceded by a short, especially an
accented short, syllable, but not otherwise, e.g. dit, fidé. This shortening
was not extended to Ablatives with long penult. So with Adverbs in -2
(originally -éd, e.g. facilumed on S. C. Bacch. of 186 B.c., C.I.L. i. 196),
Plautus scans prov, maxiimé, though an instance of the shortening of this -e is
wanting in Terence. ;

§ 45. Shortening: of final -6. In Plautus and the other dramatists final
-6 is shortened under exactly the same conditions as final -é, that is to say, only
under the influence of a Brevis Brevians, e.g. v3, which normally has this
scansion when joined closely as an auxiliary verb with an infinitive, vold-scire,
&c. This shortening of -d in some iambic and cretic words had so established
itself in pronunciation that even the later Republican and Augustan poets
admit scansions of iambic words like homé (Lucr. vi. 652), vold (Catull. vi. 16),
dats (Catull. xiii. 11), veld (Hor. S. i. 1. 104), and even of cretic words like
Poltio (Hor. S. i. 10. 42, 85; and even in the Odes, ii. 1. 14), mentid (Hor. S,
i. 4. 93), dixerd (Hor. S. i. 4. 104), quomodd (Hor. S. i. 9. 43). The shortening
of final -d, like that of final -d, and unlike that of final -2, rapidly extended
itself to all instances, even when a long syllable preceded. In Ovid we have
ergé (Her. v. 59, and elsewhere), esti, Trist. iv. 3. 72, Sulmd, Nasd, &c.; and even
“Cicero uses Veits, if his epigram is rightly quoted by Quint. (viii. 6. 73) fundum
Vetté vocat, quem possit mittere funda, &c. (On endd, see ch. ix. § 27.) But -d of
the Dat. and Abl. is not shortened till very late times. The fourth-century
grammarians speak of the final -o of Nouns (Nom. Sing.), Verbs (x Pers. Sing.
Pres. Ind.), Adverbs and Conjunctions, as universally shortened in the
pronunciation of their time, except in monosyllables and foreign words.
(Charis. p. 16. 5 K. etiam illud magna cura videndum est quod veteres
omnia vel verba vel nomina quae o littera finiuntur, item adverbia vel
conjunctiones producta extrema syllaba proferebant, adeo ut Vergilius quoque
idem servaverit, in aliis autem refugerit vetustatis horrorem, et carmen
§§ 45-49.] ACCENTUATION. FINAL SYLLABLES. 213
contra morem veterum levigaverit ...paulatim autem usus invertit, ut in
sermone nostro ‘scribo’ ‘dico’ et item talibus, ubi o non solum correpta
ponitur, sed etiam ridiculus sit qui eam produxerit .. . sane monosyllaba fere
quaecumque sunt verba mpwrdruza o littera finita tam versu quam etiam prosa
similiter productam habent: necesse non corripi, ut ‘sto’ ‘do.’ quibus si
conferatur ‘dico’ ‘curro’ ‘disco’ item producta o littera, dijudicari poterit
quam sit aliud absurdum, aliud per euphoniam gratum; cf. p. 63. 17 K.
nullum autem nomen o producta finitur nisi peregrinum, veluti ‘Ino’
‘Sappho ’ ‘Dido’ (ef. Diom. p. 435. 22 K. ; ‘Prob.’ de ult. syil. p. 220. 15 K.); Mar.
Victorinus (p. 28. 23K.) distinguishes the Verbs monstré, ostenté, &c. from monstra,
ostentd, the Dat. and Abl. cases of the Nouns monstrum, ostentum. Servius (ad
A, iv. 291) attests quandd. Priscian (i. p. 409. 16 H.) excuses vigilandé of
Juv. iii. 232, &c. on the ground that it is part of a verb (: nos in ‘do’ utimur
terminatione, quae similis est dativo vel ablativo nominis, nisi quod verbum
hoe existimantes quidam etiam corripiunt o finalem ejus.]

§ 46. Shortening of final -i. The shortening of -7 by the Brevis Brevians
Law is common in Plautus in Imperatives like abi, redi, and Perfects like
dedi; while in nouns we have domi (very frequent), eri, viri, seni, &c., with the
Ablatives avi sinistra, Pseud. 762, part fortuna, Bacch. 1108 (cf. Ter. levi sententia,
Hee. 312).

§ 47. Shortening of final-t. By the Brevis Brevians Law we have mani
in Plaut. Trin. 288, but owing to the few words with short paenultima ending
in -a, the examples are not frequent. Terence has always dit, never ditt, but
Plautus has dit (or dja?) very frequently.

§ 48. Shortening of final diphthong. A final diphthong is almost never
shortened by the Brevis Brevians Law in Plautus (e. g. novaé nuptae, Cas.
118), and never in Terence. This is rather to be referred to the infrequency
of words so ending (Noms. Plur. and Dats. Sing. in -ae), than to be quoted
as a proof of the difference in sound between a diphthong and a long vowel.

49. Shortening of long vowel before final Consonant. -1. In Plautus
we have still the long quantity, e.g. Aul. 413 aperitur Bacchanal : adest, but
in Classical poetry -dl, e.g. tribundl, Ovid (cf. Mar. Victorinus de Finalibus
p. 231. 11. K.). Subtel (apparently for *subtéle, Neut. of *subtélis, a Compound of
sub and tdlus) (but see § 30), is quoted as an instance of -é by Priscian, i.
p. 147. 11 H., and explained as 76 xotAov rod Todds.

-m. Whether the different treatment of -m after oof the Gen. Plur. and o of
the Acc. Sg. Mase. and Nom. Acc. Sg. Neut. of O-stems on the earliest
inscriptions (¢.g. C.I.L. i. 16 Suesano probom, ‘Suessanorum probum’) is
a proof that a long vowel was at the end of the third cent. B.c. not yet
shortened before final -m is uncertain (see ch. ii. §137). If Lat. -m sounded like
-w owing to the lips not being closed in pronouncing it (ch. ii. § 61), the
shortening spim may be compared with néu (ch. ii. § 34). Ose. paam ‘quam’
Acc. Sg. Fem., Umbr. pracatarum Gen. Pl. Fem. (with u the equivalent of Lat. 6)
indicate a retention of the long quantity in Umbro-Oscan.

-r. (z) Nouns like calcar(e). The final syllable was without doubt long in
Plautus, though there is no certain evidence of its quantity, or of its loss of
final -e. It is short in Classical poetry, e.g. exemplar, Hor. (but exemplare,
Luer. ii, 124), though the grammarians recognize that it ought to be long by
214 THE LATIN LANGUAGE. [Chap. III.

the analogy of the other cases -dris, -ari, &c. (Charis. exc. p. 541. 2 K. : Mar.
Victorin. de Fin. p. 232. 9 K., and similarly of -al p. 231. 10 K.).

(2) Nouns and Adjectives like sdror, maeror, minor, major, have the long
quantity invariably in Plautus, unless in cases of shortening by the Brevis
Brevians Law, e.g. sordr, Poen. 364; amir, Cist. i. 1. 69, and probably always
patér. Iambic nouns often retain the old Nominative ending -os, e.g. odos,
Pseud. 841; honos, Trin. 697. Ennius in his Annals has -dr. Whether he ever
(4. 455 M. siédér) uses -dr is doubtful ; but Lucilius has normally -dr (e.g.
stridér, inc. 90 M.), and the two probable examples of -dr in Lucilius, (délor
v. 55 M. ; piiddr xxx. 70 M.) are perhaps metrical lengthenings before the
chief Caesuras, like Virgil’s libor (@. iii, 118), dimitor (A. xii. 550), &e.
There is a lack of decisive instances in the plays of Terence (see Boemer, De
correptione, p. 25).

(3) Verbs like méror, itor Ind., morer, utar Subj., in Plautus always have
a long final, unless shortened by the Brevis Brevians Law, e. g. Rud. 1248 nil
morér ullam lucrum ; Aul. 232 utér; Bacch. 153 nil mérér. By Lucilius’ time
it is invariably short, e. g. fruniscdr xviii. 3 M.; oblindr xxx. 25 M. In the
Comedies of Terence the evidence is defective (e. g. stqudr, Andr. 819). In
Tibullus, i. 10. 13 trdhér is of course a case of metrical lengthening before the
penthemimeral Caesura. Oscan patir ‘pater,’ keenzstur ‘censor’ apparently
retain the long vowel.

-t. The shortening of a long vowel before final -t was perhaps slightly
earlier than before final -r. It is indeed not found in Plautus, except where
the Brevis Brevians Law interposes (and here the shortening is much more
frequent than with -r), e. g. cubdt, Amph. 290 ; timét, Amph. 295 ; vénit, Aul. 226;
arat, Asin. 874; solét, Merc. 696; ait, Cas. 693; but in Ennius’ hexameters,
though the long quantity is usual, we find shortening occasionally even after
a long syllable, e. g. mandebdt, A. 138 M. (but ponebdt, A. 288) ; splendtt, Sat. 14
(but jubét, A. 465) ; potesstt, A. 235 (but essét, A. 81). In Lucilius the short
quantity is normal, though we have crissavit ix. 70 M. But Terence, to judge
from the slender evidence at our disposal, seems to follow rather the usage
of Ennius’ hexameter poems, for we have more long scansions, e. g. stetit,
Phorm. prol. 9; augedt, Adelph. prol. 25, &c. beside audirét, Adelph. 453. On
a Scipio epitaph of c. 130 B.c. written in elegiac metre (C. I. L. i. 38) we have
nobilitauit, though the spelling -eit in the Perfect is found much later (e. g.
probaueit beside coerauit, in C. I. L. i. 600, of 62 B. c.) (see ch. viii. § 70). Ovid
repeatedly lengthens the -it of interiit, abiit, rediit, &c. and of petiit (see Munro -
ad Luer. iii. 1042). On attdtin the Dramatists see § 10. p. 164.

Before final -s the long quantity persisted to classical times. It is
occasionally shortened by the Brevis Brevians Law in Plautus, e. g. Mil. 325
sunt manis ; Aul. 187 habés ; and the same is true of the plays of Terence (e. g.
bonis, Eun. prol.8), which however do not offer any example of a verbal form
in -s being shortened (potés, adés are for *potéss, *adéss, not *potés, *adés). But
this shortening is very rare in both dramatists, and not at all so frequent as
the shortening by the same Brevis Brevians Law before -t, -r (see the
statistics in Leppermann, De correptione ... apud Plautwm, and in Boemer, De
correptione ... Terentiana). THorace’s paliis aptaque remis (A. P. 65) is a unique
scansion in Augustan poetry (cf. vidtn, rogdn § 42). Ennius has in his Annals
(1. 102 M.) virginés (cf. Plaut. Pers. 845) before a consonant initial (see below).

§ 50. Shortening of Final Syllable long by position. Abést in Lucilius
§§ 50, 51.] ACCENTUATION. FINAL SYLLABLES. 215

(ix. 29 M.), which seems to be the right reading, stands perhaps alone as an
instance in non-dramatic poetry of the shortening by the Brevis Brevians
Law ofa final syllable long by position. Horace allows dixeré, but not e. g.
dixertint. Ennius’ virgints in Ann. 102 M.:

uirgines nam sibi quisque domi Romanus habet sas,

shows shortening of a final syllable long both by nature (-és for -ens, ch. vi. § 2),
and by position. In the dramatists the shortening by the Brevis Brevians
Law of final syllables long both by nature and by position, or by position
only, is freely allowed in the case of dissyllables in the dialogue metres (e. g. Nil
pétést (?) suprdé, Ter. ; ex Graécis bonis Latinas fecit nén bonas, Ter.), in the
case of trisyllables, &. only (as a rule) in Anapaestic and other lyric metres
(e.g. vénerant hic, Plaut. ; odio énicis miseram, Plaut. ; qui hic liberis
virginés méreatur, Plaut.).

Final syllables which had originally a double consonant are long in
Plautus, e. g. miles for *miless, Aul. 528 milés inpransus dstat, aes censét dari,
though they may, of course, be shortened by the influence of a Brevis
Brevians, e. g. potts, Stich. 325. But after Plautus’ time they appear to be
short syllables. Ennius has not only équés (Ann. 484. 249 M.), but also milés
(Ann. 277); Terence has always adés, potés; Lucilius has milés (xi. 8 M.),
prodés (inc. 128 M.) ; Lucretius (iii. 721) exds, and so on. But final -s for -ss
never fails before an initial consonant to make ‘position’ in Early Latin
versification as original -s usually fails ; milés vult could not end an Iambic
Senarius like occidistis me, Plaut. Bacch. 313. Perhaps the reduction of the
final double consonant was proper to a position before an initial consonant,
so that the actual pronunciation would be originally, e. g. miless impransus,
miles pransus (see below § 51, on hoc(c)).

§ 51. Shortening of Monosyllables. The connexion of all
these cases of shortening with the absence of accent is seen from
the fact that monosyllabic words are as a rule not shortened,
unless they are subordinate or enclitic words. Thus a long
vowel is shortened before final -7, -2, in Classical poetry in un-
accented syllables, e.g. candér, major, fundor, caleir, tribunal, but
not in the monosyllables far, s6/, where the natural length of the
vowel is retained. The monosyllable cor, however, which repre-
sents *cord, with vowel naturally short, but long by position, is
scanned short in classical poetry, though it is long in Plautus,
Poen. 388 :

hiius cor, huitis studium, huius sduium, mastigia,

the pronunciation of his time having probably been cord huius,
cord ardet, when the next word began with a vowel, but cor calet
(like cor(d)culum, ch. iv. § 157), when the next word began with
a consonant. Similarly the more or less subordinate word ter,
216 THE LATIN LANGUAGE. (Chap. III.

older *terr (cf. terr-wncius) for *¢érs (I.-Eur. *tris, ch. vi. § 61), is
a long syllable before an initial vowel in Plautus, Bacch, 1127
(a bacchiac line):

rerin tér in duno posse has tonsitari,

while in subsequent poetry the ‘doublet’ used before an initial
consonant, e.g. ter(7) durus, like hor(r)dewm (from *horsdeum,
ch. iv. § 158), established itself before initial vowels too. Hoe
Neut. for hoce (*hdd-ce, ch, vii. § 16 ; cf. hocci-ne) retained its ante-
vocalic ‘doublet’ form in classical poetry, e.g. Virg. (A. ii. 664):

hoc erat, alma parens, &c.,

and Velius Longus, commenting on this line, tells us that the
actual pronunciation of his time was ‘hocc erat’ (54. 6 K. ergo
scribendum per duo ¢, ‘hoc-c-erat alma parens,’ aut confitendum
quaedam aliter scribi, aliter pronuntiari). Plautus uses the
proper ‘doublet’ of all these monosyllables which have a vowel
naturally short followed by a consonant that represents two con-
sonants; thus he invariably makes ¢s, ‘thou art’ (I.-Eur. *es-s(i),
ch. viii. § 2), a long syllable before a word beginning with a vowel
(unless under the operation of the Brevis Brevians Law, just as
we find 4é¢ in a line like Men. 522 quid hic ést negoti?). But
in Terence es is a short syllable, and so in Lucilius (e.g. iv. 4 M.)
(On the reduction of a final double consonant, see ch. ii. § 133).

The shortening of s7 in st-quidem, and (in the older poetry only)
of tu, té, mé, &c. before quidem (e.g. tiiguidem, Lucil. xiv. 26 M.,
Plaut. Epid. 99), is due to accentuation, and should be under-
stood in connexion with the rule that antepenultimate syllables
could not be circumflexed in Latin (ch. iii. § 2, p.153), and with the
modern Italian practice of diphthongizingaparoxytone vowel, e.g.
buono (Lat. dénus), but not a proparoxytone, e.g. popolo (Lat.
popilus). The shortening of English sheep, know in shep-herd,
shepherdess, know-ledge, is similarly due to accentual conditions,
and of Welsh brawd, ‘a brother,’ in brod-yr, ‘ brothers,’ &c.

§ 52. Loss of Final Syllable with -m. This could hardly
take place except in the case of words closely joined in ordinary
talk with a following word [e.g. noen(um) est, noen(um) habet,
nihil(um) est, nihil(um) habet, would be the ‘doublet’ forms
§ 52.] ACCENTUATION. FINAL SYLLABLES. 217

before initial vowels, noenum dat, nihilum dat, the ante-con-
sonantal doublets], so that the theory which explains adverbs in
-ter, e.g. breviter, as Accusatives Sing. Neut. of adjectives with
the ‘ comparative’ suffix -¢ero (ch. v. § 18), for dreviter(wm), &c., is
unlikely to be correct (seech. ix.§ 2). The Preposition circum, when
compounded with a verb beginning with a vowel, has its final
syllable not entirely elided but left in prosodical hiatus, e.g. cir-
cu(m)it (a trisyllable); (cf. sublatuirt for sublatum iri, ch. viii. § 89),
and it is possible that xon,n7(47)2, should not be referred to noen(um),
nihil(um), but should receive another explanation, such as zon for
noe-ne (on noenii, see ch. x. § 18), nihil for *ni-hile, Neut. of I-stem
(cf. imbellis and imbellus, subtél from *sub-téle, Neut. of an I-stem
compound of talus (?), and see § 49). But vénire seems to repre-
sent venum ire, though pessum ire did not become *pessire. The
grammarians defend the spelling sed against set by a reference
to an older sedum (Charisius, 112. 5 K.; Mar. Vict. 10. 13 K.)
(see ch. x. § 5). (Ondonec and O. Lat. donicum, see ch. x. § 12).
218 THE LATIN LANGUAGE.

By comparing the various Romance words for, let us say,
‘horse,’ Ital. cavallo, Span. caballo, Port. cavallo, Prov. cavals,
Fr. cheval, Roum. cal, &c., it is possible to conjecture the form
of the Latin prototype from which they all have descended,
caballus. In the same way we can guess at the early form, what
is called the ‘Indo-European’ form, underlying any cognate
group of words in the various Indo-European languages; e.g.
Lat. mater, Dor. Gk. warnp, O. Ind. matér-, O. Ir. mathir,
O. Slav. mater-, Arm. mair, O. Eng. médor, point to something
like *matér as their prototype. We may similarly trace back
inflexions to an ‘Indo-European’ form, and may out of these
conjectured words and inflexions construct an ‘ Indo-European ’
alphabet.

In the last two chapters we have discussed the pronunciation
and accentuation of Latin, and the phonetic changes of the
language produced under the influence of the accent, or due to
peculiarities (often local and temporal merely) of pronunciation.
In the next chapter we shall compare Latin with the other lan-
guages of the Indo-European family ; we shall investigate the form
in which the various sounds of our imaginary ‘ Indo-European ’
alphabet appear on Latin soil, and how that form differs from
the forms assumed in the various languages of Asia and Europe,
which are classed under the name ‘Indo-European.’ These
languages are: (1) the Aryan, including 1. Indian, ii. Iranian
(Zend, Persian, &c.); (2) the Armenian ; (3) the Greek; (4) the
Albanian; (5) the Italic, including i. Latin, ii. the Umbro-
Oscan dialects; (6) the Celtic, including i. Gaulish, ii. Goidelic
(Irish, Gaelic of Scotland, &c.), mi, Brythonic (Welsh, Breton,
&c.); (7) the Balto-Slavic, including i. Baltic (Lithuanian, &c.),
ii. Slavonic; (8) the Teutonic, including i, Gothic, ii. Scandi-
navian, iii, W. Teutonic (German, English, &c.), (see Introduction
to Brugmann’s Comparative Grammar).
CHAPTER IV.

THE LATIN REPRESENTATIVES OF THE INDO-EUROPEAN
SOUNDS.

A, A.

§1. A. I.-Eur. 4 is Latin @ Thus in the declension of
A-stems we have Latin familias (the old genitive preserved in
legal language, pater familias), terrdt (later terrai, terrae), praeda
(O. Lat. praidad), fihidrum, filiabus (another legal form, required
for distinction from jiliis, Dat. Abl. Plur. of jiZivs); the word for
‘mother,’ I.-Eur. *mater- (O. Ind. matér-, Arm. mair, Dor. Gk.
parnp, O. Ir. mathir, O. Slav. mati, with o as the equivalent of
I.-Eur. 4 in Lithuanian and in the Teutonic languages, Lith.
mote, ‘wife, O. Eng. médor, O. H. Germ, muoter, now Mutter,
with short vowel and double consonant instead of long vowel and
single consonant) is in Latin mdter.

I.-Eur, 4, Lat. @, is often found in developments from simple
roots like gén-, ‘to beget,’ e.g. Lat. gndtus, later ndtus, beside
indi-géna, gén-us: tel-, ‘to carry,’ Lat. ditus for *tlatus, P. P. P.
of tollo; stel- (O. Slav. stelja, ‘I spread’), Lat. datus, wide,
earlier status, stlata, sc. navis, whence the adj. stldtarius, or with
-ditt- for earlier -d¢- (ch. ui. § 127), stdatta, stlattarius (Paul. Fest.
455.1 Th. stlatta, genus navigii, latum magis quam altum, et
a latitudine sic appellatum, sub ea consuetudine, qua ‘stlocum’
pro locum, et ‘stlitem’ pro litem dicebant; Gl. Philox. stlata:
metparixod oxddovs eidos: Juv. vii. 134 stlattaria purpura); ster-
(Lat. sterno), Lat. stra-tus, stra-men; Ser-, ‘to rub, ‘wear
away,’ ‘make old’ (Gk. yépwy), Lat. grd-num ; keld-, ‘to strike’
(Lat. per-cello), Lat. clades; hert-, ‘to bind,’ ‘weave together’
(O. Ind. ecrtéti, ‘he binds,’ krnatti, ‘he spins’), Lat. erates ; ker-
220 THE LATIN LANGUAGE. (Chap. IV.

(Gk. xépas, horn), Lat. erdbro for *erds-ro (§ 152), a hornet.
The /é, ra, na has been variously explained in some or all of
these instances, as (1) long sonant or syllabic 1, r, n (§§ 81, 92),
(so Brugmann, Grundy. i. §§ 253, 306), so that, for example, Lat.
granum would represent I.-Eur. *8rno-, while Goth. kaurn, Eng.
corn, represent I.-Eur. *Sino- (ef. Lat. réd-ax, I.-Eur. *wrd-, but
Goth. vaurts, Eng. wort, I.-Eur, *wid-): (2) due to the fusion of
an e-sound with an a-sound in a grade of a dissyllabic root of the
form gena-, &c., so that e.g. Lat. gndtus would come from gena-,
the root Sen- with the addition of an a-sound, while Gk. (Att.
and Dor.) -yyyros would come from gene-, the root Sen- with
the addition of an e-sound (so Bechtel, Hawptprobleme, p. 203);
the *erds- of Lat. erdbro for *cras-ro will thus be a grade of
I.-Eur. *keras- (Greek xépas-): (3) a secondary root, formed by
the addition of a stem-suffix 4 to the weak grade of the simple
root (see Brugmann, Morph. Unt. i. p. 1; Persson, Wurzeler-
weiterung, p. 91), so that e.g, Lat. gudrus would show a root
formed from @n-, the weak grade of the root gen-, ‘to know,’
by the addition of the suffix 4, as gndtus, gndsco would show
a root similarly formed from én- by the addition of the suffix 6;
Lat. /atus, ‘carried, for *¢/dtus, will thus be like Dor. Gk. é-rAd-v
from root tel-, ‘ to carry,’ é-z7d-v from root pet-, ‘ to fly.’

Latin @ is often & lengthened by ‘compensation’ (§ 162),
e.g. qudlus for *quiis-lus (cf. qudsillus) ; halo for *déns-lo, from the
root an-,‘to breathe,’ with the addition of s, ans-, ‘ to be fragrant’
(O. Sl]. achati, ‘ to be fragrant ’), with compound &-hélo from *an-
enslo, with change of & to 2 before the vowel became lengthened
by ‘compensation’ (ch. iv.§ 162). The older spelling was with
(cf. quallus, ankellus, and anhellitus in Virgil MSS.), so that the @
is due to the shifting of the long quantity from the consonant to
the vowel. In octdvus from octé we seem to have before v an
a developed from an 6, just as in cdivus we have dv for dv (§ 19).

Unaccented @ remains unchanged, e.g. immdnis, from an old
word, mdéaus, good; but when final, or when preceding final m,
t, 7, 2, it was, like other long vowels, shortened in course of time.
When final, perhaps only in iambic words, originally (ch. iii. § 43),
ferd, noun, piiti, imper.; but this shortening was extended to all
Noms. Sing. of A-stemsand Noms. Accs. Plur. Neut, (see ch. vi. §§3
§§ 2,8.] REPRESENTATIVES OF I-EUR. SOUNDS. A, A. 221

and 45). Thus in the declension of A-stems, the ‘First Declension,’
final -a of the Nom. Sg. is even in the earliest poetry a short
vowel; the Acc. Sg. has -dim; the final syllable of the 3 Sg. Pres.
Subj. Act., e.g. mitéat, and Pass. mittar was shortened in the second
cent. B.c.; -ad (older -d/e) was also shortened (see ch. iii. § 49),
In Umbro-Oscan IJ.-Eur. 4 was likewise retained (von Planta,
1. p. 77), e.g. Ose. maatreis, Umbr. matrer ‘matris’; Ose.
fratrim ‘fratrum, Umbr. frater ‘fratres.? But final -a became
an QO-sound, written in Oscan @ (in Lat. alph. 0, in Gk. alph.
o), in Umbr. u (in Lat. alph. 0) and a, e.g. Osc. molto, Umbr.
mutu and muta, ‘multa’ (‘a fine’), Osc. vii.‘via.” It is scanned
(in the Neut. Pl. of an O-stem) as a short syllable by Lucilius
(soll6, Lucil. ine. 106 M.; cf. ch. i1. § 1), so that I.-Eur. final -& may
have been modified at a very early period in the Italic languages
(Latin as well as Umbro-Osc.), and the Latin shortening may not
have been confined originally to iambic words (but see ch. iii. § 43).

§ 2. Latin 4 for I.-Bur. 4. Other examples are (1) in suffixes, &c. : L.-Eur.
a of the Subjunctive, Lat. ramus, ferdtis, &c. ; I.-Eur. noun-suffix -tat- (e. g.
O.Ind. déva-tat-, ‘divinity,’ Dor. Gk. ved-ra7-), Lat. nivitat-, viluptat- ; I.-Eur.
adjective-suffix -ako- [e.g.Ir. buadhach, ‘victorious’ (from buaid, ‘victory’; ef.
Boudicea, wrongly called by us Boadicea), Gaulish Teuto-bodiaci, Bén-acus,
Lith. satd6kas, ‘sweetish,’0.SI. novaki ; ef. Gk. védt], Lat. mérdcus, verac-; (2)
in individual words : I.-Eur. *bhrator-, ‘brother’ (O.Ind. bhratar, Gk. pparwp,
the member of a gparpia, O. Ir. brathir, W. brawd, Goth. bripar, O. Eng.
bropor, Lith. broter-éli-s), Lat. frater ; I.-Eur. *bhago-, ‘beech-tree’ (Dor. Gk.
payés, O. Engl. bic-tréow, ‘beech tree,’ bic, ‘a book,’ lit. the runes scratched on
a piece of beech-wood), Lat. fagus; I.-Eur. *swadu-, ‘sweet’ (O. Ind. svadt-,
Dor. Gk. ddvs, O. Sax. swoti), Lat. sudvis for *suddvis; similarly Lat. clavis
(Dor. Gk. #AG(F)is), navem Ace. (O. Ind. nav-am, Hom. Gk. vf(F)-a, O. Ir. nau),
Sari, fama, fabula (Dor. Gk. pa-pi, O. Sl. ba-jati, ‘to converse’), vdtes (O. Ir. faith,
the i being due to ‘Infection,’ that is, to the influence of an 7, which was
suppressed in pronunciation in a following syllable, from stem fati-, I.-Eur.
*wati-).

§3. A. L-Eur.%is Latin d Thus I.-Eur. *486, ‘I drive’
(O, Ind. ajami, Gk. &yw, Ir. agaim, O. Isl. aka inf.), is in Latin
ago; its, derivative, I.-Eur. *igros, ‘a field’ (O. Ind. 4jra-s, Gk.
dypés, Goth. akrs, Engl. acre), is in Latin dger, stem dgro-.

J.-Eur. 4% varies with 4, and similarly Latin 4 with 4, in this
root ag-, ‘to drive’ (Lat. amb-dges, Sanscr. aji-, ‘a race, contest,’ Ir.
ag, ‘a contest ’), and in others, some of which are enumerated in
§ 56. The P. P. P. of std, from root sta-, ‘ to stand,’ is stétus (Gk,
222, THE LATIN LANGUAGE. {Chap. IV.

otards), with % for the weak grade of a. Latin & (probably
Eur. %) is also the vowel of a weak grade of 6, e.g. in a root
like d6-, ‘to give’ (Gk. da-vos, 6-pov), Latin da-tus beside dé-num.
Like Lat. dé-tus from root dd- is Lat. sé-tus from root sé-,
‘to sow’; and this @ seems to be an Eur. 4, a weak grade
of 6, eg. in root kéd- (Gr. éxexyde. and xexddovro, Lat. cédo).
This L.-Eur. vowel, found in a weak grade of roots with 4, 6, 6,
whether it was in each case %, or in some or all cases was an
indeterminate vowel (written 9 by Brugmann), appears in Latin as
a, stdtus, ditus, situs, but in O. Ind. we have i in sthitds, 4-di-ta
3 Sg. Aor., -dhitas P. P. P. of dha- (I.-Eur, dhé-), ‘to place.’ The
same Q, Ind. {is seen in words like I.-Eur. *pater-, O. Ind. pitar-,
where in the other languages we have %, Gk. zarjp, O. Iv. athir,
Goth. fadar, O. Engl. feder, as @ in Latin pater, probably
a derivative from the root pa-, ‘to protect,’ with this weak-grade
vowel. In other words, like Latin péteo, d seems to vary with
& (Gk. werdvyvpr); and in Latin we have a few instances of
d, where other languages, or kindred Latin forms, offer 8, frango
fragilis (Goth. brikan, Engl. break), flagro (Gk. pdAéyw), gradus
(Goth. grips), aver (O. Engl. eofor, Germ. Eber), They are mostly
cases of & with a liquid or nasal, and so admit of the explanation
that they are a form of the sonant or syllabic 1, r, m, n (so
Osthoff, Morph. Unt. vol. v. pref.), while aper has been explained
as ‘a contamination ’ of I.-Hur. *fapro- (Gk. xdémpos) and J.-Eur.
*epro; they have also been explained by the theory that I.-Eur.
é and 5 when pretonic became @ in Latin (Wharton, Hiyma
Latina, p. 128). The more or less complete fusion of I.-Hur.
% and 6 in other languages makes it impossible to be sure that
this use of @ in words connected with é-roots is not a peculiarity
of Latin, or rather of the Italic languages generally (cf. Ose.
patensins, Umb, abro-), depending, it may be, on the pronunciation
of Latin or Italic a (see ch. ii. $1). For Latin & for 3, under influ-
ence of », e.g. cévus, older covus, see § 19, and for ar, al, an,
from sonantr, 1, n, §§ 81, 92.

Unaccented Latin % in the posttonic syllable became at first é,
except before 7 and labials, where it became 3. This @ became,
perhaps about the end of the third century B.c., 7 in syllables

not long by position (except when it preceded r), and before xg;
§§ 4, 5.) REPRESENTATIVES OF I.-EUR. SOUNDS. £,E. 223

while this 0 became w or the #-sound, which in most cases passed
into % at the close of the Republican period. Thus the compound
of ab and cado became accédo (so spelt by Ennius), then accido ;
from iz and arma we have the compound izermis; from sub
and rapio first *surropio probably, then surripio (Plaut.), then
surrigno; from ex and frango, effringo (see ch. iii. § 18). Final Latin
& probably became @, and might be dropped (see ch. ili. § 37).

In Umbro-Oscan I.-Eur. % remains, as in Latin, e.g. Umbr.
ager, ‘afield, Osc. actud ‘agito’ third Sg. Imperat., also I.-Eur.
a, eg. Ose. pateref ‘patri,, Umbr. Iupater ‘Juppiter’ (von
Planta, i. p. 75).

§ 4. I.-Eur. 4 The I.-Eur. preposition *4pd (O.Ind. dpa, Gk. do, Goth.
af, Germ. ab, Engl. of) is Latin dp- of ap-ério, usually written db, with suppres-
sion of the final vowel ; but the form pé- of yo-situs, from pono for *pd-sino, shows
suppression of the initial vowel ; *déd (0. Ir. ad, Goth. at, Engl. at) is Latin dd.
The I.-Eur. pronoun *alyo-, ‘other’ [Gk. dAdos, O. Ir. aile, Gaul. Allo-broges,
‘those ofanother country ’(Schol. Juven. viii. 234), (as opposed to *Combroges,
‘native,’ whence Welsh Cymry), W. all-, Goth. aljis, Engl. el-se] is Latin diius.
Similarly médeo (Gk. paddw) ; silico (Gk. GAAopanr) ; ‘siilix (Ir. sail, a C-stem,
Bret. haleg-en, O. H. Germ. salaha, O. Engl. sealh, Engl. sallow) (but see
§§ 92-94) ; ddcriima, later lacruma and lacrima (Gk. ddxpv, O. Ir. dér, W. dagr,
Goth. tagr, O. Eng. téar, Germ. Zihre) ; ango, angor, angustus (QO. Ind. dhas,
‘need,’ Gk. dyxw, Ir. t-achtaim, W. t-agu, Lith. aiksztas, ‘narrow,’ O. SI.
azaki, Goth. aggvus, Germ. enge) ; arceo (Gk. dpxéw, Arm. argel, ‘ hindrance’) ;
macer, ‘thin’ (Gk. paxpés, long, Av. masah-, ‘size,’ 0. H. Germ. magar, ‘ thin’) ;
albus (Gk. ddgés, white leprosy) ; animus, dnima, ‘soul’ (O. Ir. anim, anman
Gen., ‘soul,’ Gk. dveyos, wind, from root an-, ‘to breathe’) ; cdno (O. Ir. canim,
W. canu, Goth. hana, ‘a cock,’ Engl. hen) ; dio (Ir. alaim, W. alu, Goth. ala,
‘I grow up,’ Gk. dv-aAros, insatiate) ; dqua (Goth. ahva); scdbo (Gk. oxarrw,
Lith. skabi, ‘I cut,’ Goth. skaba, ‘I shave,’ O. Engl. scafe, Engl. shave) ; dro
(Arm. araur, ‘a plough,’ Gk. dpéw, O. Ir. arathar, ‘a plough,’ W. ar, ‘ tilth,’
Lith. arit, ‘I plough,’ O. Sl. orja, Goth. arja, Engl. to ear) ; sal- (Arm. ad, Gk.
dds, O. Ir, salann, W. halen, O. SI. soli, Goth. salt, Engl. salt).

I.-Hur, & or e (see § 51).

3—6 (gee § 55) e.g. atrox and odium, acer-bus and ocris.

a—€ (see § 61) e.g. aser blood (GE. éap), sacena a priest’s knife (cf. seco).

E, E.
§5. E. I-Eur. é is Latin 2, Thus the optative-suffix, I-Eur.
-ié- (-iyé-) (O. Ind. syas, siyas, Gk. eins for *éous) is -ié- of
O.Lat.siés. From the root plé-, ¢ to fill’ (O. Ind. pri-ta- Part., ‘full,’

Arm. li, Gk. 1A7-pys, O. Ir. lin, ‘ number,’ O. Isl. fleire, ‘ more’)
comes Latin plé-nus, im-plé-tus, plért-que,O. Lat. ea-plé-nunt ; from
224. THE LATIN LANGUAGE. [Chap. IV.

sé-, ‘to throw, throw seed’ (Gk. ¢nus for *o1-07-p2, 7-yo. for *sé-mn,
O. Ix. sil, ‘ seed,’ W. hil, Goth. mana-séps, ‘ mankind,’ Engl. seed,
Lith. séju, ‘I sow,’ O. SI. séja, sé-me, ‘ seed’) Latin sé-v2, sé-men,
This I.-Eur. é is often found in developments from simple roots
with &, as, for instance, plé-, from the simple root pél-, ‘to fill’
(Goth. filu, ‘much,’ O, Tr. il), or psé- (O. Ind. psa-, ‘ to devour,’ Gk.
ariv, to rub) from the simple root bhés-, (O. Ind. bhas-, ‘to devour’),
the & being either due to the fusion of & with an e-sound, in
a grade of a dissyllabic root (thus plé- would be a grade of pele-),
or a stem-suffix added to the weak grade of the simple root (thus
psé- is ps-, the weak grade of bhes-, with the addition of the
suffix é). The same doubt we found to exist about roots with
a, like 6na- (Lat. gud-tus) from gena- or gn-i- (§ 1). Occa-
sionally é became 7 in Latin through the influence of an 7 (y)
followed by a vowel in the next syllable, e.g. f2dius for *félins.
Latin é is often @, lengthened by ‘compensation,’ e.g. duhélus
for *an-éuslo from *an-dnslo- (cf. halo), written in the older
orthography (in Virgil MSS.) anhellus, a spelling which indicates
the lengthening of the ¢ to have been a transference of the long
quantity from the consonant to the vowel. Sometimes Latin é is
due to the fusion of two vowels, e.g. préndo from prehendo,
trés from *tréyés (ch. vi. § 61).

In the unaccented syllable, Latin é remained unchanged,
e.g. concédo, accédo (see ch. iii. § 30). But when. final, it was
shortened in iambic words in course of time, so that while
Plautus scans cévé and occasionally cdévé, the ordinary pronun-
ciation in Cicero’s time was cévé only. When preceding final
m, it was shortened like other long vowels; hence the first
Pers. Sg. of the optative would be siém in Latin, unlike Gk. einy
for *é(o)inv; and before final -¢, -r, -2 it became (like d, &c.)
a short vowel in the second century B.c. (For this shortening,
see ch, ili. § 40.)

In Oscan I.-Eur é is i (the symbol also of I.-Eur. 1, § 13),
fi (Lat. alph. 2), e.g. ligattis ‘legatis, ligud ‘lege’; in Umbr.
e, sometimes i, e. g. plener ‘plenis, habetu and Aabitu ‘ habeto’ (von
Planta, i. p. 89).

§ 6. Lat. 6 for I.-Eur. 6. Other examples are: I.-Eur. dhé-, ‘to suck,’
‘suckle’ (O. Ind. dha-, dha-ri-, ‘suckling,’ dha-tri, ‘nurse,’ Arm, diem, ‘I suck,
§§ 6-8.] REPRESENTATIVES OF I-EUR. SOUNDS. #,B. 225

Gk. 67j00a1, 67-Avs, On-AH, O. Ir. dith, ‘he sucked,’ dinu, Pres.-Part., ‘a lamb,’
Goth. daddja, ‘I suckle,’ 0. H. Germ. tau, Lith. dé-lé, ‘a leech,’ pirm-délé,
‘young mother,’ O. Sl. dé-te, ‘infant’), Latin fé-mina, fello (vulgar form of félo),
Fi-lius for *felius; I.-Eur. né-, ‘to sew, spin’ (Gk. viv, vq-0w, vij-ya, vij-tpov, Goth.
né-pla, ‘needle,’ 0. H. Germ. nadela, naen, Germ. nihen), Lat. né-re, né-tus, né-
men ; I.-Eur. *sémi-, ‘half’ (O. Ind. sami-, Gk. jui-, O. H. Germ. sami-, 0. Engl.
sam_-, Engl. sand-blind), Lat. s¢mi-. Similarly Lat. rés (O. Ind. ra-s, ‘property’);
Lat. vérus(O. Ir. fir, W. gwir, Goth. tuz-vérjan, ‘to doubt,’ 0. SL véra, ‘ belief’) ;
Lat. ré-ri (Goth. rédan, ‘ to advise,’ 0. Engl. rédan, Engl. rede, Germ. rathen);
Lat. spes (Lith. spéti, ‘to have leisure,’ O. Sl. sp&ti, ‘ toadvance,’ Goth. spédiza,
‘later,’ Germ. spit). This I.-Eur. é is often a ‘doublet’ of éi (see § 47); ré-, for
example, of Lat. rés, O. Ind. ra-s, isa byform of réi-, réy- (O. Ind. ray-d4s Gen.),
and some refer the fi- of Lat. filius to an I.-Eur. dhi-, a grade of a root dhéi-,
dhéy, ‘to suckle.’ That the Romans of Plautus’ day regarded filius as a cognate
of félo (fello) appears from a line preserved only in the Ambrosian MS., Pseud.
422 iam ille felat filius, and in Umbrian the word seems to have had the
sense of ‘suckling,’ e.g. sif filiu trif, tref sif feliuf, ‘tres sues lactentes’
Ace., as wellas that of ‘son,’ e. g. fel. for felis, ‘ filius’ on an Umbrian epitaph.
(Biich. Umbr. p. 174.) (On Praenestine file(d)a, a nurse (?), see A.L.L. ii. 482).

§ 7. i for 6. Delinio, a byform of delénio; Plinius (dialectal?) apparently
from plénus ; convicium from root wéq3-, ‘to speak’ (?); suspicio from root spék-, ‘to
look,’ all seem to be examples of this change of é to 7, produced by a y-sound
in the next syllable. Filius is spelt felius on an inscription (C. I. L. xiv. 1o11),
and seems in Umbrian to have the é-sound, spelt ¢ or 7. On the spellings
Cornilius, Aurilius, which prove the affinity of Latin @ with an i-sound before
a syllable with y, see ch. ii. § 11 (Aurilius occurs on an inser. of 200 B.c. (C. I. L.
xiv. 4268, with cisdim) ] ; and on the spelling stilio, for stéllio, a newt, see Georges,
Lex Wortf. s.v. (Parodi in Stud. Ital. i. 385 gives other exx., and adds tilia, &c.)

§8.E. J[-Eur.é is Latin @: for example, in the present stem,
e.g. of the root bhér-, I.-Eur. *bhéré, first Pers. Sg. (O. Ind.
bh4rami, Arm. berem, Gk. dépw, O. Ir. berim, W. ad-feru Inf.,
Goth. baira, Engl. I bear, O. S]. bera), Lat. féro ; in the Neuter
ES-stem, e.g. of root gén-, I.-Eur. *genos, Nom. Sg. (O. Ind.
Janas, Gk. yévos, O. Ir. gein), Lat. génus; in the numeral ‘ ten,’
,L-Eur. *d&km (O. Ind. daa, Gk. déka, O. Ir. deich, W. deg, Goth.
taihun, O. H.G. zehan, Lith. dészimt, O. $1. deseti), Lat. décem.

Latin em, en may represent J.-Eur. m, n, the sonant or syllabic
nasal, e.g. I.-Eur. *kmtom (or *iemtom), O. Ind. Satém, Gk.
é-xatov, O. Ir. cét, W. cant, Goth. hund, Lith. szifitas, O. SI.
stito), Lat. centwm (see § 81). I.-Eur. éw became dv in Latin,
which in the unaccented syllable passed into u, e.g. ndvus (Gk.
véos) and dé-nuo, and similarly 1.-Eur. wé became 4, e. g. sdror
for I.-Eur. swésor-, in certain circumstances (see below), and I.-Eur.
el became 07 (i) except before e, 7 or in the group el/, e.g. volo,
but vélim, velle. Latin ¢ became 7 before ng, e. g. tingo (Gk. réyyo),

Q
226 THE LATIN LANGUAGE. (Chap. IV.

before gn, e. g. dignus from decet (? see§ 119), just as in the Teutonic
languages &§ has become i before a nasal and a consonant, e. g.
Engl. ‘wind.’ It became 7 also in open unaccented syllables, except
when final, or when preceding 7, but passed into a «-sound before
Zora labial. Thus the compound of xéco was é-nico (later spelt
eneco), the ordinal of décem was déctimus, class. decimus (ch. iii. § 18).

Latin @ sometimes represents ei (éy) before a vowel, e.g. é
from the I.-Eur. root ei-, ‘to go,’ on which see § 63. Final Latin
é may represent any I.-Eur. short vowel, as may also ¢ in unac-
cented syllables before 7 or a consonant-group, e. g. pé-pér-i from
pario, an-ceps from cdiput (see ch. iii. § 18). Final -¢ was often
dropped, e. g. née for négqué, exemplir, older exemplare, as & (and 7)
in the middle of a word might be suppressed by syncope, e. g.
surgo for sub-régo (see ch. iii. § 13). On the substitution of -ér- for
-ri- in ter, older terr (cf. terr-wucius) for *ters, I.-Eur. *tris, &c.,
see ch. ii. § 15. 8. J.-Eur. é remains in Umbro-Oscan, though
before some consonants it appears as ¥, e.g. Osc. estud ‘ esto,’
Umbr. fertu ‘ferto” The change to 5 before ] is apparently
unknown. (For particulars, see von Planta, i. p. 83.)

§ 9. Latin é for I.-Hur. &. I.-Eur. -é in the Voc. Sg. of O-stems (O. Ind.
vtka, Gk. Av«e, Lith. vilka, O. Sl. vliée) is Latin -¢ of lupé, &., as in the
Imperative 2 Sg. Act., e.g. *488 (O. Ind. dja, Gk. dye) Lat. dg?, dropped in dic,
dic, fac (ch. iii. § 36) ; the conjunction ‘and,’ I.-Eur. *q3é (O. Ind. ca, Gk. re) is
Latin qué, with -¢ dropped in née for négué, &c. ; the first personal pronoun
(0. Ind. ahém, Arm. es, Gk. éy#, Goth. ik, O. Eng. ic, Lith. &sz, 0. SI. azii)
is in Latin égo. Other examples are Lat. nébitla (Gk. vepéAy, O. Iv. nél from
*neblo-, W. nifwl, 0. H.Germ. nebul, Germ. Nebel) ; Lat. st¢quor (O. Ind. sac-,
Gk. éropat, O. Ir. sechur, Lith. seku) ; Lat. e (Gk. gre) ; Lat. médius (O. Ind,
madhya-, Gk. péooos, péoos,Goth. midjis, O. Ir. medin, ‘the middle,’ 0.81. mezda;
Lat. équus (O. Ind. dkva-, QO. Ir. ech, Gaul. Epo-rédia, W. ebol, ‘a colt,’ Goth.
aihva-tundi, ‘a bush,’ lit. “horse-tooth,’ 0. Engl. eoh, Lith. aszva ‘a mare’); 0
Lat. vtho (O. Ind. vah-, Pamphyl. Gk. Féxw, Ir. fon from *wegno-, ‘a waggon,’
Goth. ga-viga, ‘I move,’ Germ. be-wege, Lith. vezi, O. SL. veza).

§ 10. 6 for 6 with w and 1. I.-Eur. *néwn (0. Ind. nava, Gk. évvéa for
*éy-veFa, Ir, noi, W. naw, Goth. niun) is in Latin nivem; I.-Eur. *néwo-
(0. Ind. néva, Gk. véos, 0. Ir. ntie from *nowio-, Gaul. Novio-dinum, W.
newydd) is Lat. névus, Novius, while in the unaccented syllable we see % for
I.-Eur. ew in déntio. (For other examples of this u, see ch. iii. § 24.) The change
of é to 6 before w, which is shared by the Celtic languages (e.g. Gaul. Novio-
dunum, O. Ir. nie for *nowio-, W. newydd from *nawydd for *nowio- from
I.-Eur. *néwio-, cf. Gk. veds, Goth. niujis; O.-Ir. noi, W. naw from I.-Eur.
*newn), and by the Balto-Slavic (e.g. Lith. tivas for I.-Eur. *téwo-, Gk.
re(F)és; O. Sl. novi for I.-Eur. *néwo-, Gk. vé(F)os), does not affect the ev
(Latin ev, but not I.-Eur. -ew-) of words like vis (Gk, é-Aaxvs), stverus,
§§ 9,10.) REPRESENTATIVES OF I.-EUR. SOUNDS. £,#. 227

apparently from root ségh-, brévis (Gk. Bpaxvs), so that the law of change must
have ceased to operate before these words assumed in Latin this form. It is
like the change of the diphthong eu to ou in the Italic, Celtic, and Balto-
Slavic languages, e.g. O. Lat. douco for I.-Eur. *deukd (Goth. tiuha)
(see § 35).

I.-Eur. swé- appears as sd- in Latin, e.g. I.-Eur. *swésor- (O. Ind. svasar-,
Gk. éop-es, O. Tr. siur, and after a vowel fiur, W. chwaer, Goth. svistar, Lith.
sesti, O. SI. sestra) is in Latin sdror ; I.-Eur. *swéktiro- (0. Ind. Svasura-, Gk.
Fexupés, W. chwegrwn, 0. H. Germ. swehur, Germ. Schwaher, Lith. szeszuras,
O. Sl. svekrit) is Lat. sécer ; I.-Eur. *swépno- (0. Ind. svapna-, 0. Scand. svefn,
O. Engl. swefen) is Lat. simnus for *sipnus; ef. sipor. I.-Eur. kwé is said to
appear as cé- in Latin in combr-étum, a bulrush (Lith. szvefidrai Plur.) from a
stem kwéndhro-, though this may stand for *quombr-étum with the O-grade of
stem (see § 137) ; forem seems to represent dissyllabic *fwtrem. But ¢ of dwé-
remains, e. g. bellum, older duellum, béné (cf. older Duenos). Qué from I.-Eur.
qzé, kwé, &c. remains, and does not become co, e. g. -qué (I.-Eur. q28), quéror
from kw-és- (ef. Gk. xwx¥w), though quo became co in course of time, e.g. célo,
the O. Lat. form of which was quolo, as in the old inscription of the Faliscan
‘collegium cocorum,’ written in rude Saturnians, and with equally rude
spelling (Zvetaieff, Inscr. Ital. Inf. 72a):

gonlegium quod est aciptum aetatei aged[ai],

opiparum ad ueitam quolundam festosque dies,

quei soueis astutieis opidque Uolgani

gondecorant saipisume comuiuia loidosque,

ququei huc dederunt inperatoribus summeis (i. e. Jupiter, Juno
and Minerva)

utei sesed lubentes beneiouent optantis,

where also cogui is written ququei (so golunt for colunt in the Plautus Palimpsest
in Pseud. 822). The compound inquilinus, with -quil- for quél- in the unaccented
syllable, was formed before the change from quo to co, and being a legal term
kept its old spelling, unlike incila, (but see p. 229). That qué had come to sound
like cé as early as the beginning of the second cent. B.c., we may infer from
the spelling in oquoltod, for in occulto, on the S. C. de Bacchanalibus (C. I. L.
i. 196) of 186 B.c. ; for occiilo must be connected with célo, which has not the
qi- guttural (cf. Ir. célim, W. célu with the é-grade of the same root). The
analogy of quam, quem, &c. would preserve the spelling quom till a late date,
though the word was probably pronounced *com, for the preposition, I.-Eur.
*kém or *kém, is usually spelt quom till the time of the Gracchi (Bersu,
Gutturale, p. 42); and similarly loquontur, &c. would be written after the
fashion of lquantur, loquentur, with quo; so that it is not until the fifth
cent. A.p. that every qué has assumed the spelling co, e. g. cot, corum, coque
(the conjunction), condam (the adverb), locor (Bersu, p. 90). The form
quotidie is censured by Quintilian (i. 7. 6 frigidiora his alia ut... ‘ quotidie,’
non cotidie, ut sit quot diebus: verum haec jam etiam inter ipsas ineptias
evanuerunt), by Velius Longus (79. 16 K. illos vitiose et dicere et scribere
{qui potius] per ‘quo’ ‘quotidie’ dicunt quam per ‘co’ cotidie, cum et
dicatur melius et scribatur. non enim est a quoto die ‘quotidie’ dictum, sed
a continenti die cotidie tractum), and by Marius Victorinus [13. 21 K. nam
concussus quamvis a quatio habeat originem, et cocus a coquendo (v.i. quo-

Q2
228 THE LATIN LANGUAGE. [Chap. Iv.

quendo), et cotidie a quoto die, et incola ab inquilino, attamen per c quam
per qu seribuntur]. Coftidie and cotidée are the spellings of the best MSS., and
are found on inscriptions (see Georges, Lex. Wortf. s. v.), though no doubt the
older spelling would have quo-.

Lat. vo- became ve- (see Solmsen, Stud. Lat. Lautg. p. 1) in the middle of the
second cent. B.c. Quintilian tells us that Scipio Africanus (Minor) was
credited by tradition with this change of orthography (i. 7. 25 quid dicam
‘vortices’ et ‘vorsus,’ ceteraque in eundem modum, quae primus Scipio
Africanus in e litteram secundam vertisse dicitur?). (On these spellings in
inscriptions, see Brambach, Orth. p. 101.) The MSS. of Plautus show the
older spellings vorto, vorsus, and compounds, voster, voto (I-Eur. g8-), -vorro; and
invorto, divorst, vortec are found even in the MSS. of Augustan poets, like
Virgil. The grammarians of the Empire sometimes advised the retention of
these forms for the sake of distinctions, e. g. Caper, 99. 11 K. vortex fluminis
est, vertex capitis; 97. 15 vorsus paginae dicetur, versus participium est
a verbo vertor. One of the o-forms indeed, vister, was retained to the last
(perhaps by analogy of vds, or of néster) in Vulgar Latin (cf. Roumanian vostru,
Ital. vostro, Fr. vétre), aso was retained in classical vico (by analogy of vox ?),
vimo, voro (I.-Eur. g3-). O was retained before single 7 and! before another
consonant, e. g. volo, to wish, volo, to fly, vola, hollow of hand, volvo, volnus,
later vulnus, &c., and before v, e.g. voveo; though Cassiodorus, a doubtful
authority, makes convollere the old spelling of convellere (149. 17 K.). Voldterrae
for Etruscan Velaéri, Vélumnius for Etruscan Velimna cannot be quoted to
prove that Latin vé- was ever pronounced vé-. They exemplify the phonetic
law that & became 61 in Latin (see below). There is no evidence that vého
was ever *voho, or Vénus *Vonus, or véntus *vontus, or véru *voru, &e. The old
spelling vorto (I.-Eur. *wértd, Goth. wairpa) probably belongs to a period
when vo- had come to take the sound of ve- and was occasionally used as
a symbol of this sound (ch. viii. § 8); it has also been referred to the analogy
of the P. P. P. vorsus (I.-Eur. *wrt-to-), where Lat. or represents I.-Eur. r.

Oi may similarly have become ¢i after v, so that vidi may represent an I.-Eur.
*woidai (0. Sl. védé ; cf. Gk. ofSa, ch. viii. § 39) ; but the appearance of é beside 6
in Latin in words like amplector, O. Lat. amploctor (Prise. i. p. 25.15 H.; ef. below
ch, viii. § 33) is better referred to the same ‘ variation’ (Ablaut) as that seen in
tego beside toga, procus beside precor, &c., on which see § 51.

Elis found in the group ell, e. g. velle, vellem (that lJ had a more ‘ exilis’ sound
than J, in technical language was ‘front-modified,’ is attested by the gram-
marians, ch. ii. § 96), and before e, i (y) ; butin other circumstances it seems that
the character of Latin J so asserted itself as to change e¢ to 0, e. g. volo, though
there are a few exceptions to the rule, and not very many instances'. Thus
the Greek éAai(f)a, when adopted by the Romans (in the period of the
Tarquins, Plin. Nat. Hist. xv. 1), became *olaiva, then *oleiva (§ 27), ola.
A following e- or i-vowel prevents the change, e.g. vélim, mélior. The older
type of declension hilus, *héleris (from *heleses) has left traces of itself in holus,
holeris and O. Lat. helus (Paul. Fest. 71.13 Th. ‘helus’ et ‘helusa’ antiqui

 

1 Pliny contrasts the J of lectus, exactly as Russian or Gaelic 1, viz.
tectum with the ‘exilis’ i-sound of a ‘deep’ J before a, 0, u, a ‘palatal’
Metellus, so that we cannot suppose 1 before e, i.

Latin J to have been pronounced
§ 11.] REPRESENTATIVES OF I.-EUR. SOUNDS. £, i. 229

dicebant, quod nunc holus et holera; ef. the gloss ‘helitores’ hortolani Liéwe,
Prodr. p. 339), but *scdlus has not survived beside scéleris. Before a consonant
el became ol (ul § 17) (on the pronunciation of 7 before a cons., see ch. ii. § 96),
e. g. vult, older volt. (On gelu, helvus, stmel, celsus, and for other instances of the
change to ol, see Osthoff, Dunkles u. helles ‘1’ im Lat.). Inquittnus may thus
represent an older *enquelino-, incdla an older *enquold-, both from an early q¥él-.

§ 11. i for (accented) e. Other examples are: before ng, Lat. lingua, older
dingua from I.-Eur. dngh}- (O. Ir. tenge, W. tafod, Goth. tuggo) ; Lat. inguen
from I.-Eur. ng®- (Gk. d5yv) ; Lat. stringo, I bind, draw tight (O. Ir. srengim,
‘I draw’); Lat. seplingenti, confringo, aitingo, &ec. for*septengenti, *confrengo, *attengo,
&c. ; hefore gn, ignis from I.-Eur. *ngni- (O. Ind. agni-, Lith. ugnis, 0. Sl. ogni) ;
Ignatius, a late spelling of Egnatius (see Schuchardt, Vok. i. 334) ; ilignus and
iligneus from tex, but abiégnus (with 2, according to Priscian, i. p. 82.8 H.) from
abies; the old religious term for a sheep, brought with its two lambs to the
sacrifice, is given by Paul. Fest. as ambegna (4. 7 Th. ‘ambegni’ bos et vervex
appellabantur, cum ad eorum utraque latera agni in sacrificium ducebantur),
but in Glossaries as ambigna (Mai, vi. p. 506 b. ‘ambignae,’ oves ex utraque
parte agnos habentes ; and ‘ambignae,’ oves quas Junoni offerebant, quia
geminos parerent), while the MSS. of Varro give ambiegna, which may indicate
a correction of ambegna to ambigna (L. L. vii. 31 ‘ambiegna’ bos apud augures,
quam circum aliae hostiae constituuntur). We have sim- for sem- (I.-Eur.
sm-, Gk. d-rAdos, &c.), in simplus, simplex, simpludiarea funera (quibus adhibentur
duntaxat ludi corbitoresque, Fest. 498. 24 Th.) as well as singuli, sincerus,
sincinia (cantio solitaria, Paul. Fest. 500, 23 Th.) and in simul, older semul (see
Georges, Lex. Worf. s.v.), similis, simita, but e does not become i before mpl of
templum, nor before ne in jivencus, and Umbrian sumel, Gk. dards, &c. suggest
that the sim- of similis and its cognates is I.-Eur. som-, and has the #i-sound
of stimus, written also simus, luibet later libet (see ch. ii. § 16) ; sinciput is derived
from sémicaput by Velius Longus (78. 18 K.), &e. On the quantity and quality
of the vowel i before gn in dignus, &c., see ch. ii. § 144. The I.-Eur. word for
‘five’ *pénq%é (O. Ind. péfica, Arm. hing, Gk. révre, O. Ir. cdic, Gaul. mepné-
Sovda, cingfoil, O. W. pimp, Goth. fimf, Lith. penki, 0.81. peti) is in Latin
quinque [with long i indicated both by inscriptions and by the Romance forms,
a quantity which has been referred to the influence of quin(c)tus (K. Z. xxx.
501) (see ch. ii. § 144)]. In rustic and dialectal Latin ¢ before re became i,
e.g. Mircwrios, Mirqurios (C.I. L. i. 1500 and 59, both from Praeneste), stircus
(C. I. L. ix. 782, from Luceria, in Apulia on the borders of Samnium), com-
mircium, mentioned as an older form by Velius Longus (77. 12 K. ‘mium’ et
‘commircium’ quoque per i antiquis relinquamus, apud quos aeque et ‘ Mir-
curius’ per i dicebatur, quod mirandarum rerum esset inventor, ut Varro
dicit, nostris jam auribus placet per e, ut et Mercurius et commercia dicantur).
The i of country-terms like hirsutus, hirtus, &. (apparently from root ghers-,
‘to be rough,’ whence Lat. horreo, hordewm, &c.) may be explained by this
dialectal pronunciation of stircus for stercus, &c. (cf. Osc. amiricatud ‘immercato’),
In other positions than before re, &c. the ‘rustic’ pronunciation seems to have
substituted efor (ut iota litteram tollaset e plenissimum dicas, Cic. de Orat. iii. 12,
46) (ef. above, ch. ii. §17). To this confusion is perhaps due the uncertainty
in the spelling of country-terms like /iix or flix, a fern (the latter approved by
Caper, p. 106. 1 K.; see Georges 8. v.), fiber and jeder, a beaver, from I.-Eur.
bhébhr- (see Georges) ; but the byforms pinna and penna, vigeo and vegeo, villus
230 THE LATIN LANGUAGE. [Chap. IV.

M. and vellus N. have not yet been satisfactorily explained. Levir, a brother-
in-law, a word only found in late Latin writers, and so misspelt with 2 for
ae (see § 28), from I.-Eur. *daiwer- (Gk. dajp, Lith. déveris) takes 7 by anal.
of vir. (On other byforms due to the late Latin identification of ¢ with 2, and
to the change of 2 to 7 in unaccented syllables, see ch. ii. § 6).

I, I.

§12. I. I.-Eur. i has been faithfully retained by the various
languages in almost all circumstances, and is in Latin 7, though
often written in O. Lat. ¢, after the I.-Eur. ei-diphthong had
come to take the sound of 7 (ch. i. § 9). The diphthong « in
Greek developed to the same sound (thus éreioa, the proper spell-
ing, became érica), so that in Greek also es was in course of time
often written for i, e.g. woAeirys, and Ulfilas adopted this symbol
ei for the long i-sound of Gothic. For examples of I.-Eur. i we
may take the adjective-suffix in -Ino- (O. Ind. nav-ina-, ‘new,’ Gk.
ayxtor-ivos, often with names of animals, e.g. kopak-ivos, deAdax-tvn,
xolp-ivn, Goth. gulp-eins, O. H. G. guld-in, Engl. gold-en, Goth.
sv-ein, O, Engl. sw-in, Engl. swine, O. 81. mater-inii, ‘ motherly,’
sv-inti), in Latin su-tnus, div-inus, &e.; the optative-suffix i
(varying with yé, ch. vii. § 55) (O. Ind. dvigi-mahi, Gk. eideiuer
from <id«(c)-i-vev, Goth. vil-ei-ma) in Lat. s-a-mus; the adjective
*g4iwo-, ‘alive,’ (O. Ind. jiva-, Lith. gyvas, O. 81. Zivti, O. Ir. biu,
W. byw) in Lat. vivus, in old spelling veivos; Lat. vis (Gk. ts,
i-pt); Lat. virus (Gk. ids for *Fivos), Lat. vitew (Gk. iréa and
eiréa, Eng. withy, Lith. vytis). J.-Hur. iis usually a grade of an
ei-root, and it is often difficult to say whether Lat. 7 represents
the i-grade or the ei-grade.

After 7, Latin 7 appears as é in /dni-éna for *lani-ia, &e. (cf.
tonstr-ina), as we have é in sdciétas, anziétas, but % in castitas, ndvi-
‘tas, &e. (see ch. v. § 83). Latin 7 represents J.-Hur. ei in dico (O.
Lat.deico, Gk. deix-vi-p), ad-dico, &e., fido (Gk. re(Ow for * pew),
con-fido, &c. (see ch. viii. § 6), and has come from 7% lengthened
by compensation in words like widus for *nisdo- (*nizdo-) (Arm.
nist, ‘situation,’ Engl. nest). In the unaccented syllable it may
represent older ei (Latin ez, not I.-Eur. ei), viz. an I-diphthong
whose first element has been weakened; as in con-cido, older
con-ceido, from caedo, older caido, and in the final syllable of
Perfects like ¢u-d#id-a (older -ei) (O. Ind. tu-tud-é), which have the
§§ 12,13.) REPRESENTATIVES OF I.-EUR. SOUNDS. Tf, i. 231

I.-Eur. 1 Sg. Perf. Middle ending -ai (-ai) (ch. viii. § 66); oi in
the final syllable of Noms. Plur. of O-stems like populi (oldest
Lat. poploe, then populei, ch. vi.§ 40). Vidi, older veidei, from I.-
Eur. *woidai (O. Sl. védé) shows vei-, a development of an older
voi-, as versus (1.-Eur. *wrt-to-) shows ver-, a development of an
older vor- (§ 10). On the use of é for Lat. 7 (perhaps properly
only I.-Eur. ei) in rustic Latin, e. g. speca, vella, see ch. ii. § 17,
and ef. below, § 32; on @ for 6 in jilius,§ 7. I.-Eur. 1 remains
in Umbro-Ose. aaa is written in the Oscan alphabet if or i (in
Lat. alph. 2), in Umbr. i (in Lat. alph. i and ¢7), e. g. Ose. liimitti[m
‘limitum,’ Umbr. si, siz, sei, ‘sit’ (see von Planta, i. p. 102).

§13. I. 1.-Eur. i is Latin 7, sometimes written in Old Latin
e (e.g. Tempestatebus on a Scipio epitaph), after unaccented ¢ had
come to take the 7-sound (see also ch. iii. § 18). TheI.-Eur. pro-
noun *{- (O. Ind. i-d-A4m Neut., Goth. is Masc., ita Neut., Engl. it)
is Lat. 7s M., td N.; the pronoun *ki-, ‘this’ (Goth hi-mma Dat.,
hi-dré Adv., Engl. him, hither, Lith. szis, O. 8. si) appears in
Latin cis, citra; the interrogative and indefinite pronoun *q%%-
(O. Ind. ci-d, Gk. ri(3), O. SI. G-to) is Latin guts M., quid N. (cf.
Umbr. pis) ; from the root min-, ‘ to lessen,’ from the primary root
mei- (O.Ind. minémi, Gk. pwi-6e, Goth. mins Adv., O. SI.
minjiji Adj.) we have Lat. minuo, minor.

I.-Eur. i is generally the weak grade of the diphthong ei (as
min- from mei-, ‘to lessen’), and so in Latin, e. g. 2-dico beside
dico (older deico), fides (cf. Gk. é-7110-ov) beside f2do, older feido (cf.
Gk. wel0 for *pedw). For i varying with i, e.g. I.-Eur. *wiro,
‘a man,’ see§ 58. Latin 7 may represent any short vowel in an
unaccented syllable, not long by position, e.g. conctno for con-cano,
diligo for dilego, inquilinus for *inguolinus, quidlibet for quidlubet
(whence /ibet for Zubet, ch. ii. § 16). In an unaccented syllable
long by position? became 2, e. g. indéx beside indico, judéa beside
judico (cf. jurt-dicus), cémes, properly *coméss, Gen. com-it-is, and
in an open unaccented syllable before 7, e. g. ctm-er-cs from cinis,
and also when final, e.g. maré but maria, mari-timus ; lévé, Neut.
of Zevis (contrast Gk. idpt, Neut. of tSpis), ruré, Loc. of ras, in which
case it may be dropped, e. g. dntmdl for animale, Neut. of animédlis.
Latin -ri- in the unaccented syllable, when preceded by a con-
232 THE LATIN LANGUAGE. (Chap. IV,

sonant, became é, e.g. dcerbus for *acrt-bus from *acrii-dho-
(Lith. asztris, O. SI. ostrit), xcertus for *inerttus (Gk. a-xpitos) ;
and apparently this may occur in the accented syllable too, e. g.
ter, properly ¢érs (cf. terr-uncius) for *térs from *tris (Gk. rpis),
testis for *érz-stis (cf. Oscan tristaamentud, in Lat. testamento
Abl.) (but see ch. iii. § 15. 8). Similarly for -27-, when, through
syncope, the / has to play the part of a vowel (sonant or syllabic
2), as is seen in our ‘able,’ and more clearly in French able, we
find # (older 57) in Latin, e.g. facultas for *facl(t)tas beside
facilitas, stmultas for *siml(t)tas beside stmilitas.

After z we find @ not 7 in sdcietas, anaietas, &c., beside provitas,
castitas, &c. (ef. Engl. yé- for yi- in ‘yet,’ ‘yes’), as we find
léniéna beside tonstrina (§ 12). In Oscan I.-Eur. i is f (in Lat.
alph. ¢,in Gr. «.); in Umbr. itis i (Lat. alph. 2), but in O. Umbr.
ofteh e; e.g. Osc. pis, Umbr. pis ‘quis’ (von Planta, i. p. 96).

§ 14, Other examples of Lat. iforI.-Eur.i. The -is- of the I.-Eur. Super-
lative suffix is-to- (O. Ind. svad-iStha-, Gk. #5.070s, Goth. sut-ista, ‘ sweetest ’)
appears in Latin mdg-is-ter (in O. Lat. written magester according to Quintilian,
i. 4. 17 quid ? non e quoque i loco fuit ? ‘Menerua’ et ‘leber ’ et ‘magester’ et
‘ Diove Victore,’ non Diovi Victori), min-is-ter. The weak grade of an ei-root is
seen in L.-Eur. tri- from root trei-, ‘three’ (O. Ind. tri-8a Loc., Gk. rp-oi, Goth.
pri-m Dat., Lith. tri-sé Loc., 0. Sl. tri-chii) and Latin tri-bus, though in ter, terni,
&e. there is the usual change of -ri- to -er-; Lat. piscis (Goth. fisks), while O. Ir.
iasg from *peisco-shows the ei-grade ; Lat. video (O. Ind. vid-m4, 1 Pl., Hom. Gk.
Fté-nev, W. gwedd, ‘aspect,’ Goth. vit-um, 1 Pl., Engl. wit, Germ. wissen) from
root weid-, ‘to see, know,’ with Perfect-stem woid- (Gk. ofa) ; Lat. fid-i, findo
(O. Ind. bhid-), ‘ to split,’ Goth. bitum, 1 Pl. Pret., Engl. bit, Germ. bissen),

from root bheid- (Goth. beitan, Engl. to bite, Germ. beissen).
i in the unaccented syllable. See ch. iii. § 18 for other instances.

§ 15. ié, not ii. Other examples are pietus, séitietas, ébrietas and other deriva-
tives in -tat- from io-adjective stems, drietis, dbietis, &c., viiriego, &e. beside lévigo,
&c., hietare, and the earlier spellings conieciant, proiecitad, inieciatis, traiécere, &e.
(see ch. ili. § 18, p, 188).

O, O.

§16. 6. I.-Eur. 6 is Latin 6. Thus the I.-Eur. root pé-, ‘to
drink’ (O. Ind, pa-,pa-na-, Noun, Gk. 7é-rw-xa, du-7-T1s, the ebb,
Aeol. 7é-vw, Lith. pu-ta) is Latin pd- of po-to, pd-tus, po-culum ;
do-, ‘to give’ (O. Ind, da-na- and da-ti-, ‘a gift,’ Arm. tur, Gk. 86-
pov and burivy, Lith. di-tis, O. 81. da-rti, dan-ii, P.P.P.; in Celtic,
the I.-Eur. word for ‘gift’ has taken the sense of ‘an accom-
plishment,’ O. Ir. dan, W. dawn) is in Latin dd-num, do-s. On
§§14-17.] REPRESENTATIVES OF I.-EUR. SOUNDS. 0, 6. 233

the other hand, I.-Eur. 6 is often a grade of & or é (e.g. Gk.
motdouat from root met-, cAdy, a thief, from xdrAéatw, Owpds
from réOnr) (see §§ 51, 53). A root like énd- (Gk. yuwrds, Lat.
notus, older ¢ndtus, gudsco) beside Sen-, admits of being explained
either as an addition of the suffix -6- to the weak grade &n-, or as
a grade of a dissyllabic root geno-, so that gndtus from Seno-
would be like gzdrus from &ena-, &e. (§ 1).

Latin 6 sometimes represents an 0 lengthened by ‘compensa-
tion,’ e. g. pono from *p0-s(z)no (cf. po-situs) ; sometimes it is due
to crasis, e.g. cdpula for *co-apula, comburo for co-amb-uro (see
ch. 11. § 149); sometimes it is the ‘rustic’ development of Latin
ow, which in standard Latin became 4@, e.g. rdbus for *ribus (cf.
vubidus) from I.-Eur, reudh- (Goth. raups) (see § 41); some-
times it is the ‘rustic’ form of au, e.g. plostrum, a byform of
plaustrum, from plaudo (ch. ii. § 37). But 6 is also found to vary
with aw in Latin, when both are sprung from an original 6u,
e.g. dsculum, ausculum (26.).

In unaccented syllables 6 remained unaltered, e.g. con-déno
and co-gnosco, but final -6 became shortened in course of time (see
ch. iii. § 45). Octdvus from oct6 (I.-Eur. *oktd, *oktou) appears to
show dv for dw, as cavus shows dv for bw (§ 19). Far, cir
(O. Lat. gudér) seem to represent an I.-Eur. *bhor (Gk. gop,
p. 254), *q#d-r (Lith. kuf, ‘where,’ for *kir, ch. x. § 10), and
non to stand for *nin, a development of xoen(um) (ch. x. § 18).

I.-Eur. 6 is in Ose. u, uu (in Lat.alph. ~), but the endings -6s, -op
appear in Osc. as -ts, -Gd (with 4, the symbol of I.-Eur. 6); e.g.
djuunated ‘ddnavit, dunum ‘donum,’ Abellants ‘ Abellani’
Nom. Plur., Bévaianid ‘Boviano’ Abl. Sg. In Umbr. it is
usually o (in Lat. alph., for the native alphabet writes u for
both the U-sound and the O-sound, ch. i. § 1), but -On appears
as -ur, e.g. nome ‘nomen,’ postro ‘retro, arsferture ‘adfertori’
(see von Planta, i. p. 116). :

$17. 6. I.-Eur. 6 is Latin (accented) 3, e.g. 1-Eur. *3kto,
‘eight’ (O. Ind. a&ta, Arm. ut’, Gk. d«ré, O. Ir. ocht, W. wyth,
Goth. ahtau, O. Engl. eahta, Lith. asztini, O.S]. osmi), Lat.
bcto; 1.-Eur. *pdti-, ‘master’ (O. Ind. pati-, Gk. xdéo1s, Goth.
brip-faps, ‘bridegroom,’ O. Eng]. fadian, ‘to arrange,’ Lith. pats,
vesz-patis, ‘ lord’), Lat. potis, potior.
234 THE LATIN LANGUAGE. [Chap. IV.

.

I.-Eur. Sw seems to have become dv in the beginning of the
second century B.c., e.g. cdvus, older covus (Gk. «doi, cavities,
Hesych., xotAos for *xoFiAos). Latin vd- became vé- in the
middle of the second century B.c. in versus, older vorsus, &c.
(§ 10); but vdco was the older form of wiico, as we see from
Plautus’ pun in Cas. 527:

fic habeant linguém tuae aedes. Quid ita? Quom uenidm, uocent.

Latin -dvs may represent I.-Eur. -8w-, as in ndvem, novus (see
§ 10); Latin sd-, I-Eur. swé-, as in sdror, sdcer, somnus (see
§ 10); Latin 07, dr, the I.-Eur. sonant or syllabic 1, r, as in
fors, cor (see § 92). In the accented, as well as the unaccented
syllable, Latin 4 became wu, before 7 with a consonant (not J),
before m with a labial, before ngu, e.g. vult (volt) from volo, lum-
bus, unguis. Before certain other consonant-groups it tended to
the close o- or to the w-sound (see ch. ii. § 22). In the unaccented
syllable, Latin 0 offered more resistance than, for example, Latin
a, to the rule that a short vowel became @, then 7, in open syllables;
e. g. adnoto, arrdgo have not changed their vowel like adigo (from
digo). In syllables long by position, d became % about the end of
the third century B.c. ; and any d which had escaped weakening to
é,% took the same course (see ch. iii. §18). The terminations -os, -om
became -ws, -wm towards the end of the third century B.o. [Luciom,
Jilios, on one Scipio epitaph (C. J. L. 1. 32), Lucius, proguatus on
another (id. i. 30), -ws, -wm invariably inthe S.C. Bacch. of 186 B.c.,
(id. i. 196), and on the decree of L. Aem. Paulus Macedonicus of
189 B.c. (2. ii. 5041)]. But after v, u, gu, gu we find the spelling
-os, -om down to the end of the Republic. There are similar traces
in Osc. of unaccented 6 becoming U, e.g. dolom and dolum (von
Planta, i. p. 111). Final 6 became @, like final 4, 3, &., e.g.
séquére imper. for I.-Eur. *seqeso (Gk. émeo) (see ch viii. § 77).

§18. Latin 6 for I.-EHur. 6. Other examples: from the I.-Eur. root dqi-, ‘to
see’ (varying with oqi-) (O. Ind. aksi-, ‘the eye,’ Arm. akn, Gk. dupa, dopa,
Lith. akis, 0.81. oko) comes Lat. oculus; the L.-Eur. preposition *pré, ‘forth’
(O. Ind. pra, Gk. mpd, 0. Ir. ro, used like the Augment to indicate past time, e. g.
ro chan ‘I sang’ from canim ‘I sing,’ O. Bret. ro-, Goth. fra-, Lith. pra-, 0. SI.
pro-) is Lat. pré- of pré-ficiscor,&c. ; another preposition, I.-Eur.*xom ‘with ’ (Gk.
xowds from *xop-yos, O, Ir. com-, Osc. com) is Lat. com, which when unaccented,
or when preceding b, », gu, &c., became cum, its usual form in classical Latin ;
I.-Eur. *nokti-, ‘night’ (O. Ind. nakti-, Gk. vvg, O. Ir. in-nocht, ‘to-night,’ W.
henoeth, nos, Goth. nahts, O. Engl. neaht, niht, Lith. naktis, O. SI. noSti) is
§§ 18-20.] REPRESENTATIVES OF I.-EUR. SOUNDS 6,6. 235

Lat. nix, néctis Gen. ; I.-Eur. root od-, ‘to smell’ (varying with 6d-) Gk. d5un,
é{w, Arm. hot) appears in Lat. dd-or, oleo for *od-eo (§ 111).

§ 19. Lat. & for I.-Hur. 6, under influence of v. Other examples are Lat.
céiveo for *cdveo (Gk. xo(F)éw, to perceive, d-rovw), Lat. autiimo for *dvi-twmo from
*évi-tumo (Gk. diw), Lat. ldvo for *livo (Gk. Aovw). Lat. fdvilla for *fovilla from
I.-Eur. root dhegh?-, ‘to burn’ (0. Ind. dah-, Gk. rég-pa, ashes, Lith. degu,
‘I burn’). The example previously quoted, Lat. cdivus for covus, enables us to
assign a date to this change of I.-Eur. ow, Lat. ov to av. The Spanish and
Portuguese words (Span. cueva, Port. cova) show that covo-, not cavo-, was the
Vulgar Latin stem at the time when Spain was made a province. The
country-term coum, (cohum), the hollow in the plough, used by Ennius of the
innermost part of the heavenly sphere, retained the 0, as did ovis (Gk. 6(F):s),
though whether the avi- of aububulcus ‘ pastor ovium’ (Liwe, Prodr. p. 348),
avillus ‘agnus recentis partus’ (Paul. Fest. 10. 32 Th.) is better referred to
this root or to ag}- the root of agnus (Gk. dpvds for *dBvos) isnot clear. (Varro,
L.L. v. 135 explains coum as ‘sub jugo medio cavum, quod bura extrema
addita oppilatur,’ and adds ‘vocatur coum a covo’; cf. Paul. Fest. 28. 1 Th. ;
Isid. Nat. Rer. 12 cous (v.1. chous) est quo caelum continetur, unde Ennius,

vix solum complere coum (MSS. choum, cous) terroribus caeli.

Partes ejus sunt, cous (v. J, chous), axis, clima, cardines, convexa, poli, hemi-
sphaeria ; Diomedes (365. 17 K.) says that Verrius Flaccus spelt incoho, not
inchoo, for he derived the word from cohwm, the Old Latin word for mundus).
We have already found that I.-Eur. éw became 6v in Latin (e. g. Lat. novus for
I.-Eur. *newos), and that probably at a very early date, seeing that the change
is shared by other Italic languages (e. g. Osc. Navellum). If then it be the case
that I.-Eur. ow became av in Latin in the third or second cent. B.c. we must
suppose that Latin ov from I.-Eur. ew had a different sound from Latin ov from
1.-Eur. ow ; for the former ov does not undergo change to aw (e. g. novus, not
*navus; novem, not *navem). Latin ov from I.-Eur. ogh® shares the change to
av, ©. g. favilla from I.-Eur. dhégh?-, the o-grade of the root dhégh?-, ‘to burn’ ;
but nddus for *nov(e)dus from I.-Eur. *nog}-, ‘naked,’ shows that the change
of ov to av was later than the syncope of é in the post-tonic syllable.

Lat. vico (Plaut. Cas. 527) for vdico (Umbr. vageto- P. P. P., vakaze, for *vakaz
se, Lat. vacatio sit?), is probably nothing but an indication of the o-sound
assumed by a when preceded by v (see ch. ii. § 4). This o-sound in *vicitus,
the Vulg. Lat. word for ‘empty,’ must have persisted till late times, for
Italian voto, as well as O. Fr. voit, reflect this form.

§ 20. wi for 6. (1) in close syllables, unaccented (according to the early
Accent-law) : vétistus for *vetis-to- (Gk. (F)éros), and other derivatives from
Neuters in -os (class. Lat. -ws), such as angustus, vénustus, dnustus ; homillus for
*homillus from *homin-lo-, and other lo- Diminutives from én-stems, like
lenullus, whereas Diminutives from on- have -dil-, e.g. persdlla, céralla ; so also
Diminutives in -co-lo- from on-stems, like himiinculus, latrunculus ; dhimnus (cf.
Gk. rpep-dpevos) and similar formations, Vertumnus, Autumnus, cilumna, &e.

(2) before 1 with consonant (not il) : ciilmen (contrasted with coliimen) ; stultus
(contrasted with stolidus) ; pulvis (contrasted with pollen) ; fulvus (contrasted.
with solvo for *solud). In Old Latin we have o, e.g. on inscriptions, Folvius
(C. 1. L. vi. 1307, of 187 B.c. ; Eph. Epigr. viii. 476, ¢. 135 B. 0.5 CLL. i. 554 and
236 THE LATIN LANGUAGE. [Chap. Iv.

555, both of 1g0-r29 B. c. &e.) Polc[er'] (ib. i 552 of 132-131 B.c.). Priscian
G. p. 27. 33 H.) tells us that colya was the O. Lat. form of culpa; and in the
lines of Ennius about Servius Tullius (Amn. 337 M.) the corrupt reading of
the MSS. optimus for ultimus, probably indicates the spelling oltimus (cf.
Ose. Gltiumam) :
mortalem summum Fortuna repente
reddidit, ut summo regno famul oltimus esset.

On the spellings volva and vulva, Fem. of adj. *volvus from volvo, see Georges,
Lex. Wortf. s.v. Vulnusis the pronunciation of Varro (LZ. L. iii. fr., p. 148 Wilm.
vafer, velum, vinum, vomis, vulnus, where he gives examples of initial »
followed by the various vowels of the alphabet). This ol may be I.-Eur. él,
e. g. volt, vult (§ ro).

(3) before m with labial : wmbo for *ombo, like umbilicus for *omb- (Gk. déupaads);
the Greek (Thracian ?) Joyzdaia is rumpia in Ennius (A. xiv. fr. 8 M.) and Livy
(xxxi. 39. 11). Perhaps also before ms, e. g. dimérus from omso- (O. Ind. dsa-,
Arm. us, Goth. ams, Umbr. onso-).

(4) before ngu: unguo, unguen, unguentum for *ongu- (O. Ind. afij-, ‘to anoint’) ;
ungula, like unguis (Gk. dvug) ; but longus, where the g is not velar (gu), retains
the o (we find however lun[gum] beside lon[gum] in neighbouring inserr.,
C.I. L. i 1073). We have also uncus for *oncus (Gk. éyxos), (but sesconciam
C. I. L. i. 1430, in a Cremona inscription), &. On the occasional use of u for 0,
often for Greek o, before other consonant-groups, e. g. turnus beside tornus,
from Greek répvos, O. Lat. frundes (see ch. ii. § 22). Some isolated cases of
u for accented 6 have various explanations ; hiimus, for *homus (cf. Gk. xOwv),
may take its u from the analogy of himeo, for the word (not common in the
oldest writers ; humi, for example, not occurring till Terence, Andr. 726) seems
to have been first used in the sense of moist ground, clay, e. g. Laevius ap.
Prise. i. p. 269. 7 H. humum humidum pedibus fodit; Varr. Men. 531 B. in
pavimento non audes facere laconam+, at in humu calceos facis elixos; Enn.
Trag. 396 R. cubitis pinsibant humum ; Pac. Trag. 351 R. tractate per aspera
saxa et humum ; ef. Gracch. Trag. 3 mersit sequentis humidum plantas (MS.
plantis) humum ; Priscian’s ‘old Latin huminem’ (i. p. 27. 1 H.) may be an
etymological spelling to suit a derivation from humus, like the spelling colina,
adapted to the derivation from colo (Varro ap. Non. 55. 20M. ; ‘Serv.’ ad Aen.
iii. 134) ; fillica, if for *folica, either follows the analogy of filigo, or shows the
vowel of fulca, the form used by Furius Antias ap. Gell. xviii. 11. 1. 4.

(5) in syllables unaccented under the later Accent-law : the 3 Plur. suffix
-ont became -unt at the end of the third cent. B.c., though the old spelling
was sometimes retained even later. On old inscriptions we have e. g. dederont
(0. I. L, i, 181, from Picenum), cosentiont on a Scipio epitaph (i. 32) (so on the
restored Columna Rostrata, exfociont, i. 195). Festus (244. 13 Th.) quotes
praed-opiont (MS. praedotiont), in the sense of pracoptant, from the Carmen Saliare,
and nequinont, an old 3 Pl. form of nequeo, from the Odyssea of Livius Andronicus
(ap. Fest. 162. 24 Th. ; ef. Paul. Fest. 169. 14 Th.) :

partim érrant, nequinont Graéciam redire ;

the Nom. 8g. termination of Neuter ES-stems is -os on an old inscription, opos
(¢. I. L. i. 52, probably from Orvieto) (cf. Uenos on old mirrors, i. 57 and 58) ;
the Acc. Sg. termination of O-stems is -om in the older period, e. g. donom, the
Nom. 8g. is -os (see Index to C. I. L. i.), and this spelling remained after 4, v,
§§ 21-23.] REPRESENTATIVES OF I.-EUR. SOUNDS. 0,0. 237

e. g. equos, arvom, till the time of Quintilian, though the pronunciation may
have been the ti-sound (see § 70). (On this reduction of 6 to % in unaccented
syllables, see ch. iii. §§ 18, 26; another example is the verb swm, for *som, with wu
for 0, because of its usual unaccented character.) In late Latin, when % and
d had come to have nearly, or altogether, the same sound, o is often written for
u, so that the older spelling seems to be revived (see ch. ii. § 29).

U, U.

§ 21. U. I.-Eur. i is Latin #, L-Eur. *dhimo-,‘ smoke,’ from
root dheu-, ‘to move violently’ (O. Ind. dhamé-, Gk. diyds,
passion, Lith. damai Pl, O. Sl. dymit), Lat. famus; I.-Eur.
*miis-, ‘a mouse’ (O. Ind. mus-, Gk. pis, O. Engl. mis, O. SI.
mysi), Lat. mis. It is generally a grade of a eu-root as i of an ei-
root (§ 12). Latin #, older ov, may represent also I.-Eur. eu or ou,
e.g. dico, older douco (see §§ 35, 41), Latin %, older 07, oe, I.-Eur.
oi, e.g. cura (§ 38), and sometimes has arisen from % by ‘ compensa-
tion, e.g. dimus, older diésmo- (Paul. Fest. 47. 20 Th.), a spelling
retained in the proper name Dusmius; diimetum for dusm-, in Virgil
MSS. spelt dummetum, shows that -dm- is equivalent to -dmm-.
In the unaccented syllable # may represent au, e.g. défrido
from fraudo, inclido from claudo (ch. iti. § 18). On far, cir for
*for, quor, see § 16. J.-Eur. Gi is in Umbr. and perhaps in
some other dialects i, e.g. Umbr. /r7f ‘fruges’ Acc. Pl. (see
von Planta, i. 129).

§ 22. Other examples of Lat. a, I.-Hur. a. Lat. frinisci (Goth. brikjan,
“to use,’ Germ. brauchen, O. Engl. briican, Engl. to brook) ; jus, broth (0. Ind.
yiia-, Gk. (i-un for * Cuo-pyn, leaven, Lith. juszé) ; situs (O. Ind. syitd-, Gk.
veo-Kat Tiros) ; so-lutus (Hom. BovAiréy-de).

§ 23. U. I.-Eur. ti, Lat #, appears often in the weak grade of
an eu-root, e.g. I.-Eur. *ytigo-, ‘a yoke,’ weak grade of yeug-, ‘ to
join’ (O. Ind. yiigaé-, Gk. Gydv, Goth. jtik, O. Sl. igo for jtigo),
Lat. jaigum ; I.-Eur. lik-, weak grade of leuk-,‘to shine’ (O. Ind.
riic-, Gk. duqu-Aven, twilight), Lat. lécerna. I.-Eur. tt (Lat. 7)
is also the weak grade of a wé-root, e.g. I.-Eur. *péruti, ‘last
year” (Gk. wépvor), from *wetes-, ‘year, and similarly in the
unaccented syllable Latin % often appears for wé, e.g. coneutio for
*conquetio from quiitio (see ch. iii. § 25). Lat. « often represents
I.-Eur. (and older Latin) 6; for an ¢ passed at the end of the
third cent. B.c, into the sound #, when in the unaccented syllable
(unless saved by a preceding », ~), e.g. donwm from earlier donom,
2.38 THE LATIN LANGUAGE. [Chap. IV.

but eguom till the time of Quintilian ; and even in the accented
syllable 6 came to assume a w-sound before certain consonant-
groups, e.g. exdpa, older colpa, Fulvius, older Folvius (see § 20).
(On equom, divom, &c., see §§ 70, 135).

Lat. % offered more resistance than @ to the usual transition
of a short vowel in the open unaccented syllable to 7 (earlier ¢,
and always before r), e.g. sdcer (Gk. éxvpds) (ch. iii. § 18). Before
Z and labials it passed in open unaccented syllables (especially
when the next syllable contained an 7 in hiatus) into the
z-sound, which ultimately was written and pronounced 7, e.g.
minubiae, manihiae, manibus, dissupo, dissipo. That it ever had
the d-sound, the sound of Greek v, in the accented syllable of
native Latin words is doubtful. [On (quid)lubet and (quid)-
libet, &c. see ch. ii. § 16]. Before a vowel in the unaccented
syllable Latin « may represent I.-Hur. ew, ow, Lat. ov, e.g.
dénuo for dé novo, éluo for é-lavo (I.-Eur. *]5wo) (ch. iii. § 24);
before Zand labials any short vowel, e. g. oceupo from cap-, to take
(ch. ui, § 18), and before any consonant-group Latin 6, e.g.
himullus for *homén-lo- (see § 20). Final -i#, like other short
vowels, normally became é, and might be elided (ch. iii. §§ 37, 38).

A close relation exists in Latin, as in I.-Eur., between ii and
w, vocalic and consonantal u. After / and r the vowel # became
a consonant in the second cent. B.c. in Latin, e.g. dérva (Jarua,
Plaut.), arvum (druos, -a, -om Plaut.), milvus (miluos, Plaut.),
pelvis, &e. (see ch. iti. § 48). For Latin & (earlier 3), the parasitic,
or svarabhaktic vowel, in ocudus from éc-lo-, speculum from spéc-
lo-, draculum from ord-clo-, see ch. i. § 154.

I.-Eur, ti is Umbro-Ose. u, e.g. Umbr. swéra ‘supra,’ though
sometimes we find o written in Umbr. [i.e. in the Lat. alph., for
the native alphabet did not distinguish the O- and the U-sound'
(see ch. i.§1)]. In Oscan we find iu after t, d, n (s?),e.g. tiurri
‘turrim,’ Diumpais ‘Lumpis’ (‘to the Nymphs’), Niumsiefs
‘Numerii’ Gen. Sg., an affection of u which resembles Boeot.
riovxa (Att. réxn), or Engl. ‘pure,’ &c. (pronounced pju-), (see
von Planta, i. p. 122).

§ 24. Lat. i for I.-Eur. t. I[.-Eur, -t- in the U-stem suffix is Latin % of
fructis, manis, &e. ; the I.-Eur, preposition *ipé (O. Ind. apa, Gk. bao, O. Ir.
fo for *wo, with p dropped between vowels, and u turned into w, Goth. uf,
§§ 24-26.) REPRESENTATIVES OF I.-EUR. DIPHTHONGS. 239

Engl. of-ten) is Latin s-ub (see ch. ix. § 52); the preposition *ipér, *upéri
(O. Ind. updri, Gk. breip, trep, O. Ir. for, from *wer, with suppression of p,
Goth, ufar, Engl. over, o’er) is Latin s-uper (see ch. ix. § 53); I.-Eur, *sniiso-,
‘daughter-in-law’ (O. Ind. snuSd-, Arm. nu, Gr. vv(c)és, O. Engl. snoru,
Germ. Schuur, 0. Sl. sniicha), is Lat. nitrus; I.-Eur. *klito-, ‘famous,’ from Len
‘to hear’ (O. Ind. Sruté-, Gr. «durés, O. Ir. cloth, Gaul. Cluto-ida, O. H. Germ.
Hlud-olf, Hlot-hari, from hari, ‘army,’ whence Lothair, &c.), is Lat. in-chitus,
later in-clitus; I.-Eur. *ridhro-, ‘red,’ from the root reudh-, ‘to be red’ (O. Ind.
rudhira-, Gk. épv6pds, O. Sl. rudrit) is Lat. riber ; Lat. miisca (the Romance forms
prove %), Gk. puta for *nvoya, Lith. musé ; Lat. jivenis (O. Ind. yavan-), jiivencus
from I.-Eur. *ytiwnko- (O. Ind. yuvasa-, Gk. “Lax-wO0s, O. Ir. dac, Welsh ieuanc,
Gaul. Jovincillus, Goth. juggs for *juvunga-, O. Engl. geong) ; Lat. fuga (Gk.
vy) from root bheug-, ‘to flee’ (Gk. gevyw) ; Lat. wrgeo from urg-, the weak
grade of the root werg-, ‘to confine, press’ (Gk. é-(F)épyw, Lat. vergo.)

§ 25, Latin ti and Latin 6. We have seen that Latin 6 became % when

unaccented, and even in the accented syllable before certain consonant-groups,
and that in late Latin % and 6 came to have the same sound and were often
exchanged in spelling (see § 17 and ch. ii. § 29). Some doubtful cases may
be discussed here. As turbo and turba (Gk. répBn, orupBatw) come from I.-Eur.
turb-, from ttr-, the weak grade of twér-, ‘to twirl,’ so I.-Eur. q#r-, gh¥r-, the
weak grades of q¥ér-, gh3ér-, seem to appear in Latin as cur-, fur, e.g. furnus,
anoven, from I.-Eur. gh?rno- (0. SI. griint, ‘a kettle’), but formus, warm, from
I.-Eur. ghYormo- (0.H.G. warm, Engl. warm). Whether Nonius, who
derives furnus from formus, has any justification in spelling the word fornus
(531. 24 M.) is uncertain ; but fornax has 0, and Plautus Epid. 119 puns on
furno (so the MSS.), and foro. Curro may then stand for *q3rso- from the root
quer- (cf. Lat. querquérus) ; curtus cannot be the same as Gk. «aprés, from the
root ker- of Gk. «efpw, but must stand for *q#rto-, from a root qer- ; corpus
must represent *q¥orpes-, not *q¥rpes- (cf. O. Ind. kfp-); ursus may be
a loanword (Lucanian, according to Varro, L. L. v. 100), or may owe its u to
a velar guttural -rq3-; dirceus cannot be connected with orca; the spelling
forcillis in MSS. of Catullus (ev. 2) must be late, if the word is connected with
Hesychius’ govpkop, pupxos, which point to u (Cyprian popres = xapaxes shows
dialectal o for v) ; urbs (cf. Mars. en urbid ‘in urbe’) cannot be connected with
orbs, so that there is no etymological appropriateness in the play on these
words which is often found in the later poets (e. g. Rutil. i. 66).

§ 26. The Diphthongs. In its treatment of the diphthongs
ai, au, e, [(1) I.-Eur. ei and in Latin also, (2) I.-Eur. ai (oi?) in
the post-tonic syllable, (3) I.-Eur. ai, oi in the final syllable],
oi, ou, [(1) 1-Eur. eu, (2) I.-Eur. ou], Latin stands halfway
between Oscan an Umbrian. In Oscan they are all retained
intact, except that the i-element has sunk to an e-sound’. In
Umbrian they are all reduced to simple sounds, e.g. Umbr.

 

1 Perhaps both in Oscan and Latin a sound between open iand e, just
the second element of ai, &c. was as in German and English.
240 THE LATIN LANGUAGE. (Chap. IV.

dévo-, Osc. defvo-, ‘god’; Umbrian toro-, Ose. ravpo-, bull;
Umbr. toto, Osc. twFro, people, community [I.-Eur. *teuta-,
Goth. piuda, whence some derive ‘Teuton, O. Ir. tuath, Gaul.
Teuto-bédiaci, W. tud (in Gaelic tuath is used for ‘the country-
people,’ ‘the tenantry’), Lith. tauta], (see von Planta, i. p. 137).
In the Latin of Cicero’s time ae (from az) and au are the only
survivors ; and even they tend in rustic or colloquial speech to
single sounds, é, 6, e.g. pretor, plostrum. Diphthongs whose two
elements had affinity of sound, such as ei, ou, are naturally the
first to be simplified ; ei, for example, both in Greek and in the
Teutonic languages, passed early into a long @-sound. The oldest
Latin inscriptions offer with great fidelity e¢ for I.-Eur. ei, as also
for the ei to which J.-Eur, ai, oi, when unaccented, were reduced.
Thus on the 8. C. de Bacchanalibus of 186 B.c. (C. 2. L. 1, 196)
we have deicerent (I.-Eur. ei), inceideretis (I.-Eur. post-tonic ai),
foideratei (I.-Eur. final -oi). But this diphthong soon became
identical in pronunciation with the long é-vowel, so that spelling-
reformers like Accius and Lucilius used ¢ and ¢ for this long
s-sound without sufficient regard to the past history of the
sound (see ch. i. § g), and the practice grew up of using ¢ to
indicate the long vowel-sound, 2 to indicate the short. Ow is
also sometimes used for Latin @ (I.-Eur. oi), e.g. couraverunt
(C.L.L.1i. 1419, from Picenum), though not at all to the same
extent as ¢ for 7. Ow seems to have been reduced to a simple
sound at the end of the third century B.c. 4i became ae a little
later. Towards the end of the second cent. B.c. we find the
spelling ae established in use, with an occasional resort at the
transition period to a spelling aez; but the original spelling (not
pronunciation) was again brought into fashion in the reign of
Claudius and is found occasionally on epitaphs even of the late
Empire. On the diphthongal sound of ae (Germ. Kaiser is
evidence of this sound in Lat. Caesar), au, see ch. ii. § 32.

Oi passed (through oe) into % at the beginning of the second
cent. B.c., though 07, and afterwards oe, were long retained
on official inscriptions in phrases like faciundum coiraverunt
(coeraverunt), eg. C.I.L. i. 567 (Capua) of 106 B.c.: murum
et pluteum faciund. coeravere, where the spelling murum
(older moerum, moirom) shows that the pronunciation was @,
§ 27.] REPRESENTATIVES OF ‘I.-EUR. SOUNDS. AI. 241

not oe, and in some words of the official or legal style like
poena, foedus.

The long diphthongs are not common in I.-Eur., so that we
have hardly sufficient material from which to discover their
history in Latin. In the Veda, the oldest literature of India,
we see a tendency to use final -d (I.-Eur. -6) before a consonant
initial, final -du (I.-Eur. -u) before a vowel initial (e.g. déva
and dévau in Dual of dev4-, I.-Eur. *deivo-,‘ god’); and the
development of éi and 6u in roots and suffixes in the various
I.-Eur. languages suggests that doublets of this sort, 6i and 6,
ou and 6, already existed in the case of these two diphthongs of
kindred elements in what is called ‘the Indo-European period.’
Long diphthongs, composed of sounds not so nearly allied as
é and i, 6 and u, may have taken on Latin soil a different course
of development, according as they were final or not. When
final, the second element may have been suppressed (probably
after passing through the doublet-stage), just as in later times
a short final vowel has been suppressed after a long syllable in
words like ewemplar(e), név(e). When followed by a consonant
the long element must have been shortened, by the rule that any
long vowel is shortened before y, w, n,m, 1, r, &c. followed by
a consonant, so that ai (ay) would pass to the ordinary diphthong-
sound ai, au (aw) to au, éu to eu, 61 to 02; just as we have véntus
from *wént- (root wé-, ‘ to blow,’ Gk. ant) (see below, § 45).

§ 27. AI. I.-Eur, ai is Lat. ae (older ai), e.g. the L.-Eur. root
aiwé-, ‘ time, life’ (Gk. aidv, dei for *alFe(o)., with Ion. Att. a for
aif-, as in édaa for *édaiFa, O. Ir. ais, aes, Goth. aivs, O. Engl. &),
appears in Latin aevwm ; I.-Eur, aidh-,“to burn,’ (O. Ind. édhas-,
‘firewood,’ Gk. aiéw, O. Ir. aid, aed, ‘fire,’ O. H. G. eit, ‘pyre,’ O.
Engl. 4d), in Latin aedes, lit. ‘where the fire is kept up,’ aestus.
In the last root I.-Eur. ai is the weak grade of ayé, thus aidh-
of ayédh- (aye- is seen in I.-Eur. *ayes-, ‘metal, gleaming metal’),
and in the first of ai (§ 45) (cf. O. Ind. ayu-). In the
unaccented syllable ae became 7, e.g. inguiro, occtdo from quaero,
caedo (see ch. iii. § 18), or rather a2 became ¢?, then 7, e. g. inceideretis
on the S. C. de Bacch.; and in the final syllable, e.g. ¢wtude
(O. Ind. tutudé) with the 1 Sg. Perf. Middle ending -ai or -ai

R
242 THE LATIN LANGUAGE. {Chap. IV.

(ch, viii. § 76). On rustic and colloquial ¢ for ae, and on the late
Latin confusion of ae, e, o¢ (that is, oe which had been restored,
§ 38), see ch. ii. §§ 41, 44. A spurious diphthong ai is found in
the verb aio (for *ahio, § 116), ain (for *azsne), aibat (dissyllabic), &c.

§ 28, I.-Hur. ai, Lat. ae (ai). Lat. caedo from I.-Eur. root skaidh- (Goth.
skaidan, O. Engl. scddan, Engl. water-shed, Lith. skédziu, ‘I separate’); Lat.
scaevus (Gk. oxar(F)és) ; Lat. laevus (Gk. Aa(F)éds, O. Sl. lévi) ; Lat. caecus from
I.-Eur. *kaiko-, ‘blind’ (0. Ir. caech, with another sense, ‘empty,’ cf. Gael.
caoch-ag, ‘a nut without a kernel,’ Goth. haihs; in Gk. xomvaAdw, ‘to gape about,’
the a: has become o:, through the assimilating influence of the accented v) ;
Lat. haereo from I.-Eur. root ghais- (Lith. gaiszti, ‘to tarry’; Goth. us-gaisjan,
‘to frighten,’ lit. ‘cause to hesitate,’ Engl. gaze); Lat. haedus (Goth. gaits, Engl.
goat) ; Lat. lévir from I.-Eur. *daiwer-, O. Ind. dévadr-, Arm. taigr, Gk. dan,
0. H. G. zeihhur, O. Eng. tacor, Lith. déveris, 0. SI]. déveri) is a late Latin
spelling for laevir, the i being due to the analogy of vir.

§ 29. AI, AE on Inscriptions. We have ac on theS. C. Bacch. (¢. I. L. i. 196)
of 186 8. o. in aedem (along with aiquom, tabelai, data, &e.) ; Aemilius on three
inscriptions of 187 B. co. (i. 535-7), but Admilius (C. I. L. ii. 5041) of 189 B. ©. ;
aetate on a Scipio epitaph of c. 130 B.c. (i. 34), (along with guatratis) ; quaestor,
quaero, praetor are established spellings in the Lex Bantina of 133-118 B.c. (i.
197), the Lex Repetundarum of 123-122 B. ¢. (i. 198), &. The spelling aei,
which we find once or twice towards the end of the second cent. B. ¢., conquaeisivei
(in the post-tonie syllable under the older Accent Law) (i. 551, of 132 B. ¢.),
Caeicilius (i. 547 b., ‘of 141 or 116 B.c.,” Momms., and i. 1487, from Majorea),
Caeician[us] [i. 378, on a coin with an alphabet A-X, so older than the intro-
duction of Y, Z (ch. i. § 2)], Caeidia (ix. 3087, from Sulmo), seems to mark the
transition stage. Often ai and ae are found side by side, e.g. praitores acre
Martio emeru (i. 1148, from Cora) ; aetated and saip[is}wme on the dedicatory
inscription of the Faliscan ‘collegium cocorum’ (Zvet. I. I. I. 72a). The
spelling Caisar, &c. is frequent on inscriptions of Claudius’ reign, when anti-
quarian lore was in fashion (e.g. C. I. L. vi. 353), and we find on epitaphs of
the late Empire Valeriai (Rossi, i. 113, of 352 a.D.), quai and filiai (Rossi, i. 410,
of 3934.D.), &c. His not regularly exchanged with ae till the fourth cent. a. p.
on inscriptions (Seelmann, Ausspr. Lat. p. 225), but in dialectal inserr. it is of
course much earlier, e.g. cedre for caedere on an old inser. of the Umbrian
territory (C. I. L. xi. 4766), and in plebeian from the first cent. a. p. (Hammer,
Loc. Verbr. p. 11). The use of ai, ae for @ is a feature of inscriptions of the
Etruscan country (see Mem. Ist. Lombard. 1892), e. g. Puinsscos on a Praenestine
mirror (0. I. L. xiv. 4098) [cf. Saeturni on a Praenestine vase (i. 48), and perhaps
Lat. Aesculapius for ’AcxAdmés]; we have et in queistores (i. 183, Marsic).

§ 380. AU. L-EHur. au is Latin au, which in the unaccented
syllable became @ (see ch. iii. § 18), and in the accented syllable
was in dialectal Latin 6 (e.g. Plautus’ Umbrian name was Plotus
‘ splay-foot,’ Paul. Fest. 305.7 Th.). Thus the I.-Eur. root aug-,
indicating ‘ growth’ or ‘ strength,’ a weak grade of the root awég-
of Gk. a(F)éfw (O. Ind, djas-, Goth. aukan, ‘to multiply,’ Engl.
§§ 28-32.] REPRESENTATIVES OF I.-EUR. SOUNDS. AU, EI. 243

eke vb., Lith. dugu, ‘I grow, Gk. aig, abédvw) appears in Latin
aug-eo, augus-tus ; the I.-Kur. particle *au [Gk. ad, at-re, ad-r1s,
Goth. au-k, ‘also’ (with -k like Gr. ye), O. Engl. éac, Eng. eke
advb.], in Latin aw-t (Osc. avti, Umbr. ote), au-tem (Ose, avt).
This I.-Eur. *au is perhaps similarly the weak grade of *a-we,
(see ch. x. § 4). Lat. aw represents the weak grade of an I.-Eur.
du-root in ausctlum (Plaut.) beside ds from I.-Eur. *éus (cf.
caitus beside cds, § 54).

§ 31. Other examples. Lat. paucus (Goth. favai Pl. Engl. few); Lat.
aurora for *ausisa (Gk. aiprov for *adopiov, dyx-avpos vg Apoll. Rhod., ‘nearing
the dawn,’ Lith. ausz-ra, O. Scand. austr, O.H.G. dstar, Engl. east) from
I.-Eur. aus-, a weak grade of awés-, ‘to gleam’ (cf. Gk. jus for *éfws), whence
*auso-, ‘gold’ (Lat. aurwm, Lith. duksas); Lat. awris for *ausis, aus-culto (O. Ir.
au, a Neuter S-stem, Goth. ausd, an N-stem, Lith. ausis) from I.-Eur. *aus-,
a weak grade of *awés- [cf. Gk. alw (dw), I perceive ; but ots isa Greek develop-
ment of the high grade *dus; ef. O. Ind. avis, “epenly" ]. Lat. au-, away, as
in au-fero, au-fugio (Pruss. au- of au-mi-sna-n Acc., ‘washing off’; O. Sl. u- of
u-myti, ‘to wash off’) from I.-Eur. *au, a weak grade of *awe (*awo ?) (O. Ind.
ava, ‘away,’ pane ‘aufero’). It is often difficult to distinguish this Lat.
au, representing I.-Eur. au (a reduction of I.-Eur. awé) from Lat. au, a reduc-
tion of Lat. dvi, dvi, e. g. audeo from avidus. (Other examples in ch. iii. § 16. 9.)
On the occasional appearance of % for au in the accented syllable, e. g. sed frude
‘sine fraude ’ in the Lex Repetundarum, post-class. cludo, and on the plebeian
and dialectal reduction of au to 6, e.g. plostrum, Clodius, see ch. ii. §§ 36-37.

§ 32. EI. I.-Eur. ei was in Old Latin e¢; but this diphthong
became identical with the sound of long ¢, so that in inscriptions
from the latter part of the second century B.c. the symbol e is
used not only for J.-Eur. ei, but also for I.-Eur. i, and some
spelling reformers proposed to reserve the letter I for short 2, and
the diphthong-symbol EI for long 7 (see ch. i. § 9). In Oscan,
however, the diphthong is preserved. In Latin it is used in
the interjection Ae, ¢, from which comes the verb ¢ilo.
Instances of I.-Eur. ei, Latin ¢ or 2, are: I.-Eur. deik-, ‘to show,
say’ (Gk. deck-viys, Goth. ga-teiha, ‘I declare’), Lat. dico, O.
Lat. deico, Osc. deicwm Inf.; I.-Eur. bheidh-, ‘to believe, trust’
(Gr. meiOopar), Lat. fido, O. Lat. feido, Before a vowel ei, 4
became ¢ in Latin, e.g. co from I.-Eur, ei-, ‘to go,’ both in the
accented (§ 63), and in the unaccented syllable (ch. in. § 24).
O. Lat. e2, class. 2, may represent other I-diphthongs in the final or
unaccented syllables, e. g. in dutidi Perf. (O. Ind. tutudé), I.-Eur.
-ai or -ai (ch. viii. § 76); in pdpiilt Nom. Pl. (oldest Lat. poploe),
244 THE LATIN LANGUAGE, [Chap. IV.

J.-Eur. -oi; in con-cido from caedo, older caido, I -Eur. ai. The older
spelling of all these words shows -ei, tutudet, pop(u)lei, concerdo,
sometimes e, e.g. ploirume ‘ plurimi,’ Nom. PI., on a Scipio epitaph
(C.L.L. i. 32). Greek e before a vowel was written é, e.g.
Alexandréa, Daréus (shortened to ¢, e.g. balnéum), later 7, e.g.
Alexandria (shortened to 7) (cf. ch. il. § 143).

§ 33. Other examples of I.-Bur. ei. I.-Hur. *ei-ti, 3 Sg. Pres. Ind. of
ei-, ‘to go’ (O. Ind. éti, Gk. efor, Lith. citi, eft), Lat. é (with 7 in Old
Latin, but class. %,a shortening produced in the course of the second
cent. B.c. by the influence of the final -t, ch. iii. § 49); I.-Eur. *deiwo-, ‘god’
(O. Ind. dévé-, Lith. dévas, Ir. dia, Gaul. Acovova, W. dwy-fol, ‘divine,’
O. Engl. Tiwes-deg, ‘Tuesday’; but Gk. dios is for *5Fros, like O. Ind. divyé-),
Lat. divus, on the Dvenos inscription deiuo-. When 6 of the final syllable
was weakened to %, deivos became *deius (for *deivus, the v being absorbed by
the following u, § 70), deivom became *deium, which passed into de(y)us,
de(y)um as *ei-um, Ace. M. of is, into e(y)-um, *ei-o 1 Sg. into e(y)o, so that the
word would be declined deus, deivei, deivd, deum, &e.; from this variation arose
the ‘doublets’ divus, divi, divo, divum, &e., and deus, dei, deo, deum, &e. (cf.
C. I. L. i. 632 sei deo sei deivae); some grammarians of Varro’s time proposed
to restrict divus to the sense of a mortal made a god (so later divus Augustus),
but Varro contested the accuracy of this usage, showing that in old times
divus was the word for any god (Varro, L. L. iv. fr., p. 150 Wilm. ; cf. Serv. ad
Aen, xii. 139 diva deam, &c.). Reus (cf. rivalis, older reiv-), seu (cf. sivé, older
seive) are to be similarly explained, though some prefer to suppose that the
prior stages of all three words were *déus, *réus, *stu (like Alexandrta, Daréus,
&c.), and quote lévis as an example of the passage of ei before v into e (ef.
Gk. Aci(F)os) (von Planta, Osk.-Umbr. Dial. i. p. 145). Neu, ceu are most naturally
explained as shortenings from *néu (cf. néve), *ceu (cf. Gk. «7, Lat. cé-teri, B. B.
xv. 313), though they also admit of being referred to *nei-u (from O. Lat. nei,
ni, used in the sense of class. nz), and from a Locative, either Demonstr. *cei
(ch. vii. § 15) or Relative quei, qui (ch. vii. § 23). On meio see ch. viii. § 6,
ejus ch. vii. § 13, peior below § 116. :

§ 34. EI and I in Inscriptions, &c. On the S. C. de Bacchanalibus
of 186 B.c. (C.J. L. i. 196) we have deicerent (I.-Eur. ei), inceideretis (I.-Eur.
post-tonic ai), foideratei (I.-Eur. final -oi). But we have ei employed merely
to indicate long ¢ in audeire of the Lex Repetundarum of 123-122 B.o,
(i. 198); ameicitiam of the Lex Agraria of 111 B.c. (i. 200); ameicorum,
véneire (i. 203, of 78 B.C.) ; erceiscunda deividunda and feient of the Lex Rubria
of 49 B.c. (i. 205) ; esureis on a leaden bullet used at the siege of Perusia
with the cruel message carved on it, esureis et me celas (i. 692); veivos
(i. 1256), &c. ; and this seems to be the function of ¢ in the Plautine text
represented by the Codex Ambrosianus (see Index to Studemund’s Apograph,
p. 504). Even as early as the end of the third cent. B. co. we have opeinod
deuincam ted on a Praenestine mirror with a representation of a gaming-table
(Rendic. Accad. Lincei, v. p. 253, 1889). The transition stage from ei to é is
perhaps marked by the spelling ¢ in ploirwme (Nom. Pl.) on a Scipio epitaph
of the end of the third cent. B.c. (4.2. L. i. 32), conpromesise on the 8. C.
§§ 83-85.] REPRESENTATIVES OF I.-EUR. SOUNDS. EU. 245

Bacch. of 186 B.c. (i. 196; cf. ameiserunt, i. 204), though this spelling is often
nothing but a dialectal variety, e. g. uccos (Lat. vei-, vicus) on an inscription
from the Marsic territory (i. 183) (ef. Umbr. devo-, ‘god,’ Lat. divo-). Now and
then we find ei written for a short vowel, as in inpeirator on the inscription of
Aem. Paulus Macedonicus, from Spain (C. I. L. ii. 5041, of 189 B.c.), leiteras
(see ch. ii. § 130), and seine on the Lex Repetundarum (i. 198); so in the
Ambrosian Palimpsest of Plautus, ibeis, Cas. 92; curabeis, Merc. 526. And ei
appears occasionally even for the é-sound, as in pileib-, in an old inscription
found between Rome and Ostia (Eph. Epigr. i. 3), in leigibus on «a Praenestine
cippus of erratic orthography (pro sed sueq for pro se suisque, &c.) (C0. L.
xiv. 2892), and in decreivit on the Spanish inscription of 189 B.c., just
mentioned (¢. I. L. ii. 5041) (see ch. ii. § 11). The Dative forms in -e on old
inscriptions (e. g. 0. I. L. i. 1110 Iunone Seispitei Matri) are best explained
as graphical varieties of the Dat. in -ei, class. -i (see ch. vi. § 28), and similarly
the 3 Sg. Perf. Act. forms in -ed, -et, e.g. fefaced on the Praenestine fibula, dedet
(beside cepit) on a Scipio epitaph (C. I. L. i. 32), as graphical varieties of -eit (e.g.
probaveit beside coeravit, i. 600) (see ch. viii. § 70). The Plautine spelling must
have been mendico-, eira to judge from the remarks in Rud. 1305 that mendicus has -
‘one letter more’ than medicus, and in Truc. 262 that comprime sis eiram becomes
comprime sis eram by ‘taking away a single letter.’ (Cf. Early Greek E for EI).

§ 35. EU. This diphthong has been merged in ow in most lan-
guages ; but Greek, with ev and ov, and Gothic, with iu and au,
will serve as criteria. We find ew in Latin in the interjections
heu (cf. Gk. 6), ehew, hews; the pronoun neuter [a trisyllable
(ch. ii. § 32), from né and wter, with accent on the xe, whereas in
neutiquam, pronounced nitiguam (or nyiitiquam, ch. ii. § 149), ne
being unaccented, was elided]; the conjunction xew, a byform of
névé, seu, a byform of sive (older seive), cew from *cé-ve or *cei-ve
(§ 33). (On the pronunciation of ew in these words and in
Latinized Greek words, like Orpheus, see ch. ii. §§ 32, 46.) Hu
is assigned to the Carmen Saliare on the strength of the quota-
tions Leucesie (Ter. Scaur. 28. 11 K.), and cozeulodorieso of Varro,
L. L. vii. 26 (perhaps O Zeu, &e., ch. ii. § 5), but whether rightly or
not is a matter of doubt. (See Rhein. Mus, xxiv, 1 on Latin ew.)

Examples of I.-Eur. eu, Latin # (O. Lat. ow) are: Lat. dicere
(Goth, tiuhan, O. Engl. téon, Germ. nichon): Lat. jageribus Abl. Pl.
from *jagus (Gk. (ebyos); Lat. aro (Gk. etw for *etdw). The Greek
TloAvdedans is in O. Lat. *Pollouces, written in the orthography
of the early Praenestine inscriptions Poloces (C.J. L.i. 55), and
Polouces (xiv. 4094), then Polluces (so the MSS. in Plaut. Bacch.
894; cf. Varro, L.L. v. 73 in latinis litteris veteribus nomen
quod est, inscribitur ut HoAvdedxys, ‘Polluces,’ non ut nunc, Pollux).
246 THE LATIN LANGUAGE. [Chap. IV.

Before a vowel ew (ev) from I.-Eur. ew similarly became ov in
the accented syllable, e.g. ndvas from I.-Eur. *néwo- (see § 10),
and in the unaccented was reduced to x, e. g. déni6 (see ch. iil. § 24,
and for other examples Solmsen, Stud. Lat. Lautg. p. 128); in cloaca,
&e, v has been dropped before the accent (ch. ii. § 53), leaving o.

§ 36. Other examples of I.-Eur. eu. Jupiter, in the usual Latin spelling
Jitppiter (ch. ii. § 130), may have been originally a vocative like Gk. Zed mdrep,
(Can the fragment of the Carmen Saliare quoted above from Varro have 0 Zeu
with Z for the sound dy-, as in Oscan inscriptions written in Roman
characters we find zicolo- as the Diminutive of the word for ‘day,’ like Latin
dikciila 2).

§ 37. OU, U in Inscriptions. The diphthong was reduced very early to
a simple sound, as is natural where the two elements of the diphthong have
so close affinity aso and wu. We have @ for ow (I.-Eur. eu) in the name Lucius
in two of the oldest Scipio epitaphs (C. I. L. i. 32 Luciom ; i. 30 Lucius with
Loucanam, and abdoucit), not later than 200 B. c. ; deducundae, 181 B.¢. (i. 538) ;
Lucius in a dedicatory inscription of the consul Mummius, 146 B. ©. (i. 542) ;
luuci and iurarint (with iouranto, ioudicetur, ioudex, &c.) in the Lex Bantina of
133-118 B.c. (i. 197) ; turato, iudicibus, duco (with ioudiciwm, ioudicatio, ious) on
the Lex Repetundarum of 123-122 8.0. (i. 198), while the spelling with ou
is entirely discarded in the Lex Cornelia of 81 B. c. (i. 202, with iwus, ture, &e.).
Now and then ou occurs for a short vowel, e.g. éoubeatis (beside tousiset) in the
8. C. Bacch. (i. 196) (but see ch. viii. § 29), probowm on old coins (i. 16),
Laowmeda on an old Praenestine vase (xiv. 4108, or Lad-?), possibly to represent
the transition-sound between 6 and %, or in imitation of the Greek ortho-
graphy, in which ov represented the u-sound, v the ii-sound. Ow is some-
times used for @ (I.-Eur. oi), e. g. couraverunt (i. 1419, from Picenum), plowrume
(with Clowi[é] for Cloelius, i. 1297, from near Amiternum), though not at all to
the same extent as ei for 7. (On ou for I.-Eur. ou, see § 41.) The spelling
o for ou (I.-Eur. eu) is dialectal, e.g. Poloces and Losna on a Praenestine mirror
(i. 55) (ef. Umbr. toro- for Latin tauro-, bull). On Latin inscriptions o occurs
for I.-Eur. ou, and for the new ou-diphthong, which arose by syncope in words
like nov(e)ntius, cov(e)ntio, but not for I.-Eur. eu or i. The spellings pobdlico-,
puplo-, &c. (see index to C.I.L.i.), are due to confusion of the two radically
different words piibes and pip(u)lus.

§38. OI. I.-Eur. 01 was o7 till the second century, then came
to be written oe, and finally passed into the sound #, though oe
was still written in some words which belonged to legal or
official diction, e.g. foedus, a treaty, poena (but punio), Poenus,
moenia (but munio), dboedio, in the poetic words fvedus, foul,
démoenus, and in the family name Cloelius. Thus I.-Eur. *oino-,
‘one’ (cf. *oiwo-, Gk. ofos, ‘alone’) (Gk. olvn, the ace, O. Ir. oen,
W. un, Goth, ains, O. Engl. an, Engl. one, an, a, Pruss. ains,
Lith. vénas, O.SI. inti; in Greek the numeral-root used was
I.-Eur. sem- of Lat. sémel, &c., efs for *sem-s, ula for *opia, ev
§§ 86-40.] REPRESENTATIVES OF I.-EUR. SOUNDS. OI. 247

for *sem), Lat. aus, older oenus, oino-. I.-Eur. oi is a grade of
an ei-root, often seen in the Perfect Tense or in a derivative
noun, e.g. woid- in the Perfect of weid-, ‘to know’ (Gk. ofa and
nénov0a from eGo, I.-Eur. bheidh-, Lat. fido, feido), *q¥oina-, ‘an
assessment, fine,’ from q%ei-,’ ‘to value, care for’ (Gk. row7, from
rei, often written riw, borrowed by the Romans, poena, Zend.
kaena-, O. S81. céna, the k and ¢ in these two examples indicating
oi not ei) and similarly in Latin, e.g. foedus, a treaty, from feido,
fido, just as 6 appears in the similar grade of 6-roots, in pondus
from pendo. After initial v- Latin oi became e2, as 6 became @ in
vorsus, versus, &e. (§ 10), e.g. vidi in older spelling veider, from
I.-Eur. *woidai (-ai), Perf. Mid. (O. Sl. védé; cf. Gk. of8a, Goth.
wait, O. Engl. wat, Engl. wot).

In the unaccented syllable oi became ¢/, class, 7 in the nautical
term anqguina, a truss, a loanword from the Greek (dyxotvn),
perhaps adapted to Latin nouns in -ama (Non. 536. 5 M.
anquinae vincla quibus antennae tenentur), but in most cases
was as resistive of weakening as 0 (ch. iii. § 18), e. g. sé-curus from
cura, older coira; impunis (impoene Cato frag. p. 37. 21 Jord.) from
poena (cf, punto); in the final syllable ez, 2 is regular, e.g. Nom. PI.
populi, earlier pop(u)lei, from a still earlier poploe. (Fest. 244. 24
Th. quotes from the Carmen Saliare pilwmnoe poploe, a designa-
tion of the Romans ‘velut pilis uti assueti.’) A spurious
diphthong o/, oe, class. #, has arisen through composition in coepi
(older céépi), coetus for co-itus, and through loss of v before the
accent in Julius, older Iuilio- from *Jo(v)itius (§ 43) (cf. Cloelius,
older Cluilius, from the root *klew-, ‘to be heard, famous’). On cui
from quo}, see ch. vii. § 25, and on ném from noen(um), ch. x. § 18.

§ 39. Other examples of I.-Eur, oi. Lat. minus, Pl. minera, *mine, Pl.
moenia, mitnia, minicipium, com-miinis, im-minis, O. Lat. moini-cipio-, comoinem
(C.I. L. i. 196, of 186 B.c.) (Goth. ga-mains, Germ. ge-mein, ‘common,’ Lith.
mainas, ‘exchange,’ O. Ir. moini, maini, ‘ gifts’), and from the same root mito,
to exchange, mituus, lent (Sicil. Gk. pofros, requital, Goth. maipms, ‘a gift,’
Lett. meetdt, ‘to exchange’); Lat. ator, O. Lat. oit-ile (C.I. L. i, 201. 9), oeti
(i, 603. 6. 8), &c. (cf. Mart. Cap. iii. 236 ‘oisus’ etiam dicitur; sie enim
veteres usum dixere) seem to show, like Gk. ofros, fate, ‘portion,’ the o-grade
of a root eit-, seen perhaps in Ose. eitiuva-, ‘money,’ for *eitu-; ciinae shows
the o-grade of the root fei-, ‘to lie’ (Gk. xei-po ; ef. xotry).

§ 40. OI, OE, U on Inscriptions. 0: is reduced to @ in a Scipio epitaph of
the beginning of the second cent. B.c. (C. I. L. i. 33) with utier ; so usura in one
of Mummius’ tithe-dedications to Hercules (i. 542) of 146 B.c.; muru Acc. on
248 THE LATIN LANGUAGE. [Chap. IV.

a Capua inscription of ¢. 135 B.c. (Eph. Epigr. viii. 476) ; procurandae (with oina
and moinicipieis) on the Lex Agraria of 111 B. ¢. (C.J. L.i, 200). But the spelling
ot, and after it the spelling oe, long continued to be used, especially in such
phrases as faciundum coiraverunt (coeraverunt) in magisterial inscriptions (e. g.
i. 566, of 106 B. ©. cotravere and loid[os] ; i. 600, of 62 B. c. coeravit ; i. 617, of 51
B.¢. coeraver.) ; and Cicero in the laws which he draws up for his ideal state
seems to think that the official style demands the spelling oe (¢. g. ploeres, Legg.
iil. 3.65 oenus, ib. iii. 3.93 coerari and oesus, ib. iii. 4. 10), though in the Lex
Julia Municipalis of 45 8. c. (C. I. L. i. 206) we have regularly cwro, utor, munt-
cipium (once foidere), and similarly on the Lex Rubria of 49 B. ©. (i. 205), ludus
(not loedus), the form used in the Comm. Lud. Saec. and the Mon. Anc., both of
Augustus’ reign, though Virgil MSS. often show moerus for murus (see
Ribbeck’s Index). The traditional nature of this spelling, even in the second
cent. B.¢., is seen from the occurrence of spellings like murus side by side with
coeraverunt, &., €. g. i. 567, of 106 B.c., murum... coeravere ... loedos; i. 568, of 104
B.C., murum ... cotraver- ; Eph. Epigr. viii. 460, of 108 B. c., murum . . . cotraverunt,
Examples from the older literature are, oenigenos ‘ unigenitos’ Paul. Fest. 225,
2Th. ; oenus Plaut. Truc. 104 (B.) ; proilio Men. 186 (P.) ; moenis, obliging, quoted
by Nonius 23. 9 M. from Pacuvius ; moerus Accius Trag. 347 R.; moenio in the
Ambrosian Palimpsest of Plautus (see Index to Studemund’s Apograph) ;
‘loebesum ’ et ‘loebertatem’ antiqui dicebant liberum et libertatem Paul.
Fest. 86. 30 Th. (Locbasius is given as the Sabine name for Liber by Serv.
ad Georg. i. 7 quamvis Sabini Cererem Pandam appellent, Liberum Loeba-
sium ; cf. Gl. Plac. 80. 22 G.). This oe seems to have represented to the Romans
a long ti-sound, the sound, in fact, of Greek # ; and the earlier instances of u
for I.-Eur. oi probably indicate this sound. Plautus (Bacch. 129) puns on
Lydus (Gk. Avdds) and ludus :
non 6mnis aetas, Lyde, ludo cénuenit ;
and oe is often used to express Gk. 3d, e. g. goerus, coloephia (ch. ii. § 28).

§ 41. OU. IL.-Eur. ou before a vowel (ow) became, as we saw
($ 19), first ov in Latin, then in the second century B.c. av,
which in the unaccented syll. fell to w, e.g. éduo (ch. ii. § 24).
For I.-Eur. ou before a consonant we find sometimes 6, sometimes
uw. Occasionally both these spellings occur for the same word.
Thus I.-Eur. *roudho-, ‘red’ (Goth. raups), is Latin rdbus, 7db-7g0
[the form rudigo (with @ like rufus, or with @ like ruder?) is
rejected in the Probi Appendix, 199. 5 K.; cf. the gloss robigo
non ‘rubigo’ C. G.L. v. 144. 32, and see ch. ii. § 24], and ré/us, the
last being shown by its / instead of d or 6 (§ 114), to be dialectal.
Ribi-dus comes from a verb in -eo (cf. himidus from himeo,
ciladus from ecdleo, &e.), *raibeo, with J.-Eur, eu of Gk. épev6a,
while the ordinary form of the verb, ribeo, shows, like rier
(Gk. épvOpés), the ti of the weak grade of the root, I.-Eur. rtidh-.
The same variety of spelling is seen in a word indicating
unshaped metal, &c., ddus and riédus, though the normal spelling
§§ 41, 42.] REPRESENTATIVES OF I.-EUR. SOUNDS. OU. 249

is perhaps raudus (see Georges, Lex. Wort/. s.v.), the weak grade
of this stem (perhaps connected with the stem of rébus, réfus)
being apparently seen in ridis, which means rough, literally,
e. g. aes rude, or rough, metaphorically, e.g. arte rudis, in Ovid’s
criticism of Ennius : Ennius ingenio maximus, arte rudis. [Varro,
L. L. v. 163 deinde (porta) Rauduscula, quod aerata fuit. Aes
‘raudus’ dictum: ex eo veteribus in mancipiis scriptum ‘ raudus-
culo libram ferito’; Festus 356. 4 Th. rodus, vel raudus
significat rem rudem et imperfectam, nam saxum quoque raudus
appellant poetae, ut Accius... hinc manibus rapere roudus (so
the MS.) saxeum; Paul. Fest. 377. 1 Th. Rodusculana porta
appellata, quod rudis et impolita sit relicta, vel quia raudo, id est
aere, fuerit vincta]. This variety of spelling suggests that
I.-Eur. ou became in Latin an aw-sound, which was sometimes
written, like Lat. au from I.-Eur, au (§ 30), as 6, sometimes, like
Lat. ow from I.-Eur. eu (§ 35), as #; though, owing to the
scarcity of reliable instances, it is impossible to determine how
far these spellings corresponded to the pronunciation of the
diphthong at various periods, or how far they were influenced
by the analogy of other grades of the same root. For the higher
grade with I.-Eur. 6u, 6 would have in Latin 6 (§ 50), and the
weak grade of an eu-root with J.-Eur. i would have in Latin
& (§ 23), while a d@grade (I.-Eur. i, Lat. z) was also not
unknown (§ 51). The Latin diphthong corresponding to I.-Eur.
eu had thus a different sound from the representative of I.-Eur.
ou. An ow-diphthong arose in Latin also from I.-Eur. -og(h)*-,
for the velar g(h)? took in Roman lips the sound of v (w). Thus
I.-Eur. *ndg%édo- became in Latin *novedo-, *noudo-, whence
*nidus, an examplé which suggests that the ow from I.-Eur.
og(h)* had the same sound as the ow from I.-Eur. eu, and was
developed in the same way to long w. Also by syncope in words
like O. Lat. noventius (as in the prophecy of Cn. Marcius: quam-
uis nouentium duonum negumate, ap. Fest. 164. 28 Th.) with
Lat. dv, I.-Eur. éw, O. Lat. *novendinae, from ndvem (I.-Eur.
*néwn) and din-, a stem for ‘day,’ seen in O. Sl. dini, ‘a day,’
O. Ind. dina-, &e. For this spurious ow we have first 6, later w.

§ 42. Other examples of I.-Eur. ou. From the root neud-, ‘to use, enjoy’
(Goth. niutan, ‘to enjoy’) comes the Latin nitric, nitrix (in Old Latin notriz,
250 THE LATIN LANGUAGE. (Chap. IV.
Quint. i. 4. 16), perhaps showing the o-grade, I.-Eur. noud- (Goth, naubs,
O. Engl. néad, Engl. need, Germ. Noth. Cf. Latin usus est, there is need) ; Lat,
cliinis seems to represent I.-Eur. Kloun- (0. Scand. hlaunn, ‘haunch,’ but Gk.
«Aous), though the word occurs so seldom in the older writers that we cannot
say whether *clawnis, *clonis were earlier spellings ; Lat. liécus is I.-Eur. *louko-
(0.H.G. 1oh, ‘copse, brushwood,’ the -loo of Water-loo, 0. Engl. 1éah, Eng]. lea),
properly an open space in a wood, like the German Lichtung (ef. collicare,
to make a clearing in a wood), showing the o-grade of the I.-Eur. root leuk-
of Lat. liceo, &c., so that the old etymology ‘lucus a non lucendo’ had a grain
of truth after all. The O. Lat. spelling shows -ow-, e.g. in hoce loucarid on the
inscription of Luceria (C. I. L. ix. 782) ; honce loucom . . . quod louci siet on the
inscription of Spoletium (C. I. L. xi. 4766); the Perfect Part. Pass. of ldvo, to
wash, I.-Eur. léw- (Gk. Aovw) is lautus, later lotus (see Georges, Lex. Wortf. s.v.);
the Dat. (Abl.) Plur. of dds, bévis (I.-Eur. *g¥Gu-s, *gkdw-es) is bobus and bibus ;
formed from the same stem bou- after the fashion of instar (ch. iii. § 36), comes the
word spelt bustar in the Glossary of Philoxenus, and bostar in the Glossary of
Cyrillus, and stated in both glossaries to be the equivalent of the Greek Bovard-
o.ov, a word which must have belonged to the older period only, for Spanish
and Portuguese alone preserve it [Span. bostar, Port. bostal, indicating a Latin
original bastar ; bustar, a place for burning a dead body, (Charisius 38. 19 K.),
is a quite different word, connected with bistum] ; Lat. aber appears to show
the a-grade of the root, like O. Ind. didhar, 0. Engl. ider, Lith. idriiti, ‘to give
milk,’ but the ou-grade of Gk. od@ap, Gen. -aros for -ntos, meaning (1) udder,
(2) fertility of soil, may appear in the (dialectal) name of a river in Latium,
Ufens (modern Uffente) [cf. the Apulian river, Aujfidus (modern Ofanto)], from
which comes the name of one of the Roman tribes Ufentina, in Old Latin
Oufentina (see Index to C.I.L.i.), also Vofentina, and Qfentina (C. I. I. xi. 5702),
in Greek inscriptions ’Q¢-, Ovwp- (Eckinger, p. 44).

§ 43. a for older ovi, ové. By the composition of com-, co- with vir we
get the word *coviria (cf. Volsc. covehria-), which became by syncope *cou(?)ria,
ciria-. The name Jilius (written Zuilio on a lamp found in one of the oldest
graves in the Esquiline burying-place, Ann. Inst 1880, p. 260) seems to come
from an earlier *Jovillio-, from a word found in Oscan in the form diuvila-, later
iuvila-, meaning apparently a gift presented yearly by a corporation or clan
to its tutelary god. Here the older spelling Iuil- suggests that the % repre-
sents not ov(z) but o(v)i, the v having been suppressed before the accent (see
ch. ii. § 53). Jullus is the original form of the substantival name, of which Julius
is an adjectival derivative. The trisyllabic Zulus is an invention of Virgil’s
(Herm. xxiv. 155). Similarly Cluilius, the older form of Cloelius (written on an old
inscription Cloul[is], C. I. L. i. 1297), shows its derivation from the name Cluvius,
a name evidently connected with the root kleu-, ‘to be famous’ (Gk. AéF-0s)
[compare the gloss cluvior: nobilior, Léwe, Prodr. p. 364; Paul. Fest. 39. 2 Th.
refers the word to Clonius: Cloelia familia a Clonio Aeneae comite, est appellata.
The family name retained, as often happens, the older spelling with oe, Cloelius,
not Clilius]. Similarly O. Lat. coventio (couentionid on the §.C. Bacch. of 186
B.C., O.I.L. i. 196), lost its v before the accent, and the two vowels o and e
were fused into 6; contio (but see ch. ii. § 147, and below on nuntius).

§ 44. The spurious diphthong ou. Nontio, the older spelling (denontiari on
§§ 48-45.) REPRESENTATIVES OF I.-EUR. SOUNDS. AI, &C. 251

the Lex Bantina of 133-1188. ¢., C. I. L. i. 197 3 pronontiato on the Lex Repetun-
darum of 123-122 B.c., i. 198; nontiata on the Epistula ad Tiburtes of c. 100
B.0., i. 201 ; pronontiato and pronontiatum on fragments of old Laws, i. 207 and
208) became nuntio at the close of the Republican period (renuntio is the
spelling throughout the Lex Julia Municipalis of 45 B.c., i. 206 ; so nuntia-
tionem on the Lex Rubria of 49 8. ¢., i. 205), though Cicero in his Laws (ii. 21)
uses nontius. Marius Victorinus (12, 18 K.) says the old spelling had ow. For
nundinae the oldest spelling is with ou, noundinum on the S. C. Bacch. of 186
B.C. (C. I. L. i. 196, then with 0, nondinum on the Lex Bantina) (i. 197).

$45. AI. The I.-Eur. root, referred to in § 27, as aiwé-, ‘ time,
life, is perhaps more correctly aiwé- (cf. O. Ind. ayu-, ‘ life’), with
the diphthong originally long, but shortened in Latin aevum,
according to the rule that a long diphthong (including combi-
nations with a nasal or liquid as second element) shortened in
Latin its first element when a consonant followed. Final I.-Eur.
-ai, the ending of the Dat. Sg. of A-stems, shows traces in Latin
of ‘ doublet’ forms, (1) @ (with suppression of the second element,
by a similar syncope as produced exemplar from exemplare), a form
which seems to occur on a few old inscriptions, e.g. Junonet
Loucina (C. 1. L. i. 189), Iunone Loucina Tuscolana sacra (i. 1200) ;
(2) -aé (one syllable), class. -ae (presumably de ; ef. Osc. -ai, Umbr.
-e, Rustic and Late Latin e, ch. ii. § 41) (with a shortening of the
first element, which properly took place only before a consonant),
Similar doublets -é and -0i seem to show themselves for I.-Eur.
-di in O-stem Datives [ (1) class. égu6, ddmind, (2) O. Lat. populot
Romanoi, Numasiot; cf. Osc. -4i, Umbr. -e], though in O-stems
the first (not the second doublet as in the A-stems) established
itself in the classical usage. (On these Datives, see ch. vi. § 23.)
In Greek this I.-Eur. -ai became -d, e. g. xépq, later -a, but in
dialects also -a: (presumably 1), while this I.-Eur. oi became
-, e.g. lnm, later -w, in dialects -ou.

Shortening of long ‘diphthong’ before consonant. It seems to have
been a law very widely spread through the I.-Eur. languages that a long
vowel became shortened before any y (in i-diphthongs), w (in u-diphthongs),
m, n, 1, or], when this was followed by aconsonant. Thus I.-Eur. *wént- from
the root wé-, ‘to blow’ (Gk. d-7u) has become in Gk. 4-évr-, in Goth. vinds, our
‘wind,’ as in Latin véntus (Span. viento) ; the I.-Eur. word for the heel or
the ham (0, Ind. parSni-) is in Gk. wrépva, in Goth. fairzna, and in Latin
pérna (Span. pierna) [curiously enough Mar. Sacerdos (vi. 451. 5 K.)} cites
a derivative of this word, pernizx, swift (lit. ‘strong in the ham,’ and properly
used of horses and other animals, A. L. L. viii. 453) as an illustration of a short
252 THE LATIN LANGUAGE. [Chap. IV.

e, the long pronunciation of which is a barbarism: barbarismus ... fit... si
dicas pernix et ‘per’ producas, quae correpta est] ; the Dat. (Abl. Loc. Instr.)
Plur. Suffix of O-stems, I.-Eur. -dis has become in Greek -ous, as in Latin -eis,
~is (ch. vi. § 48) ; Wéntus is a cognate of lenis. (Cf. Engl. ‘kept,’ ‘wept’ from
‘keep,’ ‘ weep’).

§46. AU. The I.-Eur. long diphthong 4u is seen in the stem
nau-, ‘ship’ (O. Ind. nats, navds, Gen.; Hom. Gk. vnis, vnds Gen.,
Arm, nav, O. Ir. nau, naue or née, Gen., W. noe, ‘a dish,’ like our
‘butter-boat,’ O. Scand. nor), which is in Latin an z-stem xdvis,
like clavis (Gk, kAn(F)is). The -au- of nau-frigium, nau-stibulum
‘vas alvei simile’ (Fest. 172. 23 Th.), claudo shortens the first
element owing to the fact that a consonant follows (see above,
§ 45). [That claudo had the same au, as the equivalent of I.-Eur.
au (e.g. fraus), we see from its sinking to @ in the unaccented
syllable, e.g. ewclidere, like defrudare.| For Latin dv from
I.-Eur. dw, e.g. octévus, and perhaps flavus (beside florus, Gk.
xAwpds), see § 50. Gatus (a trisyllable till late Latin, Harvard
Studies, 1891) is the class. form of older Gavius (Osc. Gaaviis),
with suppression of intervocalic v (§ 70), a name apparently
derived from the root of gaudeo, gavisus sum (Gk. ynOéw for
*yaF eOew) (cf. Raius and Ravius) On the curious remark of Teren-
tianus Maurus about the pronunciation of Lat. au, see ch. ii. § 34.

§ 47. EI. J.-Eur. éi appears e. g. in *réi-, ‘ property ’ (O. Ind.
ras, Gen. rayds), Lat. rés, Gen. ré2 for *réy7, a root in which the
diphthong had apparently in the ‘Indo-European period’ the
doublets éi (éy) and é. The Loc. Sing. ending of E-stems shows the
second of these doublets in Latin, e.g. dié crastini, postridié, &e.

§ 48. EU. The I.-Eur. Nom. *dyéus, ‘the sky,’ shows the long
diphthong éu (O. Ind. dyaiis, Acc. divam and dy4m, diyim; Gk.
Zevs for *Znus, Lat. diés like Acc. diem; on Jévis, &c., from the
stem dyéw-, see ch. vi. § 9). A final év-diphthong arose in Latin
by the suppression of the final -@ of xéve, and produced the form
neu ; ceu is probably to be referred to an older *cé-ve (ch. x. § 11).
The Loc. Sg. ending of U-stems, if this was I.-Hur. -éu, appears
yp xnocti, where the % represents an earlier -éw with shortening
of the first element of -éu (see § 26), so that I.-Eur. *dyéus
should be Lat. *dids (nii-dius-tertius ?).

§ 49. OI. An example of final -di has been already mentioned,
§§ 46-51.] REPRESENTATIVES OF L-EUR. VOWEL-GRADES. 253

the ending of the Dat.Sg. of O-stems. This in Latin shows the
doublets, (1) -d, the classical ending, e.g. gud, dimind, (2) -0i
(presumably -32), an ending found in very old inscriptions, e.g.
Numasioi (Osc. -Gi, Umbr. -e) (ch. vi. § 26).

§50. GU. I.-Eur. du- is seen in the numeral *dktou, a dual in
form, with the sense apparently of ‘two sets of four’ (O. Ind.
aktau, a’ta, Gk. éxré, Goth. ahtau, O. Engl. eahta, Lith. aszti-ni),
in Latin octd. Duals in the Veda show generally -au (I.-Eur. -du)
before an initial vowel, -a(I.-Eur. -d) before an initial consonant,
e.g. dévat and déva, ‘twin-gods ;’ and it is probable that these
doublets existed even in what is called the I.-Eur. period, so that
the -d of Latin octo (cf. ambo, duo) will represent an I.-Eur. -,
and not -du. As I.-Eur. dw became dv in Latin (§ 19), so I.-Eur.
ow is said to have become dv in the corresponding ordinal number,
octavus for *octovus (Gk. éyd0(F)os). The long diphthong seems to
have occurred in the I.-Hur. declension of the word for ‘ox,’ stem
*g¥Ou- (O. Ind. gatis, Loc. giivi, Acc. gm, Nom. Pl. givas, &c.,
Arm. kov, Gk. Sods, Dor. Bés, according to the grammarians,
O. Ir. bou, bé, W. bu, buw, O. Engl. ci, Lett. giws); but the
Latin ds is a doubtful example, for its J- instead of the normal v-
(§ 139) suggests that it is a dialectal (or rustic) form like dvis for
*avis (1.-Eur. *Swi-, § 19), and not a genuine Latin development.

§ 51. Variation (Ablaut) of Vowels. A root like pet- of Gk.
aérecOat, to fly, O. Ind. patati, ‘ he flies,’ appears in the form pt-
in Gk. rrécOat, O. Ind. 4-pa-pta-t, ‘he flew,’ the shorter form being
a syncopated form of the other, due to loss of accent. Similarly
the root ei-, ‘to go’ (Gk. ci-o1, 3 Sg., Lith. ei-ti, Lat. it, older
ei-t, ch. viii. § 2), loses the @ of the diphthong in the P. P. P. *i-t6-
(O. Ind. -ité-, Gk. -iros, Lat. -t¢us), where the accent falls on
the suffix; and eu becomes i, through loss of accent, in I.-Eur.
*bhiiga, ‘flight’ (Gk. pity}, Lat. fia) from *bhetigs, ‘I flee’ (Gk.
gedyw); while en, em, er, el, similarly reduced, appear before
a vowel as n,m, 1,1, eg. Gk. yl-yv-ouat, Lat. gi-gn-o, beside
Gk. yév-os, Lat. gén-us, but before a consonant assumed in Greek
the forms a, pa, Aa, &.8. dards from gev-, to kill, I.-Eur. ghen-,
dpaxdy (O. Ind. dr’ént-) from bepx-, to glance, I.-Eur. derk-, in
Latin en, em, or, ol, e.g. ten-tus (O. Ind. ta-t4-, Gk. ra-rés) From
254 THE LATIN LANGUAGE. [Chap. IV.

ten-, ‘to stretch,’ fors (O. Ind. bhr-ti-, O. Ir. brith, Goth. ga-
baurps) from bher-, ‘to bear’ (see §§ 81, 92). We may call these
reduced forms pt-, bhtig-, bhr-, &c., the ‘weak grade’ of the
roots, and pet-, bheug-, bher- the normal or E-grade. We find
these roots also with their 8 replaced by 6 in such words as Gk.
yeyou-a Pft., otros, a Derivative Noun from the root ei-, ‘to go,’
yévos from the root gen-, pdvos from gh*en-; and gon-, oi-,
gh*on- may be called the O-grade of these roots. There are
also occasionally forms with 6, 6, e.g. Gk. mwrdowa: from pet-,
‘to fly.” This variation, or gradation, called by the Germans
‘ Ablaut,’ of I.-Eur. vowels has not yet been thoroughly explained
or systematized ; the relation for example of i, i to the ordinary
weak grade ti, 1 in words like L.-Eur. *kli-té- (0. H. G. hlit,
‘loud, Zend. srii-ta) beside I.-Eur. *klti-td- (O. Ind. Srii-té-, Gk.
kdb-rds) is not quite clear, nor yet that of the E- and O-forms
to the E- and O-forms, e.g. Gk. twrdopwat beside morgouar. They
are generally called ‘lengthenings’ of the weak grade (of diph-
thongal roots), ofthe H-gradeand of theO-grade, and are by German
philologists classed under the term ‘ Dehnstufe’ (‘lengthened
grade’), [On these see Streitberg in Indog. Forsch. iii. 306, who
explains them as produced by syncope of a following short vowel
in *bhér (Gk. gdp) for *bhdr(o)s, &c., *réks- (Lat. réa-i) for
#rés-8s-, &c]. Nor have the grades of other than E-roots been
properly equated to grades like pét-, pt-, pot-; the variation
of the root do-, ‘to give,’ for example, which has 6 in Gk. 3f-Be-fu1,
Lat. dé-num, but a short vowel in Gk. d0-rds, Lat. dé-tus, of the
root sé-, ‘to throw, or ‘to throw seed, with é in Gk. ¢-n-yu for
*ot-on-pt, Lat. sé-men, but with a short vowel in Gk. é-rds, Lat.
sa-tus, of the root sta-, ‘to stand, with 4 in Gk. ¢-ornw for
*ou-ord-yt, Lat. sta-re, but with a short vowel in Gk. ord-rdés,
Lat. sté-tus; similarly the variation of & and 4, eg. in the
I.-Eur. root meaning ‘to drive, %&- in O. Ind. aji-mi, Gk.
dyw, Ir. agaim, Lat. digo, 4@- in O. Ind. aji-, ‘a contest,’ O. Ir. ag,
‘a contest,’ Lat. amb-dges; and the variation of 6 and 6, e.g. in
Lat. fodio beside fod, Gk. 63-u4 beside ed-ddys ; not to mention
the variation of a and 6 in Lat. sedibo beside scdbis, deies (Gk.
dxpos, Hom. &xpis, a hill-top) beside O. Lat. deris, a rugged hill
(Gk. ékpus, a point), from the root ak-, ok-, ‘ sharp.’
§ 51.) REPRESENTATIVES OF I.-EUR. VOWEL-GRADES. 255

Examples of this variation of vowels in Latin are I. in E-roots:
(a) Weak grade in (1) Derivative Nouns with A-suffix, O-suffix,
TI-suffix, &e. (ch. v. §§ 2, 42), e.g. fliga (Gk. pvy-7) from the root
bheug-, ‘ to flee,’ jiig-wm (O. Ind. ytig-4m, Gk. @y-dv) from the
root yeug-, ‘to join’), fors (O. Ind. bhr-ti-), mens (O. Ind. ma-ti-),
mors (O. Ind. mr-ti) ; (2) in P. P. P. with TO-suffix, e.g. diic-tus
from diico for *deuco, is-tus (A. L.L. ii. 607) from dro for *euso
(Gk. ew for ethw), per-culsus for *kld-to- (cf. clades); (3) in
Reduplicated Present-stem, e.g. gi-gu-o from the root gen-, sido
for *si-sd-o from the root sed-; (4) in Nasalized Present-stems,
e.g. ju-n-g-o from the root yeug-, %-n-quo from the root leiq?-
(Gk. Acirw), ft-n-d-o from the root bheid- (Goth. beita, ‘I
bite’); (5) in some Present-stems with the YO-suffix, e.g. /tig-io
from the root bheug-.

(2) Normal or E-grade in (1) Neuter ES-stems, e.g. gen-us
from the root gen- (Gk. yév-os), decus from the root dek-, nemus
from the root nem-; (2) Present-stems formed with the Thematic
Vowel, e.g. veho from the root we&h- (O. Ind. vah-, Lith. vezi,
O. Sl. veza), seqguor from the root seq?- (O. Ind. sac-, Gk. érouat,
O.Ir. sechur), fido, O. Lat. feido, from the root bheidh- (Gk.
mlm for *pedw), dco, O. Lat. decco, from the root deik- (Gk.
detxvupn), dco, O. Lat. douco for *deuco (§ 35), from the root
deuk- (Goth. tiuha), ao for *ewso from the root eus- (Gk. eo
for *ethw), pluo, O. Lat. plovo for *plewo, from the root pleu-
(Gk, wAe(F)w). (On the w of pluo see ch. viii. § 6.)

(c) O-grade in (1) Derivative Nouns with A-suffix, O-suffix,
&e., e.g. procus from the root prek-, ‘to ask ’ (Lat. precor), domus
(Gk. 8duos) from the root dem-, ‘ to build’ (Gk. dé4), toga from
the root (s)teg-, ‘to cover, thatch’ (Gk. oréyw, Lat. tego); (2)
Causative Verbs, e.g. moneo, lit. ‘cause to remember,’ from the
root men- (Lat. me-min-i), torreo, lit. ‘cause to dry up, from the
root ters- (Gk. répoopat).

II. in E-suffixes. (1) Nouns of the Second Declension with
Voe. Sg. in -é, e.g. eque, Nom. Acc., &c. in -os, -om, e.g. equos,
equom ; (2) N-stems, R-stems, S-stems, &c., e.g. temp-iis (O. Lat.
temp-0s), temp-Or-is for *temp-ds-es Gen., temp-ér-i Adv., aug-us-tus
for *aug-ds-to-, auatlium for *aug-s-ilio-, pléb-és, déc-or (O. Lat.
dec-08) ; nd-mén, car-n-is, rati-bn-i. In the I.-Eur. declension
256 THE LATIN LANGUAGE. [Chap. IV.

of these stems the suffix may have shown in the Nom. Sg.
é when accented, 6 when unaccented, e.g. Gk. d0-ri7p, 56-rwp,
in the other ‘strong’ cases € when accented, 6 when unaccented,
e.g. Gk. na-rép-a, ppd-rop-a, and in the ‘ weak’ cases the weak
grade, e.g. Gk. wa-rp-ds, ma-rpd-ov (O. Ind. pi-tf-gu), but the
divergences of the suffix-form have been to a great extent
removed in the various I.-Eur. languages, e.g. Gk. nrépos as
well as ytpds, marépwy instead of marpdv. The alternation of
strong and weak stems in Declension, depending on the
accentuation of the stem or the suffix, has left its mark in
the divergent form of words like Gk. movs, qmodds (cf. Lat.
trt-pod-are) and Lat. pes, péd-is (cf. Gk. né(a, rpd-me(a, lit. ‘ four-
footed,’ me¢és, &c., for *ped-ya-, *ped-yo-), Lat. pecten and
Gk. xreés for *axrevs, &e.

The combination yé, wé was treated somewhat similarly to
ei, eu, being reduced by the loss of accent to 1, ti, e.g. Gk. ia-vos
for *stip-nos (O. S81. stinti) from the root swep (O. Engl. swefn,
‘a dream’), Gk. répvoi, last year, for wépur. (O. Ind. par-ut) from
the root wet- (Gk. (F)éros, a year, Lat. vétus). This root wet-
seems to have lost by procope an initial 4, and the reduced form
of 4wét- appears in Gk. éu-avr-ds, just as the reduced form of
Awég- (Gr. 4(F)ééw) in the aug- of Lat. aug-co, Gk. aifw, O. Ind.
djas-, ‘strength,’ while tig- the reduced form of wég- (with
procope of initial 4) appears in O. Ind. ukg-,‘to grow strong,’ Gk.
vyujs for *vyujs, &. And as we sometimes find i, ti beside f, ti the
weak grades of ei, eu, so we find the same long vowels in forms
of yé- and wé-roots, e.g. ki- (O. Ind. Siina-, ‘ want, emptiness ’)
from the root kwé- (Gk. «(F)evés, empty). On the variation of
i with yé (iye) in I.-Eur. YO-stems, see ch. v. § 4, of i with yé in the
I.-Eur. Athematic Optative (O. Lat. siés and simus, &c.), see
ch. vii. § 55. Latin con-céit-io for *con-quét-io from quatio, ab-te-i0
(with the first syllable short in the older poetry), for *ab-yéc-io
from jdcio, shows that the Latin language had the same tendency
as the I.-Eur. to reduce unaccented yé to i, wé to t, and in many
cases it is impossible to say whether the reduction belongs to the
“Indo-European ’ period or is a Latin development.

_ TIL. in other roots, I.-Eur. & (Lat. é) varies with Lat. % in
Lat. sé-men (Gk. 7j-ya, a casting) and sé-tus (Gk. é-rds) from the
§ 51.) REPRESENTATIVES OF L-EUR. VOWEL-GRADES. 257

root sé-, féc-2 (Gk. ¢-Onx-a) and fac-io, an extension of the root
dhé- (Gk. r(-@y-1) ; I.-Eur. 6 (Lat. 6) varies with Lat. % in Lat.
dé-num (Gk. d6pov), and di-tus (Gk. dords, ddous) from the root
do-, cds (O. Ind. %B.-, ‘to sharpen ’) and cé¢us, which in O. Lat. meant
‘sharp,’ ‘shrill’ from the root ko-; I.-Eur. a (Lat. a) varies with
Latin & in Lat. fa-ma (Gk. ¢7j-un) and fat-eor (ch. viii. § 32) (Gk.
pards), sta-re (Gk. tornue for *o1-ord-wu) and std-tus (Gk. ora-rés).
In all these cases the Latin weak-grade vowel is 4, while in Greek
we have « for E-roots, o for O-roots, ¢ for A-roots, but in Sanscrit
¥ for all roots (e.g. hi-ta- from root DHi-, di-ti- from root Dd,
sthi-té- from root sTA-), just as we have Lat. @ of pater, which is
also & in Greek and other languages (Gk. warp, O. Ir. athir, Goth.
fadar) represented by Sanscrit i (pitdr-). We find even in Greek
occasionally @ in the weak form of E-, O-roots (e.g. kexddovro
beside éxexyjdet, ddvos beside dépov), so that we are perhaps justified
in supposing a short a-sound (é) to have been the form of the weak
grade of 6, 6, as well as of 4, in European languages (see § 3).

The Latin tendency to weaken every unaccented vowel has
greatly obscured the traces of the I.-Eur. variation of vowels ;
prosperus, for example, has spd- (cf. O. Ind. sphir4-, ‘ wealthy’) the
weak grade of spé- of spés, spéro (O. Ind. spha-, O.S]. spé-ti),
with @ weakened before 7 in the unaccented syllable to ¢, and
créditus (O. Ind. $r4d-dhita-) has the same vowel weakened to %.

Words like fringo (% is shown by con-fringo for con-frengo, &e.),
with the weak grade frég- in the Nasalized Present-stem of the
I.-Eur. root bhreg- (Goth. brikan, ‘to break’), gridus, gradwor,
with a weak form gréd- from the I.-Eur. root ghredh-, ‘to step’
(Goth. grips, ‘a step’), suggest that in Latin (as perhaps in other
languages, M. U. v. pref.) ri, ma, &c. were the weak grades of
ré, 18, mé, né, although we have seen Lat. dr, o/, ém, én to be
the weak grades of ér, él, ém, én where the liquid or nasal follows
the & But we occasionally find in Latin (and perhaps in the other
Italic languages) @ in forms of E-roots where & is not preceded by
a liquid or nasal, e.g. pateo (Osc. pate-) from the root pet- (Gk.
merdvvupt); its relation to the obscure or indeterminate vowel
(like the Hebrew sh°va), written @, 0, °, &c., is not clear (see §§ 3,
83, 94).

As an I.-Eur. a is the weak grade of 6, so du may be the weak

8
258 THE LATIN LANGUAGE. [Chap. IV.

grade of du; the Plautine aws-culum, for example, may then exhibit
the weak grade of the stem *éus-(O. Ind. as-, Lat. ds-). It may
also be the weak grade of éu (Cawrus or Corus, Lith. sziguré,
beside O. SI. séverti, is quoted as an example), and of au; and
similarly di of 6i, éi, ai. It is also possible that as 4 varies with
5, SO al may vary with oi (e.g. Lat. aemidus and Gk. oiddw), and
au with ou (e.g. Gk. xavdds, a stalk, and xotdos for *xofidos,
hollow) ; and some explain in this way Latin forms like /ivo beside
Gk. Aovw (on which see § 19).

§ 52. I.-Eur. and Lat. 6 and 6. Lat. procus (Lith. praszyti, O. SI. prositi)
from I.-Eur. root prek- (Lith. perszit, and Lat. preces, precor) ; Ta, noceo, the
Causative of I.-Eur. nek- (O. Ind. nas., Gk. véeus, Lat. nex, &e.); Lat. pondo,
in weight (used with ellipse of litra, e.g. centum pondo es, ‘you weigh
a hundred pounds’), the Abl. of an O-stem *pondus, -i, beside pondus, -eris,
from pendo, to weigh; Lat. domus, an O-stem in Plautus and the writers
before Sulla, I.-Eur. *domo- (0. Ind, dima-, Gk. dduos) from I.-Eur. root dem-,
‘to build’ (Gk. 5€uw, déuas, form, Goth. tim-r-jan, ‘to build,’ Engl. tim-ber,
Germ. Zim-mer; Goth. ga-timan, ‘to be suitable, conformable,’ Germ. zie-
men); Latin réta, ritundus (O. Ir. roth, ‘a wheel,’ W. rhod, 0. H. G. rad, Lith.
rftas) from I.-Eur. root reth-, ‘to run’ (0. Ir. rethim, W. rhedu Inf., Lith.
ritt, ‘Iroll’); Latin drbus, I.-Eur. *érbho- (Arm. orb, Gk. dppavds, dppo-Béra:,
0. Ir. orbe, from stem *orbio-, ‘an inheritance,’ Goth. arbi) from I.-Eur. root
erbh- (O. Ir. erbim, ‘I entrust, bequeath’); ciinae from *coinae (cf. Gk. xoir7)
from Kei- ‘to lie’ (Gk. «etpau).

§ 53. 6 and 6. This 6 is best seen in the Noun Suffixes -6n-, -dr- which vary
with -én-, -ér- and -én-, -ér-, as well as with -6n-, -dr-, e.g. Lat. himod Nom.,
hominis for *homénis Gen., Lat. ddtor, datoris beside Gk. durwp, Swropos and Sorjp,
doripos (see ch. v. § 57); in Greek it is seen also in the Perf. of verbs with 6,
e. g. TéwxTa from Onyw, dpéwra from dpinut, &ec. The root of Lat. fds, Flora is
bhls- (Ir. blath, Goth. blé-ma, Engl. bloom ; O. Engl. blés-tm, Engl. blossom ;
0. Engl. blé-wan, Engl. to blow, of flowers), which is connected with the root
bhlé- of Lat. flémina, congestion of blood (Goth. uf-blésan, Germ. auf-blasen),
just as the root plé-, ‘to be full’ (Latin plenus, Gk. wAnpys; Lat. plébes, Gk.
mAHG0s) seems to be connected with the root plo- of Ir. lar, Engl. floor, &e. ;
but these are rather to be explained like 8nd- and gna- of Lat. gndtus, and
gnérus (§ I).

§ 54. 6-&, 6-&. From root léd-, ‘to leave, to let’ (Goth. létan, O. Engl. létan,
Engl. let), with d-grade in Goth. lai-lot Perf., we have in the weak grade Lat.
ldssus (Goth. lats, Germ. lass); Lat. catus, which Varro makes the equivalent
of dcitus, used in Old Latin and in the Sabine dialect of sounds, i.e. sharp,
shrill (Z. ZL. vii. 46 apud Ennium :

iam cata signa fere sonitum dare voce parabant.
Cata acuta; hoc enim verbo dicunt Sabini: quare: ,
catus Aelius Sextus
non, ut aiunt, sapiens, sed acutus) is in O. Ind. Sité., ‘sharp,’ from S., ‘to
§§ 52-56.] REPRESENTATIVES OF I.-EUR. VOWEL-GRADES. 259

sharpen,’ and is connected with Lat. cis, 2 whetstone (like dis from root
dé-). Similarly we have rd-tus beside ré-ri, fuic-io beside féc-i (ch. viii. § 41), and
from I.-Eur. ok-, ‘swift’ (connected with ak-, ‘sharp’ ?) (0. Ind. asa-, Gk.

avs, Lat. dcior) acupedium (presumably with d), equated in the Philoxenus and
Cyrillus Glossaries to éfunodia (cf. G1. Plac. p. 7. 40 G.; the acupedius of Paul.

Fest. 7. 19 Th. is a mistake, see Class. Rev. v. p. 9); acctpiter (écc- Ter. Maur.

1267) is probably a corruption of *acu-peter (cf. O. Ind. asu-pdtvan- for *oku-,

‘swift-flying,’ and Gk. dxv-rérys, the epithet of a hawk in Hesiod, Op. 210),
due to a popular etymology from accipio; the form acceptor, the eG of
O. Span. acetore, is used by Lucilius (inc. 123 M.) exta acceptoris et unguis,
but is censured by the grammarian Caper (p. 107.8 K. accipiter non ‘acceptor’) ;
dm4arus shows the root dm- (O. Ind. am-la-, ‘ sour’), which is usually regarded
as the weak form of 6m-, ‘raw’ (O. Ind. aima-, Gk. dds).

§ 55. 4-6. Examples of this interchange are Lat. atrox for *adrox (Arm.
ateam, ‘I hate’) and odiwm; scabo, to scrape (Gk, mata, to dig) and scobis,
sawdust ; acuo, acus, acies, acer-bus (Gk. dxpos, Hom. dxpis, a hill-top) and ocris
(Gk. éepis, a point, Sepnein; an epithet of unhewn stone in Homer), an Old
Latin word for a rugged hill. (Fest. 196. 17 Th. ocrem antiqui, ut Ateius
Philologus in libro Glossematorum refert, montem confragosum vocabant, ut
apud Livium :

sed qui sunt hi, qui ascendunt altum ocrim?

. unde fortasse etiam ocreae sint dictae inaequaliter tuberatae ; in Umbrian,
and Marrucinian the stem ocri- (Nom. Sg. ocav, in Umbrian) seems to bear the
sense of citadel]; ancus, an Old Latin word for a person with a crook-elbow
(Paul. Fest. 15. 3 Th. ancus appellatur, qui aduncum bracchium habet, et
exporrigi non potest), which went out of use c. 200 B.c., to judge from the
fact that the word survives only in Portuguese anco, ‘the elbow’ (Gk. dywv
and dyxdAn, O. Ir. écath, ‘a hook,’ from root ank-), and wncus, a hook, reduncum
bracchium, aduncus unguis (Gk. éy«os, a hook); doceo and Gk. &8donw for
*di-5ax-oxw. We have Oscan a, Latin o in tongere, a word used by Ennius,
declared by Aelius Stilo to be equivalent to noscere, and to be still employed
in the Praenestine dialect (Paul. Fest. 539. 5 Th. tongere nosse est, nam
Praenestini ‘tongitionem’ dicunt notionem. Ennius: alii rhetorica tongent.
Cf. Fest. 538. 9 Th. [tongere Aelius Stijlo ait noscere esse), appearing in
Osean in the noun tangion- with the sense of sententia, e.g. senateis tanginud,
‘senatus sententia,’ the cognate of our word ‘think,’ probably from a root
teng-. (Cf. the (dialectal ?) variation of names like Blossius, Blassius ; Fabius,
Fobius. |

§ 56. &and &, I. Eur. pig-, ‘to fasten’ (cf. pak-) (Sanger. pAda-, ‘cord,’ Dor.
Gk. rayrups and éndyqv, Mid. High. Germ. vuoge, ‘deftness in fastgning,’ &c.,
Mod. Germ. Fuge, and Goth. fagrs, ‘suitable),’ Lat. pdngo (with @; cf’ com-pingo),
pepigi (from *pe-pig-i) and com-pages, pdciscor and pac-em Acc. ; I.-~Eur. swid-, ‘to
make pleasant’ (O. Ind. svddati and svadaté, Dor. Gk. Eanes and Hom. Gk.

evabor, dBov Aor.), Lat. suddeo; I.-Eur. *nis-, ‘the nose,’ probably Nom. nas(s),
Gx, n&s-os (0. Ind. nas: and nas-, Lith. nésis and O. S1. nosti), O. Lat. nassum,
class. Lat. ndsus (ch. ii. § 129), na@ris; similarly Lat. sdgio, sagus and sdgax
(Dor. Gk. &yéopa:, Ir. saigim, ‘I seek,’ Goth. sdkja, Germ. suche, O. Engl.
séce). The Latin words sdgio and sdgax were used of hounds on the track,

82
260 THE LATIN LANGUAGE. [Chap. IV.

whence praesdgio, to ‘scent out’ the future (Cic. Div. i. 65: cf. Ennius, 4.
375 M. nare sagaci Sensit; voce sua nictit ululatque ibi acuta); sdga was
applied to an old match-maker, like Gyllis in the first Idyll of Herondas
(Non. 22. 34 M. sagae mulieres dicuntur feminarum ad libidinem virorum
indagatrices; cf. Lucil. vii.6 M. saga et bona conciliatrix). So Latin acer, stem
deri-, and dcuo (Gk. &xpos, Lith, asztri-s, &c., show I.-Eur. &); Lat. deri- beside
Gk. dxpo- reminds us of Latin sdcri- (sdcres porci, pigs for sacrifice, Plaut.)
beside sdcro-.

§ 57. 6 and 6. I.-Eur. éd-, ‘to eat’ (O. Ind. ad-, ‘to eat’ and adya-,
‘eatable,’ Gk. Z5oua: and é5y8és, Goth. itan and étum, Lith. éd-és Part., O. Sl.
jad-ii), Lat. édo and édi Perf. ; Lat. léx, lag-is and légo, &e. ; I.-Eur. rég- ‘to stretch,
rule’ (0. Ind. raji-, ‘a row,’ raj-, ‘a king,’ O. Ir. rigim, ‘I stretch,’ ri, ‘a king’),
Lat. régo, I rule, rég-,a king. This lengthening appears chiefly in Preterites,
e.g. O. Sl. nésii, ‘I carried’ (beside nésa, ‘I carry),’ Goth. sétum, ‘we sat’
(beside sitam, ‘we sit’), O. Ir. ro midar, ‘I judged’ (beside midiur, ‘I judge,
think,’ from the root méd- of Gk. pédouzac), and in the nouns derived from Verb-
stems like lax, rex. The occurrence of the long vowel in some Present-forms
like Lith. éd-mi, ‘I eat,’ is probably due to the use of a Perfect-stem as
a Present (like Gk. dywyw Pres. from dywya Perf.).

§58 iandi. I.-Eur. *wiro-, ‘a man’ (0. Ind. vird-, Lith. vyras have 1;
0. Ir. fer from *wiro-, Goth. vair from *wiro-, 0. H. G. wer, Germ. Wer-wolf,
Engl. were-wolf and Lat. vir have); .-Eur. *gtiro-, ‘lively ’ (O. Ind. jira-, Lat.
vireo). But most examples of Latin 74 are really cases of i-ei, e. g. fides and

ido, older feido (see § 13).

§59. Sand 6. TheI.-Eur. root 6q3-, ‘to see’ (Gk. dPouou and dy) appears in
Lat. dc-ulus ; the root 6d-, ‘to smell’ (Gk. 6544, Arm. hot, and Gk. 5vo-wdys, Lith.
tidziu) in Lat. ddor; ndgt, ‘naked’ (O. Ind. nagné-, Ir. nocht, Goth. naqaps,
and Lith. ntigas), in Lat. nadus for *név(i)dus; dlen-, ‘the elbow’ (Gk. wrévy,
Goth. aleina, ‘a cubit,’ ‘ell,’ O. Ir. uile, uilenn Gen., W. elin, O. Engl. eln,
Eng). ell, el-bow) is Lat. ulna for *0l(i)na. As with é (varying with é) we find
6 (varying with 6) in the Perfect-stem, in nouns derived from Verb-stems, &c.,
e.g. Gk. dn-wr-a, dy.

§ 60. & and a. I.-Eur. *ni, ‘now,’ from the root neu- of *newo-, ‘new’
(O. Ind. nu, Adv. and Particle, nandm, Gk. vi, viv, O. Ir. né and ni, a Verbal
Part. often used with the Present Tense, e.g. no chanim, ‘Iam singing,’ Goth.
nu, O. Engl. nu, Germ. nun and sometimes nu, Lith. ni-gi, 0. Sl. ny-né),
Lat. nu-dius tertius the day before yesterday, lit. ‘now the third day,’ -num in
etiam-num, &c. (Gk. rol-vuv), niin-c; L.-Eur. la- from the root leu-, ‘to loose’ [Gk.
Bovairév-de, about the time of loosing the oxen from the plough, towards
midday (Class. Rev.ii. 260; Schulze, Quaest. Ep. p. 321), and Avrés), Lat. so-litus ;
I.-Eur. st-, ‘a sow’ (O. Ind. st-kard-, Gk. ts, O. Engl. sii), Lat. sis and sii-cerdae,
‘stercus suillum’ (Fest. 432. 8 Th.; cf. Paul. Fest. 493. 2 Th.; Non. 175.
14 M.) (W. hw-ch, Goth. sw-ein, O. Engl. sw-in); I.-Eur. ti, the accented
and the unaccented form of the 2nd Personal Pronoun (Gk. r#-y-7 and ov, &c.),
Lat. ta (on ti-quidem, see ch. iii. § 51); I.-Eur. pu-, ‘to rot’ (Gk. 7#w, Goth. fals
ist, ‘he stinketh,’ O. Engl. fal, Engl. foul, Lith. pati; Gk. mos for *rtcos), Lat.
piteo and piiter ; I.-Eur. *kuti-, ‘skin’ (Gk. éy-xcvriand O. Engl. hyd, Engl. hide,
0.H.G. hat, Germ. Haut), Lat, ciitis, But usually the alternation of @ with 4 in
Latin is the alternation of I.-Eur. eu, ou with t, e. g. dco and dix, figi and
§§ 57-62.) REPRESENTATIVES OF I.-EUR. VOWEL-GRADES. 261

fiigio (see § 23). The t-grade of eu-roots and the i-grade of ei-roots are
frequent before certain suffixes, especially the TO-suffix (ch. v. § 28), e.g.
I.-Eur. *li-to- (Lat. so-létus, Hom. Bov-Avréy-de), though the i-grade and
i-grade are most common in the P.P. P., e.g. Gk. Av-rés (ib.) (ef. defratum,
must boiled down, Plaut. Pseud. 741, Mar. Vict. 24. 15 K. in defruto apicem
secundae syllabae imponere debetis, nam a defervendo et decoquendo fit tale ;
but defriitum Virg. G. iv. 269). [See Osthoff’s list of forms with i, & in Morph,
Unters. vol. iv., such as Lat. fa-mus (I.-Eur. *dhi-mo-, 0. Ind. dhii-md-), situs
(Gk. veo-cdrtiros, O. Ind. syii-ta-), pi-rus, &c.] .

§ 61. 6 and & Cognate with O. Ind. asdn-, Lettish asins, ‘blood,’ is an
O. Lat. word aser, blood, with a derivative *aseratwm, a mixture of wine and
blood (Paul. Fest. 12, 19 Th. assaratum apud antiquos dicebatur genus
quoddam potionis ex vino et sanguine temperatum, quod Latini prisci sangui-
nem ‘assyr’ vocarent; Gl. Philox. 23. 56 G. aser: aia) which appears
with an e-sound in Greek (gap in the Cretan dialect, wéAay efap Callimachus).
Another Old Latin word sacena, a priest’s knife or axe, whose byform scéna
recalls the Irish word for a knife, scian (stem scén4-) (Fest. 466. 16 Th. scena ab
aliis; a quibusdam ‘sacena’ appellatur, dolabra pontificalis; id. 488. 33 Th.
scenam genus [fuisse ferri] manifestum est, sed utrum securis an dolabra
sit, ambigitur. quam Cincius in libro qui est de verbis priscis, dolabram ait
esse pontificiam. Livius in Lydio:

corruit quasi ictus scéna, haut multdé secus)

may be, like another word for a priest’s knife, secespita (Fest. 522.4 Th. ; Paul.
Fest. 523. 3 Th. dicta autem est secespita a secando; ‘Serv.’ ad Aen. iv. 262)
derived from seco, to cut. O. H. Germ. sahs, O. Engl. seax Neut., ‘a knife,’
suggest connexion with Lat. saxum. Lat. aries, Gk. épi-gos, is in Lithuanian
éras, ‘a lamb,’ with 6, and has in Umbrian some e-sound, erietu Acc., so that
it has been suggested that Lat. d, Gk. « (but see § 51 on Kexddovro with
a from root «73-), may be weak grades of an é-root; and the same explanation
might be given of aser and sacena, In several cases of Lat. d, Gk. ¢ we have
a Gk. byform with 1, e.g. Lat. pateo, pando, Gk. merdvvups and nitynye; Lat.
quattuor (Osc. petora, Umbr. petur-), Dor. Gk. réropes, Aeol. réovpes and Hom.
nioupes; Lat. lapis, Gk. Aéwas, and in Hesych Aip. . wétpa. Other examples
are: Lat. gradior, and gressus (O. Sl. greda, ‘I come,’ O. Ir. ingrennim, ‘T
pursue,’ from root grend-, Goth. grips, ‘a step’); Lat. magnus (cf. mdgis, major,
ch. ii. § 55), Gk. wéyas (Goth. mikils, Arm. mec); Lat. nancis-cor, nactus, Gk.
é-veyceiy (Lith. naszti, ‘to carry,’ O. Sl. nesti, O. Ir. conicim, ‘Iam able’);
Lat. labium, labrum (by analogy of lambo?), O. Engl. lippa from root leb- (?) ;
Lat. glacies and gelu; Lat. alnus, Lith. elksnis, O. SL. jelicha, O. H. G. elira and
erila, Germ. Erle). Farcio (cf. frequens) shows dr (cf. confer(c)tus), perhaps by
metathesis (like Gk. ¢aprrés beside ppaxrds) (but see § 92).

§ 62. Gu-éu. Another example of *dus-, the weak grade of the I.-Eur.
stem *éus-, ‘mouth,’ is awreae, whence, by composition with dgo, auriga (Paul.
Fest. 6. 27 Th. ‘aureax’: auriga, ‘aureas’ enim dicebant frenum, quod ad
aures equorum religabatur; ‘orias’ quo ora cohercebantur, with an absurd
reference to auris, ear), a byform of dreae (Fest. 202. 23 Th. ‘oreae’: freni
quod ori inseruntur ... Naevius in Hariolo:

depréindi autem ledéni si obdas éreas,
262, THE LATIN LANGUAGE. [Chap. IV.

like our proverb ‘to beard a lion’); austium (C. I. L. i. 1463) (O. Pruss. austin,
‘mouth,’ 0. Ind. é8tha-, ‘lip’), a byform of dstium (Lith. ista, ‘mouth of river,’
Lett. dsta, ‘harbour’). (Schmidt, Pluralb. p. 221.)

Y, W.

§ 63. y. I.-Eur. initial y is represented in Greek sometimes
by the rough breathing, e.g. ‘Yd«-w00s, from I.-Eur. *ytiwnk-,
sometimes by ¢ e.g. (uydv (I.-Eur. *ytigo-), the same letter as
we find used for an initial y- or yy-sound which has developed
from an original dy-, e.g. Zeds (I.-Eur. *Dyéu-). But in Latin
both these kinds of I.-Eur. y are represented by j, as we write
the letter, but, as the Romans wrote it, i, e.g. juévenc-us, juigum.
The question whether this 7 was pronounced like a consonantal
spirant (y), or like the half-vowel 2, is discussed in ch. ii. § 48. In
the middle of a word it is often difficult to say whether the
original form of the sound, the ‘ Indo-European form,’ as we call
it, is more correctly expressed by the spirant y, or by the vowel
1 (oriy, ii, ay, &c.). This would no doubt often depend upon the
preceding consonant or consonant-group. The I.-Eur. word for
‘middle,’ for example, we write *médhyo- (cf. O. Ind. mddhya-,
Gk. yéo(c)os, Goth. midjis, O. Sl. mezda, ‘the middle’); the
word for ‘ paternal,’ *patrio- or *patriyo- (*patrijo-, *patrayo-)
(cf. O. Ind. pitriya-, Gk. mdrpios). In Latin, this suffix -yo-,
-io- appears as -i0-, médius, pitrius, though a word like sdcius
gives us a clue that -yo- may often have been the original form.
For in this word the gu of the root seq3-, ‘to follow, accompany’
(Lat. seguor), has become ¢ in the derivative with this suffix, and
with the o-grade of the root ; and this would not have happened
unless q? (Lat. gu) had preceded a consonant (§ 116) (cf. Gk.
a-ocontip from *écc0- for *soq#yo-); so that the word originally
socyo- has become in time a trisyllable, *socio-.

At a later period, owing to that wave of Syncope which, as
we saw, passed over Late and Vulgar Latin, this Adjective-
ending -iws, and similar dissyllabic endings, were reduced to
single syllables, the ¢ (now become y) merging itself in the
preceding consonant and giving it a palatal character, e.g. Titius
became *Tityus, and then something like *Zitsus; the i ‘lost
§§ 63.] REPRESENTATIVES OF I.-EUR. SOUNDS. Y, W. 263

itself in a sibilant sound,’ as a fifth-century grammarian puts it
(perdit sonum suum et accipit sibilum ; see ch. ii. § 90 ; also § 151).

It is this palatalization of a consonant before an 7, reduced to
y, Which has made many Romance words, especially French, so
unlike their Latin originals, e.g. Fr. bras from Lat. bracchium,
through *braccy-; niéce from eptia, through *netty-; ache
from dpiwm, through *apy-; rage from rabies, rabia, through
*raby- ; singe from stmia, through *simy- (see ch, iii. § 13). The
history of the suffix -yo- in Latin is very like that of the suffix
-lo-. After a consonant both developed a parasitic vowel ;
*soc-yo- became *soc-iyo-, socius, as *oc-lo- (from root oq?-)
became *oc-olo-, oculus; and this vowel was in both cases
absorbed by syncope at a later period, *soc-yus, *oclus (whence.
the Romance forms, e.g. Ital. occhio). Between vowels y was
dropped in Latin, e.g. I.-Eur. *eya-, the Fem. of the Demonstra-
tive (Goth. ija Acc.) is Latin ea (so in Umbro-Osc., Umbr. eo, ea,
Ose. iG, 70; von Planta, i. p. 175). Causative Verbs, which
ended in I.-Eur. in -eyé (e.g. O. Ind. manéya-mi from root man-,
the I.-Eur. root men-; Gk. ¢o8éw from root ¢¢8-), end in Latin
in -€6, e.g. moneo, I remind, ‘ cause to remember,’ from root men-
of mémini for *me-men-i, I remember. On the weakening of
-yé- to {in J.-Eur. and in Latin (e.g. ab-icio), see § 51, p. 256.

The intervocalic j (y) which we find in major, aio, &e. (pro-
nounced ‘miyyor,’ ‘dyyo,’ ch. ii. § 55), has arisen through
suppression (or assimilation) of 4 (for I.-Eur. Sh) before y, In
the first syllable y is not found after any consonant in Latin,
except d, and that only in the older period, e.g. O. Lat. Diovem,
classical Jovem (from dyéu-, O. Ind. Dyats, Gk. Zevs, while
Lat. diés shows a bystem *diéu-, O. Ind. Diyais), though how
far this 7 (y) of Jovem (for yy-, I.-Eur. dy-, Gk. ¢-) differed at
any time in pronunciation from the j of jévencs (I.-Eur. y-: Gk. *-)
and from the j of jugum (I.-Eur. y-, Gk. ¢-), we cannot say.
All these have developed to the same sound in the Romance
languages, e.g. Ital. Giove, giovane, giogo (with the sound of our
j or -dge in ‘ judge’).

The Oscan orthography shows both ii and i for intervocalic
i (y), though the paucity of the remains of the language makes
it impossible to determine how far this was arbitrary or reflected
264 THE LATIN LANGUAGE. [Chap. IV.

the actual pronunciation, e.g. difviiai ** diviae,’ mefiai ‘ mediae’ ;
heriiad is 3 Sg. Pres. Subj. of the verb and heriam Ace. Sg. of
the noun derived from her-, ‘to wish’ (I.-Eur. GHER-), The
presence of the y-sound is indicated by the doubling of a pre-
ceding consonant after the accented vowel, e.g. medikkiat ‘med-
diciae’ Dat. Sg., Startins ‘ Statii’ Gen. Sg., Pettio-, &e. In
the Oscan dialect of Bantia this y-sound is merged in the pre-
ceding consonant, e.g. Bansae ‘ Bantiae,’ allo ‘alia’ (see von
Planta, i. p. 165). [Cf. Pel. @ (ch. ii. § 51), written s in Musesa,]

§ 64, L.-Eur. initial y. I.-Eur. *yiwnko- (0. Ind. yuvaké-, Gk. “Tdx-w8os,
0. Ir. dac, Welsh ieuanc, Gaul. Jovincillus, Goth. juggs for *juvunga-,
O. Engl. geong), Lat. jivencus, with its cognates jiivénis (O. Ind. yavan-),
Jiwenta (Goth. junda, Engl. youth), &.; I.-Eur. *yiigo-, (O. Ind. yugdém, Gk.
(vyév, Goth. juk, Engl. yoke, O. SI. igo for *yigo, *ytigo), Lat. juigum, from the
root yeug-, ‘to join’ (0. Ind. yuj-, Gk. (evyvups, Lith. jungiu, ‘I yoke’), Lat. jungo;
I.-Eur. *yiis- (O. Ind. yiiSa-, Gk. ¢é-un for *(vo-yn, leaven, Lith. juszé), Lat.
jis, broth, while Lat. jas, law, older jous, is I.-Eur. *yeus- (O. Ind. yos,
‘welfare’); I.-Eur. ya-, ‘to go,’ formed by adding the suffix 4 to i-, the weak
grade of the root ei-, ‘to go’ (O. Ind. ya-, ‘to go,’ Lith. jéti, ‘to ride,’ 0. SI].
jad) appearsin Lat. ja-nua, a door, Januarius. (On the tendency to give Latin
a after initial j the open e-sound, whence Vulg. Lat. *Jenuarius, see ch. ii. § 1.)
Similarly Lat. jicus with I.-Eur. yé- (cf. Lith. jitkas, ‘ridicule,’ with I.-Eur.
yo-, B. B. xviii. 255) (cf. § 59 above). On rien (Plaut. ; ef. lién) and rén, see
Prisc. i. 149.7 H. On Vulg. Lat. qu(i)étus, ch. ii. § 151, and cf. ch. iii. § 11
on par(i)és.

§ 65. I-Hur. y preceded by a consonant. (1) In the first syllable :—
This y has been dropped in the Latin derivatives from I.-Eur. roots like syi-,
‘to sew’ (O. Ind. syi-, Gk. xa-cotw, a compound with xar(d), Goth. siujan,
O. Engl. seowian, Lith. siiti, O. SI. Siti), Lat. suo; *Zhyés-, ‘ yesterday ’ (O. Ind.
hyds, Gk. x6és; cf. Goth. gistra-dagis), Lat. héri, with Adj. héster-nus; though
it is possible that there were sometimes I.-Eur. byforms without y (cf. O. Ind.
sti-tra-, ‘thread,’ also used, in the sense of ‘clue,’ for ritual and grammatical
text-books, the Sitras). So Lat. spuo (Gk. rrva, Lith. spiduju, O. Sl. pljuja, &e.).
There is a similar doubt about dy-. It may have lost the dental at a very early
period, for we have Joves (‘ Jovios’?) on the ancient Dvenos inscription, and
it is not possible to prove that the form Diovem had gone out of use by the
time that the form Jévem came in (cf. Diouem, C. I. L. i. 57, Iouet, i. 56, both on
old Praenestine mirrors). Diovem may quite well have come from a byform
*diyéw-, as O. Lat. siés from *siyés (O. Ind. siyas), a byform of *syés, (0. Ind.
syds) and have been discarded in course of time for the other ‘doublet’
¥*dyéw- Jovem.

(2) In other syllables :—Y after a consonant in other syllables than the
first became vocalic in Latin. Thus after p we find y becoming 7 in Greek in
verbs formed with the suffix -yo- (-iyo-), e.g. roa-rw for *run-yw, xahen-Tw for
*xaden-yw ; but in Latin these verbs appear with -pio, e. g. céipio, sdpio; after
n we find y producing epenthesis in Greek, e. g. Baiyw for *g¥m-y6, root gkem-,
but not in Latin, e.g. vénio. But -ghy-, as has been mentioned, became hy,
§§ 64-68.] REPRESENTATIVES OF I.-EUR. SOUNDS. Y, W. 265

then y or rather yy, e. g. dio, pronounced *ayyo, and often written aiio (ch, i,
§ 7), from agh-, ‘to say.’

§ 66. I.-Hur. y between vowels. The Nom. Pl. of I-stems shows I.-Eur,
-dyés, e.g. I.-Eur. *treyes from the stem tri-, ‘three’ (O. Ind. trayas), *Zhosteyes
from the stem ghosti-, ‘a stranger’ (0. SI. gostije), in Latin -2s for -*e-es, 6. g.
tres, hostés. In the words dhenus, aheneus the letter h indicates the hiatus caused
by the dropping of y of I.-Eur. *Ayés-, ‘metal,’ *ayes-no-, ‘made of metal’
(0. Ind. dyas-), like hin the Umbrian stahu for *sti-yo, ‘I stand’ (Lat. sto).
Like Lat. sté for *sta-yo (Lith. pa-sté-ju), are fleo for */lz-yo (O. SI. blé-ja), neo for
*nz-yo, &c., where the similar vowels a and o are blended into one sound, but
the dissimilar, ¢ and 0, remain in hiatus. Similarly Lat. formo for *forma-yo,
from forma (stem *forma-), and other 1st Conj. verbs from rst Decl. nouns,
like GK. rip, for rTipd-w, *ripayw, from tiwh, Dor. ripa (stem tipd-).

§ 67. Latin j. Any j which has been developed by the phonetic changes
of the language is treated in much the same way as I.-Eur. y. Thus the group
sj- (sy-) has been produced by the union under one accent of the two words
si audes. The j (y) is dropped, like I.-Eur. y in suo, in the form sddes, a form
which seems to have come into use in the period between Plautus and Terence.
For Plautus has the full si audes, e. g. Poen. 757 mitte 4d me, si audes, hédie
Adelphasiim tuam, but Terence the shortened form, e.g. Andr. 85 dic,
sodes. Internal j (y) after a consonant becomes vocalic in nunc-iam (3 syll.
Plaut.), and between two vowels disappears in bigae for *bi-jigae from bi- and
jugum. On abjécio, abjicio, abicio see ch. iii. § 18, ch. ii. § 48, ch. i. § 7.

§ 68. W. I.-Eur. initial w is represented in Latin by the
sound which we write v, and which the Romans wrote w (On the
spelling and on the pronunciation of the letter, see ch. ii. § 48), e.g.
I.-Eur. *wegho, ‘I carry’ (O. Ind. véha-mi, Gk. (F)éxos, Pamph.
Fexw, O. Ir. fén for *fegn, ‘a waggon,’ W. gwain, Goth. ga-viga,
Germ. be-wege), Lat. vé/6. The suffix wo- after a consonant was
perhaps, like the suffix yo- (§ 63), vocalic in early Latin, e.g. furvus
for early *fusuos (§ 148; cf. fus-cus), Minerua, a quadrisyllable in
Plautus (Bacch. 893); though Syncope, like that which reduced
Titius, &e. to *Tityus, reduced druum, lérua, miluus (all trisyllabic
in Plautus) to arvum, larva, milvus. Between vowels v remains,
e.g. dvis, vis; but in the unaccented syllable we find w for dv,
év, bv, e.g. dénuo for de novo (see ch. iii. § 24), and before the accent
v is often dropped, e.g. seorswm (and sorsum) for sévorsum, especially
between similar vowels, e.g. O. Lat. da(v)inus, li(v)abrum (see ch.
ii. § 53). After a consonant in the initial syllable, w is dropped
in pius for *pw-iyo- (?) from the same root as’ purus, but remains
(like y) after d, e.g. O. Lat. duonus, Duenos, duellum ; though at
the beginning of the literary period this dv- passed into J-, e. g.
bonus, béné, bellum. Side by side with dissyllabic duonus, duelium
266 THE LATIN LANGUAGE. (Chap. IV.

we find trisyllabic duonus, duellum, just as in I.-Eur. we have
duw- and dw- in the words for ‘two’ (O. Ind. duva, and dva,
Gk. vw and mais Lat. uo and 6%s), ‘dog’ (O. Ind. Suvén- sh
8van-, Gk. xdwv and Lith. szu). And side by side with I.-Eur.
*twoi, the Locative case of the and Pers. Pron. Sg. we have the
unaccented form *toi (O. Ind. tvé and té, Gk. cof and ro:), so that
it is often difficult to say when the I.-Eur. form has dropped w,
when it has w, and when it has uw (ew), and to determine when
the w has been dropped in the ‘I.-Eur. period’ and when in the
‘Latin period” An é has been turned into 6 through the influ-
ence of a preceding w in the group swé-, e.g. sdror (I.-Eur.
*swésor-), sdcer (I.-Eur. *swékiiro-), &. (see § 10). Before a con-
sonant (1, r) I.-Eur. w is dropped in Latin, e.g. radix for *wrad- (cf.
Goth. vaurts, Engl. wort). On the weakening of unaccented wé
to ti in I.-Eur,, e.g. *périit(i) (O. Ind. parut, Gk. 7épvor) from
the root wét- (Gk. Feros, a year, Lat. véus), and in Latin, e. g.
con-ciitio for *-quetio, see § 51, and on ferbui for fervui, ch. ii, § 52.

Latin v often represents I.-Eur. g#, gh® (see §§ 139, 143), both
initial, e.g. vénto for *g%myo from the root gtem- (Gk. Balve,
O. Ind. gam-, Goth. qima, Engl. come), and between vowels,
e.g. niv-em Acc. for *nightm (Gk. vido Ace., W. nyf) from the
root sneigh?- (e. g. O. Ir. snechta, Goth. snaivs, Lith. snaigyti, ‘ to
snow,’ O. Sl. snégii, ‘ snow’).

In Umbro-Osc. there are separate symbols in the native
alphabets for the consonantal and for the vocalic U-sound
(written in our transcription v and u respectively). The rules
for the use of uv and v in Umbr. have not yet been determined ;
we have arvia and (once) aruvia, vatuva (never *vatva), &c. (see
von Planta, i. p. 180).

§ 69, I.-Eur. initial w. The I.-Eur. root weid-, ‘to know, to see’ (O. Ind.
yéda, Arm, gitem, Gk. Feria, O.Ir. fiadaim, Goth vait, O. Engl. wat, Engl.
wot, O. SL. vidati, ‘to see,’ védéti, ‘to know’) appears in Lat. video ; the I.-Eur.
*weik- (0. Ind. vi8-, veké-, Gk. aes, Goth. veihs, 0.8]. visi), in Tativ vicus ;
the conjunction *wé, ‘or’ (O. Ind. va, Gk. 7-(F)é), Lat. -vé; the root wert-, ‘to
turn’ (O. Ind. vart-, Goth. vairpa, Engl. ‘ woe worth the day,’ W. gwerthyd,
‘aspindle,’ Lith. vartat, O. SL. vratiti Inf.) in Lat. verto. Similarly Lat. vieo, vitis
(O. Ind. vi-, Lith. vyti, O. SI. viti, Ir. féith, ‘woodbine,’ W. gwydd-fid) ; Lat.
vellus, with Wl for In (§ 78), (Goth. vulla, Engl. wool, Lith. vilnos, 0. SI. vitina) ;
Lat. vérus (O. Ir. fir, W. gwir, Goth. tuz-vérjan, ‘to doubt,’ 0. Engl. wer, ‘true,’
Germ. wahr; cf. Lith. véra, ‘faith,’ O. SL véra); Lat. vespa (O. Engl. weesp and
§§ 69—71.] REPRESENTATIVES OF I.-EUR. SOUNDS. Y, W. 267

weeps, O. Sl. vosa) ; Lat. virus (0. Ind. vitd-, Gk. ids for Fioos) ; Lat. vimo (O. Ind.
vam-, Gk. éuéw for *Feu-, Lith. vemit).

§ 70. L-Eur. w (and Latin v) between vowels. I.-Eur. *g%iwo-, ‘alive’
LO. Ind. jiva-, W. byw, O. Ir. biu, Goth. qius, 0. Engl. ewicu- (with -c- deve-
loped before Ww, Engl. quick, Lith. gyvas, 0. SI. Zivii] is Lat. vivus ; I-Eur.
*néwo-, ‘new’ (0. Ind. nava-, Gk. vel Fos) O. Sl. novit) is Lat. névus ; I.-Eur.
‘rgwra, nine’ (0. Ind. nava, Gk. é-véa, 0. Ir, noi, W. naw, Goth. niun) is Lat.
nivem ; I.-Eur. *yiwnko-, (O. Ind. yovadé,, Gk. “Ya«-160s, Gaul. Jovine-illus,
W. ieuanc) is Lat. jivencus. Similarly Lat. clavis (Gk. wAn(F)is) ; Lat. levir,
better laevir, brother-in-law (0. Ind. dévar-, Arm. taigr, Gk. dap from *5acFep-) ;
Lat. aevwm (Gk. ai(f)év, Goth. aivs, O. Engl. ®) ; Lat. laevus (Gk. Aau(F)és).

The question of the change of the ending -vos to -vus and -us is a difficult
one. The most natural explanation of the change of deivos to deus (the form
in ordinary use as early as Plautus) is that given in § 33 (through *de(é)us),
which takes for granted that -vos became -vus when -os became -us (§ 17) [ef.
Flaus on a coin of 200-150 B.c (C.L. L, i. 277), Gnaeus, boum, coum and cohum
(§ 19)], and seems to conflict with the fact that the spelling -vo- (with -quo-,
-guo-) is retained to the end of the Republic. But it is by no means clear that
the spelling vo did not represent the sound vu, the spelling vo being preferred
to vv, because this last might be confused with the sound @ ‘ch. i. § g) or wv.
Velius Longus (first cent. a.p.) expressly asserts this (58. 4 K.): a plerisque
superiorum ‘primitivus’ et ‘adoptivus’ et ‘nominativus’ per v et o scripta
sunt, scilicet quia sciebant vocales inter se ita confundi non posse ut unam
syllabam [non] faciant, apparetque eos hoc genus nominum aliter scripsisse,
aliter enuntiasse. Nam cum per o scriberent, per u tamen enuntiabant.
The tendency to re-insert the v from other cases and cognate words, divo,
divos, diva, &c., would interfere from time to time with the natural develop-
ment of the sound. (Cf. § 67 on abjecio, abjicio and abicio.) On Republican
inserr. we find v in rvenva (C.J. L. i. 1202), &e., where vv (uv) is the ortho-
graphy of the Augustan age, but that this always represents the pronunciation
(as in Pacu(v)ius also written Paquius; Vésu(v)ius, ef. Galen x. 364) is unlikely.
[For examples of the spellings vo, vv, V (vo, vu, uv, u), and for a fuller discussion
of the treatment of intervocalic v in Latin, see Solmsen, Stud. Lauty. sect. iii.]

The process of Syncope affected at various periods intervocalic v in different
ways. Under the early Accent Law *dvi-spex was reduced to auspex (cf.
au-ceps), *vivita (Lith. gywata; cf. Gk. Biory) to vita, &c.; under the Paenul-
tima Law *dvidére became audere, &c.; in the period of the Early Literature
aevitas became aetas, tividus became tidus, praevides became praedes, &c.; while
four-syllabled words with the first, second, and third syllables short, which
were in the second cent. B.c. still accented on the first syllable (ch. iii. § 8),
may have suffered syncope within the literary period or at a much earlier
time, e.g. Aulius from Auilios (C.J. L. i. 83, Praeneste). (On the loss of -v-
through Syncope see ch. iii. § 16. 9.)

§ 71. I.-Eur. w after a consonant. (1) In the first syllable :—Latin sérius
(Goth. svérs, ‘honourable’ ; Germ. schwer, ‘ heavy’) is from the root swer-, ‘to
weigh’ (Lith. svefti) ; for Latin si we have in Oscan svai, in Umbr. sve;
Latin sidus is connected with Lith. svidéti, ‘to shine’; Lat. suf-fio, to fumigate,
with Gk. @dw, from I.-Eur. *dhw-iyé. But swa- is Lat. sud- in suavis (a trisyl-
lable in Vulg. Lat., cf. Ital. soave; Servius ad den. i. 357 says that many
persons in his day made suddet a trisyllable). Lat. dis-, apart, asunder, seems
268 THE LATIN LANGUAGE. (Chap. IV.

to be I.-Eur. *dis-, a byform of *dwis (Lat.° bis, O. Lat. duis), as I.-Eur. *toi
*(unaccented) was a byform of *twoi (accented), though some regard Lat. dis
as the direct descendant of I.-Eur. *dwis, and O. Lat. duis (class. bis) as the
descendant of I.-Eur. *dtiwis. Savium, a kiss, seems to represent s(w)dvium.
On sos, &c. for suos, &c., see ch. vii. § 12.

(2) In other syllables :—Internal -dw- becomes -v- in suavis for *suad-vis ;
*derviosus, later derbiosus (see ch. ii. § 52), for *der-dwi-oso- (O. Ind. dardi-; ef.
Engl. tetter). Viduus does not show I.-Eur. -dhw-, but -dhéw-, for it repre-
sents I.-Eur. *widhéwo- (0. Ind. vidhdva-, Gk. 4-(F)i@e(F)os, O. Sl. vidova,
‘widow.) Postconsonantal wu, later v, is in Vulgar and Late Latin dropped
before the accent in Jan(v)arius, Febr(v)artus, batt(v)ére, cons(v)ére, quatt(v)ordecim,
contin(v)ari, whence Ital. Gennajo, Febbrajo, quattordici, &c. (see ch. ii. § 54),
like v between vowels before the accent (see above). The suffix -wo- remains
dissyllabic in the classical form of words like mortuus (O.Sl. mritvi). So
quattuor (O.Ind. catvaras, Lith. ketveri, O. SI. tetvero) ; tenuis (O. Ind. tanvi
F.), génua (Zend. zanva, Lesb. Gk. yévva for *yovFa), though tenvia, genva are
found in classical poetry, and in the first cent. a.p. tenwis is declared to
have wavered between a dissyllable and a trisyllable (Caesellius ap. Cassiod.
vii. 205 K.). Late-Lat. mortvus, &c. (ch. ii. § 48) became mortus, &c. (Ital.
morto), as rivus, &c., rius.

dw-,duw-. Duellius (Duill- ?), consul of 260 B.c., was the first to change his
name to Bellius (Cic. Orat. lv. 153; ef. Quint. i. 4. 15); duonus seems to be a
trisyllable in the Saturnian fragments (cf. ch. ii. § 141”), viz. C.I. L. i. 32:
dionéro éptumo fuise uiro, and Naevius(?) ap. Fest. 532. 22 Th.: simul
diona eérum pértant ad nduis, duellum to be a disyllable (Aem. Lepidus’
inser. ap. Caes. Bass. 265. 25 K.: duéllo maégno diriméndo, régibus subigéndis),
as it always is in Plautus (e. g. Amph. 189; extincto duello méxumo), whereas
Ennius has (A, 168M.): pars occidit illa duellis, a scansion imitated by later
poets. (On O. Lat. duis, dui- for bis, bi- see ch. vi. § 59; on du- in glosses,
see Liéwe, Prodr. p. 363, and add Duellona, C. G. L. ii. 56. 34.)

§ 72, L-Bur. w before a consonant. Lat. liquo, liquor stand for *vlig- (O. Ir.
fliuch, ‘wet’) ; Lat. rémens may be a Pres. Part. of I.-Eur. wrép- (Gk. fémw, to
fall; ef. Lith. virpéti, ‘to tremble’) (but see ch. viii. § 18).

§ 73. M, N. In Sanserit various kinds of nasals are dis-
tinguished in writing. The palatal n (made palatal by the
preceding j) of yajfid-, ‘worship’ (Gk. dyvds, holy) is written
differently from the cerebral n (made cerebral by the preceding
r) of mrnami, ‘I crush, annihilate’ (Gk. pdpvaua: Mid.). And in
Greek, owing to the fact that y before v of ytyvopm, &c. had
come to take the sound of the Agma (see ch. ii. § 63), *yuavopae
(cf. Dor. yivoua), y was used to express the guttural nasal,
e.g. dyxupa, éyyds, though in inscriptions we often find dvxvpa,
éryds. But in Latin we have only the symbols m, ~ (see ch. ii.
§ 63, and on their pronunciation, ch. ii.§ 61). Before a consonant,
though sometimes omitted on inscriptions, they were not dropped
in correct speech. Ligii/a, spoon (cf. O. Ir. liag, W. llwy), is not the
§§ 72, 73.] REPRESENTATIVES OF I.-EUR. SOUNDS. M, N. 269

same word as Hingiila, strap; *népe in Plautus should be corrected
to nemp(e). (For other instances, see Skutsch, Vorsch. i. § 2.)

M. [.-Eur. m is Latin m, whether initial, e. g. I.-Eur. *mater-
(O. Ind. matér-, Arm. mair, Dor. Gk. parnp, O. Ind. mathir,
O. H. G. muoter, O. Engl. médor, Lith. mote, ‘ wife,’ O. Sl. mati),
Lat. mater, or internal, e.g. from I.-Eur. root wem- (O. Ind,
vam-, Gk. (F)euéw, Lith. vematai, Pl.), Lat. vdmo; I.-Eur.
*termen- (O. Ind. térman-, Gk. répuwy), Lat. termd, terminus ;
I.-Eur. rump-, ‘to break’ (O. Ind. lumpami), Lat. rwmpo, or final,
e.g. I.-Eur. -m of the Acc. Sg., as in I.-Eur. *tdm (O. Ind. tam,
Gk. rév, Goth. pan-a, Lith. t@, O.S1. tii), O. Lat. is-tom, class.
Lat. 2s-tum, Before y (2), ¢, 8, d, c we find ~ in gudniam (for
quom jam, and originally used in a temporal sense ‘ when now,’
‘now that,’ ch. x. § 13), quan-sei, C. I. L. i. 200. 27 (on quiisi, see
ch. x. § 11), altrin-séeus, centum from I.-Eur. *kmtom (Lith.
szimtas), septen-tridnes (lit. ‘the seven oxen,’ according to Aelius
Stilo and Varro, who regarded the -ér7o not as a mere termina-
tion, but as a rustic word for an ox, Gell. ii. 21), septen-décim, nunc
(from um, now, which survives in etiamnum in a temporal
sense, but usually has the interrogative sense of ‘now,’ as in
our ‘now is this the case?’, ‘now is that true?’, where ‘now’
has something of the dubitative significance of Lat. num), princeps
for *prim(i)-ceps ; though a traditional spelling is often used,
e.g. numcubi, quamtus, O. Lat. guamde, than, &c. (see ch. ii. § 61).
A p is inserted to facilitate the pronunciation of these groups in
sumptus, sumpsi for sumtus, sumsi, exemplum for *enemlum, &c.
I.-Eur. mr is Latin dr in hibernus from *hibrino- (§ 13), (Gk.
xetmepwds), but initial mr probably became in Latin fr. On the
loss of final -wm by elision in sed from older sedum, nihil from
nihilum, &e., see ch. iii. § 52.

Latin m represents an original before a labial, &c., eg.
impello for inpello, an original labial before x, e.g. scamnum from
the root skabh-, ‘to support, with Diminutive scabelluwm. It is
sometimes lost in a consonant-group, e. g. forceps for *form(i)-ceps,
from formus, warm.

J.-Eur. m is Umbro-Ose. m, e.g. Umbr. matrer, Osc. maatreis
‘matris. Final -m is usually dropped in Umbrian (see von

Planta, i. pp. 301, 570).
270 THE LATIN LANGUAGE. (Chap. IV.

§ 74. I.-Eur.m; other examples. I.-Eur. *médhyo- (0. Ind, madhya-, Gk.
Héa(o)os, Ir. medon, ‘the middle,’ Goth. midjis, O. Sl. mezda, ‘the middle’) is
Lat. médius ; I.-Eur. root men-, ‘to think’ (O. Ind. man-, ménman-, ‘thought,’
O. Ir. men-me, ‘thought,’ 0. H. G. minna, ‘remembrance,’ Lith. ment, ‘T
remember’) appears in Lat. mémini for *me-men-i, mens, &c. Similarly Lat.
mé (O. Ind. mam, Gk. we, Ir, mi, Goth. mi-k) ; Lat. semi- (0. Ind. sami-, Gk.
Hyt-) ; Lat. himo (Goth. guma, Engl. groom, bride-groom) ; Lat. cum, com-, with
(O. Ir. com-); Lat. mors, mortis (O. Ind. mrti-, Lith. mirtis, O. Sl. si-mriti ; cf.
Goth. maurpr, ‘murder’); Lat. fimus (O. Ind. dhimé-, Lith. dimai, Pl., 0. Sl.
dymit).

§ 75. nfor m. Lat. con-tra from com ; quon-dam from quom ; vénun-do, class.
vendo, from véno-, sale (cf. Gk. dvos), lit. ‘to put or make sale,’ like vénum eo,
class. véneo, lit. ‘to go to sale’ (ef. pessum do, to ruin, and pessum eo, to be
ruined). On the spellings damdum, damdam (C. I. L. i. 206. 17. 49,) &e.
see ch. ii. § 64.

§76. L-Bur.ms. In Lat. ténzbrae (Plur., like O. Ind. tamasi, Russ. sumerki)
from the I.-Eur. root tem- (Lith. témti, ‘to grow dark’, Ir. tem, ‘dark,’ temel,
‘darkness’) the n seems to be due to the influence of s, for O. Ind. témisra,
0O.H.G. dinstar, Germ. finster, show that *iemésra- or *temsra- was the old form
of the stem in Latin. The m remains in témére Loc.,‘in the dark’ (ch. ix. § 5)
(ef. O. Ind. timas- from I.-Eur. *témes-, ‘darkness,’ and Germ. Dammerung).
Another example of I.-Eur. -ms- (mes-?) is Lat. tmérus (Umbr. onso-,
O. Ind. 4sa-, Arm. us, Gk. duos, Goth. ams), and perhaps Lat. nimérus,
Numerius (ef. Numisius, O. Lat. Numasio-,Oscan Niumsio-) ; and another example
of -msr- is Lat. membrum for *mems-ro- (cf. Goth. mimz, ‘ flesh,’ O. SI. meso ; also
Gk. pnpés for pnopo-, the thigh, O. Ind. masé-, ‘ flesh,’ Arm. mis.) The fewness
and the contrariety of these instances make it difficult to decide how I.-Eur.
-ms- was treated in Latin. Latin -ms- became ns, e. g. con-sentio, con-silio, &e.,
from com-, where the m might be regarded as the final letter of a separate
word, but mps, e. g. sumpsi, dempsi, where the m could not be so regarded.
(On the spellings sumpsi, sumsi, &c., hiems, hiemps, see Brambach, Lat. Orth.
p. 248 ; the Roman grammarians approve of sumpsi, &c., but not of hiemps.)

§77. I.-Eur. mr, ml are equally difficult to trace in Latin. They are
represented by 8p, BA in Greek when initial, e.g. Bpords, BAwoxw, BAirTw (from
Herr), by uBp, wBA when medial, e.g. duBporos, wéuBrawxa, In Irish we find
initial mr- to be an early spelling, which was changed later to br-, e.g.
mraich, ‘malt,’ later braich, from a stem *mraci-, and similarly ml-, later bl-,
e.g. mlicht, blicht, ‘milk.’ The Latin fraces, olivelees, seems to be the same as
this Irish word mraich, in which case jfr- will be the Latin equivalent of
I.-Eur. mr- (for other examples, see Osthoff, Morph. Unters. v. 85), and the b of
hibernus, tuber, a swelling, a truffle (from tumeo, to swell), will be like the b of
ruber, &e. (§ 114), for which f is found in other dialects, e.g. Umbr. rufro-.
[Ital. tartufo, ‘truffle,’ lit. ‘earth-mushroom,’ from terra and tiber, « name
borrowed by the Germans in the eighteenth cent. for the potato, Kartoffel,
shows that the dialectal form of taber had f (ch. ii. § 83)]. The long vowel in
hibernus, tuber might then be explained like the long i of injero, infringo, &e.,
(ch. ii. § 144), and the original forms would be *himfrino-, timfro-. On the
other hand the analogy of other languages and the connexion of m and b in
§§ 74-78.] REPRESENTATIVES OF I,-EUR. SOUNDS, M, N. 271

Latin would make us expect to find Latin br as the equivalent of I.-Eur.
initial mr-. A further difficulty is caused by géner, a word which it is hard
to dissociate from Gk. youBpés, with n(e)r for -mr-. I.-Eur. medial -ml-
appears in exemplum for *ex-em-lo-, lit. ‘something taken out,’ with a euphonic
P inserted, while a vowel seems to have intervened between m and 1 in trému-
lus, timulus, &e. (ch. v. § 21).

§ 78. N. I.-Eur. n is in Latin ~, whether (1) initial, e.g. I-Eur.
*néwo-, *newio-, ‘new’ (O. Ind. néva-, ndévya-, Arm. nor, Gk. véos,
O. Ir, nie, Gaul. Novios, W. newydd from *noviyo-, Goth, niujis,
O. Engl. néowe, Lith. natijas, O. Sl]. novit), Lat. ndvus, Novius, or
(2) internal; e.g. the I.-Eur. root sén-, ‘ old’ (O. Ind. séna-, Arm.
hin, Gk. &m xal véa, O. Ir, sen, W. hen, Goth. sineigs, sinista
Superl., Lith. sénas) appears in Lat. sénex, senior, the root angh-,
‘to choke’ (O. Ind. ¢has-, ‘need,’ Arm. anjuk, ‘ narrow,’ Gk. ayo,
O. Ir. cum-ung, ‘ narrow,’ Goth. aggvus, ‘narrow,’ Engl. anger,
Lith. afiksztas, ‘narrow, O. Sl. aziikti) in Lat. ango, angor,
angustus, or (3) final; e.g. the I.-Eur. preposition *8n (Gk. év, O. Ir.
in, Goth. in, Lith. {) Lat. 7m, O. Lat. ex. Before a labial we find
m, e.g. impello, immito. But nm seems to have become in Latin
rm, if carmen stands for *can-men, germen for *gen-men, just as the
nm which arose at a later time from the syncope of 2 in Gnima has
become in some Romance languages 7m (e.g. Prov. anma, alma,
and arma, O. Fr. anme, alme, and arme, Catal. arma and alma,
Sicil. arma, Milanese armella). Before / it was assimilated, e. g.
corolla for *cordn-la, himiilius for *homin-lus, iligo, wlex, malluviae,
water for washing the hands, from *man-luviae (so before 7 in
irritus, &c.), and also after /, e.g. collis for *col-ni-s (Lith. kétnas ;
cf. Gk. xodwvds), vellus (Lith, vilna, O. Sl. vliina), as in Greek we
have AA for Av in éAdds, a fawn (Lith. élnis, O. S$]. jeleni), &c.
In ulna, &e. a vowel originally came between / and wn (cf. Gk.
édévy). On the pronunciation of xs as ss, or s, with lengthening
of the preceding vowel, e. g. vicessimus and vicésimus from vicensi-
mus, see ch. it. § 64. So with nf (2.).

J.-Eur. n is in Umbro-Ose. n, e.g. Umbr. nerus, Dat. PL,
Ose. nerum, Gen. Pl., from the same root as Gk. dup (cf. Nero).
Before mutes and spirants we find n often dropped, especially in
O. Umbr., e. g. iveka ‘juvencas’ (N. Umbr. ivenga), but in Umbr.
of all periods before s, e.g. aseriatu and aseriato, anseriato; in
Oscan before a mute in an unaccented final syllable, e.g. -et for
272 THE LATIN LANGUAGE. [Chap. IV.

-ent 3rd Plur., iak in Ace. Sg. for fank (Lat. eam, with the particle
-ce, ch, vii. § 15) (see von Planta, i. p. 301).

§ 79. I.-Hur. n; other examples. I.-Eur. *néwn, ‘nine’ (0. Ind. nava, Gk.
év-véa, O. Ir. noi, W. naw, Goth. niun, O. Engl. nigon, Lith. dewyni, O. SI.
deveti with d- by analogy of the words for ten, just as Vulg. Engl. ‘thrup-
pence’ takes its u from ‘tuppence,’ ‘twopence.’) Lat. névem [for noven (§ 81) :
a similar substitution of -m for -n has been found in Subj. feram, (O. Ind.
bharani)]; I.-Eur. *né, *nou, ‘we’ (0. Ind. nau, Gk. vw, O. Sl. na; ef. O. Ir. ni),
Lat. né-s; I.-Eur. *nas-, ‘the nose’ (0. Ind. nasa Du., Lith. nésis; cf. O. Engl.
nosu), Lat. naris for *nas-is; I.-Eur. *nau-, ‘ship’ (O. Ind. nau-, Arm. nav,
Gk. vais, O. Ir. nau, W. noe, ‘a dish, vessel’), Lat. navis; I.-Eur. *népot-,
‘grandson’ (O. Ind. napat-, M. Ir. niae, niath Gen., W. nai), Lat. népis, Gen.
nepotis ; I.-Hur. sneigh?-, ‘to snow’ (Zend. snaezaiti, Gk. vipa Acc., dydvvipos for
*dya-ovipos, O. Ir. snecht, W. nyf, Goth. snaivs, Lith. snaigo, Vb., O. SI. sn&git),
Lat. nix, ninguit ; I.-Eur. *8n6-to-, ‘known’ (0. Ind. jilatd-, Gk. yvords, O. Ir.
gnath, ‘accustomed’), Lat. ndtus, O. Lat. gndtus; the I.-Eur. root bhendh-,
‘to bind’ (O. Ind. bindhana-, ‘binding,’ bandhu-, ‘a relation, Gk. mevOepés,
stepfather, me(oya, a rope, for *revé-cpa, Goth. bindan, ‘to bind’) survives in the
religious term, of-fendices, the knots with which the priest’s apex was tied on,
an old word wrongly connected by some Roman antiquarians with offendo
[Festus 244. 2 Th. offendices ait esse Titius nodos, quibus apex retineatur et
remittatur. At Veranius coriola existimat, quae sint in loris apicis, quibus
apex retineatur et remittatur, quae ab offendendo dicantur. nam quom ad
mentum perventum sit, offendit mentum. Paulus Diaconus, the epitomator
of Festus, has been misled by the corruption offendimentum for offendit mentum
(Paul. 245. 1 Th.) : offendices dicebant ligaturae nodos, quibus apex retineba-
tur. Id cum pervenisset ad mentum, dicebatur ‘ offendimentum ;’ whence
the ‘ghost-word’ offendimentum has come into our Latin dictionaries ; ef.
Cc. G. L. iv. 132. 3 offendix nodus proprius quo apex flaminum retinetur (MS.
restinguitur) et remittitur]; I.-Eur. *ghans-, ‘a goose’ (O. Ind. has4-, Lith.
xasis, O. Engl. gis; cf. Gk. yqv: in Irish the word means a swan, géis from
stem *gési-, originally *ghansi-), Lat. anser, properly hanser.

The instances ofa nasal (n or m) being dropped before a consonant in Latin
without ‘compensation’ are illusory : ligiila, a spoon, from lingo, to lick, is
a different word from lingiila, a strap, lit. ‘a little tongue,’ though the two
were sometimes confused , as we learn from Martial’s lines on a silver spoon
(xiv. 120):

quamvis me ligulam dicant equitesque patresque,
dicor ab indoctis ‘lingula’ grammaticis,

(see Friedlander ad loc.) ; lanterna (from Gk. Aapzarnp) is the correct spelling
(see Georges, Lex. Wortf. s.v.), laterna being probably a corruption due to
a fanciful connexion of the word with ldtus ; typdnum is Gk. témavoy, tympa-
num, Gk, tépmavov, &e.

§ 80. nm. How far the theory that *canmen became carmen is supported
by the comparison of cancer with O. Ind. karkara-, ‘hard,’ Gk. xapsivos, and of
crépus-culum with Gk. xvépas is doubtful. For the change of r to n in the first
of these examples is due to Dissimilation, like the change of ! to r in the first
syllable of caerwleus (§ 84), and the second example, if correct, would exhibit
§§ 79-81.] REPRESENTATIVES OF I.-EUR. SOUNDS. M,N. 273

the same change of initial en- to er- for facility of pronunciation, as is seen
in dialects of Gaelic, e.g. cnu, ‘a nut’ (pronounced cru with nasalized vowel),
enoc, ‘a hill’ (pronounced similarly cr-; in Manx, cronk), gnath (I -Eur.
*8ndto-) (pronounced gra with nasalized 4). In Latin, however, I.-Eur. cn-
and gn- seem to have become n- (§ 119), so that crepus-culum, creper, ‘dark,’ if
they represent an earlier en-, must be dialectal. Varro makes them Sabine
words (LZ. L. vi.5 secundum hoe dicitur ‘crepusculum’ a crepero. id vocabu-
lum sumpserunt a Sabinis, unde veniunt ‘ Crepusci’ nominati Amiterno, qui
eo tempore erant nati, ut ‘Lucii’ prima luce in Reatino; ‘crepusculum’
significat dubium ; ab eo res dictae dubiae ‘ creperae,’ quod crepusculum dies
etiam nune sit an jam nox multis dubium). The Probi Appendix censures
the pronunciation ‘parcarpus’ instead of pancarpus (Gk. mty-capros) (197.
32 K.; the reading is doubtful, and ‘prancarpus’ has been proposed).

The evidence for the change of n to 7 in carmen, germen (cf § gt on n for r)
is thus not very strong, and a good deal may be said for the view which
refers germen to *ges-i-men (from géro for *geso, cf. ges-si, ges-tum) and gemma, to
¥*gen-ma. Carmen may be connected with O. Ind. kari-, ‘a singer,’ or with
cdéiro (in Lat. a portion of meat, flesh, but in Osc. any portion, e.g. maimas car-
neis senateis tanginud ‘maximae partis senatus sententia’), as O. Ir. drecht
means (1) part, portion, (2) song. (O. Ind. sis-man- Neut., ‘praise,’ a word
which occurs only once in the Rig Veda, comes from the O. Ind. root Sas-,
‘to praise,’ which is usually connected with Lat. censeo). In Compounds
with in, con-, &c. -nm- became mm, e. g. im-mitto, com-mitto, so that gemma seems
a natural development of gen-ma.

The connexion of n with 7 seen in Provengal alma for Lat. an(i)ma, &c.
receives an equally doubtful support from the mispronunciation censured in
the Probi Appendix (197. 24 K.) ‘cuntellum’ for cultellum, and in the curious
form of the word neptis mentioned in glossaries, ‘leptis’ (Léwe, Prodr. p. 340) ;
nuscicio explained as ‘ caecitudo nocturna’ [Fest. 180. 23 Th. ; ef. nusciciosus,
ib., nusciosus in glossaries (Léwe, Prodr. p. 17 ‘qui plus vespere videt ’)] seems
to be a popular adaptation of luscitio (-cio ?) (Paul. Fest. 86. 21 Th.), derived
from luscus, to the word nox. Conucella, the Diminutive of célus (C. @. L. iii.
322. 9) (ef. Ital. conocchia, ‘a distaff,’ from Lat. *conucwa, Fr. quenouille, &c.),
may be influenced by cénus. N appears as 1 in dialectal Greek, e. g. Adpvaé for
vapvag. (K. Z, xxxiii. 226.) On the affinity of the n-, l-, and r-sounds, see
ch. ii. §§ 61, 96, 105, 106. (Cf. Germ. Himmel, Goth. himins, Engl. heaven ?).

§ 81. The M- and N-Sonants. For the sounds to which ém,
én (mé, né) are reduced in the unaccented syllable of such a word
as I.-Eur. *tnt6- (O. Ind. taté-, Gk. rards, Lat. tentws) from the
root ten-, ‘to stretch,’ some write, m, n, others am, en (ma, na),
the a indicating an obscure vowel sound, or else *m, ¢n (m*, n°).
Jn Latin these sounds became &m, én, e.g. décém (Gk. déxa, Goth.
taihun), tentus, the @ being subject to all the changes of 2 for
I.-Eur. & (e.g. incertus, Gk. dxpiros, undécum, decumus, decimus,
&e.); in Greek they became a before a consonant [but ap, av before
vocalic i (y)]; in Teutonic um, un, and so on; so that it is only

T
274 THE LATIN LANGUAGE. [Chap. IV.

by reference to another I.-Eur. language that we can tell whether
Lat. ém, én are I.-Eur. ém, én, or I.-Eur.m,n. We occasionally
find mi, nd in Latin words from roots in mé, né, which are
probably to be explained similarly by the weakening influence of
the accent, e.g. nactus from the root nek-, nenk-, ‘to obtain’
(O. Ind. nag-, Gk. é-veyx-eiv, Mid. Ir. co-emnacar, ‘ potui,’ Lith.
nészti,‘ to carry, O.Sl.nesti). (See Osthoff, Morph. Unt. v. Pref.,
and cf. above, § 61.) Whether Gk. vd, Lat. xd of Gk. vivoa (Lat.
dnas) for *var-ya, a duck, Lat. gndtus, and Lat. an, O. Ind. a of
Lat. antae, O. Ind. jata-, ‘ born, are rightly called the long sonant
nasals from roots like gen-, &. is a point which has not yet been
settled (see § 1). J.-Eur. initial m, n have been also referred to
a weakening of an original initial am-, an-, e.g. *mbhi beside
*ambhi (Gk. duu), like *po beside *apo (Gk. ine) (ah ix. § 12).
I.-Eur. m, n are in Umbro-Ose. treated as in Latin, e.g.
Umbr. desen- ‘decem,’ ivenga ‘juvencas, Ose. tristaamentud
‘testamento’ Ab]. Sg. Why ain- should appear for Lat. 7n- (en-),
in Umbr. antakres ‘integris,’ Ose. amprufid ‘improbe,’ Umbro-
Osc. anter ‘inter, is not quite clear (see von Planta, 1. p. 315).

§ 82. Other examples of the Nasat Sonants. I.-Eur. -mn of the Nom.
Sg. Neut. of men-stems (e.g. Gk. jua) is Lat. -men, e. g. stmen; the weak
form of the root ten-, ‘to stretch,’ appears also in I.-Eur. *tend-, ‘thin’
(O. Ind. tant-; ef. Gk. ravu-, stretched, of ravv-yAwaooos, tavi-nrepos, Tavi-
memos), Lat. tenuis from the Fem. form (0. Ind. tanvi) (ch. v. § 47); I.-Eur.
*Kmtém, ‘hundred’ (0. Ind. Fatdm, Gk. é-cardv, O. Ir. cét, W. cant, Goth.
hund, Engl. hund-red, lit. ‘too-number,’ Lith. szimtas, O. Sl. stito) is Lat.
centum ; I.-Eur. *g3mti-, the Verba] Noun from the root gtem-, ‘to go’ (O. Ind.
gati-, Gk. Bdois, Goth. ga-qumps, ‘assembly,’ Germ. Her-kunft, &¢.), Lat. in-
venti-o ; I.-Eur. *sm-, the weak grade of the root sem-, ‘ one’ (O. Ind. sa-kft,
‘once,’ Gk. d-mag, a-7Ados), Lat. simplex ; I.-Eur. *mnt6-, mnti-, the P. P. P. and
Verbal Noun of men-, ‘ to think’ (O. Ind. matd-, mati- Gk. atré-paros, O. Ir.
dér-met, ‘ forgetting,’ Goth. ga-munds, ‘ remembrance,’ 0. Engl. gemynd, Eng].
mind, Lith. mifitas, 0. Sl. meti, pa-meti, ‘memory’) appear in Lat. com-mentus,
mens, Gen. mentis. Similarly Lat. ensis (O. Ind. asi-) ; Lat. ingens, lit. ‘un-
known,’ ‘uncouth’ (O. Eng. un-cid) from the root gen-, ‘to know’; Lat.
jtivencus (see § 64); Lat. lingua, older dingua (Goth. tuggd, O. Engl. tunge.)
(On nivem for *noven, see § 79.) The late retention of the m-, n-sounds (as of
the ]-, r-sounds, § 92) has been inferred from patrém (beside patris), nomén (not
-in) (but see p. 186).

§ 83. Other examples of am, an, ma, na. Lat. gndrus from root gen-, ‘ to
know’ (0. Ind. jandmi) ; anda, pillars at door of a templé (O. Ind. ata-, Arm.
dr-and) ; Lat. janitrices, sisters-in-law (O. Ind. yatar-, cf. Gk. eivdrepes, O. SI.
jetry). Lat. amare is explained by some as derived from émo, ‘I take,’ with
§§ 82-85.] REPRESENTATIVES OF I.-EUR. SOUNDS. L, R. 275

a reduction of the ém-, as dicare shows reduction of the deic- of dico (O. Lat.
dewco). (Cf. § 94 on al, ar, and § 3 on péteo (Gk. mer-), dper.

§ 84, L, R. These two sounds are often interchanged in
Latin by ‘ Dissimilation’ of /-/, e.g. caertileus from caelum,
populd-ris and austra-lis (ch. ii. § 101). The same thing is found
in other languages, e.g. O. H. G. turtula-tiba, our ‘ turtle-dove’
from Lat. turtur (Gk. xefadapyia and xepadadyia) and perhaps
occurred in what is called the I.-Eur. period, e.g. I.-Eur. g*er-
g*el-in the reduplicated root (Lat. gur-gu/-i0, O. H. G. querechela,
Lith. gargaltiju). On the parasitic vowel often found between
a consonant and / (7), see ch. il. § 103, and on the avoidance of
rr, e.g. praest(r)igiae (like Gk. dpt-paxros for dpv-ppaxtos) 78.

§ 85. L. I-Hur. 1 is Lat. /, eg. the I.-Eur. root leiq?, ‘ to
leave’ (O. Ind. ric-, Arm. lk‘anem, Gk. Aefaw, O. Ir. lécim, Goth.
leihvan, ‘ to lend,’ O. Engl. léon, Lith. lékt, ‘I leave,’ O. SI. otti-
lékti, * remainder’) appears in Latin dinguo, Pft. Zégui; the root
klei-, ‘to lean’ (O. Ind. éri-, Gk. «Atvw, O. Ir. cloen, ‘ awry, Goth.
hlains, ‘a hill,’ Lith. szlaitas, szléti, ‘to lean’), appears in Lat.
clivus, ac-clinis ; suffixal -lo, -la, as in derivatives from the root
dhé-, ‘to suck’ (O. Ind. dharé-, Arm. dal, ‘ beestings, Gk. @nAq,
OjAvs, O. Tr. dél, ‘teat,’ Gael. deal, ‘leech,’ O. H. G. tila, ‘teat,’ Lith.
délé, ‘leech’), in Lat. fellare for fé-lare, &e. (ch. ii. § 130).

We find @/ in Latin for original ld, e.g. per-cello (cf. Gk.
xAadapés, brittle, Lat. clades), In, e. g. collis (Lith. kalnas; cf. Gk.
codwvds) (ulna had I.-Hur, -lén-; ef. Gk. od€vn), 1s, e.g. collum,
O. Lat. collus (O. Engl. heals, Germ. Hals M.), velle for *vel-se (cf.
es-se), rl, e.g. stélia for *stér-la, dgéllus for *ager-lus. So dl in
compound verbs, e.g. alligo for ad-ligo. But usually a parasitic
vowel was inserted to facilitate pronunciation, when / was pre-
ceded by a consonant, e. g. pidculum for pia-clum (see ch. ii. § 154).

J.-Eur. d became / in the Sabine dialect (see J. F. ii. 157), and
this form was sanctioned in a few words in Latin, e. e. lingua,
older dingua, from I.-Eur. *dngh*a- (Goth. tuggd) (§ 111). On
fucul, &e. for *faelt, fac(ilé, &e., see § 13; on leptis, a byform
of neptis, ch. ti. § 106; on the change of ¢ to d (#) before Latin
2 [except before 7 followed by e, 7 (y) or before the group J],
§ 10, and on the various pronunciations of Latin / when alone,
when double, and when preceding a consonant, ch. ii, § 96.

T2
276 THE LATIN LANGUAGE. [Chap. IV.

I.-Eur. 1 is Osean 1, e. g. ligatiis ‘légatis. Similarly we have
Umbr. pilener ‘ plenis,’ veltu ‘vulto’ Imperat.; but initial 1 does
not occur in the Eugubine Tables, the chief record of the lan-
guage, and it is not impossible that at the beginning of a word
I.-Eur. 1 is Umbrian v in words like vapef (Lat. /épides? Ace. Pl.),
Vuvois (Lat. Licius? cf. Osc. Luvkis), vutu (Lat. /évito ?); between
vowels | became (like d) i (rs) in kat¥etu, carsitu ‘let him call’
(Gk. ckadéw, Lat. caldre), famerias ‘familiae,’ &c.; before t it was
dropped in muta ‘ multa’ (‘a fine,’ Osc. multa-), &c., while the
Umbrian name Voisieno- (Lat. Volsiéno-)! suggests that it became
i before s (see von Planta, i. p. 285).

§ 86. I.-Bur.1; other examples. The I.-Eur. root leuk-, ‘to shine’ (0. Ind.
ruc-, Gk. Aevsds, O. Ir. loche, lichet Gen., ‘lightning,’ Gaul. Leucetios, W.
Tluched, Goth. liuhath, O. Engl. léoht, O. SI. luéi) appears in Lat. lux, litceo,
liicerna ; the root Shel-, ‘ green, yellow’ (O. Ind. hari-, Gk. xAwpéds, yAdn, O. Ir.
gel, ‘white,’ 0. H.G. gelo, O. Engl. geolo, ‘ yellow,’ Lith. Zélti, ‘to grow green,’
O. Sl. zelije, ‘ vegetables,’ zelenti, ‘ green’), in Lat. helvus, hélus, and hilus. Simi-
larly Lat. sal (Arm. ad, Gk. Gas, O. Iv. salann, W. halen, Goth. salt, O. SI. soli,
Lith. salunka, ‘salt-box’); Lat. /lés (O. Ir. blath, ‘bloom,’ W. blodau, ‘ flowers,’
Goth. bloma, 0. H.G. bluomo) ; Lat. cliinis (O. Ind. éréni-, W. clun, 0. Scand.
hlaunn, Lith. szfaunis) ; Lat. plénus, pleri-que, plétus (O. Ind. pratd-, Arm. li, Gk.
wAnpns, O. Ir. linaim, ‘ Tfill,’ lin, ‘a number,’ Lith. pilnas, ‘full,’ 0. Sl. plind) ;
Lat. in-chiitus (O. Ind. #rutd-, ‘heard,’ Gk. «Aurés, in Hom. ‘heard, loud,’ e.g,
évopa KdvTév, ALpH HAvTOs, O. Ir. cloth, ‘famous’; cf. 0. Engl. hlad, ‘loud’).

§ 87. R. J.-Eur.r is Lat.7, Thus the L-Eur. *réé-, ‘a king’
(O. Ind. raj-an-, O. Ir. 11, rig Gen., W. rhi, Gaul. Catu-riges, lit.
‘kings in fight’) is Lat. réx, stem rég-; I.-Eur. *bhéro, ‘I carry’
(O. Ind. bhaéra-mi, Arm. berem, Gk. pépw, O. Ir. berim, W. ad-feru
Inf., Goth. baira, O. Engl. bere, O. SI. bera) is Lat. féra; I.-Eur.
wert-, ‘to turn’ (O. Ind. vrt-, Goth. vairpan, Engl. worth, in ‘ woe
worth the day, Lith. vefsti and vartyti, O. Sl. vratiti) is Lat,
vértere ; 1-Kur. *kipro- (Gk. «ampos, boar, O. Engl. hefer, ‘ goat’)
is Lat. ediper, stem *cépro-, I.-Eur. rs before a vowel became 77 in
Latin, e.g. dorreo for *torseo (O. Ind. trg-, Goth. paursjan, ‘ to
thirst’ ; ef. Gk. répcouat). But 7 was assimilated to a following
s in pronunciation, e.g. Persa pronounced *Pessa (see ch. ii. § 96),

 

‘ On inserr. of Asisium the Umbr. (xi. 5390=i. 1412). The Volsienus
Voisiener Gen. Sing. (C. I. L. xi. 5389 of no. 5390 seems to be the son of the
= Bicheler, Umbrica Inscr. Min. i, Volsienus of no. 5389.

p. 172) corresponds to Lat. Volsienus
§§ 86-90.) REPRESENTATIVES OF I.-EUR. SOUNDS. L,R. 277

and I.-Eur. rs before a consonant became ss, e.g. tostus for
*tosstus, *torstus from torreo. HK was also assimilated to 7 in
Latin, e.g. stéla, dgéllus for *stérla, *agérlus. The Campanian
town of which the Latin name was Ated/a has on its coins Ader/.
so that its Oscan name must have been *Aderlo, a name appa-
rently meaning ‘the little black town.’ On the metathesis by
which 7% became ér, e.g. ter, terr- (*ters) for *tris, see § 13.
Sometimes the cacophony of a repetition of 7 (§ 84) seems to be
avoided by using x for one 7, e. g. cancer for *carc- (Gk. apx-ivos;
cf. O. Ind. kar-kar-a-, ‘ hard’) (but see ch. ii. § 105). On carmen,
possibly for *can-men, see § 78, and on 7 for d before f, v, 9, e.g.
arfuerunt, arvorsum, arger,§112. Between vowels I.-Eur. s became
7 in Latin, e. g. généris from the stem Sénés- [cf. Gk. yéve(c)os], on
which see § 148). J.-Eur. r is Umbro-Ose. r, e.g. Umbr. rehte
‘recte, Osc. Regaturef ** Regatori’ (see von Planta, i. p. 285).
In Umbrian, as in Latin, intervocalic s became r, but not in
Oscan (cf. infr. § 146).

§ 88. I.-Eur.r; otherexamples. Lat. rita (O. Ind. rdtha-, ‘chariot,’ O. Ir.
roth, ‘wheel,’ Gaulo-Lat. petor-ritum, ‘a four-wheeled vehicle,’ O. H. G. rad,
‘wheel,’ Lith. ratas) ; Lat. porcus (Gk. wépxos, O. Ir. ore, O. H. G. farh, Engl.
farrow, Lith. pafszas, O. Sl. prase) ; Lat. vir (O. Iv. fér, Goth. vair; ef. O. Ind.
vird-, Lith. vyras); Lat. inter (O. Ind. antar, O. Iv. eter) ; Lat. serpo (0. Ind.
srp-, Gk. éprw) ; Lat. riiber (O. Ind. rudhira-, Gk. é-pupés, O. SI. rtidrti; ef.
QO. Ir. ruad, Goth. raups); Lat. rwmpo, (O. Ind. rup- and lup-, O. Engl. be-
réofan, ‘to bereave,’ Lith. riipéti, ‘to trouble’); Lat. pro- (O. Ind. pra, Gk. mpé,
QO. Ir. ro, used like the augment to indicate a past tense, e.g. ro alt, ‘he
nourished,’ ro char, ‘he loved,’ 0. Bret. ro, Lith. pra-, O. Sl. pro-).

§ 89. ss for rs before consonant. Other examples are: Lat. testémentum
for *tesst-, *terst- from *trist- (Osc. tristaamentud Abl. Sg.) ; céna for *cesna,
*cessna from *cersna (Ose. kersna-, Umbr. sesna-). The spelling coena, due to
the analogy of Gk. xowds, is very old. It appears on an early Praenestine
cista of the third cent. B.c. with the representation of a kitchen where cook-
ing operations are going on. One servant is saying feri porod (= feri porro) ;
another replies cofeci (=: confeci); a third orders made mirecie (= made mi regie
or made mire, cie); a fourth misc sane (= misce sane); a fifth says asom fero
(= gssum fero) ; a sixth conjice piscim. The title is coenalia or coena pia (see Mel.
Arch. 1890, p. 303). Similarly Lat. fastigium for *farst- (O. Ind. bhrsti-, ‘a
point’) ; posco for *porcsco (O. Ind. prchami, 0. H. G. forscdn, Germ. forschen).

§ 90. rr for rs before vowel. Other examples: Lat. farreus (Umbr.
farsio- and fasio-; cf. O. Sl. brasino, ‘ food’); garrio (Lith. gafsas, ‘noise’); horreo
(O. Ind. hri-) ; porrum (Gk. mpdaor) ; terreo [O. Ind. tras-, Gk. tpé(o)w, Umbr.
tursitu tremitu] ; verres (Lith. vefszis, ‘ calf’; ef. O. Ind. vr8ni-, ‘ram’) ; verriica,
‘a rising-ground’ in O. Lat. (Lith. virszis ; cf. O. Ind. varSman-) ; Maspiter,
278 THE LATIN LANGUAGE. [Chap. IV.

a byform of Marspiter ; Tuscus (Umbr. Tursco- and Tusco-). Lat. fers has pro-
bably re-appended -s to an older *fer, *ferr for *fer-s [so ul-s has appended -s
a second time to *wl, *ull for *ol-s (ch. ix. § 56).] In Umbrian inscriptions
written in the Latin alphabet, the Umbrian d-sound (see ch. ii. § 88) is
written rs, e.g. capirse (in the native alph. kapife), (Lat. capidi), Dat. Sing. of
caipid-, a bowl.

§ 91. nforr. Lat. crépuscilum, if it has cr- for cn- (compare Gk. «vépas),
with that change of cn- to cr- which we find in some languages, e.g. Gael.
enu (pronounced cru with nasal vowel), Bret. kraoun, ‘a nut,’ must be like
our ‘gloaming,’ a dialectal word: Varro (L. L. vi. 5) makes it Sabine (see
§ 80). On the spellings menetris for mereti ix, &c., see ch. ii. § 105.

§ 92. The L- and R-Sonants. For the sounds to which &,
ér (dé, ré) are reduced in the unaccented syllable of such a word
as I.-Kur. *drtd-, *drti (O. Ind. diti-, Gk. dpards and daprds,
dépos, Lith. nu-dirtas, Goth. ga-taurps) from the root der-, ‘ to
flay, some write 1, r, others al, or (la, ro), the o indicating an
obscure vowel sound (also ¢r, °l, &c.). In Latin these sounds
became 0/, dr, e. g. mors, Gen. mortis (O. Ind. mrti-, Goth. maurpr,
O. H. G. mord, Lith. mirtis, O. Sl. sti-mriti), fors, stem *forti-
(O. Ind. bhrti-, O. Ir. brith, Goth. ga-baurps, ‘ birth,’ O. Engl. ge-
byrd, ‘ fate’), the o being subject to all the changes of for I.-Eur.
5 (e. g. pudsus, Gk. tadrés from the root pel-), while before a vowel
we seem to find d/, dr in sdlia (O. Iv. sail; ef. Gk. €Xéky with
E-grade), céiro (Umbr. karu,‘a portion’; cf. Gk. xefpw) ; in Greek
they became dp (pd, e.g. narpaor), &d (AG), in Teutonic tir, il ; in
Slavonic ir, il, and so on; so that it is only by reference to
another J.-Eur. language that we can tell whether Lat. 3/, dr are
I.-Eur. 61, dr, or I.-Eur. 1, r. (Lat. 37 may also be I.-Eur. él, § 10.)
We occasionally find /é, 7@ in Latin words from roots in \é, ré,
which are probably to be explained similarly by the weakening
influence of the accent, e. o« fringo, fragilis from the root bhreg-,
‘to break’ (Goth. brikan, O. Engl. brécan) (see Osthoff, Morpi.
Unt. v. Pref.,and above § 3). Whether Gk. pw, pa, Lat. ra, of Gk.
otpwrtds, némpwrat, Kpards, Lat. stratus, clades, and Gk. op (ap),
Lat. ar of Gk. ordpyuat, Lat. pars, pdértior, quartus, are rightly
called the long-sonant vowels from roots ster-, per-, &c. is a point
which has not yet been settled (see § 1). There are some indica-
tions that vocalic 1 and r were sounds not unknown to the Italie
languages down to a fairly late period, e.g. Marrucinian paersi
‘pacer (pacris) sit, ‘may she be propitious’ (Zv. 1.1. I. 8) (cf.
§§ 91-95.] REPRESENTATIVES OF I.-EUR. SOUNDS. L, R. 279

Umbr. pacer, Lat. acer, acris Fem.), Sabine Atrno (Zv. 10), Ose.
Tantrnnaiim Gen. Pl., though how far these are merely graphic
is hard to decide (cf. ch. iii. § 14); corresponding to Gk. dypds we
have in Lat. ager, in Umbr. ager, and so on (see ch, vi. § 4). On Lat.
ter(r) for *tris, facul for fac(i)li, see § 13. I.-Eur. 1, r receive
the same treatment in Umbro-Ose. as in Latin, e.g. Umbr. orto-
(Lat. ortus), Ose. molto (Lat. mu/ta, a fine) (see von Planta, i. p. 314).

§ 93. Other examples of the liquid Sonants. I.-Eur. *prk-sko (O. Ind.
prchami, preha-, ‘enquiry,’ Arm. hareanem, hare, O. H. G. forscdn, forsca)
Lat. posco for *porc-sco; I.-Eur. *krd-. (O. Ind. hid, Gk. «apSia and xpadéy,
O. Ir. cride, Lith. szirdis, O. SI. sridice), Lat. cor for *cord ; I.-Eur. *prka-, ‘the
ridge of a furrow’ (W. rhych F., 0. Engl. furh F.), Lat. porca (Varro, R. R. i.
29. 3 qua aratrum vomere lacunam striam fecit, ‘sulcus’ vocatur. quod est
inter duos sulcos elata terra dicitur ‘ porca’; Paul. Fest. 77. 1 Th. explains the
name Imporcitor as ‘ qui poreas in agro facit arando,’ the name of a deity invoked
by the flamen in the sacrifice to Tellus and Ceres, with a number of other
agricultural divinities: Vervactor, Reparator, Imporcitor, Insitor, Obarator,
Occator, Saritor, Subruncinator, Messor, Convector, Conditor, Promitor, whose
names are given by Fabius Pictor ap. Serv. ad Virg. G. i. a1. The word
porca occurs also in a line of Accius ap. Non. 61. 19 M. bene proscissas cossi-
gerare ordine porcas) ; I.-Eur. krn- (Gk. #dépvos Hesych., Goth. haurn), Lat.
cornu (ef. Lat. cornus, cornel, Gk. xpdavos, xpavov) ; I.-Eur. *prso- (Gk. mpdoov),
Lat. porrum for *porsum. Certain examples of J.-Eur. | in Latin are not
numerous. Ulmus may stand for *]mo- (Ir. lem, Russ. ilemit) ; or for *elmo-
(O. Engl. elm); or for *olmo- (O. Seand. almr) ; mulctus, milked, for *mlkto-
(Lith. mitsztas, O. Ir. mlicht or blicht, ‘ milk’) ; oc-cultus for *-klto- from the -
root kel-, ‘to hide’; cf. sepultus from sépétio.

§ 94. Other examples of al, ar, 14, ra. Lat. palma, palm of hand (0. Tr.
lam F.; cf. Gk. maddy) ; scalpo (ef. sculpo) ; clades (cf. Gk. cAadapés, perculsus,
from the root keld- of per-cello) ; cldémo (ef. calo, cdlendae) ; flavus (cf. fulvus,
O. Lat. Folvius) ; latus, carried (ef. tollo) ; ldtus, wide, earlier silatus, from the
root stel-, ‘to extend’ (O. SI. stelja) ; ardea (cf. Gk. é-pwiids) ; armus (O. Ind.
irmas, Goth. arms, O. Sl. rame ; cf. Pruss. irmo) ; carpo from the root kerp-,
‘to eut’ (Lith. kerpt) (cf. Gk. «aprés, Engl. harvest) ; fastigium for *farst- (cf.
O. Ind. bhriti-, ‘ point,’ 0. Engl. byrst, ‘ bristle’), or *frast- (see ch viii. § 18, on
farcio for *fracio) ; crabro for *crd-sro (cf. Lith. szirszi, O. SI. sriigeni) ; cratis and
cartilago (ef. Gk. #dprados, basket, O. H. G. hurt, Engl. hurdle) ; fraxinus and
Jarnus (O. Ind. bhirja-, ‘a birch-tree ’) from the root bherg-, (O. Engl. beore,
Lith. bérZas, O. SI. bréza); granum (O. Ind. jirnd-, ‘worn out’; cf. Goth. kaurn,
‘corn,’ Lith. Zirnis, ‘pea,’ O.S1. zrtino, ‘a grain’) ; gratus (O. Ind. girta-, ‘ wel-
come’); radix for*vra- (cf. Goth. vaurts, Engl. wort); stramen (Gk. o7p@pa) from
sterno ; lana for *vléna (O. Ind. tirna, Gk. odAos for *FoAvos), On Lat. dil, dr for
al, er before a vowel, see M.S. L. viii. 279, Osthoff, Dunkles w. helles 1, p. 52,
(palea, parens, varix, &e.), and cf. § 3 on piteo (Gk. mer-), &e,

§ 95. Tenues, Mediae, and Aspirates. In Sanscrit we have
four varieties of each class of mutes or stopped consonants,
280 THE LATIN LANGUAGE. (Chap. IV.

(1) tenues, or unvoiced, (2) mediae, or voiced, (3) tenues aspiratae,
or tenues followed by / (like our ¢/ in ‘ant-hill’), (4) mediae
aspiratae, or mediae followed by % (like our dé in ‘sandhill’).
But these varieties are not kept distinct in other L-Hur.
languages. In Greek tenues aspiratae take the place of the
Sanscrit mediae aspiratae (e.g. diyds, O. Ind. dhiimdés); in the
Celtic and Slavonic families the mediae and mediae aspiratae of
Sanscrit are merged in mediae (e. g. Lith. dti-ti, ‘ to give,’ O. Ind.
di-, O. Ir. dan, ‘a gift, an accomplishment’; Lith. délé ‘ a leech,’
O. Ind. dha-, ‘to suck,’ O. Ir. dél, ‘ teat’). The tenues aspiratae
of Sanscrit are especially difficult to trace in the other languages ;
nor is it always easy to say whether they are due to some
phonetic law peculiar to Sanscrit, or represent J.-Eur. tenues
aspiratae. In the O. Ind. root stha-,‘ to stand,’ where the dental
tenuis is found in all other languages (Gk. formu, Lat. sto, &c.), it
seems probable that the I.-Eur. form of the root was sta-, whereas
in O. Ind. nakhé-, ‘ a nail, claw, for which we have an aspirate in
Gk. drvy- the tenuis aspirata may be original. Tenues aspiratae
have been with more or less probability conjectured for such
words as I.-Eur. *konkho-, ‘a shell’ (O. Ind. farakhé-, Gk. xdyxos,
Lat. congius, a quart) ; I.-Eur. skhi(n)d-, ‘ to split’ (O. Ind. chid-,
Gk. cyiGe, sywdadyss, a splinter, Lat. seiado, O. H.G. scintan,
Germ. schinden); the suffix of the 2 Sg. Pft. Ind. (O. Ind.
vét-tha, Gk. ofc-da, Lat. vid-is-ti, Goth. las-t, &c.); and the
same hypothesis has been used to explain the anomalous cor-
respondence of Latin /- and Goth. h- in the verb, ‘to have,’ Lat.
hibet, Goth. habaip (I.-Eur. khabhé- ?), &e.

These I.-Eur. tenues, mediae, and aspiratae were liable to
change their character under the influence of an adjoining
consonant. A media became a tenuis before an unvoiced con-
sonant, e.g. ].-Hur. *yukto-, P.P. P. from the root yeug-, ‘ to
join’ (O. Ind. yukta-, Lat. gunctus), A tenuis similarly became
a media before a voiced consonant, [ef. I.-Eur. *si-zd-o, the
reduplicated form of the root sed-, ‘ to sit’ (Lat. s#dv)|. A media
aspirata before ¢ or ¢ is in Latin and other languages treated
like a tenuis, e.g. vectus, veai (veces), from Lat. vého (I.-Eur.
root wegh-), but whether this was the case in what is called ‘ the
Indo-European period’ is not certain. More plausible is the
§§ 96-98.] REPRESENTATIVES OF I.-EUR. SOUNDS. P, B, BH. 281

theory that t before t or th,and perhaps d before d, dh, produced
already at this period some sibilant sound; for a trace of this
appears in every branch of the I.-Eur. family, e. g. from the root
sed-,; with the P. P. P. suffix -to-, we have Zend ni-Sasta-, Lat.
-sessus, Lith. séstas, O. Scand. sess) (for other examples in Latin,
see § 108). The occasional confusion, too, which we find between
tenues and mediae (e.g. Gk. cxardvn beside Lat. scabo), mediae
and aspiratae (e.g. Gk. oréusw beside aoreygrjs; O.Ind. aham
beside Gk. éyw, Lat. ego, &c.) may in many instances date from
the same early time.

In Umbrian a tenuis becomes a media before r in swbra ‘supra,’
podruhper ‘utroque,’ regularly after n, e.g. venga ‘juvencas,
ander ‘inter’ (Osc. anter) (von Planta, i. p. 547); and we have
both in Osean and Umbrian many instances of a tenuis appearing
for a media, which are by some explained as miswritings, due to
the earlier use of the tenuis-symbol only in the native alphabets
(derived from Etruscan, ch. i. § 1), but are by others considered
as a proof that the Italic mediae were not voiced (ch. ii. § 77).
(For instances, see von Planta, i. p. 555.)

- § 96. Media or aspirata assimilated to unvoiced consonant in Latin.
Other examples are cet-te for *cé-ddte, the plural of cédé, give, lit. ‘give here’
(ch. vii. § 15), hoc(c) for *hod-ce (wh. vii. § 16)’; topper for *tod-per (ch. ix. § 7) ;
ac-tus from dgo; scriptus from scribo ; vec-tus from veho, &c. On spellings like
optenui (Scip. Ep.\, apscede, urps, see ch. ii. § 80.

§ 97. Tenuis assimilated to voiced consonant in Latin. On ob-duco
beside op-tenui, see ch. ii. § 73. I.-Eur. d becomes t before y (unvciced ?) in
Latin, e.g. dtrox, from ad-, a byform of the root od- of ddium (§ 113).

Interchange of tenuis and media in Latin. (See ch. ii. §§ 73-77.:

Interchange of media and aspirata in Latin. On g occasionally appear-
ing for I.-Eur. gh, see § 116.

P, B, BH, PH.

98. P. I.-Eur. p is Lat. p, e.g. I.-Eur. *péku- N. (O. Ind.
pésu, Goth. faihu, Germ. Vieh, Engl. fee), Lat. péeu, I.-Eur.
*séptm (O.Ind. sapté, Arm. evtn, Gk. énra, O. Ir. secht,
W. saith, Goth. sibun; cf. Lith. septyni). P becomes m before
n, e.g. somnus for *sop-nus ; it is assimilated before labials, e. g.

summus for *supmus, suffio for *supfio.
282 THE LATIN LANGUAGE. (Chap. Iv.

The Labial Tenuis receives a similar treatment in Umbro-
Oscan, e.g. Umbr. patre, Osc. paterei ‘patri’; but pt is in
Ose. ft, in Umbr. ht, e.g. Ose. serifto-, Umbr. sereihto- ‘seripto’
(see von Planta, i. p. 424).

§ 99. Other examples of I.-Eur. p. I.-Eur. *péter-, (O. Ind. pitdr-, Arm.
hair, Gk. narnp, O. Ir. athir, Goth. fadar) Lat. pater ; O.-Eur. “pro (0. Ind.
pra, Gk. mpd, O. Iv. ro, O. Bret. ro, Lith. pra-, O. Sl. pro-), Lat. pro- ; I.-Eur.
*serpd (O. Ind. sérpa-mi, Gk. £p7m’, Lat. serpo ; I.-Eur. root spek- (O. Ind. spa8-,
0.H.G. spehon, Engl. spy), Lat. au-spex, con-spicio ; from I.-Eur. root prek-
(O. Ind. pra3-nd-, ‘a request,’ Goth. fraihnan, Germ. fragen ; Lith. praszyti,
0. SI. prositi) come Lat. précor, and pricus ; I..Eur. *népot- (0. Ind. napat-,
M. Ir. niae, niath Gen., 0. H. G. nefo), Lat. népis ; I.-Eur. root pak-, pag-
(O. Ind. paia-, ‘string,’ Dor. Gk. 7éyvuju, Goth. fahan, ‘to catch,’ Germ. Fuge)
Lat. pax, pagina; I.-Eur. *kipro- (Gk. «ampos, O. Engl. heefer, ‘goat’), Lat.
cdiper.

$100. B. I.-Eur. b is Lat. 4, e.g. I.-Eur. *pibo, ‘I drink’
(O. Ind. pibami, O. Ir. ibim), Lat. 0240 for *pibo (§ 163). Lat. 4,
whether from I.-Eur. b or bh, becomes m before x, e.g. scamnum
beside scabellum. Latin 6 may represent I.-Eur. -bh-, e. g. seribo
(cf. Gk. cxapipdopat, scratch) (§ 103), -dh-, e.g. riber (Gk. épvOpds
(§ 114), m before r, e.g. dibernus (Gk. xemepivds) (§ 77), s before
r, e.g. céreb-rum for *cerés-rum (§ 152), dw-, e.g. bis for *dwis
(Gk. dés) (§ 68). On the confusion of 4 and v in IJate spellings,
see ch. 11. § 52, and on the substitution of -d for -p in sud, ab,
ch. ii, § 73.

§ 10]. Other examples of I.-Hur. b. Lat. balbus (O. Ind. balbala-, Gk.
BapBapos) ; Lat. lambo (O. H. G. laffan, O. Sl. lobiizit; cf. Gk. Aawrw) ; Lat.
libricus (Goth. sliupan, ‘to slip’); Lat. laébiwm (O. H. G. lefs, O. Engl. lippa).

§ 102. mn for bn. ammnis from abh- (0. Ind. dmbhas-, ‘ water,’ Ir. abann,
‘yviver’). On the spellings amnuere in glosses, see Liwe, Prodr. p. 421; ef.
amnegaverit ‘C.I.L. vi. 14672). Similarly bm of *gliib-ma (root gleubh-, Gk.
yAdgw) became mm, and was reduced to m (ch, ii. § 127), glima,

§ 103. BH. [.-Eur. bh when initial became Latin 7, when
medial 4, e.g. I.-Eur. root *bher-, ‘ to carry’ (O. Ind. bhar-, Arm.
berem, Gk. ¢épw, O. Ir. berim, W. ad-feru Inf., Goth. baira, Engl.
I bear, O. SI]. bera) is Lat. féro; I.-Eur. *orbho- (Arm. orb, Gk.
éppavds, O. Ir. orbe, ‘ inheritance,’ Goth. arbi, Germ. Erbe, ‘ heir’),
Lat. orbus ; I.-Eur. root bheu- (O. Ind. bhi-, Gk. giv, O. Ir.
buith, Lith. baé-ti, O. Sl. by-ti), Lat. fui, fiutirus; I.-Hur.
*bhrator- (O. Ind. bhratar-, Gk. ¢pdrwp, O. Ir. brathir, W.
§§ 99-105.] REPRESENTATIVES OF I.-EUR. SOUNDS. T, D, DH. 283

brawd, Goth. bropar, Lith. broter-élis, O. SI]. bratrti), Lat. frader.
A good example of the different treatment of initial and medial
bh in Latin is the word for a beaver, I.-Eur. *bhébhru- (O. H. G.
bibar, Lith. bébrus, O. SI. bebrti ; O. Ind. babhri-, ‘ brown’),
Lat. filer. On mn for bn fwitth 6 from J.-Eur. b or bh),
e.g. scamnum from skabh- ‘to support, with Dim. scabellum
(-22um) (also scamillum Ter. Scaur. 14. 6 K.), see § 102, and on
dialectal 7 for 6 (e.g. Alfius, a byform of Albius), see ch. ii.
§ 83. J.-Eur. bh is Umbro-Ose. f, whether initial or internal,
e.g. Umbr.-Ose. fust ‘ erit,’ Umbr. a/fo-, Osc. Alafaternum ‘ albo-.’
On Faliscan hata for Lat. fiiba, see § 121, ch. i. § 57.

§ 104. L.-Hur bh; other examples. Lat. nébiila (Gk. vepédn, O. Ir. nél for
*neblo-, 0. H. G. nebul, Germ. Nebel), Lat. ftiteor, fama (Gk. gypi); Lat.
sorbeo (Gk. fopéw, Arm. arb-enam, Lith. srebit) ; Lat. wmbo, umbilicus (O. Ind.
mabht. nabhila-, Gk. dugadds, O. Ir. imbliu, O. H. G. naba, nabolo, Pruss.
nabis, Lett. aba Lat. lds (O. Ir. blath, ‘bloom,’ Goth. bloma)}; Lat. albus (Gk.
cdpés, white leprosy); Lat. ambi- (O. Ind. abhi, Gk. dui, O. Ir. imme for
imbe, Gaulish Ambi-gatus’ O. Engl. ymb, Engl. ember-days, from 0. Engl.
ymb-ryne ‘running round, circuit,’ Germ. um for umb) ; Lat. ambo (O. Ind.
u-bha-, Gk. dupw, Goth. bai, baj-dps, Engl. b-oth, Lith. abu, O. Sl. oba); Lat.
Jindo from I.-Eur. root bheid- (0. Ind. bhid-, Goth. beita, Engl. I bite); Lat.
Sagus (Dor. Gk. payés, O. Engl. bie). Barba for *farba (I.-Eur. bhardh-, 0. Sk.
brada, Lith. barzda, Engl. beard) is due to assimilation (see § 163).

T, D, DH, TH.

§105. T. J[.-Eur. t is Lat. ¢; e.g. the I.-Eur. root ten-, ‘ to
stretch ’ (O. Ind. tan-, Gk. reévw, O. Ir. tennaim, Goth. uf-panja,
Germ. dehnen ; O. Ind.-+tant-,‘ thin, Gk. ravi-arepos, O. Iv. tana,
W. teneu, O. H. G. dunni, O. Engl. pynne, Lith. dial. tenvas, O. SI.
tintkit) appears in Latin ¢en-do, ténuis ; I.-Eur. *wert-, ‘ to turn’
(O. Ind. vartaté 3 Sg., Goth. vairpan, Germ. werden, Lith. veiszti,
vartyti, O. Sl. vratiti, W. gwerthyd, ‘ spindle’) is Lat. verdére.
Before 7 it is dropped when initial, e.g. datws for *tlatus, P. P. P. of
féero, til, but becomes ¢ when medial; thus the suffix -tlo-,
which indicates the instrument with which an action is per-
formed, or the place of its performance, appears in Latin as
-clo-, e.g. véhiclum or vehictilum, ‘that by which one is carried’ ;
poclum or pocilum, ‘that out of which one drinks,’ ciidiclwm or
cubtcitlum, ‘ the place where one lies down,’ &c., by the same
change as is seen in later Latin veclus for vé#(a%)Jus, in the
284 THE LATIN LANGUAGE. [Chap. IV.

English mispronunciation ‘acleast’ for ‘at least,’ in Mod. Gk.
cetkdo from Gk. cedrdov, &c. Before ¢ it is assimilated, e.g.
ac for *acc from *ate, atqu(e), while ¢¢ became ss, after a con-
sonant s,e.g. passus for *pattus from patior, salsus for *salttus
from sadlo, *saldo, versus from verto, &e.

Final -nt seems to have become -zs in the Umbro-Oscan
languages, for the 3 Plur. Act. ending of Secondary Tenses
(I.-Eur. -nt) appears in them as -ns, e.g, Osc. fufans ‘erant’
quasi ‘fubant’ (ch. vil. § 73). The same may hold of Latin,
for gudtiens, titiens suggest the O. Ind. suffix -yant of ki-yant,
‘how large,” &c. (See 4. L. L. v. 575). Latin ¢ represents
an original d before 7, e.g. Gtrow beside ddinm (see § 113). On
the loss of ¢ in consonant-groups, e.g. now for *noc(t)s, vermina
for *ver(t)mina, see § 157, and on xn for tn, e.g. annus, § 161.

In Umbro-Ose. I.-Eur. t remains, e.g. Umbr. tota-, Ose.
twFro, a community, but tl became (as in Latin) kl, e.g. Umbr.
pihaklu, Osc. sakaraklim (cf. Pelignian sacaracirix ‘ sacratrices,’
pristafalacirix ‘ praestabulatrices’).

§ 106. Other examples of I.-Eur. t. I.-Eur. *imtom (O. Ind. Satém, Gk.
é-carév, O. Ir. cét, W. carft, Goth. hund, Lith. szithtas, O. Sl. stito) is Lat.
centum ; I.-Eur. *6kt0(u) (O. Ind. a&ta, Arm. ut, Gk. durd, O. Ir. dcht,
W. wyth, Goth. ahtau, 0. Engl. eahta, Lith. asztf-ni, O. SI. os(t)-mi), Lat.
octo; J.-Eur. *mater- (O. Ind. matdr-, Arm. mair, Dor. Gk. parnp, O. Ir.
mathir, O. Engl. mddor, O. Sl. mater-, Lith. motera, ‘ woman’), Lat. méter ;
I.-Eur. *wétos, ‘a year’ (O. Ind. vats-d-, ‘calf,’ lit. ‘yearling,’ Gk. (F)éros,
O. SI. vetiicht, ‘old’), Lat. vttus-tus, vétus ; I.-Eur. *és-ti (0. Ind. dsti, Gk. gor,
Goth. ist}, Lat. est ; I.-Eur. root steig-, ‘to pierce’ (O. Ind. tij-, tigmd-, ‘sharp,’
Gk. orifw, orvyun, Goth. stiks, ‘point of time,’ Germ. Stich), Lat. instigo; the
‘Comparative’ Suffix -téro- (O. Ind. katard-, Gk. mérepos, O. Ir. -ther, Goth.
hvajar, Engl. whether, Lith. katras, O. Sl. kotory-ji) is in Lat. -téro-, e. g. titer.

§ 107. I.-Bur. tl. Lat. léquor is I.-Eur. tlog3- (O. Ir. atluchur for ad-tluchur
in the phrase atluchur bude, ‘ I thank,’ ‘ad-loquor gratias’), On the spellings
stlis, sclis, sltis for class. lis, see § 150.

§ 108. I.-Eur. tt. Verbs whose stem ends in a dental show ss (after
a diphthong, &c. reduced to s, ch. ii. § 127) in their P. P. P., formed with the
suffix -to (e. g. fissus from findo, ausus from audeo, fisus from fido, &e. (see § 155
and ch. v. § 28). So in formations with the suffix -himo-, e. g. vicensumus for
*vicent-tumo- (ch. v. § 14), Adjectives in -dsus (ch. v. § 65), &e.

The combination t# in Latin appears where a vowel has been dropped by
syncope, e. g. cette for *cé-ddte 2 Plur. Imper., egret(t)us, adgret(t)us (§ 109).

$109. D. Of I.-Hur. d, Lat. d examples are: I.-Eur. *dékm,
§§ 106-110.] REPRESENTATIVES OF I.-EUR.SOUNDS. T,D, DH. 285

‘ten’ (O. Ind. déSa, Arm. tasn, Gk. 8éxa, O. Ir. deich, W. deg,
Goth. taihun, O. H. G. zehah, Lith. dészimt, dészimtis, O. SI.
deseti), Lat. décem; I.-Eur. root weid-, ‘ to see, know’ (O. Ind.
véda, Arm. git-em, Gk. ofa, O. Ir. ad-fiadaim, ‘I narrate,’ Goth.
vait, Engl. wot, O. SI. védéti Inf.), Lat. vdeo, vidi. In the Sabine
dialect d became 1 (J. /. ii. 157); and we find the spelling 7
sanctioned in a few Latin words at the beginning of the literary
period, e.g. ddéertma, older dacrima, dacruma (Gk. daxpv, Welsh
dagr, Goth. tagr), (Paul. Fest. 48.15 Th. ‘dacrimas’ pro lacrimas
Livius saepe posuit). D became 7 before gy in mergo, mergus (O.
Ind. madgii-, ‘a waterfowl’), &c., and before v and f, though the
old forms arger, arvorsus, arfuerunt, had their @ restored from ad
at the beginning of the second cent. B.c. Initial dw- became
b, e.g. bellum (older duedlum, always a dissyllable in Plautus), Lis
for *dwis (Gk. dis), (see § 68), and initial dy- became j- (y), e.g.
Jovis, O. Lat. Diovis (see § 63), but internal -dw- leaves v in
sudvis (§ 71). D is assimilated to a following m in rdémentum from
-rado, caementum from caedo, c in O. Lat. reectido, hoc for *hod-ce,
¢ in relligio, pelluviae, water for washing the feet, graddae, stilts,
from gradior, but assimilates a preceding / in percello for *-celdo
(cf. clades), &c. (see ch. viii. § 33), sallo for *saldo (Goth. salta).
It became ¢ before 7, e.g. dtrow (cf. ddiwm), and before ¢, this ¢¢
becoming’ ss (see § 108), e.g. egressus, adgressus. The forms egrettus,
adgrettus (Paul. Fest. 55.3 Th. quotes egretus, adgretus, apparently
from some early writer, earlier than the practice of writing the
double consonant) seem to stand for*egred(i)to-,adgred(i)to-(§ 108).

[-Eur. d remains in Umbro-Osc., e.g. Umbr. devo-, Osc.
deivo- ‘divus,’ but between vowels became in Umbrian a sound
(a4? ch. ii. § 88) which is expressed in the native alphabet bya sign
conventionally written by us ¥ and in the Latin alphabet by rs,
e.g. te¥a, dersa ‘det’ (quasi *dédat) (Pel. dida; cf. Osc. didest).
For nd we have in Umbro-Osc. nn as in -nno- the Gerundive
ending, e.g. Umbr. pihano- ‘ piandus,’ Osc. ipsanno- ‘ operandus.’

§ 110, Other instances of I.-Bur. d, Lat. d. I.-Eur. root deik-, ‘to point’ (0.
Ind. di8-, Gk. defnvuy, Goth. ga-teihan, ‘ to proclaim’), Latin dicere, indicare ; I.-
Eur. root déks-, ‘right hand’ (O. Ind. dakiina-, Gk. Beguds, det: repés, O. Ir. dess,
Gaul. Dexsiva, W. deheu, Goth. taihsva, Lith. deszine, 0. SI. desinit), Lat.

deater ; I.-Eur. root sed-, ‘to sit’(O. Ind. sad-, Gk. 0s, O. Ir. sudim, sadaim,
W. seddu Inf., Goth. sita, 0. SI. sedlo,‘saddle’), Lat. stdeo. Similarly Lat. scindo
286 THE LATIN LANGUAGE. (Chap. IV.

(0. Ind. chind-, Gk. cyivd-adpds, a splinter) ; Lat. divus (O. Ind. déva-, Lith.
dévas}; Lat. pes, Gen. pédis (O. Ind. péd-, Gk. rods, roids Gen., wé(a, Goth.
fotus, Lith. péda, ‘ footprint’); Lat. suddeo (O. Ind. svadi-, ‘sweet,’ Gk. #5us,
Goth. suts) ; Lat. édo (O. Ind. ad-, Arm. ut-em, Gk. é5w, Goth. ita, Lith. édu,
O. SI. jadi, ‘ food’).

§111. Lat.1 ford. Lautia, which usually occurs in an alliterative formula,e.g.
Liv. xxx. 17. 14 aedes liberae, loca, lautia legatis decreta; xxviii. 39. 19 locus
inde lautiaque legatis praeberi jussa) was in Old Latin dautia, a form quoted
from Livius Andronicus by Paul. Fest. 48. 16 Th. ‘dautia’ (Livius saepe posuit’,
quae lautia dicimus, et dantur legatis hospitii gratia; lingua, for I.-Eur.
*dnghta- or *dnghwa- (Goth. tuggo ; cf. O. Ir. tenge) was in Old Latin dingua
{nosnunc .. . linguam per 1 potius quam per d (scribamus), Mar. Vict. 9. 17 K.;
communionem enim habuit littera (1 cum d) apud antiquos, ut ‘dinguam’ et
linguam, et ‘dacrimis’ et lacrimis, et ‘Kapitodium’ et Kapitolium, id, 26. 1 K.).
(Was Aquilonia the Oscan Akudunnia-, now Cedogna? ef. O. Umbr. aked@unia-.)
Pompey, according to Mar. Vict. 8. 15 K., affected the old spelling and
pronunciation kadamitas for cdldmitas (perhaps from the same root as Oscan
cadeis amnud, ‘with intent to injure,’ ‘out of malice,’ Gk. rexadfjoa, explained
by Hesychius as BAdyar, kax@oa, orepjoa); the di Novensiles appear on an
old inscription from Picenum as DEIV. NOVESEDE (C.J.L. i. 178) (ef. the
Marsic esos novesede, Zv. I.I.I. 39). In many or all of these words the
preference of the byform with J can be explained by false analogy ; in lingua
by the analogy of lingo (ef. O. Ir. ligur, ‘ the tongue,’ Arm. lezu, Lith. lézuvis) ;
in calamitas by the analogy of calamus; in lautia by the analogy of lavo, &e.
Similarly O. Lat. délicure with the sense of indicare (e.g. Plaut. Mil. 844) may
owe its 7 to the influence of déliquare, to clarify. Perhaps Lat. lympha,
a Graecized form (cf. Gen. Pl. lymphon, Varro, Sat. Men. 50 B.) of *dumpa, (Ose.
Diumpa-), with Lat. Ulysses, does not exhibit a Latin change of d to 1, but
a change which had already been made in Greek ; for Quintilian, i. 4.16, quotes
dialectal Gk. ’OdAvaceds ; and on Gk. vases (see Kretschmer, Gr. Vas. p. 146) we
have ’Odvoeus (C.I. G. 7697), frequently ’OAureus (C.I.G. 7383, 7699, 8185,
8208), while Varro, L.L. vii. 87, quotes Auppddnmros (cf. O. Lat. Thelis for
@éms, Varro, L. L. vii. 87). The exact relation of simila, similago, fine wheaten
flour (cf. Germ. Semmel), to Gk. cepifadis is doubtful, as also that of casila
(Paul. Fest. 33. 22 Th.; for cassilla?} to cassis, cassida (an Etruscan word
according to Isidore, Orig. xviii. 14. 1). The form reluvium mentioned by
Festus (370. 17 Th. rediviam quidam, alii ‘reluvium’ appellant, cum circa
unguis cutis se resolvit, quia luere est solvere) never ousted the form redivia.
It does not appear to have been a phonetic variety of redivia, but rather
a separate word, perhaps a grammarian’s coinage, derived from luo, as redivia,
red-uvia was derived from *évo of ind-uo, ex-uo, extiviae. Some of the forms with
i mentioned above, e. g. calamitas, ‘injury to crops,’ may similarly have been
different words from the d-forms. [So consilium, &c. beside praesidium, &c. ;
ef. Mar. Vict. 9. 18 K. praesidium per d potius (secribamus) quam per 1;
considium Plaut. Cas. 966 (see below on solium)]. Some appear to be dialectal ;
e. g. lepesta or lepista [Greek deneora (but cf. AenacrH), according to Varro, L. L.
v. 123] was Sabine (Varro, J.c.), and perhaps Novensiles (id. v. 74) (ef. Paul.
Fest. 77. 7 Th. ‘inpelimenta’ inpedimenta dicebant) ; so probably ldria for
*darix (O. Iv. dair for *darix, ‘an oak’), and possibly Jaurus for *daurus (O. Ir.
daur for *darus, ‘an oak *) ; Melica (gallina) for Médica (Varro, R. R. iii.g. 19) (ef.
§ 111.] REPRESENTATIVES OF I.-EUR. SOUNDS. T,D,DH. 287

Paul. Fest. 89. 27 Th.) ; and the few modern Italian words which show this
change of d tol may be dialectal too, e.g. cicala from Lat. cicada, ellera from Lat.
hedera, trespolo from Lat. ts and pédes, as tiepolo is a dialectal variety of tepido
(Lat. tpidus) (ef. the Digentia, now Licenza, in the Sabine district). It is
possible, but unlikely, that Lat. publicus (Umbr. pupdiko-), Pablius from pies,
a name whose spelling was often altered after Poplicola and other names derived
from pépiilus (ef. Umbr. Puplecio-), may stand for *pabdi- and exhibit that
change of d to 7 after a labial which is seen in these Italian forms tiepolo and
trespolo (cf. the byform impélimenta). The town-name Telésia shows d on
Osean coins with Tedis (Zv. I. I. I. 262). Other examples of J for d are levir,
better laevir, a brother-in-law for *daever (O. Ind. dévar-, Gk. dayp for *daFnp,
Lith. déveris, O. Sl. déveri), which Nonius (557. 6 M.) explains : quasi laevus
vir; 6leo, which stands in puzzling contrast to ddor from the root od-, ‘to smell’
(Arm. hot, Gk. é84q, Lith. iid%iu) (odefucit of Paul. Fest. 193. 21 Th. ‘odefacit’
dicebant pro olfacit, may be a grammarian’s coinage). But the other
instances usually quoted are doubtful: digo may come from *dvilis as well as
from dvidus ; sélium and O. Lat. silum (e. g. Enn. Ann.93 M. scamna solumque)
may be from a root swel- (Gk. céAya, Lith, stilas, ‘a bench’), and not represent
*sodium (O. Ir. suide) from the root sed-, as sdlum, the ground, the sole of the
foot (ef. solea, a slipper) stands for *swolo-, (O. Ir. fol, ‘the base, foundation °) ;
mdlus might be from *maz-lo-, as Engl. mast is from *maz-do-. The inter-
change of d and lin lacruma, lingua, &c. was well known to the native Latin
grammarians, and was often appealed to by them to support theoretical
etymologies of words with i. Thus they explained sella (really for *sed-la ;
Lacon. Gk. éAAa ; ef. Goth. sitls, Engl. a settle, O. Sl. sedlo, ‘a saddle’) by this
interchange (Mar. Vict. 26. 3 K.); similarly ancilia from ambe-cido (Varro,
L.L. vii. 43) ; séliquastra from sédeo (Fest. 508. 10 Th.) ; méditor from pererdw
(Serv. ad Eel. i. 2); délicatus from dédicatus [Paul. Fest. 49. 17 Th. ; who quotes
a (suppositious) delicare, 51. 35 Th. ‘delicare’ ponebant pro dedicare; cf.
GI. Plac.16.11G. delicare: deferre, quod et ‘ dedicare’ dicebant pro commercio
litterarum]. This delicare for dedicare is either an etymological coinage to
explain delicatus, or the Old Latin delicare, to explain, inform, which was
mentioned above. Modern etymologists have adopted a similar course to
explain some difficult words like mutilier, miles, silicernium, lidus (for other
examples, see Wharton: On Latin Consonant Laws, in the Phil. Soc. Trans.
1889 ; and Conway, Indogerm. Forsch. ii. 157). But the evidence rather points
to this J-like pronunciation of d@ having asserted itself at the beginning of the
literary period, but not having gained admission into the literary language,
except in the case of a few words where the / was supported by analogy or
other causes. (Latin baliolus, from bddius, is a ‘ghost-word,’ the true reading
in Plaut. Poen. 1301 is batiolus, the old spelling of bajulus).

Whether an original 1 is ever represented by din Latin is doubtful. The
mispronunciation alipes for ddipes (Prob. App. 199. 3K.) was probably influenced
by the Gk. dAega, but is no argument that the Greek and Latin words are
connected ; the O. Lat. sedda for sella, quoted by Ter. Scaur. 13. 14 K., seems to be
a grammarian’s coinage to illustrate the etymology from sédeo, though it might
possibly be the same dialectal sedda, with a peculiar form of d, which is still
heard in S. Italy and Sardinia, where every Italian Wis replaced by this dd-sound
(see ch. ii, § 85). (On medipontus and melipontus, see Keil on Cato R. R, iii. 5.)

§ 112. Lat.r for d. Priscian (i. p. 35. 2H.) tells us that the ‘antiquissimi’
288 THE LATIN LANGUAGE. [Chap. IV.

used arvenae, arventores, arvocati, arfines, arvolare, axfari, also arger; and other
grammarians mention arventum (Mar. Vict. 9. 17 K.), arvorsus, arvorsarius Vel.
Long. 71. 22K.), arveniet G1. Plac.), axferia( Paul. Fest. 8. 32 Th.; Gloss. ap. Léwe
Prodr. p. 13 vas vinarium quo vinum ad aras ferebant; cf. adferial, Gl. Cyr.).
In Cato we find arveho (e. g. R. R. 135. 7; 138); and on inscriptions arfuerunt,
arfuise and arvorsum (C. I. L. i. 196) in the S. C. de Bacchanalibus of 186 3. c.,
arvorsario beside advorsarium \i. 198), in the Lex Repetundarum of 123-2 B.¢.,
arvorsu (ix. 782). Apur (Mar. Vict. 9. 17 K.), apor (Paul. Fest. 19. 34 Th.), the
old form of apud, was probably a ‘doublet’ used before a word beginning
with g, », f (cf. apur fem in an old inscription from the Marsic territory, Zv.
I.1.1. 45); but quirquir in the augur’s formula for marking out a templum (Varro,
L. L. vii. 8) is a doubtful example ; and arduuitur (Legg. XII Tabb. 10. 7 Br.),
ar me Lucil. ix. 30 M. unlikely readings. The affinity of the sounds is seen
in the (dialectal) mispronunciation ‘in usu cotidie loquentium’ (Consent.
392. 15 K.) peres for pédes (cf. arvenire Diom. 452. 29 K.); and r (or else d, the
th-sound of our ‘this,’ ‘then’) takes the place of Italian d in the dialects of
S. Italy and Sicily, e.g. dicere and ricere for Ital. dicere in the Abruzzi;
Neapolitan rurece for Ital. dodici, and (like Consentius’ peres) pere for Ital.
piede, &c. Arger persisted in Vulgar Latin (Ital. argine, ‘a dam,’ Span. arcen,
‘a parapet.’) Arbiter, from the root gtet- of Lat. vito (0. Scand. at-kvaeda,
‘a decision’) seems to be dialectal (0. Umbr. adputrati, in Lat. ‘ arbitratu’) (ef.
monériila, the form used by Plautus, classical ménéditia with the termination of
acrédula, ficédula, querquédula, nitédila, alcédo ; and the glosses maredus for mddidus,
solerare for silédare, marcerat for marcidat (Lowe, Prodr. 352; Opusc. 142). The
change of d to r in O. Lat. arfuerunt (preserved in the formula scr. ARF.
scribendo arfuerunt), &e. was often mentioned by the native grammarians, who
regarded a reference of any r to an original d as a legitimate device in
framing etymologies. Thus Velius Longus (71.23 K.) derives aurictila from audio,
and méridies (older meridie Ady.) from médius and dies. This explanation of
meridies is probably right, the r being due to dissimilation (cf. Cic. Orat. xlvii.
157 jam videtur nescire dulcius. ipsum meridiem cur non ‘medidiem’; Varro,
L. L. vi. 4, says that medidies was the old form, and that he had seen it on
a Praenestine sundial: meridies ab eo quod medius dies. d antiqui, non r,
in hoc dicebant, ut Praeneste incisum in solario vidi), though a good deal may
be said for the derivation from merus (ef. mero meridie, Petr. 37. p. 25.1 B.), and
the Praenestine D seen by Varro may have been merely an old form of the letter
R, as LADINOD, LADINEI on all coins \c. 250 B.c.) of Larinum (modern Larino)
(C. I. L. i. 24) may show the Oscan D, the symbol of r, as R was of d. Isidore
(Orig. xii. 7. 69) similarly explains méviila : merula antiquitus ‘medula’ voca-
batur, eo quod moduletur ; and Varro (L. L. v. 110) derives perna ‘a pede.’

Modern etymologists explain in the same way gldrea (Gk. xAj5os), possibly
a dialectal form, simitur (ch. ix. § 8) beside simitu, and other words of doubtful
origin (see Wharton, Latin Consonant Laws, in Phil. Soc. Trans. 1889, on céreo,
ploro, &c.}. But the available evidence hardly allows us to ascribe any r to
an original d in a Latin word except before g, v, f. Cddiiceus, a loanword from
Gk s«npdxoy (Dor. xdp-) may owe its d to a fanciful connexion of the word with
cddicus. In Umbrian ar- is found for the Preposition ad in compounds
perhaps only before f-, v-, e.g. arveitu ‘advehito,’ axfertur (and arsfertur)
‘adfertor’ (von Planta,.i. p. 408). (On Lat. arcesso and other doubtful exx.
of ar- for ad-, see Schoell, xii Tabb., p. 81.)
§§ 112-115.] REPRESENTATIVES OF I.-EUR. SOUNDS. T, D, DH. 289

§ 113. tr for dr. Lat. citrus was the old form of Gk. #é5pos, for Naevius has
citrosa vestis, while cedrus is not found till Virgil ; Cassantra and Alexanter were
the old forms of Kaczdvdpa, ’Adééavdpos, according to Quint. i. 4. 16 [we find
Alixentrom (0. I. L. i, 50), Alixente(), Casenter(a) (i. 15301), on old inscriptions from
Praeneste]. So Gk. ‘Y8pois, -odvros, Lat. Hydruntum, is modern Otranto, Lat.
nittrio is for *noud-rio (§ 42) (Lith. nauda, ‘use,’ Goth. niutan, ‘to enjoy,’ Germ.
geniessen) ; taeter for *taed-ro- (cf. taedet) ; so titer, Gen. titris (Gk. b8pia) ; litra,
an otter, may be a malformation of *iéra (O. Ind. udra-, Lith. udra, Engl.
otter), The group dr is not found in Latin, except in quadru-, quadra, &c. (but
ef. triquetrus), The name Drasus (cf. Gloss ap. Lowe, Prodr. p.398 drusus: patiens,
rigidus, contumax) is declared by Suetonus (Tid, iii.) to be a Gaulish name :
Drusus, hostium duce Drauso comminus trucidato, sibi posterisque cognomen
invenit ; andruare and drua (Paul. Fest. 7. 15 Th.) are very doubtful spellings
of antruare (antroare) and trua (ib. 1. 17).

y

§ 114. DH. I.-Eur. dh became f/ in Latin, which in proximity
to 7 became 4; but in the middle of a word between vowels
d is found; e.g. L-Eur, *dhimo- (O. Ind. dhimé-, Gk. 6dpés,
Lith. déimai Pl.,O. Sl. dymit), Lat. fimus; I.-Eur. riidhro- (Gk.
é-pvpds, O. Sl. riidri; ef. O. Ind. rudhiré-), Lat. rer, stem
rubro-; I.-Eur. root bheidh- (Gk. wei@m for *peidw), Lat. fido.
This f, 4, d may have all three developed from an older @-sound,
the sound of our ¢é in ‘this,’ ‘that,’ ‘then;’ but whether Sicilian
Nitpa (for Lat. Zébra) is a survival of the dstage is uncertain.
The change of dr to fr, dr may be compared with the change of
sr to fr, br in tenebrae for *tenes-rae, &e. (§ 152).

I.-Eur. dh is in Umbro-Ose. f, whether initial or internal, e.g.
Umbr. facia, Ose. fakiiad ‘faciat, Umbr. rufra ‘rubra,’ Osc.
mefiai ‘mediae’ (von Planta, i. p. 451). This f for Lat. (internal)
a, 6 is found in some dialectal words, e.g. erefrare for cribrare,
like dialectal f for Lat. (internal) 4 from I.-Eur. bh (see ch. ii.

§ 83).

§ 115. Other examples of I.-Eur. dh. (1) Initial: I.-Eur. root dhé-, dhéx-,
‘to place, do’ (Gk. &0y-v, é-Onx-a, Arm. d-nem, ‘I place,’ Goth. ga-dé-ps, Engl.
deed, Lith. dé-ti, O. Sl. dé-ti, ‘to lay’), Lat. fdc-io, but with dh medial, condo
(fac- shows the weak grade of dhéx-); I.-Eur. root dhé-, dhéy-, ‘to suck,’ (O. Ind.
dh4yami, Arm. diem, Gk. 6700, O. Ir. dith 3 Sg. Pret., Goth. daddjan, ‘to
suckle,’ 0. Sl. d&te, ‘an infant’), Lat. félo, usually spelt fello, filius(§ 7). Simi-
larly Lat. forum (Lith. dvaras, ‘court,’ 0. Sl. dvort) ; Lat. foveo, favilla from the
I.-Eur. root dhegh-, ‘to burn’ (0. Ind. dah-, Lith. degt, Gk. réppa for *éppa,
ashes) [the original meaning of foveo was ‘to warm’ (see the dictionaries,
and cf. Paul. Fest. 60. 15 Th. a fovendo, id est calefaciendo) ; hence fdculum,
a fire-pan, e. g. Plaut. Capt. 847 foveri foculis ferventibus].

(2) Medial; from I.-Eur. root reudh-, ‘to be red’ (0. Ind. rohita-, ‘red,’ lohi-,

U
290 THE LATIN LANGUAGE, (Chap. IV.

‘metal,’ Gk. é-pev6w, O. Ir. ruad, ‘red,’ ‘strong,’ Gaul. Roudos, W. rhudd, Goth.
raups, ‘red,’ O. H. G. rost, ‘rust,’ O, SL. ridéti, ‘to blush,’ ruda, ‘ metal,’ rida,
‘rust’) come Lat. raudus, rédus and ridus, unshaped metal, riidis, unshaped, and
with b, besides riiber mentioned above, ritbeo, robus, robur, robigo, while réfus with f,
is dialectal ; from I.-Eur. root bhendh-,‘to bind’ (0. Ind. bandhana-, Gk. metopa
for *revOopa, mevOepés, Goth. bindan), Lat. offendia (§ 79) with d, while Lat.
lumbus shows b for dh (or dhw?) (O. SI. ledvija, O. Engl. lenden), and 6
appears for dh of the suffix dhlo- (see ch. v. § 26), in tri-bulum, stdbulum, &c. as
-bro- for I.-Eur. dhro- in crt-brum, vertt-bra, &e. ; I.-Eur. *médhyo- (0, Ind.
mddhya-, Gk. péo(c)os, Goth. midjis), Lat. médius ; I.-Eur. root aidh-, ‘to burn’
(O. Ind. édha-, ‘firewood,’ Gk. aiéw, O. Ir. aid, ‘fire,’ O. Engl. ad, ‘ pyre),’ Lat.
aedes, house, lit. ‘hearth.’ Similarly Lat. vidua (O. Ind. vidhava, O. Ir. fedb,
Goth. viduvs, O. Sl. vidova) ; fidélia (Gk. miOos) ; grdédus (Goth. grips) ; vddes
(Goth. vadi, Germ. Wette, Lith. vadtiti, ‘to redeem’) ; aber (O. Ind. idhar, Gk.
ovéap, O. Engl. ider) ; combréfum, a bulrush (cf. Lith. szvefidrai Pl.) ; barbda for
*farba (§ 104) (Engl. beard, O. Sl. brada, Lith. barz-da) ; arbos (O. Ind. ardh-,
‘to grow, thrive’) ; gldber (O. H. G. glat, ‘smooth,’ Engl. glad, Lith. glodis,
‘smooth,’ 0.81. gladtikt) ; verbum (Goth.vaurd, Neut., Lith. vafdas, ‘a name’).

§116. The Gutturals. There are three series of Gutturals, viz.

. +n Pg ki
(1) Palatals (in some languages Sibilants, e.g. O. Ind. Satém, Lith-
szimtas, O. SI]. sito, ‘hundred’ corresponding to Lat. centwm) ;
(2) Velars, better called Gutturals proper (Gutturals in all
languages, e.g. the onomatopoetic name of the cuckoo, O. Ind.
koka-, Gk. xékxvg, Lat. cucilus ; ef. Lith. kuktiti, ‘to ery cuckoo’) ;
(3) Velars with Labialisation, i.e. followed by a w-sound (in
some languages Labials, e.g. Hom. Gk. micupes, W. pedwar, Ose.
petora, ‘four,’ corresponding to Lat. guattuor). They are most
conveniently written, (1) k, &, &c., (2) k, g, &c., (3) q¥, gt, &e.,
while the symbols 4, g, &e. may be reserved for Gutturals
whose exact nature is doubtful. So far as Latin is concerned,
we might write (1) and (2) as 4, g, &., and (3) as kw, gw, &e..
for the same letter ¢ represents the k of centum and the k of
carpo (O. Ind. krp-), nor does the kw of eguus, &e. (O. Ind.
L .

agva-) present a different appearance from the qg* of quattuor,
sequor (O. Ind. sac-). J.-Eur. k and k became Lat. ¢ (4), & and
g Lat. g, Sh and gh became a guttural spirant, which was
written g with a consonant, elsewhere 4; I.-Eur. q# is Lat. qu
(which we might write fv, hw), g is Latin gu, gv (gw), which
became g before a consonant, but lost its g when initial just as
dj- (dy-) when initial became j- (y-), or as Teut. gw from I.-Eur.
gh* became w in ‘warm, ‘snow, &c. It lost its y also when
medial between vowels. I.-Eur. gh* became a guttural spirant
§ 116.) REPRESENTATIVES OF I.-EUR. GUTTURALS. 2Q1

followed by a w- sound, which was written with a consonant gw or
g, but elsewhere became iw, this Aw developing at the beginning
of a word into f [just as the Greek Aw-sound from I.-Eur.
initial sw- seems to have developed into some /-sound (ch. i. § 3)],
but in the middle of a word into v (w). (For examples see below.)

The I.-Eur. Gutturals offer considerable difficulty. It is not
only that we find occasionally the confusion, found with every
species of Mute, between Tenuis and Media (ch. ii. § 75), Media
and Aspirate, &e.; e.g. the guttural Media seems to replace the
Aspirate in Lat. Zégurrio (cf. lingo), from the root leigh-,‘ to lick ”
(Gk. Aciyw), Lat. adagio, prodigium, from the root agh-,‘ to say’
(O. Ind. ah-), Lat. ftgira (cf. fingo) from the root dheigh-, ‘ to
mould,’ (O. Ind. dih-, ‘ to smear,’) just as we find the labial Media
replacing the Aspirate in Gk. oréuSw beside doreupys. We
find also apparent confusion of one series of Gutturals with
another; thus in Greek (and perhaps in other languages) the
proximity of the vowel w~ seems to change a Guttural of the
third into a Guttural of the second series, e.g. yuvij! (Boeot.
Bavd, O. Ir. ban), AvKos (dial. Lat. Ziipus); and very often the
want of a cognate word in a language which treats one series
differently from another, prevents us from ascertaining to which
series a Guttural properly belongs, e.g. whether the ¢ of coddum
is a palatal or a true guttural. We are also confronted with an
apparently I.-Eur. dialectal change of q* to p, perhaps made in order
to avoid that similarity between two successive syllables which
was so sought after in Latin ($163). Thus the I.-Eur. word for
‘five’? may have been *qénq*é in one dialect (O. Ind. panca,
Gk. wévre, &e.), *q2énq*é in another (Lat. guinque, O. Ir. céic) ;
the root meaning ‘to cook, peq?- (O. Ind. pac-, Gk. wéoow),
qteq?- (W. pobi, with p- from I.-Eur, q*, Lat. cdguus), and even
qtep- (Lith. kept). What of Gk. dpro-«émos?). The same
explanation has been suggested for the q* of Lat. guercus beside
the p- of O. Engl. furh, Engl. fir, and for the p- of Goth. fidvor,
Engl. four, beside the q* of other languages, e.g. Lat. guattuor.

A Latin Guttural, to whatever series it belongs, combines
with a following s into 2, e.g. veai from vého (I.-Eur. gh), before
¢ becomes the group ct, e.g. vectus, actus, which in late Latin

1 So w is dropped before the u-sound in Engl. ‘ two.’
U2
292 THE LATIN LANGUAGE. [Chap. IV.

came to be pronounced /¢ (Ital. atto), before x, m becomes the
group gn, gm,e.g. tignus from ilex, or the loanword cygnus
from xv«vos. Initial gn- became x- at the beginning of the
second cent. B.c., e.g. zdtus. But dina does not stand for *luwena,
but for *Zucsna (Zend raox8na-, ‘shining,’ Pruss. lauxnos,
‘ stars’), as we see from the old form on a Praenestine mirror,
Losna (C. IL. i. 55), just as vélum, a sail, stands for *vex-/um
(O. SI. veslo, ‘a rudder’), as we see from the Diminutive form
vexilium, so that Gk. dAvyvos has been declared to represent
*vksvos (M. §. L. vii. gt). Another instance of Greek xv
appearing in Latin as » with long vowel is the (loanword ?)
aranea (Gk. dpdyvn). Ewdmen beside agmen, amb-dges (on
exagmen see Class. Rev. v. 294), contaémino beside contagium ;
[contrast propagmen (Enn. Ann. 587 M.) beside propago, and cf. sub-
témen from texo beside subtegmen, e.g. Virg. Aen. il. 483], have
been variously explained by hypotheses that have as yet failed
to establish themselves, such as (1) that the Guttural is assimi-
lated (like d@ in caementum from caedo) after a long vowel,
*eadg-men becoming *exdmmen, examen, while ag-men remains
agmen ; (2) that the Guttural is assimilated in the unaccented
‘syllable (under the earlier Accent Law), whence éxdmen but
dgmen; (3) that the unassimilated forms had originally a con-
necting vowel, e.g. dg(t)men) (cf. jugumentum, a joining, Cato,
R, R. xiv. 1 and 4). Another theory, that these forms add an
s to the final Guttural of the root, examen for *ex-ax-men from
*ags-men- (cf. *ax-la, ax-is), connects the forms with a known
law of Latin phonetics, and is preferable on that account.

The cognates of major, aio show g in Latin, eg. magnus,
migis, adagio, prodigium, but the guttural Aspirate in other
J.-Eur. languages (O. Ind. mah-, ah-); so it is better to refer
them to the ordinary law that Latin 2 may be dropped between
vowels, than to posit a new law that the guttural Media was
dropped before y. Major will thus come from an older *méhior
(cf. O. Ind. mahiyas-) ; piléium, fleabane, apparently from pélez,
a flea, may be dialectal, like Umbr. muieto P. P. P., with i for
palatalized g, beside mugatu Imperat.; brévis (Gk. Bpaxvs) will
represent an older trisyllable brehuis, and lévis (Gk. édaxds:
what of édadpds?) an older *lehuis, while pinguis (Gk. rats),
§§117-119.] REPRESENTATIVES OF I.-EUR. GUTTURALS. 293

where the Aspirate is preceded by the consonant », will represent
an older trisyllabic form *pinguis (see § 127).

Aspirate Tenues, which are difficult to trace (see § 95), have
been found in waguis (cf. O. Ind. nakha-, Gk. dvvx-, O. Ir. inge,
Lith. nagas, O. Engl. negel) (see B. B. xvii. 133), congius (O.
Ind. Sarkha-, ‘a shell, Gk. «dyn, Lett. sence). On dat. cl for
I.-Eur, -tl-, e.g. pde(u)lum see § 105 ; on ce for te, e.g. ae for *ate,
atgue, hoc for *hod-c(e), § 109; on the mispronunciations ss, sc
for 2, e.g. cowim, ascella, ch. ti. § 117.

In Umbrian the combination kt, when due to Syncope, seems
to be differently developed according as the k represents on the
one hand an I.-Eur. q® or on the other an L.-Eur. k (or k); in
the former case it becomes kt, e.g. fiktu (Lat. figito for jivito,
ch. viii. § 7), in the latter, it, e.g. deitu (Lat. dicito). In Ose.
both are kt, e.g. fruktatiuf ‘fru(v)itationes, factud <facito,’
while I.-Eur. q*t (not due to Syncope) is Umbro-Osc. kt, and
I.-Eur. kt, kt (not due to Syncope) is Umbro-Ose. ht (for
examples see Buck, Vocalismus Osk, Sprache, p. 145).

§ 117. x for Guttural with s. Vexi from vého (I.-Eur. root wegh-, ‘to carry’);
panxi (cf. Gk. myyvupe) ; finxi from jfingo (I.-Eur. dheigh-) ; nix for *(s)nigh3s ;
coxi from céquo ; lixa, pro-lixus, é-lixus from liquor for *vliquor (O. Ir. fliuch, ‘ wet,’
W. gwlyb). Before most consonants « became s, e.g. sesceni for *sex-cent, (as
after r in mers for merx, sparsi, tersi, &c., § 158), but not before ¢ in dexter, dextra,
settus, &c. (but lustro, to illuminate, dlustris, &c. from *lucs-tr- ; on Sestius, mistus,
see ch. ii. § 125), while before m, », 1 it was (like s for I.-Eur. s, e. g. qudlus
for *quas-lus, cf. quisillus, § 151) dropped with ‘Compensation,’ e. g. palus for
*pax-lus (Dim. paxillus) ; dla, wing, shoulder, for *ax-la (Dim. azilla), like axis, an
axle, from the root ag- of dgo (O. H. G. ahsala, ‘shoulder,’ W. echel, ‘an axle,’
Engl. axle) ; télum, for *tex-lum, from texo, to shape (O. H. G. dehsala, ‘ an axe,’
O. Sl. tesla) ; téla, for *tex-la, from texo, to weave (Ter. Heaut. 285 texentem
telam) ; télus, for *tax-lus (Dim. tuwillus) ; soni for *sea-ni; aula, Dim. aucilla.

§ 118. ct for Guttural with t. Vectus, with vectis, a lever, from vého (Eh);
panctum and pactum from pango (§?) ; fictum from fingo (gh) ; luctus, grief, from
lageo (8); coctum from cdquo (q%); nicto, to wink; cf. co-niveo (ghx). After
a consonant c was dropped in course of time (§ 157), e. g. fortis, in O. Lat. forctis
{in the XII Tables forctes was the name given to the loyal neighbours of Rome,
sanates to those who had swerved from their loyalty, but had returned to it,
Fest. 524. 15 Th.; cf. Paul. Fest. 59. 26 ‘forctes’ (leg. ‘forctis’) frugi et
bonus, sive validus]; quintus, in the older spelling quinctus, the older form
being long retained in the names Quinctilius, Quinctius, Quinetilis, &e. (see
Georges, Lex. Worlf.s. v.). On Vulg. and Late-Lat. t# for ct, v. g. brattea, see
ch, ii. § 95, and cf. Rhein. Mus. xlv. p. 493.

§ 119. gn, gm foren,cm. Like ilignus from tex are larignus from liria, salignus
from sdliz; similarly segmentum from séco ; dignus, usually explained as *dec-nus
294 THE LATIN LANGUAGE. [Chap. IV.

from décet, but better as *dic-nus (O. Scand. tiginn, ‘high-born,’ tign, ‘rank ’)
a P.P. P. NO-stem from the root deik- of dicere, indicare, &e. Aprunus is a late
spelling of aprugnus; aprinus is an entirely different formation (see Georges,
Lex. Wortf. s.v.). After a consonant the Guttural is dropped, e. g. quernus for
*querc-nus from quercus ; farnus for *farg-nus, a byform of fraxinus, though notin
Compounds with ad, &c.,e. g. agnosco. When the consonant is a Nasal, it seems
to combine with the following Nasal, e. g. quini for *quinc-ni, so that contamino
might stand for *con-tang-mino as well as for *con-tags-mino. Compounds with
in, con show different spellings ; in- with gnosco gives ignosco, con- with gnosco
both cognosco and conosco, as con- with necto gives cénecto, &e. (ef. ch. ii. § 130).

Initial Latin gn became at the beginning of the second cent. B.c. n (as in
Engl. ‘gnat’), v. g. ndsco, older gnasco (gn-), ndtus, older gnatus (Gn-) (gnatare
madonojoa, C.G.L. ii, 35. 10), nixus, older gnixus (kn- ; ef. O. H. G. hnigan,
Germ. neigen), narus, older gnarus (gnaritur yvwpifera, C. G. L. ii. 35. 12), navus,
older gnavus, Naevius (cf Gk. Naos on the Mon. Anc.) beside Gnaeus (cf. Gnaivod,
Abl., on a Scipio epitaph, C.J. LZ. i. 30). For instances of the older forms, see
Georges, Lex. Wortf. s. vv., and Lowe, Prodr. 354: e.g. gnoscier on the 8. C. Bacch.
of 186 B.c., C.I.L. i. 196. 27, but nationwm, noverit (beside gnatus Part.) on the Lex
Repetundarum of 123-122 B.u., i, 198; natus Part. on the Sent. Minuciorum
of 117 B.c., i. 199. Plautus and Terence use, as a rule, natus for the Participle,
gnatus for the Substantive ; the g of these forms appears in the compounds
agnosco, agnatus, ignarus, ignavus, &e., and strangely also in agnomen, cognomen ;
for nomen (so in S. C. Bacch.) was not originally *gnomen (cf. O. Ind. nama-,
and for the various forms of the root, see B. B. xvii. 132). On crépusculum and
Gk. xvépas, see § 80.

§120. Lat. h dropped between vowels. Bimus for *bi-himus, lit. ‘two winters
old,’ like Engl. twinter, a two-year old beast (ef. xéuapos, goat, lit. ‘a winter
old,’ Scotch gimmer, a yearling lamb) ; nil (so always in Plautus), for nihil,
nihilum from né and hilum (cf. Enn. A. 8 M. nee dispendi facit hilum ;
Lucil. xiv. 11 M. hilo non sectius vivay; Luer. iii. 830 nil igitur mors est ad nos
neque pertinet hilum), hilum being explained as ‘ quod grano fabae adhaeret’
(Paul. Fest. 72. 10 Th.) ; praebeo for praehibeo ; cors for cdhors. On the h of dhénus
(aenus) for *iyés-no-, and on the question whether h was used in vehemens,
&e. to indicate a long vowel (as in Umbrian, e.g. comohota ‘commita’),
see ch. ii. § 56.

§ 121. Dialectal f for h. In Spanish, Latin f has become h, e.g. hablar, ‘to
speak’ (Lat. fabulari, O. Lat. fabulare), and an interchange of h and fshows traces
of itself in the dialects of Italy. We find the form fusena for hiséna ascribed to
the Sabine dialect by the grammarians (Vel. Long. 69.8 K.), along with fircus
(ef. the name of a citizen of Reate mentioned by Varro, Fircellius) and fedus.
Similar forms roughly classed by the grammarians as ‘Old Latin’ we may
believe to have been dialectal, e. g. fordewm for hordewm, folus for hélus, fostis for
hostis, fostia for hostia, &c., though some of them may be mere coinages to
strengthen the argument for the spelling with h- (see Quint. i. 4. 14; Ter.
Seaur. pp. 11, 13 K.; Vel. Long. p. 81 K.; Paul. Fest. 59. 21 Th. &c.).
A Faliscan inscription has foied for hodie (Not. Scav. 1887, pp. 262, 307): foied
uino pipafo kra karefo ‘hodie vinum bibam, eras carebo,’ but a Sabine inscrip-
tion has hiretum, apparently from the root Bher- (? gher-) (Osc. heriiad, Gk.
xaipw, &e.), and Ter. Scaurus (13. 9 K.) quotes haba (Lat. faba, O. Sl. bobu,
§§120—-128.] REPRESENTATIVES OF I.-EUR. SOUNDS. K, 6, GH. 295

I.-Eur. bh-) as Faliscan. (See von Planta, i. p. 442; Lowe, Prodr. p. 426; and
on the interchange of f and h in Etruscan inscriptions, Pauli, Altitalische
Forschungen, iii. p. 114). Lat. fel has been explained as a dialectal form for
*hel (of. Gk. xéAos), and fovea for *hovea (Gk. xed) (cf. the gloss ‘fuma’ terra,
C. G. L. v. 296. 50).

§ 122. The Palatal Gutturals: K, G, GH, KH. These
were in Latin, as in Greek, Celtic, and Teutonic, guttural sounds,
while in the Asiatic languages and Slavonic they were sibilants.

K. I.-Eux. k, Lat. c, is seen in I.-Eur.*kmtom,‘ hundred’ (O. Ind.
katém, Gk. é-xardv, O. Ir. et, W. cant, Goth. hund, Lith. szimtas),
Lat. centum ; *swéktiro-,‘ step-father’ (O. Ind. SvéSura-, Arm.
skesur F., Gk. éxupds, O. Corn. hwigeren, hweger F., Goth. svaihra,
Germ. Schwiher, Lith. szeszuras), Lat. sdcer ; *3icta(u), ‘eight’
(O. Ind. a&tau, aSta, Gk. dxré, O. Ir. ocht, W. wyth, Goth. ahtau,
O. Engl. eahta, Lith. asztiini, O. Sl. osmi), Lat. dct. I.-Eur.
kw was merged in gu, the representative of I.-Eur. q*?; thus
I.-Eur. *ékwo-, ‘horse’ (O. Ind. déva-; ; ef. Lith. aszva, ‘ mare,’
&e.) is Lat. equus.

I.-Eur. k is Umbro-Ose. k, e. g. Umbr. kletram Acc.,‘ a litter”
(Gk. kAivw, &e.), Kluviier, Osc. Kluvatiium from the root kleu-
(Lat. clweo, Gk. xdéos, &e.), (von Planta, i. p. 326). In Umbrian
k (whether from I.-Eur. k or k) was palatalized before e, i, and
was written in the native alphabet by a sign which we conven-
tionally express by ¢, in the Latin alphabet by # (sometimes s),
e.g. gersnatur, esa (Lat. céna, Osc. kersna-) (ib. p. 359).

§ 123. Other examples of L.-Eur, k. From I.-Eur. root weik-, ‘to enter”
(O, Ind. vié., ves. M., ‘a tent,’ Gk. Fotxos, Alb. vis M., ‘a place,’ Goth. veihs,
‘a village,’ Lith. vosziti, ‘to be lodged,’ O. SI. visi, ‘ fare ), Lat. vieus ; I.-Eur.
*yiiwnko- (0. Ind. yuvasé., Gk. ‘Yax-w6os, O. Ir. dae, W. ieuanc, Goth. juggs),
Lat. jiivencus ; I.-Eur. *krd-, ‘heart’ (O. Ind. grad- dha., ‘confidence,’ Arm. sirt,
Gk. xapdia, O. Ir. cride, Goth. hairts, Lith. szirdis, 0. Sl. sridice), Lat. cor, crédo ;
I.-Eur. root kleu-, tbe hear’ (O, Ind. sru-, Gk, «Avw, O.-Ir. cloor, cluinim, W.
clywed Inf, Goth. hliu-ma, ‘hparing,’ O. SI. sluti, ‘to be famous’), Lat. clueo ;
I-Eur. *pirko- (Gk. wéptos, O. Ir. ore, O. Engl. fearh, Engl. farrow, Lith. pa¥szas,

O. SL. prase), Lat. porcus ; I.-Eur. ki-, a Demonstrative Pronoun-stem (Arm. -s,
ce -«t, Alb. si-, O. Ir. e6, Goth. jhi-mma, Lith. szis, O. Sl. si), Lat. cis, citer, citra ;

I.-Eur. klei-, ‘to lean’ (0. Ind. bri, Gk. «Atv, O. Ir. cloen, ‘awry,’ Goth. hlains,
‘a hill,’ Lith. sztaitas), Lat. ac-clinis, chvus ; I.-Eur. +dékm, ‘ten’ (0. Ind. dasa,
Arm, tasn, Gk. 5éa, O. Ir. deich, W. age, Goth. taihun, Lith. dészimt, O. SI].
deseti), Lat. décem; I.-Eur. root dciks ‘to point, ee (O. Ind. di8-, Gk. Setevupi,
Goth. gateihan, ‘to proclaim),’ Lat. dione, indicare ; I.-Eur. root peek, ‘to ask’
(O. Ind. prag-n4-, ‘a question,’ O. Ir. imm-chom-arcim, ‘I ask,’ 0. W. di-er-
296 THE LATIN LANGUAGE. (Chap. IV.

chim, Goth. fraih-na, Germ. frage, Lith. praszat, O. Sl. prog), Lat. précor.
Similarly Lat. craébro for *crasro (ef. Lith. szirszi, O. SL srtigeni, Engl. hornet) ;
dicus, acer’, &c. from the root ak-, ‘to be sharp’ (O. Ind. aéri-, ‘edge,’ Arm, aseAn,
‘a needle,’ Lith. asztris, ‘sharp,’ O. Sl. ostrii, &.) ; ctrébrum for *certs-rum (ef.
O. Ind. Siras-, N., ‘head,’ Gk. «dpa, &c.).

§ 124, L-Eur. kw. Lat. combretum, bulrush, from the stem kwéndhr., seen
in Lith. szvefidrai Pl., has been compared to séror for *swésor- (§ 68) ; but it
more probably shows the O-grade of the stem, *quombr-etum (cf. O. Seand. hvdnn
‘angelica’), with reduction of quo- to co- as in cilo for older quolo (ef. § 137).

§125. G. Of L.-Eur. &, Lat. y, we have examples in the I.-Eur.
roots gen-, 6nd-,‘ to know, learn’ (O. Ind. ja-na-mi, jfiaté-, Arm.
can-eay Aor., Gk. yi-yvd-cxw, yvwrtds, O. Ir. gnath, ‘ accustomed,’
W. gnawd, O. Engl. enawan, Engl. to know, Lith. zin-dti, O. SL.
zna-ti), Lat. gnd-sco, gnd-tus; melé-,‘ to milk’ (O. Ind. myj-, ‘ to
wipe off, Gk. duéeAyo, O. Ir. bligim, O. Engl. melee, Lith.
métzu, O. SI. mltiza), Lat. mulgeo.

I-Eur. &w would be indistinguishable in Latin from I.-Hur.
gt; thus avidus (cf. Gk. typés), if connected with the root weé-,
‘to be strong’ (cf. Gk. dyiujs), shows v between vowels for gw.

In Umbro-Ose. I.-Eur, & is g, e.g. Umbr. ager ‘ager,’ Ose.
aragetud ‘argento’ Abl. (von Planta, i.p. 329). In Umbrian g,
whether I.-Eur. & or g, suffers before e, i palatalization, and is
written i, e.g. muieto, P. P. P. of a verb whose Imperat. 3 sg. is
nugatu (von Planta, i. p. 372).

§ 126. Other examples of I.-Eur. & I.-Eur. geus-, ‘ to taste’ (O. Ind. jus.,
Gk. yedw for *yevow, O. Ir. to-gu, ‘I choose,’ Goth. kiusa), Lat. giistus ; I.-Eur.
reg-, ‘to stretch, rule’ (O. Ind. 1/-, raj- or raj-an-, ‘king,’ 6-péyw, O. Ir. rigim,
rig Gen., Gaul. Catu-riges, W.rhi, Goth. uf-rakja, Lith. r§zau), Lat. régo, regis
Gen. ; I.-Eur. gen-, ‘to beget’ (O. Ind. jan-, jinas-, Arm. cin, Gk. yéyvopar, yévos,
O. Ir. génar Pft., gein, W. geni, genid, Goth. kuni, ‘race,’ Engl. kin), Lat.
gi-gn-o, génus. Similarly Lat. argentwm (Zend er*zata-, Arm. arcat'; cf. O.
Ind, drjuna-, ‘white,’ Gk. dpyjs); Lat. glos (Gk. ydAows, O. Sl. zltiva) ; Lat.
dgo (O. Ind. aj-, Arm. acem, Gk. dy, O. Ir. ag-, O. Scand. aka), Lat. granum
(O. Ind. jirné-, ‘crushed,’ Goth. kaurn, Lith. Zirnis, ‘a pea,’ 0. SL zrino) ; Lat.
génu (O. Ind. janu, Arm. cunr, Gk. -yévv, Goth. kniu N., 0. Engl. enéo N.) ; Lat.
végeo, végétus (O. Ind. vaj-, ‘to be strong,’ Zend vaz-, Gk. iyejs, Engl. I wake,
Germ. wacker) ; Lat. gélu (O. Sl. 2lédica) (on é instead of 61, see § 10).

§ 127. GH. L.-Eur. Sh is in Latin 4, but g before or after
a consonant, e.g. I.-Eur. *ghirto- (Gk. xdpros, O. Ir. gort, Lith.
zatdis), Lat. hortus ; I.-Hur. root wegh- (O. Ind. vah-, Gk. dyos,
Goth. ga-viga, Germ. be-wege, Engl. waggon, Lith. wezu,
§§ 124—129.] REPRESENTATIVES OF I.-EUR. SOUNDS. K,G,GH. 297

O. Sl. veza), Lat. vého; I.-Eur. root angh- (O. Ind, éhas-,‘ need,’
Arm, anjuk, ‘ narrow,’ Gk. éyxw, O. Ir. cum-ung, ‘ narrow,’ Goth.
aggvus, O. Engl. ange, Germ. eng, O. SI. azikti), Lat. ango, angor,
angustus ; Lat. grando (O. Ind. hradini-). A good example of the
rule for g and 2 is mingo beside mejo for *meiho, from the root
meigh- (O. Ind. mih-, Lith. mjzai, Gk. 6-pix¢éw).

I.-Eur. Shw was in Latin merged in IJ.-Eur. gh*, e.g. I.-Eur.
*hwér-, a wild animal’ (Gk. 6p, Thess. deip, Lith. Zvéris, O. S1.
zvéri), with which is connected Lat. férus, Fem. féra, a wild
animal, (But Engl. deer, Germ. Thier, Goth. dius point to some
I.-Eur. original like *dheusd-, cf. Lat. firo for *fuso ?).

I.-Eur. Sh is in Umbro-Osce. h, as in Latin, e.g Umbr.
hondra, Osc. huntro- from the root of Lat. Zémus (von Planta,
i, p. 436). On h for f in Sabine fasexa ‘harena,’ &c., see
above § 121.

§ 128. Other examples of I.-Hur. gh. Lat. hiems (O. Ind. himé-, Arm.
jiun, Gk. yay, xetudv, O. Ir. gam, O. W. gaem, Lith. Zéma, O. SI. zima) ;
Lat. hiimus (O. Ind. jma- F., Gk. xapai, Lith. zémé, O. Sl. zemlja) ; Lat. hélus,
older helus, helvus (§ 10) (O. Ind. hari-, ‘yellow,’ O. Ir. gel, ‘white,’ 0. Engl. geolo,
Engl. yellow, Lith. Zelu, ‘I grow green,’ O. Sl. zelije, ‘ vegetables’); Lat. lingo
(O. Ind. lih-, Arm. lizum, Gk. Acixw, Goth. bi-laigd, Lith. léziu, O. Sl. liza) ;
Lat. (A)anser (O. Ind. hasd-, Gk. xqv, O. Ir. géis, ‘a swan,’ Engl. goose, Lith.
Zasis); Lat. fingo (O. Ind. dih-,‘ to smear,’ déhi,‘a wall,’ Arm. dizem, Gk. reixos,
O. Ir. dengaim, ‘I fasten,’ Goth. deigan, ‘I mould,’ daigs, ‘dough,’ Germ.
Teig) ; Lat. hiare (O. H. G. gién, Engl. to yawn, Lith. -Zidti, O. Sl. zijati).

§ 129. The Gutturals Proper: K, G, GH, KH. These
appear as Gutturals in all the I.-Eur. languages. The fact that
the I.-Eur. onomat8poetic name for the cuckoo shows this form
of Guttural (O. Ind. kéka-, Gk. xéxxvg, Lat. cucilus, O. Ir. cuach,
W. cog; ef. Lith, kukuti, ‘ to ery cuckoo’) indicates what sort of
Guttural it was.

K. I.-Eur. k is Lat. ¢, e. g. I.-Eur. root kert-, ‘to plait’ (O. Ind.
ert-, kdta-, ‘ mat,’ Gk. xdptados, basket, Goth, haurds F., ‘ door,’
Engl. hurdle), Lat. erates, cartilago ; I.-Eur. kerp-, * to cut, reap’
(O. Ind. krpana-,‘ a sword, Gk. xapzds, fruit, Engl. harvest, Lith.
kerpu, ‘I cut’), Lat. carpo (on ar, see § 3).

In Umbro-Ose. I.-Eur. k remains as in Latin, e.g. Umbr.
kanetu ‘canito’ (von Planta, i. p. 327), though in Umbrian k
suffers palatalization before e, 1 (see above § 122).
298 THE LATIN LANGUAGE. [Chap. IV.

§ 130. I.-Hur.k; other examples. Lat. cruor, cridus (O. Ind. kravi8- N., ‘raw
meat,’ Gk. «péas, O. Ir. cri, ‘gore,’ W. crau, Lith. kratijas, O. Sl. krivi, 0. Engl.
hréaw, ‘raw ’); Lat. collis (Goth. hallus M., Lith. kdInas; cf. Gk. coAwvds) ; Lat.
clavis, clavus (Gk. «Anis, O. Ir. cld M., ‘a nail,’ Germ. schliessen, O. Sl]. kljudi,
‘a hook, a key’); Lat. ancus, wncus (O. Ind. araka-, Gk. d-yxmy, dyros, O. Tr. cath) ;
Lat. cdéipio (Arm. kap, ‘a fetter,’ Gk. «a7, O. Engl. heeft, ‘ captive,’ Lett. kampu,
‘T seize’) ; Lat. coxa (O. Ind. kakia-, O. Ir. coss, ‘the foot,’ W. coes, ‘the leg,’
M. H. G. hahse, ‘bend of knee, hough’); Lat. cia (O. Ind. kipa-, ‘a pit,’ Gk.
xia); Lat. séco (O. Scand. sigdr, ‘a sickle,’ O. Engl. sage, ‘a saw,’ O. SI. séka, ‘1
cut’), O. Lat. clepo (Gk. «Aém7w, Goth. hlifa, Engl. shop-lifter, Pruss. au-klipts,
‘hidden’); Lat. vinco (O. Ir. fichim, ‘I fight,’ Goth. veiha, Engl. wight, Lith.
véka, ‘strength,’ ap-veikit, ‘I compel’); Lat. scando (O. Ind. skanda-mi,
‘I spring,’ Gk. cxavddéAnOpov, a springe, O. Ir. ro-sescaind, ‘he sprang’).

$131. G@. Of I.-Eur. g, Lat. gy, examples are: I.-Hur. root
gar-, ‘to shout’ (O. Ind. gr-, Gk. ynptw, O. Iv. gair, ‘a shout,’
W. gawr, O. H. G. chirru, ‘I shout,’ Lith. gaisas, ‘ noise’), Lat.

arrio; .-Hur. *ytigo-, ‘a yoke’ (O. Ind. yugé-, Gk. Cydv,
g YUgO-, y yuga-, Y.
Goth. juk, O.S]. igo), Lat. yuigum.

I.-Eur. g is g also in Umbro-Ose. (von Planta, i. p. 330) ;
but an Umbrian g, as we have seen, is palatalized (written i)
before the vowels i, e, &c. (§ 125).

§ 132. Other examples of I.-Eur. g. Lat. gris (Arm. ktunk, Gk. yépavos,
Gaul. Tri-garanus, W. garan, O. Engl. cran, Lith. gérvé, O. Sl. %eravi) ; Lat.
tégo (O. Ind. sthdga-mi, Gk. oréyw, oréyos, and réyos, O. Ir. teg, O. W. tig,
Engl. thatch, Germ. Dach, Lith. stégas) ; Lat. augeo (O. Ind. djas-, ‘strength,’
Gk. aigavw, O. Ir. dg, ‘entire,’ Goth. auka, ‘I multiply,’ Engl. eke, Lith.

augu, ‘I grow’), O. Lat. @-rago, of which e-ructo is the Iterative form (Paul.
Fest. 58. 30 Th.), used by Ennius, Ann. 593 M. :

contempsit fontes quibus ex erugit aquae uis,

comes from the I.-Eur. root reug- (Gk. épevyoua, Lith. rigiu, O. Sl. rygaja).

§133. GH. I.-Eur. gh, like J.-Eur. &h, became 4 in Latin,
except before or after a consonant, when it became gy. Examples
are: L.-Eur. *ghdsti- (Goth. gasts, Engl. guest, O. Sl. gosti),
Lat. hostis, hos(ti)pes (O. Lat. hostis, ‘stranger, Varro L. L. v. 3);
I.-Eur. root ghred- (O. Ir. ingrennim, ‘I pursue,’ Goth. grips, ‘a
step, O. SI]. greda, ‘I come’), Lat. grddior (on a, see § 3), gradus.

In Umbro-Ose. also I.-Eur. gh is h (von Planta, i. p. 438).
On f for h in some dialects, such as the Sabine, see above, § 121.

§ 134. I.-Hur. gh: other examples. Lat. pré-hendo (Gk. xavddvw, Alb. fen,
‘T find,’ gendem, ‘I am found,’ Goth. bi-gita, Engl. I get); Lat. hordewm (Arm.

gari, Germ. Gerste); Lat. haereo (Goth. us-gaisja, ‘I frighten,’ Engl. gaze,
Lith. gaiszti, ‘I tarry’).
§§180-135.] REPRESENTATIVES OF I.-EUR. SOUNDS. Q*,G*,GH*. 299

§ 185. Velar Gutturals with Labialisation. These appear
as Gutturals in some languages, and as Labials in others,
and show this divergence even on Italian soil, e.g. Umbr.,
Ose. pis, Lat. guis. (On the Italic treatment of the Gutturals
of this series, see von Planta, i. pp. 331 sqq.). Q* I-Eur.
q* is Lat. gu (but goi for classical gui on the Dvenos inscrip-
tion). Before ~ we find ¢, e.g. séciitus from seguor, a change
which may have been very ancient (see § 116). Before o this
gu, though often retained in writing, seems to have come to
sound like ¢; hence guéquo-, a cook, was written coguo- as well
as guoguo-, and on the other hand the Preposition cum, older
cém [for kom or kom (Ose. kim, Umbr. -kum)], was written
quom till the time of the Gracchi, and the P. P. P. of occilo, from
a root kel- or kel- (W. celu), appears with the spelling oguoltod
on the S. C. de Bacch. of 186 B.c. (C. LZ. Z. i. 196). When in
the eighth cent. a. u. c. 0 before a final consonant, came to be
universally changed in spelling (see § 20) to #, we find the spelling
quo (guo) replaced by cu (gu), so that éguos became ecus (Gen.
equi), quoquos or coguos became cocus (Gen. cogui). The gram-
marians of the first cent. a.D. were puzzled by the want of
analogy between ecus Nom.,and equi Gen., &c., and reconstituted
the Nom. as eguus, &c. Instances of I.-Eur. q* in Latin are:
I.-Eur. root seq* (O. Ind. sac-, Gk. rouat, O. Ir. sechur, Lith.
seki), Lat. seguor ; I.-Eur. root leiq?- (O. Ind. ric-, Arm, e-lik’,
‘he left, Gk. Aelaw, O, Ir. lécim, Goth. leihva, ‘I lend,’ Germ.
leihe, Lith. lékt), Lat. 4inguo ; I.-Eur. *q*i- (O. Ind. -cid Neut.,
Gk. ri for *r8, O. SI. &i-to ‘what?’), Lat. guid Indef.; I.-Eur.
*q®8 (O. Ind. ca, Gk. re, O. Iv. -ch, W.-p, Goth. -h), Latin -gué.
Before a consonant this gu became ¢, e. g. sdcius, older *socyo-
from the o-grade of the root seq*, with the adjectival suffix -yo-
(O. Ind. saicya-, Gk. d-oco-nryp, with. oo for xy); dctilus, older
*oclo- from the root oq2-, ‘to see’ (Gk. doce, with oo for xy, dupa
for *é-ya, Lith. akis, O. Sl. oko) with the suffix -lo.

I.-Eur. q*w has been postulated for the initial « (v) of wz
(Ose. puf, Umbr. pufe), a, aitz (cf. Osc. puz, Umbr. puze) (with
cu in the middle of a word, e. g..s2-ciibi), viipor (Lith. kvapas),
in-vitus and in-vito (Pruss. quaits, ‘ will’; Lith. kvécziu, ‘I
invite’), &c. (see K. Z. xxxil. 405).
300 THE LATIN LANGUAGE, (Chap. IV.

In Umbro-Ose. I.-Eur. q* is p. (On the date of the change,
see von Planta, i. p. 331). Latin pdpina, liipus, &c. are dialectal,
just as Pontius and Pompeius are the dialectal names correspond-
ing to Lat. Quintius, and Petreius to Lat. Quartius. (A full
list of examples in von Planta, /. c.).

§ 136. I.-Hur. q?, Lat. qu: other examples. Lat. quattuor (O. Ind. catvaras,
Arm. éork’, Ion. Gk. réocepes, Aeol. Gk. wécupes, O. Ir. cethir, W. pedwar,
Gaulo-Lat. petor-ritum, Lith. keturi, O. Sl. éetyrije) ; Lat. quingue (O. Ind.
pafica, Arm. hing, Gk. wévre, O. Ir. coic, W. pump, Lith. penki) ; Lat. qui,
quam, &e. (O. Ind. ké-, Gk. ré6ev, 77, O. Ir. cia, W. pwy, Goth. hvas, hvé, Lith.
kas, O. SI. kii-to).

§ 137. c for qu. (1) Before uw; arcus (Goth. arhv-azna, ‘an arrow,’ O. Engl.
earh), beside arquitinens, arquites, the old word for sdgittarii (Paul. Fest.
15. 32 Th.) ; the change to -cu- appears to be Italic and not merely Latin, if
Umbr. arélata- (cf. Paul. Fest. re. 15 ‘arculata’ dicebantur circuli, qui ex
farina in sacrificiis fiebant) comes from arcus ; from quinque come quincuna,
quincuplex ; from sesque comes sescuplus (but cf. Liéwe, Prodr. p. 403).

(2) Before a consonant: Lat. née for néqué, ac from *atce for atque, with
Syncope of -é before an initial consonant (ch. iii. § 36) ; torciilus from torqueo ;
céctilum from céquo. On ct for q}-t, x for q'-s, see § 116.

(3) Before 0: célo (older quolo: we have golunt in the Ambrosian Palimpsest
of Plautus, Pseud. 822, and quolundam on the inscription of the Faliscan
‘collegium coquorum,’ which also has ququei for cogui, Zv. I. I.I. 72) beside
inquilinus ; cdlus, a distaff (Gk. méAos, an axle). The fact that quo (I.-Eur. q'o
or kwo) had come to be pronounced like co (I.-Eur. ko or ko) explains why qu
seems not to offer the same resistance to the Brevis Brevians law in the Early
Poets when it precedes 0, as when it precedes other vowels, e.g. cogud but
only logut (ch. iii. § 42). But the indiscriminate spelling of every quo as co is
not found till the fifth cent. a.p., e.g. cof, corum, condam, locor (see Bersu, die
Gutturalen, p. 90) and the analogy of the other cases and persons kept quo as
the spelling in the Nom. Sg. of equos, &c., and the 3rd Pl., sequontur, &c., until
the 0, hitherto preserved in spelling by the preceding u, became in the eighth
cent. a.U.c. u (ch. iii. § 17), when ecus, secuntur were adopted as the proper
spelling. In words where the analogy of other forms played no part (e.g.
sesconciam, C.I. L. i. 1430) the spelling co is found much earlier. Similarly the
first syllable of the stem coquo- shows co earlier than the second syllable ; we
have qu- however in the older period (e. g. in all the MSS. of Plaut. Pseud. 382 ;
ef. ququei on the inscription of the Faliscan ‘collegium coquorum,’ Zv. I. I. I.
72). Puns are unsafe evidence of pronunciation ; but the punning reply may
be quoted of Cicero to the cook’s son who asked for his vote: ego ‘quoque’
tibi favebo (Quint. vi. 3. 47). The spelling equus, &c. was instituted by Velius
Longus in Trajan’s time. (On this transition of orthography -quo-, -cu-, -quu-,
see Bersu, die Gutturalen, who quotes a large number of instances of these
spellings, as also of the use in the time of the Gracchi of g for ¢ before u, e.g.
oqupare, pegunia, &c., and has collected those passages of the grammarians
which bear on the subject. A list of the instances of the spelling quom for
cum in the MSS. of Plautus is given by Probst, Gebrauch von ‘ut’ bei Terenz,
p.178n.) I.-Eur. qué did not, as is often stated, become qué, cd in Latin (as
§§136—140.] REPRESENTATIVES OF I,-EUR.SOUNDS.Q?*,G",GH*. 301

it did in Celtic, e.g. Ir. cic, W. pump, but Lat. quinque) ; and though -wé-
after other initial consonants appears as 3 in Latin sicer (swek- -), sdror {swés-),
&e., it probably did not after a palatal ; for the evidence points to kw, gw,
&c. having been merged in q}, gin Latin. Combretum, bulrush, may show
the O-grade of the stem laanehr of Lith. szvendrai Pl. (ef. O. Scand.
hvénn ‘ angelica’).

§ 138. Lat. qu of other origin. We have already seen that I.-Eur. kw
became gu in Latin, e. g. I.-Eur. *8kwo-, ‘horse’ (O. Ind. dva-, ef. Lith. aszva,
‘mare’), Lat. équus. The guttural of ldcus (Gk. Adios, ‘a tank,’ O. Ir. loch),
lactinar, is not I.-Eur. q3, but when followed by a consonantal u we find qu in
aquear. The occasional spelling sterquilinium (see Georges, Lex. Wortf. s. v.)
may be like that usage of Late Latin orthography, whereby qui is written for
Greek xv, e.g. quinicus, helguisticon, liquiritia (see ch. ii. § 28), the ui being
meant to express the ti-sound of Greek v, or possibly the qu being meant to
indicate the hard unpalatalized guttural. In Italian, Latin qu before e, ¢ has
this hard sound (written ch), e.g. chi, che, retaining the labial affection (the
following w-sound) before a, e.g. quale (see ch. ii. § gz), though in cinque,
‘five,’ where two Latin labialized velars stood in successive syllables, the first
seems to have lost its labialization in Vulgar Latin.

’

§ 189. Gu. I.-Eur. g* is in Latin v, but after a consonant gz,
and before a consonant g. Thus I.-Eur. *g*iwo-,‘ alive’ (O. Ind.
jiva-, O. Ir. biu, Lith. gyvas) is Lat. vtvus ; I.-Eur. root ong?-,
‘to anoint’ (O. Ind. anj-, O. H. G. ancho, ‘ butter’ ; ef. O. Ir. imb,
‘butter,’ W. ymen-yn) is Lat. waguo; Lat. gravis is cognate with
O. Ind. guré-, Gk. Bapi’s, Goth. kaurus. Before « I.-Eur. g*
was replaced by g in Latin, a change probably of a very early
date, e.g. gurges (Gk. t1d-Bpuya, tro-Bpdxtos), though the w may
be often regarded as a weak form of the wé of gwé (see § 51).
Before 0, Latin gu (gv) seems to have come in time to sound like
Latin g, as guo came to sound like co ($137). After o in ter-
minations had come to be written uw, even when preceded by »,
u, we find spellings like distingunt, extingunt, for which distin-
guunt, extinguunt, &c. were afterwards restored by the analogy of
the other persons, distinguimus, &c. The grammarians of the
Empire have difficulty in determining the proper spelling of
verbs in -guwo and -go, and generally follow the rule of writing -go
when the Perfect ended in -ai, e. g. extingo, ungo.

§ 140. I.-Eur. g?, Lat. v: other examples. Lat. vénio (0. Ind. gam-, gach-,
Arm. e-kn,‘hecame,’ Gk, Baivw, Béoxw, Goth. qima, Engl. Icome, Germ. komme;
Lat. véro (O. Ind. gr-, Arm. ker, ‘food,’ Gk. Bopd, BiBpdoxw, Lith. geri, ‘I
drink,’ O. SI. Hira, ‘I swallow’); Lat. véru (0. Ir. bir N.,a U-stem, W. ber) ; Lat.
niidus for *novedo- (Goth. nagaps ; ef. O. Ind. nag-nd-, Lith. ntigas, O. SI. nagii).
302 THE LATIN LANGUAGE. [Chap. IV.

An intervocalic gw of later origin is similarly treated in mavilo (#mavvolo) from
*mag(e)-volo.

§ 141. Dialectal b. In Umbro-Osc. I.-Eur. g3 is b, e. g. Umbr. benust, Osc.
ce-bnust from the root of Lat. vénio. So Lat. bis, &c. seem to be dialectal or
rustic (see von Planta, i. p. 335).

§ 142. g for I.-Hur.g%. (1) Before consonant: Lat. agnus (Gk. dprds for
*dBvés, O. Sl. jagne) (on dvillus, see § 19); Lat. migro (Gk. d-peiBw, O. SI.
miglivii, ‘mobile’) ; Lat. glans (Gk. Badavos, Arm. kadin, Lith. gilé, O. Sl.
zeladi).

(2) Before u. (On the spellings distingunt, &c., see Bersu, die Gutturalen).

(3) Before o. (On the spellings distingo, &c., see Bersu, die Gutturalen).
Similarly gu from I.-Eur. gh® is written g before o in ningo.

§ 143. GH*. I.-Eur. gh* is Latin f, when initial, but between
vowels v, after a consonant gu (before w reduced to g), and before
a consonant yg. Thus Lat. formus (O. Ind. gharmé-,‘ heat, Arm.
jerm, Gk, depuds, Eng]. warm for *gwarm, Pruss. gorme, ‘ heat’) ;
nivem Acc., O. Lat. nivit [Gk. via, vider (vet-) ; of. Zend snaeZaiti,
Goth. snaivs, ‘ snow, Lith. snaigyti Inf., O. Sl. snégii, ‘ snow,
W. nyf]; xinaguit (Lith. snifiga) from the root sneigh*-,‘ to snow.’

I.-Eur. gh? is f in Umbro-Ose. whether initial or intervocalic,
&e. (see von Planta, i. p. 447, for examples).

§ 144, I.-Hur. gh% in Latin: other examples. Lat. cd-niveo from the root
kneight- (Goth. hneivan, Germ. neigen) ; tergus (Gk. oréppos, réppos) ; Lat.
foveo, the original meaning of which is ‘to warm’ (e.g. Plaut. Capt. 847;
foueri foculis feruentibus) from the root dhegh?., ‘to burn’ (Lith. degu), and
from the same root, Lat. fivilla (Gk. réppa for *0éppa).

§145. The Sibilants: S, Z. In Sanscrit, besides the §
(I.-Eur. k), which corresponds to a guttural in Greek, Latin,
&e. (e.g. O. Ind. ’atém, Gk. é-xardy, Lat. centum) we have s
(I.-Eur. s) which corresponds to s in other languages (e.g.
O. Ind. sapté, Lat. septem, O. Ir. secht, Goth. sibun, Lith. sep-
tyn}), and § (like our s/) which appears after i- and w-, 7- and &-
sounds, e.g. ugta-, Lat. dstws. Sanscrit k§ is the equivalent of
Greek xr in tk$a-, Gk. dpxros, &c., of Greek £ in 4kSa-, Gk.
&éwv, &c., and even of Greek x6 in k’am-, Gk. xodv. The exact
number and nature of the I.-Eur. sibilants have not yet been
determined, but we can at least discriminate an unvoiced and
a voiced sibilant, which we may call S and Z (cf. Engl. ‘use’
§§141-146.] REPRESENTATIVES OF I.-EUR. SOUNDS. §, Z. 303

Noun and ‘use’ Verb), without precluding the possibility of this
S and Z representing more than one variety of sibilant.

§146. S, Z. Latin s was, as we have seen (ch. ii. § 117),
unvoiced or hard. The voiced or soft sibilant, for which the
symbol Z may have been used in early times (ch. i. § 5) passed
in the fourth cent. B.c. into 7 between vowels (cf. Engl. ‘ for-
lorn, Mid. Engl. forloren, beside ‘lost’). Before a consonant the
voiced sibilant was dropped with lengthening of the preceding
vowel, e. g. I.-Eur. *nizdo (O. Ind. nidé-, Arm, nist, ‘ situation,’
O. Ir. net M., ‘ anest,’ Eng]. nest), Lat. xt¢dus. An initial sibilant
was often dropped in I.-Eur.; thus we have a root teg-, ‘ to
cover, roof’ (Lat. tégo, Gk. réyos, O. Ir. teg,‘ house,’ Engl. thatch),
as well as a root steg- (O. Ind. sthag-, Gk. créyw), the roots
without initial s- being perhaps those used after words ending
in -s (ef. raroreyats for rais oréyais on the Gortyn inscr., dvorn-
vos for *8uc-ornvos, &c.) ; and it is not always easy to say whether
Latin words, which lack an initial sibilant that is found in cog-
nate words of other languages, have lost it through the phonetic
laws peculiar to Latin, or represent an I.-Eur. ‘doublet. Ldtus,
bread, from the I.-Eur. root stel-,to extend’ (O. SI. stelja), appears
in Old Latin in the form s¢/dtus, stlataria or stlattaria navis (ch.
ii. § 130); and we have s¢lo-cus, stlis (slis) as the old forms of
l6cus, lis, just as in Greek cpixpds, &e. are the older forms of
puxpds, &e, An initial sibilant is not found in Latin before m,
e.g. mordeo from the root smerd- (Gk. cpepdadréos, O. H. G.
smerzan, ‘ to feel pain,’ Engl. to smart), before 1, e. g. niéirus from
I.-Eur. *sntiso- (O. Ind. snuga, Arm. nu, Gk. vo(o)és, O. H. G.
snur), before /, e. g. dado for *slabo (O. H. G. slaf, ‘loose,’ Lith.
slabnas, tweak’). Before 7 in the middle of a word a sibilant
becomes 4, e. g. sobrinus for *swesrinus (Lith. seserynai, Pl.), from
I,-Eur. *swésor-, ‘a sister,’ probably from an earlier f (§ 114);
whether it becomes fat-the beginning of a word, e.g. frigus
(Gk. giyos), or is dropped, e.g. répo for *srépo (cf. serpo), or
becomes str-, e. g. stringo (O. Ir. srengim, ‘ to draw’), is uncertain.
After 7 and Z it is assimilated, e.g. verres for *verses (Lith. vefs-
zis; cf. O. Ind. vf%a-), coldum, O. Lat. collus, for *colso- (O. Engl.
heals, Germ. Hals). Initial ps-, 4s- appear to have become s-,
304 THE LATIN LANGUAGE. (Chap. IV.

e. g. sabiilum, sand (Gk. Wwaypos for *yad-pos), dis-stpo (O. Ind.
ksip-, ‘ to throw ’).

Latin s often represents an original dental sound, e. g. adgres-
sus formed from the stem of adgrédior with the participial ter-
mination -¢o-. This change was probably very old, so that the
form adgrettus (written in the earlier orthography adgretus),
quoted from Ennius by Paul. Fest. 5. 6 Th., probably represents
*ad-gred(i)tus (like cette for *cé-dite,§ 108). A double ss (gener-
ally arisen from #¢, ¢s; ef. our ‘gossip’ for god-sip) was after
a diphthong or long vowel (see ch. ii. § 129) reduced in the
Early Empire to single s, e.g. fiisus, older fussus, asus, older
ussus. In sescéni, from sex, the second s represents an original #
(i.e. e-s) (see § 158); and # reduced to s is first assimilated, then
dropped with lengthening of the vowel, in words like téum, older
tellum, for *testum, *tewlum (§ 117). On initial s- for sy-, e. g.
suo, see § 65; for sw-, e.g. siidor,§ 71. On ns, see ch, ii. § 66,
on L.-Eur. ms, above, § 76.

In Umbro-Osean I.-Eur.s remains when initial, e.g. Umbr.
sent, Osc. set ‘ sunt,’ but when intervocalic became first voiced s
(written in Oscan in the native alphabet s, in the Lat. alph. z,
e.g. Fluusaf ‘ Florae, eymazum* rerum’), which in Umbr. passed
into r, e.g. kuratu ‘ curato ’ (Pel. coisatens ‘ curaverunt’), I.-Eur.
ss became tt if the Perfect ending -atted 3 Sg., -attens 3 Pl.
(e.g. Ose. préfatted ‘ probavit, prdafattens ‘ probaverunt’) is
rightly compared with the Lat. Future in -sso, e.g. amasso (ch.
vill. § 3); sr probably became fr (Lat. dr) (ef. mod. Neapolitan
Uttrafe for Ital. Ottobre, October’); sn, sm, sl remain, e.g. Umbr.
snata P. P. P. (Lat. xare), Osc. Slabiis ‘ Labius’ (cf. Lat. S¢/abo-
rius); YS appears in Osc. sometimes as r with ‘compensatory ’
lengthening, e.g. teerim ‘terram,’ sometimes as rr, e.g. Kerri,
in Umbrian sometimes as rs (s), e.g. ¢ursitu, tusetu (cf. Lat.
terreo), sometimes as rf, e.g. Cerfu-; ns in the middle of a word
became nts, e.g. Umbr. menzne ‘ mense’; when final it is in
Ose. -ss, in Umbr. -f, e.g. Osc. viass ‘vias, Umbr. turuf ‘ tau-
ros’; final -nts is in Umbr. (and Osc. ?) -f, e.g. Umbr. zetef
‘sedens.” (On the treatment of I.-Eur. s in Umbro-Oscan, see
von Planta, i. p. 472.)
§§ 147, 148.] REPRESENTATIVES OF I.-EUR. SOUNDS. 5S, Z. 305

§ 147. I.-Bur. s, Latin s: other examples. J.-Eur. *séno-, ‘old’ (O. Ind.
sdna-, Arm. hin, Gk. évn «al véa, ‘the old and new day,’ i. e. the last day of
the month, 0. In. sen, W. hen, Goth. sinista Superl., Lith. sénas), Lat. sén-ior,
sén-ex, Gen. senis; I.-Eur. root wes-,‘to clothe’ (0. Ind. vas-, vastra-, N., ‘cloth-
ing,’ Arm. z-gest, Gk. éo-64s, Goth. vasjan, vasti F.), Lat. ves-tis ; I.-Eur. *aks(i)-,
‘axle’ (O. Ind. dkga-, Gk. dfwy, O. H. G. ahsa, Lith. aszis, O. Sl. osi), Lat.
axis; I.-Eur. *potis Nom. Sg. (O. Ind. patis, ‘master,’ Gk. méais), Lat. potis.

Ny

§148. Lat. r forintervocalic sibilant. Intervocalic s became h in Greek,
e.g. ew for edhw (I.-Eur. *eusd), and was dropped, e. g. yéveos, yévous Gen.
(I.-Eur. *génés-6s), but in_ Latin it appears as r, e.g. tro, généris, having
probably passed through the stage of voiced s (z), a stage at which the sibilant
remained in Oscan, e. g. ezum ‘esse,’ while it suffered rhotacism in Umbrian,
as in Latin, e. g. erom. The grammarians often quote Old Latin forms with
intervocalic s, e.g. lases, Valesti, Fusii (Quint. i.4. 13; ef. Ter. Scaur. 13. 13 K.
Fusius, asa, lases); dasi, arbosem, robosem, helusa ‘holera,’ loebesum ‘liberum’
(Paul. Fest. 48. 19; 11. 20; 71. 12; 86. 30 Th.); r pros littera saepe antiqui
posuerunt, ut matosibus, meliosibus, lasibus, fesiis (id. 359. 1 Th.) ; pignosa (id.
260. 11 Th.) (for other passages see Miiller ad Paul. Fest. p. 15), and often
refer similar forms to the Sabine dialect (e.g. Paul Fest. 6. 36 Th. aurum...
alii a Sabinis translatum putant, quod illi ‘ausum’ dicebant; id. 18. 3 Th.

. Aureliam familiam ex Sabinis oriundam a Sole dictam putant, quod ei
publice 4 populo Romano datus sit locus, in quo sacra faceret Soli, qui ex
hoc ‘Auseli’ dicebantur, ut ‘Valesii,’ ‘Papisii’ pro eo quod est Valerii,
Papirii ; Varro, ap. Vel. Long. 69. 8 K. (ef. L. L. vii. 27) gave fasena as Sabine
for Lat. harena, Varro (L. L. vii. 26) quotes examples of this older spelling from
the Carmen Saliare; Livy speaking of Sp. Furius Fusus, the consul of 464 B.c..
says that some of his authorities spelt the name Fusio- (iii. 4. 1 Furios ‘ Fusios’
seripsere quidam) ; on the inscription with the Carmen Arvale (C. I. L. i. 28)
we have Lases ‘ Lares’: enos, Lases, iuuate; in the most ancient piece of Latin
preserved for us, the Praenestine fibula, Numasioi ‘Numerio’ (xiv. 4123 Manios
med fefaked Numasioi), and in the Dvenos inscr. Tottesiai ‘Tuteriae.’ But words
of the literary period with intervocalic s are either (1) dialectal, e.g. dmasius,
a gallant (Sabine ?, see Nettleship, Contributions, x. v.), or (2) foreign loanwords,
e. g. gaesum (Gaulish ; ef. gaesati, Gaulish mercenaries, C. G. L. v. 71. 23, O. Iv.
gai) (so dsinus, laser, résa, siser, &c.), or (3) had originally ss, whether derived
from I.-Eur. tt, e.g. caesus for *caet-tus from caedo, from ns (I.-Eur. ntt, &e.),
e.g. vicésimus older vicensumus (ch. ii. § 66), formisus, older formonsus (ib.), from
I.-Eur. ss, ce. g. quaeso, older quaes-so, a different word from quaero (ch. viii. § 33),
ndasus, older nassum, or from some other consonant-group. This older ss was
after a long vowel or diphthong written s after the close of the Republic, but
Quintilian tells us that caussae, cassus, divissiones, &c. was the spelling of Cicero
and Virgil (i. 7. 20 quid quod Ciceronis temporibus paulumque infra, fere
quotiens s littera media vocalium longarum vel subjecta longis esset, gemina-
batur? ut ‘caussae, cassus, divissiones’: quomodo et ipsum et Vergilium
quoque scripsisse manus eorum docent), and this spelling is by no means
uncommon in the MSS. of Plautus, Virgil, &c. (see ch. ii. § 129). After a short
vowel ss remained, e.g. fissus. Quisillus, piisillus (ef. pisus) are said to show the
same reduction in the pretonic syllable as Gfella (beside offa), mdmilla (beside
mamma) (ch. ii. § 130); rather the Dim. quasillus was formed from *quas-los

Xx

  
 
306 THE LATIN LANGUAGE. [Chap. IV.

after Rhotacism had ceased to operate ; miser (on the spelling myser see ch. ii.
§ 16, p. 29) has been explained as a loanword from the Greek (svoapés),
like other adjectives expressive of nuances of feeling, e.g. hildris older
hilarus (Gk. idapés). An initial s is not rhotacized when it comes after the
final vowel of a preposition, &c., in a compound, e.g. pd-situs from po-,
a byform of ab (I.-Eur. *ipé) and the P. P. P. of sino, but the final s of
a preposition, &c., in a compound is rhotacized before an initial vowel, e. g.
dir-imo from dis- and &mo, diribeo from dis- and hébeo. Furvus points to an
earlier trisyllabic fus-wo- (cf. arvum from trisyllabic aruwm, § 68), a by-
formation of fus-cus, and Minerva to *Menes-ua (the word is a quadrisyllable in
Plaut. Bacch. 893, Attius, Trag. 127 R), but before consonantal v we have s
dropped with ‘Compensation’ in di-vello, da-rendo, &e. Other examples of forms
with 7+ beside forms with s are maereo (maestus), géro (ges-si, ges-tum), haurio
(haus-(s)t, haus-tum), quaero (quaes-tus; but quaes(s)ivi, quaes(s)itum come from
quaes(s)o, ch. viii. § 33), Etruria (Etrusci), auris (aus-culto), nares (nds-(s)um), and ob-
lique cases of S-stems, ev. g. frineris (sunes-tus), Sneris (onus-tus), verberis (subverbus-
tam Plaut. ap. Fest, 444.15 Th.), honoris (honestus) ; by analogy of these oblique
cases + has found its way into the Nominative of honor (older honos), arbor
(older arbos; ef. arbustum and arbérétum), odor (older odos), &c..(ch. vi. § 7).
(For a fuller list of examples of the Latin and Umbro-Oscan treatment cf
I.-Eur. intervocalic s, see Conway, Verner’s Law in Italy). The change of inter-
vocalic s to + is a common occurrence in Janguage. English r corresponds
to Gothic z in words like ‘ ore’ (Goth. aiz-, I.-Eur. *ayes-, Lat. aes, aeris Gen.),
and in Polish a word like mode, ‘ can,’ has a trilled sound of the voiced sibilant
that is hardly to be distinguished from 7 (see B. B. xv. pp. 270 sqq.).

§ 149. Initial Sibilant before Consonant: (1) before unvoiced consonant :
I.-Eur. root sta-, ‘to stand’ (O. Ind. stha-, Gk. ordi, Goth. staps, 0. S1. stati
Inf. ), Lat. stare, stétio ; I.-Eur. root sper-, ‘to strike with the feet’ (0. Ind. sphur-,
Gk. cnaipw, Engl. spurn, Lith. spirit), Lat. sperno ; I.-Eur. root skand., ‘tospring’
(O. Ind. skand-, Gk. cxavdéAn@pov, a springe, O. Ir. ro se-scaind, ‘ he sprang’),
Lat. scando, Similarly we have sir-, e.g. in Lat. stratus, stramen (cf. Gk. orpwrds,
aTpwud), spr- in sprétus, scr- in scrdbis, a ditch (Lett. skrabt, ‘to scrape’), while
stl- of O. Lat. stldtus, stlataria (stlatt.) navis has become class. Lat. I- of latus, broad. .

(2) Before voiced consonant: I.-Eur. root sneigh®-, ‘to snow’ (Zend snaeZaiti,
GE. veipe:, dya-veipos for *d-ya-ovios, O. Ir. snechta, W. nyf, Goth. snaivs, Lith.
snégas, O. SL. snégit) Lat. nivem Acc.; I.-Eur. root sna- (0. Ind. sna-, ‘ to
bathe,’ Gk. v7xw, I swim, O. Ir. sndim), Lat. nére; I.-Eur. root slég- (Gk.
Aayya¢w, to slacken, Anyw, to cease, O. Ir. lac, ‘weak,’ W. llag, O. H. G. slach,
Engl. slack), Lat. langueo; I.-Eur. root sleub- (Goth. sliupan, Engl. to slip),
Lat. labricus; I.-Eur. root smerd- (Lith. smirdéti, ‘to stink’), Lat. merda;
Gk. (a) ptxpés, Lat. mica. Ifwe may infer from the treatment of an internal
sibilant before a voiced consonant, e. g. nidus for *nizdo-, it would seem that
the initial sibilant was first assimilated, *nnix for *snix (cf. Gk. pido-ppedjs
for *q.A0-cperdys) then dropped, nix.

(3) Before r, The use of ¢ for I.-Eur. d with + in Latin, e.g. der, stem dtro-
for *ddro-, suggests that Latin r was not voiced, so that it is better to consider
separately the treatment of an initial sibilant beforer. The instances are
unfortunately few and uncertain. Lat. frigus goes naturally with Gk. éryos,
but it has also been connected with Gk. ppicow, while Lat. rigor, rigidus has
been assigned to fiyos; Lat. fraga, strawberries, has been referred by some to
§§ 149-151.] REPRESENTATIVES OF I-EUR. SOUNDS. S§, Z. 307

Gk. Adé, a grape, by others to fragro, and certainly Lat. rdcemus goes more
naturally with fag than fraga. Whether Greek fcyos, fd¢ originally began
with o or f is a moot point. Lat. répo goes naturally with serpo, but Lith.
réplisti, Zend rap-, ‘ to go,’ suggest an I.-Eur. ‘doublet’ without the initial
sibilant. The I.-Eur. root sreu-, ‘to flow’ (O. Ind. sru-, Gk £é(F), Lith. sravit,
Ir. sruaim, ‘a stream,’ 0. H. G. stroum), has been sought in the Latin words
ramen, (ficus) Riimina, Rumon, the old name of the Tiber (Serv. ad Aen. viii.
63. 90) and in the name Roma itself; Lat. rétis has been connected with séro,
Lat. riibus with Germ. Ge-striipp, and so on (for other examples see Osthoff,
M. U. v.62). On the other hand Latin forms with initial r which have in
other languages a sibilant before the r, may come from an I.-Eur. ‘doublet’
which lacked the sibilant, as t#go comes from I.-Eur. teg-, a byform of the
root steg-, ‘to cover,’ roof. Another possibility is that str- may be the Latin
equivalent, as in our ‘stream,’ &c. It is not always easy to decide where
Lat. str- and str-, sr-, in other languages, represent an original str- or an
original sr- ; and similarly O. Lat. stl-, of stlécus, stlis and stis may have been
originally si- and not stl-.

§ 150. O. Lat. stl, sl, sel. Quintilian (i. 4. 16) quotes sflocus and stlites as
O. Lat. forms. The old form sélis was retained in the legal phrase decemviri
stlitébus judicandis in Cicero’s time (Cic. Or. xlvi. 156); on the Lex Repetundarum
of 123-122 B.c. (GC. I. L. i. 198) we have once slis but usually lis, and sL. IvDIK
on a Scipio epitaph of c. 130 B. c. (i. 38), sclitib.. . (x. 1249); cf. stloc[us] (v. 7381).
Stlembus, slow, is quoted from Lucilius (Paul. Fest. 455. 4 Th.); scloppus (v. 1.
stloppus), is used by Persius (v. 13) to indicate the sound of slapping the cheek
when distended (cf. Ital. schioppo) :

nec scloppo tumidas intendis rumpere buccas.
In dialectal names these combinations are preserved, e. g. Stlaccius (C. I. L. vi.
26863, &c.) (cf. Lat. lacca, a swelling on the leg ?), Stlaborius (Wilm. 1913, Pom-
peii) (cf. Lat. lébor ?), Oscan Slabio-. In Latin the ¢ of sti- (or c, for i became
cl, § 105) oe be dropped, as it is in the name Foslius (cf. Fostulus, Faustulus),
C.I.L. i?. p. 130) (ef. for(c)tis, § 157), leaving sl-, which would become *l-
(see aes then /-.

§ 151. Sibilant before voiced consonant in middle of word. Lat. audio
for *aus-dio, from *aus, a byform of auris; Lat. pedo from I.-Eur. pezd- (M. H. G.
fist, Pruss. peisda, ‘ podex’) with o-grade of root in podex for *posd-ex ; Lat. sido
for *si-sdo from the weak grade of the root sed-, ‘to sit,’ with i-reduplication
(ch. viii. § 9); Lat. sidus for *susedus from the root saus-, ‘to dry’ (O. Ind.
Sus-, Gk. atos, Lith. satisas, Engl. sear); Lat. malus for *mas-lus or for *mas-dus
(§ rr1t) (Engl. mast); Lat. damus (O. Lat. dusmus), dtimetum, dummetum in the
earlier spelling, e.g. in Virgil MSS. (see Ribbeck’s Index) for *dusmetum (cf.
O. Ir. doss, ‘a bush’) (the dialectal name Dusmia is found on inscriptions, Eph.
Enigr. viii. 128. 820, both from Teate Marruc). Lat. cdmis (cosmis on the Dvenos
inser.) ; Lat. primus for *prismus (cf. priscus, pristinus, Pelign. prismo-); Lat.
pomerium for *pos(t)moerium (Varro, L. L. v. 143; ef. Paul. Fest. 327. 13 Th.) ;
Lat. prélum for *pres-lum (cf. pres-si: a Latin *preso must have existed beside
premo, as Gk. rpé(c)w beside rpéuw); Lat. qudlus for *quas-lus (cf. qudsillus ;
Lith. kdszius, O. Sl. koai) ; Lat. cénus for *casnus (cf. cascus, Osc. casnar, an
old man); Lat. fanum for *fdsnum (cf. Ose. fiisna-, Umbr. fesna-, from stem
*fosnd-; on d-2, see § 54); Lat. fes-tus, feriae from fes-iae, Osc. fiisia-); Lat.

X 2
308 THE LATIN LANGUAGE. [Chap. IV.

pone for *pos(t)-ne; Lat. pono for *po-s(i)no, a compound of pi-, a byform of ab,
ap- of aperio (cf. O. Ind. dpa, Gk. do), and sino (cf. the P. P. P po-situs, and
Pft. Ind. originally po-sivi, then by false analogy of pos-dtus, posui (ch. viii. § 39)] ;
aénus, ahénus for *ayes-no-, Umbr. ahesno-), and so with many stems in -éno-,
-ino-, -élo-, &e. In some of these examples the sibilant is a development from
an earlier group of sounds, e.g. from st in pomerium, pone, from s(t) in pono;
similarly from (x) cs, « in lina, written Losna on an old Praenestine mirror
(C2. L. i. 55), for *lux-na (Zend raoxina-, ‘ shining,’ Pruss. lauxnos, ‘stars’) ;
séni, stmenstris (ef. ses-ceni). for *seani, *sex-mensiris; tela from texo, &c. (other
examples in § 162); (2) ns in (h)alo for *anslo (O. SL. achati, ‘to be fragrant’) ;
pilum from pinso. The older spelling showed a double consonant in these
cases. Thus vélatura, the carrying trade, for *vex-latura, from veho, was probably
spelt vellatwra by Varro in a passage (R. R.i. 2. 14) where he connects the word
with vella, the rustic form of villa; a sibilant which came at a later time to
stand before a voiced consonant was similarly treated, we. g. divello for dis-vello,
dimota for dismota (S. C. Bacch. C. I. L. i. 196), damitte (‘dismitte’ non dicas,
Caper, 97. 7 K.), di-numero, di-luo, In the same way the form 2 arose from ex
in collocations like é-vello, é-moveo, é-mitto, é-numero. -luo (see ch. ix. § 29) ; vidén for
vidés-ne was shortened to vidén (ch. iii. § 42). Cémillus (Camelio on old Praenestine
epitaphs, C.J. L. i. 74; 1501 a) was derived by the Romans from a Greek
(or Etruscan ?) xéopiAos, meaning a servant of the gods (see Varro, L. L.
vii. 34; Maer. iii. 8.5: Paul. Fest. 44. 33 Th.; cf. Virg. Aen. xi. 542). Varro
refers Céimena to an earlier Casmena, which he connects with carmen (L. L.
vii. 26). How Casmillus and Casmena (if the word ever existed in this form)
failed to become *Camillus, *Cémena is not clear. The group rsd became rd
(through *rrd, for 7s becomes rr), e.g. hordeum (cf. O. H. G. gersta, Germ.
Gerste) ; twrdus (Lith. strizdas, O. Ir. truit, Engl. throstle). The I.-Eur. proto-
type of custos (Goth. huzd, Engl. hoard), hasta (Goth. gazds, ‘a sting,’ Germ.
Gerte, Engl. yard, O. Ir. gat) may have had sth-, not -zdh-.

Quasillus from *quas-los (class. qudlus) shows that st remained later than the
change of intervocalic s to r (§ 148). Dusmus Adj. occurs in Liv. Andronicus
(end of third cent. B.c.) (rag. 39 R.) dusmo in loco. Plautus’ viden for
videsne, ain for aisne, &c. show that the Jaw was operative in his time.

§ 152. Sibilant before rin middle of word. Lat. cértbrum for *cerésrum
(0. Ind. Sivas-, ‘the head’) ; Lat. crabro for *crasro (Lith. szirszu, O. SI. sriiSeni) ;
Sibra for fisra (ef. filum for *fslum, Lith. gysla, ‘a sinew’); Lat. ténébrae for
*tenésrac (O. Ind. tamisra, from témas-, ‘darkness,’ Germ. Dammerung); fandbris
for *funés-ris, .

§ 153. Assimilation of sibilant to preceding r, 1. Lat. farreus for *farseus
(Umbr. farsio-, cf. O. SI. braSino, ‘food’; Goth. barizeins, ‘made of barley’) ;
Lat. torreo for *torseo (O. Ind. tri-, Gk. répoopar); Lat. ferre for *fer-se; Lat. velle for
*velse; Lat. erro for *erso(Goth. airzjan, ‘to mislead,’ Germ. irren); Lat. garrio for
*garsio (Lith. gafsas, ‘noise’); Lat. porrwm from I.-Eur. *prso- (Gk. mpdcov) ;
Lat. terruncius for *ters-, older *tris- (Gk. rpts). This rr from rs was when final
reduced to 7, v. g. ter (scanned as long by position in Plautus) (ch. ii. § 133), far,
Gen. farris. Before tthe s kept its place, and ther was dropped, e. g. testémentum
for *tersta- from older *#rista- (Ose. tristaamento-) ; tostus for *torstus.

An s (ss), arisen out of an earlier ts, &c, was not assimilated, ev. g. versus for
*verttus, rursus and reversus for *reverttus, ars beside far. Latin rs was
§§ 152-157.] REPRESENTATIVES OF I.-EUR. CONS.-GROUPS. 309

pronounced like ss, as we see from the pun in Plautus, Pers. 740 Persa me
pessum dedit, and often came to be written ss, and after a long vowel, s;
hence russus and rusus, introsum, prosa, &c. (see ch. ii. § 129).

§ 154. Assimilation of preceding dental to the sibilant. Lat. sudsi for
*sudssi from suddeo; conctissi from conciitio, &e.; pissum for *pit(e)sum.
Similarly in the final syllable, hospes for *hospets, miles for *milets ; this -es is
short in classical poetry, but probably long by position in Plautus (ch. ii. § 133).

§ 155. Lat. ss for tt. Before r we find st for #, e. g. pidestris for *pedet-tris
from jedes, Gen. peditis, assestric Fem. of assessor, and perhaps at the end of
a word, e. g. est, 3 Sg. Pres. of édo. But in other cases tt became ss, e. g. dsus,
older tissus, from utor (older oitor, oetor) for *ut-tus, asio for *ut-tio, *oit-tio (Ose.
oittiuf ‘usio,’ beside Pel. oisa ‘usa’ is best explained as *oit(i)tions) ; so fassus
from fiteor, stssus from sédeo, morsus from mordeo, perculsus from per-cello, *per-celdo,
&c., all formed by adding the participial TO-suffix (see ch. v. § 27) to the root of
the verb. In the second cent. B. c. some verbs whose root ended in a guttural
followed the analogy of these verbs, owing to the similarity of their Perfect
Indicative Active, e.g. spargo, sparsi made sparsus, as ardeo, arsi made arsus;
tergo, tersi made tersus, as mordeo, morsi made morsus. But in the period of the
older literature these false forms in -sus had not established themselves ;
Paul. Fest. quotes mertat for mersat (57. 16 Th.; cf. 89. 26); Quintilian
(i. 4. 14) says: ‘mertare’ atque ‘pultare’ dicebant ; and Nonius (179. 4 M.)
quotes from Varro tertus for tersus, and from Accius mertare for mersare. Exfuti,
explained by Paul. Fest. 57.16 Th. as ex/fusi (cf. con-fitto, futtilis, fiitilis) has been
referred to *fatus, P. P. P. of a verb *fuo, to shake (O. Ind. dhi-); ifit comes from
fundo it must represent an older *fud(i)tus, a byform of *fud-tus as al(i)tus of
al-tus ; so mattus, drunk (the Romance forms attest tt) for *mad(i)tus, like
adgrettus (§ 109). Estis, este, &c. from edo, to eat, must be due to the analogy of
dmda-tis, ama-te, &e. ; so com-estus beside comessus. Férunto, &c. for *feront-tod (ch.
viii. § 57), vehementer, if for *vehement-ter (ch. ix. § 2) retain ¢ in the same way.

§ 156. Other groups with a sibilant. When a sibilant came between two
labials or gutturals, the first was dropped, e.g. asporto for *abs-porto, disco for
*dic-sco (cf. di-dic-t), sescéni for *sex-ceni, *secs-ceni. Similarly pst becomes st in
ostendo (but O. Lat. obstinet) beside obstrido, obstino.

As rs became rr, and Js became J, so rs, ls before a consonant become 7, I,
e.g. hordeuwm for *horsdeum, ainus for *alsnus (Lith. elksnis), perna, the ham,
from I.-Eur. *pérsna-, ‘the heel’ (Gk. arépva, Goth. fairzna, O. Engl. fyrsn,
Germ. Ferse ; ef. O. Ind. parini-), but an unvoiced consonant preserves the s at
the expense of the 7, J, e. g. toslus for *torstus, posco for *por'sco (0. H. G. forscin,
Germ. forschen), properly for *prksko from the root prek- of precor, &e. (On
these groups see the next paragraph.)

§ 157. Loss of Consonant in Group. It is convenient here
to bring together the various examples of the loss of consonants,
when they occur between two other consonants, or in some un-
pronounceable combination (cf. Engl. ‘ hal(f)penny, ‘ Satur(n)-
day,’ ‘ be(t)st,’ ‘ cas(t)le, ‘ go(d)spel’). It is not always possible
to decide whether the consonant was already ejected in what we
310 THE LATIN LANGUAGE. (Chap. IV.

call the‘ Indo-European period) e. g. misk-,‘ to mix’ (Lat. mesceo,
O. Ir. mescaim, W. mysgu Inf., O. H. G. miseu), for *mik-sk-,
from the root meik-,‘ to mix’ (O. Ind. mi8-ré-,‘ mixed, Lith. su-
miszti, ‘ to get mixed’), with the addition of the Inceptive suffix
(ch. viii. § 21), or whether its ejection is due to the phonetic laws
of Latin, e.g. dana (on an early Praenestine mirror Losva) for
*lu(e)sna (Zend raox&na-, * shining,’ Pruss. lauxnos, ‘ stars’).

A consonant between two others is dropped in such groups as:

(1) 1(e)t, 1(e)s, r(e)t, r(e)s, r(t)e, s(e)t, s(t)l, s(e)1, e(t)s, r(t)s
or r(d)s, e.g. ultus for *ulctus, mulsi for *mulesi, fortis, O. Lat.
foretis (§ 118), tortus for *torctus, torsi for *torcsi, corculum for
*cort-culum from cor(d), pastum for *pasctum, O. Lat. sds, class. lis
for stlis (selis) (§ 150), now for *nocts, ars for *arts, arsi for *ardsi.

(2) r(g)n, r(g)m, r(d)n, r(d)m, r(b)m, e. g. wrna for *urgna (cf.
urceus), tormentum for *torgmentum, from torgueo, orno for *ordno
(cf. ordino), vermina, gripes (Gk. orpddos, Paul. Fest. 571. 12 Th.)
for *verdmina from verto, sarmentum for *sarhmentum from sarpo.

The first consonant is dropped in groups like:

(3) (t)se, (c)se, (p)sp, (pst, (p)se, (8)ps, (nen, (x)st, (x)se,
e. g. esca for *etsca from édo, disco for *diesco (§ 156) (cf. dé-die-2),
asporto for *apsporto from abs (ch. ix. § 12) and porto, ostendo for
*onstendo from obs and tendo (but obstindtus, O. Lat. obstinet, &c.),
Oscus, older Opsceus (Obscus) (see Fest. 212. 24 and 234. 29 Th.),
ipse for *ispse, ignis for *engnis (I.-Eur. *ngni-, O. Ind. agni-,
Lith. ugnis, O. Sl. ogni. See WM. 8. L, viii. 236), fastigium for
*farstigium (cf, O. Ind. bhrsti-, ‘a point, Engl. bristle), Zuscus
(Umbr. Zursco- and Tusco-).

(4) (e)sn or (g)sn, (c)s1 or (g)sl, (e)sm or (g)sm, e.g. lana for
*Zusna (Praen. Losna) for *lucsna, télum for *teslum for *tecslum,
dla for *asla for *agsla, subtémen for *subtesmen for *subtecsmen.

The group act is preserved in junctus, defunctus, anteclass, quinc-
tus, but drops the ¢ in class. guintus, late Lat. defuntus (C. I. L.
ili. 2137), santus (v.81 36), nantus (iii. 1635. 4), &e. (see ch. ii. §§ 70,
95), and cf. conctione miswritten for contione on the Lex Repetun-
darum, i. 198. 18). The group zes remains, e. g. planwi, lana.

§ 158. Other examples. On the forms cals for *cal(c)s, cal, and mers for
*mer(c)s, merz, see ch. ii. § 125; they are like farsi for *faresi, fulst for *fulcsi, mulsi
for *mulesi. Like for(c)tis is fertwm, O. Lat. ferctum, a sacrificial cake, from a lost
§§ 158, 159.] REPRESENTATIVES OF I.-EUR. CONS.-GROUDPS. 311

verb *fergo, to bake (ef. O. Ir. bairgen ‘bread ’) ; also fartus from furcio, O. Lat.
tertus (§ 155) from ergo, sartus from sarcio. Cf. fulmentum for fulcmentum, quernus,
for *querenus. The b of *ambe, around, is dropped in am-termini, am-cacsa, am-stgétes,
&c. (see ch. ix. § 16), and the loss of the d in indu- in similar circumstances prob-
ably led to its being ousted by in, e.g. imperator, ingredi (O. Lat. induperator,
indugredi, ch. iii. § 15). The sibilant is dropped in hordewm for horsdewm (0. H. G.
gersta), turdus for *tursdus (Engl. throstle ; cf. Lith. strazdas), perna (Goth.
fairzna, ‘the heel’ ; O. Ind. par%ni-), alnus (cf. Lith. elksnis for *elsnis) (see
§ 156); also in inquam, coinquo if these stand for *ind-squam, co-ind-squo, but
remains in exta if this stands for encsta (Lith. inkstas, ‘kidney’). Like asporto,
&c. are suscipio for *sups-cipio, astilit (Charis. 237. 2 K.) for apstulit (abstulit).
Posco represents *por-sco (O. H. G. forscon, Germ. forschen), I.-Eur. pr(k)-sko-
like *mi(k)-sko-. The group nst remains in monstrum, &c. but becomes si
between vowels, e.g. tell Like jastigium is testémentum for *ter tum,
(Ose. tristaamentud AblL.).

The weakening of a root often produces an unpleasing consonant-group
which has to be changed and often becomes unrecognizable. Thus the I.-Eur.
weak-grade of dek- (of the numeral ‘ten,’ I.-Eur. *dékm) appears in the word
for hundred as dk-, which is changed to k- (I.-Eur. *kmtom for *dkmtom,
ch. vi. § 76; Lat. centum beside décem) ; the weak-grade of the root gen-, ‘to
be born,’ appears in Lat. gndtus, which in class. Lat. lost its initial g (§ 119).
Similarly *tddatus from the root tel- of tollo, &c., became ldtus, ‘ carried’ (§ 105),
and stlatis, from the root stel- of O. SL. stelja, ‘I extend,’ was reduced to the
same form ldtus, ‘extended, broad’ (§ 146). Other initial consonant-groups
avoided in Latin are dr-, cn-, dl-, sm-, wl-, wr-, &e (see this chapter passim).

 

§ 159. Assimilation of Consonants. The loss of a consonant
in a group is often really due to assimilation. Thus the loss of
s in hordeum for horsdeum can hardly be separated from the
assimilation of s to 7 in the group 7s, e.g. horreo for *horseo. In
the case of Assimilation, as of Ecthlipsis, it is often difficult to
say whether the Assimilation already existed in ‘the [-Eur.
period’ or not. Assimilation plays a great part in the com-
pounding of Prepositions with verbs, e.g. aecurrere for ad-cur-
rere, O. Lat. ommentans for ob-mentans, pellége (Plaut.) for per-
lege, al-ligare for ad-ligare (the assimilated form had so established
itself by the time of Pliny that he treats it as a simple verb and
re-compounds it with ad, ad-alligare), but the unmodified forms
of the preposition were often restored in spelling at least (thus
Servius ad Aen. i. 616 says that applicat was the spelling for-
merly in vogue, adplicat the spelling of his own day), a restora-
tion which went hand in hand with the restoration of the
unweakened form of the vowel in verbs like é-neco (older enico),
intel-lego, &c. (see ch. iii. § 31). Examples of Assimilation are:
312 THE LATIN LANGUAGE. {Chap. IV.

PC, &. g. 0¢-caeco, suc-curro; pf, e.g. of-ficina (O. Lat. dpi-ficina),
sufficio; bg, e.g. og-géro, sug-gero; bm, e. g. ommentans (quoted
from Liv. Andron. from Festus 218.14 Th., and explained by obmd-
nens, ‘waiting’; cf. C. G. L. v. 37. 3 ommentat: expectat), summitto,
amitto (for ammitto, ch. ii. § 127) (but dmitto), glitma for *gliib-ma
from glito. Bis assimilated tor in Prepositional Compounds like
surripio, and before » becomes m in scamnum (cf. scabellum),
amnegaverit (C. I. L, vi. 14672), &e. (§ 102), though in Preposi-
tional Compounds the spelling with 4 is usually retained, e. g.
ab-négo; ef (rather e(s)f, § 157), e.g. eféro from ex-fero ; te, e.g.
ac-curro, hoc for *hod-ce; tf, e.g. affero; dg, eg. ag-gero; dl,
e.g. al-luo, pel-luviae, water for washing the feet, /épil/us for
*lapid-lus; dm, e.g. idmentum from rddo; dn, e. g. an-nuo, mer-
cénnarius from mercéd-; tp, e.g. ap-pdreo; tq, e.g. quicquam ;
dr (tr), e.g. ar-rideo ; ds (ts), e. g. as-stdeo ; In, e. g. codlis (Lith.
kinas), but «dma had originally a short vowel between / and x
(Gk. Ad); 1d, e.g. per-cello (cf. clades), but valdé from vilide,
calda from célida; 1s, e.g. collum, O. Lat. collus (Goth. hals
Masc.), velle for *vel-se (cf. es-se); m and » are assimilated in
Compounds of the Prepositions com-, in, and the Negative Prefix
in-, e.g, col-laudo, il-libor, illaudabilis, cor-ruo, tr-ruo, w-ritus,
con-necto or conecto, im-mitto, im-mémor, and similarly the final -m
of &tiam,tam, &c. was often written x (ch. ii. § 65) before an initial x,
e.g. etian-num tan-ne (ch. il. § 135); nl (asin Engl.‘ eleven, Mid,
Engl. enleven), e.g. cdrolla for *cordn-la, himullus for *homén-
lus ; vl,e.g. Atella (Osc, Aderl-), dged/us for *agerdus (cf. supellex
non ‘superlex,’ Probi App. 198. 14 K.); rs, e. g. torreo for *torseo’
(cf. Gk. tépcopat), ferre for *fer-se (cf. es-se). On the Assimila-
tion of s to a following voiced consonant, and the consequent
lengthening of the preceding vowel by ‘ Compensation,’ e.g.
qualus (older quallus) for *quiis-lus (cf. quiisillus), see § 151, on
a like treatment of » before s or 7, id., and on the Assimilation
of Mediae to Tenues (e. g. scriptus for *scribtus), cét-te for *céd(2)te),
Tenues to Mediae, e. g. ab-duco from ap- (I.-Eur. *apd), see § 95.
In dialectal Latin nd became nz as in Osc, ipsanno- ‘ operando-,’
whence dispennite and distennite (Plaut.) (see ch. ii, § 71); on
the pronunciation ss for rs (cf. russus for rursus), see ch, i.
§ 104.
§160.] REPRESENTATIVES OF I.-EUR. CONS.-GROUPS. 313

§ 160. Assimilation in Preposition compounded with Verb. The passages
of the Roman grammarians dealing with this subject are enumerated by
Brambach, Lat. Orth. pp. 294 sqq. Lwucilius declared it to be immaterial
whether one wrote d or c in adcurrere, accurrere (ix. 25 M.):

‘adcurrere’ scribas
dne an c, non est quod quaeras eque labores,

but seems (though the reading is doubtful) to have insisted on the necessity
of distinguishing ad-bifere (from ad and baeto) and ab-bitere (from ab and baeto)
(ix. 27 M.):
‘abbitere’” multum est
d siet an b;

(absimilis seems to have been discarded in Latin for dissimilis, through fear of
confusion with adsimilis) ; he pronounces in favour of pellicio (ix. 32 M.):

in praeposito per
‘pelliciendo,’ hoe est inducendo, geminato 1.

Similarly Priscian (i. 50. 7 H.) quotes pellege, pellucet from Plautus.
The MSS. of Plautus and Terence show great prevalence of Assimilation ;
Plautus puns on adsum and assum, Poen. 279:

Milphio, heus ubi tu és? Assum apud te eccum. At ego elixus sis uolo,

where however the MSS. read adsum, and Stilo (end of second cent. B. c.)
derived as-siduus ‘ab asse dando’ (Cic. Top. ii. 10). It is quite a mistake to _
suppase the unassimilated forms t to be the older, and the assimilated the more
recent (see Dorsch in the Prager philol. Studien, 1887). In the Herculanean~
papyri the preposition is generally not assimilated, e.g. ‘adsiduo,’ ‘inridens,’

‘inlita,’ ‘adfini,’ but ‘imminet,’ ‘imperiis’ (Class. Rev. iv. 442). The byform
a of ab originated in an assimilated form, e.g. before f- in the verb d-fluo, to be
abundant (cf. ab-undo), often confused in MSS. with af-fluo, to flow to (see
Nettleship, Contributions, s.v. afluo), and before m-, v-, e. g. d-mitto, G-vello, the
forms ammitto, &c. heing avoided apparently through fear of confusion with
compounds of ad ; in classical spelling ad is assimilated usually before c-, e. g.
ac-cipio (sometimes before q-, e. g. ac-quiro), before g-, e.g. aggrédior and ad-gredior,
before J-, e.g. alligare but adluere, adliqui (Velius Longus, p. 61 K.), before p-.
e. g. ap-pono, rarely ad-pono, before r-, e.g. arripio and ad-ripio, before s, e. g.
as-sideo and ad-sideo, ad-sum, before t-, e. g. at-tribuo ; com- is assimilated before
l-, e.g. col-légium, col-léco and con-loco, before r-, e.g. cor-rigo, and becomes con-
before c-, d-, f-, g-, J-, N-, -, 8-, t-, v-; on ex see ch.ix.§ 29; in- is assimilated before
m-, &.g. im-mitto (becoming im- also before b-, p-), occasionally before r-, e. g.
ir-ruo and in-ruo, not so often before l-, e. g. in-lado and il-ludo ; ob- is assimilated
before c-, e.g. oc-curro, before f-, e. g. of-fendo, before g-, e. g. og-géro, before p-, e. g.
op-périor, and occasionally before m-, e. g. ob-mdneo, O. Lat. om-mentare ; per- is
assimilated before [-, e.g. pel-licio, pel-légo and per-lego; sub is assimilated before
c-, & g. suc-curro before f-, . g. suf-féro, before g-, e.g. sug-géro, before p-, e. g.
sup-pono, and optionally before m-, e. g. sum-mitto and sub-mitto, and r-, e. & sur-
ripwi (contracted surpui) and sub-ripui ; trans- often becomes tra- before j-, d-, I-,
M-, N-, ©. g. tra-do (trans-dere attested by Donatus for Terence, Phorm. 2, where all
our MSS. have tradere) ; before j- we find co- in coicio, &c., pe- in pejérare, a later
spelling of perjerure (see Georges, Lex. Wortf. 3. v.) (ef. pothert Plaut. Truc. 612 (B))
314 THE LATIN LANGUAGE. [Chap. IV.

(ef. Ital. Gennajo for Lat., Januarius). (See Brambach, Lat. Orth. pp. 296 sqq.
on the Assimilation of Prepositions on Inscriptions, and the Indices to C.I. L.)

§ 161. Other examples of Assimilation. If Festus (252. 7 Th.) is right in
saying that both petna and pesna were O. Lat. words for ‘a wing,’ we must
suppose penna to be the development of the former, while the latter (from
*petsnd-) would become *péna (ef. luna for *lucsna); annus is most naturally
derived from *at-no- (Goth. apn Neut., ‘a year’). The assimilation of ¢ to
a following ¢ was a feature of dialectal (e.g. blatta for *blacta, Lett. blakts,
‘a bug’) and Late Latin (see ch. ii. § 95). Like glama from glubo is rumentim
(glossed by ‘abruptio’ Paul. Fest. 369.12 Th.) from rumpo ; like r@mentum from
rado is caementum from caedo. (On the reduction of mm after a long vowel or
diphthong to m, e.g. *caemmentum to caementum, see ch. ii. § 127). For dp we have
O. Lat. topper (see ch. ix. § 7) for *tod-per (on quippe, quippiam, see ch.x.§ 7). Idcirco
is sometimes spelt iccirco (see Brambach, Hiilfsbiichiein, s.v.). Whether nm
became mm, ve. g. gemma, or rm, e.g. germen, is discussed in § 80, and whether
examen represents *exagmen or *ex-ags-men in § 116. In the Probi App.
(198. 26 K.) we have: amygdala non ‘amiddula’; the gd of frig(t)dus
(frigda Probi App. 198. 3 K.) became dd (ef. fridam, 0. I. L. iv. 291; Ital.
freddo, &c.).

§ 162. Lengthening by Compensation. Closely connected
with the Assimilation of Consonants is what is called the ‘ Com-
pensatory’ Lengthening of Vowels, where the assimilated con-
sonant lends itself rather to increase the length of the preceding
vowel, so that the loss of the consonant is, as it were, compensated
by the additional quantity of the vowel. (English examples are
‘lady, ‘maid, ‘rain’ ‘thane.”) Qudlus, for *gucis-lus (cf.
quas-illus, § 148), is in the older spelling guadlus, dnhélus for
*anhenslus is anhellus, vélum for *vealum (cf. vexillum) is vellum,
aula, a pot (later olla), for *auala (cf. auxilia) is aulla, &c.; the
Adjective ending -dsus for *o-went-to- (ch. v § 65) is in the older
spelling -onssus, -ossus (see Brambach, Orth. p. 268, and the
Indices to Ribbeck’s Virgil and Studemund’s Apograph of the
Ambrosian Palimpsest of Plautus); dimetum for *dusmetum is
in Virgil MSS. dummetum, as diminuo is in Plautus MSS. dim-
minuo (cf. dirrumpo, Bacch. 441 (C D), but disr- (B)], and so on.
(On the spellings with double consonant in the MSS. of Plautus,
Virgil, &c., see ch. ii. §§ 127-133, where the question is discussed
how far a long vowel with a single consonant might be substi-
tuted for a short vowel with a double consonant in Latin.
On the lengthening of a vowel before ns, see ch. it. § 144, and for
additional examples of the loss of s, # with ‘ compensatory’
lengthening, § 151 above.)
§§ 161-163.] REPRESENTATIVES OF L-EUR. SOUNDS. 315

§ 168. Assimilation of Syllables. The change of the older
Perfect-forms cécurri, mémordi, péposci, pépiigi, &e. to cticurri, mi-
mordi, poposct, piipugi, &e. (see ch. vill. § 43) shows the partiality of
Latin for the complete assimilation of two neighbouring syllables.
The I.-Eur. dissimilation of *q?enq*e, ‘five, to *penq*e ($ 116)
(O. Ind. pafica, Lith. penki, &c.) is not seen in Lat. quingue ;
quercus (for *guerquus), querquétum may be another example, for
O. Eng]. furh, our ‘ fir,’ points to I.-Hur. *perg?- (cf. 6¢20 from
I.-Eur. pib-, O. Ind. pibami, O. Ir. ibim). The same similarity
of initial and following syllable, whether an original similarity
preserved in Latin or first produced by the Latin partiality for
a repetition of the same sound, is seen in words like cincimnus,
a curl (Gk. kixwwvos), guisquitiae, shreds (Gk. xooxvApdria), barba
for *farba (Engl. beard, O. SI. brada), quergquéra, ague, mur-
mur (Gk. popptpa), tiptipa (Gk. érow), furfur, tintinno, ciictiimis,
turtur, &. But in Vulgar Latin we find gw becoming ¢ when
a following syllable has gu, e.g. cimque for quingue, cesquo for
guiesco {Bersu, die Gutturalen, p. 98).
CHAPTER V.
FORMATION OF NOUN AND ADJECTIVE STEMS.

$1. I. STEM-SUFFIXES. We have seen how the several
sounds of the Latin language were written (ch. i.), and pro-
nounced (ch. ii.), and what original or ‘ Indo-European ’ sounds
they represent (ch. iv.). We have now to see how Latin words
were formed, and how the Latin process of formation was related
to the ‘ Indo-European.’

For the forming of words we find sounds combined into roots,
and these developed into stems; thus the sounds t, e, and g, are
combined into the root teg-, ‘ to cover’ (Lat. tég-0, teg-men, tectus
for *teg-tus, tdég-a with O-grade of root), which is further
developed into the stems tiga- (Lat. Nom. Sg. toga, earlier *¢oga,
Gen. Pl. togd-rum, &c.), tegmen- (Lat. Nom. Sg. tegmen, Gen.
tegminis, earlier *tegmen-es, &c.) by the addition to the root of
the stem-suffixes -4-,-men-. It is these stem-suffixes, used in
the making of Nouns and Adjectives, which will be the subject
of this section.

§ 2, Suffixes ending in -6, -§ (Nouns and Adjectives of the
First and Second Declension). -O-,-A-. -O-, which should
rather be called the &-G-suffix, since it alternates with & (e.g.
I.-Eur. Voc. Sg. of Mase. o-stems ended in -8, *ékwé, ‘O horse,’
Gk. ine, Lat. equé, &c.), is associated with the Masc. and
Neut. Gender. -A-; which should rather be called the d-suffix,
since it alternates with 4% (e.g. I.-Eur. Voc. Sg. of Fem. a-stems
ended in -&, *ékwi,‘O mare’; cf. Hom. Gk. vipa), is associated
with the Fem. Gender. Hence the 6- and 4-suffixes were used
§§ 1, 2.] NOUN AND ADJECTIVE STEMS. -O-, -A-. 317

in Adjectives, e.g. I.-Eur. néwo-, Mase. and Neut., *newa-,
Fem. (Gk. vé(F)os, vé(F)ov, ve(F)a, Lat. ndvos, novom, nova, &c.).
Special circumstances have however produced a few instances
of Fem. o-stems and Masc. a-stems. Thus Lat. fagus, Gk.
gnyos are Fem., being names of trees; and Lat. agricola, when
it passed from its original abstract sense of ‘ field-tillage’ into
the concrete sense of a ‘field-tiller,’ became Mase. (cf. Gk.
*yeavid-, youth, veavids, a youth). (See ch. vi. § 1.)

Of the many uses of the 5-suffix, two may be selected for
particular notice: (1) in Abstract Nouns (Nomina Actionis),
these having the accent on the root, e.g. L-Eur. *&éno-, ‘ pro-
duction ’ (O. Ind. janam, Gk. ydvos), from root Sen-,* to produce’ ;
(2) in Nomina Agentis, these having the accent on the suffix,
e.g. I.-Eur. *t6rd-, ‘a piercer’ (Gk. ropés), from root ter-, ‘to
pierce’; I.-Eur. *préko-, ‘an asker’ (Lat. prdcus, a suitor), from
root prek-, ‘to ask.’ The root in all these examples shows the
o-grade (ch. iv. § 51).

The 4-suffix is similarly used in Abstract Nouns (Nomina
Actionis), e.g. I.-Eur. *bhiga, ‘the action of fleeing’ (Gk. guy7,
Lat. figa), from the weak grade of the root bheug-, ‘ to flee.’

How far these simple suffixes -d- and -a- have been combined
with others to form the large number of suffixes which end in
the letter 6, or the letter 4, e.g. -io-, -14-, -to-, -ta-, -tuo-, -tua-,
-tro-, -tra, &c. need not be discussed here. In Latin we find
them more used in the older stages of the language, while fuller
suffixes seem to be required in the classical period; thus per-
vicus (from the root weik-, ‘to fight,’ Lat. vinco) is O. Lat. for
per-vicéa, and squdlus, Enn., became squalidus. The Verbal Noun
used as Infinitive by the Umbro-Samnite nations was probably
a Neuter 5-stem, e.g. Osc. ezwm, Umbr. erom from root és-, ‘ to
be,’ Lat. esse; Osc. detewm corresponds to Lat. dicere, Osc.
moltaum to Lat. multare); and at all periods of Latin we see
a tendency to make rough-and-ready coinages of words with the
help of these simple suffixes, e.g. Carna, from *cdr(o)n-, flesh,
the goddess of the vital organs, to whom a temple was dedicated
by Junius Brutus in 510 8.c., Carda (or Cardea), from *cardon-,
a hinge, the goddess of hinges, nola, ‘a say-no’ from xé/o, in
Caelius’ punning description of Clodia (Quint. viii. 6. 53).
318 THE LATIN LANGUAGE. [Chap. V.

§ 3. Latin 6- and 4-suffixes; other examples. Lat. uncus from *oncos
O. Ind. arkas, Gk. éyxcs) from the root ank-, ‘to bend’ ; Lat. délus (Gk. 5éA0s),
perhaps the O-grade of a root del-; Lat. jigum (O. Ind. yugim, Gk. (vyéy,
Goth. juk Neut., O. SI. igo Neut.) from the root yeug-, ‘to join’; Lat. plaga
(Gk. wAnyn) from the root plag-, ‘ to beat’ ; con-viva from vivo, to enjoy oneself
(e.g. Catull. v. 1; Plaut. Pers. 30 uiues mecum; so vita Plaut. Trin. 477; ef.
Non. 14. 16 M. sicuti qui nune est in summa laetitia, ‘vivere’ eum dicimus) ;
parcus from parco ; with lengthened root col-léga (cf. tég-tila, rég-tila, and célare,
if from *téga, *réga, *céla, Verbal Nouns from the roots tég-, rég-, kél-; cf. § 23).
Of early forms, and occasional coinages, may be mentioned : condus and promus
from condo, promo, v.g. Plaut. Pseud. 608 condus promus sum, procurator peni;
trdha, a harrow, for which Virgil substituted (invented ?) the form ¢trahea (cf.
the note of Servius or G. i. 164 traheaeque: Epenthesin fecit causa metri, ut
‘navita.’ traha autem vehiculum est a trahendo dictum; nam non habet
rotas), from traho, like séra, the bolt of a door, from sero, to join ; from ato was
formed Aius Locutius, the god to whom a temple was dedicated in gratitude
for the supernatural warning against the attack of the Gauls, 390 B.c. ; from
pando, Panda, the goddess of opening, after whom was named the Pandana
porta (Varro, L,L. v.42), the Osean name of the goddess being Patana-.
Nonius quotes pervicus, stubborn, persistent, from Accius (Trag. 158 RB.) :

sed péruico Aiax dnimo atque aduorsdbili,
derived from pervinco, to be stubborn, persistent, as we may see from the
assonance of Ennius (Trag. 408 R.):
peruince pertindci peruicdcia ;

coa and nola, from coco and nolo are quoted from Caelius by Quintilian (viii. 6.
53): quadrantariam Clytemestram, et in triclinio coam, in cubiculo nolam ;
so perhaps confeta sus, for sus cum fétu, explained by Paul. Fest. (40. 28 Th.) :
quae cum omni fetu adhibebatur ad sacrificium ; Domi-duca, Pro-ntiba (an

epithet of Juno), Juga (another epithet of Juno), Lua, Vica Pota, noctt-lica,
sangur-stiga. Scriba was the early word for ‘a poet’ (Fest. 492. 19 Th.).

§ 4, -I0-, -IA- (-YO-, -YA-). It is difficult to distin-
guish in Latin the I.-Eur. suffixes (1) -yd-, -ya-, (2) -tyd-,
-lya- (by some written -ayo-, -eya-), for as we have seen (ch. iv.
§ 65), y after a consonant in the middle of a word became
vocalic ¢ in Latin, so that Lat. médéus from JI.-Eur. *médh-
yo- (O. Ind. madhya-, Gk. pé(o)oos) is, unlike the O. Ind.
and Greek forms of the word, a trisyllable. The weak grade of
I.-Eur. -yé-, -iyd- (or -yé-, -iyé-, § 2) seems to have been -1-,
-I-; e.g. Goth. briks, ‘ useful,’ for *brikis, I.-Eur. *bhrig"-,
*bhrig*ys- (-yé-), from the root bhreug*-, ‘ to use, enjoy’ (Lat.
Jruor) ; Goth. hairdeis, ‘a herd, shepherd,’ for *hairdis ; and this
opened the way to a confusion of io- stems with i-stems. Another
byform seems to have been -iy6-, -Iya-, e.g. O. Ind. trt-fya-, ‘third,
§§ 3,4.) NOUN AND ADJECTIVE STEMS. -IO-, -1A-. 319

Hom. Gk. zpoOvp-tn, a form which would in Latin shorten the i
before the following vowel,and become identical with I.-Eur.-iyo-.

These IO- suffixes have three chief uses in I.-Eur.: (1) to
form Verbal Adjectives, especially Gerundives, the Neuter and
Fem. being often employed as Verbal Nouns, eg. I-Eur.
*sdq*-yo-, ‘requiring help or company’ (Lat. sdcius; cf. O. Ind.
saciya-, Gk. d-ooontip from *écc0-) from the root séq?-, ‘to
accompany’ (O. Ind. sac-, Gk. érouar, Lat. séguor). Similarly
Latin extmius in the sense of eximendus (e.g. Ter. Hec. 66 utin
eximium neminem habeam? ‘am I to make no exception ?’) ;
Lat. stidium from studeo, Lat. extiviae from exuo; plitvia from
pluo; (2) as a secondary suffix of Adjectives, the Neuter and
Fem, being often used as Abstract Nouns; often too in Com-
pound Adjectives; e.g. I.-Eur. *pétriyo- (O. Ind. pitriya-, Gk.
mdrpios, Lat. patrius) from the noun *péter- (I.-Eur. pitdér-, Gk.
aatyp, Lat. pidter); Lat. somnium (O. Ind. svépnyam, O. SL
sunije, stinije) from somnus, stem *somno- (0. Ind. svdpnas,
O. Sl. stnti); Lat. falsi-jairius from falsus and jus, discordia from
discors ; (3) in Adjectives which have a sense of comparison or
distinction, indicating a special locality, direction, &c., e.g.
I.-Eur. *médhyo- (O. Ind. madhya-, Gk. é(o)oos, Lat. médius) ;
J.-Eur. *alyo- (Arm. ail, Gk. dAAos, Lat. dius, Goth. aljis);
Gk. deéids has this suffix, while Lat. dexter (Gk. de&repds) has
the -tero- suffix, which has the same force (§ 16). So in some
Ordinal Numbers, e.g. Lat. tertius (ef. O. Ind. trtiya-, Goth.
pridja, O. SI. tretiji, ch. vi. § 61).

A notable use of this suffix in the Italic languages is in the
formation of Proper Names. While in all, or most, of the other.
J.-Eur. languages Compounds were used for Proper Names, the
son taking a Compound slightly varied from the father’s (e.g.
Gk. Awo-xpdrys, son of Auwvo-xdjs, Teut. Walt-bert, son of
Wald-ram), the Italic stocks employed simple stems with this
IO- suffix, e.g. Lat. Licius, Statius, &e., which correspond to
some contracted or ‘pet’-names in the other I.-Eur. nations,
e.g. Gaul. Toutius, a familiar shortened form of Toutio-rix, Gk.
Zevélas, for the more ceremonious Zevé-in7os, &c., Aedkis beside
Aed-intos, &e, (see Fick, Personennamen).

The Oscan inscriptions enable us to distinguish two varieties
320 THE LATIN LANGUAGE. [Chap. V.

of this suffix in Patronymics (or family names), which cannot
so easily be distinguished in Latin: (1) -yo-, in Patronymics
derived from praenomina (what we call ‘Christian names’) in
-o, eg. Ose. Uhtavis, Lat. Octdvius, the patronymic derived
from Lat. Octavus; Osc. Statis from a praenomen *Stato-.
The Oscan suffix is in native characters written -is (i.e. -is), in
Latin characters -is, in Gk. -1s: (2) -iyo-, in Patronymics
derived from praenomina in -yo-, e.g. Statiis, a patronymic
from the praenomen Statis (stem *Statyo-). This suffix is in
Oscan characters -iis, in Latin characters -ies, in Gk. -tes. To
these we may add a third variety, -iyo-, apparently the un-
shortened form of -iyo-. This is used in ceremonious language on
inscriptions bearing the names of magistrates, &c., and is written
-iis, Gk, -eves, e.g. Viinikiis, ’AFderes, both quaestors. How far
Latin spellings like Clodeius, Publeius, Vareius, if they are genuine
Latin forms and not dialectal, may be distinguished from the
normal forms Clodius, Publius, Vartus is hard to determine.
The diphthong ¢ in O. Latin may, as we have seen, represent
the weakening of an original -ai- (-oi-) in the unaccented
syllable, e.g. occeido, as well as an original -ei-, e.g. deico; it
may also be a graphic expression of the long simple vowel 2,
for this -ei- came to be pronounced, and in time spelt, in the
same way as 7; and before another vowel 2 would be shortened
to % Thus Osc. Bovaiano- was in Latin Bovianum through
*Bovianum from *Boveianom; Osc. Paimpaiians is Lat. Pom-
peianus; Osc, Maraio-, Falisc. Mareio-, is Lat. Martus (cf.
Umbr. pernaio-, ‘in front,’ postraio-, ‘behind,’ in Lat. anticus,
posticus), Analogous to the Oscan -iyo- as opposed to -iyo- is
perhaps the Latin use of the full ending -ws, as opposed to the
shorter -7s or -7 (with -2- like Osc. -is ?), in names of magistrates ;
for example, on the S. C. de Bacchanalibus (C.J. Z. i. 196) the
consuls’ names are Marcius and Postumius, but the names of the
clerks who ‘scribendo arfuerunt’ are Claudi, Valeri, Minuci.

As the IO-suffix is often added as a secondary suffix to Verb-
stems (e.g. pinsio and pinso, ch. viii. § 15), so it is added to
Nouns. O-stems either drop their final vowel before it, e.g.
somn-ium, or show -éyo- which became -&o- (ch. iv. § 66), eg.
aureus. (On rustic -eo- for -io-, see ch. ii. § 10). This ending
3 4.) NOUN AND ADJECTIVE STEMS. -I0-, -IA-. 321

was often assigned to other stems, e.g. flammeus (A-stem), corneus
(U-stem), vitexs (I-stem), to denote material. The TER-stems
augmented by -io- produce in Latin a numerous class of Neuter
Nouns indicating the place or instrument of an action, with the
ending -‘6rio- corresponding to Greek -rnpio-, e.g. audt-torium,
‘the place of hearing,’ deversorium for *devert-torium, ‘a lodging-
place,’ scalp-torium, ‘an instrument for scratching’ (Martial
xiv. 83). (Cf. Gk. Bovdev-ripiov, épyac-rijpiov, both indicating
place, xavo-rjpiov, ‘instrument for burning,’ xpi-rjpov, ‘ means of
deciding’). (On Fem. -¢oria in Late Lat. see Ronsch, Codlectanea,
p- 197.) Similarly we have -mdniwm, -monia from MEN-stems,
e.g. dlimonia and dlimonium (cf. alimentum), flamonium (on the
spelling, see Nettleship, Contributions, s.v.) from flamen, an ending
extended to ¢r2stt-monta, sancti-monia, &e. The Adj. ending -drius
(from dsios, ch. iv. § 160) was to some extent supplanted by -dris in
later Latin (ch. iil. § 16), though the popular speech retained the
older forms (e.g. vinarius), forms which should perhaps be restored
to various lines of Plautus (see Langen, Beitrdge, p. 3243; e.g.
militarus, Pseud. 1049). The same may be true of -d/ius and
-dlis, &e., e.g. minualium: éyxepldvov (Gl. Cyrill.), dvitinm (ib.)
A common ending of Abstract Nouns is -t¢ia (often -itées), e. g.
laetitia from laetus, like militia from miles (see also YE-stems,
§ 51), -téium, e.g. servitium, flagitium, lanitium, From N-stems
we have, e.g. cdlonia, in O. Lat., ‘a dwelling-place’ (Plaut. Aw/.
576: ut conmutet coloniam).

-Ario- and -ari-, -Alio- and -ali-. Caper (p. 103. 9 K.) approves the old form
vinarius :
Vasa istaec vinaria sunt, vinaria cella:
Vulgus adhuc retinet de prisca verba loquella ;

and similarly atramenturium (p. 108. 3K.) ; but Probi Appendix (p. 198. 7 K.)
primipilaris, non ‘ primipilarius.’ (For other examples of -arius, -alius, see
Ronsch, Collectunea, pp. 196, 208; Neue ii*. p. 158.)

Other examples of IO-stem Compounds. From jis and dgo was formed
jir(i)-gium ; so litigium ; from dico, jidicium, indicium ; from eo, cimitium, exitium,
initium (ef. § 77); similarly praemium from émo, incendium, suspirium, discidium,
conniibium, subsidium, aedi-ficium, lecti-sternium, stilli-cidium, obséquium [wrongly
declared (Cicero ap. Quint. viii. 3. 35 ; but cf. Lael. xxiv. 89) to be a coinage of
Terence, for it is used by Plautus (Bacch. 1082) and Naevius (Don. ad Ter.
Andr. i. 1. 40)] and exstquiae, exctibiae, suppttiae, vindémia from vinum and démo,
inciiria from cira, &e.

Y
322 THE LATIN LANGUAGE. (Chap. V.

§ 5.-U0-, -UA-. Here again the two I.-Eur. forms of the suffix,
(1) -w6-, -wa-, (2)- tiw6-, -tiwa-, (or -awo-, &c.) are difficult to dis-
tinguish in Latin, where w after a consonant in the middle of a
word became vocalic w at first, though it might afterwards become
a consonant by the process of Syncope described in ch. ii. $ 13,
e.g. furvus, which must have been trisyllabic, fusuos (cf. fus-cus),
at the time when s between vowels became 7 in Latin. An
original -iwd-, -8w6-, -6w5- would also become -#0-, and in time
-vé- in Latin (ch. iil. § 24,p. 174), so that the exact origin of the
Latin suffix -io-, -vd- is often doubtful. Another element of
confusion is that the weak grade of the I.-Eur. suffixes -wo-,
-tiw6- (-wé-, -iiwé-, see § 2) was u, which opened the way to
these stems coalescing with t-stems.

The UO-suffix is much used in Latin and in Teutonic in
adjectives denoting colour, e.g. Lat. helvus (O. H. G. gelo, Engl.
yellow, from *&hél-wo-), furvus, flavus, révus, &c. Greek Verbal
Adjectives in -reos for *-refos, with Gerundive force, e.g.
duoxréos, requiring to be pursued, capable of being pursued (cf.
O. Ind. kartva-, kértuva-, ‘requiring to be done’), are in Latin
represented by formations in -wus, e.g. caeduus, excipuus, prae-
cipuus, conspicuus, while another class of Verbal Adjectives,
denoting state or condition, end in -dévus, e.g. nativus, captivus,
votivus. These Adjectives in -tivus seem to be derived from
Verbal Nouns with a TIO-suffix or a TI-suffix (cf. furtivus
from *furti-, a stem seen in Adv. furtim, ch. ix. § 4; sémentivus
from sementis, Late Lat. sementium, Ronsch, Collect. p. 209),
though some have tried to connect them with Sanskrit Gerund- «
ives in -tavya-, e.g. O. Ind. kartavya-, ‘requiring to be done.’
(See Thurneysen, Verda auf -10, p. 41; von Planta, Gramm. Osk.-
Umbr. i. p. 169), while Verbal Adjectives of the same sense in
-twus, e.g. réctdivus [cf. cadivus (morbus), ‘the falling sickness,
in Gaulish Latin, e.g. Mare. Emp. xx. 93], sub-sécivus, O. Lat.
vocivos from vocare (classical vicare), may come from Verbal
Nouns with an IO-suffix. The forms dé-cidwus, vécuus are not
phonetic developments of these, but follow the analogy of
Gerundive Adjectives like caeduus, excipuus, &e.

§ 6. I.-Hur. Stems in -w6-. I.-Eur. *g%i-w6-, ‘alive, lively’ (O. Ind. jiva-,
0. Ir. biu, W. byw, Goth. qius, Lith. gyvas, 0. Sl. xivii), Lat. vivus ; I.-Eur.
§§ 5-7.) NOUN AND ADJECTIVE STEMS. -UO-, -UA-. 323

*laiwo-, ‘left’ (Gk. Aai(F)és, O. SI. lévii), Lat. laevus, probably connected with
Engl. slow, from Teut. *slaiwa-. Similarly Lat. calvus (O. Ind. kulva-) ; Lat.
scaevus (Gk. oxa(F)és) ; Lat. clivus (Goth. hlaiv Neut. ‘tomb,’ 0. Engl. hlaw,
hl#w, ‘hill,’ esp. ‘grave-hill,’ Sc. law) from the root klei-, ‘to lean, slope.’

The thematic vowel is inserted in I.-Eur.*widh-é-wo-, ‘unmarried, widowed’
(O. Ind. vidhava-, Gk. 4-i@cos ; cf. Goth. viduvo (n-stem), 0. Sl. vidova, ‘a
widow,’ O. Ir. fedb, W. gweddw], Lat. viduus, from the root weidh-, ‘to
separate,’ of Lat. di-vido, &e.

§ 7. Latin Verbal Adjectives in -uus, -ivus, -tivus. Relicwus (rather
relicuos) is a word of four syllables in Plautus, and indeed in all the Republi-
can literature, though it afterwards became religuwos and finally relicus ;
similarly délicuus, &e. (see Bersu, die Gutturalen, p. 59). This ending -uus (-uos),
indicating state or condition, is seen in continuus, ingénuus, assiduus (whence
the Adverb assiduo, for the sake of a pun with which Plautus coins the form
acctibuo, Truc. 422), exiguus, ambiguus {though we find other Compounds like
prodigus from dgo (prodigivus in the Comm. Lud. Saec.), indigus from égeo with the
O-suffix] ; from first conj. verbs we have drwus, irriguus (irrigivus Cato), vdcuus
(voctvus Plaut.) ; from ineo we have Inuus. Derivative 10-stems from these
are e.g. reliquiae, deliquium. Examples of Adjectives, &c. in -wus, derived from
nouns, are: annuus from annus (O-stem), Minerva for *menes-ud (ef. O. Ind.
manas-vin-, ‘intelligent ’) from *menes-, ‘intelligence’ (O. Ind. mdnas-, Gk.
uéveo-), a quadrisyllable in Plautus (ch. iv. § 148), strénuus (ef. Gk. otpiivos,
health, Engl. stern, Pruss. sturnawiskan, ‘earnest’, patrwus from pédter (R-stem).
Another example of a Gerundive Verbal Adjective in -wus is pascwus, fit for
pasture, intended for pasture, with which Plautus contrasts druus, fit for
ploughing (Truc. 149):

non aruos hic, sed pascuos ager est ;

(ef. Cie. de Rep. v. 2. 3 agri arvi et arbusti et pascui), whence arvum, with
O. Lat. aruae Plur., a field for ploughing. Of Nouns and Adjectives in -ivus,
-wa, which some would make Derivative IO-stems of Adjectives in -wus (reci-
divus from recidwus ; but ef. deliquium, reliquiae from delicuus, relicuus), examples
are: internicivus (ef. internecio and internecitum) from nécare, subsicivus [ef. i(n)si-
cium, feni-sicium] from sécare. (Subsicivus denotes what remains over and above
a division of land, &c., hence subsicivus ager, spare land, subsicivum tempus,
spare time, whence the proverb subsicivis operis, Cic. de Orat. ii. 89. 364).
Another word often confused with this last, viz. succisivus, from succido, shows
the more usual mode of derivation, from a Verbal Noun TI- or TIO-stem (cf.
succisio), like passivus, fugitivus. Lixivus, whence the derivative I0-stem lixivius,
comes from lixius, derived from lixa, water, lye, *lixare (Ital. lessare), to boil,
words connected with the root wleiq? of Lat. liquor, O. Ir. fliuch, ‘ wet’ (see
Class. Rev. v.10). The O. Lat. word sonivius (Paul. Fest. 409. 6 Th. ‘sonivio,’
sonanti) used in the augur’s phrase sonivium tripudium (Serv. ad A. iii. go ;
ef. Fest. 422. 19 Th.), will, if the second syllable is long, be similarly related
to sénare, as lixivus to *lixare, subsicivus to subsicare, &c. [Nocivus, Plin. Phaedr.,
&e. from niceo, I.-Eur. *ndkéyd (ch. viii. § 23), has been compared to O. SI.
chodi-vi, ‘ wandering,’ from chodi-ti, ‘ to go,’ ljubi-vi, ‘loving,’ from ]jubi-ti,
‘to love’). :

The I.-Eur. suffixes -twé-, -twa-, or -ttiw6-, -ttiwi-, closely connected with
the Verbal Noun suffix -tu- (§ 47), are frequent in O. Ind. and Slav. (e.g.

Y2
324 THE LATIN LANGUAGE. [Chap. V.

O. Ind. kar-tuva-, kdr-tva-, ‘requiring to be done,’ kar-tva-m, ‘a task’; ef. Gk.
-Te(F)os of diwx-réos, &c.), but hardly appear in Latin ; e. g. mor-tuus (O. SI. mri-
-tvii) ; Fa-tuus another name of Faunus, the god of prophecy, derived from
fa-ri (but féituus, foolish, with short a, means literally ‘ gaping,’ from fitiscor,
&e.). In mituus the t belongs to the Verb-stem (cf. Lett. meetot, ‘to exchange,’
Goth. maipms, ‘a gift’).

§ 8. -NO-, -NA-. The I.-Eur. suffix -né-, Fem. -na-, seems
to vary with the higher grades -éné-, -éna-, and -dnd-, -dna- ;
sometimes a vowel-sound seems to precede the nasal, repre-
senting some such variation as -ond-, -ana- (ch. iv. § 81). Its chief
use is in the formation of Verbal Adjectives, usually with the
force of a Perfect Participle Passive ; thus in Sanscrit a certain
number of Verbs have P. P. P. in -n4-, the others in -té- (I.-Eur.
-to-, $27), and likewise in Teutonic and Balto-Slavic, e. . O. Ind.
purné-, ‘ filled,” O. Engl. bunden, ‘ bound, O. 8]. dani, ‘ given,’
In Latin, as in Greek, the P. P. P. suffix is -to-, but traces of
a similar use of -no- are found in words like péléwus, full (cf. im-
plétus) ; while of Verbal Nouns formed with this suffix we have,
e.g. dénum (O, Ind. dana- N., O. Ir. dan), somuus from the root
swep-, ‘to sleep’ (O. Ind. svapna- M., Arm. k'un, O. Ir. suan,
O. Engl. swefen, Lith. sapnas ; cf. Gk. tavos, O. SI. stint), An
N-stem which passes into the O-declension shows this suffix ;
thus régnum (ch. it. § 144) may be from stem *régen- (cf. O. Ind.
rajan-, ‘rule’). The suffix is preceded by s in dima for *luana (ef.
Josna on an old Praenestine mirror, C. J. L.i. 55) (Zend raox8na-,
‘shining, Pruss. lauxnos Pl., ‘stars’). In Greek we find some
Adjectives of Time in -ivos, derived from a Locative Case ending
in i, e.g. xeusepi-vds (Lat. hibernus for *himrt-no-, ch. iv. § 77), éapt-
vds (Lat. vernus), éomepi-vos (cf. Lat. vesperna), &e. Adjectives in
-tmus in Latin like faginus, junctnus show an I,-Eur, suffix -iné-,
denoting material or origin, e.g. Gk. yy-ivos, made of beech-
wood, BYBAtvos, made of papyrus, &. In Latin, owing to the
weakening of vowels in unaccented syllables, -7~ws may represent
an older -ind-, -8nd-, -dnb-, &c., as well as -indé- ; and, owing to
the syncope of such vowels, -2«s may represent the same forma-
tions. It is however often possible to distinguish between
original -no- and original -ino-, &c.; thus papudnus, made of
poplar-wood, must have had originally a vowel between the /
and the x, for original Zu becomes 2 in Latin (e.g. collis for

@aediag
§ 8.] NOUN AND ADJECTIVE STEMS, -NO-, -NA-. 325

*colnis ; cf. Lith. katmas; see ch. iv. § 78). The suffix -ivos in
Greek, denoting species, occurs frequently with names of animals,
e.g. deAax-ivy from déAgaé, xopax-ivos from xépa€ ; and similarly
in Latin we ‘have ddvinus, equinus, suimus (Goth. sv-ein, Engl.
swine, O, SI. sv-ini), fibrinus (O. H. G. bibir-in; ef. Zend bawr-
aeni8, Lith. bebr-inis), &c., the feminine often being employed
with ellipse of céro, as vitilina, veal, suina, pork, &e. Latin
-inus is often due to the addition of the NO-suffix to IO-stems,
e.g. Litinus from Latium (though, when the suffix -ino-, and
not -no-, is added, we have -sénus,e. g. dlienus from alius, laniena
from danius, with the same dissimilation of the 7- and e-vowels,
as in pretas instead of *piztas, médietas instead of *medtitas, &e.,
Engl.‘ yet,’ ‘ yes, for ‘yit,’‘ yis’), or to I-stems, e.g. marinus from
mare, piscina from piscis, omnind from omnis. It is often seen in
the transference of an ION-stem into the O- or A-declension
(ef. § 55 on epitlonus beside epulo), e.g. in names of gods like
Jugatinus, ‘ qui conjuges jungit,’ from jzgdtio, Potina, the god-
dess worshipped when a child first took milk (Non. 108. 17 M.),
from potio, &e. Lat. -ino may also represent an earlier -aino-,
&c., for ai in the unaccented syllable, became e7, which passed
into 2, but Osc. deiv-ino-, &e. proves an original -ino- for
divinus, &e. The suffix -tinus of diu-tinus, cras-tinus, pris-tinus,
&c., corresponding to the O. Ind. suffix -tna-, -tana- used to form
Adjectives from Adverbs of Time, e. g. diva-tana- and diva-téna-,
‘daily,’ nii-tna-, ni-tana-, ‘ of the present time, pra-tn4-, ‘former,’
&c., may be connected with O. Ir. tan, ‘time, and so be more
strictly the second element of a compound than a mere suffix,
just as -gnus in privi-gnus, bignae, twins (Paul. Fest. 24. 25 Th.),
represents the root Sen-, of génus, gigno, &e. The -gnus of ili-
gnus, silignus, lérignus on the other hand shows the suffix -no-,
the g being the development before w (ch. iv. § 119) of the final ¢
of the stems i/éc-is, salic-is, Jaric-is, and was by their analogy ex-
tended to other tree-adjectives like abiegnus from dbiés, Gen. abiét-is
[ ferriginus (cf. aurigineus, filigineus) adds the O-suffix to the
stem of ferrigin-is|. Similarly -dnus, the ending of Adjectives
formed with the NO-suffix from A-stems, e. g. si/vanus, arcinus,
is extended to Adjectives from other stems, e.g. wrbdnus. By
the addition of this NO-suffix to Nomina Agentis in -or we get
326 THE LATIN LANGUAGE. [Chap. V.

-urnus, e.g. téetturnus (and -urnius, e.g. Plausurnius) ; while
ES-stems give -énus; e.g. ahénus, Umbr. ahesno- from aes,
I.-Eur. *ayes-; vénénum, lit. ‘ philtre” ‘love-potion,’ for *vénés-
no- (cf. Vénus); egénus (cf. eges-tas), &e.; we have -dna from
a U-stem in Jécina. A common use of the NO-suffix in Latin
is to form Distributive Numerals, e. g. gudterni, bini, trint, terni
(see ch. vi. §§ 59, 61, 63).

§ 9. I.-Hur. NO-suffix. I.-Eur. *oi-no-, ‘one’ [Gk. otvy, the ace on dice,
(though ofos, Cypr. ofFos, alone, has the WO-suffix, like Zend aeva-, diva-),
O. Ir. oen, W. un, Goth. ains, Lith. vénas, O. Sl. ini), Lat. dus. Similarly
Lat. cdénus for *casnus (cf. cascus), Ose. casnar, an old man; Lat. urna for
*urena (cf. urceus) ; Lat. quernus for *quercnus; Lat. agnus (Gk. dpvés for
*4B-vos, O. Ir. uan, W. oen ; ef. O. Sl. jagne) ; Lat. grénum from root ger-, ‘to

rub down, wear out’ (O. Ind. jirnd-, ‘rubbed down,’ Goth. kaurn, ‘corn,’
O. Sl. zrino).

§ 10. Latin -nus. Other examples are péter-nus, mater-nus, from R-stems ;
alter-nus, infer-nus, exter-nus ; pronus from the preposition pro, as Osc. amno-,
‘a circuit,’ from the preposition am- (Lat. ambi-), Osc.com(o)no-, Umbr.kumno-,
corresponding to the Latin cdmitium, from the preposition com ; mdtérinus from
materiés (-tn- probably) ; from U-stems tribiinus, (cf. péciinia). The suffix -éno-,
-dno- appears in O. Lat. Duenos, béne (cf. bénignus, bellus for *ben-lus), O. Lat.
duonus, binus from the same root as O. Ind. duvas-, ‘honour.’ From sareio
we have sarcina, from pango (compdges), pagina, from ango, angina, from sto, de-
stina, a prop (ef. destindre, like lancindre, &e., ch. viii. § 10).

§ 11. Latin -inus. From A-stems, names of animals, we have dquilinus, formi-
cinus, noctuinus, mustélinus, cdlumbinus, vipérinus, &c., and from names of persons,
Agrippina, Jiigurthinus, Messilina, Sibyllinus, &c. Other examples are: from
O-stems, div-inus, vicinus; from an R-stem, sobr-inus (for *sosr-inus from séror,
I.-Eur *swésor-); from U-stems, génuinus dens, from *genus, the jaw (Gk.
yévus), veruina from véru ; from Verbs in -io, officina, fodina. (For a list of
Nouns in -ina, see Rénsch, Collectanea, p. 199).

§ 12. Latin -Anus. (See 4.L.L. i. 177.) From town-names/of the first
declension we have Romanus, Ctipwanus, &e. Names of persons lin -dnus are
usually derived from place-names, and often preserve the names of lost towns ;
e.g. Apscillanus points to a town *Apscilla (Eph. Epigr. ii. pp. 25-92). Deriva-
tive Adjectives from the fem. of ordinal numbers show -dnus, e. g. undécimani
from undecima, se. légio, cohors. So decimanus from decima, sc. pars, the tenth
part, tithe, e.g. ager decimanus, land paying tithes, an adjective which some-
how acquired the sense of large, huge, e.g. decumana scuta, decumanus fluctus,
decumana ova, all quoted by Paul. Fest. (3. 31 5 50. 27 Th.), dectmanus acipenser,
Lucil.iv.6M. The same ending appears in some names of gods which are
derived from Verbs of the first conjugation, e.g. Levana from lévare, to lift,
the goddess who protected the newly-born child when first lifted from the
ground, Tutuna from titari, Praestana from praestire, &c. The ending -idnus,
properly affixed to id-stems, e.g. Octavianus from Octévia, sc. gens (the cognomen
of a person who had passed by adoption from the gens Octavia to another
§§ 9-13.] NOUN AND ADJECTIVE STEMS. -MENO-, -MENA-. 327

gens), was much affected with N-stems, e.g. Cicerdnianus, Pisdnianus, which
seem to have pleased the Roman ear more than *Ciceron-anus, *Pison-anus, and
was in time extended to other Proper Name-stems, e.g. Caesdrianus (but
Caesarinus in Cicero, &c.). The ending -itdnus, e.g. Abdéritanus, was produced
by adding the Roman termination to the Greek -irns, e.g. ’"ABdnptrns ; so
Nedpotitanus, Pénormitanus, &. The ending -icénus often denotes a resident
alien as opposed to a native, e.g. Africanus, an Africander, opposed to Afer,
Gallicanus, to Gallus. (Varro, L. L.i. 32. 2: legumina Gallicani quidam ‘legarica’
appellant.)

The Romance languages point to a great extension of the -dnus endings in
Vulgar and Late Latin, e.g. Fr. certain from *certénus, moyen from *medidnus.

$18, -MENO-, -MENA-. Other grades of this suffix were
-m6n6-, -mén4-, and -mné-, -mnaé-, &c. In Latin -méno- and
-méno- would both become -mino- or (by Syncope) -mno- (cf.
lamna, earlier lammina). The suffix was used in the Middle or
Passive Participles of Thematic Tenses of the I.-Eur. Verb (e. g.
O. Ind. bhdra-mana-, Gk. epd-yevos); and although the Pres.
Part. Passive was lost in Latin, traces of this formation remain
in the 2 Pl. Pres. Ind., e.g. légimini for legimini estis (while
legimini, 2 P\. Pres, Imper., may equally stand for the Inf., Gk.
Aeyéwevat, used in Imperatival sense ; see ch. vill. § 81), dumnus,
6 tpepdpevos (sometimes a nurse, as in the Ciris, 441: commu-
nis alumna omnibus, of the earth), /é-mina from the root dhéy)-,
‘to give suck, Vertwmnus, the god of the changing seasons, from
verto, Volumnus, the deity who guarded new-born children, from
vblo, *calumnus (cf. cdliimnia) from calu-or, calvor, to deceive.
Analogous, but irregular, formations seem to be O. Lat. pilumnoe
poptoe, from pilum, a javelin, used of the Romans in the Carmen
Saliare (Fest. 244. 24 Th.), like classical pi/ani, and the names of
deities, Pilumnus, from pilum, a pestle, Vitumnus, * per quem
vivescat infans, from vita, &c. We find -mo- (from an earlier
-mno- ?) in Umbro-Oscan Imperative forms like Umbr. persni-
himu * supplicato,’ Osc. censamur * censetor’; cf. Lat. praefa-
mind, antestdmind (see ch. viii. § 60). Sometimes the suffix is
used in the transference of a MEN- or MON-stem into the 6- or
a-declension, e.g. cd/umna, beside cdlimen (cf. columella for *colu-
men-la), terminus, beside termen and termo (quoted by Festus, 550.
22 Th., from Ennius, e.g. 4. 591 M.: qua redditus termo est).

The ending -mnus in Latin often arises from the addition of
the suffix -no- to a stem ending in a labial consonant, e. g. som-
328 THE LATIN LANGUAGE. [Chap. V.

nus for *sopnus (cf. spor), damnum (cf. Gk. dardvn), scamnum (cf.
scabellum), &e.

§14 -M6-, -MA-. This suffix was used to form Adjectives
(e.g. Gk. pvépos from piéis, MWoyos from Advous) and Nouns,
especially Masculine nouns, but sometimes Fem. (e.g. Gk.
Abstracts in -yds like At-uds, Aow-pds). Examples are I.-Eur.
*dhimés (O. Ind. dhimés, Gk. duds, Lith. damai Pl. O. Sl.
dymii), Lat. fzmus, from the root dheu-; I.-Eur. *gh*ormés,
*gh*ermos (O. Ind. gharmés, ‘ warmth,’ Arm. jerm, ‘ warm,’ Gk.
Oepyds, Eng]. warm), Lat. formus, from the root gh*er-.

It was also used to form Superlatives (with Comparative in
-tro-, ch. vi. § 52), e.g. Lat. swmmus for *s-up-mo, with Comp.
s-tipero- (O. Ind. upamé-, with Comp. tipara-; ef. O. Engl. yf(e)m-
est). In Latin pulcerrimus for *pulcersimus, *pulcrisimus, it is
affixed to the Comparative suffix -is- of mdgis, &e. The more
usual Superlative suffix however was -temo- (-t*mo-) (with ‘ Com-
parative ’ in -téro-, ch. vi. § 52), e.g. Lat. in-timus, with ‘ Comp.’
intero- (O. Ind. én-tama-, with ‘Comp.’ 4n-tara-). (On the
Lat. Superlative see ch. vi. § 54.) But originally this. suffix,
like the Comparative -tero- (ch. vi. § 52), had the sense rather of
likeness (O. Ind. gé-tama-, lit. ‘like an ox’), or position (Lat.
mirvitimus, older mari-tumus, lit. ‘placed by the sea’). The
Ordinal Numeral ending -mus of décimus, &e. may owe its m
to the final of the Cardinal Numeral stem ; but -t*mo- appears in
vicésimus, &e. (ch. vi § 74).

§ 15. Other Examples: (1) of the Noun- or Adjective-suffix. Lat. dnimus
(Gk. dve-pos) ; Lat. fama (Gk. ¢7-un) ; Lat. pal-ma (Gk. wadd-yn, O. Ir. 1i-m
Fem., O. Engl. folem Fem.) ; Lat. culmus (Gk. #dda-yos, O. Engl. healm, Lett.
sal-ms, O. Sl. sla-ma Fem.) ; Lat. dimus, O. Lat. dusmus Adj. (Liv. Andr.
dusmo in loco) (cf. O. Ir. doss, ‘a bush’); Lat. limus (O. H. G. lim, Engl. lime).

(2) Of the Superlative suffix, (a) alone: Lat. minimus, briima from brévis for
*breghu- (ef. Gk. Bpaxus) ; Lat. infimus beside inférus. (b) With -is-: célerrimus,
Sicillimus, simillimus, &.. The suffix -temo- (-temo-) has its original sense in
fini-timus, legi-timus, aedi-tumus, a temple attendant, which was changed in
Varro’s time to aedi-tuus, through a false reference of the word to tucor (Varro
R. R. i. 2. 1) ab aeditimo, ut dicere didicimus a patribus nortris, ut
corrigimur a recentibus urbanis, ab ‘aedituo’; cf. Gell. xii. 10). Ultimus
(Ose. Ultiumo-) is Superl. of ulterior, citimus of citerior, &e.

§16. -RO-, -RA-. This suffix in its various forms was used
to form Adjectives and Concrete Nouns, e.g. I.-Eur. *riidhré-,
§§ 14-16.) NOUN AND ADJECTIVE STEMS. -RO-,-RA-. 329

‘red,’ from the root reudh- (O. Ind. rudh-iré-, Gk. é-pv6pds, O.SL
riidrti ; ef. O. Scan, rddva Fem., ‘blood’), Lat. river ; 1.-Eur.
*igro-, ‘a field, from the root ag-, ‘ to drive’ (O. Ind. Ajra-, Gk.
dypés, Goth. akrs), Lat. ager. In Latin the ending -7ds (and
-ris, § 40) became -er, as is seen in these two examples [cf.
ch. ii. § 15 (8)]. The suffixes -éré- and -t@rd-, which in Latin
might through Syncope lose the @ and appear as -ro-, -tro- have
been already mentioned as Comparative Suffixes, corresponding to
Superlatives in -mo- (-mmo-), -tmmo- (-t*mo-), Their original
sense however was rather that of likeness, of equal than of
greater degree, e. g. O. Ind. vatsa-taré-, lit. ‘like a calf’ (cf. Lat.
matertera), and similarly O. Ind. -tama in gé-tama-, lit. ‘like an
ox,’ &c.; and in O. Ir. the suffix -tero- retains this sense, e. g. dem-
nithir, ‘equally certain’ (not ‘more certain’), from demin, ‘certain,’
while in O. Ind. and Greek it has developed into a regular Com-
parative suffix (but cf. Hom. @ndvrepos, dypérepos). Latin nouns
like ftliaster, matraster, patraster have this suffix with a prefixed
-as- (see Ascoli, Suppl. Arch. Glott. i), while in még-is-ter, min-
is-ter, the Comparative sense belongs to the suffix -zs-, not to
the suffix -ter ; citer, eater, &e. are not Comparatives ; ci-ter-ior,
eu-ter-ior, &c. are. (On the Latin Comparative, see ch. vi. § 53).
Latin Adverbs in -?¢er have probably this suffix, e.g. dréviter
(see ch. ix. § 2); though some have explained -iter as the noun
iter, a way, so that drev-iter would correspond to the German
adverb kurz-weg. It is used in Possessive Pronouns in Latin,
e.g. vester, as in Gk., e.g. tuérepos, and in various pronominal
and locative Adjectives, with the sense of ‘like, ‘in the direction
of, e.g. al-ter, t-terum, stip-erus, to which Adverbs with -(t)ro-
correspond, e. g. intrd (ch. iii. § 15). This suffix -tero-, in Lat.
-tero- or -tro-, must be distinguished from the I.-Eur. suffix -TRO-,
which was used to form Neuter nouns indicating an instrument,
&e., e. g. dré-trum, ‘an instrument for ploughing,’ a plough, and
from the Latin suffix -cro- which represents the stem céro-,
making, from the root ker-, ‘ to make’ (cf. Lat. Cerus, creare, &c.),
e.g. ladé-cer, or stands by assimilation for -co- (I.-Eur. -tlo-) when
an Z precedes, e.g. nvdlii-crum for *involu-clum (ch. ii. § 101). This
-clo- (I.-Eur, -tlo-) is a suffix closely associated with -tro-, form-
ing Neuter Nouns which indicate a tool or instrument. Another
330 THE LATIN LANGUAGE. [Chap. V.

suffix, -dhro-, is used in the same way, though in Latin it affects
also the Feminine gender, -d7a for *.frd, *-dhra (ch. iv. § 114), e.g.
térébra, a gimlet, borer, from tero, to bore; ertdrum, a sieve, from
cerno, to sift. Latin -drus, of hondrus, décdrus, cdndrus, Sdorus,
&e. stands for -ds0- ($74); we have -érus in e. g. stvérus, procérus.

§ 17. Other examples of the RO-suffix. Lat. pro-sper, stem pro-spéro- for
*pro-spéro-, with the weak grade (ch. iv. § 51) of the root of spés (O. Ind.
sphird-, O. Sl. sporti); Lat. vir (O. Ir. fer, Goth. vair, Engl. wer-wolf; ef.
O. Ind. vird-, Lith. vyras); Lat. tndbrae for *temes-rae (O, Ind. tamis-ram,
tamis-ra, O. H. G. dinstar); Lat. céiper (Gk. xda-pos, O. Engl. heefer) ; Lat.
méicer (Gk. pax-pds) ; Lat. pla-rus, gnd-rus, in-tiger, gldber for *gladhro- (cf. O. Sl.
gladiki, ‘ smooth’), cérébrum for *ceres-rum (cf. O. Ind. sivas-, ‘head ’), mati-rus
(cf. pentria), (On Fut. Part. in -tirus, see ch. viii. § 86.)

§ 18. Examples of I.-Eur. -téro- and -Sro- in Latin: (1) attached to
Nouns. Adjectives formed with this suffix from Nouns seem to have passed
into i-stems in Latin (cf. § 34) ; they have often a locative sense and corre-
spond to Adjectives in -timus like médritimus, finitimus : e.g. camp-és-ter, silv-es-ter,
&e. which take -es- by the Analogy of Neuter ES-stems (cf. Gk. dpéa-repos, &c.) ;
éques-ter for *equit-tri-, pédester for *pedit-tri- ; péiluster for *palud-tri-, tellis-ter. The
ending -aster of Nouns or Adjectives, especially in Vulgar or colloquial Latin
(ef. Ital. giovinastro, poetastro), derived from Nouns or Adjectives, implies
likeness, and is often used contemptuously in the sense of ‘a poor imitation
of,’ e. g. péditaster, of which Plautus uses the Diminutive in the sense of ‘ tag-
rag and bob-tail soldiery’ in Mil. 54:

at péditastelli quia erant, siui uiuerent,

Antoniaster (Cie. fragm. orat. pro Vareno, 10); dleaster, wild olive, and simi-
larly dpiastrum, wild parsley, &c. ; filiaster, a stepson, mdtrastra, a stepmother,
patraster, a stepfather; surdaster, with other Adjectives indicating bodily
defects, such as calvaster, claudaster, and the diminutive ravastellus from rdvus,
grey (v. 1. gravastellus ; cf. Gk. ypats?) in Plaut. Epid. 620. Derivative Adjec-
tives show -ast(r)inus, e.g. médiastinus and mediastrinus, oleastinus, filiastinus,
(On these formations in -aster, see A. L. L. i. 390.)

(2) Attached to Prepositions, &c.: Lat. stipero- (O. Ind. tpara-, Gk. tzrepos,
O. Engl. ufer-ra with -ra for Goth. -iza), siip-er and sub-ter; Lat. inter-ior
(O. Ind. antara-, Gk. évrepor, the entrails, O. Sl]. jetro Neut., ‘the liver’) ; Lat.
exter (O. Ir. echtar), postero-, postri-dit, contra, praeter (ch. ix. § 2). Other locative
words are ci-ter (Goth. hi-dré, ‘ hither’), dex-ter (Gk. defi-repds), sin-is-ter.

(3) Forming Pronouns: Lat. t-ter, Ose. potoro- (cf. O. Ind. katard-, Gk.
nérepos, Goth. hvafar, Lith. katras) ; iterum, for another time, again, from ttero-,
other (O. Ind. itara-, ‘other’). The suffix often expresses that a pair of persons
or things is spoken of, e. g. al-ter, the other (of a pair), but dilius, another (of
many). The Possessives ‘our,’ ‘your’ take -tero- in Latin, nos-ter, ves-ter, and
Greek 7yé-repos, tué-repos, but -ero- in Teutonic, e. g. Goth. unsar, Germ. unser.

§ 19, I-Eur. -tro-. Lat. drd-trum (cf. Gk. dpo-rpov, Arm. aror, Ir. arathar,
W. arad(r) from aro (ardtus); spectrum from spécio (spectus) ; riitrum from ruo
§§17-21.] NOUN AND ADJECTIVE STEMS. -LO-,-LA-. 331

(riitus) ; vérétrum from vereor (veritus) (so fulgetrum from fulgeo ; ef. ténitru from
tono, tonitus A. L.L. i. 111); mulctrum, a milkpail, from mulgeo (muletus) ; rostrum
from rédo. We find -stro- in Lat. cépistrum, a halter, from cpio (or for *capit-
trum from cdput?), monstrum from mineo (ef. O. H. G. gal-star Neut., ‘a song,’
from galan, ‘to sing’), &. The Dim. of monstrwm is mostellum (ch. iv. § 158).

§ 20. I.-Eur. d-hro-. The Greek and O. Ir. cognates of Lat. trtbra show
the suffix -tro- (Gk. réperpoy, O. Ir. tarathar Neut.); so O. Ir. criathar, ‘a
sieve,’ for *kreitron (cf. Lat. palpébra and palpétra). Other examples of Lat.
-bra are déld-bra, from doldre, whence the name (originally a nickname) Déla-
bella, ldété-bra, lit. ‘a place for hiding,’ from lateo, vert-bra from vertére, pel-lécé-bra
from lacio, to allure (cf. «A-Opov from xndéw). Examples of Lat. -brum are
Jld-brum from flare, O. Lat. polubrum (quod Graeci xépy:Bov, nos trullum vo-
camus, Non. 544. 20 M.) from *pi-luo in the sense of ab-luo (cf. ch. ix. § 12),
delubrum, ventild-brum from ventild-re. (On the possibility of referring all these
forms to the suffix -dhlo-, see § 26). Latin -bro-, -bré- also represent an
original -s-ro-, -s-rA- (ch. iv. § 152), e.g. ténébrae (O. Ind. témisra-) from *temis,
*temus darkness (cf. témére), cérébrum (cf. O. Ind. diras-, ‘the head’) ; and some
would explain terebra, &c. as *teres-ra, from the Verbal Noun-stem teres-
seen in Inf. terere (§ 71).

§ 21. -LO-,-LA-. This suffix in its various forms was used
for Nomina Agentis (Nouns and Adjectives), and often came
to denote an instrument, while as a secondary suffix it was
specially used to form Diminutives. Thus Engl. shovel, literally
‘an instrument with which one shoves,’ meant originally ‘the
shover’; Engl. throstle, is a Diminutive. Latin examples are
légiilus, a picker, from J/ego, pendilus, hanging, from pendo,
pendeo, vinciilum, a bond, ‘an instrument for binding, from
vincio, mensiila, a little table, from mensa. In Latin, since every
short vowel in a syllable which had not the accent under the
early Accent Law (ch. iii. § 5) became before 7 the short
u-vowel, it is impossible to distinguish -élo- (e.g. Gk. vepédn,
Lat. nébila), from -tilo- (e.g. Gk. maxvdds, O. Ind. bahulé-),
&e. Further, owing to the tendency to insert a short w-vowel
between a consonant and 7 to facilitate pronunciation (ch. il.
§ 102), it is not always possible to decide whether the original
suffix was -lo- or -élo-, -tilo-, &¢., though in words like exemplum,
&e. (as contrasted with words like trémulus, &c.) it is clear that no
vowel intervened between the final consonant of the root and
the LO-suffix. The wave of Syncope which passed over the
Latin language reduced all these formations in Late and Vulgar
Latin to -dus, -da-, -lum (e.g. auri-cd-la, formed by adding the
332 THE LATIN LANGUAGE. [Chap. V.

vy
Diminutive Suffix -/a to the Diminutive Suffix -co- (§ 31), in
classical Latin auricula, became auricla, dricla, whence Ital.
orecchia and orecchio, Fr. oreille, &e. (ch. iti. § 13). From these
formations with the two Diminutive Suffixes -co- and -/o- we
must distinguish Neuter nouns formed by the suffix -tlo-,
denoting the instrument with which an action is performed, or
the place of its performance. This took in Latin the form
-clum, or with parasitic vowel -culwm, e.g. véhi-clum (véhiculum),
‘that by which one is carried? pd-clum (pdculum), ‘that out of
which one drinks,’ cibz-clum (ciibiculum), ‘the place where one
lies down’; and we have seen (ch. ii. § 154) that Plautus
generally makes this suffix monosyllabic, and the Diminutive
-co-lo- dissyllabic, e.g. véhiclum, périclum, but corctilum, uxorcila.
The suffix -dhlo- (in Latin -dii/um) had much the same function
as -tlo-, e.g. sidbulum, ‘a place for standing. The presence of
an Zin the stem of the word causes a dissimilation of -clwm to
-crum, -blum to -brum in ambiila-crum, ‘a place for walking,’ &c.
(ch. iv. § 84). Beside Neuters in -di%/~m we have Passive Adjec-
tives in -d3/is, with much the same sense as the Passive Adjectives
in -is, e.g. dgt-bilis, that can or ought to be driven, from ago,
like dgilis, ‘that can easily be driven, nimble.

§ 22. Adjectives formed by the LO-sufix. Other examples of Adjectives
expressing the action of a Verb are Lat. bibiilus from bibo, crédiilus from crédo,
trémitlus from trémo, garriilus from garrio, éminiilus from émineo, pdtulus from pdteo.
With a passive sense they become I-stems in Latin, and indicate capacity,
suitability, &c., e.g. dgilis (O. Ind. ajird-), ‘easily driven,’ nimble, from ago,
to drive, décitis from diceo, frigilis from frango, bibilis from bibo. This -li- suffix
is often added with the same sense to P. P. P. stems, e.g. coctilis from coctus,
fissilis from fissus, flexilis from flexus.

In Adjectives derived from Nouns, &c., we find -li-, e.g. hiimilis (Gk. xéapa-
Ads) from huimus, herdilis from herba, similis (Gk. dpadds) (on these see § 40).
Active Verbal Adjectives in -lo- are used in O. SI. in the periphrastic perfect

‘tense, e.g. zna-lu jesmi, ‘I have known,’ lit. ‘I am acquainted,’ from znati,
‘to know.’

§ 23. Nouns denoting the Agent or the Instrument, e.g. Lat. figiilus,
a potter, from jingo ; tzgitla from tego; régiila from régo (unless these are Diminu-
tives of *téga, *réga ; cf. col-léga ; cf. § 3); cdpiilus from cdpio; spéciila, a place of
outlook, and spéculwm, a looking-glass, from spécio, to look ; torciilum (with Adj.
torculus), a wine-press (later torcular), from torqueo; cingtilum and cingulus,
a girdle, from cingo ; jdciilum, (1) a javelin, (2) a throw-net (rete iaculum, Plaut.
Truc. 35), and jaculus, a kind of snake, from jédcio ; sella for *sed-la (Lac. Gk.
éAAd) from sédeo ; grallae from grddior: pilum for *pinslum, a pestle, from pinso.
§§ 22-25.] NOUN AND ADJECTIVE STEMS. -LO-,-LA-. 333

Often we have the terminations -slo-, -sl4-, e. g. -dlum for *ans-lum, from ans-
(cf. 0. Sl. ach-ati, ‘to be fragrant’), a byform of the root an-, ‘to breathe’ ;
balus for *pax-lus (ef. Diminutive paaillus), from pango ; vélum, a sail, for *vexlum
(cf. vexillum), from vého (cf. 0. 81. veslo, ‘a rudder’). Nouns in -éa may be derived
from Neuter es-stems (e g. quéréla for *queres-la, séquéla for *seques-la) or may have
had originally 6 (@. g. ci-cindéla, w glow-worm, from candeo, cundé-re, like Gk.
pupndds from yipéouar). These nouns in -2la were in Late and Vulgar Latin
confused with Diminutives and became querélla, sequélla, &e. (see ch. ii. § 1 30),
just as camélus became caméllus, cuctilus became cucullus, anguila (A. LL. viii. 442)
became anguilla, We have -rum for -lum by dissimilation of 1 in scalp-rum
from scalpo, &e.

§ 24. Diminutives. Lat. cistula; auxilla from aula (olla) for *aus-la; porculus
(Germ. Ferkel) ; servolus; filiolus; Umneola; lactucula from lactiica ; Wqude-ulus ;
misellus (occasionally misérulus); Atella (Osc. Aderla-); asellus from dsinus ;
gemellus from gémini ; lapillus from léipid-. Sometimes this termination is added
a second time, e. g. cistella from cistula ; ollula from olla ; porcellus from porculus ;
asellulus from asellus ; gemellulus from gemellus. Sometimes it is added to the
Diminutive suffix- co-, e. g. olli-cu-la, servi-cu-lus, cor-cu-lum, légiun-cu-la, cdni-cu-la
(ef. febri-culisus, méti-culdsus), dpi-cu-la, vallé-cu-la, dié-cu-la, corpus-cu-lum, arti-cu-lus.
The Diminutive retains the Gender of the simple Noun, unlike Greek Diminu-
tives in -.ov, which are Neuter (A. L. L. iv. 169). This suffix -ciilus gives to
Adjectives the sense of ‘somewhat,’ e. g. mélius-culus, ‘somewhat better,’ and
other Comparatives like plus-culus, majus-culus, &c. ; also grandi-cu’us, dulei-culus,
lévi-culus, &. With the Diminutive suffix -lo-, familiar or pet names are often
formed in I.-Eur. languages, ve. g. Gk. Opacd-Aos, the familiar form of @paav-
paxos, Goth. Vulfi-la. The gradual weakening of the diminutive force of
these suffixes, which is to some extent accountable for the doubling of the
suffix in puellula, cistella, &c. (cf. anellus from dnulus, Dim. of anus, a large ring,
e. g. Plaut. Men. 85 anum lima praeterunt) is seen in words like ancilla (Fem.
of serrus), which had ceased to be a Diminutive as early as the time of Plautus.
Adulescentulus always differs from ddi'escens in Plautus, but in Terence is hardly
distinguishable. Diminutives were a feature of Vulgar Latin, as we see from
the forms censured in the Probi Appendix : juvencus non ‘ juvenclus’ (197.
29 K.); catulus non ‘catellus’ (198. 2), auris non ‘oricla’ (198. 11; ef. Ital.
orecchio, Fr. oreille) ; fax non ‘ facla’ (198. 23); neptis non ‘ nepticla,’ anus
non ‘anucla’ (199. 1); mergus non ‘mergulus’ (199. 7). The ending -éllus,
as we have seen, may denote a Diminutive of a LO-Diminutive, e.g. dnellus
(on -ell-, see ch. iv. § 10), Dim. of dnulus, or the Diminutive of a Noun with a
RO-suffix, e.g. dgellus from dgro- for *agré-lo- (ch. iii. § 15. 8), or of a Noun
with @ in the penult, e.g. femella from femina (earlier -mena). Similarly we
have -illus for -id-lus in lapillus, lla for -én-la in catella, -illum for -in-lum in villum,
-tillus (older -dllus) for -dn-lus in hémullus, -dlla for -dn-la in corolla, -cillus for -tin-
lus in illus, -dllus for -dn-lus in Hispallus, &e. (cf. nitédula and nitélla).

§ 25. Neuters formed with the Suffix -tlo-. Lat. discerniculum (‘acus quae
capillos mulierum ante frontem dividit: dictum a discernendo,’ Non. 35.
29 M.); pideclum (pideulwm), ‘a means of appeasing the gods,’ a victim, then
‘a sin for which the gods must be appeased,’ from pidre ; récepticulum from
receptére ; pavicula, a mallet, from pdvire; sediculum from sédeo (‘sediculum *
334 THE LATIN LANGUAGE. [Chap. V.

sedile, Paul. Fest. 500. 9 Th.) ; dperculum from operio (opertus) ; stpulerum, from
stpelio (sepultus). We have -crum by dissimilation of 1-1 in léud-crum, ‘a place for
bathing,’ from lavare ; involiicrum from involvo, &e.

§ 26. The suffix -dhlo-. Lat. ldtibulum, ‘a place for hiding,’ from Jateo ;
venabulum, ‘an instrument for hunting,’ a hunting-spear, from vendri ; vecté-
bulum (quod nune vehiculum dicitur, Non. 54. 26 M.) ; conciliabulum (locus ubi
in concilium convenitur, Paul. Fest. 27. 9 Th.) ; ef. fabula, siibula. We may
have -brum by dissimilation of 1-1 in lévabrum ; and it is possible that some, or
all, of the examples of Lat. -brum (I.-Eur. -dhro-}, quoted in § 20, had
originally -blo- and not -b70-, e.g. po-lubrum. Just as I-stem adjectives with
a Passive sense like dgilis, ‘ easily driven,’ are connected with Neuter Instru-
mentals in -lo- like dgitlum (‘agolum,’ pastorale baculum, quo pecudes aguntur,
Paul. Fest. 21. 37 Th.), so we have Passive I-stem Adjectives connected with
the suffix -dhlo-, e. g. amé-bilis, horri-bilis, flé-bilis, vili-bilis, mo-bilis, intelligi-bilis,
&e. (ef. Umbr. facefele ‘facibile’), sometimes derived from the P. P. P. stem,
e.g. persudsi-bilis, flexi-bilis, sensi-bilis, and sometimes showing -bris (-ber) for
-bilis, when an 1 precedes, e. g. dilébris (and dlibilis), anclabris. For a list of Adjs.
in -bilis in early authors (e.g. ndbilis, known, Plaut. Pseud. 1112 neque illis
nobilis fui; cf. Pacuv. Trag. 221 R.), see Hanssen in Philol. xlvii. 274, who denies
that they ever have a transitive sense, e.g. incogitabilis, ‘thoughtless,’ Plaut.

§ 27. -T0-, -TA-. This I.-Eur. suffix was used to form
(1) Verbal Adjectives, which in Latin and some other languages
have the function of perfect participles passive, e.g. géni-tus from
the root Sen-, while with the negative particle prefixed they
may express incapacity, e.g. I.-Eur. *nmrto-, ‘incapable of
being killed,’ immortal (O. Ind. amfta-, Gk. &uBporos); (2)
Ordinal Numbers, and when added to the Comparative suffix
-is- (§ 76), Superlatives ; e.g. I-Eur, *stksto- (O. Ind. 3a8-thé-,
Gk. ék-ros, Goth. saihs-ta, an N-stem), Lat. sextus ; I.-Eur. *Ok-
is-to- (O. Ind. a-iktha-, Gk. é«ioros). In Latin this formation
of Superlatives is not found, but another, e.g. dcisstmus (ch. vi.
§ 54). Abstract Nouns in -ta- are found beside Verbal Adjectives
in -to-, eg. Gk. yevery, birth, Goth. junda ‘juventa’1, which
occasionally pass into a concrete sense and become Masculine.
e.g. yevérns, tmadrns (ef. Lat. égues from an earlier Abstract
Fem. *equita?), though the suffix in Latin was ousted by -tas,
-tido, &e. (§ 67). I-Eur. -to- is often seen added to the MEN-

 

1 Lat. jiventa seems to be aforma- sense of ‘period of youth’ and ‘a
tion on the analogy of stnecta (se. number of young men.’ (Fleck. Jahrb.
aelas), for juventis is the form used Suppl. 1891.) We have aetate iuenta on
ly the oldest writers, both in the an inscription (C. I. L. i. 1202).
§§ 26-28.] NOUN AND ADJECTIVE STEMS. -TO-,-TA-. 335

suffix (§ 54) in neuter nouns, e.g. Lat. cognd-men-tum beside
cogno-men, such forms being apparently the Neuter of Participles
or Verbal Adjectives, formed not from verbs but from nouns,
e.g. *cogno-mentus beside cogno-mindtus, like scélestus beside
scelératus (cf. Eng]. compounds like ‘ bare-footed,’ ‘ black-headed,’
where the participial suffix is added to the nouns ‘foot,’ ‘ head’).
The forms with -mentum are, as a rule, those used by prose
writers, the forms with -menx being relegated to poetry. With
-mentum, Plur. -menta has been compared Gk. Plur. -yara, e.g.
otpépara (Lat. strdmenta), cacovpara (cf. Lat. assimenta).

§ 28. Participles in -tus. The weak grade of the root is used with
J.-Eur. verbal adjectives in -to-, and the suffix is accented, e. g. I.-Eur. *klti-t6-,
‘heard, heard of, famous,’ from the root kleu-, ‘to hear,’ [O. Ind. gruté-, Gk.
«dtTés, Which perhaps retains the old sense of ‘heard,’ ‘loud’ in such Homeric
phrases as «Aura pda, O. Ir. cloth, from *cluto-, O. H.G. Hlot-hari (from
O0.H.G. hari, Germ. Heer), the name Lothair, corresponding to Greek
KaAuré-orparos ; ef. Zend srita-, O. Eng. hlid, ‘loud’], Lat. in-cliitus ; though
with the noun we often find the high-grade and the accent on the root, e.g.
GE. xoirn from xefua, ofros from ef, &e. In Latin -sus replaced -tws when the
verbal stem ended d or ¢ (ch. iv. § 155), e.g. salsus from sallo for *saldo ; occa-
sionally the P. P. P. took -sus when the Perfect Ind. had -si, e. g. tersus from
tergeo (Perf. Indic. tersi), though the older spelling appears in the earlier
literature, e. g. tertus (Varro) (see ch. viii. § 92; ch. iv. § 155).

Examples of Latin participles in -tus are stratus (from the root ster-,
O. Ind. strta-, Gk. orpwrds, O. SI. -stritii) ; malctus from the root melg- (Lith.
milsztas) ; com-mentus from the root men- (O. Ind. matd-, Gk. aird-parTos,
Goth. munds, Lith. mifitas, O. Sl. metii); Lat. gndtus from the root gen-,
*gna- (O. Ind. jaté-, Goth. -kunds); Lat. situs for *sytito- (O. Ind. syité-,
Gk. veo-xdrriros, Lith. sidtas, O. Sl. Siti ; Lat. gnotus (O. Ind. jhatd-, Gk. yards,
O. Ir. gnath); Lat. junctus (cf. O. Ind. yuiktd-, Gk. (everds). The different
treatment of the stem vowel before the suffix is exemplified by dmédtus, démitus
from first conj. Verbs; viétus, vigétus, exercitus (but ar(c)tus), monitus (but
Monéta), mulctus from second conj.; dlitus and altus, cultus, factus (but ficetus)
from third conj. ; finitus and dpertus from fourth conj. (On these Verb-stems
see ch. viii.) ; aegrd-tus, argii-tus, inclii-tus, citus and citus. Words like pilati,
-armed with the pilum,’ barbatus (O. Sl. bradatti), auritus, cinctutus, do not of
course imply the existence of verbs, *pilare, *barbare, *aurire, &c. Of similar
formations from Noun- or Adjective-stems examples are: Lat. libertus beside
libératus ; Onustus beside onératus ; stnectus from senex ; and the words indicating
a place planted with trees, &c., e. g. arbus-tum, siilic-tum, the Neuters of arbustus,
‘provided with trees ’ (arbustus sive silvestris, Columella), &e. (Arbiretum is
a quasi-participle from *arboreo, arboresco, like décétum from acesco ; so ntic-etum,
pin-etum, iic-etum. The -cétum of ilicetwm, &c. and the -ctum of salictum, &e.
were extended by false analogy, e.g. bucetwm, virectum). For a list of Adjec-
tives in -estus, -ustus, -titus with this sense of ‘provided with,’ see Rénsch,
Collect. p. 217, and cf. Plaut. Capt. 392: qui me honore honestiorem semper
336 THE LATIN LANGUAGE. (Chap. V.

fecit et facit). Latin participles in -to- have often become nouns, e.g. tectum,
legatus, répulsa, sinecta, or Adjectives (ch. viii. § 92), ©.g. sanctus, latus, broad
(for *stlatus, from the root stel-, ‘to extend,’ O. SI. stelja).

§ 29. Abstract Nouns in -ta (-sa!. The Fem. of the Participles répulsus,
déprensus, &e. is used in an abstract sense: repwsa, ‘defeat at an election,’
deprensa, ‘genus militaris animadversionis, castigatione major, ignominia
minor’ (Paul. Fest. 50.30 Th.). These Abstracts must be distinguished from
Concretes like torta (se. pldcenta), a roll, expensa (se. péciinia), « sum expended
(for a fuller list, see Ronsch, Collect. p. 195). The Fem. Abstract *equita, ‘horse-
manship,’ seems to have been made a Mase. Concrete in O. Lat. with the
sense also of ‘a horse’; thus Ennius (4. 249 M.), describing « charge of
cavalry and elephants, says:

denique ui magna quadrupes eques atque elephanti
proiciunt sese,
a usage imitated by Virgil (@. iii. 116):

equitem docuere sub armis
insultare solo et gressus glomerare superbos,

and commented on by Aulus Gellius (xviii. 5; cf. Non. 106. 24 M.).

§ 30. Neuters in -mentum. Other examples are augmentum beside augmen
(poet.): fundamentum beside fundamen (poet.) ; intégiimentum beside tegumen,
tégimen, tegmen ; cognomentum beside cognomen ; termentum and trimentum beside
térimen ; argamentum, from arguo,‘a making clear,’ a proof, then ‘the subject of
a story, picture, &c.,’ e.g. Virg. A. vii. 791 arguinentum ingens; Prop. ili. 9.
13: argumenta magis sunt Mentoris addita formae ; vestimentum, from vestio ;
caementum for *caed-mentwm (ch. iv. § 161) from caedo; jugmentum (et paries,
C.I.L. vi. 24710), and jugwmenitum (Cato) ; jiimentum from juvo, according to
Augustine, Quaest. in Heptat. iii. 2 and v. 38; mdmentum beside mdmen (poet.)
from méveo. Nien, crimen, culmen (ef. cihiimen), &c. have no byforms in -mentum ;
implementum, incrémentum, mintimentum, &c. have no byforms in -men.

§ 31, -KO-, -KA-. I.-Eur. -/o-, the -y of Engl. ‘stony,’
‘angry,’ &., is rarely -ko- (with palatal k), e.g. L-Eur. *yittwnko-
(O. Ind. yuvasd-, O. Ir. dac, W. ieuanc, Goth. juggs), Lat.
jiévencus, but usually -ko- or -q#o-. It is used as a primary
suffix, e.g. cascus (cf. cdnus for *casnus, Ose. casnar) fuscus (cf.
furvus for *fusuus), but mainly as a secondary suffix employed
in the formation of Adjectives from Adverbs, e.g. anticus (of
place), antiguas (of time) from ante (cf. O. Ind. antiké-), Nouns,
e.g. bellicus, civticus, and Adjectives, e.g. O. Ind. ndégnaka-,
‘naked,’ beside nagna-, having often a diminutive significance,
which is in Latin denoted by -cii/us (§ 24), the addition to -co-
of the other diminutive suffix -lo- (§ 21), e.g. xigriculus beside
§§ 81-38.] NOUN AND ADJECTIVE STEMS. -K6-,-KA-. 337

niger, Svicula (0. Sl. ovica), [adbicare and nigricare are Verbs
with the Diminutive KO-suffix (ch. viii. § 33 (7)], (cf. 20mun-c-io
and homun-cu-lus, sénéca, Non. 17.18 M., senecio and seniculus), as
in Gk. by -toKo- of madicxos, &c. We also find it preceded by
1, e.g. elpwv-txds, Lat. histridn-icus (without the vowel we should
have had *istriuncus), often -tico-, e.g. rus-ticus, herba-ticus,
errd-ticus after the analogy. of Participle-stems in -to- (§ 28);
by 1, e. g. Goth. mahteigs,‘ mighty,’ Lat. dmicus, piidicus (7 not ei
is attested for mendicus by Plaut. Rud. 1305; see ch. iv. § 34); by
a, e.g. méracus. Adjectives in -ako- had the sense of English
adjectives in -ish, e. g. Lith. saldékas, ‘ sweetish, and came in some
languages to acquire the force of Comparatives, e.g. Lett.
salddks, ‘ sweeter, W. glanach, ‘ fairer.’

In Latin the -ko- and -q*o- suffixes seem often to have been
confused, e.g. anticus and antiquus, tesca or tesqua, lit. ‘dry
places,’ for *¢ersc-, from the root ters-, ‘to dry’ (cf. torreo, Gk.
répcouat), and, as in Greek, &c. there are often byforms of the
Consonantal declension, e. g. Libdw, &e. beside merdeus, felix, &e.
beside pudicus, &c., like Greek AE and 7Alkos, wetpa€ (O. Ind,
maryak4-). For -icus we have -icius in aedili-cius, tribinicius, pa-
tricius, adventicius, commendaticius, &c.; for -deus we have -dceus
in herbaceus, gallindceus, drundindceus, &e.

§ 32. Adjectives with the KO-sufix. From the adverbs ré- and pré- we
have réci-précus, while procul adds the lo- (li-) suffix to *procus (0. SI]. prokt)
(Rhein. Mus. xliii. 402) ; posticus is the opposite of anticus ; pris-cus (cf. primus
for *pris-mus). From the numeral anus we have unicus (Goth. ainahs, O. Sl.
inoktt; cf. O. Ind. dvika-, Gk. dads for *5ficos). From nouns : patricus (ef.
patricius); hosticus, used in Plautus like hostilis (hostica manus, Capt. 246 ;
hostilis manus, Capt. 311) ; civicus (beside civilis). (Cf. Porph. ad Hor. ¢. iia.a
adtende autem non ‘civile’ sed ‘civicum’ dixisse antiqua figura. illi enim
‘civica’ et ‘hostica,’ deinde ‘civilia’ et ‘hostilia’ dicebant. denominationes
autem hae fere liberae sunt apud doctos). Other noticeable forms are
ciidacus, mandicus (ef. fidie-ia from *fidiicus, and panniicia) ; lingtildca, verbendca ;
hiulcus, pétulcus from hiare and petere with the (diminutive ?) LO-suffix. (For
a list of Nouns in -aea, e. g. verruca, a wart, in OQ, Lat. a hill (Cato ap. Gell.
iii. 7.6; Quint. viii. 3. 48 and viii. 6. 14), from the root wers- of Lith. virszis,
‘the top,’ O. Ind. vargman, ‘a height,’ see Stolz, Bettriige, p. 6, who points out
the connexion of this ending with the ending -ago, e.g. Verrugo, a Volscian
town in Latium.] A gloss has : manubrium, quod rustici ‘manicum’ dicunt
(C. G. L. v. 115. 47° (manico- in Romance, e. g. Ital. manico, Span. mango, Fr.
manche) ; Paul. Fest. quotes olentica ‘mali odoris loca’ (223. 4 Th.).

§ 33. Adjectives in 4Acius. Denominatives (i.e. derivatives from Nouns

' Zz
338 THE LATIN LANGUAGE. [Chap. V.

or Adjectives) have -i-, e. g. patricius from pater, natalicius from nidtdlis ; Deriva-
tives from P. P. P. in -to (Vb. Nouns in -tio-, -ti-) have -1-, e.g. dediticius from
deditus, insiticius (cf. insitivus) (also ndvictus). (For a full list, see A. L. Lv. 415.)

§ 34. Suffixes ending in i (Nouns and Adjectives of third
Declension). -i-. In the declension of these stems 1 varies with
ei and oi (see ch, vi.) The [-stems are often confused with IO-
stems (§ 4), with I-stems (§ 51), and with YE-stems (§ 51).
Examples of the primary suffix -i- are Lat. anguis (Lith. angis,
O. Sl. aZi, Arm, auj), and another I-Eur. word for a snake,
*8Shi- (O. Ind. hi-, Zend aZi-, Arm. i2, Gk éy:s).

Neuter I-stems in I.-Eur. (like Neuter R-stems, &c., see § 56)
seem occasionally to show a heteroclite declension, a nasal
replacing the vowel in oblique cases, e.g. O. Ind. ksi, ak$nds
Gen., ‘ the eye, Asthi, asthnds Gen.,‘ a bone, which may have led
to a confusion of I- and N-stems in such words as I.-Eur.*aksi-,
‘an axle’ (Lat. awis, Lith. aszis, O. Sl: osi; but Gk. dfwv). As
a secondary suffix -i- is used in many languages to give an
adjectival sense, especially in the formation of Compound
Adjectives from Nouns, e.g. Lat. ewsomnis from somuus, O. Ir.
essamin for *exomni-, from omun (*omno-), ‘fear, though we
often find in the early Latin literature the O-stem, e.g. tmermus,
class. inermis (so Gaulish Lxobnus, Exomnus for O. Ir. essamin).

In Latin the I-declension has been greatly extended; thus
consonantal stems of the third declension often take the i-stem
case-suffixes, e.g. péd-t-bus, férent-ium, &c.; and a stem-suffix
ending in 6 or ti passes readily, if it be an adjective, into an
I-stem, e.g. similis (Gk. duadds), hiimilis (Gk. xPapadds), avis
(I.-Eur. *nau-, O. Ind. nats, Gk. vais), brevis (cf. Gk. Bpaxds), levis
(cf. O. Ind. raghus, Gk. é-Aayds, Lith. lenguris). The Greek
loanword Ailirus (tdapés) is also hilaris by the time of Terence.

§ 35. Other examples of I-stems. Lat. ensis(O. Ind. asi-) ; Lat. triidis from
trido (root treud-, Goth. us-priutan, ‘to trouble,’ Engl. thrust, threat, O. Sl.
trudt, ‘toil’); Lat. riidis, unworked, rude, originally of metal, from the root
reudh-, ‘to be red’ (ef. raudus, unworked metal, ch. iv. § 41) ; jagis, continual,
from jungo; Lat. scébis from scdébo; Lat. médre (O. Ir. muir Neut., for *mori,
0. H.G. meri Neut., Germ. Meer Neut., Engl. mere).

§ 36. Adjective I-stems from O-stems. Other examples are: O. Lat.
sublimus, e.g. Enn. Tray. 2 R. deum sublimas subices, Luer. i. 340 sublimaque
caeli (see Munro’s note) ; O. Lat. stérilus, e.g. Lucr. ii. 845 sonitu sterila (ef.
Paul. Fest. 463. 1 Th. ‘sterilam’ sterilem). Greek aternpés had the same
tendency to the I-declension in Latin as Greek iAapés, for Caper gives
§§ 34-40.] NOUN AND ADJECTIVE STEMS. -I-, -RI-, -LI-. 339

a caution against the form austeris (p. 108. 4 K.). Nonius (494. 26 M.) quotes
pronis from Varro (Men. 391 B.). On O. Lat. forctus, see ch. viii. § 92.

The Noun torris from the root ters- (Gk. répoopat ; cf. Lat. torreo for * torseo,
ch. viii. § 23) was in O. Lat. torrus (Non. 15. 22 M.; Serv. ad Aen. xii. 298
‘hic torris’. .. ita nune dicimus: nam illud Ennii et Pacuvii penitus de usu
recessit ut ‘hic torrus, hujus torri’ dicamus), but is rather to be explained as
an E-stem become an I-stem (like sordes, § 51). The form torres F., a burn-
ing, read by Lachmann in Lucretius iii. 917 (MSS. torret) is doubtful (A. L. L.
viii. 587). On hilarus beside hilaris, see Neue, ii’. p. 149.

§ 37. -NI-. This suffix, varying with -éni-, -ini-, &c. is more
common in those languages which have extended the use of the
P. P. P. in -no- (§ 8) than in Latin. Examples are: Lat. agnis
(ef. O. Ind. agni-, Lith. ugnis F., O. Sl. ogni M.), Lat. clénis

S . . . . .
(O. Ind. Sréni-, O. Scand. hlaunn, Lith. szlaunis). Adjectives in
-nis may have been originally NO-stems (§ 36), e.g. immanis
from O. Lat. manus, good (with a bystem in -ni-, Manes, lit.
‘the good deities’); cf. Janis, a byform of Janus, in Carm. Sal.
(Tert. Apod, 10).

§ 38. Other examples of Latin -nis. Lat. amnis for *ab-nis (ef. O. Ir.
abann), Lat. crinis for *crisnis (cf. Lat. crista), panis for *pasnis (cf. Lat. pastillus).

§ 39. -MI- is a still rarer suffix than -ni-. An example is
Lat. vermis (O.H.G. wurm), a word the relation of which to
I.-Eur. *q2rmi- (O. Ind. kfmi-, O. Ir. cruim, W. pryf, Lith.
kirmis) is not quite clear. Latin Adjectives in -mis like iermis,
sublimis were originally -MO-stems (cf. § 36); thus Nonius
489. 7 M. gives some examples of sublimus from the older
literature (cf. Georges, Lew. Worlf. s.v.). °

§ 40. -RI-, -LI-. These are not nearly so common I.-Eur.
suffixes as -ro-, -lo-, e.g. O. Lat. dcris, a hill (in Umbrian, &c.
used of the citadel) (Gk. dxpis, a point, depis, a hill; cf. O. Ind.
é&ri-, but also Gk. dxpos), Lat. td/is, qualis (O. Sl. tol Adv., koli
Adv.; cf. Gk. rnAl-kos, mmAi-cos); but -li- is fairly frequent in the
Slavonic languages, where the P. P. P. in -lo- is much in vogue.
In Latin, Adjective -RO- and -LO- stems often show -7is, -/is, e.g.
O. Lat. saeres, used of animals for sacrifice (e.g. Plaut. Rud.
1208 stint domi agni et pérci sacres) beside sicer (cf. Manes
beside O. Lat. manus, &00d), similis (Gk. duadds), humilis (Gk.
x9apadds). And the use of -h- and «ri- for Adjectives derived
from Nouns is very widely extended in Latin, far more widely

2
340 THE LATIN LANGUAGE. [Chap. V.

than in any other I.-Eur. language, e.g. vitdlis from vita, alaris
for *aalis(?) from dla, so that the Greek Adj. dayArjs, when
borrowed by Latin, assumed the form dapsilis. Dialectal
examples are Osc. luisarifs, ‘ *lusaribus,’ Sab. Flusare ‘ Florali,’
while Ose. Fiuusasiais ‘ Florariis’ has -dsio- (Lat. -ario-, § 4).

§ 41. Other examples of Latin -li-, -ri-. ‘iter, a skin, for *id-ri-, beside
idérus (O. Lat. uterum) ; deer, sharp (cf. Gk. dpos). Verbal Adjectives in -tis
are, as we have seen (§ 22), byforms with Passive sense of Active Verbal
Adjectives in -lus, e.g. dgilis, ‘easily driven,’ bibilis, ‘easily drunk’ (beside
dibiilus, ‘easily drinking,’ inclined to drink), décilis (like décibilis) ; sometimes
formed from the Perf. Part. Pass., e.g. j/ictilis, fissilis, flevilis (and flexibilis),
hamatilis et saxdtilis (Plaut. Rud. 299), missilis, The ending -iis of aedilis has
I,-Eur. i, not ei, to judge from old inscriptions, v. g. C. I. L. i. 61 aidilis (ef. ib. 31).
This -zis (proper to I- and 10-stems) as well as -alis (proper to A-stems) is often
extended by analogy ; e.g. cinilés (for *analis) by analogy of sénilis ; vernilis is from
verna, verndlis from ver ; from libra, a pound, we have librilis and libralis ; from
scurra, scurriis ; from manu not *manilis, but manudlis (so dorsudlis from dorso-).
Examples of -élis are fidélis from fide-, crudélis (ef. famé-licus, contiimé-lia) as well-
as patrudlis and matruélis. Like tribiilis (from tribu-), idalis (from tdu-), is édiilis.
From O-stems we have puérilis, virilis, hérilis, servilis, &e. ; anndlis, fatilis, &e. ;
from Cons.-stems cépit-dlis, virgin-dlis, hospit-dlis, jivéndlis and jitvénilis, &e.

§ 42. -TI-. As -to- was the I.-Eur. suffix of Verbal Adjectives,
especially of the Perfect Participle Passive, so -ti- was the suffix
of Verbal Nouns (Nomina Actionis), e.g. Gk. miords and tiotis.
These nouns were of the feminine gender, had the weak form of
the root, and are accented sometimes on the suffix, and some-
times on the root, e.g. O. Ind. mati- and mAti- for I.-Eur.
*mn-ti- (Lat. mens) from the root men-,‘to think.’ In Latin, as
in Celtic, we find this suffix enlarged by an EN-suffix, e. e. Lat.
mentio, Acc, mentionem, O, Ir. -mitiu-, -mitin Acc., and this
compound suffix, which in Latin supplanted almost entirely the
older -ti-, appears also in other languages, e.g. Goth. rapjé (Lat.
ratio), Gk. dwrivn (cf. Lat. détio, Acc. dationem), The sup-
planting may be accounted for, wholly or partly, by the change
which the phonetic laws of the Latin language would produce
in the TI-suffix, a change which would often make the suffix
unrecognizable. From the root men-, the Verbal Noun *mnti-
became mens in Latin and from the root bher-, the Noun *bhrti-
became fo7's, while from méto we have messis for *met-tis, from
the root wes-, vestis, from siétidre, sitids, &c., forms whose
common formation has been obscured past recognition. A similar
§§ 41—44.] NOUN AND ADJECTIVE STEMS. -TI-. 341

explanation has been offered for the fact that in Teutonic also
the same suffix -ti- ceased to be a living suffix, namely that
under the working of the Teutonic phonetic laws it would
assume the various forms -pi, -di, -ti, -si, &. The older suffix
remains in Adverbs like raptim, furtim, &e. (see ch. ix. § 4).

Occasionally the feminine abstract passed into a concrete
noun and might change its gender. Thus Lat. Zostis (Goth. gasts,
‘a guest, O.S]. gosti) may have been originally abstract, just as
Lat. agricd/a, a field-tiller, meant originally ‘ field-tillage’ (§ 2).

The secondary suffix -tati- (or -tat-), used to form feminine
Abstract Nouns, derived from Adjectives and Nouns, and the
similar suffix -titi- (or -tiit-) exhibit this suffix -ti- added to the
suffixes -ta- and -tu-, e.g. Lat. juventds beside juventa (see § 27).
The suffix -tati- (-tat-) is found in O. Ind., Greek, and Latin, while
-tiiti- (-tit-) is found in Latin, Celtic, and Teutonic, e.g. Lat.
novitas (Gk. vedrys), Lat. anitis (O. Ir. oentu), Lat. gaventis
(O. Ix. ditiu). In Latin, in addition to -titi- (-tit-), which is
much less in use than -tati- (-tat-), we find a form augmented
by an N-suffix, -t#do, Gen. -tidinis, e.g. servitado beside servitis,
hilirttudo (so in Plaut., not hilaritas), bedtitudo, which (with
beatitas) was a coinage of Cicero (Quint. vill. 3. 32).

§ 43. Other examples of the suffix -ti- in Latin. Latin vectis, a lever (cf.
vectio), from vtho; Lat. vitis from the root wei-, ‘ to plait, weave’; Lat. messis (cf.
messio) from méto; Lat. fors (O. Ind. bhrti-, O. Ir. brith, Goth. ga-baurps,
0. H. G. giburt, 0. Engl. gebyrd, ‘ fate’) from féro; Lat. mens (O. Ind.
mati- and mati-, Goth. gamunds, ana-minds, Lith. at-mintis, O. Sl.
pameti) from the root men-, ‘to think’ (ef. mentio) ; Lat. gens, from gigno
(cf. natio for *gnatio); Lat. mors (O. Ind. mrti-, 0. Lith. mirtis, 0. SI.
si-mriti) from mérior; Lat. dds (O. Ind, dati-, Lith. di&tis, O. Sl. dati)
from root do- (ef. détio, Gk. dwrivn) ; so Lat. cis beside cdtus (ch. iv. § 54); Lat. ars
(O. Ind. rti- means ‘attack ’) ; Lat. pars (cf. portio) ; Lat. grates beside gratus ;
Lat, quiés beside quiétus (inquies for inquietus is due to the tendency to turn
Compound Adjectives into I-stems) ; Lat. vestis from the root wes-, ‘to clothe,’
meant originally, like our word ‘ clothing,’ the act of arraying oneself, then
the raiment itself.

§ 44. Examples of Lat. -tion-. Lat. visio (ef. O. Ind. vitti-, O. Sl. -visti
and vésti) from the root weid-, ‘to see, know’ ; Lat. -ventio (ef. O. Ind. gati-,
Gk. Baors, O. H. G. cunft, Lith. -gimtis) from the root g%em-, ‘to come’; Lat.
siitio (cf. Gk. dv-eors, Goth. -séps, seed,’ Lith. séti) from the root sé-, ‘to throw,
throw seed’ ; Lat. con-ditio for *con-diitio (cf. 0. Ind. -hiti-, Gk. 6éo1s, Goth. -déps,
‘a deed,’ O. SI. -d&ti) from the root dhé-, ‘to put, place’; Lat. dd-eptio for *ad-dptio
(ef. O. Ind. apti-) beside aptus ; Lat. ea-plétio (cf. O. Ind. prati-, Gk. mAjais),
beside ex-plétus ; Lat. ndtio for *gnotio (cf. O. Ind. -jnati-, ‘the act of knowing,’
342 THE LATIN LANGUAGE. [Chap. V.

jfati-, ‘an acquaintance,’ Gk. yv@ous, O. H. G. ur-chnat, O. Sl. po-znati, Goth,
ga-kunps, ga-kunds, Germ. Kunst, 0. Sl. zeti, ‘son-in-law’) beside (g)notus; Lat.
dictio (ef. O. Ind. di&ti-, Gk. Seifis, 0. H. G. -ziht) from the root deik-; Lat.
stitio (ef. O. Ind. sthiti-, Gk. ordows, O. H. G. stat, O, Sl. -stati) from stare ;
Lat. in-tentio (ef. O. Ind. tati-, Gk. réos) from the root ten-, ‘to stretch’ ; Lat.
junetio (ef. O. Ind. yukti-, Gk. (edgis, Lith. junkti) from jungo.

§ 45. Adjectival -ti- for -to- in Latin. The same tendency to turn O-stem
Adjectives into I-stems, which we have seen in similis (Gk. dpuadds), hiimilis
(Gk. x@apadds), appears in an adjective like fortis, in O. Lat. forctus (Paul.
Fest. 73.9 Th. horctum et forectum pro bono dicebant), originally a P. P. P.
from the root dhergh-, ‘ to establish’ (O. Ind. drdha-, ‘ stablished, firm’). The
older spelling forctis occurs in the clause of the XII Tables quoted by Festus
(524. 15 Th.), which provided: ut idem juris esset Sanatibus quod Forctibus,
the Sanates being allies who had revolted but had returned to their allegiance,
‘quasi sanata mente,’ the Forctes being those who had never broken faith (ef.
Paul. Fest. 59. 26 Th. forctis (MSS. forctes), frugi et bonus, sive validus]. The
word Saneates shows a similar transference to the I-declension of the P. P. P. of
sdnare, and throws light on the termination, -ds O. Lat. -dtis (see ch. iii. § 16),
which indicates the country or the party to which one belongs, e.g. opti-
mates, infimatis (Plaut. Stich. 493), Arpinates.

§ 46. Other examples of Lat. -tat(i)-, -tiit(i)-, -tidin-. The tendency of
the I-stems to encroach on the Consonant-stems in Latin (mentioned above,
§ 34) makes it difficult to decide when -tat-, -tiit-, and when -tati-, -tiati-
were the suffixes used, e.g. Gen. Pl. civitatum and civitatium. In O. Ind. -tat-
(e.g. sarvatat-, ‘completeness ’) is rarer than -tati- (e. g. sarvatati-), while in
Greek only -tat- is found. Other Latin examples of -tds (see A. L. L. viii. 321)
are scaevitas (Gk. cxardrys), comminiias (Goth. gamaindtps), binitas from bonus,
lbertas from liber, cwitas from civis, ubertas from aber, ftécultas and facilitas from
ficitis, viluptas from voliip(e), tempestas from tempus (ef. tempéri), viluntas from
vilens. Examples of -tiis: sénectus from sénex, virtus from vir; O. Lat. tempestus
(Varro L. L. vii. 51 libri augurum pro tempestate ‘tempestutem’ dicunt
supremum augurii tempus). Examples of -tédo: altitudo from altus, latitudo
from latus, sollicitudo from solltcitus.

§ 47. Suffixes ending in -i (Nouns of fourth Decl.).
-U-. In the declension of these stems ti varies with eu and
ou (see ch. vi.), I.-Eur. Adjectives in -u- usually show the weak
form of the root, and are accented on the suffix, e.g. *plti-,
‘broad’ (O. Ind. prthi-, Gk. aAards, Gaul. litu-). They formed
their Nom. Sing. Fem. in -wi, e. g. *swadu-, ‘ sweet,’ with Nom.
Sg. Mase. *swadus (O. Ind. svadis, Gk. j3¢s), Nom. Sg. Fem.
*swadwi (O. Ind. svadvi). In Latin these Adjectives, aided by
the analogy of the Fem., have followed the tendency of O-stem
Adjectives, and have passed into the I-declension (as in Teutonic

*tnu- ‘thin’ became punni- by influence of the fem. *t°nwi); thus
suavis for *suddvis, ténwis (O. Ind. tané-, Gk. ravi-yAwooos,
§§ 45-49.) NOUN AND ADJECTIVE STEMS. -U-. 343

O. Sl. tinti-kti), /évis (cf. O. Ind. raght-, Gk. é-Aayvs, O. SI.
ligti-kti), brévis (Gk. Bpayvs), grivis (O. Ind. guré-, Gk. Bapis,
Goth. kaurus). Of nouns with the U-suffix examples are I.-Eur.
*petku- (O. Ind. pasa- M., Goth. faihu N., ‘property, O. H. G. fihu,
O. Engl. feoh, Germ. Vieh, Engl. fee), Lat. pécw and pecus ;
Lat. dicus (O. Ir. loch, a Neuter U-stem) ; with the rarer NU-
suffix, Lat. cornu and cornus, mdénus, pinus, The -ru- of Gk.
ddxpv is augmented by the suffix -ma- in Lat. dacriima, lacriima.
The Romance languages show us that U-stems had come in
Vulg. Lat. to be merged in the O-declension ; and even as early
as Plautus Noun-stems in -u retained little of a distinctive
declension in ordinary speech.

-TU-. As we have seen -ti- used to form feminine Verbal
Nouns (Nomina Actionis), connected with the P. P. P. in -to-,
similarly -tu- was used for mase. nouns of the same kind, some-
times with accent on the root (strong grade), sometimes with
weak grade of root and the root unaccented, e. 2. O. Ind. étum but
itva, from the I.-Eur. root ei-,‘to go.’ The Latin first Supine
is the Accusative of a TU-stem, used with a verb of motion, e. ¢.
visum it (cf. O. Ind. véttum éti, O. Sl. vidétii ideti), and the
second Supine is another case (ch. viii. § 88) of a similar stem.
Occasionally these Abstract: Verbal Nouns become concrete, e. g.
Lat. mdgistratus, a magistrate. In Greek (they are mostly con-
fined to the Ionic dialect) they are feminine, e.g. dpris, a fitting,
a connexion (Lat. artus, M., a limb) (cf. O. Lat. metus F.).

§ 48. Other examples of U-stems in Latin. Lat. génu (cf. O. Ind. janu, N.
Gk. yévu N., Goth. kniu N.); véru (O. Iv. bir, a Neut. u-Stem, W. ber).

§ 49. Interchange of U- with O-stems. This interchange is found also
in Umbro-Ose. Thus Ose. senateis shows the same Gen. Sg. suffix as an
O-stem (e.g. sakarakleis ‘sacraculi’), Umbr. maronato ‘magistratu,’ the O-stem
Abl. In Plautus the U-stems appear mostly in the Nom. Acc. Abl. Sg., and
so are hardly to be distinguished from O-stems ; the Dat. Sg. is usually in -ui
(ch. vi. § 27), but the Gen. Sg. is normally in -¢ (ch. vi. § ar) ; in the Plural
instances are comparatively seldom and almost only in the Nom. Acc. and
Abl.; the only certain instance of a Gen. Pl. is mille passum. On the 8. C.
Bacch. of 186 B.c. we have Gen. Sg. senatuos, but at the end of the same
century senati (C.I. L. i. 199, 200, 547), and Quintilian (i. 6. 27) states that it
is impossible to say whether séndti or sindtus is the Genitive form. Nura for
mirus, socra for socrus, forms censured in the Probi Appendix (198. 34 and
199. 1K.) occur on late inscriptions (nura, 0. I. L. viii. 2604. 4293, &e. ; nurud,
V. 24523 socrd, ii. 530, 2936 ; fii. 655; viii. 2906, 3994 ; xiv. 526, &e. ; socera, iii.
3895, Vii. 229, &c.) and néra, sdc#ra or socra are the prototypes of Ital. nuora,
344 THE LATIN LANGUAGE. [Chap. V.

suocera, Span. nuera, suegra, &e. On Greek inscriptions the only fourth
Deel. Nouns found are idus, tribus, and the forms used are Nom. Pl. eido: (i80),
Gen. Pl. eiSvey (second cent. B.c., but later cidwv) Abl. Sg. rpBou (see
Eckinger, p. 134).

Dimus is declined only according to the second Decl. in Plautus and till
Sulla (see Langen, Anal. Plaut. ii. p.5). Quintilian (i.6. 5) mentions itas a word
about whose declension there might be doubt. We find in the older writers
humu (Varro, Men. 422, 531 B.; the gender of hitmus is Mase. in O. Lat., see
Georges, Lex. Wortf.s. v.), lectus (fourth Decl., Georges s. v.), sénus (fourth Decl.,
Georges s. v.), and in a quotation from a Law of Numa (ap. Fest. 212. 17 M.)
occurs the form Janui Dat. Sg. : Janui Quirino agnum marem caedito. Priscian,
discussing the interchange of second and fourth Decl. forms (i. p. 256 H.),
mentions the variants fastos and fastus in Hor. C. iii. 17. 4, cibus Gen. Sg.
(doubtful), arci Gen. Sg. (Cic. Deor, Nat. iii. 20. 51; arcus is Fem. in O. Lat.),
as well as the parallel forms sptcus and specum, pinus and penum, fico and
ficu, &c. For statistics of the O- and U-declension of such words see Neue, i?.
pp. 509 sqq. Names of trees like laurus, fagus, &c. show this uncertainty
(Varro L, L, ix, 80 alii dicunt cupressus, alii cupressi, item de ficis platanis
et plerisque arboribus), an uncertainty perhaps due to the rarity of fem. O-
stems (e. g. fagus, Gk. @yyés F.) in I.-Eur. (see ch. vi. § 1. p. 369).

§ 50. Other examples of -tu-stems. Lat. ésus from edo (O. Ind. attu-,
Lith, ésta, O. Sl. jastt) ; Lat. vitus, the felly of a wheel (Gk. irus F.); Lat.
actus, (1) a driving, (2) a road for driving (Paul. Fest. 13. 17 Th. iter inter
vicinos quattuor pedum latum), (3) the space over which something is driven,
a measure of land, like our ‘plough-gate,’ (Plin. xviii. 9 in quo boves
agerentur cum aratro uno impetu justo) ; Lat. fétus, a brood (cf. Gk. rpopy,
e.g. Soph. 0. T.1) ; O. Lat. metus F., e.g. Enn. A. 526 M. nec metus ulla tenet.

§ 51. The Suffixes -YE- (Nouns of fifth Decl.) and -I-. The
Stems in -E. As the feminines of O-stems were formed with
the suffix -a- (¢ 2), so the feminines of other stems appear with
the suffix -yé-, e.g. Lat. tempériés from the HS-stem tempus, or
-i-, e.g. O. Ind. datri, fem. of datér- (in Latin with ¢ added, e.g.
ditri-a, stem ditric-; cf. Gk. atAnrptié- with added dental and
short i), or -ya-, e.g. Lat. temperia (cf. Gk. dAjOeva for *ddnbec-
ya, ddrecpa for * dSorep-ya). How far these differences are due to
a variation of -yé- with -1-, as in the Optative, e. @. Lat. siés, sitis
(ch. viii. § 55), or to a confusion of I-stems (thus O. Ind. napti-,
Lat. zeptis, may be an I-stem corresponding to the U-stem, O. Ind.
évakra-, Lat. socrus, O. Sl. svekry), has not yet been satisfactorily
determined. It is possible that the é-vowel of Latin and the
Balto-Slavic languages (e.g. Lith. Zem-é- for *%em-ié-, ‘land,
O. SI. zém(1)-ja- for *zem(l)jé-, whence the name Nova Zembla,
connected with Lat. Atimus and Gk. yédév) may be a modification
§§ 50, 51.] NOUN AND ADJECTIVE STEMS. -YE-, -L-. 345

of an original 4 under the influence of the preceding y-sound (cf,
Vulg. Lat. Jenuarius for Jdnuarius, ch. ii. § 3), so that O. Lat.
hervem Acc., a word occurring in the liturgical formula Zeriem
Junonis (Gell. xiii. 23. 2), will exactly correspond to Osc. heriam
[on an execration-tablet, Zvet. J. I. I. 129. 1; the i (not ui) of
the Oscan word probably indicates a y-sound (ch. iv. § 6 3)|. Verbal
Nouns with -yé- are a feature of Latin, e.g. pro-génies, ribies,
scabies, pernicies, and permities [in facies, species the i (y) appears
also in the Pres. stem].

Other examples of these suffixes are: of U-stems, O. Ind. svadvi,
Gk. id¢eia for *oFadeFyd (cf. Lat. sudvis M. and F., § 47), fem. of
I.-Eur, *swidu-; of N-stems, O. Ind. yani, fem. of yivan-,
‘ young’ (the Plautine iwuenix, Mil. 304 points to Lat. janiw being
rather a contraction of a stem *ytiweni- than a direct develop-
ment of the stem yiimi-), O. Ind. rajfi (mod. Ranee), fem. of
rajan-,‘a king’; of NT-stems, O. Ind. bhéranti, fem. of the Pres.
Part. bhérant- (Gk. ¢épovea for *pepovryd ; cf. Lat. praesentia,
&e.).

Other fifth Decl. stems are suffixless, e.g. 7é-s (O. Ind. ra-s, ray-ds
Gen. from the root réi- (ch. iv. § 47)!; diés from a stem d(i)yéw-
(O. Ind. dyaés Nom. Sg.), a byform of d(i)yéw-, ‘ sky, day’ (ch. iv.
§ 48). Some seem to have the same Verbal E-suffix asis seen in
verbs like edilé-facio, candé-facio [ch. viii. § 33 (9)], e. g. flim-2-s
(cf. Gk. é-yxdv-n-v from xaivw for *xayyw), which is really a fifth
Decl. noun like /7d-é-s (cf. Gk. é-i9-n-v), Another noun declined
according to the fifth Decl. is sordés, Abl. sordé, Lucr. vi. 1271,
Gen. Pl. sordérum, Plaut. Poen. 314 (cf. sordeo). So from *dleo,
to grow, préles for *pro-dles, sub-dles, ind-dles (cf. ol(e)-facio from
leo, to smell); from @uo ues, from struo strues from ruo rues
[‘rues’ ruina, C. G. LZ. iv. 281. 5; ef. lue rue, ‘luem ruem’ (?)
on the Carmen Arvale]. Against the association of /aes with
libe-facio is the different quantity of the root-vowel. A long
vowel is seen also in méles (mdles-tus), sédes (sédeo), amb-ayes (ago),

 

' snés may be of similar formation
(cf. O. Ind. sphayaté, ‘he extends him-
self, increases’); it is used by Plautus
only in Nom. Acc. Abl. Sg. and in
Nom. Acc. Pl. (spes). But Ennius

treats it asan S-stem in Ann. 448 M.:
spero, si speres quicquam prodesse
potis sunt (cf. 119 M.). With the
variants spés and spérés, compare vis
(Nom. Pl.) and vires,
346 THE LATIN LANGUAGE. [Chap. V.

contiges (tiéngo), propages and compages (piingo), ripes (rtimpo).
The cognates mdles-tus, Gk. eos, vépos have suggested the refer-
ence of mdlés, sédés, niles (though the @ is difficult to explain) to
ES-stems (§ 71), with the same -és that we see in Gk. ddyOns
from djj0os, Dor. A&os (cf. EA%Hov), But perhaps a better ex-
plation of sédés is to regard it as a Plural of an I-stem sédis, used
as a Singular (cf. O. Ind. vayas, ‘ a bird, properly ‘ birds,’ and for
the long vowel Gk. pis, &e.); for a good many of these nouns
in -és were more used in the Plural than the Singular, e. g. am-
Lagés (O. Ind, aji-, * contest,’ Ir. ag, ‘ contest’), aedes, vepres, and
byforms of the Nom. Sg. occur with -2s, e. g. nubis, Plaut. Mere.
880 (cf. nubs, Liv. Andr.), saeps, Cic. (cf. prae-saepe Neut.),
vatis (cf. O. Iv. faith for *wati-), véhis, cladis, aedis (C. I. L.i.
206. 30), caedis, molis, sedis. Many of them belong to poetical
diction ; e.g. neither Plautus nor Terence use sedes (the usual
word for a seat being sddium or sella), rupes, tabes, caedes, &e. ;
and they must have been liable on that account to confusion of
Number. <A certain amount again of Nouns in -és are fem."
names of animals, e. &. cdinés (F., beside céints M., in O. Lat.), félés,
mélés, vulpes (cf. vudpécula), pidumbés. The last two, perhaps all,
are dialectal, so that this -és may be the dialectal expression of
-eis, a byform of the Nom. Sg. suffix of I-stems. P/éés has
been called an ES-stem like wAj0os, but it is more persistently
declined after the fifth Decl. (cf. plebeius), than any of the
others, whose Ab]. is often the only Singular case found in use
(e. g. ambage, vepre, Ovid, &e.); pubes (cf. piber-tas, im-piibér-es
Nom. Pl.) has a better claim, although the Abl. pudé, Plaut.
Pseud. 126, would have to be explained as due to a false analogy
of the Nom. Sg., such as has produced reguié Abl., requiem Acc.
from requiés, -étis.

The declension of a good many of these Nouns with Nom. Seg.
in -és wavers between the third and fifth Decl. Plautus makes
the Gen. Pl. of sordés sordérum (Poen. 314), like rérum, but we
find sordium in late Lat. Cicero (Top. vii. 30) says that specierum,
speciebus are impossible forms (nolim enim, ne si Latine quidem

 

1 Priscian (i. pp. 168. 15. 169, 9, and 321. 19 H.) seems to think that
adip- F. implies a Nom. Sg. adipes.
§§ 52, 53.] NOUN AND ADJECTIVE STEMS. -VE-, -I-. 347

dici possit, * specierum’ et ‘ speciebus’ dicere), and so prefers
Jorma to species as a translation of the Greek efSos. Yet specie~
rum and speciebus are common in late authors, e.g. Apuleius, and
Priscian (i. 367. 23 H.) quotes facierwm from a speech of Cato.
Quintilian (i. 6. 26) expresses the doubt felt about the declension
of progenies and spes: quid progenies genetivo singulari, quid
plurali spes faciet? (For statistics of fifth Decl. forms, see
Neue, i”. pp. 370 sqq.) And the use of -iés itself in the Nom. Sing.
seems to have been a matter of gradual extension. Plautus has
Sicies, permities, matéries (but materiam on the Sententia Minu-
ciorum of 117 B.c.,C.. L. i. 199), segnities, vastities, and possibly
caestiries, inlitvies, intempéries, ribies (but effigia, &e.), to which
Terence adds mollities, Lucretius ndtities, spurcities, &e. These
yé-forms are generally confined to the Nom. Acc. Sg., while in
other cases ya-forms are preferred, e.g. intemperiae Nom. Pl.
materiae Dat, Sg., lumiiriae Dat. Sg., mollitiis Abl. Pl., in Abl.
Sg. barbirid, Poen. 598, mollitia, Vidul. 35. Still Plautus
adheres to the fifth Decl. type for facies (Nom. Gen. Acc. Abl.
Sg.), and apparently dcies (Acc. Abl. Sg.), permities (Nom. Acc.
Abl. Sg.), species (Acc. Abl. Sg.), though he does not use these
words in the Plural. Pliny (ap. Charis. p. 118.15 K.) allowed plani-
ties, Luxuries, mollities, but not dmicities (amicitiem, Lucr. v. 1019),
‘because the Plural is amicitiae,’ while Charisius (p. 57. 3 K.)
confines cdnities to poetry, and makes canitia the proper prose-
form. (For fuller statistics, see Neue.) The -¢ié-stems are
found also in Umbr., e. g. uhtretie ‘ auctoritate,’ kvestretie, ‘ in
the quaestorship.’

§ 52. Other examples of Latin Fems. in -1, -ic, &e. Latin cornix (cf. Gk.
Kopwrn) ; génétria \O. Lat. janitri, Gk. yevére:pa); from méreor comes mérétrix, but
from médeor the name of the deity Méditrina, whose festival, the Meditrinalia,
is mentioned by Varro (L. L. vi. 21) and Paul. Fest. (88. 36 Th.) in connexion
with the curious Roman custom of hallowing the first taste of new wine
with the words: novum vetus vinum bibo; novo veteri morbo medeor;
ctinicula (cf. O. Ind. Sunt) ; clavis, -im Acc., -% Abl. (ef. Gk. «Anis, stem «Ant-8-).

§ 53. -yé- and -i-. Acisculus, a small pickaxe (there was a Roman family-
name Valerius Acisculus), has been referred (with supposed 7) to dcies (Rhein.
Mus. 1891, p. 236); the Probi Appendix (p. 198. 12 K.) censures a large
number of forms in -is beside forms in -es, e. g. vatis beside vates, and among
them facis for facies (cf. facitergium Isid. beside facietergium Greg. Tur.).
348 THE LATIN LANGUAGE. [Chap. V.

§ 54, Suffixes ending in -n (Nouns of third Decl.).
-EN-, -YEN-, -WEN-, -MEN-. Beside I.-Eur. Mase.
stems in -o- (-e-), -yo- (-ye-), -wo- (-we-) (§$ 2, 4, 5), we
find masc. stems in -en- (-on-), -yen- (-yon-), -wen- (-won-).
The relation between the two seems to be that the O-stem is
the Adjective, but becomes an N-stem when a Substantive,
e.g. Lat. multi-libus Adj., but b7/6 Subst. An adjective
which is restricted to denote one individual, in other words
which becomes definite from indefinite, takes this suffix, e.g.
Lat. réfus, red, but Rafé, ‘the Red’ (cf. Gk. orpaBds and Erpd-
Bwv); and this process, which has been more consistently carried
out in the Teutonic languages than in any other, is still seen in
the German ‘ weak declension’ of the Adjective, e.g. rothe Nom.
Pl. indef. (strong decl.), die rothen Nom. Pl. def. (weak decl.)
(Goth. raudai and pai raudans). This early connexion between
O- and N-stems explains why an N-stem in a Compound is
often replaced by an O-stem, e.g. Gk. dxyd-Oeroy, stithy, com-
pounded of dkpov and 7/61, and why an N-stem in one lan-
guage often corresponds to an O-stem in another, e.g. Gk.
ai(F)sév to Lat. aevom (also explained as *aiwén), Goth. ga-juk-a
(N-stem) to Lat. con-jtig-us (O-stem), or in the same language,
e. 2. Lat. dénio to Lat. linius, inctibo to imcitbus. In the declen-
sion of these Masc. N-stems, n varies with en, on, én, 6n, &c.,
yen also with in, and wen also with tn. They are sometimes
transferred by the addition of -o- or -a- into the O- and A-
declension, e. g. Lat. gpitlinus beside épiild, terminus beside termd
($ 13). Neuters in -n are connected with neuters in -r, &c., the
n apparently being proper to the Oblique Cases, e. g. Lat. femur,
Gen. feminis (§ 56). They are therefore better considered in the
paragraph which deals with the R-suffix § (56).

Neuters in -men are Verbal Nouns (Nomina Actionis), and
are used in various languages as Infinitives (like -sen-, e.g.
O. Ind. né$4ni from ni-), e.g. O. Ind. vid-mén-é, Hom. Gk. t-
pev-at (perhaps Lat, 2 Pl. Imper. Pass., e. ¢. /2yimini, is a similar
Infinitive form with Imperative sense ; see ch. viii. § 81). They
have usually the E-grade of root, e.g. teg-men, but the Latin law
of Syncope seldom allows us to decide when there was a connect-
ing vowel, when an s preceded, &c. (cf. regimen, integumen-tum,
§§ 54-56.] NOUN AND ADJECTIVE STEMS. -N-, -R-. 349
*

jugmen-tum and jugumen-tum, augmen-tum, siimen from. siigo, and
see ch. iv, § 116 on agmen and exdmen). Their byforms with
-mentum in Latin, e. g. ddtmentum, have been mentioned in § 30,
and the extension of MEN-stems by the YO-, YA- suffixes, e. g.
alimonia, alimonium, in § 4.

Fem. Verbal Abstracts in -yen-, -tyen- (-tien-) in Lat., Celt.,
Teut., e.g. Lat. dsi-céipio, captio have become fem. by the
analogy of other abstract nouns. They sometimes appear as
A-stems with the ending -ina, -tina, e. g. Potina, the goddess
worshipped when a child first drank (Non. 108. 15 M.) from pétio,
Statina, ‘statuendi infantis,’ offtcina (dpi-ficina Plaut.), répina,
ruina (cf. § 8).

§ 55. Mase, en-stems in Latin. Lat. lién, Gen. lignis shows -én, and pectin,
Gen. pectinis (Gk. xreis for *rerevs ?) shows -tn, but the usual form is - in: the
Nom. Sg., -on- or -in- (with % for older 6 or 2) in the oblique cases, e.g. edd,
Gen. edonis from the verbal root ed-, ‘ to eat,’ himd, Gen. hominis from the noun-
stem humo-, the ground. The Fem. cdro, Gen. carnis, is exceptional in reducing
the stem-suffix in the oblique cases to n. In Umbr.-Osc. the word has the
same declension and gender, but retains the older sense of ‘a part,’ a ‘share,’
e.g. Osc. maimas carneis senateis tanginud, which would be in Latin, ‘ maximae
partis senatis sententia’; and originally the word seems to have been
a Verbal Abstract (Nomen Actionis), ‘the act of cutting or dividing,’ whence
the feminine gender. (On the variation of gender in cardo, margo, ctipido, grando,
tiligo, farrdgo, see Neue, i*%. p. 654; Abstracts in -go, like drigo, are fem., and
Nouns indicating defects or ailments like ferrigo, weriigo, caligo, priirigo).

Nouns in -d, -dnis Gen., used of persons, belong as a rule to plebeian or
colloquial Latin and express contempt, e.g. dileo, bibo, Capito, Naso ; in the earliest
period they are derived only from Adjectives, e.g. strdbo, but later from
Nouns and Verbs, e.g. commitlito from miles. lénio from ldnius, erro from errare,
stitiirio from sdtirire. The more respectful formation was in -énus, v. g. patronus
(but tarpwr always on Greek inscriptions, as early as 150 B. c., Eckinger, p. 135);
Paul. Fest. quotes O. Lat. epolonus for gpiilo (55. 15 Th.), centiirionus, ciirionus and
déctirionus (34. 36 Th.) ; so Dor. Gk. Adr& became Latona. (On these nicknames,
&c. in -d, -dnis, see Fisch, Lat. Nomina Pers. auf -0, -onis, 1890.)

§ 56. Suffixes ending in -r (Nouns of third Decl.).
-R-. J.-Eur. Neuters in -r (-r) seem to have substituted n for r
in the oblique cases, e.g. Lat. femur, Gen. féminis, O. Ind. udhar,
‘ the udder,’ idhnas Gen., which has led to confusions of R- and
N-stems, e.g. Lat. jécur, but O. Ind. yakén-, and to such curious
declensions in Latin as it-%n-ér-is, jéc-¥n-dr-is, and joc-in-er-is, &e.

§ 57. Neuter R-stems. Many of these are names for parts of the body,
e. g. femur, jécur, aber (Amer. Journ. Phil. xii. 1). On O. Lat. aser, blood (O, Ind.
Asr-k Nom., asn-ds Gen.) and on other Neut. R.-stems, see ch. vi. § 15.
350 THE LATIN LANGUAGE. (Chap. V.

§ 58. -ER- and -TER-. As the Comparative suffix is some-
times -ero-, sometimes -tero- (§ 16), so we find -er and -ter as
the suffix for forming words of relationship, and masculine No-
mina Agentis [with fem. in -(t)zi, -(t)ria, &e., § 51]. This -(t)er-
varies with -(t)or-, -(t)ér-, -(t)dr-, -(t)r-, &., e.g. Gk. aarip,
Acc. rarépa, and @parwp, Acc. pparopa, Gk. dorip beside dérwp,
Borip beside Bdrwp, &e. The Nomina Agentis in Latin show
-tir-, e.g. actor, Gen. actéris (Gk. dxtwp, -opos and én-axtyp,
-fjpos). Abstracts in -or, -dris Gen., often connected with Verbs
in -eo and Adjectives in -Zdus (e.g. céilor, beside caleo, callus),
are not R-stems but S-stems, and had in O. Lat. -ds in the Nom.
Se., e.g. calos, though in the classical period the 7, into which
intervocalic s in the oblique cases had passed by the phonetic law
of Latin (ch. iv. § 148), forced its way into the Nom. Sing. also
(see ch. vi. § 7).

§ 59. Nouns of relationship. Lat. pdter [O. Ind. pitdr-, Arm. hair, Gk.
natnp, O. Ir. athir, Goth. (rare) fadar] ; Lat. mater [O. Ind. matdr-, Arm. mair,
Gk. porn, O. Ir. mathir, O. Engl. médor (in Goth. aipei, ‘mother,’ atta, ‘father’)
O.SI. mater-]; Lat. frater (O. Ind. bhratar-, Arm. eAbair, Gk. ppatwp and ppaérnp,
a elansman, O. Ir. brathir, ‘a brother,’ W. brawd(r), Goth. bropar ; ef. O. SI.
bratrti) ; Lat. sdror (0. Ind. svasar-, Arm. k‘oir, O. Ir. siur and fiur, Lith.
sesi; cf. Goth. svistar, O. Sl. svestra) ; Lat. lzvir better *laevir, for *laever by
Anal. of vir (I.-Eur. *daiwer-, O. Ind. dévdr-, Arm. taigr with g for w,

Gk. danp for *SaFnp, O. Engl. tacor, O. H. G. zeihhur; ef. Lith. déver-is,
0. Sl. déver-i).

§ 60. Latin Nomina Agentis. Lat. praetor for *prac-itor (O. Ind. pura-étar-,
“he who goes before,’ ‘a guide’); junctor (O. Ind. yoktdr-, Gk. Gevernp) (cf.
junctus) ; con-ditor for *con-ddtor from the root dhé-, ‘to put, place’ (O. Ind.
dhatér- and dhatar-, Gk. @erfp) (ef. conditus) ; potor (O. Ind. patér- and patar-,
Gk. worjp) (cf. potus) ; génitor (O. Ind. janitar-, Gk. yevergp and -yevérwp) (cf.
génitus) with Fem. génétrix (O. Ind. janitri, Gk. yevérerpa) (on @ see ch. iii. § 19);
teator (O. Ind. ta%tar-, ‘ carpenter’) (cf. textus) ; ésor (O. Ind. attar-, Gk. dnornp)
(cf. @sus).

§ 61. Suffixes ending in -t (Nouns and Adjectives of
third Decl.). -T-. Latin stems in -¢ are probably of various
origins. A compound like com-es (stem *cdm-tt-) from cum and
ire, which has the sense of a Participle Active,‘ going along with,
reminds us of the use of this suffix in O. Ind. with verbal roots
ending in vowels, &c., when these roots form the second part of

4 . . oe
a compound, e. g. O, Ind. visva-jit-, ‘all-conquering,’ from ji-, ‘ to
conquer.’ To this category belong Latin ¢-stems like anti-stes
§§ 58-62. ] NOUN AND ADJECTIVE STEMS. -T-. 351

from the root sta-,‘ to stand,’ sticer-dds from the root dé-,‘ to give.’
On the other hand adjectives of passive meaning like man-sués,
Gen. mansuétis, ‘ accustomed to the hand, tame (beside mansuétus),
in-gens, Gen. ingentis, ‘not known, huge (cf. Engl. uncouth),
may be for *mansuetis, *ingentis with that transference of the
Adjective to an I-stem which we have seen in O. Lat. Sanates,
beside sanatz, forctis (class. fortis) beside forctus (§ 45). It has
further been suggested (§ 29) that nouns like égues, Gen. equitis,
may be connected with Greek formations like taadérys, the fem.
Abstract *eguita, horsemanship, becoming masc. eqgues, a horse-
man, just as antistita fem. is related to antistes masc.

Again Feminine ¢-stems like téyes, Gen. tegétis, a mat, lit.
‘a covering, séges, Gen. segétis, a crop, lit. ‘a sowing’ (W. hau
for *sog-, ‘to sow’), merges, a sheaf, lit. ‘a dipping’ (if merga,
a pitchfork, be rightly explained by Paul. Fest. 89.13 Th.: quia
. .. messores eas in fruges demergunt, ut elevare possint manipu-
los), or ‘a plucking’ (ef. Gk. d-yépyw), may have been originally
Fem. TI-stems (Nomina Actionis), like messes for *met-tis,
‘a reaping,’ harvest (§ 42), and may show that confusion of the.
JTI- and the T-suffix which is shown by words formed with
-tat(i)-, -tat(i)-, (§ 46), and bystems like Gk. vu«r-, Lat. noeti-.
But Greek nouns like «éAns, -nros, a horse, lit. ‘runner,’ from
the root kel-, ‘to run, go quickly ’ (O. Ind. car- ; ef. Lat. celer),
and adjectives like dpyfr- and dpyér-, bright, from the root arg-,
‘to shine’ (O. Ind. arj-; cf. Lat. argentwm), rather point to by-
forms in -t, beside the usual Pres. Part. stems in -nt, just as
Greek compounds like dyvss, -@ros, (1) unknowing, (2) unknown,
beside ayvwros and d&yvworos, mpoBArs, -fjtos, a headland, beside
mpéBanros, thrown forth, suggest that Latin mansués, &c. may
be quite separate forms from mansuétus, &c.,and not sprung from
*mansuetis, still less produced by Syncope of the last syllable of
mansuetus (cf. ch, iti. § 16).

§ 62. Other examples of Lat. T-stems. Lat. téres, Gen. terttis, from tero ;
hébes, Gen. hebitis ; ales, Gen. alitis, with the sense of aldtus ; tudes (cf. ttidttare),
an old word for a hammer (Fest. 530. 30 Th. ; Paul. Fest. 531. 12. Th.), from
tundo, is a rather doubtful form ; and the origin of 1.-Eur. *népot- (O. Ind.
napat-, M. Ir. niae, Gen. niath, W. nai; cf. Hom. Gk. vérodes), Lat. népis, with
its cognate *népti (O. Ind. napti-, Goth. ni(f)pjis, ‘kinsman,’ 0. Sl. ne(p)tiji-,
‘nephew,’ &e.), Lat. neptis is not clear. The termination -es was much in vogue
352 THE LATIN LANGUAGE. [Chap. V.

in the earlier period, e. g. caeles, -itis, circes, -ttis, From it were formed Abstract
Nouns in -ttia, -ities, &c. (see §§ 4, 51) (cf. axittidsi, Paul. Fest. 2. 34 Th. : a. factiosi
dicebantur, cum plures una quid agerent facerentque, from avites id. 3. 1 Th.
a. mulieres sive viri dicebantur una agentes. The adj. axitidsus is quoted by
Varro, L, L. vii. 66 from the Asiraba of Plautus :

axitiosae annénam caram e uili concinndnt uiris).

§ 63. -NT-. The suffix -ent-, -ont-, -nt-, &c. plays an im-
portant part in the I.-Eur. languages, being used in the form-
ation of all Active Participles, except the Perfect (ch. vill. § 89).
In the thematic conjugation we have -ont- in all the cases in
Greek (dépovt-os, pépover for *pepovror, &c.) and other languages ;
and O, Lat. forms like fleawntes (?),as well as the classical ewntes
where ¢ precedes, have been quoted as proof that Lat. ferent-is,
_Serent-em, &c. represent an older *feront-, with change of the
short vowel before double consonant in the syllable unaccented
under the early Accent-law (ch. iii. § 5) into 2 (but see ch. viii.
§ 90). These Participles have often become adjectives and nouns,
-e.g. Engl friend (Goth. frijonds, lit, ‘loving ’), fiend (Goth. fijands,
lit. ‘ hating,’ Germ. Feind), Gk. dpywv, Lat. radens, sometimes
with transference to the O-declension, e.g. Lat. véntus (Goth.
vinds) beside the Participle, O. Ind, vant-, Gk. deés for *aFevrs.

§ 64. Other examples of Lat.-ent. Lat. dgens (O. Ind. djant-, Gk. dywv) ;
Sérens (O. Ind. bhaérant-, Gk. pépwyv, Goth. bairands, O. Sl. bery); rudens
(O. Ind. rudant-), (1) ‘roaring,’ (2) ‘a rope’; vékens (O. Ind. véhant-, Goth. ga-
vigands, Lith. vez@s, O. Sl. vezy) ; bénévélens is often a Noun (cf. bénémérens) in
the Comedians, and is closely connected with the Adjective benevolus, so that:
benerolentior and benevolenter have taken the place of the Comparative and
Adverb of the latter, as magnificentior, magnificentissimus supply a Comparative
and Superlative to magnificus. Other Nouns are pdrens (beside the verb piri),
serpens, ddiilescens; other Adjectives, ddéquens, sdpiens, innicens. The addition
of the ya-suffix, which forms the fem. of these participles in Greek (e.g.
pépovoa for *pepovrya) forms Abstract derivatives in Latin, e.g. benivolentia,
praesentia, eloquentia (all of these first used by Terence), confIdentia, midlivolentia,
pitientia, scpientia, pollentia, &e. (all used by Plautus), sententia (beside the verb
sentio). On the rare ending -entium, e.g. silentium, see Rénsch, Collect. p. 208.
Fluentum and cruentus probably originated in Neut. Pl. fluenta, cruenta (cf.
silenta loca Laev.), a formation like Gk. ¢épovra, &e. The ending -lentus of
optilentus (beside opulens), vidlentus (and violens), triiciilentus, esciilentus, virtilentus, &e.
comes from an Adj.-stem in -lo (-li § 21); ef. grdcilentus (the o of vindlentus,
somnilentus has been explained by ‘popular etymology’ from vind lentus, somnd
lentus ; see ch. iii. § 21), ,

§ 65. -WENT-. This suffix, like our ‘ -ful,’ added to Noun-
§§ 63-67.] NOUN AND ADJECTIVE STEMS. -WENT-, -D-. 353

stems to form Adjectives, with the sense ‘ possessed of,’ ‘abounding
in, and occasionally ‘resembling, e.g. O. Ind. agni-vant-,
‘provided with fire,’ xapies for *xapiFevts, possessed of grace,
graceful, crovdes (Corcyr. orovofecav, Acc. Sg. Fem.), woeful,
is in Latin augmented by the TO-suffix, e.g. délosus (Gk.
dodrdeis) for *dolo-venssus from *dolo-went-to- or *dolo-wnt-to-,
nivosus (Gk. wodes). In O. Ind. it was often added to the
P.P.P. TO-stem, e.g. krta-vant- from krté-, P.P.P. of kr-,
‘to do,’ and came to be used in the sense of a Perfect Indicative
(with omission of the Substantive Verb), e.g. sa tad krtavan,
“he has done this,’

§ 66. Other examples of Lat. -dsus. Lat. virosus (ef. O. Ind. vi8d-vant-) ;
vnosus (ef. Gk. oivdes), (1) full of wine, (2) like wine; cédéddvérosus, like a dead
body. From U-stems, -wosus, e.g. aestuosus (but fastdsus, and from the stem
mont-, montudsus), but from I-stems -osus, e.g. piscosus (Virgil’s rendering of
Homer’s ix@vées), (but biliosus). Incitriosus (from inciria) produced ciiriosus
(instead of *ciirdsus) ; céiléimitdsus stands for *calamitdtosus (ch. iii. § 13, p. 176).
On the change of *-ovenssus to -dsus (cf. retrorsum, retrosum from retroversum, &c.),
see ch. ii. § 53. The older spelling is -onssus, -ossus (see Brambach, Orth. p. 268,
and the Indices to Ribbeck’s Virgil and to Studemund’s Apograph of the
Ambrosian Palimpsest of Plautus). Gellius (ix. 12) comments on the
Active and Passive sense of these Adjectives: ut ‘formidulosus’ dici potest et
qui formidat et qui formidatur, ut ‘invidiosus’ et quiinvidet et cui invidetur,
ut ‘suspiciosus’ et®qui suspicatur et qui suspectus est, ut ‘ambitiosus ’ et qui
ambit et qui ambitur, ut item ‘gratiosus’ et qui adhibet gratias et qui
admittit, ut ‘laboriosus’ et qui laborat et qui labori est, &e.

§ 67. Suffixes ending in -d (Nouns of third Declension).
D-suffixes are not frequent enough in the I.-Eur. languages to
enable us to determine the formation of Latin d-stems like
lipis (cf. Gk. Aéras ?), mereés, hérés, nor to trace the passage of
original D-stems into other stems. The -ad- of Gk. xouipds,
a reef, ‘like a hog’s back,’ weAeids, a wild dove, has been
declared to be the first part of the Latin -astro- (for *-ad-tro-) of
poétaster, ‘like a poet,’ Sleaster, ‘a wild olive,’ &c. Lat. -édila
occurs in names of birds, &c., e.g. ftcédula (and jicélla), a beccafico,
acrédula, querquédula, nitédula and nitélla (cf. alcédo); Lat. -édo
in names of ailments like frigédo, ruibédo, grivédo (and gravido)
resembles the -nddv of Gk. dyOnddv, xarpyddv. Lat. -dus of
Adjectives beside Verbs in -eo and Abstract Nouns in -or, e.g.
palledus (beside palleo and pallor), splendidus (beside splendeo and

Aa
354 THE LATIN LANGUAGE. [Chap. V.

splendor), squalidus (O. Lat. squalus) (beside squdleo and squalor),
has been referred to the root dé-, ‘ to give’ (ef. O. Ind. jala-das,
lit. ‘giving moisture,’ like Lat. zméri-dus, artha-das, ‘ giving
benefit,’ but see ch. iti. § 15. (8)) ; and the termination of Gerun-
dives like daudandus, &c. has been similarly explained (*laudam-
dus, ‘ praise-giving ’) with the Acc. Sg. of a Verbal Noun as the
first part (cf. vin-dex) of the compound (see ch. viii. § 94), as well
as Verbal Adjectives in -bundus, e.g. errabundus, pidibundus, firt-
bundus, gémébundus, méribundus, lasctvibundus (-ib- probably) and
-cundus, e.g. jucundus, ribicundus, vérécundus, fecundus, wracundus,
facundus. (On these also, see ch. viii. § 94.) The termination
-tado, Gen.-tédinis, has been called the amplification of the stem
-tiit- by an N-stem, *tiit-n- producing *tiid-n- (but see ch. iv.
§ 161), but it may also be derived from a tu-stem, as ribédo
from an é-stem.

§ 68. Other examples. Lat. péicus, -tidis beside pecus, -dris ; Lat. cassis, an
Etruscan word, according to Isidore (Orig. xviii. 14. 1), with transference to
the A-declension in the byform cassida.

§ 69. Suffixes ending in a Guttural (Nouns and Adjectives
of third Declension). These also are infrequent in I.-Hur.
languages, and often have, as we have seen (§ 31), byforms
with added -o-, e g. Gk. adéan€ (O. Ind. lopi€i-), Gk. petpag
(O. Ind. maryaké-), Gk. véaé (O. SI. novakii), Lat. sénex (O.
Ind. sanaké-, ef. Goth. sineigs; perhaps with the Diminutive
suffix -ko-, so that Lat. senea may be for *sénécis, ‘ oldish,
with that transference to the I-declension so frequent in Adjec-
tives ; ef. senz- from I.-Eur. *séno-, O. Ind. séna-, Gk. évos, O. Iv.
sen, Lith. sénas, &c.). Latin Adjectives in -da express tendency
or character, e.g. Libdw, dicda, ripax, pervicdx (O. Lat. pervicus) ;
of adjectives in -27 we have félia from *féla, Gk. Ondy, the
breast, perniw (properly of horses, &. d. L. L. viii. 453) from
perna ; -tria is the fem. ending (O. Ind -tri, Gk. -rpyd) of Mase.
Nomina Agentis in -tor, e.g. génétria (O. Ind. janitri, Gk.
yevéterpa, § 51); -de appears in the derivatives from Adjectives,
Jéerox (from férus), solloe (from sollus), dtror (from *étro-; cf.
ddium), and in vélom, céléx, which suggest connexion with dcior.
We have O-stem Adjectives with -dceus, e.g. mérdcus, and -icus,
e.g. mendicus, dmicus (§ 31), and A-stem Nouns with -ica, e.g.
§§ 68-71. ] NOUN AND ADJECTIVE STEMS, -ES-. 355

lectica, and -iica, e.g. lactiica (§ 32). The Latin termination
-igo, in names of ailments, e.g. vertigo, dépétigo, impétigo, caligo,
may be due to the addition of an N-suffix to Adjectives in -ir,
*vertic-n-, *calic-n-, &c., and the somewhat similar -igo of
ferrigo (ck. rabigo), aeriigo, laniigo, and in names of plants, e.g.
molligo (-igo), aspériigo may similarly represent -de-n-, the
becoming g before x by the phonetic law of Latin (ch. iv. § 119).

§ 70. Other examples. Lat. piiléz, -icis, a flea, but pulegium, fleabane ; Lat.
latéx, Pl. latices (but Gk. Adrayes; cf. Gk. éprvy- but O. Ind. vartaka-); Lat.
vertée from verto; Lat. vervéx, « wether ; Lat. ndtrix, -icis, a watersnake (0. Iv.
nathir, nathrach Gen.).

Words like aureax (a. auriga, Paul. Fest. 6. 27 Th.), aurifex, &e. do not show
the Guttural suffix, but are Compounds of ago, fucio (§ 78).

The number of words ending with -x that indicate parts of the body is
noticeable, v. g. calx, faux, coxendix, cervix (or rather cervices Plur., for the Sing.
was a poetical usage; cf. Varro, L. L. viii. 14 and Quint. viii. 3. 35), mdtrix,
pantex, pidex, also bodily marks, e. g. varix, vibix, famex, &e.; bird-names in -ix
are, ©. g. cornix, coturnix, spinturnix. Adjectives in -dx from second Conj. Verbs
are audax, ténax (cf. rétindculum), mordax, &c.; from I10-Verbs cépax, efficax,
perspicax, With fallax cf. falla, O. Lat. for failécia; with dicax ef. dicd-re beside
dicére ; with cdpasx, oc-ciipa-re.

§ 71. Suffixes ending in -s (Nouns and Adjectives of
third Declension). -ES-. The suffix -es-, varying with -os-,
-s-, &c. was used to form Neuter Abstract Nouns, with E-grade
of root and accent on the root, e.g. I.-Eur. *Sénos Nom, 8g.,
*Senes- in Oblique Cases, from the root gen- (O. Ind. janas,
Gk. yévos Nom., yéve(o)-os Gen.), Lat. génus Nom., genér-is for
*genes-es Gen. These Neuter-stems became adjectives by trans-
ferring the accent to the suffix and substituting -és for -ds in
Nom. Sg. Masc., -és in Nom. Sg. Neut., e.g. Gk. ed-yenijs
Masc., -vés Neut. Beside them we find occasionally Mase. or
Fem. Nouns with -és in Nom. Sg., a formation which came into
great favour in Latin, e.g. ¢énor Mase. for *¢ends, beside tenns
Neut. (Gk. révos,a string), from the root ten-,‘to stretch’; tépor
Mase. for *tepds (cf. O. Ind. tapas Neut, for *tépos), from the
root tep-, ‘to be warm.’ A case of these Neuters is used as the
Infinitive in various languages, e.g. Lat. vivére Loc., O. Ind.
jivasé Dat., ‘to live’ (see ch, viii, § 83).

§ 72. Neuter ES-stems in Latin. The original declension with -os Nom.,
-es- in Oblique Cases and Derivatives, is sometimes departed from, e.g. tempus,

Aa 2
356 THE LATIN LANGUAGE. [Chap. V.

-dris (but ef. temptri, tempes-tivus) ; dnus, dnusius (but ef. oneris, onerare) ; Velius
Longus (p. 73. 1 K.) calls attention to the discrepancy between faenoris and
Sfueneratorem, facinoris and facinerosus. We have the O-grade of the root, e. g. in
modes-tus (by analogy of the O-stem médus) from *modes-', from the root med-
(but Umbr. mers for *med(o)s Neut., the right, the due, is normal), foedus [also
Jidus, i. e. *feidos, to judge from Varro L. L. v. 86 per hos (Fetiales) etiamnunc
fit foedus quod ‘ fidus’ Ennius scribit dictum ; cf. Paul. Fest. 64. 3 Th. fidusta
a fide denominata, ea quae maximae fidei erant] from the root bheidh-, pondus
by analogy of pondo- (Abl. pondd) from pendo. Minerva was in earlier times
*Menes-ua (§ 7), a derivative of the Neut. stem *ménes- (O. Ind. manas-, Gk.
pévos), from the root men-. From other than E-roots we have, e.g. dpus
(O. Ind. dpas and dpas, ‘work,’ especially ‘a religious performance’). The
ending -nus is seen in facinus (cf. Gk. 5d-vos), &e. Tho -r of rébur (0. Lat.
robus, see Georges, Lex. Wortf. s. v.) may be due to the Mase. byform robor (ef.
O. Lat. robosem Acc., quoted by Paul. Fest. 11. 20 Th.); ef. cdélor Neut. (Plaut.
Mere. 860). The weak grade of the suffix, -s-, is seen in the derivative
O-stem O. Ind. vats-d- for *wetso-, ‘a calf,’ lit. ‘a yearling,’ from *wetos,
‘a year’ (Gk. Féros), &. Other examples are Vénus (originally Neuter and
meaning ‘glamour,’ like 0. Ind. vanas-; cf. vénenum tor *venes-num, properly
‘a philtre’) ; tergus (Gk. oréppos and répos).

§ 73. Adjective ES-stems. Of the transference of a Neut. S-stem into an
Adjective a good example is Lat. vétus, which is nothing but I.-Eur. *wétos
(Gk. Feros, a year) used in apposition as a predicate (cf. Hom. Gk. miap, Lat.
tiber, &e.). The ordinary way of making an Adjective from a Neuter S-stem
was by adding the suffix -to- (§ 27), e. g. vétustus, to which however, owing to
the cacophony of the two similar syllables, vetus was preferred in the Positive
and veterrimus in the Superlative, though vetustior was the Comparative in
vogue (cf. Varro, L. L. vi. 59 a vetere vetustius ac veterrimum), dniis-tus,
scélés-tus, fiinéstus, jiis-tus. Perhaps another way was to add the suffix -o-, v. g.
O. Lat. scélérus, and perhaps finérus (beside funereus) (see Fleck. Jahrb. 1891,
p. 676; both forms are doubtful) (cf. décérus from décor, § 74).. The veter used
by Ennius (Ann. 16 M. cum ueter oceubuit Priamus sub Marte Pelasgo) and
Accius (Trag. 48t R.) may be of this formation, or of the third Decl. like
puber (beside pubes Adj.) and the Cumpounds dégéner, bicorpor, &e.

§ 74. Masc. (and Fem.) ES-stems. The usual termination is -or Nom.,
-oris Gen. (older -ds, -dsis), e.g. décor (beside decus; with Adj. decdrus, O. Lat.
decirem Accus. ; cf. indecérem beside indecirum) ; angor (O. Ind. ahas Neut. ; with
Adj. angustus) ; hinor (with Adj. honestus) ; arbor F. (with Adj. *arbustus, cf.
arbustum, § 28); paedora is Vulg. Lat. for paedores (C. G. L. iv. 270. 4 ‘paedora’
aurium sordes). The -s remains in flés [like Flora F. beside lds M. is auréra
beside I.-Eur. *ausds (Gk. 7jus F.)]; while some have found -és in the Fem.
nouns sédés, beside I.-Eur. *sédég (O. Ind. sidas, Gk. 50s), plebés (Gk. 7700s),
and perhaps aedés beside aedis (see Georges, Lex. Wortf. s.v.) (O. Ind. édhas,
Gk. ai@os) (but see § 51). These Mase. Nouns in -or are widely used as Verbal
Abstracts, connected with Verbs in -eo and Adjectives in -idus, e.g. pallor
(palleo, pallidus), timor (timeo, timidus) (see § 67; and for a list of examples

 

1 Plaut. Curc. 200 immodestis tuis modereris moribus.
§§ 73-77.1|_ NOUN AND ADJECTIVE STEMS. -VYES-. 357

A. L. L, viii. 313). The Verb in -eo is not found with Jluor, fluidus, cruor, criidus
(ef. O. Ind. kravii-, ‘raw flesh,’ Gk. xpéas); the Adj. in -tdus is not found
with favor (fiveo), a coinage of Cicero's time (Quint. viii. 3. 34).

§ 75. Other S-stems. Lat. cértbrum for *ceresrum points to an S-stem like
O. Ind. giras-, ‘the head,’ Gk. «épas, a horn; Lat. ténzbrae for *tenesrae (cf. O.Ind.
tamis-ra beside tdimas, ‘darkness’; cf. Lat. témére, lit. ‘in the dark ”); Lat.
cinis M., sometimes F., resembles Gk. «évis F. in its termination ; Lat. far,
Farris Gen. for *fars, *farsis has a derivative farina for *farisna, Fem. of an Adj.
*farisnus (ef. Goth. bariz-eins, ‘made of barley’).

§ 76. -YES-. This suffix, used to form Comparatives (cf. the
‘Comparative’ use of -yo-, § 4), appears in Latin as -ior Nom.
M., F., -doris Gen., -ius Nom. N. (older -ids, -idsis, -i08), e.g.
sudvior, suavidris, sudvius. The variations of the suffix are not
easy to determine amid the variety of forms in the various
languages (e.g. O. Ind. svad-iyas-, with Nom. svdd-iyan, Gk.
ndlov, Acc. 7diw for *jd.0ca, Goth. sutiza for *sut-izen-, Lith.
sald-és-nis, &c.), but it is certain that the weak grade -is- was
prefixed to the suffix -té- (-thd-?) to form Superlatives in various
languages (O. Ind. svad-i§tha-, Gk. jdictos, Goth. sutists),
though in Latin this formation was ousted by -dssimus (cf.
however mdgis, magis-ter, &c.). These suffixes were originally
affixed to the root without the suffix of the Positive stem (so
Lat. de-ior, O. Ind. a&-iyas-. Gk, dx-iwy from a Positive stem
*dku-), but we have also in Latin, e.g. tduu-ior, sudvior for
*suadv-ior, aspér-ior, &e.

For other details of the formation of Comparatives and Super-
latives in Latin, see ch. vi. section 2.

§ 77. Suffixless Forms. Nouns formed directly from the
root, without any suffix except those of the cases, are especially
frequent as the second element of a Compound, and take in this
position the function of a Nomen Agentis, e.g. Lat. au-spex,
‘seer of birds,’ from the root spek-, ‘ to see,’ parti-ceps, ‘ taking
a share.” They are also found independently, e.g. O, Ind. spés-,
‘a spy,’ often with a high grade of vowel,e. g. Gk. oxéy, an owl
(cf. ch. iii. § 51, p. 254), both from the same root spek- (skep-).
The passage of such forms into the vowel declension is a very
near one, so that we find bystems with and without a suffix in
the same language, e.g. Lat. auspea beside exti-spicus (a late
358 THE LATIN LANGUAGE. [Chap. V.

form of exti-spex), or in different languages, e.g. O. Ind. nat-,
Gk. vais beside Lat. ndv-i-s, O. Ind. tid-atic- beside Gk. 108-a7-d-s,
Lat. prép-inqu-o-s, The Latin Inf. Pass., e. g. ag’ seems to be a
case of a suffixless Verbal Noun (cf. 0. Ind. nir-4jé, dr8é, Inf.
of dr&-, ‘ to look *), just as the Inf. Act., e.g. dgéré, seems to be
the Locative case of a Verbal Stem in -es- (see ch. vill. § 83).

§ 78. Suffixless stems at end of Compounds in Latin. From jungo,
con-jux (cf. O. Ind. sa-yuj-, Gk. od-Cug) and con-junx (O. Ind. yunj- beside
yuj-) ; from ftcio, artifex, carnifex, dpifex, &c. ; from ccipio, manceps, auceps, forceps
for *formi-ceps, princeps, &c.; from dico, jidex, vindex ; from sédeo, déses, réses ;
‘from céino, cornicen, fidicen ; from dgo, remex, aureax, &e. Beside these we have
often O-stems, &c., e. g. jivi-dicus beside ju-dex, prid-igus beside rém-ex, O. Lat.
hosti-capas (hostium captor, Paul. Fest. 73. ro Th.), wrbi-cdipe Voc. (Plaut. Mil.
1055) beside auceps, &e. The presence of suffixless stems in the second part of
Compounds in other I.-Eur. languages forbids us to regard the third Decl.
forms as due to Latin syncope of the O-suffix in the Nom. Case, *avi-cap(o)s,
&c. (see ch. ili. § 16). But in some adjectives original O-stems may have
become I-stems (cf. § 34), and the Nominative form may be due to syncope
or to the analogy of consonant-stems (just as pars Nom. for *parti-s, with
Gen. partis, is due to syncope or to the analogy of r@x Nom. with Gen. rég-is,
&c.), e. g. praecox beside praeciquis and praecoquus, cohors for *cohortis from hortus.
The Feminines add u, ev. g. exti-spic-a, fidi-cin-a,

§ 79. Latin Independent suffixless stems. Lat. réuv(O. Ind. raj-, O. Iv. rig
Gen.) from the root reg-, of régo ; léx from légo ; pés Nom., péd-is Gen. (cf. O. Ind.
pad-, pad-am Ace., Dor. Gk. rovs Nom., 708-ds Gen., &.); fur for *for (ch. iv. § 16)
(Gk. gwp) from the root bher-; 7% (O. Ind. ras) ; hiems (Zend zya, Gk. xuav) ;
ds, the mouth (0. Ind. as-); miis (O. Ind, mi8-, Gk. pis, 0, H. G. mis) ; dic-is
Gen., in the phrase dicis causa, for form’s sake (O. Ind. dis, ‘direction’) from
the root deik-, of dico; nix (Gk. vip-a Acc.) ; sal (Gk. GAs) ; siis, stbus Dat. Pl.
(Gk. os, O. H. G. sii); dita from dico; préc-es Plur. from root prek-, ‘to ask’ ;
vow (O. Ind. vae- F.) beside vico. (On the long vowel, see ch. iv. § 51, p. 254.)

§ 80. II. COMPOSITION. Compounds are seldom resolv-
able into two intact words like Gk. Atéo-xovpot, dpn{-paros, slain
in battle, Lat. séatds-consultum, patres-fimiliarum, jiris-jirandi,
res-publica, O. Engl. Tiwes-deg, ‘Tuesday.’ Sometimes, as in
reduplicated words, one element is reduced almost beyond
recognition, either the second element (in ‘broken’ or curtailed
Reduplication), e.g. Lat. bal-b-us (cf. O. Ind. bal-bala-karémi),
gur-g-es, % whirlpool (cf. O. Ind. gaér-gara-), or more usually
the first, e.g. ci-cindé-la, a glow-worm, while Lat. gur-giil-io,
the throat, guer-quér-us, mur-mur, &c. give equal prominence to both
elements. But generally the full stem without the case suffixes
§§ 78-80.] NOUN AND ADJECTIVE STEMS. COMPOUNDS. 359

is used in the first part of the Compound, e.g. patré-cida (con-
trasted with putres-familiarum), jart-dicus (contrasted with juris-
jurandi). The treatment of these stems which begin the Com-
pound is the special subject of this section.

Owing to the weakening of unaccented vowels in Latin,
every vowel in the final syllable of such a stem was liable to
change under the early accentuation of the first syllable of each
word (ch. ili. § 5). It is therefore often necessary to call in the
aid of other I.-Eur. languages before one can determine the
original vowel in a Latin Compound, e. g. d/é-ger from dla may
be shown to have been originally *alé-ger by Greek sdo-rédpos,
&c. from try (§ 82), though, so far as the Latin form goes, it
might equally well have been *ali-ger, *alé-ger, &c., while in
forceps for *formi-ceps the vowel has been suppressed altogether
by Syncope (ch. iii. § 13), in arctbii for arci-cubii by Dissimi-
lation (id. p. 176). O is the I.-Eur. ‘Composition-Vowel’ par
eaceltence. Not only does it appear in O-stems, but it is often
added to Consonant-stems, and sometimes takes the place of the
-i of A-stems. In Latin post-tonic 5 (like % and other short
vowels) became, as was shown in ch. iil. § 18, % before labials
(later 7), and 7 before other single consonants (except 7), so that
7 is the ‘ Composition-Vowel’ of Latin as 6 of I.-Eur. (cf. Uno-
mammia Plaut., Oimu-mama on a Praenestine inscription, ani-
mamma, an Amazon).

The second part of a Compound is often scarcely to be dis-
tinguished from a suffix. Thus the second part of imbri-dus (cf.
O. Ind. jala-da-, lit. * moisture-giving, § 67) is often called the
‘ suffix’ do-; and the form taken by the final vowel of the stem
before a suffix is often determined by the same laws as before
the second element of a Compound (cf. /abd-ginus with faba-ceus,
faba-tus, faba-rius, faba-lis; imbri-dus with imbri-cus). For
these laws with suffixes, laws often disturbed by the influence of
analogy, e.g. dnilis (from anu-) for *anilis by analogy of sénilis
(from seni-), see the preceding section.

Composition does not play so great a part in Latin as in
Greek (cf. Liv. xxvii. 11. 4 quos ‘androgynos’ vulgus, ut
pleraque, faciliore ad duplicanda verba Graeco sermone, appellat).
The early dramatists and other imitators of Greek poetry incurred
360 THE LATIN LANGUAGE. [Chap. V.

the censure of Quintilian for their attempts to reproduce Greek
compounds like xvpravxnv in Latin (Quint. i. 5. 70 sed res tota
magis Graecos decet, nobis minus succedit: nec id fieri natura
puto, sed alienis favemus, ideoque cum «upravxeva mirati simus,
‘ineurvicervicum’ vix a risu defendimus, alluding to Pacuvius’
line:

Nérei repdndirostrum inctruiceruicim pecus) ;

and Virgil uses a periphrasis like (Averna) sozantia silvis (A. ili.
442) where an earlier poet might have employed a compound like
silvisOnus (cf. silvifrdgus, used by Lucretius, who however com-
plains that the ‘ patrii sermonis egestas’ prevented him from
reproducing the Greek compound éoopdpera, i. 832). The
compound Proper Names of other I.-Eur. languages are, as we
have seen (§ 4), replaced in Latin (and Umbro-Oscan) by
Adjective IO-stems, such as Liecius [on the compound Opiter,
see ch. iii. $16 (9)]. The Latin language does not therefore give
the same occasion as the Greek for a study of the I.-Eur. types
of Compounds, or the various irregularities which disturbed the
normal course of Composition; and a brief account of these
types and irregularities will suffice.

I.-Eur. Compounds, Nouns and Adjectives, are sometimes
classified according to the scheme of the Sanscrit grammarians
[Dvandva or Collectives, Bahuvrihi or Possessives, Tatpurusha
or Determinatives (including Karmadharaya formed of Adj. and
Noun, and Dvigu formed of Numeral and Noun), Avyayibhiva
or Adverbial Compounds], sometimes by the more rational
criterion of the change or retention of the meaning of the
second element (thus the Possessive longi-médnus, ‘ possessing long
hands,’ “long-handed,’ changes its second element from a Noun
to an Adjective, while the Determinative perennt-servus, ‘a con-
stant slave, retains the Noun-meaning of servus), these two
main classes being subdivided according to the nature of the
first element [a Noun or Adj. stem as in longi-manus, a Particle
as In in-certus, vé-sdnus, a Preposition as in con-servus,a Noun or
Adj. Case as in vin-dew, legislator, jurisdictio (contrast jurt-dicus),
the last being, as we have seen, rather Word-groups than Com-
pounds ; cf. parcé-prémus Plaut., b&né-vilus, paen-insiila (contrast
§ 80.] NOUN AND ADJECTIVE STEMS. COMPOUNDS. 361

lévi-densis, solli-citus)], &c. Of Possessive Compounds (Sanscr.
Bahuvrihi) examples are angui-pes, ‘ possessing a foot which is
a snake,’ ‘snake-footed’!; pudédri-cdlor, possessing the colour of
shame,’ ‘ shame-coloured,’ sicc-dcdlus, ‘ possessing dry eyes,’ ‘ dry-
eyed.’ Of Determinatives with first element consisting of (1)
a governed Noun (Sanscr. Tatpurusha): v7ti-sdtor, ‘ planter of
the vine, arti-fer, iqui-lem; (2) Adjective qualifying a Noun
(Sanscr. Karmadhiraya): O. Lat. add-gdlérus, the white cap of
the Flamen Dialis, sudvi-sdvidtio Plaut., vivé-rddi« Cato, lati-
elavus (the Adj. contrary to rule follows the Noun in Pliny’s
equifer, a wild horse ; cf. ovdfer, a wild sheep); (3) Numeral, ¢77-
nummus. (For other examples of Numeral Compounds, see ch.
vi. sect. iii.) Coordinate Compounds (Sanscr. Dvandva, e.g. agni-
dhiman, ‘fire and smoke’) are not found in Latin, except in
Derivatives, e. g. su-dvi-taurilia (but not *su-ovi-taurus), a sacri-
fice of a swine, a sheep, and a bull, séru-fer(c)tarit, those who
offered ‘struem et fertum.’ Scaliger’s rule that a Noun is
never compounded with a Verb was a law of I.-Eur. Noun-
compounds, and is not broken in good Latin, though Tertullian
coins vinci-pes by false analogy of nudi-pes, whose first element
he conceives as a verb, ‘qui pedes nudat’ (de Pall. 5 quem enim
non expediat in algore et ardore rigere nudipedem quam in
calceo vincipedem ?), and Verti-cordia was a name under which
Venus was worshipped. As we have seen-in the chapter on the
Latin Accent (ch. iii.), the line is often hard to draw between
a Word-group (united under a single accent) and a Compound,
e.g. affatim (from ad fatim), denuo (from de novo), Juppiter (from
*Ja- pater, Voc., ch. vi. § 32). From the group per xoctem has been
formed the Compound Adj. pernoz, much as méri-die [a single
word like O. Lat. dieqguinte or diequinti, ch. iii. § 12% (g)], which is
only found in this form in the earlier writers, gave rise to the
Compound Noun meridies; from Sacra Via we have the deriva-
tive Sacravienses, from quarta décima (sc. legio) the derivative
quartadecimini (cf our ‘get-at-able’ &c. formed from the

 

1 These Possessives were originally retained in the Fem. of Compounds
used in Apposition, e.g. angui-pes, like fod0-5dervAos ('Hus), ‘Rose-
‘Snake-foot.? Hence in Greek -os is finger.
362 THE LATIN LANGUAGE, [Chap. V.

phrase ‘to get at’). Holusatrum differs from a word-group by
the fact that the first element remains undeclined in /olusatri
Gen., &. (§ 85); vin-dex, by the fact that -dea (-dia) is not used
as a separate word, though the first element shows the Noun in
its proper Case (for a similar explanation of laudan-dus, &e.,
see ch. vill. § 95). Similarly the Case instead of the Stem
appears in the first element of centum-péda (cf. Gk. éxardu-medos),
beside centi-peda (but see ch. vi. § 76), while the reverse is seen
in the forms multi-mddis (for multis modis), omni-modis (for
omnibus modis), which are used by Plautus. Plautus is especially
fond of whimsical compounds coined on the Greek type, and
often half-Greek, half-Latin, e.g. Pers. 702-5:

Uaniloquidorus Uirginesuendénides

Nugiepiloquides Argentumextenebronides

Tedigniloquides Nummosexpalpénides

Quodsémelarripides Namquameripides: ém tibi.
Caivaedium may stand for cav(um) aedium, domnaedium Accus. for
dimin(um) aedium, as dnimadverto for anim(um) adverto (ch. ii.
§ 52) (dnimaequitas has similarly been referred to anim(1) aeyuitas) ;
Lucr. uses ordia prima for primordia ; and summopere, magnopere,
&e. represent summo opere, magno opere, and the like. Greek
compounds like dpiord-xeup (with a Superlative as first element),
atro-xeip (with a Pronoun) are alien to the spirit of the Latin
language. i

These Noun and Adjective Compounds sometimes retain the
stem of their second element unchanged, e. g. sicc-dciilus, some-
times add a suffix (on dpificus, &c. beside dpifew, &e., see § 77).
Compound O-stem Adjectives, as we have seen (§ 34), tended to
become I-stems, e.g. O. Lat. txermus, class. inermis; and Com-
pound Nouns (and Adjectives) affected the IO-suffix, e. g. /ati-
clavium beside laticlavus (so the Vulgar Compound formed from
the Word-group ¢errae motus assumed the form terrimotiwm,
Prob. App. 198. 32K.). Often the selection of a suffix fora
Compound is determined by the usage in Derivatives from the
simple word, e.g. transmdrinus from trans mare, like marinus
from mare, dibingénidsus from dubius and ingenium, like ingeniosus
from ingenium, simplididrius (cf. ludiarius),
Compound Verbs have normally as their first element a Prepo-
§§ 81, 82.] NOUN AND ADJECTIVE STEMS. COMPOUNDS. 363

sition (see ch. ix.). Whether the Negative Particle 7x might be
used, e.g. ig-nosco, ‘not to notice” to overlook or pardon, is
doubtful (see ch. x. § 18; cf. ne-scio, ne-queo, &e.). But Com-
pounds like aedif‘icare from aedes and facio are really Derivatives
from Compound Adjectives or Nouns, aedificus or aedifer; so
that Scaliger’s law (see above) is not violated (cf. Gk. olxodoudw,
&e. from oixodsuos); so navigare from *navigus, Spitilari from
*opitulus (like philosophari from philosophus). (On these Deriv.
Verbs, see ch. viii. §§ 21, 33.) Bénéfticio, miléfacio, &e. are
really word-groups, and so are cdilefacio, arefacio, &e. (cf. facit aré
Luer.). Crédo comes from an J.-Eur, word-group (O. Ind. ’réd
dadhami, ‘I set the heart to,’ see ch. viii. § 27). Compound
Adverbs like dérépenté, désitibito are discussed in ch. ix., and also
Compound Prepositions like Vulg. Lat. dante (Fr. avant). (For
a fuller treatment of Latin Compounds than can be permitted
within the limits of this book, see Skutsch, Nom. Lat. Comp.)

§ 81. Reduplicated Nouns and Adjectives in Latin. Lat. gur-g-es,
whirlpool, gur-giil-io, throat (O. Ind. gaér-gar-a-, ‘whirlpool,’ Gk. yap-yap-edv,
uvula, yép~yep-os, throat, O. H. G. quer-chal-a and quer-ch-a, Germ. Gurgel) ;
bal-b-us (O. Ind. bal-bal-a-karomi, ‘I stammer,’ Gk. Bdp-Bap-o-s) ; quer-quér-u-s,
cold, shivering (frigidus cum tremore, Paul. Fest. 343. 5 Th., who quotes from
Lucilius febris querquera, the ague) (cf. Hom. Gk. xapaipw), has reduplicated
form like other words for trembling, shivering, such as Germ. zittere from
*ti-trd-mi ; can-cer (cf. O. Ind. kar-kat-a-, Gk. wap-xiv-o-s) ; mur-mur (O. Ind.
mar-mar-a-, Gk. pop-pip-w, Lith. mur-m-iu); ta-ta, like mamma, children’s
words mentioned in Martial’s witty epigram (i. 100) :

‘mammas’ atque ‘tatas’ habet Afra; sed ipsa tatarum
dici et mammarum maxima mamma potest,

and found on children’s epitaphs, e.g. C.I. L. vi. 25808 destituisti, Vitilla
mea, miseram mammam tuam (ef. ch. ii. p. 118 ), qui-squil-iae, shreds of
leather, &c. (Gk. xo-cxvaA-paria) (Caec. Com. 251 R. quisquilias uolantis, uenti
spolia) ; fiber (the I.-Eur. name was *bhe-bhr-u-, O. Ind. ba-bhr-u-, ‘brown,’
0. H. G. bi-bar, ‘a beaver,’ Lith. bé-br-u-s, O. Sl. be-br-ti). Reduplication is
common in onomatopoetie words, e. g. ci-cilus (cf. Gr. xdxxvé, Lith. ku-kii-ti,
‘toery cuckoo,’ &e.), uip-tip-a (cf. Gk. én-op), til-dil-a (cf. O. Ind. ul-ul-i-, Gk. da-
od-iyh), tur-tur. The Latin tendency is to assimilate the first to the subsequent
syllable, e.g. cin-cin-nus (Gk. si-u-vos) (ef. mi-mord-i for earlier mé-mord-t, &c..
see ch. iv. § 163). Often a reduplicated Noun is a derivative from a redupli-
cated Verb-form, e. g. ci-cindé-la from *ci-cinde-o, a reduplicated form of candeo,
like Gk. 5:-8ax-7 from 6:-Sdoxw, Ke-xpary-nds from Ké-«pay-a. (On these redupli-
cated Verb-forms, see ch. viii. § 9.)

§ 82. A-stems show sometimes 4, but usually 6 in I.-Eur. compounds,
364 THE LATIN LANGUAGE. [Chap. V.

(e.g. GK. mxn-pdpos and Nixd-payos, O. Ind. urvara-jit- and ukhia-chid-, Lith,
szikszné-sparnis and gatva-raisztis, Gaul. Teuté-bodiaci, Goth. airpi-kunds),
so that Lat. ¢ of ttbi-cen, Gli-ger, &c. must have been originally 0. The long a
is probably seen in fiéba-ginus, dled-ginus, as before suffixes like Romd-nus, ala-
ris. Stems in -ia seem to show a similar divergence of forms, tibi-cen with
7 for id (§ 4), and before a suffix vid-ticus,

§ 83. O-stems appear with -3 in Gk. inmé-dapyos, Gaul. Dévo-gnata, Epo-
rédia, Teut. Austro-valdus, &c., so that Lat. ¢ of belli-ger, magni-ficus, earlier %
before a labial, e. g. magnu-ficus, is probably a weakening in the unaccented
syllable of original 6. O. Lat. spellings (mostly before a labial), like Uno-
mammia (Plaut. Curc. 445 ; ef. Oinu-mama, an Amazon, on an old Praenestine
cista, C. I. L. i. 1501), sescento-plagus (Plaut. Capt. 726), albo-gdlerus (Paul. Fest. 8.
6 Th.), Ahéno-barbus may thus be genuine relics of the oldest spelling, though
late compounds like mdlo-grandtum must be imitations of the Greek. Before
« vowel this -6 is elided in Latin (as in Greek, &c., e.g. inm-aywyés), ©. g.
magn-dnimus, aequ-dnimitas, forms like multi-angiilus being late. Lat. -I0-stems
show -ié- for -ii- by Dissimilation (ch. iv. § 13) before a suffix in sdcte-tas, anaie-
tas, &e., but -i- in Lucilius, &c. (ef. hostt-lis from the I-stem hosti-). Like sescento-
plagus, &c. is O. Lat. sicio-fraudus or sociu-fraudus in Plaut. Pseud. 362. (The
Palatine MSS. have o, the Ambrosian Palimpsest u.) Latin ro-stems show
-ér- by rule (ch. iii. § 15. (8)], e. g. sacer-dds, but sometimes -ri-, e. g. sacri-fex,
after the law by which -ri- became -ér- had been forgotten.

§ 84. I-stems had i in I.-Eur. compounds (e.g. O. Ind. tri-pad-, Gk.
tpimous, Gaul. tri-garanus, O. Engl. pri-féte, Lith. tri-kdjis, O. Sl. tri-zabii),
and so Lat. tri-ennium, tri-géminus, though by rule -ri- [at least unaccented -ri-,
ch. iii. § 15. (8)], became -ér- in Latin, c. g. ter-geminus, and before a labial 7 was
by the older spelling properly w, e.g. dcriifolios Cato (R. R. xxxi. 1), as before r
it was ¢, e.g. légériipa Plaut. Before a vowel this i (now become y) might be
dropped, e.g. fun-ambulus (coined by Messalla to express Gk. cxorvoBarns,
Porph. ad Hor. S. i. ro. 28), from *funy-ambulus, sem-ermis for *semyermis.

§ 85. U-stems had originally v (e.g. O. Ind. svadu-rati-, Gk. #5u-(F ems,
Gaul. Catu-riges ; O. H. G. Hadu-mar, Lith. virszi-kalnis), so that the older
spelling méinii-festus preserves the earliest form (class. méni-festus). From
diphthongal U-stems we have nau-frdgus (see ch. iv. § 46), bi-caeda beside bévi-
cidium, bu-star and bo-star, an ox-stall (ch. iv. § 42), ja-glans (Gk. Ads Badavos)
(Macr. iii. 18. 3). Dies-pitris (C.L. L. xi. 3259), Dies-pitri Arnob. ii. 70 (ef.
Dies-pitrem, Macrob. i. 15. 14 ‘ut diei patrem’) may be a case of the declension
only of the second part of a word-group, like holus-atri for holeris atri (§ 80), or
pronouns like alter-uter, alter-utrius Gen. (ch. vii. § 29). On Juppiter, see ch. vi.
§ 32.

§ 86. N-stems show, as we have seen (§ 54), their close connexion with
O-stems. by substituting -6 for -n whether in the first half of a compound,
e.g. Gk. dxpd-erov, stithy, from dxuwr, Lat. himi-cida, from homo (cf. Goth.
guma-kunds, ‘ of male sex’), or in the second, e. g. Gk. Sparuos beside dpuaipor.
We find also n in Gk. évopd-«dvros, so that the en of Lat. nomen-clitor, &c. was
originally n; also the ‘Composition vowel’ -6- added to the stem, e. g. Gk.
prev-o-BhaBhs, Lat. tmagin-i-fer.
§§ 83-90.] NOUN AND ADJECTIVE STEMS. COMPOUNDS. 365

§ 87. R-stems, like N-stems, take the weak grade of the stem suffix before
a consonant, r (e.g. O. Ind. pitr-Srivana-, Gk. rerpd-yvos, Goth. bropru-lubd),
before « vowel, r (e.g. O. Ind. pitr-artham, Gk. nat p-wvtpios). and often add
the ‘Composition vowel’ 5 (e.g. Gk. natpo-pévos), Since Latin -ér- (for
I.-Eur, 1), -ri- (for L-Eur. -rd-} would generally become in the unaccented
syllable -é-, it is not easy to decide on the origin of -ér- in each case. For
patri-cida, patri- cus, &e. we should expect *patercida, *patercus ; cf. acertas on the
Aes Italicense. (0.1. L. ii. 6278, 1. 36, of 176-180 A.D.)

§ 88. Dental and Guttural Stems. The frequent interchange of conso-
nantal with I-stems in Latin (§ 34), suggests that the -i- of dent-i-frangidiilus,
péd-i-stquus, 1¢g-i-fiigium, may have been original i. It may also have been the
6, which is often used as ‘ Composition vowel,’ with these stems, in other
languages [e.g. Gk. Spaxovr-d-yardos, vid-d-Bodros, Gaul. Carant-o-magus,
Cinget-o-rix (lit. ‘king of warriors’ ; ef. 0. Ir. cing, ‘a warrior,’ from cingim,
‘I march’)]. Before r we should have 2, e. g. légériipa (Plaut.), and before
a labial % (6) in the older spelling (see ch. iii. § 18).

§ 89. S-stems. The S-stems, like other consonant-stems in Latin, often
show i, e.g. miir-t-idus, jir-i-dicus, whether the i of I-stems or the ‘Compo-
sition vowel’ 6 it is not easy to decide. But we find also the normal stem,
e.g. mis-cipiila (cf. Gk. pio-pdvos). The ES-stems took -és- in I.-Eur. com-
pounds [e.g. Gk. caxeo-pdpos (the poetical form), Goth. sigis-laun}, and so
before suffixes in Latin tempes-tivus, hines-tus, &c., though -iis-, by Analogy of the
Nom. Sing., is found in dnus-tus, &. In Latin compounds the stem suffix is
either displaced by % [the 6 of Gk. éro-moids (the prose form), dAj0d-pay7is, &e.
O. SI. tudo-totini), e.g. foedi-fragus, or augmented by it, e. g. foedér-i-fragus,
hinor-t-ficus (cf. O. Sl. Gudes-o-toéint, Goth. aiz-a-smipa).

 

§ 90. Stem-suffixes and Composition in Romance. The Latin suffixes
have for the most part remained productive in Romance (e. g. Ital. cannonata
with the TO-suffix), though their meaning has sometimes undergone a change.
Thus -ino-, which has ousted -e(y)o-, the suffix denoting material (§ 4), e. g.
Fr. ferrin, ivoirin, has acquired in Italian and Portuguese a Diminutive
sense, e.g. Ital. tavolino, Port. filhinho ; -dceo- has taken the sense of large
size or inferior quality, e.g. Ital. corpaccio, acquaccia. And new suffixes have
been gained from other languages, such as -issa (from the Greek), a fem. suffix
used especially in titles, e.g. Ital. duchessa, Fr. duchesse ; -itto- with Diminu-
tive sense, e.g. Ital. biglietto, Fr. amourette ; -ia (from the Greek), denoting
Abstracts, e.g. Ital. villania, Span. villania, cortesia, and soon. (For a full
account of the Romance suffixes, see Meyer-Ltibke, Rom. Gram. ii. pp. 448 sqq.)
Of Composition these varieties are noteworthy : Word-groups like Fr. pour-
boire ; Bahuvrihi Compounds like Fr. rouge-gorge; Dvandva Compounds
like Ital. acqui-vento, ‘wind and rain’; Verb with Noun (violating Scaliger’s
rule), e. g. Fr. garde-robe, Ital. guarda-boschi, becca-fico, lit. ‘ peck-fig,’ with
the Verb apparently in the 2 Sg. Imperat. (see Meyer-Liibke, ib. pp. 577 8qq.).
CHAPTER VI.
DECLENSION OF NOUNS AND ADJECTIVES.
COMPARISON OF ADJECTIVES. NUMERALS.

$1. I. DECLENSION OF NOUNS AND ADJECTIVES.
The I.-Eur. Noun had three Genders, Masc., Fem., and Neut.,
three Numbers, Sing., Dual, and Plur., and at least eight Cases,
Nom., Gen., Dat., Acc., Voc. (if the Voc, may be called a Case),
Abl., Instrumental, and Locative (the Dat., Abl., Instr., Loc, are
not always easy to discriminate in the Plural). The three
Genders are retained in Latin, but of the Numbers the Dual has
disappeared, though traces of it remain in the Numeral forms dué
(§ 59), 00/6 (lit. ‘two sets of four’ ?), and in the Pronoun form
ambé (ch. vil. § 29)1. (On vi-ginti, see § 74.) Of the Cases the
Voe. hardly survives except in O-stems (§ 31), and the Instr. has
left only doubtful traces of itself in some Adverb forms (§ 36) ;
the Locative became by the operation of the phonetic laws of the
language indistinguishable in A-stems from the Gen. Sg., while
in O-stems it seems to have ousted the Genitive (§ 17), and in
Cons.-stems the Ablative (§ 33).

The I.-Eur. Cases were indicated sometimes by the addition of
suffixes, e.g. -s for Nom. Sg. Masc., -m for Acc. Sg., sometimes
by modification of the stem, e. g. *patér Nom. Sg. of stem *patér-,
‘a father,’sometimes by both, e. g. *patr-ds, *patr-dm Gen. Sg.,
Pl. of the same stem. This modification (called by the Germans

 

' With the gradual loss of these a suffix still retained in ‘oxen,’ &e.
peculiar suffixes (in Plautus ambos The Dual is rapidly disappearing
has begun to oust ambd Acc., and in at the present time in the Prus-
Late Latin ambis ousts ambobus) we sian dialect of Lithuanian, though
may compare our plural ‘shoes,’ with in the other dialects it is better pre-
disuse of the older suffix of ‘shoon,’ served.
§ 1.] DECLENSION OF NOUNS, ETC. 367

‘Abstufung ’) of the stem is due to the different accentuation of
the different cases; and the J.-Eur. cases have been divided into
(1) ‘Strong’ Cases, viz. the Nom., Voc. M. and F. of all
Numbers, and the Acc. M. and F, of the Sing. and Dual, along
with the Loc, Sing.; (2) ‘ Weak’ Cases, where the accentuation
of the suffix weakened the stem, e.g. *patr-6s (Gk. warp-ds),
This alternation of unweakened and weakened stem is a feature
of the Sanscrit declension, but has been effaced in most other
languages by the natural tendency to make one Case like another
in everything but the suffix (cf. Hom. Gk. warépos on the
analogy of warépa), Lat. patr-em on the analogy of patr-is), and
often the only trace left of it is the existence of varieties of the
same stem ; e. g. the varieties polen- (Lat. polen-ta) and poll- for
*poln- (Lat. poll-en) may be due to a former declension with the
stem polen- in the strong, and the stem poln- in the weak cases.
[The variations homén-, e.g. himinis, hemon-, e.g. O. Lat. hémonem
(Paul. Fest. 71. 18 Th.), and homd(x)- have been similarly ex-
plained.| The appropriation of different stems to different cases
leads to what is called Heteroclite declension ; thus 7¢ér- is the
stem appropriated to the Nom. Sing., t¢inér- the stem appropri-
ated to the other cases in Latin [so séuée- (ch. v. § 69) Nom. Sg.,
but sén(z)- in the other cases, siipelleg-, for super-leg-, a suffixless
stem (ch. v. § 77) in the Nom. Sg., supellectili- for super-lectilc- (an
Adjectival LI-stem, ch. v. § 40) in the oblique cases]. Lat. fémir
Nom., féminis Gen., &c. retain a very ancient type of heteroclite
declension of Neuter Nouns, in which the consonant R was the
mark of the Nom., Acc., the consonant N of the Gen., Dat., &c. (see
ch. v. § 56). In discussing the declension of the Latin Noun it will
be better to put aside the conventional division made by the native
grammarians, and to class nouns rather according to the final letter
of their stems as A-stems, O-stems, &e. The heterogeneous com-
position of the fifth declension (rés stem 7¢i-, fidés stem Jjid-é-,
tristities stem tristitid- (?)) has been already pointed out (ch. v.
§ 51); also the close connexion of Consonant and I-stems (ch. v.
§ 34; ef. below, § 46, on civitatium, férentinm, audacium, &e.), and
the absorption of the U- into the O-declension (ch. v. § 49).
O-stems had, as we saw (ch. v. § 34), a tendency, when used
as Adjectives, to become I-stems, e.g. tnermis, O. Lat. mmermus,
368 THE LATIN LANGUAGE. (Chap. VI.

from the stem armo-; and in Vule. Lat. we find a similar confu-
sion of I-stem Adjs. with O-stems, e. g. tristis non ‘tristus,’ Prob.
App. 198. 3K. (cf. érista N., Rossi, i. 842, of 472 a.D., Ital. tristo).

In the Romance languages the Latin Declensions have been
‘levelled’ to a much greater extent than the Latin Conjugations,
owing to the fact that the distinction of the Cases came to be
expressed rather by Prepositions than by Case-suffixes. The
sign of the Genitive was the Preposition de, of the Dative ad,
and so on; and these Prepositions ceased to retain their classical
construction [as early as the first cent. a.D. we have on a Pom-
peian graffito (C. J. L. iv. 275) Saturninus cum discentes]. Thus
the Cases have been reduced in most languages to one, though
O. Fr. and O. Prov. retain the distinction of the Nom. and Ace.
{e.g. O. Fr. chars Nom., char Acc. of Lat. cérus; suer Nom.,
serour Acc. of Lat. séror), and in Roumanian we see the Dat.
Fem. in roase Dat. (Lat. rdsae) beside roas’ Nom. (Lat. rdsa), &e.
More important was the distinction of Singular and Plural; and
so the two Numbers are always distinguished, except (in pronun-
ciation) in French, though a Latin Plural has often become
a Romance Singular, e. &. Zigna, Ital. legna, dibtia, Ital. bibbia.
The fourth Declension has been merged in the second, a process
which shows itself very strongly even in the conversational Latin
of Plautus (ch. v. § 49), the fifth in the third or first ; and forms
like Span. polvo (Vulg. Lat. *pudvus for pulvis), Ital. serpe (Vulg.
Lat. serpi- for serpens) illustrate how the ‘levelling’ influence of
Analogy gradually removed the distinctions of declension. (For
particulars of the Romance declension, see Meyer-Liibke, Gram.
Rom. Sprach, i. pp. 1 sqq.).

As regards Gender, the laws according to which one Noun was
Masculine, another Feminine, and a third Neuter in I.-Eur., have
not yet been determined. The Neuter Gender seems to have
been restricted to things without life. Thus while names of
trees were mase. (in O, Ind., e. g. bhiirjas, ‘a birch ’) or fem. (in
Greek and Latin, e. g. dyyds, fagus, ‘a beech’), names of fruits
were Neuter (e.g. O. Ind. amram, ‘ mango-fruit,’ beside amras,
‘mango-tree,’ Lat. mdlum beside mdlus) (see Delbriick in Brug-
mann’s Grandriss, iii. ch.i.); the difference between the gender of
names of rivers in O. Ind. (fem.) and in Gk. and Lat. (masc.)
§1.] DECLENSION OF NOUNS, ETC. GENDER. 369

may be due to the different gender of the word for ‘river’ in
these languages (O. Ind. nadi F., Gk. worayds M.), just as the
names of the months, winds, &c. in Latin are really Adjectives
agreeing with mensis, ventus (e. g. Jdnuarius, Februdrius, Martius ;
Auster, Caurus, Févdnius). But the proneness of nouns to take
a new gender by analogy of a noun which had a similar termin-
ation, or a kindred meaning, or with which they were often
jomed in speech, makes it impossible to trace the original
gender of each and every noun. The feminine gender seems to
have been associated with Abstract Nouns, e.g. Lat. optio,
‘choice. But if an Abstract Noun came to be used as a Con-
crete, it might change its gender; and so optio in the sense of
‘a centurion’s assistant’ was masculine. Similarly agr%edda, lit.
‘field-tillage, became mase. in the sense of ‘a field-tiller’; and
this is probably the reason why A-stems, which were associated
with the feminine gender in I.-Eur., are often mase. in the
various I.-Eur. languages (ch. v. § 2). O-stems were similarly
associated with the masculine (with Nom. Sg. in -ds) and the
neuter gender (Nom., Ace. Sg. in -6m); but we have feminine
O-stems in the various languages, e. g. Gk. 636s, xéAevOos, Aios,
wWigos, and names of trees like gnyds, &e. In Latin (as we have
seen, ch. iv. § 49) we seem to detect a tendency of these feminine
O-stems to pass into the fourth declension. Thus fagus is treated
like a U-stem (fourth Declension) in the Culex, 1. 139 : umbrosae-
que patent fagus (cf. Varro ap. Charis. p. 130. 5 K.); and the
declension of ddmus (an O-stem in the earlier literature, ch. v.
§ 49; cf. Gk. dduos?), clus (Gk. dros), nitrus (Gk. vuds for *vucds)
has been so explained.

The confusion of masculine and neuter O-stems may be illus-
trated by the words col/um, which in Plautus is coldus (I.-Eur.
kolso-, Germ. Hals M.), and #érus, which in Plautus is wterwm
(cf. O. Ind. udéram). (Other exx. in Neue, 1’. p. 529.) (On
licus, loca Plur., &e., see § 45.) In Late and Vulgar Latin the
masculine seems to oust the neuter in these stems, as we see from

 

1 Thus the fem. gender of diés 2 0. Slav. domi is a U-stem, but
has been ascribed to the analogy of in Slavonic the O- and U-declensions
nox. have been mixed up, as in Latin.

Bb
370 THE LATIN LANGUAGE. (Chap. VI.

the precepts of the grammarians [e. g. Caper (first cent.) censu-
res pratus 105. 6 K., solius 94. 19 K., and expresses himself
forcibly about cereber 103. 6 K.: hoc cerebrum est nam ‘ cereber’
qui dicunt sine cerebro vivunt], and from the ‘ sermo plebeius’
of Petronius (e.g. fatus, 42, p. 28.13 B.). On plebeian epitaphs
we find collegius (e.g. C. I. L. xi. 4579. 4749), monimentus (e. g.
ib. vi. 19319), and especially fatus in the stock-epitaph of the
lower classes (like our ‘ Affliction sore long time he bore’):

noli dolere mater eventum meum.
properavit aetas: hoc voluit fatus mihi.

(On the disuse of the Neuter in later Latin, see Appel, De genere
neutro imtereunte in lingua Latina, Erlangen, 1883).

In the Romance languages the Neuter has disappeared, though
it has influenced the formation of the Plural. Thus O. Prov.
pratz points to pratus for pratum, but Ital. tempora Plur. (tempo
Sg.), braccia Plur. (braccio Sg.), retain the Neut. Plur. formation‘
tempora,bracchia. (On change of gender see Meyer-Liibke, ii. 416.)

The Adjective Declension took advantage of the connexion of
O-stems with the masculine and neuter, and of A-stems with
the feminine gender, e.g. I.-Eur. *néwis M., *néwa F., *néwom
N. (Lat. ndvus, -a, -um) (see ch. v. § 2). Consonant-stems’
formed their feminine with the I-suffix, e.g. I.-Eur. *népti F.
beside népot- M. (Lat. neptis beside népds.) (On the I-suffix, see
ch. v. § 51, and on other feminine formative suffixes, as in Lat.
gall-ina beside gallus, rég-ina beside réx, consult the same chapter.)
The Greek extension to the Feminine of the masculine suffix -os
in Compound Adjectives, which were originally Nouns in Appo-
sition, is not found in Latin, e. g. fododdxrvAos Hus, lit. ‘ Dawn
Rose-finger’ (ch. v. § 80), though we have in O. Lat. lupus
femina, agnus femina, &c. (e.g. Ennius, A. 59 M., in the story of
the nursing of Romulus and Remus, has :

indotuetur ibi lupus femina).

But a usage peculiar to Latin is the extension to the Neuter of
the S-suffix of the Masc. and Fem., not merely in Present Parti-
ciples like férens, where ferens Neut. may represent an older
*ferent (ch. iv. § 105), and in vétus, which was probably originally
§ 2.) DECLENSION OF NOUNS, ETC. NOM. SING. 371

a noun (Gk. Féros, § 55), but also in Adjectives like audax (faci-
nus audax, Plaut.), dives (dives opus, Ovid). It is to be com-
pared with the extension of Derivatives in -trix (properly
feminine, e.g. victrices lauros, Virg.) to Neuters, e. g. victricia
arma, Virg., though victrix Sing. is not used with a Neuter Noun
till Late Latin (see Neue, ii*, p. 40; and cf. below, § 16). The
distinction in RI-stems! between the Nom. Sg. Mase. in -er, and
the Nom. Sg. Fem. in -ris is not always found in the earlier
authors (e.g. Ennius has somnus acris, A. 400 M., and acer hiemps,
A. 471 M., and Virgil himself has a/acris Masce., 4. vi. 685), and
is not rigorously enforced even in classical Latin, e.g. mediocris
Mase., illustris Mase. (see Neue, ii*, p. 15). (On the Romance
declension of Adjectives, e.g. Ital. buono M., buona F., Span.
bueno M., buena F., O. Fr. bon, bone, &c., see Meyer-Liibke,
Rom. Gram. ii. p. 75.) (On the Pronominal declension of certain
Adjectives in J.-Eur., see ch. vii. § 29.)

§ 2. Nom. Sing. I. Mase., Fem. A-stems took -a in
I.-Eur, (e.g. O. Ind. Aéva, ‘a mare, Gk. ydpd). By the time of
the oldest Latin poetry this 4 has in every Nom. of an A-stem
been shortened to @, a shortening which may have begun in dis-
syllables with short first syllable, e.g. Zera, where the working
of the Law of Breves Breviantes would shorten the fittal syllable,
héva like edvé, hivé (ch. iii. § 40); though an early reduction of
this -a, as well as of the -a of Nom. Pl. Neut. of O-stems (§ 45),
on Italian soil is indicated by the fact that in Oscan both have
been replaced by an o-sound (in Oscan alph. G, in Lat. 0, in Gk.
o), and in Umbrian are written sometimes -a, sometimes -o (in
Umbr. alph. -u), e.g. Ose, vio ‘ via,” Umbr. mutu and muta
‘mulcta’ (see ch. ii. §.1)*. O. Lat. Aosticapas (hostium captor,
Paul. Fest. 73. 10 Th.), paricidas quoted from the Laws of Numa
(Paul. Fest. 278. 10 siqui hominem liberum dolo sciens morti
duit, paricidas esto) may be analogous to the Greek usage of

 

' Many were originally RO-stems Ital, allegro ; ef. Ital. campestro, &c.
(ch. v. § 40; cf. O. Lat. hildérus), They 2 For the Umbro-Oscan forms of
are often RO-stems in Late and _ these and the other case-suffixes, see
Vulgar Latin, e.g. Vulg. Lat. dilécer, Class. Rev. ii. pp. 129, 202, 273.

Bb2
37% THE LATIN LANGUAGE. {Chap. VI.

adding -s to a fem. Abstract d-stem when used as a masc. Con-
crete, e.g. veavids, a youth, from *veavid, youth, though the
usual practice in Latin is to retain the ordinary Nom. form, e. g.
agrt-cola, a field-tiller, originally ‘ field-tillage.

YA-stems, the fem. of consonant-stems, &c. (ch. v. § 51), which
in O, Ind. take -i, e.g. bhéranti, Pres. Part of bhr-, ‘to carry,’
napti, Fem. of ndpat-, ‘ grandson,’ in Greek -.d, e. g. pépovea for
*epovtyd, show in Latin -ia, e. g. prae-sentia, possibly also -%s,
e. g. *ferentts which became ferens, neptis (older Lat. -7s?). Beside
-ia (first Decl.) we find -7és (fifth Decl.), e.g. mdtér-ces beside
materia. The exact relation between O. Ind. -i, Gk. -.a, Lat.
-ia and -iés has not yet been determined (see ch. v. § 51).

O-stems took -os in I.-Eur. (e.g. O. Ind. vfkas, ‘a wolf” Gk.
Avcos, Gaul. tarvos,‘a bull’), and in Latin, e.g. Zupus, taurus,
older */upos, *tauros (ch. iv. § 19). In RO-stems the final -ros,
when preceded by a consonant, was changed by a phonetic process
common to Latin with other Italic languages to -er, e.g. Lat.
dager for *agro-s (Gk. dypés), Umbr. ager; even when a short
vowel precedes, we find, e. g. Lat. sdcer (in Plautus socérws) for
*soctiros, *socéros (I.-Eur. *swékivos), viv for *vtros, sdtur for
*satiuros, ;

YO-stems, whose suffix in I.-Eur. seems to have varied with i
(e.g. Gothhairdeis, ‘a herdsman ;’ Lith. gaidfs, ‘a cock,’ beside
svéczias, ‘a guest’) show in Latin usually -2ws, but in familiar
language also -is, e. g. Coruélis and other proper names.

All other stems took -s in J.-Eur. and in Latin, e.g. dvis
(O. Ind. dvi-s, Gk. 6(F)is, Lith. avis), médniis (O. Ind. svadis,
‘sweet, Gk. jdvs, Lith. stints, ‘a son’), vis (Gk. is), sts (Gk. os),
rés (O. Ind. ras), milés for *milets (the last syllable is scanned
long by Plautus, *miless, ch. ii. § 133), més for *mis-s, réx, &e.
S-stems have -és (e.g. I.-Eur. *dus-menés, O. Ind. dur-manas,
Gk. dve-yerjs) or -ds (e.g. L-Eur. *ausos, Hom. Gk. jos; cf.
O. Ind. u¥as), and so in Latin, e.g. pibés, hinds M., later honor,
ténor M. (beside tenus N.). But N-stems, which in J.-Eur.
showed (1) -dn, -én,(2)-6 (-é)[e. g. (1) Gk. kd, worpijy, (2) O. Ind.
iva, é dog, O. Ir. ci, Lith. s2i], show -d in Latin, e. g. hdmé, riitis,
uirco * virgo’ on the very ancient Dvenos inscr. I.-Hur. R-stems
had similarly (1) -Gr, -ér, (2) -6, -6 [e.g. (1) Gk. unrnp, dérwp,
§ 3.] DECLENSION OF NOUNS, ETC. NOM. SING. 373

(2) O. Ind. mata, dita, Lith. moté and mété, sesti, ‘ sister”), but
display only the first formation in Latin, e.g. mater, ditir, sdror
(in O. Lat. *matér, dator, sorér ; see ch. iii. § 49). So in Umbrian
karu, ‘a part’ (Lat. cérd), with u as equivalent of Lat. 6, but in
Ose. statif ‘ statio, fruktatiuf ‘ fruitatio, Gittiuf ‘ utitio’ with -f
for ns, the ns being perhaps a re-formation just as carnis some-
times replaces caro in Latin (Prise. i, p. 208. 19 H.). Umbro-
Oscan R-stems form their Nom. like the Latin, e.g. Umbr. ars-
Jertur, Osc. censtur ‘censor’ (both with uw, the equivalent of
Lat. 0),

§ 8. Nom. Sing. of A-stems in Lat. We have seen in ch. iii. § 43 that
all supposed instances of -@ Nom. Sgein early poetry are illusory [aguild, Enn.
A. 149 M. is a case of metrical lengthening of a short syllable before the
penthemimeral caesura ; copia, Plaut. Mil. 1226 shows ‘ syllaba anceps’ at the
end of the hemistich ; familia, Trin. 251 is a proceleusmatic ( fiimilid) represent-
ing an anapaest, ana soon]. The only genuine instances are Greek words
with -d}, which are long in later poetry too, ev. g. Nemed (Stat. Theb. vi. 516).
Greek Nominatives in -as, -ns were especially in the older literature changed
to the ordinary Latin Nom., e.g. Anchisé (Enn. A. 19 M.), Aenea (Quint. i. 5.
61 ne in a quidem atque s litteras exire temere masculina Graeca nomina
recto casu patiebantur, ideoque et apud Caelium legimus ‘ Pelia cincinnatus’
et apud Messalam ‘bene fecit Euthia,’ et apud Ciceronem ‘Hermagora,’ ne
miremur, quod ab antiquorum plerisque ‘Aenea’ ut ‘Anchisa’ sit dictus),
and similarly in classical Latin postd, nauti, biblidpold, &e. (cf. Atridd, Propert.
li. 14. 1; Marsyd, Hor. S. i. 6. 120; and for other instances see Neue, Formenl.
i, pp. 31 sqq.) ; though they usually in the classical literature retain -ds, -és,
just as Greek Noms. in -7 retain @, e. g. Andrimédché, or take the Latin suffix,
e.g. epistiild. Hosticapas, quoted by Paul. Fest., is a strange form. Compounds
of cépio usually show -ceps, e.g. miini-ceps, while Plautus has urbi-cépe Voc.
(Mil. 1055). Paricidas (i.e. parricidas, for the double consonant was not written
double till Ennius’ time, ch. i. § 8) is indeed in the Republican and Classical
period an A-stem (e.g. parricida Voc., Plaut. Pseud. 362; but adjectivally
mitri-cide homo Voc., ‘ you coward,’ Epid. 333) ; however, if these Masc. A-stems
originally took -as in Latin, they had conformed to the ordinary usage of
Fem. A-stems as early as the beginning of the second cent. s.c., for Plautus,
Rud. 652, has légiripa (better legerupa, ch. iii. § 20) Nom. Sing. with its last
syllable elided :

légerupa, inpudéns, inpurus, inuerecundissimus,
where legerupa, originally ‘the act of law-breaking,’ a fem. Abstract, then

mase, and Concrete, ‘a law-breaker,’ cannot have been written by Plautus
legerupas (ch. ii. § 137). The names on Oscan inscriptions (Map)as (Zv. I. I. I.

 

1 agoea in Enn. A. 567 M.: multa word seems to have been dyua, and
foro ponet et agoea longa repletur, one MS, reads ponens ageaque.
is not a certain example. The Greek
374 THE LATIN LANGUAGE. (Chap. VI.

253 from Messana), Maras (I. F. ii. p. 437 from Puteoli or Cumae), whence the
derivative Osc. Maraio-, Falisc. Mareio-, Lat. Mdrius (ch. v. § 4), and Tanas
(Zvet. 102 from Samnium) suggest that Noms. in -as (for -As would probably
be syncopated in Oscan) were used in Oscan like Noms. in -ds, -ys in Greek.
But they may belong to some un-Italic dialect. Hav6ias is Osc. Santia
(Zvet. 228).

§ 4, RO-stems. The substitution of -er for -ros is extended even to Gk.
loanwords, e.g. Alexander, though the usage varied, e. g. Euandrus and Euander
in Virgil (see Neue, Formeni. i, p. 77). It is not found in Latin words when
a long syllable precedes -r0s, v. g. séve-rus, stisur-rus, médtii-rus, nor in these
tribrach words hiimérus, niimérus, titérus. Plautus has the tribrach stem sécéro-
(I.-Eur. *swékiro-, O. Ind. évadura-, Gk. éxvpds, Lith. szeszitiras) with Nom.
socerus (Men. 957), and Priscian (i. 231. 13 H.) says that puerus was used by the
older writers, though he is perhaps referring to Voc. puére, probably the in-
variable form of the Voc. Sg. of puer i Plautus, for puerus is not found in
any extant literature. It is not however certain that puer was originally an
O-stem, for the cognate Greek word is ma(F)i5-, and the old Saturnian poets
used puer as fem. as well as mase. (see Charis. 84. 5 K.; Prise. i. p. 232 H.),
e.g. Naevius, Bell. Pun. ii.:

prima ineédit Céreris Prosérpina pier,

while the curious compound slave-names Marcipor, Gaipor, Quintipor, &c. are
I-stems, Plur. Marcipores (e.g. Plin. xxxiii. 26). Inférus, with siipérus, is used in
Livius Andronicus’ translation of the Odyssey (at least in the later dactylic
version of it):

inferus an superus tibi fert deus funera, Ulixes?,

but infer and super by Cato (R. R. cxlix. 1 ubi super inferque uicinus permittet),
like citer (Cato, Orat. fr. lxii. p. 65 J.). Of compounds ending in Verbal
Adjective in -rus we have always mori-gérus (e.g. Plaut. Capt. 966), prd-pérus,
(jini-perus is by some explained as jun(on)i-pirus, ‘Juno’s pear,’ like ja-glans,
‘Jove’s acorn,’ by others is derived from *jinus, cf. jun-cus, and pério), but the
normal usage favoured -ger, -fer, &c. (for details, see Kiihner, Lat. Gram. i.
pp. 278 sqq.). All this points to -er having originally been substituted for -ros
only when a consonant preceded, e. g. dger for *ag-ros. After the vowel i we
find -ros curtailed to -r in vir (but pirus beside pirum) ; after the vowel u, in
sdtiir (as early as Plautus; see A, L. L. v. 34), although volttrus (class. vultur,
-tiris Gen.) is used by Ennius, 4.138 M. The tendency to curtail -rus to -r
increased in Vulgar Latin, as we see from the Probi Appendix 197. 30 K.:
barbarus non ‘barbar,’. although the opposite tendency (due to Greek
influence of 8. Italy ?) is also mentioned (ib. 198. 26): teter non ‘tetrus,’
aper non ‘aprus.’ In the Umbro-Oscan dialects -los was similarly changed
to -el, e.g. Osc. famel, apparently for fam-lo-, ‘a dweller’ (cf. Osc. faamat,
‘he dwells’) with the same formation as Lat. figulus, stem fig-lo- from figo, &c.
(ch. iv. § 51). But this was not the Latin usage, e. g. fdmiilus, bibiilus, mascilus,
&e. (ef. ch. iv. § 10), though Ennius borrows from his native Osean the form
Jamul in his reference to Servius Tullius (A. 336 M.):

mortalem summum fortuna repente
reddidit, ut summo regno famul oltimus esset,
§§ 4-6.] DECLENSION OF NOUNS, ETC. NOM. SING. 375

in which he is imitated by Lucretius, who echoes the rhythm of this passage
(iii. 1035) :

Scipiadas, belli fulmen, Carthaginis horror,

ossa dedit terrae proinde ac famul infimus esset.

The S. Italian forms jigel, mascel found their way into plebeian Latin under
the Empire (Prob. App. 197. 28 K. figulus non ‘ figel,’ masculus non
‘mascel’); but the usual Vulgar Latin form masclus is mentioned in the
same treatise (197. 20 K. speculum non ‘speclum,’ masculus non ‘masclus,’
vetulus non ‘veclus,’ vernaculus non ‘vernaclus,’ articulus non ‘ articlus,’
baculus non ‘baclus,’ angulus non ‘anglus,’ jugulus non ‘juglus.’) (On the
curtailment of -ris and -lis, see ch. iv. § 13; the restriction of -er to Mase.,
-ris to Fem. Nom. Sing. is not observed in the older literature ; e.g. Ennius
has somnus acris and. acer hiems ; ef. § 1, p. 371).

§ 5. YO-stems. The -is of the Nom. Sing. of Oscan YO-stems, e.g. Pakis,
Lat. Pacius (-is would be syncopated in Osean, e.g. cevs, Lat. civis), is perhaps
indicated for Latin by the occasional spellings with -es on Greek inscriptions,
e.g. ‘Ekas (Lat. <Aelius) (I. I. S. 928, Ostia, very late), Merpwves (Lat.
Petronius) ; the usual Greek transcription, however, is -1s. (Neither -s nor
-eas is found till the beginning of the first century a.p.; see Eckinger,
Orthographie p. 56). Alts (if we may infer this quantity from dlid, Luer. i.
263, &c.) may have its final syllable shortened by the Law of Breves Brevi-
antes, like cavé, havé, &c. (ch. iii. § 42). In the 8. C. de Bacchanalibus (C. I. L.
i. 196) the consuls’ names are written in what we may suppose to be the
ceremonious form, Marcius, Postumius, while the secretaries’ names have the
is-ending, Claudi, Valeri, Minuci. Ritschl in a paper entitled ‘ De declinatione
quadam latina reconditiore’ (Opusc. iv. 446) has collected a large number of
these proper names with -és or -i Nom. (less certainly -is Gen., -i Dat., -im Acc.)
from Latin inserr. The form alis,quoted from the older writers (e. g.Catull. Ixvi.
28) by the grammarians (see the passages mentioned by Ritschl, ib. p. 452),
may have been specially used in collocations like alis alium (so in the Vulgar
Latin of the Itala; cf. C. I. L. ii. 2633, of 27 a. D.: eique omnes alis alium...
receperunt), alis dlibi .e.g. Sallust fragt. ap. Charis. p. 159. 31 K.), where the
two words formed a single word-group like our ‘one another.’ This byform
of the Nom. (Acc. &¢.) must have led to confusion with I-stems, of which we
have perhaps a trace in the gradual ousting of the second Decl. suffix -drius by
the third Deel. -dris (ch. v. § 4), and in the remark of Caper (112. 2 K.) that
vates (an I-stem, like O. Iv. faith for *wati-, Gaul. oda@res Plur.) was in O. Lat.
vatius.

§ 6. I-stems, The -is of the Nom. Sing. of I-stems is syncopated in the
Umbro-Oscan dialects, e.g. Ose. cevs (Lat. civis), like the -és of the Nom. Sing.
of O-stems, e.g. Umbr. emps (Lat. emptus), Ikuvins (Lat. Iguvinus), Osc. hirz
(Lat. hortus), Bantins (Lat.Bantinus), and the -és of the Nom. Plur. of Consonant-
stems, e. g. Ose. peddeg for *med-dik-és Nom. Plur. of meddix, censtur for *cen-
stor-és, Lat. censdrés. We have seen (ch, iii. § 16) that it is very difficult to
prove a similar treatment of -is in Latin ; for Noms. like pars (for partis), Gen.
partis, may have dropped i not by Syncope, but by the Analogy of Cons.-stems
like réx, Gen. régis, léx, Gen. légis,&e. Like -rés, however, -ris was reduced to
-er in Latin as in Umbro-Oscan, e.g. Lat. dcer for deris, Umbr. pacer for *pac-
376 THE LATIN LANGUAGE. [Chap. VI.

ris, ‘propitious,’ connected with Lat. pax. Ennius, who coined famiil after the
type of the Oscan fumel (§ 4), used debil (Voce. Sing. ?) for debilis (A. 341 M.
debil homo), perhaps after Osc. aidil (Lat. aiditis), &e. Nouns with Nom.
Sing. in -1 (consul, praesul, exul, piigil, vigil, mugil, &c.) are declined as Consonant-
stems in Latin (Gen. Plur. consul-wm, pugil-um, vigil-um, mugil-wm) (see Neue,
Formeni. i2, p. 153), though the line between Nouns in -1 and Adjectives in -lis,
e.g. debilis, is, as might be expected, often passed over. Thus Juvenal, (x. 317)
has mugilis (but Mart. Cap. iii. 294 si ‘mugilis’ esset ... ‘mugilium’ faceret.)
For vigil, pugil we should expect *vigulus, *pugulus, like bibiilus, figiilus (ch.v. § 22),
or with adjectival i (ch. v. § 34) *vigilis, *pugilis, although -ilis has properly
w passive sense, e.g. bibilis, ‘drinkable, easily drunk,’ dgilis, ‘easily moved,’
hdbilis (whence débvilis for *de-hibilis), ‘easily handled’ (ch. v. § 41). Beside -is,
the usual Nom. Sing. of Mase. and Fem. I-stems in Latin, we find occasionally
-és, e.g. cdinés F., the O. Lat. form which had been replaced by canis by the
time of Varro (Z. L. vii. 32). The -2s of ambagés, &c., we have seen to be really
the -és of the Nom. Plur. of I-stems, viz. -éyés (e.g. Lat. trés for *tréyés, O. Ind.
traéyas, Cret. Gk. rpées for *rpeyes, O. Sl. trije, &e.) (§ 40); it must not be
confounded with an early spelling like aidiles for aedilis on a Scipio epitaph
(C. I. L. i. 31, but aidilis on another Scipio epitaph, i. 32), where the e (pro-
nounced @) is merely an expression of the i-sound in an unaccented syllable,
like the third e of Tempestatebus for Tempestdtibus (C. I. L. i. 32) (see ch. iii. § 22).
This use of -2s in the Nom. Sing. of I-stems led to the diversion of other stems,
which took -és in the Nom. Sing., into the I-declension. Thus pilébés, if an
ES-stem by origin like Gk. wepi-wAnO4s (beside mAnG0s) (ch. v. § 74), Should have
made its Gen. Sing. *plébéris, but was led by the analogy of canés, &c. into
taking a Gen. pleb-is, as on the other hand the analogy of stems like rés,
Gen. ré, réi supplied the Gen. plebai, pleb& ; and the tendency of Vulgar Latin
to replace every Nom. Sing. -es by the more familiar -is was perhaps the cause
of the forms cautis, plebis, vatis, tubis, nubis, subolis, vulpis, palumbis, luis, vepris, fumis,
cladis, prolis, censured in Prob. App. pp.198-9 K. For the late byform molis for
miles (an ES-stem, cf. miélés-tus), see Georges, Lex. Wortf. s.v., and for other
examples, Ritschl, Opusc. ii. 654. Beside plebzs, nubés, cawtes, saepes, we have
plebs, nubs (used by Liv. Andronicus, according to Servius ad A. x. 636, and
frequent in the Itala), cds, saeps ; also trabs for older trdbes (Varro, L. L. vii. 33
sic dictum a quibusdam ut una ‘canes,’ una ‘trabes’.. . cujus verbi singu-
laris casus rectus correptus ac facta trabs.) (On these Noms. in -és, see ch. v.
§ 51.) O. Lat. sortis (class. Lat. sors) and the like are discussed in ch. iii.
§ 16; messis of Plaut. Rud. 763 (AP) is changed by editors to messis.

§ 7. S-stems. Masc. and Fem. ES-stems, connected with Neuter ES-stems
(Nom. -és, ch. v. § 71), took -ds M. (class. -or), -és F. in Latin, e.g. hinds, class.
honor (cf. honés-tus), plebés (cf. Gk. 7AH00s). As Adjectives (cf. Gk. wepi-tAnOns
beside wAfO0s, Wevdns beside Pevdos) they show -er in piber (also pubzs, -éris
Gen., impubés, -¢ris Gen.), dé-géner ; -or in con-cilor, bi-corpor (ch. v. § 73). The
-os of honos, ldibos, célos, vipos, &e. was not quite ousted by -or (taken from the
oblique cases, honoris, hondrem, &c., where s came between two vowels, ch. iv.
§ 148) till the Augustan period (for details, see Neue, Formenl. i?. p. 167).
Sallust, according to Servius ad A. i. 253, almost always used the form labos ;
and -os was persistently retained in monosyllables, e.g. /lds, ros.

§ 8. N-stems. We find -en in pecten M. (Gk. x7reis M. for *nxrevs, Gen.
§§ 7-10.] DECLENSION OF NOUNS, ETC. NOM. SING. 377

«revés), flamen, and lien (the @ is attested by Prisc. i. 149. 7 H., Mart. Cap. iii.
279). Sanguis M., which often has its last syllable scanned long by the Latin
poets (always sunguén or sanguis in Lucretius, Munro ad Luer. i. 853), may
represent *sanguins (ch. ii. § 144), a patchwork of the old Nom. *sangui, with
the oblique cases *sanguén-es Gen., &c. (For this declension of some I.-Eur.
neuters, e.g. O. Ind. dksi Nom., akSnds Gen., ‘the eye,’ see ch. v. § 34.) The
scansion sanguis brought with it the treatment of the word as an I-stem,
sanguem Acc., sanguis Gen. &c. (see Georges, Lex. Wortf. s.v.).

§ 9. Diphthong Stems. I.-Eur. *nau-s (O. Ind. nat-s, Gk. vad-s) is Latin
navis ; I.-Eur. *gtoas (O. Ind. gat-s, Gk. Bots) is Lat. bts, but the b- points
to the form being dialectal; the stem réy- (O. Ind. ras Nom., ray-ds Gen.)
probably formed its I.-Eur. Nom. Sing. as *ré-s, and so in Lat., 7s, I-Eur.
*d(i)yéu-s, ‘the sky, day’ (O. Ind. d(i)yats, Gk. Zeds) has in Latin in the sense
of ‘day* the Nom. dié-s, while for the name of the sky-god a compound is
used, Juppiter (the correct spelling, ch. ii. § 130. p. 116) for *Jeu-pater, probably
in the Voc. case, unless Jeu- be the stem (ch. v. § 85; cf. Jani-patri, C. I. L. xi.
5374). The grammarians point out the incongruity of a declension like Jup-
piter Nom., Jovis Gen. (‘as absurd as Phoebus Nom., Apollinis Gen.’ Mar. Sacerd.
473. 1 K.), and tell us that in the old liturgical books the word was declined
Juppiter Nom., Juppitris Gen., &e. (Pompeius 172. 25 ; 187.9 K.), or Jovis Nom.,
Jovis Gen., Prise. i. 229. 10 H.). We have tovos Nom. on an old Praenestine
cista (C. I. L. xiv. 4105), also [Die ]spater (Bull. 1887, p. 232), Diesptr (C. I. L. i. 1500) ;
and in Plautus, &c. Diespiter is not unknown (see Georges, Lex. Wortf. s.v.
Juppiter). Dies is like the Accus. diem. For the Nom. we should expect *dieus
(Gk. Zeds), with @ from eu, which would become in Latin dis (ef. nti-dius-
tertius ?).

§ 10. Nom., Ace, Sing, II. Neut, Neuter O-stems in I.-Eur.
have their Nom. Sing. in -ém, the suflix of the Acc. Sg. Mase. ;
all others use the bare stem, ES-stem Nouns taking however -és,
N-stems -n, NT-stems -nt, &. So in Latin, e.g. jagum (older
jugom). (1.-Eur. *ytigom, O. Ind. yugam, Gk. Coys), mite
(older *miti; see ch. ii. § 37) (cf. O. Ind. Sci, * pure, Gk tp.) ;
nomen with -en for I.-Eur. -n (O. Ind nama; cf. Gk. évoua); ferens
from *ferent (ch. iv. § 105) with -ent for I.-Eur.-nt (O. Ind. bhérat),
génus (older gends) (L. -Eur. *génds, Gk. yévos), cor for *cord (O.
Ind. hfd ; ef, Gk. xfjp for *«xnpd). U-stems have in I.-Eur. -i,
e.g. *médht,‘ mead,’ *swadii, ‘sweet’ (O. Ind. madhii, svadt, Gk.
peds, 780) ; and similarly Latin Neuter Nouns have -u, e. g. pécz,
cornu (U-Stem Adjectives passed into the I-declension in Latin,
e.g. suave; see ch. v. § 47). But there is a doubt with regard
to the quantity of the -v. Most grammarians declare it to be
short, while Priscian (i. 362. 11 H.) controverts their opinion,
and proves by quotations from the poets that it is long. (For
378 THE LATIN LANGUAGE. (Chap. VI.

details, see Neue, Forment. i. p. 345). The existence of byforms
like pecus, cornum, &c, obscures the question, but there seems to
be little doubt that with the Augustan poets cornd, &c. was the
recognized scansion, The long vowel has not yet been satisfac-
torily explained (see Brugmann, Grundr. ii. § 223 for the
various theories, and cf. below, § 45). The Neuter Nom. of
consonant-stem adjectives has been assimilated to the Masc. and
Fem. form, e. g. félia for *felic (cf. allée N., alléw F.), du-plex for
*du-pléc (Umbr. tu-plak) (cf. above, § 1, p. 370).

§ 11. O-stems. We find -wm lost in nihil, nil, from nihilum, 1 compound of
né and hilum (quod grano fabae adhaeret, Paul. Fest. 72. 10 Th.), a loss
which seems due to elision before a vowel (ch. iii. § 52). From phrases like
nihil(um) hoc est on the one hand, and nihilum dicit on the other, the ‘ doublets’
nihil and nihilum would come into use, and no doubt existed for a long time
side by side till the less cumbrous nihil, nil ousted its rival. Similarly non for
ne-oenum, like our ‘nought’ for ‘ne-aught,’ 6 being substituted for % (older oe)
because of the monosyllabic form or the unaccented character of the Con-
junction (but see ch. x. § 18). The Umbro-Oscan neuters have -ém, Osc. saka-
raklam ‘ sacraculum’ (‘a shrine’), dunum ‘dénum,’ Umbr. esonom, a sacrifice.

§ 12. I-stems. Final éis dropped by Syncope (ef. ch. iii. § 36) in Neuters
like facul, an O. Lat. form of facile, e. g. Accius, Trag. 460 R..

érat istuc uirile, ferre aduérsam fortundim facul,

volup, Neut. of a lost Adj. *volupis, e.g. Plaut. Cas. 784 facite nostro animé
uolup. (On the question whether volupest is rightly divided into volupe est or
into volup est, see Georges, Lex. Wortf. s. v.). Similarly the old Nom. lacie, with
the I-stem form (e.g. Plaut. Bacch. 19. 1134, Men. 1089, Mil. 240) became
lact [Plaut. Truc. 903 (?), Varro, L L. v. 104], classical lac (see Georges s. v.).

§ 13. U-stems. The uncertainty of the grammarians of the Empire about
the quantity of -u of fourth Decl. Neuters may be due to the fact that in later
Latin the fourth Decl. was being supplanted by the second Decl. (ch. v.
§ 49). Priscian elsewhere (i. 161. 26 H., &c.) corrects the extraordinary
statement of Charisius (fourth cent.) (22.15 K.) and others, that the -u of
-usin fourth Decl. Noms. masculine was pronounced long, a quantity indicated
neither by poetry nor by the orthography of inscriptions.

§ 14. S-stems. The I.-Eur. Nom. Sg. -ds of Neut. ES-stems, and -és of
Mase. ES-stems, remained distinct in Latin, e.g. O. Lat. opis (C. LL. i. 52),
honos. In course of time Neut. -ds sank to -tis, dpus (see ch. iv. § 20), Mase. -ds
became by Analogy of the oblique cases (hindr-is, hondr-em, &c.) -dr, then -ér,
owing to the difficulty of sounding a long vowel before a final -r (ch. ili. § 49).
There are a few indications of a temporary formation of Neuter Noms. in -or,
e. g. calor (Plaut, Merc. 860 nec calor nec frigus metuo), prior bellum (Claudius
Quadrigarius ap. Prise. i. p. 347. 7H.), bellum Punicum posterior (Cassius Hemina
§§11-17.] DECLENSION OF NOUNS, ETC. GEN. SING. 379

ap. eund.) (ef. § 53 below). Priscian (i. c.), who quotes some instances of
Neut. Comparatives in -or from the old historians, says: vetustissimi etiam
neutrum in or finiebant, et erat eadem terminatio communis trium generum,
thus hinting that the justification of this usage was the Analogy of Adjectives
of one termination for Masc., Fem., Neut., like auddx, felix. And it is possible
that a Neuter in -tir (older -dr like rdbur (robor Acc., Varro, R. R. iii. 7. 9) took
-r for -s in the Nom. from the oblique cases robir-is, robér-i, &c., for Cato (R. R.
xvii. r) uses the form robus, or from an Early Latin Mase. byform [if we may
trust Paul. Fest. 11, 20 Th. ‘robosem’ pro robore (dicebant antiqui)]. But in
spite of these occasional deviations, Latin writers hold with great persistence
to the rule that a Neuter ES-stem hasa Nom. in -is (older -ds), a Mase. ES-stem
in -6r (older -ar, -ds), e.g. ténus N., tenor M., décus N., décor M., frigus N., frigor M.
The -us, not only of Neut. Nouns, but also of Neut. Comparatives, is invariably
short in Plautus and the older poetry (Miiller, Plaut. Pros. p. 55).

§ 15. R-stems. Neuter R-stems show usually -ir (older -ér), apparently
representing I.-Eur. -r, but occasionally -é, e.g. aber (O. Ind. tidhar-, Gk.
ov6ap), in Latin. The obsolete word aser, blood (ef. Cret. Gk. gap, O. Ind.
asr-k, asn-d4s Gen., Lett. asins), is of doubtful spelling [ef. Paul. Fest. 12. 19 Th.
‘assaratum’ apud antiquos dicebatur genus quoddam potionis ex vino et
sanguine temperatum, quod Latini prisci sanguinem ‘assyr’ vocarent; Gl.
Philox. asaer (leg. -er): afya].

§ 16. S in Nom. Sg. Neut. of Adjectives. This, if we may believe the
MSS., is as old as Plautus, e. g. faucinus audax, Aui. 460 (so Ter. Phorm. 233, &c.),
duplex (sc. aurum), Men. 546, sagaa nasum, Cure. 110, and occurs in the ancient
phrase quod bonum faustum felix fortunatumque sit. (For examples, e.g. dives opus,
Ovid, pondus iners, Cic., see Neue, ii*, p. 22). Similar is the extension of the
suffix -tric-, properly fem., to neuter Adjectives, e. g. victricia arma, Virg. A. iii. 54
(though victrix Sg. is not used as neut. till Late Latin). [Does concapit of the
XII Tables (ap. Fest. 556. 27 Th. tignum iunctum aedibus uineaue et concapit
ne soluito) point to an earlier use of the bare stem for the Neuter ?].

§17. Gen. Sing. A-stems took in I.-Eur. -as (e. g. Gk. yopas,
Goth. gibds, Lith. rafikos), and similarly in the Umbro-Oscan
languages, e.g. Umbr. tutas, later ¢otar ‘civitatis, Osc. eituas
‘pecuniae,’ and in O. Lat. e.g. escas, Liv, Andr, - But a rival
formation, of doubtful origin, which appears in the oldest literature
as dé (dissyllabic) ultimately established itself in exclusive use in
the form -ae. In Greek we find Masc. A-stems taking the
O-stem suffix, e.g. Hom. ’Arpeiddo like Alddoo, and it has been
suggested that Lat. -di began in Mase. Nouns such as agricd/a,
advéna, &c. (§ 2), and took its -2 from the Gen. of the second
Decl. Similarly fifth Decl. stems show -d, later -e7, -7,e. g. fide
O. Lat., fide class., 7é, ré2, and ré in the Dramatists, diz, Virg.
A. i, 636, also (like -@s from A-stems), O. Lat. /aciés, diés.
380 THE LATIN LANGUAGE. (Chap. VI.

O-stems, which in Umbro-Oscan show the I-stem suffix, have
in Latin, as in Celtic, a long i-sound, e.g. Lat. xdti, virt, Gaul.
Ate-gnati, O. Ir. eich for *eci, ‘ of a horse,’ which one would have
no difficulty in regarding as the Locative suffix of O-stems, -e/
(§ 37), were it not that it is written -¢ and not -e in the
oldest Latin inscriptions. For IO-stems indeed a Gen. -? would
naturally go with a Nom. -is (§ 5); and it is possible that
the suffix, or at any rate the spelling of the suffix, has been
extended from these over all O-stems.

I-stems have -eis, -ois in various languages (e. g. Goth. anstais
for I.-Eur. -ois) ; and in Umbro-Oscan we have -eis, e.g. Umbr.
ocrer, of the citadel (Lat. deris), a suffix extended to consonant-
stems and even to O-stems, e.g. Ose. carneis, of a part (Lat.
carn-is), sakarakleis, ‘of a shrine’ (Lat. *saerdeili). But in
Latin the consonant-stem suffix has enforced itself on the I-stems
too, e.g. partis like reg-is (cf. partus like Castorus, C.I. L.
1. 197).

U-stems seem similarly to have had -eus, -ous (e.g. Goth.
sunaus, ‘ of a son,’ for. I. Eur. -ous), and so in Umbro-Oscan, e. g.
Umbr. trifor ‘tribus,’ Ose. castrovs ‘ fundi,’ Lat. ménis, all prob-
ably with -ous from I. Eur. -eus (ch. iv. § 35). Whether the Gen.
form of ddmus affected by Augustus, viz. domos, points to the
coexistence in Latin of I.-Eur. -ous is uncertain (cf. ch. iv. § 41,
on Lat. 6 for I.-Eur. ou). A common formation, perhaps the
usual one in the careless talk of every-day life, in which the fourth
Decl. seems to have been greatly merged in the second (ch. v.
§ 49), was -7, the O-stem genitive. This is the normal genitive in
the Dramatists of the Republic; and even Quintilian in the first
cent. a. D. declares it impossible to decide whether senati or
senatis is the proper Gen. of séxdtus. Occasionally the
Dramatists have -ais, the suffix proper to i-stems like sis, socrits,
as -*zis, -is to I-stems like vis (O. Lat. Gen. vis).

Consonant-stems show -és in some languages (e. g. O. SI]. din-e,
‘ofa day,’ with -e from -és), -3s in others (e.g. Gk. zoiev-os).
Latin -is, on old inscriptions -es, shows the former suffix, while
the -ws occasionally written on inscriptions seems to be a
relic of the latter. To make -is a weakening of earlier -us
(-os) is an unlikely theory, seeing that -ws (-os) of the Nom.
§ 18.] DECLENSION OF NOUNS, ETC. GEN. SING. 381

Sg. of Neuter ES-stems, &c., e.g. genus, opus, was not weakened
to -es, -ds.

The A-stem Gen. -aes (pronounced -és with the open E-sound,
ch. ii. § 32) is probably a feature of the Italian-Greek patois, for
it is practically confined to epitaphs of the uneducated classes
(from the last century of the Republic). It is merely an ex-
pression in Roman letters of the Greek Genitive-ending -ns
(with open E). Hedonét (C.1.L xi. 3316 Forum Clodi), may
be an example of a Greek name in -n taking a Genitive after
the analogy of Latin fifth Decl. stems (or for Hedonii ?)

§ 18. A-stems. (1) In -as. This form is proper to the Saturnian and
earliest Epic poetry. Thus escas, Ménétas, Latinas are quoted by Priscian
(i. p. 198 H.) from Livius Andronicus, Terras and fortinas from Naevius, vias
from Ennius. (For other passages of the grammarians treating of this
Genitive, see Neue, Formeni. i. p. 5.) Servius favours the reading auras for
aurae in Virg. A. xi. 801, and in his note on the passage mentions that some
interpreted custédias as a Gen. Sg. in a passage of Sallust: castella custodias
thensaurorum in deditionem acciperentur. A relic of the old usage survived
in legal phraseology, so conservative always of old words and ceremonies, in
the terms piter fimilias, mater familias, filius (-a) familias. But this form is
unknown to the conversational language of the Dramatists [Alctiménas in the
Argument (post Plautine) of the Amphitruo, 1. 1, is an imitation of the
antique]|, and must have been in their time out of use. (A contrary view is
stated in Studem. Stud. ii. p. at.)

(2) In -a, class. -ae. Dissyllabic -@ is not infrequent in Plautus, and is
perhaps found in Terence (Rhein. Mus. 1893, p. 305), while in Lucilius it
is allowed in hexameters only (e. g. Tirésiai, v. 43 M.), not in the dramatic
metres. Lucretius is especially fond of this early form; and it is used
occasionally by Cicero, Virgil, and other Epic writers (for instances, see
Neue, i? p. 12). To Martial it seems typical of the uncouth early Latin

poetry (xi. go. 5):
attonitusque legis ‘terrai frugiferai,’
Accius et quicquid Pacuviusque vomunt.

The rarity of the elision of the final -¢ of -a, as of -a% [Plaut. Bacch. 307
Diana(i), Ephesiae; Pers. 409 pecunia(i) accipiter, are more or less doubtful
instances], may be an indication that the ending had already at the begin-
ning of the second cent. B.c. ceased to be quite two distinct syllables, though.
it is scanned as a spondee. The change to -ae would probably begin by the
shortening of the @ before the following vowel, so that ai (classical -ae)
would differ from -d@i in Plautus very much as his pronunciation Chius (Adj.
pius from Chius, pius (ch. ii, § 143). Though written -ai on early inscriptions
(for example on the old Praenestine vases and mirrors) it need not have been
pronounced otherwise than the diphthong ai of aidilis, &e. (ch. iv. § 29), pre-
cisely as the archaistic spelling of a later.metrical inscription (C. I. L. vi. 555)
offers as a spondee ripai, Another inscription of no early date (i. 1202),
382 THE LATIN LANGUAGE, [Chap. VI.

seems to show -ai with -i elided: non aevo exsacto vitai es traditus morti),
[For passages of the grammarians referring to this Genitive in -ai, see Neue,
i. p. 9, e.g. Quint. i. 7. 18 unde ‘pictai vestis’ et ‘aquai’ Vergilius aman-
tissimus vetustatis carminibus inseruit. Servius on A. vii. 464 says that Virgil
ended the line with ‘aquae amnis’ (ley. vis?), which was changed by Tucca
and Varius to aquat].

A list of ‘Greek’ genitives in -aes from plebeian epitaphs is given by Neue,
i, p. 13. On two bricks of the same year (123 a.p.) from the manufactory of
Flavia Procula we have (C. I. L. xv. i. 1157-8) Flaviaes Proclaes and Flaviae Procule.
[Cf. no. 1425 Seiaes Isauricae (123-141 A.D.), but usually Seiae Isauricae. }

Some would connect Lat. -di, -we with O. Ind. Gen. -ayas, Dat. -ayai (used in
the Brahmanas for the Gen.) of A-stems, the yo f which forms is of doubtful
origin. The derivation of Lat. -di from an earlier *-ais is impossible. The
supposed ‘ Prosepnais ’ofa Praenestine mirror (C. I. L. i. 57) is really Prosepnat,
and is a Dative, not a Genitive (see Rhein. Mus. 1887, p. 486).

§ 19. Fifth Decl. Stems. The Genitive of these stems is discussed by Aulus
Gellius in the fourteenth chapter of Book ix of the Noctes Atticae. In old copies
(aliquot veteribus libris) of the History of Claudius Quadrigarius he found
facies Gen., sometimes with facit added in the margin: meminimus enim in
Tiburti bibliotheca invenire nos in eodem Claudii libro scriptum utrumque
‘facies’ et ‘facii” Sed ‘facies’ in ordinem (in the text) scriptum fuit, et
contra (in the margin) per i geminum ‘ facii.’? He quotes dies from Ennius
(Ann. 433 M.) and from Cicero, pro Sest. xii. 28: equites vero daturos illius dies
poenas (where our MSS. read diei, but where Gellius found dies in the older
copies: inpensa opera conquisitis veteribus libris plusculis), and mentions
a report that in a ‘liber idiographus’ of Virgil the line (@. i. 208) was
written :

Libra dies somnique pares ubi fecerit horas’.

He adds examples of -ii (Nom. -iés), -i (Nom. -2s) from early literature, fami
from Cato and Lucilius, pernicii from Sisenna and Cicero, progenit from Pacu-
vius, acii and specii from Matius, luxurii from C. Gracchus, and supports the
reading dii in Virgil, A. i. 636: munera laetitiamque dii (quod inperitiores
‘dei’ legunt, ab insolentia scilicet vocis istius abhorrentes). Finally he
summons the authority of the great Dictator for die, specie, &e. : sed C. Caesar
in libro de Analogia secundo ‘hujus die’ et ‘hujus specie’ dicendum putat,
and supports this form from an old MS. of Sallust: ego quoque in Jugurtha
Sallustii summae fidei et reverendae vetustatis libro ‘die’ casu patrio scriptum
inveni. (The passage is Jug. xevii. 3, where two of our MSS. have die, the rest
diei.) (For the remarks of other grammarians on this point, see Neue, i’. p. 375.)
From his account we gather that forms like diz, specié, were grammarians’
coinages designed to restore the actual forms dii, specii to the proper e-type
of stem. At the same time the tendency to Dissimilation, which in the
middle of a word turned ii to iz in dlienus, &., may have been to some extent
operative in certain collocations of these words, e. g. dii-festi, dii-natalis, &c.
[see ch. iii. § 12a (9)]. The spelling diei, speciei, &c., in early literature and
inscriptions, may often have represented dii, specii, the -ei being diphthongal as

 

1 The form dies would, however, produce cacophony with its repetition of
-s in the Cicero and Virgil passages.
§§ 19, 20.] DECLENSION OF NOUNS, ETC. GEN. SING.

383
in the old spellings deico, feido (ch. iv. § 34). Gellius mentions (i. c.) the theory
of some grammarians that die, specie were Ablative forms used as Genitives,
and modern philologists have made them Locatives, like die crastini, &e. (§ 37).

The rule of the grammarians of the Empire is that in the approved Gen.
form, dissyllabic -ei, the e is short after a consonant, long after a vowel,
e. g. fidéi, dizi. In Plautus and Terence we rarely find ré (e.g. Plaut. Men,
494), but usually monosyllabic ré, hardly ever the ceremonious form rei,
ve. g. Mil. (prol.) 103 magnéi réi publicai gratia (post-Plautine ?); similarly
Jidei is always dissyllabic in Terence and usually in Plautus (but twice Side) ;
spei is never a dissyllable. (Seyffert, Stud. Pl. p. 25.) (Compare the usage of
the dramatists with regard to the Pronoun Dat. Sg. @, %, and &, ch, vii. § 19.)
The normal shortening of 2 in hiatus (ch. ii. § 143) would be hindered when
¢ preceded, e. g. diei. Of the elision of the final -¢ of dissyllabic -et examples
(more or less doubtful) are: Plaut. Aul. 68 Malaé rei euenisse, Poen. 479 Quoi
réi? Ad fundas uiscus ne adhaerésceret, &c. (Other examples of all these
forms of the Gen. of fifth Decl. stems, v. g. rdbiés in Lucr. iv. 1083:

quodcumque est, rabies unde illaec germina surgunt,

die in Varro, Ep. ad Fufium: meridiem die natalis, fidé in Hor. C. iii. 7. 4 con-
stantis juvenem fide, see in Neue, 7. ¢.)

§ 20. O-stems and IO-stems. The grammarians tell us that Vdléri, Vergili,
&c. were accented on the second syllable, that is to say they were accented as
if they were contractions of Valerii, &c. (ch. iii. § 10. 4), though whether this
accentuation was due to tradition or to grammarians’ rules is open to question.
Lucilius’ rule for the use of the single symbol i for a Singular case, v. g. pueri
Gen. Sg., and of the double symbol ¢ for a Plural, e. g. puerei Nom. Pl. has
been mentioned in ch. i. § 9.

The earliest form of the O-stem Gen. Sg. suffix is -i, e.g. Saelurni pocolom
C.I.L. i. 48; from the time of Lucilius to the end of the Republic -e, which
had come to be an expression of the long i-sound (ch. i. § 9) is also found,
e.g. populi Romanei on the Lex Agraria of 111 B.c. In Faliscan we have -oi in
the one instance of the Gen. Sg. of an O-stem, Zextoi ‘ Sexti’ on a rude inser.
on a tile (Zvet. I. I. I. 73) ; 10-stems (with Nom. in -io or -es) have -i, e.g.
Acarcelini (ib. 62), Caui (ib. 49) (also -es?). On the use in the Gen. Sg. of
I10-stems of -i (the older form) and -ii (Propertius, Ovid, &c.), see Neue,
Formenl. i?. pp. 85-94. The passages which he quotes from the grammarians
make it clear that -ii was a grammarian’s restoration! on the Analogy of

 

1 The suggestion of -ii seems to
have been made as early as Lucilius,
who proposed to distinguish in this
way the Gen. of Numerius from the
Gen. of numerus. The phrase servandi
numeri should, he said, mean ‘ for the
purpose of keeping tune’ (inc. 66 M.):

‘seruandi numeri ’:—numerum ut
seruemus modumque.

This use of the Gen. of the Gerundive

to indicate purpose (cf. Aegyptum
proficiscitur cognoscendae antiqui-
tatis, Tac.) is a genuine Latin con-
struction (see Weisweiler, Der jinale
Gen. Gerund. 1890), and is found in
Umbrian, e.g. esono- ... ocrer pihaner
‘sacrificium arcis piandae’ (Tab.
fg. vi. A. 18), verfale pufe arsfertur
trebeit ocrer peihaner ‘templum (?) ubi
flamen versatur arcis piandae ’ (ib.
vi, A. 8).
ZX

384 THE LATIN LANGUAGE. [Chap. VI.

O-stems, -i the actual historical development. Adjective I1O-stems have -ii,
e.g. patrii sermonis, Lucr. [See Neue, ii*. p. 44; jfluvit of Virg, A. iii, 7o2
(Gela fluvii cognomine dicta) has been explained as an Adj.]

§ 21. U-stems. On the S. C. de Bacchanalibus of 186 B. 0. (C. I. L. i. 196)
we have séndtuos, but in inscriptions of the latter part of the second cent. B.c.~
senati (i. 199, of 117 B.u.3 i, 200, of 111 B.C. ; i. 547, Of 141 or 116 B.C.), (cf. ldci
i. 584, of 82-79 B.c., and see Mommsen’s note) ; and in the Comedians and
Tragedians -i is the usual form (cf. Prise. i. 257. 18 H.), occasionally -uis
(dissyll.), e.g. Ter. Heaut. 287 eius anuis causa. Gellius (iv. 16. r) tells us that
Varro and Nigidius (first cent. B.c.) approved -wis, e. g. senatuis, dimuis, a form
which sticklers for Analogy defended by the Dat. Sg. senatui, since patri, diici,
caedi had as Genitives patris, ducis, caedis. According to Mar. Victorinus
(9. 4 K.), Augustus used domos for domus Gen. (divus Augustus genetivo casu
hujus ‘domos’ meae per o, non ut nos per u litteram scripsit. Cf. Suet.
Aug. 87). (For other passages of the grammarians dealing with the Genitive
of u-stems, see Neue, i% p. 352). The -wos of senatuos must be the i-stem Gen.
with the I.-Eur. Gen. suffix -63 (see below), as the -uis of anuis is the t-stem
Gen. with the I.-Eur. Gen. suffix -és. But the -is of senatis, dnis can hardly
be derived by the ordinary processes of phonetic change from either ; for -uos,
-uis would naturally become -vis (-iis), -vis (ef. miluos, Plaut., milvos in class.
Lat., mort(v)iis in Late Lat., ch. iv. § 71).

In Faliscan we have (Zvet. I. I.I. 70): de zenatuo sententiad, where the
final -s of zenatuos (Lat. senatuos) has been dropped before the following
initial s-. Oscan senateis shows the same Gen. suffix as the O-stems.

§ 22. Consonant-stems. The frequency of the Gen. -us in S. Italian
inscriptions, e.g. Vénérus, C. I. L. i. 565 (Capua, 108 B.c.), Eph. Epigr. viii. 460
(Capua, 108 B.c.), C. I. L. i. 1183 (Casinum), i. 1495 (on a tile, now at Naples),
‘Cerérus, i. 566 (Capua, 106 B.c.), i. 568 (Capua, 104 B.c.), Hénorus on the Lex
Pariet. Fac. i. 577 (Puteoli, 105 B.¢., a copy), may be due to the influence of
the Greek Gen. in -os [so régus (with sécittatis) on a bilingual Greek and Latin
inser. of 81 B.c., Not. Scav. 1887, p. 110], but this Latin suffix cannot have
been merely a usage of Italian-Greek patois, like -aes in Gen. of A-stems. It
is found on so early inscriptions as the S. C. de Bacch. (i. 196) with nominus,
and the old Praenestine cippus (xiv. 2892) with Sdlutus, and on various
official inscriptions, v.g. the Epistula ad Tiburtes (i. 201, of ¢. 100 B.c.) with
Kastorus, the Lex Agraria of 111 B.c. (i. 200) with himinus, praevaricationus, the
Lex Bantina (i. 197, of 133-118 B. ¢.) with Castorus and even partus (an I-stem),
and may be the correct reading in Lucil. ix. 28 M.: foris subteminus panust.
(Other examples in Neue, i’. p. 191, such as the soldier’s message of defiance
cut on a glans used at the siege of Perusia: L. Antoni ecalve, peristi C.
Caesarus victoria, C. I. L. i. 685).

The Genitive in -es on old inscriptions may sometimes be dialectal with -és
for -eis (the I-stem Gen., extended in the Umbro-Oscan languages to Con-
sonant-stems), e.g. Umbr. matrer, nomner (cf. Osc. maatreis ‘matris’), but is
more naturally regarded as -és, the older spelling of classical -%s (as early as
¢. 180 B. u., fldminis, C.I.L. i. 33) (ef. ch. iii. § 18). Examples are C.1.L. i. 49
(Orte) Salutes pocolom, i. 187 (Praeneste) Apolones dedert, i. 811 (Rome ?) [Clereres.

On a possible byform -s, of the Gen. Sg. suffix, seen in the O. Lat. Adverb
nox, ‘by night’ (Gk. vu«rds), see ch. ix. § 3.
§§ 21-23.) DECLENSION OF NOUNS, ETC. DAT. SING. 385

§ 23, Dat. Sing, The Dat. Sing. of A-stems had in I-Eur.
the long diphthong -ai (e.g. Gk. xpd). In Latin and the
Umbro-Oscan languages we find the ordinary diphthong -ai
(Ose. -ai, Umbr. -e, Lat. -ae, older -ai), Whether in O. Lat. -a
existed beside -aé is not quite certain. If it did, we must suppose
-¢ and -ai to have been doublets, both sprung from original
-ili, just as atgue and ac (for *ate) were doublets, the one re-
presenting the sound which at with the enclitie gue took before
a word beginning with a vowel, the other its sound before a
consonant (ch. 11. § 136; ch. iv. § 45).

O-stems had similarly in I.-Eur. the long diphthong -di (e. g.
Gk. inz@). In the most ancient Latin inscriptions we have -oi
with the quantity of the o unascertainable (Umbro-Oscan seem
to have had the ordinary diphthong -oi, in Oscan -tii, in Umbrian
-e), but in all other inscriptions and in classical Latin, -d. This
-oi and -6 are generally regarded as doublets, like -d and -ai of
A-stems, the long vowel having survived the struggle for
existence in the one declension, the diphthong in the other.

As regards Fifth Decl. Stems, we have seen (ch. iv. § 47) that
the doublets -éi (the long diphthong) and -é probably existed in
I.-Eur. times. If Gellius is right in saying that facie, &c. were
regarded as the correct forms by the older writers, this may indicate
that the latter gained the day in Latin. The alternative Dative
which he mentions, /acii, may then be the Genitive form (facit
for older faciéi from facié?), which was adapted to the dative use
on the Analogy of third Decl. datives in -2 (just as the classical
faciét seems to be a Genitive form), though some prefer to regard
it as a relic of the I.-Eur. ‘ doublet ’-suffix é-.

I-stems have in Latin -2, older -¢2, probably (like the Genitive
in -ts, older -es, also -us, § 22) a loan from Consonant-stems,
and so originally -ai, The Umbro-Oscan termination was -ei
(Ose. -ei, Umbr. -e), as in Consonant-stems.

U-stems have -w2 in Latin, which is equally traceable to
either of the I.-Eur. suffixes, -éwai and -wai (e.g. O. Ind.
sundvé, ‘ to a son, and Siév ,‘toachild.’ The occasional Latin
forms in -#@ are (cf. Umbrian ¢rifo ‘ tribui’) really Locatives
(§ 37), according to some, Instrumentals (§ 36).

Consonant-stems had -ai (-ai? ch. iv. § 3) in I.-Eur. (e. g. O. Ind.

ce
386 THE LATIN LANGUAGE. (Chap. VI.

San-é, ‘ to a dog, dé-man-é Inf., Gk. 86-wev-ae Inf., Lat. légimini
Inf. used as Imper. (?), ch. viii. § 81), in Latin -2 from older -e7
(sometimes written -e), the diphthong -ai being weakened
first to -e2, then to -? in the unaccented syllable, as ai of oc-cardo
to ei, occeido, and 2, occido (ch. iii. § 18). In Umbro-Ose. this
weakening does not seem to have taken place, so that their -ei
(Ose. -ef, e. g. medikei ‘ meddici,’ chief magistrate, [A]meAAouvn: ;
Umbr. -e, e.g. patre, womne) can hardly represent I.-Eur. -ai.

§ 24, A-stems. Dissyllabic -az is not found in the Dat. of A-stems, but
only in the Gen. (terrai frugiferai, Enn. A. 605 M. is, like Virgil’s aulai medio,
with which Charisius couples it, a Genitive, so correct ‘dativo’ in Char.
19. 1 K.; widi sternenda@t, Lucil. xi. 5 M., even if the reading is right (MSS. rim
sternenda et), is anything but a certain example (see L. Mueller’s note), so that
Priscian’s remark that the Nom. and Voc. Plur. ending of the first declension
did not admit of ‘divisio,’ as the Gen. and Dat. Sg. did, cannot be quite
accurate (Prise. i. p. 291. 17 H. nominativus et vocativus pluralis primae
declinationis similis est genetivo et dativo singulari. Nam in ae diph-
thongum profertur, ut ‘hi’ et ‘o poetae’; sed in his non potest divisio fieri,
sicut in illis). Gellius (xiii. 26. 4) tells us that Nigidius (first cent. B. c.)
approved -ai (presumably the diphthong) in the Gen., -ae in the Dative. (On
Lucilius’ practice see L. Mueller’s note on Lucil. ix.6.) The -e found on some
inscriptions is dialectal (cf. Umbr. -e) and rustic, e.g. Diane (C. I. L. i. 168,
Pisaurum), Fortune (i. 64, Tusculum), Victorie (i. 183, Marsi). Of the ‘ Datives
in -«,’ only found on very old inscriptions, most of the apparent examples
come from Pisaurum (C. I. L. i. 167-180), where -e (Diane just quoted) was the
Dat. suffix of A-stems, and may be Genitives in -ds with omission of the final
s (ef. Nom. Pl. matrona Pisaurese for matronas Pisaurenses, i. 173 ; 80 Gen. Sg. Coira
pocolo, Eph. Epigr. i. 6), or else a mere dialectal variety, which would prove
nothing for the Latin dative. Others, viz. Fortuna (i. 1133, Praeneste) ; Fortuna
(Bull. 1885, p. 62, Signia) ; Fortuna Diouo fileia primogenia (xiv. 2863, Prae-
neste) ; Diana (xiv. 4182 a4 and 4184 a, Nemi) are open to similar doubts. The
strongest instances are: [Me]nerua dono d... (Not. Scav. 1887, p. 179, Rome) ;
Iunonei Loucina (C. I. L. i. 189, loc. ine.) ; Iunone Loucina Tuscolana sacra and
(Palle [Tusclolana sacra (i. 1200-1, Capua) [cf. Faliscan Menerua gsacru (Zvet.
I.1. I. 70)}.

Examples of -ai are Dianai donum dedit (C. I. L. xiv. 4270, beg. of second cent.
B.C.), Meneruai donom port- (C. I. L. i. 191), [Iunon]e Loucinai (i. 813). We have
-ai even on inscrr. of the Emperor Claudius, e.g. Antoniai Augustai matri
(Orelli 650).

§ 25. Fifth Decl. Stems. Gellius (ix. 14): in casu autem dandi qui purissime
locuti sunt non ‘ faciei,’ uti nune dicitur, sed ‘facie’ dixerunt. He then quotes
two examples of facie from Lucilius (vii. 9 and vii. 7 M.), and adds; sunt tamen
non pauci, qui utrobique ‘facii’ legant. In Plautus the treatment of the

. Dative Sg. of these stems is the same as that of the Genitive (see § 19). In
Umbr. ri ‘rei’ the i may correspond to Lat. -é, as in pru-sikurent ‘ pronun-
tiaverint’ with the 6-grade of root seen in Lat. sedi, &c. (ch. viii. § 39).
§§ 24-29.] DECLENSION OF NOUNS, ETC. ACC. SING. 387

§ 26. O-stems. The suffix -of (mentioned by Mar. Victorinus 17. 20 K.:
‘populoi Romanoi’ pro populo Romano solitos priores scribere) is found on
the very ancient Praenestine fibula (C. I. L. xiv. 4123): Manios med fefaked
Numasioi (=Manius me fecit Numerio), but -o on the Dvenos inscription
(Zvet. I.I. I. 285), if the words: die noine med mano statod, be rightly read and
interpreted ‘die noni me Mano stato’ (ef. Numisio Martio donom dedit mere-
tod, Not. Scar. 1890, p. 10; Lebro ‘Libero’ C. I. L. i. 174, from Pisaurum. The
Vestine dialect had -o, e.g. Herclo Iowio (Zvet. I. I. I. 11).

§ 27. U-stems. Senatuei (C. I. L. i. 201, of c 100 B.c.). Gellius (iv. 16)
informs us that Varro and Nigidius used senatui, domui, fluctui, &c. in the
Dative, and senatuis, domuis, fluctuis in the Genitive, but gives examples of -z
from Lucilius (iv. 8 M.; dnu, iv. 9, ef. vii. 21) and Virgil, and clenches them
with the authority of Caesar: C. etiam Caesar, gravis auctor linguae Latinae,

- in libris Analogicis omnia istiusmodi sine i littera dicenda censet. In
Plautus -ui is the usual form, e. g. quaestui habere, extersui, usui esse, and with
the force of a second supine, Bacch. 62 quia istaec lepida sunt memoratui ;
but -u is also found, e.g. Rud. 294 sunt nobis quaestu et cultu.

§ 28. Consonant-stems. The so-called ‘Datives in -é@’ in Latin poets
(ef. Servius ad A. x. 653 conjuncta crepidine saxi, A. x. 361 haeret pede pes, and
L. Mueller’s note on Ennius, Ann. 395) are really Locatives or Instrumentals
(see Neue, i?.p.195). The -e which we find (along with -e7) on old inscriptions
is -2, possibly in some cases (e. g. C.I.L. i. 1170, Iowe, Marsic), a dialectal form
(ef. Umbr. patre), but certainly in others a mere graphic variety of -ei, later
-Z, just as the e of ploirwme on the Scipio epitaph (C. I. L. i. 32 hone oino
ploirume cosentiont) represents no different sound from the usual -e7, later -7
of the Nom. Pl. of O-stems. Instances of Dat. -ei and -e are: Hercolet Gi. 1503,
Rome, 217 B.c.); Martei (i. 531, Rome, 211 B.C.) ; Hercole (Ann. Epigr. 1890,
no. 84, Rome) ; Hercole (C. I. L. xiv. 2891-2, Praeneste). We have the three
spellings of the suffix side by side on a freedman’s inscription from the
Roman district (i. 1110): Iunone Seispitei Matri, and the two older in i. 638:
[D}iouet Uictore (Rome, c. 180 B.c.), (Quintilian (i. 4. 17) mentions Didve Uictore
as an old form], and in xi. 4766 Ioue... Iouei (Spoletium in Umbria). (For
other instances of these old spellings, see Index to C. I. L. i.)

§ 29. Acc. Sing. To form the Acc. Sing. Masc. and Fem.
the suffix -m was added, which in the case of Consonant-stems
took the form -m (e.g. O. Ind. matér-am, Gk. pnrép-a, Lat.
matr-em). "The Acc. Neut. was the same as the Nom. (§ 10).

Thus A-stems had -am (e.g. O. Ind. éévim, Gk. ydpav), which
in Latin would become -dm (ch. 1i1. § 49), equim. The long vowel
is indicated by Osc. paam (Lat. guam), the Acc. Sg. Fem. of the
Relative Pronoun, but the usual spelling is e. g. Osc. tovtam, Umbr.
totam, the community. O-stems had -ém, in O. Lat. .om, class.
-um (ch. iv. § 20); IO-stems, -iom, -iwm, perhaps also in the

ce2
388 THE LATIN LANGUAGE. [Chap. VI.

‘familiar’ declension (§ 5), -im; I-stems, -im, which is found in
the older literature and in many examples, ¢uwrrim, &c., in the

classical period, though -em, the Consonant-stem ending, has

usually supplanted it ; U-stems, -tim ; I-stems, -im (also -iym, e.g.
O. Ind. dhiyam, ‘ thought’), which in Latin would become -im ;
U-stems, -uwm, Lat. -wem, e.g. suem, also -im, Lat, -#m, e.g.
socrum. Consonant-stems take in Umbro-Oscan -om, the O-stem
Accusative, e.g. Osc. medicatin-om (Lat. *meddicationem from
med-div, a magistrate). Of Latin ES-stems some are regular,
e.g. dégénérem from *degenes-em; others follow the analogy of
E-stems, as in the Gen. and other cases, e.g. plébem (ch. v. § 51).

§ 30. The endings -im and -em. We can hardly say that -ém arose from
-im by ordinary phonetic change, seeing that final -im remains in so many
words, e. g. Adverbs in -im like dlim, as well as Accusatives like clavim. The
change is rather due to that intermixture of I- and Consonant-stems which
was the despair of grammarians as early as Varro (L. L. viii. 66), and which
led to the substitution of -é for -2 in the ‘ Ablative’ (see below, § 33). A list
of Accusatives in -im, with references to the Latin grammarians who discuss
this question, is given by Neue, i*. p. 196, to which may be added piscim on an
old Praenestine cista (Mél. Arch. 1890, p. 303), and the instances from the
Ambrosian Palimpsest of Plautus given in Studemund’s Index, e. g. imbrim,
Pseud. 102. Sometimes the use of -im indicates an I-stem, e.g. vim, some-
times a Greek loanword, e.g. turrim (?); it is retained in Accusatives used
adverbially, e.g. partim (ch. ix. § 4). [Cluudi, &e. (C. I. L. iv. Ind.), if for Clau-
di(uym, may be dialectal. On Gk. Angew for”Ammov Nom. Sg. (1.1.8. 1411), &e.,
see Eckinger, Orth. p. 56.]

§ 81. Voc. Sing. In the Plural and Dual, and in the Neut.
Sing., the Nom. form was used also for the Voc. in I.-Eur., and
even in the Masc., Fem. Sing. the same thing is often found (e.g.
a : & = . a
in the Veda, Vayav Indraé ca, ‘OV ayu and Indra’; in Hom., Zed
marep,... Heédids re; in Plautus meus ocellus,. .. mt anime). The
rule however was that in the Singular the bare stem was used
(accented on the first syllable, e.g. O. Ind. pitar, Gk. aarep,
unlike the Nom., O. Ind. pita, Gk. aarjp). A-stems had
a short A-vowel (Gk. décmora, &c.), O-stems -é (O. Ind. viki,
Gk. Avce, Lat. déipé) and so on. In Latin I-, U- and N-stems
substitute the Nom. form for the Voe., e.g. civis, manus, himo
(contrast Gk. é1, mhxv, kvov); and the same was done in R-
stems even earlier than the shortening of long vowels before
final -r in the second cent. B.c. removed the distinction between
I.-Eur, -ér and -ér, -dr and -dr, to judge from scansions in
§§ 30-32.) DECLENSION OF NOUNS, ETC. VOC. SING. 389

Plautus like Mere.:800 Uxér, heus uxor; probably also in
S-stems. Final 4 was shortened in Nouns of the first declension
still earlier, before the literary period (ch. iii. § 43), so that it is
impossible to say whether egud Voe. is the Nom. form (originally
equd), or is a special Voc. form. It cannot be the I.-Eur.
Vocative, if final I.-Eur. -% became -é in Latin (ch. iii. § 37),
so that I.-Eur. *ekwi would sink to egué, and would be indis-
tinguishable from the Voc. of O-stems, I.-Eur. *ekwé, Voc. of
*ekwos. The levelling process to which the other Latin
Vocatives have submitted makes it likely that the Nom. was
used for the Voce. in the A-declension too; and that the Umbrian
language, in which a distinction between the Nom. (in -o, § 2), and
the Voe. (only in -a, e.g. Serfia), of A-stems is clearly apparent,
has retained the I.-Eur. -% of the Voe. (cf. ch. iii. § 18, p. 191).
Latin [O-stems show in the Voce. -é, e.g. Véléri; but this form
is hardly found except in proper names [which, as we saw before,
(§ 5), admitted the ‘familiar’ declension, -2s Nom., -dm Acc. (?), &c.],
and the word of everyday life, fai, so that the -2 need not be
a contraction of an older -ie, but may be the Voc. byform
corresponding to the Nom. byform -is (cf. Lith. gaid¥ Voc. from
gaidys Nom.,*a cock’; see § 5). RO-stems which took -er in
the Nom. retain this in the Voc. too, though pwére, and not
puer, seems to be the form always used by Plautus.

§ 32. Other examples. Jii-piter (better Juppiter, ch. ii. § 130, p. 116) cor-
responds exactly with Gk. Zed mdrep, and might be a Vocative used as
a Nominative, just as Homer’s vepeAnyepéré Zevs, pntiérd Zevs, have heen
explained as obsolete Vocative forms, preserved only in certain liturgies, and
treated by the poet as Nominatives through a similar mistake to ours in using
‘cherubim’ as a Singular. But Ji-piter may also be a correctly formed
Nominative with the stem *Dyeu-, Lat. Jov- (cf. O. Lat. Jovis Nom.) as the
first part of the Compound (ef. ji-glans, ch. v. § 85). The same double expla-
nation is possible for Dite pater Voc. (C. I. L. i. 818). Dite Voe. (Eph. Epigr. viii.
529) reminds us of Gk. équ, &e.

The Vocative formation for 10-stems is discussed by Gellius (Noct. Att. xiv. 5),
who describes a battle royal waged in his presence between two grammarians
about the proper Voce. of egregius, without satisfactory result (non arbitratus
ego operae pretium esse, eadem istaec diutius audire, clamantes compugnan-
tesque illos reliqui). Priscian (i. p. 301. 19 H.) says that the early writers
used -ie as well as -i in the Voce. of proper names: haec tamen eadem etiam
in e proferebant antiquissimi, ‘O Virgilie,’ ‘Mercurie’ dicentes, though all
that he quotes is a couple of instances of Laertie, which is a Greek word, and
as much an Adjective asa Noun. He adds that the classical form -? must be
390 THE LATIN LANGUAGE. (Chap. VI.

a contraction of this older -ie (as Arpinds, &c., of older Arpindtis, &c.) because
Voces. in -2 were accented on the paenultima, e.g. Valéri Voc. (like Valéri
Gen.). Gellius (xiii. 26) tells us that Nigidius Figulus (first cent. 8. c.) wished
to distinguish Valeri Voc. from Valéri Gen., but says that in his time both Voe.
and Gen. of 1O-stems were invariably accented on the paenultima (ef. ch. iii.
§ 10. 4). (For other passages of the grammarians, see Neue, Formenl. i?. p. 82.)
Priscian elsewhere (i. p. 305. 9 H.) quotes jilie (apparently the more ceremonious
form) from Livius Andronicus :
pater néster, Saturni filie,

but almost the only instance of the Voc. Sg. of a masc. IO-stem to be found
in Plautus and the older writers is volttri (for which some would read voiture),
‘you vulture’ (Capt. 844). Publi Cornelt occurs on a Scipio epitaph of c. 180 B.c.
(AIL. i, 33). Adjective IO-stems take at all periods -ie, though there is
evidently a reluctance on the part of good writers tu use these forms (see
Neue, Formeni. ii®, p. 42).

§ 83. Abl. Sing. The Ablative suffix, ending in -d, appears to
have been used in J.-Eur. only in O-stems, which formed their
Abl. Sg. in -dd and -éd (the latter suffix being reserved in the
Italic languages for Adverbs, ch. ix. § 1), [O. Ind. yugat, ‘ from
a yoke, O. Lat. jugod, Falisc. rected, class. Lat. jad, recté, final
d being dropped after a long vowel at the close of the third cent.
B.C. (ch. ii, § 137)]. In other stems the Genitive ending -és or
-ds was used (O. Ind. navas, ‘from a ship, Hom. Gk. yn(F)és) ;
and in the Greek language this Ablatival use of the Genitive
was extended to O-stems too. In the Italic languages on the
other hand the A-, ., U- and E-stems acquired Ablatives in -d
on the Analogy of the O-stems; and the Consonant-stems
availed themselves in Latin of the I-stem Abl., in Umbro-
Oscan of the O-stem Abl. (e.g. O. Lat. air-id, Ose. ligud, ‘ by
law’), The Locative Case in -é (possibly Instrumental, § 36)
of these Consonant-stems competed with this I-stem Abl. not
only in Consonant-stems (e. g. aeré and aerz), but also in I-stems
(e. g. civé and cévi), so that Varro declares that ove was heard in
his time as often as ovi, ave as avi. The confusion of cases was
increased by the circumstance that when -d of the Abl. was
dropped, nothing remained to distinguish Abl. -é(d) of E-stems,
-i(d) of I-stems, -a(@) of U-stems from Loe. -é, -2, - (see § 37), 80
that the dvz, dvi of Varro’s time have as much right to be called
Locatives as Ablatives. (On the question whether these forms
can have been Instrumentals, see § 36.)
§§ 33,34.) DECLENSION OF NOUNS, ETC. ABL. SING. 391

aa

The use of -é in the‘ Abl’ Sg. of I-stems was certainly not
so far advanced in the time of Plautus as in the classical period.
He uses only sorti, bili, civi, fusti, navi, &e., not sorté, &e. Priscian
however attests ree (presumably ré2, like forté, Most. 694) in Rud.
1020, and remarks (i. p. 331. 16 H.): vetustissimi solebant hujusce-
modi ablativum etiam in -e proferre. Of Consonant-stems. with
~i we have, e. g. pariéti (MSS. -e), Cas. 140, piimici (MSS, -e), Pers.
41, obiéci, Pers. 203. (For details of the use of -i and -e forms
by the Latin authors, and the rules laid down by the native
grammarians, see Neue, i*. pp. 212 sqq.). An early example of
the Ablatival use of -e (presumably -é) is the line of the Saturnian
Scipio epitaph (C. I. L. ic 30, ¢. 200 B. 6. ?);

Gnaiuod pitre prognatus fortis vir sapiénsque,

(Cf. aire moltaticod i. 181, Picenum) ; somewhat later are i, 198
(Lex Repet.) maiore parte diei; i. 199 (Sent. Minuc.) de maiore
parte; 1. 603 (Lex Furf.) mense Flusare (dial. ? Cf. Sab. mesene
Flusare). We have -ei, apparently a graphic variety of 7 (ch. i.
§ 9), in the Scipio epitaph of c. 130 B.c. (i. 34, along with aetate) :

is hic situs quei ninquam uictus est uirtutei

(cf. ab fontec on the Sent. Minuc., i. 199. 7); -¢ in the Lex Agr.
(i. 200. 23 ab eo herediue eius . . . testamento hereditati dedi-
tioniue obuenit, and again: curatore herediue), in the Lex Jul.
Municip. (i. 206 ubi continenti habitabitur), the Lex Rep. (i. 198.
56 de sanctioni, but also adessint for adessent). These are clearly
the later forms of the older -2d of i. 61 airid, i. 186 (S. C.
Bacch.) couentionid, xi. 4766 bouid.

§ 34. O. Lat. Abl. with -d. Ablatives with -d seem to be unknown in
Plautus and the earliest Dramatic literature, so that the final dental must
have dropped out of the spoken language before the end of the third cent. 8. c,,
though it is found in the Saturnian poetry (e.g. Naevius, Bell. Pun. 7 M. néctu
Troiad exibant capitibus opértis; C. I. L. i. 30, one of the oldest Saturnian
Scipio epitaphs : Gnaiuod patre (probably -é} prognatus fortis uir sapiénsque
(but SamnioAbl. on the same epitaph), and is persistently written in the 8. C. de
Bacchanalibus (i. 196) of 186 B. 0. (sententiad, couentionid, exstrad, suprad, oquoltod,
preiuatod, poplicod, facilumed, &c.). It ix as persistently omitted in a nearly
contemporary inscription (6. I. L. ii. 5041, Spain, of 189 B.¢.). [For other
examples of Abl. -d on inscriptions, and for passages of the Latin grammarians
referring to this form, see Neue, i”. p. 2, Ritschl, Neue Excurse, i., and add porod
(for class. porro) on a Praenestine cista, Mél. Arch. 1890, p. 303.) In Oscan the
-d remains, e.g. tristaamentud ‘testamenté,’ akrid ‘acri,’ egmad ‘re’ (cf.
39% THE LATIN LANGUAGE. [Chap. VI.

Faliscan sententiad), but in Umbrian it has been dropped as in Latin, e. g. poplu
‘populo,’ re-per ‘pro re,’ ocri-per, vea ‘via’ (cf. Pelignian oisa ‘us&’); similarly
with Adverbs in -éd, e. g. Osc. amprufid ‘improbe,’ Umbr. rehte ‘recte ;’ Cons.
stems show the O-stem Abl., e.g. Ose. ligud ‘lege,’ or the Loc. in -i, e. g. Pelign
aetate, Umbr. nomne.

§ 35. I-stem and Cons.-stem ‘Abl.’ in -iand -e. There is no evidence
of an old Cons.-stem -éd, later -2, corresponding to I-stem -id, later -7. The
dictatored (also navaled, but marid) of the Columna Rostrata (C. I. L. i. 195) is
probably a mistake, for the inscription is not the actual inscription of 260 B.c.,
but a copy made in the time of the Empire; and the instances in MSS. of
Plautus and the old poets with final -2 may be due to that ‘imperitia’ on the
part of scribes which Priscian (i. p. 345. 1 H.) blames for the change of civi, &c.
to cive in MSS. of Cicero. The reading of the best Palatine MS. (B) in Plaut.
Pseud. 616 is mititite, which points to a correction in the archetype of militi to
milite (the Ambrosian Palimpsest seems to have mitliti); and the MSS. often
vary between -i and -e, e.g. Naev. Bell. Pun. 14 M. pietati (v. 1. -te), Enn. A.
486 M. monti (MSS. montis and monte).

§ 36. Instr. Sing. The Latin grammarians knew nothing of
an Instrumental Case. Quintilian indeed (i. 4. 26) suggests
that a seventh case is required in Latin for such a phrase as
hasté percussi, where hasté is not a real Ablative ; though of the
previous existence of an Instrumental Case in the Latin language
he has no conception. But in various I.-Eur. languages we find
an Instrumental, and also a Locative Case; Sanscrit, for example,
has, in addition to the Abl. dévat, ‘from a god, the Instru-
mental! dévéna, ‘with a god’ (in Vedie also *déva), and the
Locative dévé,‘in a god.’ And the suffixes used in these languages
to form their Instrumentals and Locatives it is possible to find
also in Latin, though the weakening process which attacked every
Latin final syllable has made them indistinguishable from other
Case suffixes. To form the Instrumental Singular there seem to
have been originally two methods used in the Indo-European
language : (1) the addition of -é [according to some -&, which
would in Latin become -é (ch. iii. § 37)]; in Cons.-stems this -é
is found unchanged, in A-stems we have -a, in O-stems -6 or -é
and so on; (2) the addition of -bhi (e.g. Hom. Gk. i-ge), or -mi
(e.g. Lith. sinu-mi, ‘ with a son’); and various modifications of
these suffixes are found in the different languages. Of A-stem

? The Instrumental, it may be remarked, is the case used after the phrase
arthd bhavati (Lat. opus est).
§§ 35, 36.] DECLENSION OF NOUNS, ETC. INSTR. SING. 393

Instrumentals with I.-Eur, -a (and -am ?) (e.g. Gk. AdOpa ?) there
are no certain examples in Latin, for the Adverbs supra, extra,
&c. are written in the S.C. de Bacchanalibus suprad, exstrad,
and are therefore Ablatives. Some make O. Lat. contré (the
invariable scansion in O. Lat. poetry, Skutsch, Porschungen, i.
p- 3) an Instrumental, with the same shortening of -d as is seen
in the Nom. Sing. of A-stems, e. @. ferrd (ch. iii. § 43), contra,
the classical form being adapted to the type of supra, extra, &e.
But contri may be an Acc. Pl. Neut. form, and frustrié (the
O. Lat. quantity, e.g. xe frustré sis, Plaut.) need not be an
Instrumental either. The Oscan preposition contrud (i.e. *con-
trod ; ef. Lat contré-versia) is an Ablative.

The O-stem suffix -d would by the second cent. B.c., when -d
was dropped after a long vowel, be identical with the Abl., so
that modo, cito (usually with -d by the Law of Breves Breviantes,
ch. 11. § 42) may be either Instrumentals or Ablatives. Porro is
shown to be an Ab]. by the old spelling porod, mentioned above
(§ 34). The other O-stem suffix -6 may indeed appear in Jéné,
mélé, since the shortening of their final syllable by the Breves
Breviantes Law (e. g. Plaut. dtd for *datod) has advanced more
rapidly than in the case of -é¢; but on the other hand this might
be referred to their greater use in everyday life (cf. havé but moné
in Quintilian’s time, ch. iti. § 42), and their more frequent occur-
rence in word-groups, e. g. lene-rem-geras, male-ficio (cf. diéquinte
but fidé, ch. iii. § 44). (The scansion dené, malé in Plaut. is
doubtful.) Superlative Adverbs in -é have lost a final d, as is
shown by. faci/wmed on the 8. C. de Bacchanalibus. Similarly -é of
Fifth Decl. stems, e. g. ftéic?é, ré, may be Instrumental -é or Abla-
tive -éd ; -2 of I-stems may be Instr. -i or Abl. -id; -# of U-stems
may be Instr. -i or Abl. -iid; they may also be Locative -é (e.g.
postri-die), -yi- (Lat. -e7, class. -2, e.g. Nedpodi), -eu- (Lat. -w,
e.g. noctit) (see § 37). The -é of Consonant-stems, e.g. paéré,
cannot be an Abl. suffix, but either Instrumental -é (-4?), or
Locative -i (Lat. -2, e.g. Tibéré), used ablativally and instru-
mentally, as Loc. -i in Gk. Cons.-stems was used to express all
the meanings of the Greek Dative case, e.g. tarp-t.

It thus appears how difficult it is to establish by certain proof
the presence of Instrumental forms in the Latin declension, owing
394 THE LATIN LANGUAGE. [Chap. VI.

to the lack of a sufficient number of examples from the earlier
inscriptions in which Abl. forms have not yet lost their final d
(e.g. C. J. D. xi. 4766 bouid piaclum datod ; i. 61 airid [coir]au-
[it]; i. 181 aire moltaticod ; Zvet. J. I. I. 72 opidque Uolgani)?,
and Loe. -¥ would be distinguished from Instrumental -é (-4 ?).
Yet the evidence of cognate languages shows that Instru-
mental case-forms must have been a living part of Latin at
some period, however remote; and when we come to examine
the formation of Latin Adverbs we shall find that some of them
are believed with a fair amount of probability to be Instrumentals.
The evidence that we can draw from forms on Oscan inscriptions
(they are not very numerous), is all in favour of the supposition
that in the declension of the Noun the Instrumental forms had
quite dropped out of use. The Oscan language, unlike the Um-
brian, does not drop final d; and mdubitable Ablative forms
with -d are used in all the senses of the Latin ‘ Ablative,’ to
express our prepositions ‘from,’ ‘ with, ‘by,’ &c., e.g. kamben-
nieis tanginud, ‘by decree of the assembly,’ eftiuvad, ‘ with
money,’ tp efsid sakaraklid (Lat. apud id sacellum). (The doubt-
ful eitie of Zv. 1. [. I. 89 : stvad eftie upsed (Lat. sua pecunia opera-
tus est), requires confirmation before it can be used as evidence that.
IE-stems used an Instr. or Loc. -ié instead of Abl. -iéd. The
absence of an Abl. in -iéd from the early Latin inscriptions can
be explained by the comparative paucity of Fifth Decl. stems.)
In the Pelignian dialect, a variety of Oscan, we have in the few
inscriptions preserved an ‘ Ablative Absolute,’ oisa aetate (Lat.
usd aetate, with passive sense of the Deponent, ‘his life having
been exhausted’) (cf. forte, of doubtful meaning, on the same
inscription), and an Ablative of uncertain construction, suad (?)
aetatu firata fertlid (Lat. sud aetate . . . fertili), with apparent
dropping of -d before initial f; and this evidence, so far as it
goes, points to an Abl. of Cons.-stems in -ud (i. e. -dd, the O-stem |
Abl. suffix), beside another case in -e (presumably -8). If how-
ever this -e represents an original -i (cf. Pel. ae for ai, Ose.
Bansae Loc.) the case will be a Locative, not an Instrumental ;
and this view is favoured by the fact that other stems have in

 

1 The (Adverb ?) méritd is spelt meretod or meritod on the oldest inscriptions.
§ 37.] DECLENSION OF NOUNS, ETC. LOC. SING. 395

Oscan a Locative as well as an Ablative Case, e.g. A-stems:
viaf mefiai Loc., beside eftinvad Abl. (and etwas Gen.) ; O-stems :
Ladinei Loc., beside~tristaamentud Abl. (and sakarakleis Gen.).
An isolated example of an Adverb formed apparently by the
Instr. suffix in Osc. is suluh ‘omnino’ (Zvet. J. /. [. 129), though
this stands on a carelessly written inscription, a leaden execra-
tion tablet, and is not free from the suspicion of being meant for
sullud, a form which seems to occur (the last letter is unfortu-
nately not quite legible) on another tablet of the same kind (J. F.
il. 435; cf. von Planta,i. pp. 577-80). The conclusion therefore
which the scanty evidence at our disposal entitles us to draw is
that Instrumental formations, though they may be found in some
Latin Adverbs, are not found in the declension of Latin Nouns,
the case-forms which competed with the Latin Ablative (especi-
ally in Cons,-stems) being Locatives and not Instrumentals.

§ 87. Locative Singular. Locatives in J.-Eur. seem some-
times to have had a final -i (e. g. O. Ind. mirdhén-i and mirdhn-i,
‘on the head’), sometimes not (e. g. O. Ind. mirdhan, Gk. dduev
Inf., a Loc. as dduevar is a Dat.). Of Locatives without -¥ in
Latin there are only uncertain traces, such as Prepositions like
péenés (Loc. of penus, with -es not sunk to -2s, possibly because
the accent rested on it in collocations like penés me, penés te, ch. iii.
§ 12a. 3), and Adverbs like xoct#. The predominant formation
is with -7. Of these i-forms, A-stem locatives show I.-Eur. -ai
in O. Ind. 48vay-am, Lith. rafikoj-e, with Postpositions -am and -e,
but in Greek the ordinary diphthong -ai (a * doublet’ of ai, ch. iv.
§ 45), e.g. OnBar-yerys ; and this is also the Italian form (O. Lat.
-ai, class. Lat, -ae, Ose. -ai, Umbr. -e). O-stems took -oi and
-ei (e.g. Gk. ofxor and ofxe:). By the phonetic laws of Latin
both these suffixes would become -ei, class. -2 (ch. iii. § 18) ; so
the origin of the suffix of Cdrinthi, &c. is, so far as Latin is con-
cerned, doubtful. But in Oscan we have ~ei, which must be
I,-Eur. -ei (e.g. mtinikef teref, ‘on common ground’) (Umbr. -e,
e.g. destre onse,‘ on the right shoulder,’ may be -oi or -ei). Fifth
Decl. stems took -éi, which already in the ‘I.-Eur. period ’ had a
‘doublet’ -é. It is this latter form which appears in Latin, e. g.
die in the phrase postri-dié, die crastini, &e. (unless die has lost
396 THE LATIN LANGUAGE. (Chap. VI.

a final 7; cf. Faliscan foied ‘ hodie’). I-stems had -éyi (e. g.
Hom. Gk. a7dd¢i), which in Latin would become -e7, class. -é,
and would be merged in the Dative (§ 23) [possibly Instrumental
(§ 36)] suffixes. U-stems had -éwi (e.g. Hom. Gk. adorei), but
Latin U-stem Locatives show -#, the i-less formation, e. 9. nocti,
mentioned above. I-stems showed -tyi, Lat. -i, as U-stems -tiwi,
Lat. -we, e.g. sue. Consonant-stems had -i (used in Greek as
Dative suffix, as well as Locative), which in Latin became -é,
e.g. Carthaging, réré, and Infs. Act. like dgere, viveré (contrast
O. Ind. jivds-é, Lat. agz, which are Datives), though by false
analogy of O-stems (or I-stems ?) we sometimes find -2, e. g. ruri,
perhaps introduced to discriminate the locative from the ablative
use, e.g. rurt esse, from ruré venire.

§ 38. Locatives in -iand -ein Latin. Hériis by modern editors written
with -e when the last vowel has to be scanned short, otherwise with -7, The
scansion hert (by the Brevis Brevians Law, ch. iii. § 42) is common enough in
the early Dramatists (e. g. Caecil. Com. 197 R. heri uero, where heri has abun-
dant MS. authority), while the spelling here is established for passages like Plaut.
Mil. 59 (quantity of final vowel doubtful), where the Ambrosian Palimpsest
has here and the Palatine MS. hercle (cf. Pers. 108). Quintilian (i. 7. 22) says:
‘here’ nunc e littera terminamus: at veterum comicorum adhuc libris invenio:
‘heri ad me uenit,’ quod idem in epistulis Augusti, quas sua manu scripsit
aut emendavit, deprehenditur. [On his remark (i. 4. 8), in ‘here’ neque
e plane neque i auditur, see ch. ii. § 16.] The spelling of these forms, especi-
ally in the early writers, is often doubtful, and so it is difficult to prove with
certainty such a theory as that only rwri is used for ‘in the country,’ and
usually rure for ‘from the country’ in Plautus (Langen, Beitriige, p. 308). Cha-
risius (p. 200. 12 K.) attests heri for Afranius Com. 71 R., peregri for Naev. Com.
93 R., but peregre for Naev. Com. 84 R., as prae-fiscine for Afranius Com.36R. The
long quantity of the final vowel of pérégre (so both the Ambrosian Palimpsest
and the Palatine family) is required by the metre in Plaut. Truc. 127, an
anapaestic line, and peregr@ has been explained as the suffixless Locative of an
I-stem peregri- (ch. v. § 34) with the I.-Eur. ending -é, a doublet of -éi (ef.
O. Ind. agna, Loc. of agni-, ‘fire’). Vesptri, the form always used by the early
writers, is naturally referred to the O-Stem vespero-. Tempéri (-ori, see Georges,
Lex. Wortf. 3. v.) may be related to temporé as faeniri to faenoré, majort to majoré
(§ 33), and so with rw (e. g. Ter. Phorm. 363, Plaut. Cist. 226), Carthagini (e.g.
Plaut. Poew. 1056 AP), Accherunti Plaut., while médne (if not an Adverbial Ac-
cusative) may show conversely » Cons.-stem ‘ Ablative’ suffix applied to an
I-stem muani-, Manes Pl. (ch. v. § 37). The close connexion of the Ablative and
Locative, already mentioned in § 33, is seen in phrases like mane sane septimi,
Plaut. Men. 1157 ; luci claro, Plaut. Awl. 748 (the use of claro for clara or clarae is
due to the fact that luci being an Adverb does not have the fem. gender of lux).
(For fuller details about these Locatives, see Bell on the Latin Locative; Neue,
Formenlehre, ii®. p. 640, i%. p. 242).
§§ 38-40.] DECLENSION OF NOUNS, ETC. NOM. PLUR. 397

§ 89. A-stems, &. O. Lat. -ai is never dissyllabic, like -ai of the Genitive.
In Plautus a common A-stem Loe. is riciniae, as in the phrase proxumae uiciniae,
‘next door.” We have Romai on a very early inscription, C.I.L.i. 54: med
Romai fecid. Die quinti occurs in Cato’s account of Maharbal's boast to Hanni-
bal: mitte mecum Romam equitatum ; die quinti in Capitolio tibi cena cocta
erit (ap. Gell. x. 24. 7).

$40. Nom. Plur. I. Masc., Fem. The I.-Eur. -és, which
appears in Cons,-stems as -és (e. g. O. Ind. matdr-as, Gk. unrép-es),
in A-stems as -ds, in O-stems as -és, and so on, is the suffix in
use among the Umbro-Oscan dialects (e.g. Umbr. frater for
*fratr-és, Osc. censtur for -rés, aasas * arae’), but in Latin is hardly
found except in I-stems whose -és represents I.-Eur. -éyés (e. g.
from the J.-Eur. stem tri-,‘ three,’ O. Ind. tréyas, Cret. Gk. rpées
for *rpeyes, Att. rpeis) Lat. trés. In O-stems this I.-Eur. suffix -ds
is replaced in many of the I.-Eur. languages by -oi, the Nom. Pl.
suffix of the Pronominal Declension (e.g. J.-Eur. *toi, O. Ind.
té, Hom. Gk. rof, Lat. 2s-t7 from -tei from original -toi). Thus
in Greek we find -o1 (e.g. Avxou), in Celtic -oi (e. g. Gaul. Tano-
taliknoi; O. Iv. fir Nom. Pl., ‘men, points to an original
*wiroi, as does Lat. virz, while firu Voc. Pl. is either the I.-Eur.
Nom. Pl. in -ds or the Acc. Pl.), Teutonic -ai (used in Adjectives,
L.-Eur. -ds being used in Nouns), Balto-Slav. -ai (e.g. Lith.
vitkai, O. SI. vliici, ‘ wolves’). Similarly in Latin we find -7
from -ei. from still earlier -oe or -oi, e.g. O. Lat. poploe (Carm.
Saliare), puplei, class. popiilz, which is thus distinguished from
Ace. Pl. populés as Nom. Pl. ist@ (originally -toi) from Ace. Pl.
istos. The prevalence of this Pronominal oi-suffix among the
European languages suggests the possibility that Umbro-Ose.
-6s (e.g. Umbr. Atiersiur ‘ Attiedii,’ Osc. Navlanis ‘ Nolani’)
may have had at one period a struggle for existence with -oi,
and may have owed its acceptance into use to the analogy of A-
stem Noms. Plur. in -4s. The Latin A-stem suffix -ae, e.g.
arae (O. Lat. -ai, never dissyllabic, according to Prise. i. p. 221 H.),
is an example of the contrary change from a prehistoric -as, of
which no traces remain, to a new formation made on the model
of the -oi of O-stems. It resembles Greek -d (e.g. xGpar), but
must have been originally -di, since the ordinary diphthong -ai
would become -2 in the unaccented syllable in Latin (cf. occtdi for
ée-caido, ch. iii. § 18). Lat. 2é-stems with Nom. Pl. -iés may
398 THE LATIN LANGUAGE. (Chap. VI.

have the I.-Eur. suffix, but U-stems with -7s show the Acc. Pl.
suffix, instead of the I.-Eur. Nom. Pl. -éwés, which would be in
Latin -ié%s, -itis. (This would hardly contract into -ds, as we saw
before, § 21). The -is which is occasionally found for -és in I-stems
is also an Ace. Pl. suffix (§ 51); and the I-stem Nom. Pl. vis (so
in O. Lat., but in class. Lat. vir-és for *vis-és, an S-stem) is pro-
bably an Accusative form. The -és of Cons.-stems, e. g. matr-és,
censor-és may either be the I-stem Nom. Pl. ending, since there is
so much interchange of Cons.- and I-stems in Latin (§ 30), or the
Ace, Pl.; and the same is true of the U-stem -wés, e.g. sués.

§ 41. A-stems. Ritschl (Neue Excurse, i. p. 118) proposed to avoid hiatus in
some passages of Plautus by the change of Nom. Pl. -ae to -ds, reading e. g.
dlternas in Trin. 539:

nam falguritae sunt alternae arbores,

where all the MSS., the Palatine family as well as the Ambrosian Palimpsest,
read alternae (which may be right, the hiatus being palliated by the allitera-
tion, though editors prefer alternis, or alternas, an Adverb like dias, altéras).
He quoted in support of this change a line from one of the Atellanae of
Pomponius, ¢. 90 B. c. (Com. 141 R.):

quot laetitias insperatas médo mi inrepsere in sinum,

where laetitias insperatas is now usually explained as Acc., governed by inrepsere,
though it may quite well be a dialectal form, for the Atellanae in imitating
the manners of country life may also have imitated its language. Dialectal
-as Nom. Pl. is found in the old inscriptions of Pisaurum with the s dropped
(¢C. I. L. i. 173 matrona Pisaurese dono dedrot; 177 dono dedro matrona),
along with Gen. Sg. -d(s) [or Dat. Sg. -a(i), § 24]. In early inscriptions we
have -ai for class -ae, e. g. tabelai, datai on the 8. C. de Bacch.

§ 42. O-stems. Piluwmnoe poploe was a phrase used of the Romans in the
Carmen Saliare (Fest. 244. 25 Th. velut pilis uti assueti) ; ef. fescemnoe (qui
depellere fascinum credebantur, Paul. Fest. 61. 10 Th.; should we read Fesceni-
noe, class. Fescennini?). On early inscriptions we have -ei, wv. g. foideratei, uiret,
oinuorsei on the 8.C. de Bacch., sometimes written -e (cf. ch.iv. § 34), e.g. ploirume
(¢. I. L.i. 32). A Nom. Pl. of an IO-stem with -is occurs on an inscription of
the first cent. a.p. (C. I. L. i. 1541), filis. It is impossible to say whether
Jilei (i. 1272), feilet (i. 1284) (ef. socei, i. 1041) is meant for this form (cf. Clodé for
Clodis Nom. 8g.), or is a misspelling of filiet (i. 1275) or a contraction of it (like
gratis for older gratiis), O-stems show -eis, also written -es, -is, in some inscrip-
tions of the end of the second or beginning of the first cent. B.v., e.g. magistreis
(C. I. L. i. 565, Capua, 108 B.c,: heisce magistreis Uenerus Iouiae muru aedi-
ficandum coirauerunt), lanies (vi. 168, Rome), violaries rosaries coronariis (vi. 169,
Rome). (For other examples, see Ritschl, Opuse. ii. 646, and add heisce magis-
treis, Not. Scav. 1893, p. 164, from Capua, mustae picis, C. I. L. iii. Suppl. 12318,
from Samothrace.) This form is attested for the pronoun hic by Priscian
(i. p. 593. 5 H. inveniuntur tamen etiam nominativum ‘hisce’ proferentes
antiqui), and is found in the Nom. Pl. Mase. of hie, ile, iste in the Dramatists
§§ 41-45.] DECLENSION OF NOUNS, ETC. NOM. PLUR. 399

before a word beginning with a vowel, when the particle -ce is added, e. g.
Plaut. Mil. 374 :

non possunt mihi miniaciis tuis hisce oculi exfodiri,

(similarly ilisce and istisce are the forms used in Plautus before a word begin-
ning with a vowel, never illic, istic; Studemund in Fleck. Jahrb. 1876, p. 57),
though probably never in the Nom. PI. of O-stem nouns. In the pronouns it
seems to be due to the addition of the plural suffix -s to the already formed
plural in -2 (older -e/) ; in the Noun O-stems it may have the same origin,
though it is not unlikely that the 10-stem formation mentioned above had at
least some share in bringing it into use.

Deus has two Nom. Pl. forms, dei, a dissyllable (probably the more cere-
monious form), and di, also written dii,a monosyllable. (Dii and dei are
compared to ii and e by Prise. i. p. 298 H.).

§ 43. I-stems. Varro (Z. L. viii. 66) says that puppis and puppes, restis and
restes were rival forms in his time, like Abl. dvi and ove, dvi and ave. On early
inscriptions we have usually -es, e. g. aidiles, C. I. L. i. 187, Eph. Epigr. viii. 676,
but ceiveis on the Lex Repetund. of 123-122 B.c. (C. I. L. i. 198. 77), fineis and
Jiis in the Sent. Minuciorum of 117 B.c. (ib. i. 199), and pelleis on the Lex
Furf. (i. 603) (ef. cogues atriensis on a Praenestine inscr., i. 1540). (See Neue, i’.
p. 246.) The O. Lat. Plur. of vis was vis (Prise. i. p. 249. 9 H.).

§ 44. Cons.-stems. Lat. quattwor appears to be a relic of the -és formation,
for *quattuor-és (§ 63), I.-Eur. *q3étworés (0. Ind. catvaras, Dor. Gk. 7éropes,
O. Ir. cethir), though some make it represent I.-Eur. *q¥etwor, supposing this
to be a byform of the ordinary Neuter, which would be in Latin *quattuora
(Ose. petora or *petoro). Plautine scansions like céinés, turbinds (Trin. 835) are of
course mere examples of the Law of Breves Breviantes, like the Imperatives
cdivé, pitd, &c. (ch. iii. § 42) (cf. Ace. Pl. liberts virginés, Pers. 845) and are no
evidence of the use of the suffix -és.

§ 45, Nom,, Acc, Plur, II. Neut, In the Italic, Balto-
Slavic, and Teutonic languages all Neuter stems form their
Nom. and Ace. Plur. in -4, while in Greek we have -d. This
a seems to have been originally peculiar to O-stems, and to be in
reality the same as the Nom. Sg. Fem. suffix. Prof. Johannes
Schmidt, in his book on the Indo-European formation of Neuter
Plurals (Die Pluralbitdungen der Indogermanischen Neutra,
Weimar, 1889) has mustered an array of facts from the various
I.-Eur. languages, which point to the Neut. Plur. having been
originally a Collective Fem. Sg. like Lat. /dmi/ia in the sense of
fimili, so that, e.g. Lat. juga originally meant what the
Germans would express by ‘das Gejiche,’ the yoke-material.
The use of a Singular Verb with a Neut. Plur. subject in
400 THE LATIN LANGUAGE. [Chap. VI.

Greek, O. Ind. (Vedic) and Zend may be explained by this
hypothesis, e.g. pijpa in Hom. J7. i. 464 eel ward pijp éxdn
differs from pnpot of 1. 460 pnpods 7 e€€rauov in signifying the
mass of meat as opposed to the thighs separately. And the
change of Gender in Lat. caementum N., caementa F., mendum N.,,
menda F. may be due to the fact that a Collective Sing. Fem.
caementa, menda, being treated as a Plural (a ‘ Nom. Plur. Neut.’),
developed a new Singular, caementum, mendum (Nom. Sing.
Neut.). Other illustrations of the connexion between a Col-
lective Sing. Fem. and a Plur. Neut. may be seen in Prof.
Schmidt’s book, e. g. Lat. dpéra Sg. Fem. and dpéra Plur. Neut.,
Hom. Gk. ra julia, Att.) jvia, with plur. ai jvia. (The Latin
examples of change of Gender like docus Sg., loca Pl. are to be
found in Neue, i?. p. 540.)

The suffix -& appears, as we have seen, in the Nom. Plur. of
all Neuter-stems in the European languages. But in the oldest
Indian and Zend literature we have relics of an earlier state of
things, viz. -a for O-stems, -i for I-stems (a Latin relic of -i is
tri-ginta, lit. ‘ three tens,’ § 74), -i for U-stems (perhaps the long
vowel of Lat. pécu, génu, cornu, &c. may be due to their having
been originally Neut. Plur.; another suggestion is, that geni,
cornit, &c. were Duals, like I.-Eur. *siint, Nom. Dual of the
stem *siinti-, ‘a son’); while Cons.-stems lengthened the
vowel of their final syllable [e.g. Vedic nama, ‘names,’ for
*nomo(n)?; Prof. Schmidt compares Lat. quattuor from I.-Eur.
*q*etwor, but see § 63]. All these formations Prof. Schmidt
identifies with Fem. Sing. Collectives.

Like the -a of the Nom. Sing. of A-stems, Neut. Plur, -a
appears in Latin as -& in the earliest literature, and in Umbro-
Oscan becomes an o-sound, Osean ui, Umbr. u, (a), both written
in the Latin alphabet 0, which is scanned by Lucilius as a short
syllable in the Oscan loanword sollo (Lat. tota Neut. Pl.) (ine.
160 M.):

uasa quoque omnino dirimit, non sollé dupundi,
and is written -a by Festus in the Oscan numeral petora, four
(250. 30 Th. petoritum et Gallicum vehiculum esse, et nomen
ejus dictum esse existimant a numero quattuor rotarum. alii
Osce, quod hi quoque ‘ petora’ quattuor vocent). In Umbrian
§ 46.] DECLENSION OF NOUNS, ETC. GEN. PLUR. 401

it appears that the Nom. and Acc. Pl. of Neuter-stems were
distinguished by the addition to this -o of the -r (-s) and the -f,
which are the final letters of the Masc. Nom. and Ace. Plur.,
though it is possible that this -r and -f were not pronounced,
but were used merely as graphic criteria of the two cases, e.g.
tuderor Nom., verof Acc. Whether it is merely accidental, or
not, that the ordinary forms in -a, -o are found in the Acc.
along with forms in -of, but not in the Nom. along with forms
in -or, the limited material does not allow us to decide.

‘With that interchange of the Cons.- with the I-declension
mentioned in § 50, we have e. g. pliria (cf. compluria) and plura,
forms discussed by Gellius (v. 21), who tells us of a letter
written by Sinnius Capito to Pacuvius Labeo to prove the
thesis: ‘pluria non plura dici debere” We have -ia in the
Neut. Plur. of Adjective Stems like dérétia, auddcia, victricia,
ferentia, &c. (see Neue, i?, p. 121), but always vééra from
vetus, which was originally a Noun (Gk. (F)éros, a year) (§ 55).
In O. Lat. we have silenta, for silentia, quoted by Gellius
from Laevius (Gell. xix. 7. 7 ab eo quod est sileo ‘silenta
loca’ dixit et ‘ pulverulenta’ et ‘ pestilenta’) (see ch. v. § 64).

§ 46. Gen. Plur. The suffix -dm is indicated by most of the
I.-Eur. languages (e.g. Gk. tmwv, unrép-wv), which would in
Latin become in time -dm (ch. ili, § 49), then -dm (e.g. socium
on the S.C. de Bacch.) (cf. Ose. Naivlanam, Mapeprivovp,
Aovxavoy, Umbr. Atiersio). A-stems took in Greek and in the
Italic languages -asom, e.g. (Hom Gk. Oedwv, Att. dedv, Lat.
dedrum, Osc. egmazum ‘rerum,’ which was the Gen. Plur. Fem.
suffix of Pronouns (e.g. O. Ind. tasam, Hom. Gk. rdwv, Lat.
is-tarum), and after this model a Gen. Pl. of O-stems was formed
in Latin with the suffix -drwm (O. Lat. -orom), a suffix not found
in Umbro-Oscan, which by Cicero’s time drove the older -om, -am
off the field. On the same model the Fifth Decl. stems formed
their Gen. Pl., e.g. faciérum. The use of -cwm in Gen. Plur. of
Adjectives like férentium, auddcium, and of -wm in Gen. Pl.
dpum, volucrum, vatum, &e. (see instances in Neue, i’. pp. 258 sqq.,
e.g. clvitdtum and civitatium), is due to that confusion of Cons.-
stems with I-stems, which played so great a part in the Latin

pd
402 THE LATIN LANGUAGE. [Chap. VI.

declension, and which occupied a great deal of the attention of
the native grammarians

§ 47. -um and -orum in O-stems. Cicero’s remarks on these suffixes are
worth quoting (Orat, xlvi. 155): atque etiam a quibusdam sero jam emendatur
antiquitas, qui haee reprehendunt ; nam pro deum atque hominum fidem
‘deorum’ aiunt. Ita credo. Hoc illi nesciebant? an dabat hance licentiam
consuetudo? Itaque idem poeta (Ennius) qui inusitatius contraxerat : Patris
mei, meum factum pudet, pro ‘meorum factorum,’ et: Texitur, exitium
examen rapit, pro ‘exitiorum,’ non dicit ‘liberum,’ ut. plerique loquimur,
cum ‘cupidos liberum’ aut ‘in liberum loco’ dicimus, sed ut isti volunt:
Neque tuum unquam in gremium extollas liberorum ex te genus. Et idem:
Namque Aesculapi liberorum. At ille alter (Pacuvius) in Chryse non solum :
Ciues, antiqui amici maiorum meum, quod erat usitatum, sed durius etiam :
Consilium socii, augurium atque extum interpretes ; idemque pergit: Post-
quam prodigium horriferum, portentum pauor. Quae non sane sunt in
omnibus neutris usitata. Nec enim dixerim tam libenter ‘armum judicium,’
etsi est apud eundem: Nihilne ad te de iudicio armum accidit ? quam cen-
turiam, ut censoriae tabulae loquuntur, fabrum et procum audeo dicere, non
‘fabrorum’ et ‘procorum.’ Planeque ‘duorum virorum judicium’ aut
‘triumvirorum capitalium’ aut ‘decemvirorum stlitibus judicandis’ dico
nunquam. Atqui dixit Attius: Uideo sepulera duo duorum corporum ;
idemque : Mulier una duum uirum. Quid verum sit intellego, sed alias ita
loquor, ut concessum est, ut hoc vel pro deum dico vel pro deorum, alias, ut
necesse est, cum triumvirum, non ‘virorwm,’ cum sestertium, nummum, non
‘nummorum,’ quod in his consuetudo varia non est. Similarly Varro (LZ. L. viii.
71): quaerunt, si sit analogia, cur appellant omnes aedem Deum Consentium
et non ‘Deorum Consentium’? Item quor dicatur mille denarium, non
‘mille denariorum’; est enim hoe vocabulum figura ut Vatinius, Manilius,
denarius ; debet igitur dici ut Vatiniorum, Maniliorum, denariorum; et non
equum puplicum mille assarium esse, sed mille ‘assariorum’; ab uno enim
assario multi assarii, ab eo assariorum. (A list of Genitives Plur. of O-stems
in -umis given by Neue, i*. 103.) Nostrum and vestrum, Gen. Plur. of noster, vester,
established themselves in class. Lat. as Gen. Pl. of nos, vos (ch. vii. § 9).
A-stem Genitives like agricilum (Lucr. iv. 586) follow the analogy of O-stems,
as do vectigdliorum, anctliorwm, &c., and perhaps currum, &c. 5 anphérum and
drachmum follow the Greek. On the (restored) Columna Rostrata (C. I. L. i.
195. 10) we have the Pronoun olorom ‘illorum’; on a Scipio epitaph of
perhaps the end of the third cent. B.c. (i. 32), the Adj. duonoro ‘ bonorum,’
but -o(m) on the earliest coins, e.g. C. I. L. i. 15 Caleno (with Neorodtwy), i. 16
Suesano (with NeomoA:twv}. (See ch. iii. § 49.) Ini. 24, of end of third cent.
B. C., LADINOD is usually read Lavinor., a Gen. Pl.

§ 48. Dat., Abl., Loc., Instr. Plural. These four cases must
be considered together; they are so intermingled in Latin and
in other I.-Eur. languages. Latin A-stems show -?s, older -e7s,
which has come from an earlier -ais (cf. Oscan -a{s, Greek -ais),
a suffix apparently formed in imitation of the -ois of O-stems.
§§ 47-49.] DECLENSION OF NOUNS, ETC. DAT., ABL. PLUR. 403

This O-stem suffix, in Latin -is, earlier -eis, and still earlier -ces
or -o1s (Osc. -tis, -cis) is regarded by some as a Locative, by
others as an Instrumental (see Brugmann, Grwndriss, ii: §§ 357,
380). The suffix -diis, O. Lat. -20s, shown by other stems (e.g.
ewit-bus, lég-t-bus with thei of I-stems, /écii-bus or lict-bus, sii-bus
or sé#-bus), comes from an original -bhés. In Umbro-Oscan, as
usual, the short vowel of the final syllable is syncopated, e.g. O.
Ose. Juisari-fs (in Lat. *isdribus), Osc. teremn-{-ss (in Lat.
terminibus), lig-i-s (in Lat. lég-t-bus), Umbr. fratr-u-s, with s
for ss and so not changed to r, and preceded by a vowel which
may be the I.-Eur. ‘ Composition Vowel’ o (ch. v. § 80), though
this is quite uncertain. This suffix was in O. Lat. em-
ployed in the Dat., Abl. Plur. of A-stems, e.g. deatrabus (cf.
Gaul. MadzpeB8o NapavoixaBo), but in the classical period this
form was retained only in legal language, for the purpose of
distinguishing A- from O-stems, e.g. filiis e¢ filiabus, just as we
retain the old Plural suffix in * oxen’ but have dropped it in ‘ shoes’
(earlier ‘shoon’). Adverbs like aids, O. Lat. altéras, fords may
show the I.-Eur. Locative of A-stems (Gk. -doi, e.g. Qvpdor;
O. Ind. -asti, e. g. a$vasu) (but see ch. ix. § 4).

§ 49. A- and O-stems. The old form privicloes (privis, id est singulis), the
Dat., Abl. Plur. of a diminutive of privus, is quoted from the Carmen Saliare
by Festus (244. 21 Th.), and Paul. Fest. 14. 17 Th. has: ab ‘oloes’ dicebant
pro ab illis, but -eis, sometimes written -es (ch. iv. § 34), is the spelling of the
older inscriptions: e. g. uieis, leibereis, populeis, aedificieis, agreis, loceis, on the Lex
Agraria of 111 B.¢. (C. 1. L. i. 200) ; soweis nuges on an old epitaph of a mimus
(i. 1297) :

plouruma que fecit populo soueis gaudia nuges,

which has é for ei also in the Nom. Sing. Masc. of the Relative, que for quet
(class. gui); de manubies (Eph. Epigr. viii. 476, Capua, 135 B.c.). We find -iis
contracted in course of time into -is; thus grdtiis (always with -tis, and
similarly ingratiis, in Plautus and Terence ; ef. gratiis in a line of Pomponius,
v. 90 B.C., Com. 110 R.) became gratis in classical Latin ; provincis, &c. beside
judiciis (the long i being indicated by the tall form of the letter), occur on the
Mon. Ancyranum ; and of 10-stems we have, e.g. Januaris (C.I. L. vi. 543, of
115 A.D.), Junis (vi. 213, of 131 A.D.). (For other examples, see Neue i*. p. 31.)
An example of -dbus in O. Lat. is quoted by Nonius (493. 16 M.) from Livius
Andronicus’ translation of the Odyssey: déque manibus dextrdbus. The
passages of the grammarians bearing on this form, and details of the use
of deabus, filiabus, and libertabus, the most frequent words of the kind, are given
by Neue, i”. pp. 22 sqq. We have Mase. -dbus with Fem. -dbus in the Duals
duobus, wmbobus, though the rarity of the formation led to the latter being

pda
404 THE LATIN LANGUAGE. (Chap. VI.

replaced in Vulg. Lat. by ambis Masc. and Fem. (Caper 107. 14 K. ambobus, non
‘ambis’ et ambabus). O-stems sometimes take the I-stem and Consonantal
-ibus in Late and Vulgar Latin, e.g. C. I. L. vi. 224 dibus omnibus deabusque
(197 A.D.); 15267 amicibus; 17633 alumnibus. Pomponius, the writer of
Atellanae, uses pannibus (Com. 70 R.) for pannis in imitation of the rustic mode
of speech ; and in the Sermo Plebeius of Petronius we have diibus (Sat. 44.
p. 29, 35 B. ita meos fruniscar, ut ego puto omnia illa a diibus fieri). In the
0. Lat. inscription (C.I. Z. i. 814): devas Corniscas sacrum, found in the
‘Corniscarum divarum locus trans Tiberim’ (Paul. Fest. 45. 16 Th.), the two
first words may be Gen. Sing. If Plural, they are Locatives like alias. The
instances of dialectal and Lat, -os are all doubtful (Class. Rev. ii. p. 204).

§ 50. Other stems. O. Lat. -bos of trebibos on a bronze vase in the British
Museum (Eph. Epigr. ii. 299 Q. Lainio Q. f. praifectos protrebibos fecit), corre-
sponding to class. -bus, indicates a short vowel, for -bds would have retained 6
(ch. iii. § 18). The few apparent examples of its being scanned as a long
syllable by the Dramatists (collected by C. F. Miller in his Plaut. Prosodie,
p. 53; add Naev. Trag. 57 R.) must be illusory (many of them are cases of
sylJaba anceps at a pause in the line, e.g. Plaut. Merc. goo, Rud. 975).

$51. Ace. Plur. The I.-Eur, suffix was -ns, after a con-
sonant -ns. Thus A-stems ended in -ins, which became -as,
O-stems in -6ns, which became -6"s, I-stems in -ins, R-stems in
-rns, S-stems in -sns, and so on. Latin exainples are vids, updos,
JSratres [with -és from -éus (ch. ii. § 64) from I.-Eur. -ns], /ondrés
for -dséms. I.-Eur. -ns, -ns becomes in Oscan -ss, in Umbr. -f,
e.g. Ose. viass ‘ vias,’ fefhiiss, walls (cf. Gk. retyos), Umbr. vitlaf
‘ vitulas, or with loss of -£ witla, toru for *toruf ‘ tauros, avef,
avif, and aveif ‘ aves,’

I-stems in Latin should show -7s (from -ins, ch, ii. § 64); and.
this is the usual form in the best MSS., though we often find
-¢s, the Nom. Pl. ending or the Cons.-stem ending. Thus wrédzs is
attested for Virg. G.i. 25: urbisne invisere, Caesar, &c., but wrbes
for A. i. 106: centum urbes habitant magnas, res for A. x. 350,
but ¢/7s for the following line (Gell. xiii. 21); so on the (restored)
Columna Ros . 1. 195) [c]lasesque nauales .. . claseis
Poenicas . . eee tacinienseis ...naueis. (For statistics
of thapse of -és and -és, see Neue, i 1% p. 245.)

$52. II. THE COMPARISON OF ADJECTIVES. The
J.-Eur. suffixes used to form the Comparative and Superlative of
Adjectives have been already mentioned in chap. iv. For the
Comparative, (1) -yés-, with weak grade -is- (e.g. O. Ind. svad-
iyas-, Gk. 76fm Acc. for *jd.0ca, Goth, sutiza, Lith. sald-és-nis),
§§ 50-52.] DECLENSION OF NOUNS, ETC. COMPARISON. 405

Lat. sudvior, older *suavids, magis Adv.; (2) -téro- and -éro-, the
original sense of which was rather that of likeness, of equal, than
of greater degree (e. g. O. Ind. vatsa-taré-, lit. ‘like a calf,’ Lat.
mater-tera, lit. ‘like a mother, Ir. demnithir, ‘equally certain,’
from demin, * certain’); for the Superlative, (1) -is-to- (-is-tho-?),
(e.g. O. Ind, 4&-i&tha-, Gk. dxuoros; O. Ind. svad-iktha-, Gk. #3-
totos, Goth. sut-ists, Eng]. sweet-est), apparently composed of the
weak Comparative suffix -is- and the TO-suffix (ch. v. §27); (2)
-temo- or -t*mo- (-tmmo-) and -emo- or -*mo- (-mmo-) (ch. v. § 14)
(e.g. O. Ind. dn-tama-, Lat. in-¢imus; O. Ind. upamé-, Lat. swmmus
for *sup-mus). This last Superlative suffix was, like the Com-
parative -tero-, -ero- (Ascoli, Suppl. Arch. Glott. Ital. i. 53),
originally a suffix denoting likeness (e.g. O. Ind. gé-tama., lit.
‘like an ox’) or position, and it has this force in Latin words like
aedi-tumus, lit. ‘living in a temple’ (later corrupted to aedi-tuns,
as if from ¢weor, ‘ guarding a temple,’ Gell. xii. 10; Varro, FR. R.
1.2. 1); fint-timus, miiri-timus, &c., so that e. &. ct-timus probably
meant originally ‘ near in position’ (cf. dex-timus, ‘on the right,
not ‘ most on the right ’), and is not properly a Superlative. _ To
give Superlative sense, the weak Comparative suffix was added,
-is-emo- (-is-*mo-), e. g. maximus for *mdgis-imus, sdcerrimus for
*sacris-imus, ficillimus for *facilis-imus (on the change of -ris-
to -ers-, -err-, &e., see ch. iv.§13). The origin of the usual Latin
Superlative ending -dssimus, older -issumus, has been matter of
much discussion. One theory makes the I.-Eur. suffix of
O. Ind. a8-istha-, Gk. @x-taros -istho-, not -isto-, and explains
Lat. -issimus as this suffix augmented by -emo- (-mo-). But
the change of I.-Eur. sth into Latin ss is not satisfactorily
proved by Lat. ossi-, bone (O. Ind. asthan-, Gk éoréov; see
ch. iv. §.95), and it seems safer to analyze -2ss¢mus into the suffixes
-isto- and -temo- (on ss for (s)tt, see ch.iv. §108)1. The -zs- of
-issimus, like the -is of magis, had short 7,a fact attested not only
by grammarians [Mar. Victor., p. 242. 24 K.; Vergilius, p. 189.
17 H. (Suppl.)], but by late spellings like merentessemo, haressemo

 

+ Or -issimo-, an ending peculiar, sime) obscured the presence of this
so far as is known, to Latin, may suffix, and may have led to a re-
contain the suffix -is- twice. The formation from ocius (*ocis), viz. ocis-
change of *dcis-wme(d) to oxime (*oc- — sime.
406 THE LATIN LANGUAGE. [Chap. VI.

(C.I.L. ii. 2997). (The tall form of I in some late inscriptions,
e.g. caRIsstmo C. I. LZ. vi 5325, does not prove that the vowel
was long; see ch. i. § 9).

The irregular Comparison of simple Adjectives like ‘ good,’
“bad? (e.g. Linus, mélior, optimus; Gk. dyadds, duetvov ; Engl.
good, better) is a relic of a very early time when different roots
were used to express a Positive, a Comparative, and a Superlative
notion,—bonus (older duonus) from dwen- (cf. O. Ind. divas-,
‘honour’), melior, optimus (C.J. L. 1. 1016 has opetuma, an
archaism) from the root op- of opto, dpes, &. On the Comparison
of Adverbs, see ch. ix. § 1.

In the Romance languages Comparison is expressed by the use
of the descendants of Lat. plus (Ital., French, &c.), magis (Span.,
Port., &c.), e. g. Ital. pid ricco, ‘ richer, il pid ricco, * the richest,’
except in these simple Adjectives like ‘good, ‘bad,’ which
retain their old irregular Comparison, e.g. Ital. migliore and
ottimo, ‘very good’ (il migliore, ‘the best’), peggiore and
pessimo, ‘ very bad’ (il peggiore, ‘ the worst’), &. Ital. -issimo
(e.g. ricchissimo, ‘ very rich,’ not ‘ richest’) shows itself by its
-is- instead of -es- (for Lat. -Zs-, ch. ii. § 14) to be a late innovation,
and no transmission from ancient times. (See Meyer-Liibke,
Rom. Gram, ii. p. 83.)

§ 53. The Comparative Suffixes. The suffix -yes- appears in Latin as -ior
Nom. 8g., M., F., -ddris Gen. Sg., -ius Nom. Sg. Neut., of which the older forms
were -ids, -idses, -iis. (On the change of s to 7, see ch. iv. § 148). Varro (L. L.
vii. 27) quotes from early Latin (from the Carm. Sal. ?) meliosem, and Paul. Fest.
(359. t Th.) maiosibus, meliosibus (his ‘meltom’ meliorem dicebant, 87. 25 Th., may be
a corruption of a gloss like ‘ melios’ melior ; see Class. Rev. v. 10 ; so in a Glos-
sary ‘meliosa’ meliora, Liéwe, Opusc. p.170). Priscian (i. p. 347. 2 H.) quotes
from the earlier historians Neuter forms like prior, posterior; thus from
Valerius Antias: hoe senatusconsultum prior factum est; from Cassius
Hemina: bellum Punicum posterior; from Claudius Quadrigarius: prior
bellum quod cum his gestum erat; and: foedus prior Pompeianum (on calor
Neut. in O. Lat., see § 14); though it is conceivable that the actual forms
used may have been priés, posteriis, which must have been the predecessors of
pritis, posteriis (ch. iii. § 18)}.

The yes-suffix was originally affixed to the root without the suffix of the
Positive stem, as in Lat. dc-ior, O. Ind. db-yas-, Gk. dx-iov from a Positive

 

' The supposed examples of -i#s in Plautus are illusory ; see Miiller, Plaut.
Pros. p. 55.
§§ 58-55.] DECLENSION OF NOUNS, ETC. COMPARISON. 407

stem *oku-, but Lat. sudvior for *suadv-ior, ténutor, aspérior, &c. start from the
Positive stem *swadu-, &. On the other hand the suffix -tero-, when added
to an Adjective, was affixed to the Positive stem, e. g. d«v-repos. This -tero- is
often added to Prepositions, e.g. ex-ter, ci-ler, postero- (similarly -ero- in
sup-ero-). In Latin it has not Comparative sense unless augmented by the
YES-suffix, e.g. ea-ter-ior, ci-ter-ior, dex-ter-ior, sinis-ter-tor ; dciter, &e. being
apparently the Latin equivalent of the Gk. dv-repo-, &e. (ch. ix. § 1).

§ 54. The Superlative Suffixes. The suffix -temo- or -temo- (-tmmo-) with
-emo- or -°mo (-mmo-) is closely associated with Comparative -tero-. -ero-,
ev. g. ci-timus goes with ci-tero-, in-timus with in-tero- (O. Ind. dn-tama- with
an-tara-), summus from *sup-mus with sup-ero-, infimus (imus seems not to occur
in Plautus) with infero-. The old augural term sollistumum tripudiwm may
combine it with the Comparative -is-. Its original form is difficult to
ascertain. The spelling on Republican inscriptions is -tumus, -wmus (ch. iii.
§ 18). Umbr. hondomu suggests an older -tomo-, -omo-, while Ose. iltiumam
(Lat. ultimam) shows an affection of t that is usual before a u-sound (cf. Ose.
tiurri-, Lat. turris), The form without t is seen in minimus, brima from brévis
for *breghu- (Gk. Bpaxds), (ef. Osc. maimo- ‘maximus’), and apparently
attached to a case-form, in supre-mus, extré-mus, postré-mus (cf. postumus Virg. A.
vi. 763), but it is usually combined with the Comparative -is-, e.g. maximus
for mag-is-imus (Falisc. Maxomo-), O. Lat. oxime (Paul. Fest. 225. 1 Th.) for
*oc-is-ime, medioximus from the stem medioc- seen in médiocris. This was the
formation adopted by Adjective-stems ending in -li-, -ri-, -ro-, e.g. facilis,
Jacil-limus for *facli-simus ; dcer (deri-), acerrimus for *acri-simus; misero-, miser-
rimus*, &e., though we have sévérissimus, matirissimus and maturissime, but
usually maturrime, &c. (see Neue, ii®. pp. 187 sqq.). With the last we should
probably compare O. Lat. piirime in the phrase purime tetinero, explained in
Paul. Fest. 335. 7 Th. as purissime tenuero. Ennius (according to Charisius 83.
22 K.) wrote equitatus celerissimus, and minerrimus is quoted by Paul. Fest. 88.
11 Th.: ‘minerrimus’ pro minimo dixerunt.

§ 55. Some irregular Comparatives and Superlatives. Vetustior appears
as the Comparative of vétus, because vetustus with its ill-sounding repetition of
the syllable -tus- was discarded in the Positive for vetus, apparently the I.-Eur.
Noun *wétos, *wetes- (Gk. éros,a year); in the Superlative veterrimus and
retustissimus (in Livy and later writers) are both found. Minus, with -us not
-ius, has been similarly explained to have originated in a Neuter Noun, meaning
‘the less quantity,’ and to have produced the declension minor M., minor F.,
minus N.; the Oscan equivalent of minor is minstro- (Lat. minister), e.g. ampert
minstreis aeteis eituas moltas moltawm licitud ‘dumtaxat minoris partis pecuniae
multas multare liceto’ on the Law of Bantia (cf. Umbr. mestro- ‘major,’ Lat.
migister). The coexistence of such forms as bénévilus and benevolens (ch. viii.
§ 90) produced a type of Comparison like magnificus, magnificentior, magnificentis-
simus; while fragi, which was a Dative Case of a noun [frugi (bonae) se.

 

1 Tt isa mistake to supposethat the antepenultimate of -issimus is ever
seansion mistrrimus, simillimus, &c. is shortened (Class. Rev. vi. 342).
found in O. Lat. poetry, or that the
408 THE LATIN LANGUAGE. (Chap. VI.

faciendae aptus; cf. Plaut. Pseud. 468 tamen ero frugi bonae; Poen. 892
erus si tuos uolt facere frugem) had recourse for its Comparative and Super-
lative to the Adj. friigalis. The retention of v in the Positive with its suppres-
sion in the other degrees, causes the anomaly in the Comparison of dives,
(but cf. Ter. Adelph. 770 dis quidem esses, Demea), ditior, ‘ditissimus ; jivinis,
Jinior, From plé-, a development of the root pel-, ‘to fill’ (Lat. piz-nus, replé-tus,
plé-ri-que, Gk. wanpys, &e.), were formed Greek mdelwy, mAeioros ; from plé-,
another grade of plé- (ch. iv. § 53), the Latin plas for *plo-is (ploera Cic. Legg.
lil. 3. 6), plirimus for *plois-omo- (ploirume Nom. Pl. Mase. on a Scipio epitaph of
the end of the third cent. B.c., C. I. L. i. 32:

héne oino ploirume coséntiont R{émai]
duénéro éptumo fuise uiro,

‘hune unum plurimi consentiunt Romae bonorum optimum fuisse virum’);
the plous of the 8. C. Bacch. (C. I. L. i. 196. 19 and 20) may with its ov merely
represent the t-sound which the diphthong oi had by this time assumed (ch.
iv. §§ 37, 38), and is hardly sufficient evidence of a formation *plo-us like minus ;
similarly plowrwma on the epitaph of a mime (C. I. L. i. 1297, in dactylic
hexameters :

plouruma que fecit populo soueis gaudia nuges,

‘plurima qui fecit populo suis gaudia nugis’) is a misspelling of ploerwma or
pliruma, The pleoris of the Carmen Arvale (C.J. L. i. 28), a hymn preserved in
a late and wretchedly spelt inscription :

neve luae rue, Marma, sins incurrere in dleores,
neve lue rue, Marmar, sins incurrere in pleoris,
neve lue rue, Marmar, sers incurrere in pleoris,

‘neve luem ruem, Marmar, sinas (siveris?) incurrere in plures,’ may be
a mistake for ploeres, and the plisima quoted from the Carmen Saliare by Festus
(244. 17 Th. ‘plisima’ plurima; but in Varro’s account of the same Carmen
(L. L. vii. 27) plusima is the reading of the MS.] should perhaps be corrected to
ploisuma (ploisoma), though some regard these forms as evidence of Latin deri-
vatives from the root plé-, like Greek mAciwy (mAéwv) and mAcioros. Major is
perhaps best referred to an older *mahtor (cf. O. Ind. maéhiyas-) with h for the
Guttural Aspirate, while magis, maximus show another form of root with the
Guttural Media (ch. iv. § 116), (but ef. Osc. mais for *mahis, maimas for *mahimas,
Umbr. mestru for *mahistro-) ; the relation however of the Italic forms to
Goth. maiza, ‘more’ Adj. (1.-Eur. *ma-is-), O. Ir. mia, md, O. W. moi, all of
which point to a root ending in a long vowel, ma- or m3-, is not perfectly clear.

§56, III. NUMERALS. Of the Latin Cardinal Numbers only
1-3 are inflected, auus -a -um, dud -ae -6, trés -és -ia, not 4 (O. Ind.
catvaras, cétasras, catvari, Gk. récoapes -es -a), nor 5 (O. Ind.
pafica Nom., paficanim Gen., Gk. mévre, Lesh. réunwv Gen.).
The Numeral Adverbs from 5 upwards end in -iens or -iés (on
the spelling, see Brambach, Lat. Orth. p. 269; Neue, ii. p. 335;
the Mon. Ancyranum has -iezs), an ending which is also found
§§ 56,57.) DECLENSION OF NOUNS, ETC. NUMERALS. 409

in ¢otie(n)s, quotie(n)s, and which has been connected with the
ending of O. Ind. kiyant-(Adj.), ‘ how great ?’, fyant- (Adj.), ‘so
great,’ probably I.-Eur. -yént. Umbr. nuvis ‘novies,’ Ose.
pomtis ‘ quinquies ’ seem to show the same ending, with ias weak
grade of y& (ch. iv.§ 51). [On the change of I.-Eur. -nt to -ns in
Lat, and Umbro-Oscan, see ch. iv.§ 105; in late Lat. inscriptions
we often find -is (cf. ch. ii. § 6), e.g. guinquis, Rossi, [. Chr. i.
508, of 402 a.D., sexis, id. i. 530, of 404 a.D., decis C.I. L. xii.
2087, of 559 4.D., also -es (cf. quetus for quiétus, ch. ii. § 149),
e.g. guinques, Rossi, i. 510, of 402 a.D., deces C. I. L. xii. 2086,
of 558 a. D., vices xii. 2187, of 564 a.D.|

Fractions are expressed by divisions of the as (=12 unciae),
e. g. uncia, *one-twelfth, quincuna, * five-twelfths, septuna, ‘ seven-
twelfths,’ dewna, ‘ eleven-twelfths,’ lit. ‘minus an ounce.’ The
I.-Eur. word for ‘half? *sémi- (O. Ind. sami-, Gk. qu,
O. Engl. sim-, whence our ‘ sand-blind’), is in Latin sémt-, the
declinable form sémis, Gen. semissis, &c., being apparently a
compound of sémi- and as with the -yé- of *semyéssis weakened
to % (ch. iii, § 18, p. 188); similarly the -wé- of *centu(m)-essi-,
*décu(m)-essi- is weakened to % in centussi-, decussi-. The origin
of the -us (for -nts) of triens, ‘one-third, seatans, ‘ one-sixth,
quadrans, * one-fourth, dodrans, ‘ three-fourths, deatans, < five-
sixths,’ for *dé-sextans, lit. ‘minus one-sixth,’ is not clear ; bess7- is
usually explained as *dw-essi- (on b- from dw-, see ch. iv. § 71),
but it means not ‘two asses’ but ‘two-thirds of an as’!. The
Adjective for ‘ half’ is di-midius from dis- and medius (ch. i.
§ 18). ‘One and a half’ is sesqui-, usually explained as *séms-
gue, with the same syncope as is seen in sestertius, ‘two and a
half, for sémis-tertius (cf. Germ. drittehalb).

§ 57. One. I.-Eur. *oi-no- (Gk. olvy, an ace, O. Ir. oen, W.
un, Goth. ains, Lith. v-énas, O. SI. i-nii; ef. O. Ind. éka- for
*oi-ko-, Cypr. Gk. of-Fos, Att. ofos, alone, for *oi-wo-), Lat. anus
(-a -um), O. Lat. oinos. Another [.-Eur. word for ‘one’ was

 

1 The Osean word, often compared All that the inscription shows is
with besst-, viz. diasis (Zv. I.I.I. ... jasis.
154), has secant claims to existence.
410 THE LATIN LANGUAGE. [Chap. VI.

*sém-, which is used in Gk., efs for *sem-s, ula for op-ta, év for
*sem, and in Arm., mi for *sm-i; and in Derivatives and
Compounds in all languages, e.g. *sem-, *sm- in O. Ind. sa-ktt,
‘once, Gk. d-mag, a-mAds, &-rAdos, Lat. sin-gitli, simplus, sim-plex,
sémél, sem-pér, sincinia : cantio solitaria, Paul. Fest. 500. 23 Th.,
simpludiarea funera: quibus adhibentur duntaxat ludi, Fest. 498.
24Th., Goth. simlé,‘ once.’ For the Ordinal was used a derivative
from the I.-Eur. root per- (cf. Lat. prd, prae, &e.; Gk. mépuor
for wép-vr1, ‘in the previous year, O. Ind. par-ut, from pér-
and the root of *wétos,‘a year, Gk. éros), in Latin *pris- (cf.
prius) with the suffix -mo-, primus for *pris-mus (Pelign. Prisma- ;
ef. Lat. pris-cus, pris-tinus), in O. Ind. a derivative with one
Superlative suffix, pra-thama-, in Teutonic with another, O. H. G.
fur-ist (cf. Germ. Furst), Engl. first.

The Adverb is in Latin sémé from the root sem-, just
mentioned, a byform of which furnished the Indefinite Pro-
noun ‘any, ‘some,’ in various languages (O. Ind. sama-, Gk.
aud-, Goth. sums, Eng]. some; K. Z, xxxii. 373); the Adjective,
sim-plus (Gk. a-mAés) or sim-plex (from *sem-plax, cf. du-plex
§ 59); the Distributive sin-gidi.

§ 58. Unus. 0. Lat. oinos appears on the proud epitaph of L. Cornelius
Scipio (C.J. L. i. 32), written in Saturnian metre :

héne oino ploirume coséntiont R[émai]
duénéro éptumo fuise uiro,

‘hune unum plurimi consentiunt Romae bonorum optimum fuisse virum ’ ;
cf. oinuorsei ‘universi’ on the 8. C. de Bacch. (i. 196), and Oinumama ‘ Uni-
mamma ’ (an Amazon) on an old Praenestine cista (i. 1501), oiné Adv. in the
Lex Agraria of 111 B.c. (i. 200. 21); oenus in Plaut. Truc. 103, Cic. Legg. iii.
3.9. But we have unus in the Lex Repetundarum of 123-122 B.c. (i. 198).
The Neuter, with the Negative particle nz prefixed, was used as the ordinary
Negative, noenwm (for *ne-oinom), later non (ch, iv. § 16), like our ‘not’ and
‘nought,’ Germ. nicht and Nichts, from Goth. ni waihts (see ch. x. § 18),
while to express ‘nothing’ the Romans used a compound of ne and hilum
.quod grano fabae adhaeret, Paul. Fest. 72. 10 Th.), nihilum, later nihil, nil
(ch. iii. § 52). The plural of wus is found with Nouns whose Plural is used
in a Singular (Collective) sense, e.g. wna custra, and in the sense of ‘ only,’
‘alone’ (ef. Gk. ofos), e. g. tres wnos passus Plaut. In the Romance languages
the Indefinite Article is formed from Lat. wnus, as the Definite from Lat. ie ;
and we see traces of this use in colloquial Latin, ©. g. wna adulescentula, Ter.
alndr, 118.

$59 Two. [-Euv. *dtiwo- and *dwo- with Dual declension,
§§ 58-60.] DECLENSION OF NOUNS, ETC. NUMERALS. 411

*duwo(u) M., *duwai F., *duwoi or *duwei N. (O. Ind. dvat
and dva, older duvati and duvi M., dvé, older duvé F., N., Gk.
dvw and dvo, also 4(F)o-, O. Ir. dau and da M., di F., W. dau M.,
dwy F., Goth. tvai M., tvos F., tva N., Lith. du M., for *dvii,
dvi F. for *dvé, O. Sl. dva and diiva M., dvé and diivé F., N.),
Latin duo M., N., duae F., with Dual declension, which however
became intermixed with Plural forms, e. g. duds Acc. M. beside
duo. In Derivatives and Compounds the I.-Eur. stem dwi-
appears (O. Ind. dvi-pad-, Gk. 8(F)i-7ovs, O. Engl. twi-féte), Lat.
Si-pes, bi-dens (O. Lat. dui-dens, Paul. Fest. 47. 8 Th., ef. dui-
census : cum altero, id est cum filio, census, id. 47. 5; duicensus

.. devrepov amoyeypappevos GI. Philox.), while in the Italic
languages we have also dti- (Lat. du-plus, du-plex, du-centi,
Umbr. du-pursus ‘ bipedibus’), apparently the weak grade of an
I.-Eur. *dwé- (Lat. du-bius has the same root; cf. Hom. doi,
doubt, Germ, Zwei-fel, Zend dvai-di). For the Ordinal the
Romans used sécundus, lit. * following, from séguor, or alter,
‘ the other of two, from the same root as d/-ius,‘the other of
many’ (O. Ir. aile, W. ail, which have also this sense of
‘second’; in O. Engl. oper had this numerical sense); for the
Adverb I.-Eur. *dwis or *dtiwis (O. Ind. dvis, Vedic duvis, Gk.
a(F)is, M. H. G. zwis, Goth. tvis-, ‘apart’; Engl. twis-t), dis (O.
Lat. duis, Paul. Fest. 47. 6 Th. ; cf. duidens ‘ hostia bidens,’ and
duicensus * cum altero, id est cum filio census’ quoted above) ; for
the Adjective dii-plus (Gk. 81-mAds, Umbr. du-plo-) or di-plex
(Gk. d¢-mAa€; cf. Umbr. tu-plak N.); for the Distributive J7-ni
from *dwiz-no- (O. Scand. tvenner) or from *dwi-no- (Lith.
dvynt Du., ‘ twins’).

§ 60. Duo. The original quantity of the final vowel of duo in Latin is
difficult to establish from poetry. We cannot assign much weight to the
precept of the grammarians (e.g. Charisius 35. 25 K.) which distinguishes
duo M. from dud N., nor to the scansion dud in the Christian poets (Neue, ii®.
277; similarly egé for egs). In classical poetry the scansion is invariably dué
(ef. duddéeni), but in the old Republican poets we find duo Acc. distinguished
from duos, not by the quantity of its final syllable, but in being treated as
a monosyllable or the equivalent of a long syllable; for example, duo Acc. is
never allowed to end an iambic line, which points to *dvé rather than to *dié
(A. LL. iii. 551). The shortening of the final vowel can be easily explained
by the Law of Breves Breviantes, which reduced hévé to hdvé (ch. ili. § 42) ; for
a similar doubt with reference to the pronunciation of scio as scid or *s¢jé, see
412 THE LATIN LANGUAGE. (Chap. VI.

ch. ii. § 151). The Fem. duae may retain the old suffix of the Nom. Dual of
A-stems, -ai (e.g. O. Ind. Agve for L-Eur. *ikwai, ‘two mares’), which would
be retained in monosyllabic *dvai without sinking, as in the unaccented
syllable, to -7 (e. g. occido from éc-cuido, ch. ili. § 18). The termination -dbus of
duobus M., N. is shared only by the other Dual-form ambo, though -dbus F. was
a common (Dat., Abl. or Instr.) Plural ending of A-stems, retained in legal
language especially in the words deabus, filiabus, libertabus (§ 48). But the
Plural declension encroached more and more on these Dual forms; diids
Acc. M. competes, as we have seen, in the older literature with dud, while
duas F, is a Plural, as are also the Genitive forms duorum M., N. (older duwm,
e. g. duumvirum used by Cicero, § 47), duarwn F. A Nom., Acc. Neuter dua
appears on inscriptions (e.g. C. I. L. v. 1102; other instances in Neue, ii*. p.
277), though it is called a barbarism by Quintilian (i. 5.15 nam ‘dua’ et ‘tre’
diversorum generum sunt barbarismi, at ‘duapondo’ et ‘trepondo’ usque ad
nostram aetatem ab omnibus dictum est, et recte dici Messala confirmat),
which seems to have been supplemented by a Nom. Mase. *dui in Vulgar Latin,
to judge from the Romance forms (e.g. Ital. due, older dui, O. Fr. dui, doi, &c.).
In Umbrian the word shows Plural declension, dur Nom., tuf Ace., tuva Nom.
Ace. Neut. (On the declension of Lat. duo, see Neue, ii*. pp. 276 sqq.). Late
compounds like diloris, dinummium are hybrid formations with Greek 6i-
instead of Latin bi-. The relation of I.-Eur. *dwi- to the Latin preposition
dis-, apart, has not been established, nor yet to I.-Eur, *wi- of Dor. Gk. Fi-«art,
Lat. vi-ginti, &c., perhaps connected with O. Ind. vi, ‘apart.’

§ 61. Three. I.-Eur, *tri-, Nom. Masc. *tréyés (O. Ind. tréyas,
Gk. rpeis, Cret. Gk. rpées, O. Ir. tri, W. tri, Goth. preis, Lith.
trys, O. SI. trije), Lat. #és M., F., tia N. (cf. Umbr. trif or
tref Acc., triia Neut.). The stem tri- appears in Gk, rpi-ros, rpi-
nous, Lat. tri-pes, &c., but a stem tré- in Lat. tré-centi, tre-pondo,
Lith. tré-czias, ‘third, O. SI. tre-tiji, &. The Ordinal sértins
(Umbr. tertio-) probably shows this stem tré- with metathesis of
rt; the Adverb ¢é, for ¢err (in Plautus scanned as a long syllable,
cf. ¢err-uncius) from *ter-s, comes from the same stem, or, like
O. Ind. tris, Gk. rpis, from the stem tri- (ch. iii. § 15. 8); the
Adjective is tri-plus (Gk. rpi-mdds), tri-plew; the Distributive
trinus (see on himus above), and ter-nus. We find ¢ri- and ter-
interchanged in Compounds like ¢ri-géminus and ter-geminus,
tri-vénéficus and ter-veneficus, tri-vium and ter-vium (C.I.L. ix.
2476), Lerventum now Trivento; éré- and dré- in tri-média and
tre-modia (Varro, Men. 310 B.).

§ 62. Tres. The grammarians prescribe 2s in the Nominative and tris in
the Accusative (Neue, ii, p. 284), as is the rule in all I-stems (§§ 40, 51),
though Virgil, as Gellius (xiii. 21. 10) points out, uses tres Acc. for the sake of
variety in A. x. 350:
§§ 61-63.] DECLENSION OF NOUNS, ETC. NUMERALS. 413

tres quoque Threicios Boreae de gente suprema
et tris, quos Idas pater et patria Ismara mittit,
per varios sternit casus,

as he uses elsewhere urbes Acc. in a context where the Roman erities found
that this form gave greater melody to the line (4A. iii. 106): centum urbes
habitant magnas (Gellius J. c. quotes the remarks of Probus on this form : hic
item muta ut ‘urbis’ dicas, nimis exilis vox erit et exsanguis, and his reply
toa caviller: noli igitur laborare, utrum istorum debeas dicere ‘urbis’ an
‘urbes.’ Nam cum id genus sis, quod video, ut sine jactura tua peeces, nihil
perdes, utrum dixeris!). Tris Nom. is found on late inscriptions, e. g. Eph.
Epigr. iv. 420 (other examples in A. L. L. vii. 65).

§ 63. Four. The I.-Eur. stem q*étwer- had various grades,
q®etwor-, q*etur-, qetru-, &c. Its Nom. Masc. *q*etwores (O. Ind.
catvaras, Dor. Gk. réropes, Att. rétrapes, Lesh. méoupes, O. Ir.
cethir, W pedwar, Goth. fidvor, O. Engl. féower; cf. Lith.
keturi, O. SI. Getyre) probably appears in Lat. qguattuor (some
make this an I.-Eur. Neuter *q*etw6r, §§ 44,45), apparently for
*guotvor(é)s, with ¢ doubled before the w-sound (ch. ii. § 130) and
-atv~ for -otv- like -av- for -ov- in cdévus, &e. (ch. iv. § 19). The
Oscan word was petora (so spelt by Festus 250. 33 Th., but
probably better *petoro). In Compounds and Derivatives we
find a stem q*etru- (Zend capru-, Gaul. Petru-corius; cf. Umbr.
petur-pursus ‘ quadripedibus’), which in Latin is guadru- (with
a& for @ by analogy of guattuor, quartus?), where the d is
puzzling, for -dr- seems to become -é7- in Latin, e. g. dtro- for
*ddro-, nitrix for *niudria (ch. iv. § 113), so that -tr- should not
change to -dr- [ Wharton, Htym. Lat. p. 83 suggests that quadra, a
square, means literally ‘ pointed’ and comes, not from quattuor,
but from a root q#ad-, ‘to point, sharpen,’ Engl. whet, the usual
change of -dr- to -¢7- being seen in ¢r7-quetrus, triangular.
Quadra, which retained d (by analogy of other words from
the same root?), may have been the cause of *q*etru- taking
the form guadru-; cf. K. Z, xxxii. 565]. The Ordinal qguartus
with its long @ (indicated by an apex over the letter on
inscriptions; see Christiansen, de Apicibus, p. 52) has not yet been
satisfactorily explained. On an inscription of Praeneste we
have Quorta (cf. ch. ii. § 4); the Oscan word truto- is inter-
preted variously as ‘ quartus’ for *ptru-to-, and as ‘certus.’ The
Adverb qudétér stands for *q*etrii-s (Zend capru’; cf. O. Ind.
eatar), as ager for *agrds, dcer for *deris (ch. in. § 16); the
414 THE LATIN LANGUAGE. (Chap. VI.

Adjectives guadri-plus, quadit-pler show the curious change of
-tr- to -dr-, remarked on above, from which quater is free; the
Distributive gudternus for *quatri-nus (like sdcerdos for *sacridés,
ch. iii. § 16) from *q*etrii-no- (ch. i. § 18), or from the Adverb
quater with the suffix -no (see above on dinus, § 59).

§ 64. Quattuor with double ¢ is the spelling of the best MSS. and in-
scriptions, such as the Monumentum Ancyranum (see Georges, Lex. Worlf. s.v.).
The form quattor, found on late inscriptions (e.g. C.I. L. viii. 5843; other
examples in A. L. L. vii. 65) has been already explained from *quatt(v)drdecim,
where the w-sound would be dropped before the accent, asin Jan(v)drius, Feb-
r(v)drius, &e. (ch. ii. § 54). Vulg. Lat. *guattor-decim is indicated by the Romance
words for ‘ fourteen,’ e. g. Ital. quattordici, Fr. quatorze, but Vulg. Lat. *quat-
tro by the words for ‘ four,’ e.g. Ital. quattro, Fr. quatre (Sic. battor, how-
ever, from quattor). This Late and Vulg. Latin quattor can hardly be assigned
to early Latin authors; so retain quattuor in Plaut. Most. 630 and scan qualtudr,
like enictis, Rud. 944 as a dactyl at the beginning of an iambic line ; in Enn. 4.
90 M. quattudr like virginés A. 102;. in Enn, A. 609 read feré quattuér partum (?).
Petreius, Petronius are dialectal Proper Names derived from this numeral, as
Pompeius, Pontius (Lat. Quintius) from the numeral ‘ five.’

§ 65. Five. I[.-Eur. *pénq*é (O. Ind. péfica, Arm. hing, Gk.
mévte, TeuT-OPodov, Lith. penki) shows in Teutonic assimilation
of the second syllable to the first, *pempe (Goth. fimf), in Lat.
and Celtic of the first to the second *q%enq"e (Gaul. reume-dovda,
quinquefoil, O. W. pimp ; in O. Ir. with o for e in the first syllable,
cdic; ef. Umbro-Oscan pump-, Pompeius), Lat. quingue with e
changed to 7 before a nasal and guttural (like ¢émgo for *tengo,
ch. iv. § 11), and the i lengthened (by analogy of guintus?). The
Ordinal guintus, older guinctws, follows the rules of Latin phonetics
that gu becomes ¢ before a consonant (ch. iv. § 137), and that in
this group of three consonants the middle one is dropped (ch. iv.
§ 157). On the lengthening of the 7, see ch. i. §. 144. The original
form was *q*enc-to- in Italic for I.-Eur. *peng?-to (or *pnq?-to-?)
(Gk. réurros, Lith. pefiktas, O.Sl. peti; Osc. Pintiis, also Towmrves
with -mp- from the Cardinal form), The Adverb is quinquies,
older guinguiens, on which see § 56, the Adjective quinqui-plex [or
with -cu- for -quu- (ch. iv. § 137) quincit-plea], rarely quincuplus or
quinguiplus ; the Distributive guinus for *guine-nus (ch. iv. § 157).

§ 66. Quinque. The long quantity of the i of quingue is attested by the
use of the tall form of the letter on inscriptions (for instances, see Christiansen,

de Apicibus, pp. 45-46. On Vulg. Lat. cinque (Ital. cinque, Fr. cing, &c.),
see ch. iv. § 163).
§§ 64-70.] DECLENSION OF NOUNS, ETC. NUMERALS. 415

§ 67. Six. Lat. sew points to I.-Eur. *séks or *hstks (7.8. L.
vil. 73), as Greek é&, Féé to *sweks or *kswelks (cf. Zend x$va8, O.Ir.
sé and *fé,W. chwech). The Ordinal sewtus has -st- in Umbro-
Oscan, e.g. Umbr. sestentasiaru ; Sestivs is a dialectal byform
of Seatins. The Adverb is sewiés, older sewiens (see § 56); the
Distributive séeus for *séa-nus (as lina for liw-na, ch. iv. § 162).

§ 68. Seven. I.-Eur. *séptm (O. Ind. sapté, Arm. evt‘n,
Gk. énra, O. Ir. secht, W. saith, Goth. sibun ; cf. Lith. septyn-i),
Lat. septem, with Ordinal *septmo- (O. Ind. saptamé-, Gk.
€Bdouos, Pruss. septmas and sepmas), Lat. septimus, older septu-
mus. The Distributive septénus stands for *septen-no- (ch.ii.§ 130).

In later Latin we find ¢ for pt, e.g. Settembris (C.I. L. xi.
288 5, Setebres 4075); cf. Ital. sette, Settembre. In Compounds
we find the stem seplem-, e.g. septempédalis Plaut., septemplex
Virg., septemgéminus Catull., but Septi-montium, and on the
restored Columna Rostrata septe-resmos (C.I. L. i. 195); septu-
ennis Plaut., &c. (cf. septud-gintd, below), for *septum-ennis,
shows the usual weakening of unaccented é before m to w, and
the same loss between vowels of final m of the first member of
a compound as is seen in c7reu(m)ire, &c. (ch. iii. §§ 18 and 52).

§ 69. Bight. The O.-Ind. Dual form *dkté(u), cleverly
explained by Fick as ‘the two sets of pointed’ (i.e. the fingers,
without the thumbs, of both hands), from the root ak-, ok-, ‘to
be sharp, pointed’ (O. Ind. a&ta and a&tad, Arm. ut’, Gk. éxrd,
O. Ir. ocht, W. wyth, Goth. ahtau, Lith. asti-n-i) is in Latin
octo, with Ordinal octavus (cf. Ose. Uhtavis ‘ Octavius’) for
*octovus (ch. iv. § 16), and Distributive octd-nus. In later Latin ef
became #4, Ottobres C.I. LZ. xi. 2537 (cf. Ital. otto, Ottobre), and
the final -6 is shortened, like every final -é in the poetry of the
Empire, e. g. octd Mart. vii. 53.10. In Compounds it is treated as
an O-stem, e. g. octii-plus, octi-plex, octi-pes, oct-ennis,

§ 70. Nine. I.-Eur. *néwn (O. Ind. nava, Gk. évvéa for
*éy-veFa (?), O. Ir. ndi, W. naw, Goth. niun; cf. Lith. devyn-i
with d- instead of n-) is in Latin névem with ov regularly enough
for -ew- as in Celtic (ch. iv. § 10), but with -em instead of the
416 THE LATIN LANGUAGE. [Chap. VI.

normal -en. The proper Nasal appears in the Ordinal ndnus,
older xoino- (if the Dvenos inscription is rightly read dienoine,
on the ninth day), from *ndvéno-, where the 6 is difficult to
explain, though it seems to have a parallel in xdu for noen(wm),
Failing this explanation, we may suppose that *nouno- or
*nownno- was the J.-Eur. form, with the O-grade of the root,
and understand Lat. 6 as an expression of I.-Eur. ou as in 7dbus,
&e. (ch. iv. § 41). (On ndndinum and noundinum, old spellings
of néndinum, for *no(v)éndinum, from novem and the root din-,
seen in O. SI. dini, ‘a day, see ch. iv. § 44.) Umbr. nuvimo-
shows the form rather to be expected in Latin, *ndvimo-
(O. Ind. navamé-), but in Pelign. we have Novnis ‘ Nonius,’
The Adverb is novés (Umbr. nuvis); the Distributive novenus
from *noven-no- (ch. ii. § 130).

§ 71. Ten. I.-Eur. *dékm (O. Ind. daa, Arm. tasn, Gk. dé¢xa,
O. Ir. deich, W. deg, Goth. taihun; cf. Lith. dészim-t, O. SI.
dese-ti), Lat. décem, with Ordinal déetmus (O. Ind. daSam-) (cf.
the proper name Decius), Adverb deciés, and Distributive dénus
for which we should expect *decénus like septénus (cf. § 74).

§ 72. Eleven to Nineteen. These Numerals were denoted
in I.-Eur. by Compounds, expressive of the addition of the
smaller unit to ten. These Compounds consist of the two
Numerals themselves (not their stems) placed together, the
smaller unit preceding the ten, e.g. I-Eur. *tréyés-dékm,
‘thirteen ’ (O. Ind. tr4yé-daga), The Latin Compounds (in which
the final -em of decem sinks to -im, ch. ii. § 18) are, wn-decem
for *édn(i)-decim (on Vulg. Lat. *indecim, see ch. il. § 147), duo-
decim, tré-decim for *trés-decim (like nidus for *nis-dus, ch. iv.
§151), quattuor-decim, quin-decim for *quin(que)-decim (ch.ii1.§ 13),
sé-decim (the correct spelling) for sex-decim (like lina for *lia-na,
ch. iv. § 162), septem-decim, Octé-decim and novem-decim were re-
placed by duo-dé-viginti, un-dé-viginti for *un(i)-de-viginti, or by
octo et decem, decem novem, a mode of expression which is found
in the other numerals too, e. g. decem duo (Umbr. desen-duf Acc.),
and which is used in Greek exclusively for the numbers above
twelve, e.g. tpeis kal déxa and déka rpeis, and optionally for
§§ 71-74.] DECLENSION OF NOUNS, ETC. NUMERALS. 417

twelve, 36-dexa or déxa d5¥0. In I.-Eur. it was used for numbers
above twenty, and so in Lat. guattuor et viginti or viginti quat-
tuor, &e. The Ordinals are wndecimus, duodecimus (O. Ind.
dvidagama-), &e.; the Adverbs wnzdecies, duodecies, the Dis-
tributives wnudénus, dwddénus, and so on.

§ 73. O. Lat. duovicesimus for class. duo et victsimus, twenty-second, is
attested by an interesting chapter of the Noctes Atticae (v. 4), which illustrates
the pains taken in the Imperial period to secure correct texts of early authors.
Gellius there tells us of a MS. of the Annals of Fabius: bonae atque sincerae
vetustatis libri, quos venditor sine mendis esse contendebat. A grammaticus
who was asked to inspect the MS., on the absolute correctness of which the
bookseller was willing to stake any amount of money (grammaticus quispiam
de nobilioribus, ab emptore ad spectandos libros adhibitus, repperisse se unum
in libro mendum dicebat ; sed contra librarius in quodvis pignus vocabat, si
in una uspiam littera delictum esset), declared that duovicesimo anno in Book iv.
was a mistake of the copyist for duo et vicesimo anno, but was finally forced to
admit, on being referred to other passages of ancient authors, that duovicesimus
was a genuine Old Latin form.

§ 74. Twenty to Ninety. These Numerals are denoted in
I.-Eur. by Neuter Compounds, ‘two decades, ‘three decades,’
&e., the word for ‘ decade’ being *dekm-t-, changed in Compo-
sition into *(d)kmt- [or *(d)komt-? ; ef. Greek -xovr- in 30-90,
Tptd-KovTa, Tecocapd-Kovra, or Tetpé-Kovta, &c., which suggests
that this is the plural stem, the other the dual]. In Lat. vi-ginti
vi- is probably Neut. Dual Nom. of L-Eur. *wi-; the -giné#,
with 7 for ¢ owing to the popular accentuation *vigenti, *trigenta,
*quadrdgenta, &e., which takes its y apparently through influence
of the d of I.-Eur. *(d)kimt- (but vicesimus; so Alb. -zet, ‘a gross,’
points to & not k), is also Neut. Dual Nom. In ¢ri-ginta
the frz- is Neut. Plur. Nom. of the I.-Eur. stem *tri-, the -ginta
shows the original quantity of the Neut. Plur. suffix, which has
by the time of the earliest literature been shortened in Nouns,
&e. to -% (ch. iii, § 43). In quadra-ginta the quadra- (on the
form of the stem, see § 63) may correspond to Gk. rerpw- of
rerpé-xovra (see ch. iv. § 92),or may have the Neut. Plur. ending
-d,an ending assumed by quingue, sex, septem, novem in a somewhat
haphazard way in the Compounds quingud-ginta, sexd-ginta,
septud-ginta (for *septu(m)ayinta ; see above on septuennis, § 68),
nond-ginta, while octé in octoginta retains its ordinary form. The
Ordinals are formed with the suffix -tmo- or -t*mo- (ch. v. § 14),

be
418 THE LATIN LANGUAGE. [Chap. VI.

e.g. vicésimus, older ricensumus (this spelling is more frequent
than vigesimus) for *vi-cent-tumus, tricésimus and tr —
older -ensumus for *tri-cent-tumus (O. Ind. tri- Sat- tama-); the
Adverbs are viciés, older -ens, triciés, older -ens ae
trigies), like decies, older -ens (§ 56); the Distributives vicénus
(with byform vigénus), tricénus (with byform trigénus), &c., not
like the abnormal /énus, but like *decénuws (§ 71).

§ 75. Viginti, &c. Viginti is spelt veiginti (C. I. L. i. 1194 ; x. 6009), but the
ei may merely indicate 7, for the inscriptions are not of great antiquity (ch. i.
§ 9). The late spelling vigenti (C. I. L. v. 1645, &c.) points to viginti (cf. quinqua-
genta xii. 482, and other examples of -genia quoted in A. L. L. vii. 69-70). In
Late and Vulgar Latin the g was dropped (ch. ii. § 94), whence the form
vinti; e.g. Wilm. 569:

et menses septem diébus cum vinti duobus;

ef. Sard. vinti, Ital. venti, &e. The same thing happened to triginta, producing
the form trienta (C.I.L. xii. 5399, &c.), and in Romance, Sard. trinta, Ital.
trenta, &.; also to quadraginta (a hexameter line on a late epitaph ends,
quadraginta per annos, vi. 28047), which had become quar(r)aginta (Fabretti,
iv. 134), whence Ital. quaranta, Fr. quarante. On the accentuation of these
three numerals, see ch. iii. § 11. 4. Septudginté is the scansion required in
metrical epitaphs (C. I. L. vi. 22251. 29426). Its analogy produced in Mediaeval
Latin the form octuaginta, which found its way into some early editions of
Latin authors (Skutsch, Forsch. i. 24). Octaginta occurs sometimes in the Edict
of Diocletian (C. J. L. iii. pp. 810, 811) and elsewhere (see A. L. L. vii. 70). (For
other examples of the scansion -gintdi in late poetry, see Neue, ii*. p. 2go.)

§ 76. The Hundreds. The I.-Eur. expression for 100 was
apparently ‘ a decade of decades’ (like the Compound later
evolved in Gothic, taihunté-hund (?)), viz. *(d)kmtém, probably
an old Gen. Plur. of the stem *dekmt (§ 46), treated as a Nom.,
like sestertium, -i Gen. for mille sestertium, ‘a thousand of
sesterces, in Latin. This explains why the full form, and not
the mere stem, is used in Compounds in Greek and Latin (e. g.
éxatou-Bn, ExaTdy-xeipos, centum-plex, centum-geminus, centum-
pondium); though some prefer to regard *(d)kmtim as Nom.
Sg. of a Neut. O-stem,‘a decade, just as O. Ind, dagati- means,
(4) a decade, (2) a hundred, i.e. ‘a decade (of decades’), The
d of *(d)kmtém shows its influence in Latin in the presence of
gy, instead of, or along with c, in the expressions for the various
hundreds. In O. Lat. we find Neuter Compounds, diicentum
(with di- the weak stem of J.-Eur. dwé-, a byform of dwi-,
§ 59), tré-centum (with I.-Eur. tré-, a byform of tri-, § 61),
§§ 75-78.] DECLENSION OF NOUNS, ETC. NUMERALS. 419

nongentum, &c., used with the Genitive of the thing specified,
e.g. argenti sescentum, Lucil. xxx. 22 M.; but these Neuter
Compounds, when referring not to a mass, but to a number of
individual things, became in course of time declined as Adjec-
tives, dii-centi -ae a, tré-centi -ae -a, e.g. trecentae causae Plaut.
(cf. nongéentus, Plin. xxxili. 2.31). (So in Greek YO-stem Adjec-
tives were formed, in Att. -kdov01(with -o1- for -r1-), in Dor. -xdrv01.)
They formed their Numeral Adverbs, Adjectives, and Distribu-
tives by analogy of the tens, e.g. ¢récent-ies (like tricies), trécent-
esimus (like tricesimus), trécent-éni (like triceni). The forms guin-
genti (for *guinc-genti, ch. iv. § 157), septin-genti (for *septem-genti,
*septen-genti like tingo for *tengo, ch. iv. §11),apparently influenced
quadrin-genti (for earlier guadri-genti), octin-genti, and even nonin-
genti (beside the usual non-genti). Ses-centi for *se(c)s-centi
obeys the same phonetic law as disco for *di(c)-sco (cf. di-die-i ;
see ch. iv. § 157. 3).

§ 77. Centum, &e. The stem *centd- (-%) appears in centi-ceps Hor., centi-
minus Hor., centi-péda (and centwm-peda) Plin., &c., but centom- in centumpondium
Plaut., Cato, centumplex Plaut. Pers. 560, centum-géminus Virg., &c., so that the
latter is the older formation. For examples of ducentum, &c., with Gen., from
the older writers and in legal phraseology, see Neue, ii®. p. 298. The usage is
confined to phrases like ducentum auri, argenti, vini, &c.; Plautus has ducenti
-ae -a, &e., asin classical Latin. A transitional construction is seen in C. I. L.
iv. 1136 nongentum tabernae, nine hundred shops. As to the form of the several
numerals ; duocenti is found in the late Latin of the Itala; quadrigenti is the
Plautine form, though the MSS. have usually (not, however, universally in
Bacch. 1183) changed it to quadringenti ; for the second syllable is always short,
and the Breves Breviantes Law is inoperative when a Mute and Liquid follow
the short vowel, so that -dri- not -drin- must have been the second syllable of
the word used in Plautus’ day ; quadrigenti, quadrigenus, &c. are probably also
the classical forms (Neue, ii®. p. 297), but on Mon. Ancyr. iii. 8 quadringenos ; of
quingentum Festus tells us that the pronunciation before his time was quincentum
(Fest. 338. 19 Th. ‘quincentum’ et producta prima syllaba, et per ¢ litteram
usurpant antiqui, quod postea levius visum est, ita ut nune dicimus, pronun-
tiari) ; sescenti is the correct spelling, not sexcenti (see Neue, ii®. p. 297); septigenti
for septingenti occurs on the Edict of Diocletian ; noningenti is a late and rare
form (see Neue, J. c.); Caper censures the form noncenti (104. : K. nongentos
non ‘noncentos’ dicendum est), and it seems to be the rule that after n the
voiced (g), and not the unvoiced (c) guttural is used in these words.

§ 78. The Thousands. The Latin mi//e, in O. Lat. a declin-
able Neuter Noun taking a Genitive of the things specified, e.g.
mille hominum oceiditur, milli (Abl.) passum vicerit, is probably

Ee2
420 THE LATIN LANGUAGE.

unconnected with the Gk. yidwou, Lesb. yéAAsor, from *Shéslio-
Adj., and the O. Ind. sa-hdsram from *sm-&héslom, a Neuter
Noun-compound, of which the first part is the root sem-, * one,’
literally ‘one thousand,’ as Gk. éxaréy for *é-xardv (?), ‘ one
hundred,’ ‘ one (decade) of decades (?).’ [Some suppose that it
represents an original *s-esli, ‘one thousand, and that this
became *elle, *méle, as *quaslus (ck. gudsilius) became gudlus
(older guadlus, ch.iv. § 162); the Plural *mélia becoming milia as
*Plénius became Plinius (ch.iv.§ 7). O. Ir. mile would then be
a loanword from Latin, for I.-Eur. sm- would remain in Ivish.|
It is cognate with the Celtic word for ‘thousand ’ (O, Ir. mile,
W. mil), just as the Teutonic and Slavonic numerals are cognate,
Goth. pisundi F., Lith. tikstantis, O. Sl. tyseSta or tysasta F.

For the thousands, the units dwo, tria, &e. are prefixed in Latin
as separate words, duo milia (on the spelling mlza beside mile,
see ch, ii. § 127), tria milia, &e. The Ordinals and Adverbs are
formed like those of the hundreds, mi//-ésimus, miliés, &c., like
cent-ésimus, cent-iés.

§ 79. Mille. Gellius (i. 16) quotes a number of passages to show that Cicero,
as well as the older writers of the Republic, used mille as a Neuter Singular
Noun, e. g. Cic. Mil. 53 mille hominum versabatur. So mille passuwm (earlier
passum), a mile. [Other examples are given by Neue, ii*. p. 303, to which add
Plaut. Bacch. 928 milli (MSS. mille) cum numero ndéuium]. In the spelling
meilia (probably for meillia, for double consonants are written single on this
inscription, e. g. redidei, tabelarios) of C.I.L. i. 551 (Lucania, 132 B.c.), the e
may indicate merely the long i-sound. Lucilius (ix. 21 M.) seems also to
recommend ¢ in both singular and plural (cf. ch. i. § 9) (so meitlle in the
Ambrosian Palimpsest, Plaut. Stich. 587).

§ 80. The Numerals in Romance. The Cardinals are mostly retained,
though dicenti, &c. have become Fr. deux cents, Span, doscientos, &. But
only Italian keeps the Ordinals unchanged, primo, secondo, terzo, quarto,
&c. French has premier for primus, and for the others uses the suffix -iame.
Spanish has primero (like French) for primus, tercero (with the same suffix)
for tertius, and uses for ndnus noveno (the Lat. Distributive névenus), and
similarly for décimus deceno. (For a fuller account see Meyer-Liibke, Rom.

Gram. ii. pp. 590 sqq.)
CHAPTER VIL

THE PRONOUNS.

$1. I. THE PERSONAL PRONOUNS AND THE RE-
FLEXIVE. 1. Sing. Latin ¢3, O. Lat. 290 represents I.-Eur.
*866 (Gk. éyd), of which byforms were *egd (Goth. ik, Lith. esz
and asz), *egom (Gk. éydv), *egdm (O. SI. azti; ef. O. Ind. ahém
with Aspirate instead of Media). It is strengthened by the addi-
tion of the particle -mét, e. g. egomet, mihimet, and in certain of its
eases by the particle -pré(§ 20), e. g. mahipte, while in the Accusa-
tive we find the Pronoun doubled for emphasis, mémé. For the
Genitive, mez is used, apparently the Gen. Sg. Neut. of the Posses-
sive, ‘of mine,’ but in O. Lat. we have mits (with the Gen. és-
suffix ?); for the Dative mthi (mi) for *mehei or *mehoi, with the
I.-Eur. Locative ending (Umbr. mee; of. O. Ind. m&hy-am) and
with mt- instead of mé- because of the Pronoun’s want of accent
(ch. ui. § 18); for the Accusative O. Lat. méd, by the time of
Terence always mé, owing to the Latin phonetic law that final @
was lost after a long vowel (ch. i. § 137); this 4 is either the
Ablative d, with a strange confusion between Acc. and Abl.
functions (so in Engl. ‘him’ Dat. has become Acc.), or the I.-
Eur. particle -id, often used in the Veda to strengthen Pronouns,
*mé-id (cf. Dor. Gk. éuel Ace. for *eue-id ?), just as the particle
*gé is used to distinguish the Acc. in Teutonic (Goth. mi-k,
Germ. mich; cf. Gk. éué-ye); for the Ablative, O. Lat. méd
from I.-Eur. *méd (cf. O. Ind. mad from I.-Eur. *méd), which
with méd Acc., became mé in the second cent. B.c.; the Locative
[I.-Eur. *mei or *moi, O. Ind. (Vedic) mé Loc., used also as
Gen. and Dat., Gk. oi Dat.] and Instrumental (perhaps I.-Eur.
*mé or *md) cannot be identified in Latin. On the affix -gh-
of mihi, see ch. x. § 1.
422 THE LATIN LANGUAGE. (Chap. VII.

§ 2. Egd is the invariable scansion of classical poetry, and the almost
invariable scansion of Plautus and the early Dramatists; egd is found occa-
sionally in late poetry (Neue, ii®. p. 346), where it may possibly be influenced
by Gk. éyd, and in Plautus (e. g. Poen. 1185), &c., where it must be a relic of
the older quantity (Klotz, Altrim. Metrik, p. 51; Miiller, Plaut. Pros. p. 30), not
yet shortened by the Breves Breviantes Law (ch. iii. § 42). The preponder-
ance of the shortened form is due to the enclitic character of the word (even
egimet), for Lat. eg’ cannot represent I.-Eur. *e85, if final 6 became é in
Latin (ch. iii. § 37). Miéhipte is quoted from Cato by Fest. 144. 11 Th.; Paul.
Fest. 145. 5; mepte occurs in Plaut. Men. 1059:

quin certissumumst
mépte potius fieri seruom, quim te umquam emittdém manu.

For examples of meme (a somewhat doubtful form), see Neue, ii’, p. 355. Mis
Gen. is quoted by Priscian from Ennius (A. 145 M.):

ingens cura mis cum concordibus aequiperare,

and probably occurs in Plaut. in Poen, 1188, beginning: rebus mis agundis
(anapaestic), and in other passages, though the MSS. have usually changed
it to the more familiar mei. It is often mentioned as an O. Lat. form by the
grammarians (see the references in Neue, ii*. p. 347), and may represent an
original *més, as Salutis is in O. Lat. Salutes (ch. vi. § 17). For the Dative, we
have on old inscriptions mihei (C. I. L. i. 1016. 1277), and with e to express
the ei-diphthong (ef. ch. iv. § 32) mihe (i. 1049). A still older form mehe (cf.
mehi Plaut. Men. 925 (P.)) seems to be mentioned by Quintilian (i. 5. 21 nam
‘mehe’ pro mi apud antiquos tragoediarum praecipue seriptores in veteribus
libris invenimus), though, if we retain the MSS. reading ‘ pro me,’ we must
suppose mehe to be a mere graphic expression of mz, as in Umbrian a long
vowel is indicated by repeating it and inserting h, e. g. comohota ‘ commaota’
(whether this was a Latin practice is doubtful, see ch. ii. § 56). The final vowel
of mihi, &. had quite become a short vowel in Quintilian’s time, and is
described by him in the phrase commonly applied to 7, as ‘a sound between
e and i’ (see ch. ii. § 16); and even in Plautus and the early Dramatists mihi
is the usual scansion (Leppermann, De correptione, p. 9). Mihi is often contracted
to mi, as nihil to nil (ch. ii. § 58), so that we have side by side in Plautus
divergent treatments of this Dative, such as (1) dissyllabic mihi, e. g. Truc. 77
Nam mihi haec méretrix, &., (2) the same with elision, e.g. Stich. 427 mih(i)
expedi, (3) monosyllabic mi, ve. g. (elided) Truc. 173 Sunt m(i) étiam. (For
instances of mi, see Neue, ii*. 349, and on Nigidius’ distinction of Gen. and
Dat., Gell. xiii. 26.)

The remark of Festus (156. 6 Th.) ‘me’ pro mihi dicebant antiqui, illustrated
by «a quotation from Lucilius (inc. 98 M.): quae res me impendet, means, of
course, merely that in O. Lat. another case was used (in this example, the
Accusative ; cf. Lucr. i. 326 mare quae impendent saxa), where the classical
construction had the Dative; and similarly vae te Plaut. Asin. 481 (cf. vae me
Seneca, Apoc. 4) is probably nothing but an unusual employment of the Accu-
sative. The Vocative of the Possessive Pronoun mi, e.g. mi fili, mi vir, has
been regarded as a Locative-Dative by some and compared with Homer's
pyrépt pot, the Tragedians’ & yivar po, Vedic mé giras, ‘my hymns.’ But it is
more likely to be a byform of *mie (with unaccented e sunk to i, ch. iii. § 18),
§§ 2-4.] THE PRONOUNS. PERSONAL. 423

as fili may be of filie (but see ch. vi. § 31), for it is always joined with a Voc.,
and is used in good writers only with a masculine noun (with a Fem. after
Apuleius, Neue, ii®. p. 368) ; and this is the explanation given by the native
grammarians [see below, § 12; there is no *#i (Gk. ro, cor) nor *si (Gk. of)].
The Accusative med is found on the very earliest Latin inscriptions which we
possess, the Praenestine fibula (C. I. L. xiv. 4123), with Manios med fefaked Numa-
sioi, and the Dvenos inscription (Zvet. I. 1.1. 285, Rome), with Dvenos med feked
(or feced) ; and in Plautus méd and mé are doublets, in the Acc. as in the Abl.,
mé often being shortened by prosodical hiatus to mé before an initial vowel,
while in Terence mz, &c. Acc., Abl. has ousted méd, &e. Whether Plautine mz
Abl. is ever an Instrumental form, and has not lost a final-d, it is impossible
to say ; but there is no indication of its being anything but a phonetic variation
of an original méd (cf. ch. ii. § 137). On mé-quidem Plaut., see ch. iii. § 51.

§ 3. 2 Sing. Lat. ¢# is from I.-Eur. *ti (Hom. Gk. ri-vy, O. Iv.
ti, O. H. G. da, O. SI. ty), as Dor. Gk. rd from I.-Eur. *tti, In
the oblique cases the I.-Eur. stem was *twé (or *tw3d) and *té (or
*t5), &e. [e.g. Gk. o€ for Cret. rFé, O. Ind. (Ved.) tva Instr.,
tvé Loc., té Dat., Gen.], often enlarged by an affix -bh-, like the
-&h- of the 1 Sing. Pron. (e. g. O. Ind. téi-bhyam Dat.), For the
Genitive in Latin the Gen. Sg. Neut. of the Possessive is used,
twit, * of thine’ (see above, on mez), and a Gen. tis (? ¢¢s) appears
in O. Lat.; for the Dative, ¢i07, older tibed, for *tebhei (Umbr.
tefe ; cf. Pruss. tebbei), with 7%- for ¢e- owing to its unaccented
character; for the Acc., as for the Abl., O. Lat. ¢éd; in the
earliest literature ¢éd and ¢é appear as ‘doublets’ (like méd and
mé, § 2), but by the middle of the second cent. B.c. ¢éd is out of
use ; the Locative and Instrumental cannot be identified in Latin.
A strengthened form of the Nom. is ¢i#-/é, of the Acc. and Abl.
té-té ; and with addition of the particle -mét we find ¢#-tt-met
(cf. ch. ii, § 39), t2b2-met, té-met.

§ 4. Tis (like mis, § 2) is attested for O. Lat. by the grammarians (see
Neue, ii’. p. 347), and occurs in Plaut. Mil. 1033 (an anapaestic line):

quia tis egeat, quia té careat: ob edm rem huc ad te missast,

where, if we keep the reading of the MSS. (the evidence of the Palimpsest is
wanting), we must scan fis, unlike mis. (Shall we read quia tis ea egeat ?) ;
Trin. 343 né tis alios misereat (so in the Palimpsest, but the other MSS. have
changed the unfamiliar form to tui), and possibly elsewhere (see Neue, J. c.
and add Plaut. Cist. 457). The Dative is spelt ¢ébei, C. I. L. i. 542. 1453, but
tibe in all the MSS. of Varro R. R. iii. 7. 11, and in C.J. ZL. i. 33 (one of the
Scipio epitaphs in Saturnian metre, c. 180 B.c.), The Ace. is spelt te (probably
with elision) on the same epitaph :

quare lubens t(e) in-grémium, Sefpio, récipit.
424 THE LATIN LANGUAGE. {Chap. VII.

Tu-met is not allowed by Priscian (i. p. 591. 5 H.’. An example of tute is the
famous alliterative line of Ennius (4. 108 M.) :

O Tite tute Tati tibi tanta tyranne tulisti.

(For other examples of tute, tutimet, tibimet, &c., see Neue, ii, pp. 361 sqq.; on
tiiquidem, téquidem, see above. ch. iii. § 51.)

The Umbro-Osean forms for the Accusative (and Nom.?) point to an
original tiom (from *twiom as fio from *fwio ; or with i for L.-Euwr. a ?), Umbr.
tiom and tio, Osc. tiium (usually taken as Nom.).

§ 5. Reflexive. The I.-Eur. stem was *swé (or *sw6) and *sé
(or *s), &e. (e.g. O. Ind. sv4-, ‘own, Goth. svés, O. Sl. svoji;
Goth. si-k Acc., O. Sl, sebé Dat.), often with the same affix as
2, Sg. *twé, *té (see above), viz.-bh-. The Latin Gen. is sw, the
Gen. Sg. Neut. of the Possessive (like mez, tui; see above), the
Dat. sist for *sebhei (Pelign. sefei, Ose. sifei ; ef. Pruss. sebbei),
with s?- for sé- owing to the unaccented use of the Reflexive; the
Ace. and Abl. sé (O. Lat. séd, which went out of use with med,
ted, § 2), often doubled for emphasis, sésé. The particle -met is
added for the same purpose to se, sti, viz. semet, sibimet. Se-pse

is read in Cic. Rep, ili. 8. 12: quae omnis magis quam sepse
diligit.

§ 6. The spelling sibei is found on CI. L. i. 38 (an Elegiac Scipio epitaph,
¢. 130 B.c.) ut sibei me esse creatum Laetentur; i. 196 (the S. C. Bacch. of
186 Bc.) ; i. 198 (the Lex Repetundarum of 123-122 B. ¢.); i. 200 (the Lex
Agraria of 111 B. c.); i. 205 (the Lex Rubria of 49 B.c.), &e. ; the spelling sibe
was found (with quase) by Quint. (i. 7. 24) in several MSS. (sed an hoc voluerint
auctores nescio), and was affected by Livy (T. Livium ita his usum ex Pediano
comperi, qui et ipse eum sequebatur). Livy’s spelling was probably designed
to express short unaccented i, the sound which the grammarians describe as
being ‘between an e and an i’ (see ch, ii. § 16). Sibi (like mihi, tibt) is the
usual scansion in Plautus and the early Dramatists, but sibi is by no means
rare, and is normal in the phrase suus sibt, ‘his very own’ (so meus mihi), e. g.
suo sibi gladio hunce jugulo ; ef. O. SI. pisacha svoja si réti, ‘scribebant suam
sibi linguam’). There is no Genitive form in O. Lat., *sis, corresponding to
mis, tis (Priscian ii. p. 2. 29 H.). Sed, Ace., occurs on the 8. C. Bacch., of
186 B.c. (C. I. L. i. 196 inter sed), and on the Lex Bantina of 133-118 B. o,
LL. 197 apud sed: along with sese, seese); sesed on the inscr. of the Faliscan
cooks (Zvet. I. I. J. 72 a). Sese is much more frequent than mémé and téte.
(For examples of its use, see Neue, ii®. p. 355.)

The Oscan Acc. is siom (cf. above, § 4, on Umbro-Ose. tiom ‘te’), the
Umbrian Dat. seso.

§ 7.1 Plur. Of the two I.-Eur, stems, *wé- (or *wé-) (e.g.
O. Ind. vay-4m Nom., Goth. veis, Engl. we, Lith. vé-du Dual,
O. 81. vé) and *né- (or *né-) (e. g. O. Ind. nas, the enclitic Gen.,
§§ 5-9.] THE PRONOUNS. PERSONAL. 425

Dat., Acc., O. Ir. ni, Goth. uns for *ns, Engl. us, O. Sl. nasti ;
Gk. vé. Dual, O. Sl. na; with weak grade ns- and an affix
-smé- in O. Ind. oblique cases, asman Acc., asmabhis Instr., &c.,
Lesb. Gk. dupes, Att. ucts), the latter appears in Latin nds Ace.
(Zend na Acc.). Ns is also the Nominative form. The Gen. is
nostrum (O. Lat. nostrorum, -arum) and nostyi, the Gen. Plur. and
the Gen. Sg. Neut. of the Possessive, ‘of ours,’ the Gen. Pl.
being used when the idea of plurality was specially prominent,
e. 2. omnium nostrum, pars nostrum, but amicus nostri; the Dat.,
Abl. is n6vis (earlier 16-dei-s), apparently with the plural suffix s
added to a case suffix like that of Dat. Sg. tibez, side: ; in O. Lat.
nis seems also to have been used (Paul. Fest. 33.6 Th.), with the
ordinary Dat., Abl. Plur. suffix (see ch. vi. § 48). To strengthen
the pronoun, the particle -met is added ; nosmet (always Nom. in
Plautus), nobismet.

§ 8. Enos Acc. is found in the ancient Hymn of the Arval Brothers, preserved
in a carelessly written inscription, containing the minutes of the Priesthood
for the year 218 a. p. (C. I. L.i. 28): enos, Lases, iwuate, and enos, Marmor, iuuato.

If the lines have Saturnian rhythm, the accent must fall on the first syllable
of enos (see ch. ii. § 141, p.128 7.) :

énos, Ladses, iuudte . .

énos, Marmor, iuuato ;

but this, and indeed every fact about the lines, is doubtful. The form enos is
generally explained as having a particle *& prefixed to nos (cf. é-ué). Nosmet
is, like vosmet, always Subject in Plautus, but also Acc. in Terence, Phorm. 172,
and the classical writers. The use of nostrum (Partitive Gen. and with omnium)
and nostri is discussed by Gellius (xx.6). For instances of nostrorwm, -drum for
nostrum in Plautus, &c., see Neue, ii*. p. 359. The spelling nobeis is frequent
in the Ambrosian Palimpsest of Plautus (see Studemund’s Index, p. 505).

§ 9. 2 Plur. There are two L.-Eur. stems, *yu- (O. Ind. ya-
yém, Goth. jis, Engl. you, Lith. jis; O. Ind. yuvaém Dual,
Lith. ji-du) and *wé- (or *wé-) (O. Ind. vas, the enclitic Gen.,
Dat., Acc., Pruss. wans, O. 81. vy vasii; O. Ind. vam Dual, O. SI.
va). The former, with the affix -smé- is found in Gk. tyeis,
O. Ind. yugman Acc., yuSmabhis Instr, &c.; the latter is
the stem used in Latin vos Acc. (Zend va Acc.). 6s is also the
Nominative form ; vestrum (O. Lat. vostrorum, -arum) and vestri
(O. Lat. vostri) the Gen., with the same usage and origin as
nostrum and nostri (see above) ; vdlis is the Dat., Abl., older vobeis
(see above on xobis). The strengthening particle used is -met :
426 THE LATIN LANGUAGE. [Chap. VII.

vosmet (only Nom, in Plautus, but afterwards Acc. as well),
vobismet ; and -pte in O. Lat. vopte (Paul. Fest. 578. 21 Th.
‘ vopte’ pro vos ipsi Cato posuit)).

§10. For examples of Gen. vestrum (vostrorum) and vestri, see Neue, ii’.
p. 359; and ef. nostrum, nostrorum, nostri, above. The spelling vobeis is extremely
frequent in the Ambrosian Palimpsest of Plautus, and occurs in the S.C.
Bacch. (C. I. LZ. i. 196. 29), the Epistula ad Tiburtes of ¢. 100 B. c. (i, 201), &e.

§ ll. Il. THE POSSESSIVE PRONOUNS. Inthe I.-Eur.
languages there is a close connexion between the Possessives and
the Gen. of the Personal Pronouns. Thus O. Ind. (Vedic) tvé-s is
Possessive ‘thy,’ tdva is Gen.,‘ of thee’ (I.-Hur. *téwé); Dor.
Gk. reds is Possessive, réo (I.-Eur. *té-syd) is Gen. Latin ¢uus,
O. Lat. /ove- (I.-Eur. *téwo-) must be similarly derived from the
I.-Eur. Gen. *téwé, Latin swus, O. Lat. sove- (I.-Eur. *séwo-)
from a corresponding Gen. *séwé, while meus (*meyo-) may come
from the Locative-Genitive *mei (O. Ind. mé Gen., Dat.). The
same connexion is seen between cujus, the Gen. of the Interroga-
tive (§ 13), and cajus -a -wm the Possessive (§ 23). Some
purists objected to this Possessive, and parodied Virgil’s line
(B. iii. 1) by:

dic mihi, Damoeta, ‘cujum’ pecus, anne Latinum ?

under the idea that cujus -a -wm was merely a vulgar inflexion
of a Gen. suffix -us, treated as if it had been -ws of the Nom.
Sing.

Beside the Reflexive Possessive stem *séwd- (Gk. éds), there
was another I.-Eur. form, *sw6- (O. Ind. sv4s, Gk. és for *oFos),
The latter would be in Latin so- (ch. iv. § 68), and appears in
the O. Lat. forms sam for suam, sas for suas, sos for suos, sis for
suis, which must not be confounded with the Demonstrative stem
so- on the one hand (O. Lat. sos for eos, swm for eum, &e.), nor
on the other with the monosyllabic suas, suos, suis (pronounced
swas, swos, swis), where, through the unaccented use of the Poss.,
the % has been turned into a consonant (w) before a long vowel
(see ch. ii. § 149), just as the 2 (%) of unaccented meus has been
turned into y in monosyllabic meas, meos, meis, med, med. Mieis
was the older spelling of meis (ch. ii. § 9), and in the Voc. Sing.
§§ 10—-12.] THE PRONOUNS. POSSESSIVES. 427

Mase. we have, besides meus, the form mi, e.g. mi fili, mi homo,
me vir.

For the Plural Pronouns the suffix -téro- (ch. v. § 16) was used
(cf. Gk. tyué-repos, suc-repos), 1 Pl. ndster, 2 Pl. O. Lat. voster,
which by the phonetic laws of Latin (ef. véto from older vito,
ch. iv. § 10) became vester about the middle of the second cent.
B.C. [Osc. nestro- (?), Umbr. vestro-.]

The Gen. Sg. Neut. of these Possessives was in classical Latin
used for the Gen. of the Personal Pronouns, mei, tui, sui, nostri,
vestri, and, where the notion of plurality was involved (as in
the Partitive Genitive or with omnium), the Gen. Pl. nostrum,
vestrum (in O. Lat. nostrorum, -arum, vostrorum, -arum) (see § 8).

The strengthening particles -mé?, -pté are added to the Posses-
sives, e.g. meamet, swismet, meapte, suopte, nostrapte.

§ 12. The grammarians speak of an O. Lat. spelling mius, from which they
derive Voc. mi (as Laeli Voc., Laelius Nom.) (e.g. Charis. p. 159. 17 K., Vel.
Long. p. 77. 12 K.; other references in Neue, ii®. p. 366), though the only
instance which they quote is miis in Ter. Heaut. 699 (probably a dissyllable) :

at enim istoc nihil est magis, Syre, miis nuptiis aduérsum,

where miis (a spelling retained in the MSS. of Terence), like mieis (probably
a monosyllable) on a Scipio epitaph of c. 130 B. c. (C. I. L. i. 38):

uirtutes generis mieis moribus accumulavi,

and mieis (probably a monosyllable) in Plaut. Men. 202: una uiuis mieis mori-
gera moribus (probably mieis, Truc. 709, where the MSS. read miles), shows that
spelling of i for e (I.-Eur. ey) which was the rule in Republican Latin in an
unaccented syllable before -ts (-eis), o. g. abiegnieis (but abiegnea Acc.), aesculnicis
on the Lex Parieti Faciundo of 105 B.v. (C. I. L. i. 577) (see ch. ii. § 9). [For
other passages in Plautus where miis is indicated by the MSS., see Neue, J. ¢. ;
in Trin. 822 mis may be Gen. Sg. of the Personal Pronoun, mis (§ 2), or Abl.
Pl. of the Possessive : bonis mis quid foret aut meaé uitae (anapaestic)]. For
the Vocative Sg. Masc., meus was used with an O-stem Noun in the Nominative
form, e. g. Plaut. Asin. 664 :

da, méus ocellus, méa rosa, mi anime, mea uoluptas,

Cas. 137: .
sine, amébo, ted amari, meus festis dies,
meus ptllus passer, méa columba, mi lepus.

(Notice mi lepus in Plaut., where the Noun is not an O-stem. But Virg. has
meus sanguis.) On the theory that mi is I.-Eur. *moi or *mei Loc., see § 2 above.

Mi is not found with a Voce. Fem. Sg. or Mase, Pl. till Late and Vulgar Latin,
e.g. mi parens, my mother, Apul. Met. iv. 26; o mi, inquit, hospites, Petron.
428 THE LATIN LANGUAGE. [Chap. VII.

116, p. 82. 25 B. ; (mi homines, mi spectatores, Plaut. Cist.678, should probably
be mei, as in Mil. 1330 O mei oculi, O mi anime). (See Neue, ii*. pp. 368-9.)

We find tou[am] for tuam on an old inscription (¢. I. Z. i. rago quei toufam]
pacem petit adiouta); soweis for suis on the Lex Repetundarum of 123- 122 B.c.
(i. 198. 50, beside swe, suae, suo, sua Abl.,) and on i. 1258, i. 1297 (where it is a
monosyllable :

plouruma que fecit populo soueis gaudia nuges),
souo i, 1007 (beside swom, both dissyllabic :

sudm mareitum cérde dilexit souo),
sowom Gen. Pl. Mase. i. 588, of c. 81 B.C.

The unemphatic Possessive is a monosyllable in the early Dramatists (by
Synizesis), in iambic forms, e.g. més, sts, which should probably not be
scanned according to the Breves Breviantes Law *mids, *stivs (see ch. iii. § 49).
(Instances from Plaut. and Ter. in Neue, ii®. p. 371.) How far the Vulgar Latin
unemphatic Possessive *mus, *mum, *ma (seen in Fr. mon, ma, and in Vulg.
Ital. ma-donna, padre-mo, &c.) should be referred to O. Lat. so- (I.-Eur. *swo-),
&c., or to these colloquial forms, is a doubtful point. (Of. tis for tuis on a late
metrical inscription, Orelli 4847 :

cum vita functus jungar tis umbra figuris.)

O. Lat. so- Possess. is attested by Festus and distinguished from O, Lat. so-
Demonstr. (sas Fest. 476.17 Th. ‘sas’ Verrius putat significare eas, teste Ennio,
qui dicat in lib, i.:

uirginés nam sibi quisque domi Romanus habet sas,

cum suas magis videatur significare, sicuti ejusdem lib. vii. fatendum est
eam significari, cum ait :

nec quisquam sapientia quae perhibetur
in somnis uidit prius quam sam discere coepit,
idem cum ait ‘sapsam,’ pro ipsa nec alia ponit in lib. xvi. :
quo res sapsa loco sese ostentatque iubetque,
et Pacuvius in Teucro :
nam Teucrum regi sdpsa res restibiliet ;
sis Fest, 428. 11 Th. ‘sos’ pro eos antiqui dicebant, ut Ennius, lib. i. .
constitit inde loci propter sos dia dearum,
et lib. iii. ;
circum sos quae sunt magnae gentes opulentae,
lib. vii. :
dum censent terrere minis, hortantur ibe sos,
lib xi..
contendunt Graios, Graecos memorare solent sos,

.
interdum pro suos ponebant, ut cum per dativum casum idem Ennius effert :

postquam lumina sis oculis bonus Ancus reliquit ;
§ 13.] THE PRONOUNS. DEMONSTRATIVES. 429

sam Paul. Fest. 33. 6 Th. antiqui dicebant... ‘sam’ pro suam). On the
other hand Festus quotes in illustration of O. Lat. puellus part of a line of
Ennius with monosyllabic suos (so spelt in the MSS. of Festus 324. 17 Th. and
Paulus 325. 6):

Poeni soliti suos sacrificare puellos,

and the MSS. of Lucretius spell suo (monosyll.) in i. 1022, v. 420:
ordine se suo quaeque sagaci mente locarunt,

but sis in his quotation of Ennius’ line, just mentioned, iii. 1025:
lumina sis oculis etiam bonus Ancus reliquit,

so that monosyllabic sis, &e. of everyday speech, a scansion discarded by the
Augustan poets as unsuitable to the dignity of poetry, were different forms
from O. Lat. sos, &e. We have monosyllabic suo, tud (see ch. ii. § 65) on the
dedicatory inscription of Mummius, the conqueror of Corinth (CLL. i. 542,
of 146 B.C.) :

uisum animo suo perfecit, tua pace rogans te.

Gen. Pl. mewm, tuwm, nostrum are attested by Priscian (i. p. 308. 23 H.), and
oceur along with swum (ef. souom, C. I. L. i. 588, of ¢. 81 B. c.) in the Republican
Dramatists, &c. (see instances in Neue), though an instance of vostrwm (vestrum)
is difficult to find. It occurs in the elegiac epitaph of Ennius quoted by
Cicero, Tusc. i. 15. 34 :

hic uestrum panxit maxima facta patrum.

The particle -met is not found with the Possessives in Cicero or Caesar, and is
not common in other authors. We have méamet culpa, Plaut. Poen. 446, sucimet,
sudmet, sudmet, suismet in Sallust. It was thus more a particle for Personal than
for Possessive Pronouns. But -pte is very common, though almost exclusively
with the Abl. of the Possessive, which precedes its noun, e.g. suapte manu,
Cic. ; meopte ingenio, Plaut.; nostrapte culpd, Ter. Phorm. 766. (See examples and
references to the native grammarians, in Neue, ii*. p. 373). Still the usual
method of emphatic expression was sud ipsius manu, meo ipsius ingenio, nostra
ipsorum culpa in literary Latin, and in colloquial speech swus sibi, meus mihi,
&c. (§ 6). Examples of Possessives from the Umbro-Oscan dialects are, Ose.
tuvai ‘tuae,’ suvam ‘suam,’ stivad ‘sua,’ suveis ‘sui’ (perhaps Pel. svap
‘sua,’ Osc. nistrus ‘nostros’”); Umbr. tua and twa ‘tua,’ tuer and tover ‘tui,’

vesira ‘vestra.’

§13. III. DEMONSTRATIVES. There were several Demon-
strative Pronoun-stems in I.-Eur., some of which appear in some
languages as Adverbs and Particles only, but as Pronouns in
others. Thus the stem ke-, (ko-) appears in Latin in the
Adverb ce- of cédo, ‘ give here,’ and in the Particle -ce of hujus-ce,
illis-ce, &e., but in Oscan eko- is the Pronoun in common use,
which corresponds to the Latin /zc; and on the other hand the
430 THE LATIN LANGUAGE. (Chap. VII.

stems Sho-, She- (ef. Shi-) of Lat. A2-c (O. Lat. é-c), ho-e appear
as particles in O. Ind. ha, hi, Gk. od-x/, val-xu.

The Latin Demonstrative stems are :

(1) I.-Eur. sé-, (sé-), which seems to have been originally con-
fined to Nom. Sg. Mase. and Fem., a stem té- being used
elsewhere (O. Ind. s4 and sds M., sii F., tad N., Gk. 6 and ds M.,
» F., 7d, for *rod, N.) So- is the stem of O. Lat. sam ‘ sat
sos eos, sum *eum’; to- of the second element of the compound
is-te for *ts-t8 M. (ch. iii. § 37), ts-da F., is-tud N., from the
stem *i-, *ei- (see below), though some regard the first part as
the stem es- (Umbr. es-to-?),a byform of the stem so- (se-),
with i for e owing to the unaccented use of the Pronoun (ch.
iii. § 18). By a further reduction of the vowel iste became in
course of time ste.

(2) L-Eur. ghi- (&hé-) supplies Lat. 27-c from O. Lat. 2-c M..,
haec for *hai-ce, hoc for hoce for *hod-c, ‘ this,’ all augmented by the
particle *ke. The 7 of dic is due to its unaccented nature (see
ch. iii. § 18).

(3) I.-Hur. ol- (Lat. i-tra, ul-timus; see ch. ix. § 56), a
grade of the root al- (al-) of Gk. &Ados, Lat. alins, &e. (§ 29), with
I_-Eur. s6- appended. The Nom. Sg. Masc. was *ol-s3, O. Lat.
olle, or *ol-sés, O. Lat. of/ws, Fem. *ol-sa, O. Lat. ol/a, with 7
by the Latin phonetic law for original ls (cf. velle for *vel-se),
Neut. *ol-téd, which should have been in Latin *oltud, *ultud,
but which was adapted to the Mase. and Fem. forms and
became oldu. In class. Lat. owing to the unaccented use of
the Pronoun apparently, though the change is a curious one, the
6 became 2%, ilde (id/us being dropped), idda, wlud.

(4) From I.-Eur. 1- (ei-) and eyo- (O. Ind. id-4m N., im-4m
Ace. M., Lith. jis M., O. SI. -j{, &.) comes the Latin ‘ anaphoric’
pronoun (i.e. the pronoun which refers to something previously
mentioned), ts M., ea for *eya F., t¢ N. Augmented by a
combination of the particle pé (ch. x. § 1) with the Pronoun-stem
so-, a combination which expresses ‘self, it forms the Latin
Pronoun of Identity, ipse M. for *¢-p(e)-s0 (ch. iii. § 37), ipsa F.,
ipsum N., in the O. Lat. also ea-pse with flexion of the first
element of the Compound only, and ea-psa with flexion of both
elements. Augmented by the particle -dem (ch. x. § 1) it
§ 13.) THE PRONOUNS. DEMONSTRATIVES. 431

expresses ‘ the same,’ 7-dem for is-dem M. (like audio for *aus-dio,
ch. iv. § 151), ea-dem F. (For a fuller list of the I.-Eur, demon-
strative stems which appear in Latin, see the chapters on the
Adverb and the Conjunction.)

The Declension of these stems differed originally from that
of Noun-stems, though it became more and more assimilated in
course of time, and in Greek had come to be almost identical.
The Nom., Acc. Sg. Neut. was formed, not like O-stem Nouns in
-m, e.g. donum, but in -d (often written -z, ch. ii. § 73), eg.
illud, istud (but ipsum instead of *i-ptud), id; the Dat. Sg. of
all Genders has -2, older -ei (-07), like the Locative -ei (-oi)
sufix which appears in O-stem nouns such as Crinthi, e. g.
ali, istt, ipsi, huie for O. Lat. hoi-ce, éi, and & for *ey-ei; the
Gen. Sg. of all Genders is formed by the addition to this
Dative-Locative form of the Gen. suffix -ds, -ws, as in O. Lat.
nomin-us, e.g. illt-us, isti-us, ipsi-us (with shortening of vowel
before vowel, ¢dius, &c. ch. ii. § 143), Aajus (pronounced *htiyyus)
for hoi-us, ejus (pronounced *éyyus) which when unaccented
became in the rapid utterance of ordinary speech *7//2s, *istis,
*insis, *hiis, *éis, e.g. illi(u)s modi Ter. Ad. 441, or in one
word, il/imodi, with s dropped before m as in primus for *prismus
(ch. iv. §151). The Nom. Plur. Mase. suffix of O-stem pronouns,
viz. -oi, was in Latin (as in Greek) borrowed by O-stem nouns
(see ch. vi. § 40), and so is not distinctive of the Pronoun in Latin,
but the Nom. Sing. Fem. (and Nom. Ace. Pl. Neut.) in -ai is
a feature of the pronominal declension which remains in Latin
haec for O. Lat. hai-ce, ilaec for *illai-ce, istaec for *istai-ce,
though without the particle -ce the two last take the -a of
Noun-stems, é/a, ista (and so ipsa). Before the Dat., Abl. Plur.
suffix the stem appears with an appended i (O. Ind. té-bhyas,
Goth. pai-m, Lith. té-ms, O. SI. t8-mit), eg. Lat. A7-bus; cf.
i-bus for *ei-bus (O. Ind. é-bhy4s), but gui-bus, not *gui-bus.
There is the same difficulty with Latin Pronouns as there is
with Nouns (ch. vi. § 36), in assigning any ‘ Ablative ’ forms to
an original Instrumental. Thus Osc. svai puh, ‘si quo’ (cf.
Umbr. pu-e ‘ quo’), in the sense of ‘sive’ (but see ch. ix. § 5),
suggests that Lat. gud in some of its uses may be an Instru-
mental case form and represent an original *q*6 not *q*dd. But
432 THE LATIN LANGUAGE, {[Chap. VII.

Adverbial ed in praeter-ea is shown to be an Ablative by aruorsum
ead on the 8. C. Bacch. (C. 1. L. i. 196. 24); cf. eod die, x1. 4766.
The Locative had also the Dat. and Gen. functions, which we
found to belong to the Loc. of the Personal Pronouns, e.g. Gk.
ayrépe pou (§ 2). Its suffix in O-stems was -ei or -oi, e.g. Gk.
not, Dor. wet, éxel; the former is the suffix used in Oscan, e.g.
eisef, which uses in the Fem. the A-stem Loc. suffix, e. g. efsai.

§ 14. O. Lat. so-. Festus (428, 11 Th.) quotes sos for eos from three lines of
Ennius (see § 12); elsewhere (476. 17 Th.) he cites sam for eam, sapsa for ipsa
from the same author, and (426. 2 Th.) sum for ewn: ‘sum’ pro eum usus est
Ennius lib. i. ;

astu, non ui, sum summam seruare decet rem,
et lib. ii. :
ad sese sum quae dederat in luminis oras.

[The gloss (Léwe, Prodr. p. 350) ‘soc’: ita is doubtful; cf. Umbr. e-soc ‘sic.’]
But neither the Demonstrative so- nor the byform of the Reflexive Possessive
so- (I.-Eur. swo-) are found in the conversational language of Plautus and
Terence. The Neuter *tod is preserved in an O. Lat. Adverb topper for *tod-per,
on which see ch. ix. § 7.

§ 15. The particle -ce. This particle belongs to the stems ké- (ké-), (also
ki- and kyo-) (Gk. éxe? Loc. Advb., xeivos, Ose. eko- (often with this particle
added, e. g. ekask, ‘ hae’), Lat. cé-ter, ci-tra, O. Ir. cé, ‘on this side,’ Engl. he,
him, Germ. heu-te, Lith. szis, ‘ this,’ O. Sl. si], and appears as an Adverb or
Preposition in Lat. cé-do, ‘give here’ (more frequent in Terence than in
Plautus), (O. Ir. cit ‘da’ (?)), and probably in Ose. ce-bnust, ‘huc venerit.’ It
is said to be employed as an enclitic in other languages too, e.g. Arm. tér-s,
‘this person,’ Goth. pau-h, Engl. though (?). Its widespread use with Demon-
stratives in Latin and Umbro-Oscan (Lat. hi-c, ille, O. Lat. illic, iste, O. Lat.
istic, Osc. eisa-k and Umbr. erak, ‘ea,’ Osc. ekask, ‘ hae,’ exac, ‘hac, it-k,
‘ea,’ and soon) may be compared with the Romance forms which have prefixed
Lat. ecce to certain Pronouns and Adverbs, e.g. Fr. ca, ‘ there,’ from ecce-hac ; Fr.
ci, Ital. ci, ‘here,’ from ecce-hic ; Fr. ce-, Ital. cid, ‘this,’ from ecce-hoc; Fr. celle
from ecce-illa, Fr. cette from ccce-istu, &.), (ef. eccillwm uideo and eccistam wideo
Plaut.). It is the same tendency of expression which has reasserted itself ;
and it has its equivalents in vulgar English ‘this here’ for ‘this,’ ‘that
there’ for ‘that,’ and so on, The usage with these forms with -ce differed in
early and in classical Latin; for while in early Latin they are more or less
arbitrarily used, in classical Latin their use is stereotyped in most pronouns.
Thus class. Lat. illic is the Adverb (Locative}, while ill is the Dative (so istic
and isti); hae is the Nom. Pl. Fem., haec the Nom. Pl. Neut.; but in
Plautus ili and illic, istt and istic, are equally Adv. and Dat. Pron., hae and
haec (illae and illaec, istae and istacc) are both used for the Nom. Pl. Fem., ila
and illaec (but only haec, istaec) for the Neut., while illic, istic, forms not found
in class, Lat., are equivalents of ile, iste, illuc of illud (but only istue in Terence,
and perhaps also in Plautus). Still there are rules observed even so early as
§§ 14-16.] THE PRONOUNS. DEMONSTRATIVES. 433

Plautus’ time ; for example, the Nom. Pl. Mase. of hic, ille, is before a word
beginning with a vowel hisce, illisce, but before « consonant hi, ill, and in
general the forms with -ce are found in use before an initial vowel, hosce, hasce,
hisce Dat.-Abl., illisce, istisce Dat.-Abl., though before a consonant horine, harinc
are employed when the verse-ictus has to fall on the final syllable.. In the
Latin of ordinary conversation, as represented by the language of Plautus’
plays, the particle -ce has not in these pronouns the worth of a syllable,
a dissyllabic haece, &c. being proper to a more elevated style of poetry, e.g.
Ennius, A. 294 M. haece locutus vocat, at the beginning of a hexameter line,
so that hacetenus and hocedie, quoted as O. Lat. forms by Mar. Victorinus
(9. 19 K,), must come from an Epic poem, or else from some official inscription
or legal document. When the interrogative particle -né is added, the -ce is
preserved as a full syllable, e.g. hoc-ci-ne, si-ci-ne with % not &, because the
vowel has now become medial and not final (ch. iii. § 39). The full form
only is used in Plautus and Terence, but in later authors we find also, e. g.
hicne (Stat. Theb. i. 189; other examples in Neue, ii’. p. 422). The Interjection
em seems always to take the forms with -c, e.g. em illic, em istoc, but not the
Interjection ecce, e. g. ecc-illam, ecc-istam, ecc-am (for ecce *ham?, § 16). Prefixed
to quidem the c is not used in hiquidem, &c. of the Dramatists (but istucquidem,
not istudquidem, in Plautus, as istuc not istud in the simple Pronoun).

§ 16. Hic. The old form hee (for *Bhé-ke) appears on the earliest Scipio
epitaph in Saturnian metre (C.J. L. i. 32):

hée cépit Cérsica Alériaque urbe,

which has, however, ¢ for original i in Tempestatebus, and in the previous line
shows the spelling hic (probably not the Adverb, for this would naturally be
spelt heic in so early an inscription) : :

ceénsol, cénsor, aidilis hic fuet a[pid uos].

The spelling hec seems to be a retention of the old form, due to the emphatic
position of the Pronoun at the beginning of the line, if it is not merely an
example of the early interchange of ¢ andi in the unaccented syllable (e. g.
aidiles for aedilis on the still older Scipio epitaph, i. 31 ; see ch. iii. § 22). The
Neuter Nom.-Ace. hoc should be *hoce (for *Shdd-ke), but though we are
expressly told that it was so pronounced before an initial vowel, as in Virgil’s
line : hoe erat, alma parens, &c., there is no evidence that it was ever written
with -cc, for Priscian’s statement (i. 592. 22 H.): in antiquissimis codicibus
invenitur bis ¢ scriptum, is illustrated only by the form hoccine. Hoc is
always long by position in Latin poetry, as hic Abl. is long both by nature
and by position (for *hdd-ce), and hic M. is distinguished from hic Ady. in
Plautus and the early Dramatists (spelt heic in Plautine MSS., Merc. 307, Men.
375, &c.), though, like other long syllables, hic, \hic) may be shortened, when
unaccented, after a short syllable by the Law of Breves Breviantes (ch. iii.
§ 42), e.g. quid hic est ?, what is here ?, (quid hic est?, what is this?). But in
classical poetry ric M. is more often scanned as a long than asa short syllable;
and the grammarians assert of hic in a line like Virg. A. xi. 16: manibusque
meis Mezentius hic est, that it was pronounced *hice (Mar. Victor. 22. 17 K.;
other references in Neue, ii’. p. 411), just as they declare hoc erat to have been
pronounced hoce erat ; s0 that it is probable that hic M. came to be pronounced
*nice by the Analogy of hoc (for *hodc) N., which was pronounced *hécc. (One

Ff
434 THE LATIN LANGUAGE. [Chap. VII.

theory supposes hic with long i to have been a byform of hic, Rev. Philologie,
1892.)

In the Genitive we have, e.g. hotusce on the Lex Repetundarum of 123-122 B.c.
(C.I.L. i. 198. 56), hoiusque (i. 603, of 58 B.c. hoiusque aedis ergo), hoius Plaut.
Pseud. 271 (holus A, hujus P), huiius Most. 664 (A, iwjus P). Plautus is said not
to admit the form with -ce (Poen. 1257) into the colloquial language of his plays.
It occurs, however, in Terence (Andr. 439, Phorm. 827), and the phrase hujusce
modi is very common in Cicero and Sallust.

In the Dative hoic occurs in the Lex Bantina of 133-118 B.o. (C. 1. L. i. 197. 26
hoice leegei). Mar. Victorinus (12. < K.) quotes ‘ex libris antiquis foederum
et legum, qui etiamsi frequenti transcriptione aliquid mutarunt, tamen
retinent antiquitatem,’ the form hoic, a form affected in pronunciation by
a few in the time of Velius Longus (first cent. a.p.) (p. 76.3 K.). In later
poetry huic (like ct? in the poetry of the first cent. a.p., § 25) is scanned
asa dissyllable (twice in Statius, Stlv. i. 1. 107; i. 2. 135; and even hiiic in Ter.
Maurus, 1. 1375. For other examples, see Neue, ii’. p. 415). In the Accusative
the oldest spelling hon-ce appears in the Lex Spoletina (C.J. L. xi. 4766 honce
loucom nequs uiolatod), hance in the Lex Bantina (i. 197), hoce for *hocce on
the S. C. Bacch. of 186 B.c., where double consonants are written single,
(i. 196. 26 atque utei hoce in tabolam ahenam inceideretis) (cf. i. 1291 itus
actusque est in hoce delubrum Feroniai). We have honc on the Scipio epitaph
Gi. 32), in the Saturnian line:

héne oino ploirume coséntiont R[émai],
‘hune unum plurimi consentiunt Romae.’ Like hunc from hone M. is huc(c)
from hoc(c) N., a spelling found in an inscription of somewhat irregular
orthography (i. 603, of 58 3.c., ad huc templum), and in the Falisco-Lat. huc
dederunt (Zvet. I. I. I. 72.4), but in classical Latin reserved to discriminate the
Adverb from the Pronoun (see ch. ix. § 10). In the Abl. we have the full
form hoce, e.g. in the magisterial proclamation cited above for hoce Neut. Acc.
(C. I. L. i, 1291 ex hoce loco; ef. Orell. 3857), zace in the Lex Bantina (i. 197. 7)
and throughout the Lex Repetundarum (i. 198) [cf. hocedie and hacetenus,
mentioned as O. Lat. forms by Mar. Vict. (9. 19 K.)). (For the Locative, see
the Adverb hic, heic, ch. ix. § 10). Nom.-Ace. Plur. Neut. haice is found in the
8. C. Bacch. (¢. 1. L. i. 196. 22 haice utei in couentionid exdeicatis, ‘haec uti
in contione edicatis’) ; haece has been already cited from Ennius, A. 294 M.
haece locutus ; hisce or heisce, the Nom. Pl. Masc. form used by Plautus before a
word beginning with a vowel, occurs without the particle in two inscrip-
tions of no great antiquity (C. I. L. i. 1059 heis sunt horti; i. 1071 heis sunt
duo concordes), and, according to the grammarians, in Virg. E. iii. roa:
his certe, neque amor causa est, vix ossibus haerent ;
we have heisce further in some Capua inscriptions of 108-71 B.c. (C.1.L.
i. 565 heisce magistreis Uenerus Iouiae ; i. 566 heisce magistreis Cererus ;
i. 567 heisce magistrei ; i. 569 heisce mag.; i. 573 heisc. magistr.; Not. Scav.
1893. p. 164 heisce magistreis); cf. C.I.L. i. 1478 (Cartagena), heisce magistris.
For examples of haec F., see Neue, ii. p. 417; e.g. Virg. G. iii. 305:
haee quoque non cura nobis leviore tuendae.

The form seems to have been used both by Caesar and Cicero. In the Geni-
tive the full form is found in Cato, e.g. harumce (R. R. 139 harumce rerum
ergo), and in the old oath administered to soldiers taking furlough (ap. Gell.
§ 17.) THE PRONOUNS. DEMONSTRATIVES. 435

xvi. 4. 4 nisi harunce quae causa erit, funus familiare, feriaeve denicales,
&e.). In the Dat.-Abl. hisce is used by Plaut. and Ter. before vowels, by
Cicero, &c. before consonants too (examples in Neue, ii®, p. 419 ; ef. heisce in
the Lex Repetundarum, C.J. L. i. 198«8 de heisce, dum, &c.), and the same
holds of Ace. hosce and hasce (cf. hasce sedes, C. I. L. iii. 7230).

The stem hé- (hé-) is augmented by 7 (ike the’ Pronoun Datives Pl., O. Ind.
té-bhyas, Goth. pai-m, Lith. té-ms, 0. Sl. t&-mit), in Dat. Pl. hibus (obsolete in
Varro’s time; cf. Varro, L. L. viii. 72), attested by Priscian, ii. p. 10. 15 H. (and
Charis. p. 54. 19 K.) in Plaut. Cure. 506 :

eodem hércle uos pono ét paro: parissumi estis hibus,

like ibus (for *eibus) in Plaut. Mil. 74 (see § 19); it follows the analogy of
Noun A-stems in Dat. Sg. F. hae in Cato (R. R. 14. 3 hae rei materiem.. .
dominus praebebit). The Nom. Pl. M. heis-ce, already mentioned, is, like
Nom. Pl. eis of the stem i-(§ 19), due to the addition of the Plural suffix -s to
the already formed plural hei, hi. (On the declension of hic in Plautus, see
Studemund in Fleckeisen’s Jahrbiicher, 1876, p. 57, and on its use, Bach in
Studem. Stud. ii). Hic always refers to the first Personal Pronoun in Old Latin,
and means ‘this that is near me, that belongs to me,’ &e. ; so hic homo for ego
in the Comedians. It is sometimes used for is, to refer to something already
mentioned, in the classical historians and Epic poets, but rarely earlier, e. g.
Ter. Phorm, 866-9 has hic where ibi would be used by Plautus :
ad fores
sispenso gradi placide ire pérrexi, accessi, dstiti,
hic pulchérrumum
ficinus audiui.

The Adverb hé-die (Falisc. foied) seems to show the bare stem (cf. ch. ix. § 5).
A form without the enclitic has been claimed for O. Lat. eccum, e.g. sed eccum
Palaestrionem, ‘but see, here comes P.’, Plaut., though the word may be
analyzed into ecce eum as well as into ecce *hum (§ 15). On Faliscan he cupat,
‘here lie,’ ‘here lies’ (for *hei or for *heic?), and on Late Lat. hi jacei, see ch.
ix. § 10, We have Dat. hui in late inscriptions (Henz. 7339, Rome ; perhaps
also in C. I. L. x. 7297, Palermo).

§ 17. Iste. Examples of the various cases of this Pronoun with -c(e) in
Plautus and Terence are given by Neue, ii®. pp. 398 sqq., istic Nom. Sg. M.,
istaec F., istuc N., istic Dat., istunc Ace. M , istanc F., istuc N., istoc Abl. M., istac
F., istaeo Nom. Pl. F., N., istosci-n Acc. Pl. M., istaec N., also of the later cur-
tailed form ste (cf. Ital. stasera, from Lat. (i)sfa sera) as in the Itala sta nomina ;
see I. F, Anz. ii. 153) (ef. ch. iii. § 12. 4). (The declension of iste in Plautus
is treated by Studemund in Fleckeisen’s Jahrbiicher, 1876, p. 57, and its use By
Bach in Studemund’s Studien ii.) Iste refers to the second Personal Pronoun,
‘that of yours,’ as hic to the first and ille to the third. In classical Latin, but
probably not in the earlier literature, it came to acquire a contemptuous
sense. We find istéce in an old ritual mentioned by Cato (R. R. 132. 2):
Iuppiter dapalis, macte istace dape pollucenda esto. The influence of the
Noun declension is seen in Late Lat. isto Dat. Sg. (Apuleius), istum Neut.

-(Vulgate) ; istae Dat. Sg. is the reading supported by the MSS. in Plaut, Truc.
790 (see Georges, Lex. Wortf. s.v.). In colloquial Latin ecce is sometimes pre-
fixed to the Acc. case of iste, and forms a compound, e. g. eccistam uideo Plaut.

= Ff2
436 THE LATIN LANGUAGE. [Chap. VII.

§ 18. Ile. Nom. ollus M., olla F. are the words used in two ancient
formulae preserved by Varro (L.L. vii. 42), the announcement made at a ‘funus
indictivum,’ viz. ollus leto datus est, and the proclamation by the herald at the
comitia, in which olla centuria and not ¢illa c.’ was used, just as we keep up
the old Norman French ‘ oyez, oyez’ in Royal proclamations. The same
Nom. Fem. seems to occur in the formula, also preserved by Varro (L. L. vii. 8),
which the augur used on the citadel in marking off a ‘templum,’ though the
reading of the MSS. is corrupt: ullaber arbos quirquir est... ollaner arbos
quirquir est. The immediate precursor of ille, the Nom. Mase. olle, is found
in the Law of Servius Tullius, quoted by Festus (290. 15 Th.) : si parentem
puer verberet (MS, -it), ast olle plorassit, puer divis parentum sacer esto,
A very old form of the Abl. Plur., oloes, is cited by Paul. Fest. 14.17 Th.: ‘ab
oloes’ dicebant pro ab illis; antiqui enim litteram non geminabant ; and in
the inscription on the Columna Rostrata (a restoration made in the Imperial
period) we have the Gen. PI. olorom (C. I. L. i. 195 praesente[d Hanibaled]
dictatored olforjom) (cf. olarum on an early inscr., Marini, Act, Arv. p. 233).
Macrobius (iii. 9. 10) quotes an old ritual with ollis legibus, and on the Lex
Furfensis of 58 8. 0. (C.J. L. i. 603) the same phrase occurs, olleis legibus (followed
immediately by dieis regionibus), similarly on the Lex Cornelia of 81 B. c. [i. 302.
(1). 6 olleis hominibus]. Cicero in his De Legibus employs in the archaic language
of his laws oili Dat.Sg., ollis Dat. Pl., olli Nom. Pl., olla Neut. Pl., ollos Acc. Pl. In
the early Dramatists the form is not found, so that it must have been already
relegated to the legal style by the beginning of the second cent. B.c. Ennius
in his Annals knows only olli Dat. Sg. and Nom, PL. and ollis Dat.-Abl. Plur.
(ili and illis are apparently not used by him), and these are the only forms of
colle found in his imitator, Virgil, and the later Epic poets, while Lucretius
confines himself to ollis. Quintilian mentions olli among the happy archaisms
of Virgil (viii. 3. 25 ‘olli’ enim et ‘quianam’ et ‘mocrus’ et ‘pone’ et
‘porricerent’ adspergunt illam, quae etiam in picturis est gratissima, vetus-
tatis inimitabilem arti auctoritatem). The old spelling may have lingered
longest in these forms through the analogy of the Adverb lim, which comes from
the stem dl- (Umbr. ulo ‘illue’) a byform of dl- (ch. iv. §§ 45,59), and does not
exhibit a substitution of d- for dU- (ef. ch. ii. § 127). Servius (ad A. i. 254 and
v.10) mentions a theory that olli in these two lines of Virgil was not the Dat.
Sg. of the Pronoun, but an Adverb with the sense ‘ tune’ (cf. the glosses oili
‘ili... aut tune’; oll ‘ibi, interdum, illi, vel illine,’ C. G. L. v. 229. 4-5),
and olli-c is quoted by Paul. Fest. 231. 2 Th. as an O. Lat. form of illic (ef.
Lucil. ine. 152 M. : uelut olim Auceps ille facit, where the MSS. offer olli). The
mispronunciation, ‘ oli’ for olim, censured in Prob. App. 199. 16 K., is merely
an example of the tendency of Vulg. Lat. to omit final -m, like the mispronun-
ciations ‘pride,’ ‘passi,’ ‘numqua,’ ‘ide,’ censured on the same page. From
olim was formed Late Lat. olitanus, ‘ of former time,’ and the curious phrase,
found in Petron. 43 p. 29. 2 B. olim oliorum, ‘long long ago.’ [Another theory
connects O. Lat. ollz, ‘then,’ with O.SI1. lani, ‘last summer,’ and analyzes oilo-
into *ol-no-, not *ol-so- (I. F. iii. 264).]

For instances of the addition of the particle -ce to ille in its various forms in
Plautus, &c. (illic Nom. Sg. M., illace F., itluc N., illic Dat., illunc Acc. M., illanc F.,
illoc Abl., dlac F., illisce Nom. Pl. M. (before a vowel), illaec F., illaec Neut., illisce
Dat.-Abl. (before a yowel), see Neue, ii. p. 427. In old rituals mentioned by
Cato in the Res Rustica we find illiusce (139 illiusce sacri coercendi ergo), ilace
§§ 18, 19.] THE PRONOUNS. DEMONSTRATIVES. 437

(1g2. 1 eius rei ergo macte hac illace dape pollucenda esto), illisce (141.4 Mars
pater, siquid tibi in illisce suouitaurilibus lactentibus neque satisfactum est,
te hisce suouitaurilibus piaculo) ; and in legal formulae in Varro’s Res Rustica,
illosce (ii. 5. rr illosce boues sanos esse noxisque praestari, . . . illosce inuencos
sanos recte deque pecore sano esse noxisque praestari spondesne ?), ilasce (ii.
4. § illasce sues sanas esse). The Analogy of O-stem Adjectives has produced
ilum Neut, in Late and Vulgar Latin (e. g. in the Vulgate, S. Marc. iv. 35), illae
Dat. in Cato (R. R. 153 and 154 illae rei) and Plaut. Stich. 560 (filiae illae ; so
the MSS.), ilo Dat. in late authors (e.g. Apuleius) ; to the Analogy of I-stems
is due illibus, if it be a genuine form, ascribed to O. Lat. by Serg. in Donat.
Pp. 547. 37 K.

Lucilius seems to have proposed the spelling ili Dat. Sg., illei Nom. Pl.
(ix. 15 M.):

‘hoe illi factumst uni’: tenue hoc facies i.
“haec illei fecere’: adde e, ut pinguius fiat,

on which see ch.i.§9. According to Diomedes (332. 11 K.) the strengthening
particle -met is added to ile, and produces illemet.

In colloquial Latin ecc- was prefixed to the Acc. case of ille and formed
a compound, e.g. eccillum uideo, Plaut. By prefixing em [the old deictic
interjection, for which én was substituted in class. Latin (ch. x. § 19)], was
formed ellum, ellam of the Comedians, which survives in the exclamation used
in the dialect of the Abruzzi at the present day, ello (with open e).

On the pronunciations iltus and ilius, see Neue, ii*. 518, and cf. below,
§ 22,

§19. Is. The stem i- appears clearly in is Nom. Sg. M., id N. But the
encroachment of the O-declension shows itself in the numerous cases formed
from a stem éyé- (e.g. eum for *éydm Ace. Sg. M., which ousted the old Acc.
im; eis for *eyois Dat.-Abl. Pl., which ousted the old ibus for *ei-bus (O. Ind.
é-bhyds), with stem augmented before the Dat. Plur. suffix as in hibus (§ 16).
Another old I-stem form may be Abl. 7 in 7-lico, which is used in O. Lat. in
the sense of ‘in loco’ (‘in eo loco,’ Non. 325. 6 M., who quotes Accius (Trag.
373 R.): .
ilico, inquam, habitdto, nusquam prépius],
while the class. Abl. Sg. M. is ed for *eyod. licd is however better explained
as *in sloco, O. Lat. for in léco (see ch. ix. § 7).

It may be owing to this stem eyo- that we find the natural Fem. *ia
replaced by ea. The Acc. form iam preserved in the MSS. of Varro, L. L. v. 166
and viii. 44 (ef jam, the Adverb, ch. ix. § 10. 8 ; ium on a Lueeria inscr., C. I. L.
ix. 782), may be a relic of the old and correct spelling, for there is no
indication that there ever existed a spelling *eiam, of which some have sup-
posed it to be a corruption (cf. ch. ii. §.9). (In Umbr. we find eam Ace. 8g.
F., eaf Ace. Pl. F.)

The particle -ce is not added to this Pronoun-stem until Late Latin, e.g.
ejuscemodi (Jerome, Epist. 82.6) by Analogy of hujuscemodi. [Posteac in Clau-
dius’ edict on the Anauni (Hermes, iv. 99, 1. 13) is an isolated form.] But in
Oscan we have iz-ic ‘is,’ it-k ‘ ea,’ id-ik ‘id,’ and in Umbrian (with intervocalic
s become r) er-ec ‘is,’ ed-ek ‘id,’ &c., as well as forms like Ose. id-ad ‘ad id,’
Umbr. eo Acc. Pl. M., eu Ace. Pl. N., &e. :

The various case-forms that call for notice are these. Nom. Sg. M. eis
appears three times beside the usual és on the Lex Repetundarum. It may be
438 THE LATIN LANGUAGE. [Chap. VII.

a Nom. formed from the strong stem ei- (cf. Dat. Pl. i-bus), but is as likely to
be a mere mistaken use of ei for i as in seine and leiteras on the same inscrip-
tion, an inscription not older than the last quarter of the second cent. B.c.
(On eisdem, see § 21.) In the Dat. Sg. ied is written in the Lex Rubria (1 L.
i. 205) of 49 B.c. (beside ¢#), by the same orthography as micis, abiegnicis, aesculnicis
(see ch, ii. § 9, and ef. tei Nom. Pl., ieis Dat. Pl. below). (So Umbr. ie-pru,
ie-pi). The Acc. Sg. M. was in O. Lat. im (Charis. 133. 1 K. ‘im’ pro eum.
nam ita Scaurus in arte grammatica disputavit, antiquos ‘im,’ ‘ques’... et
declinari ita: is, ejus, ei, eum vel im; Paul. Fest. 73. 29 Th. ‘im’ ponebant
pro eum, a nominativo is; ib. 33. 7 antiqui dicebant... ‘im’ pro eum ; Gl.
Philox. im: adréy, els abrév 3 Gl. Cyrill. abrév, rodrov: im (MS. eim)]. Macro-
bius (i. 4. 19) quotes a law of the XII Tables: si nox furtum factum sit, si im
oceisit, iure caesus esto, and Cicero (Legg. ii. 24. 60) another: cui auro dentes
iuncti escunt, ast im cum illo sepelirei ureiue se fraude esto (cf. Fest. 322.
13 Th.).

This form is often given as em, instead of im, perhaps by analogy of Noun
I-stems which substituted the Acc. suffix of Consonant-stems for that of I-
stems, e. g. turrem for turrim (see ch. vi. § 29); so Paul. Fest. 54. 20 Th. ‘em’ pro
eum, ab eo quod est is, Another law of the XII Tables is cited by Porphyrio
in his note on the ‘licet antestari ?’ of Hor. 8. i. 9. 76: si in ius uocat, ni it,
antestamino, igitur em (MSS. en) capito; and Festus (298.15 Th.) quotes, from
a speech of Cato, si em percussi. Similarly the Adverb im (ch. ix. § 10) (ef.
O. Ind} im, originally an Acc. of the ‘anaphoric’ pronoun, then a mere
particle), so written in Gl. Philox. im: Aomév, 45, is mentioned by Paul. Fest.
as em (53. 37 Th. ‘em,’ tum); the Acc. of an O. Lat. derivative of is in the
sense of idem is imeum in the Philoxenus Glossary (imeum: 7dv aitéy), but
emem in Paul. Fest. (54. 2 Th. ‘emem,’ eundem) (cf. the form in the Glossary
of Placidus, emdem (MSS. hendem) : aeque, similiter, C. G. L. v. 73. 19].

The Dat.-Loc. Sg. *eyei, written eiei! (along with ¢) in the Lex Repetundarum
of 123-122 Bc. (0.2. L. i. 198), has in Plautus and the older poetry three
scansions, @, & and &, in Ovid and later poets one, & (cf. r& and ré Plaut.,
7&@ class, ch. vi. § 19). This *éyéi is thought to have become & much as
-éyés became -és (e.g. Lat. trés for I.-Eur. *tréyés, ch. iv. § 66), which in
what we may call ‘ceremonious’ speech would preserve its full sound, and
for a time resist the Latin tendency to shorten every long vowel before
another vowel (ch. ii. § 143), while after the shortening set in, %, through its
unaccented character, would in the rapid utterance of everyday life become
a mere monosyllable é, as tii became tii, &c. (§ 12). (Priscian, ii. p. 10. 2 H,
speaks as if ei were as much a monosyllable as huic, cui in the ordinary
pronunciation of his time), To suppose that eiei was pronounced *ey-yei as
*Troia was pronounced *Troy-ya, pijor, *pey-yor (ch. ii. § 55), does not account
for the fact that these words are always scanned with the first syllable long
in poetry, while eum for *eyom is never scanned with its first syllable long.
(See below, however, on the doubt attaching to the antiquity of the scansion
&.) But the Genitive ejus, formed by the addition of the Gen. suffix -us to

 

' This would most naturally be form of the i-stem had in Latin the
explained as a Reduplication of ei, notion of identity, e.g. em-em (gl.
and some explain 27, & as produced eundem). Still the theory given
in this way. But the reduplicated above is far from certain.
§ 19.] THE PRONOUNS. DEMONSTRATIVES. 439

the already formed Locative had this pronunciation *éy-yus [Caesell. ap.
Cassiod. 206. 6 K. ‘Pompeiius,’ 'Tarpeiius’ et ‘eiius’ per duo i seribenda
sunt, et propter sonum (plenius enim sonant), et propter metrum. Num-
quam enim longa fiet syllaba nisi per i geminum scribatur}]. This pronun-
ciation is indicated by the spelling zrIvs on inscriptions (Neue, ii’. p. 376)
and MSS. (Index to Studemund’s Apograph of the Ambrosian Palimpsest of
Plautus). The treatment of Nom. Pl. *eyei for *eyoi and Dat.-Abl. Pl. *eyeis
for *eyois is not the same as that of Dat.-Loc. Sg. *eyei ; for we have in Plautus
and the older poetry no instance of *éi, as, but only & or 4, 7, and éis or &s, Ts.
Where ¢ is followed by a long syllable it passed in unaccented usage into y by
Synizesis, e.g. &, és, drum Plaut. (not és, &e., by Law of Breves Breviantes;
see ch, iii, § 49).

The old Abl, Sg. -d is seen in eod die of the Lex Spoletina (C. I. L. xi. 4766),
and in Adverbial ead of the S. C. Bacch. (i. 196. 25 quei aruorsum ead fecisent,
‘qui adversum ea fecissent’).

In the Nom. Pl. the spelling iei is, in accordance with the orthography of
the first cent. B. c. (ch. ii. §.9), found on the Lex Cornelia of 81 B.c. [i. 202. (1). 7,
along with ei), on the Lex Antonia de Termessibus of 71 B.¢. (i. 204, passim,
never ¢), on the Lex Rubria of 49 B.c. [i. 205. (1). 48], on the Lex Julia
Municipalis of 45 B.c. (i. 206. 24, usually ¢), and in Varro (L. L. ix. 2 and 35).
We have also i in MSS. (often confused with hi), and inscriptions (examples
in Neue, ii®. p. 382); and the e of Republican inscriptions (ibid. p. 383)
admits of being taken to denote the simple long i-sound. Priscian (i. 298.
9 H.) seems to distinguish monosyllabic i (for 2) from dissyllabic &, and
similarly tis (ts) from és, dit (di) from dé, diis (dis) from dé&s (see Neue, 1. ¢.).
The Nom. Pl]. Mase. form used in the Lex Bantina of 133-118 B.¢. (i. 197) and
the Lex Repetundae of 123-122 B.c. (i. 198) is eis (ef. i. 199. 29 ?), a form which
appears in the earlier spelling eeis in the 8. C. Bacch. of 186 B.c. (i. 196. 5,
the only occurrence of a Nom. Pl. of is on this inscr.) (possibly in i. 185 éeis),
and which should be compared with his (older heis) and hisce for hi, illis (ileis)
and ilisce for illi (§§ 16,18). (The reading its in Pacuv. Tray. 221 R. ap. Charis.
133. 4 K. is very doubtful). In the Gen. Pl. the form eum, mentioned by Paul.
Fest. (54. 20 Th. ‘eum’ antiqui dicebant pro eorum), is found once (usually
eorum) on the Lex Julia Municipalis of 45 8.c. [C. I. L. i. 206. 52 eum h(ac) l(ege)
n(ihil) r(ogatur)}. The orthography of the Dat.-Abl. Pl. resembles that of the
Nom. Pl. We have ieis on the Lex Antonia de Termessibus, the Lex Rubria,
the Lex Julia Municipalis, and other inscriptions of the first cent. B.v. (see
Index to C. I. L. i.), but ecis in the early spelling of the 8. C. Bacch. of 186 B.c.
(i. 196), while the earliest form of all, eieis, reappears on an inser. of circ.
100 2.c., the Epistula Praetoris ad Tiburtes (i. 201 de eieis rebus af uobeis
peccatum non esse. Quonque de eieis rebus senatuei purgati estis, &c.). We
have also is in MSS. (often confused with his) and inscriptions (Neue, ii%.
p. 383), and the eis of Republican inscriptions (see Index to C. I. L. i.) admits
of being taken as an expression of the sound is. Priscian, as we have just
seen, appears to distinguish iis (pronounced is) from és, as diis (pronounced
dis\ from dissyllabie déis.

The O, Lat. I-stem Dat.-Abl. ibus is mentioned by Nonius (486, 11 M. ‘ibus’
pro is minus latinum putat consuetudo, cum veterum auctoritate plurimum
valeat), who quotes Plautus, Mil. 74:

latrénes, ibus dinumerem stipéndium,
440 THE LATIN LANGUAGE. [Chap. VII.

(the MSS. of Plautus have latronisbus and latronibus), and other instances from
the Comedians. The Placidus Glossary (C. GL. v. 75. 9) has ‘ibus,’ iis, illis,
with quotation ofthe same line of Plautus. For other lines of Plautus where
some editors read ibus, and for Lachmann’s proposal to read %bus (though only
ibus, hibus are known in Plautus, &c.) in certain lines of Lucretius, see Neue,
ii*. p. 386.

The analogy of the A-declension has produced eae Dat. Sg. F., a form found
in Cato (R. R. 142 quo modo uilicam uti oportet, et quo modo eae imperari
oportet) and probably in Plaut. Mil. 348: hic eae proxumust, but not in use in
Varro’s time (Z. L. viii. 51); so in Late Latin ewm for id (see Georges, Lex.
Wortf. s.v.). Cato also uses eabus (R. R. 152 facito scopas uirgeas ulmeas
aridas. . . eabus latera doliis intrinsecus usque bene perfricato), which is
quoted from Cassius Hemina (pro eis differentiae causa in feminino) by
Priscian (i. 294. 4 H.).

The obscure form necerim mentioned by Festus (160. 21 Th.) and Paul. Fest.
(x61. 11 Th.), and explained by them as ‘nec eum,’ has been supposed by
some to be a wrong reading for nece im (nec eim?), by others to contain the
Pronoun-stem es- of iste for *este (§ 13).

The scansion & is allowed by most editors of Plautus in the Dat. Sg., but
a certain example is very difficult to find, except in the Prologues ; and these
cannot be quoted as Plautine. Thus in Rud. 392 where editors end the line
with : ne cépia esset éi, the MSS. have eius ; in Bacch. 525 : mendadcium ei dixit,
only the Ambrosian Palimpsest has ei, the Palatine MSS. have ili [editors
similarly prefer ilis the (probable) reading of the Palimpsest to iis of the
other MSS. in Rud. 219 : neque quicquam umquam illis préfuit] ; in Cure. 544,
for which we have only the testimony of the Palatine family of MSS. :

is Summanum sé uocari dixit; ei réddidi,

& might be easily emended to & ego, and so on (e.g. Cist. 138). But the
scansion é in the Dat. Sg. is certain for Terence, e.g. Andr. 443:

dum licitumst 6i dumque aetas tulit,

and is very common in Lucretius, e. g. vi. 674:

scilicet et fluvius quivis est maximus ei
qui non ante aliquem majorem vidit ;

so that 7 (like supera for supra, ch. iii. § 15) may be a form that is not genuinely
antique. The Dat, Sg. is avoided by the Augustan poets, Horace for example
in his Satires and Epistles using ejus, ewm, eo but not ei ; but in Ovid, Halieut. 33,
we have é (see Neue, ii’, p. 378, who, however, quotes as examples of a from
Plautus lines where it can be scanned é& in prosodical hiatus, e.g. Cure. 603
mater &% uténdum dederat, like méi honoris, Aul. 463).

§ 20. Ipse. We have already seen the particles -pte and -pse added to
various pronouns to give the sense of ‘self,’ ‘own,’ e. g. meopte ingento like meo
ipsius ingenio, sepse like se ipsam. These particles seem to be composed of a par-
ticle pe- (pi-), seen in quis-piam, quippe, &¢., and the Pronoun-stems so- and to-,
which alternate in the I.-Eur. Demonstrative with its Nom. *sd M., *sa F.,
*tod N., and its oblique cases formed from the stem to- (te-). (Lat. -pte should
therefore not be compared with Lith. pats, ‘self,’ from patis (Lat. potis, Gk.
méas, &e.)]. Similarly we find the Pronoun-stem i- augmented by -pse in the
Latin pronoun of identity, i-pse, which is further strengthened in the O. Lat.
§§ 20, 21.] THE PRONOUNS. IPSE. IDEM. 441

form, or forms, given by Paul. Fest. as psippe (74. 37 Th. ‘ipsippe,’ ipsi, neque
alii), in the Glossary of Philoxenus as ipsipte (‘ipsipti’: adroi, C. @. L. ii. 87. 26
and 44) and ipsipse (‘ipsipse’: adrés, ib. 91. 35), and augmented by -pte in
eopte (Paul. Fest. 78. 16 Th. ‘in eopte,’ eo ipso).

The original declension of ipse we may suppose to have been: Nom. Sg.
*is-pse \i-pse) M., ea-psa F., *id-ptod (*i-ptod) N., Ace. Sg. *im-ptom M., &ec., until
the inconvenience of the alternation of -p-so- and -p-to- brought about
a ‘levelling’ process. (Jpsud is not found till late Latin, see Georges, Lex.
Wortf. 3. v.) The declension of both elements of the Compound may still
appear in isolated forms indicated by the MSS. of Plautus, e.g. eapsa, Cas. 602
{so the Ambrosian Palimpsest (A), but ea ipsa in the Palatine MSS. (P), as in
v. 604 for eapse of A]; eaepsue, Pseud. 833 (eaepse A, eae ipsae P) ; ewmpsum non
eampsam, Truc. 133 (eum ipsum non eam ipsam A, eum ipsum non ea ase P);
eumpsum, Truc, 114 (eum ipsum A, umsum P), which have been, perhaps
unnecessarily, changed by editors to eapse, eaepse, ewmpse, eampse.

The declension of the first element appears in eapse, mentioned by Paul.
Fest. (54. 28 Th. ‘eapse,’ ea ipsa), and this and similar forms are sometimes
preserved in the MSS. of Plautus (e. g. eapse, Trin. 974, Truc. 24, Curc. 161, 534),
though usually ipse is written for the unfamiliar -pse and is sometimes
declined, e.g. ewmpse, Pers. 603 (eum ipse P) ; eampse, Poen. 272 (eam ipse P) ;
eapse, Cas. 604 (so A, but eaipsa P). (On sirempse, see ch. ix. § 8.)

In the Nom. Sg. Masc. we have a byform ipsus in O. Lat. (¢psos in a Law of
Numa quoted by Paul. Fest. 4. 29 Th.: si quisquam aliuta faxit, ipsos Ioui sacer
esto), which 1s the form used by the Comedians before a Reflexive Pronoun,
e.g. ipsus sibi, ipsus suam rem, &c. Ipse is not a development of ipsus. The two
are separate forms, ipse for -sd, ipsus for -sos, corresponding to the I.-Eur. by-
forms *36 and sés (O. Ind. sé and sds, Gk. 6 and és) (§ 13). The influence of
O-stem Adjectives produced in Late Latin ipso Dat. Sg. (Apul.) and ipsae
(Apul.), of I-stems ipsibus Dat. Pl., ascribed to O. Latin by Serg. in Don.
(547. 37K.). The Plautine ipsissumus (Trin. 988), like Aristophanes’ airéraTos,
seems to have been, in the form ipswma or ipsima, actually used in the collo-
quial Latin of the Empire in the sense of domina (Petron. 69. p. 46. 16 B. ; 75.
p. 51. 23 B., &.). (On the colloquial pronunciation tsse, issa, for ipse, ipsa, see
ch. ii. § 81.) Ipsemet is found occasionally (see Georges, Lex. Worl. s. v.).

§ 21. Idem. The affix -dem of idem, O. Lat. is-dem, tanti-dem, tantum-dem,
expressing the idea of ‘ precisely,’ ‘ exactly,’ is thought by some to have been
originally idem, for the Oscan equivalent of Latin idem is is-idum. This idem,
formed by adding the affix -em to the Neut. Demonstr. id, is used in Latin as the
Neut. of the Pronoun expressing ‘the same,’ but in O. Ind. id-dm is the Neut.
of the Demonstr. ‘ this,’ much as the Acc. Masc. with the same affix, *im-em,
is used in O, Ind. (im4m) as the Acc. Mase. ‘this,’ but was in O. Lat. the
equivalent of eundem (Paul. Fest. 54. 2 Th. emem, ‘eundem’; Gl. Philox.
imeum : 7dv adrdv; cf. GI. Place. (hjemdem, ‘ aeque, similiter’; these readings
leave it doubtful whether the word was formed by adding the affix -em to the
Acc. im, or, as seems more likely, by doubling the Acc.]. Priscian (i. 589.
14 H.) derives i-dem from is and demum.

The Nom. Mase. is discussed by Cicero in a passage of somewhat uncertain
text (Orat, xlvii. 157) : ‘idem campus habet’ inquit Ennius, et in templis EIDEM
PROBAVIT; at ‘isdem’ erat verius, nec tamen ‘ eisdem’ ut opimius: male
442 THE LATIN LANGUAGE. [Chap. VII.

sonabat ‘ isdem’; impetratum est a consuetudine, ut peccare suavitatis causa
liceret. He seems to say that idem [also written with ei for the long ¢-sound
(ch. i. § 9) eddem] was the form in use, with s dropped with compensatory
lengthening before d (ch. iv. § 151), but that some purists insisted on the spell-
ing with s, isdem or even eisdem. (For examples of these spellings, e.g. eisdem
C.I.L. i. 5763 577. (2). 9) 11, 133 1468; 1470, &e., soe Neue, ii*®. p. 390.)

In the Dat. Sg. we have gdem and Adem (like ¢ and &, § 19), but adem is
not found. On later inscriptions idem is very frequent (examples in Neue, ii*.
Pp. 390), and even isdem (Neue, !. c.), a confusion with the Nom. Sg., which had
the (archaic) spelling isdem, but the pronunciation idem. In the Nom. and
Dat.-Abl. Pl, Priscian (i. 589. 29 H.) seems to distinguish &%-dem, tis-dem from
zidem (pronounced and usually spelt 7-dem), tésdem (pronounced and usually
spelt is-dem). (For instances of the spelling 7-dem and is-dem, the usual forms,
see Neue, i.c.) The eidem of Republican Inscriptions (Neue, l.c.; e.g. C. 1D.
i. 197. 17 and i. 202) may represent the pronunciation idem. Like eis, older
ecis, as Nom. Plur. of is (§ 19), we find eisdem Nom. Pl. (C.I.L. i, 198. 27
eisdem ioudices; often in the phrase cisdem probauerunt, &e., e.g. i. 11493
i. 1187; i. 1192 ; ef. i. 1143), and in later inscriptions sometimes isdem (Neue,
p. 394). In the Dramatists, &c. with a long second syllable we find the
word pronounced with Synizesis, e.g. &dem, éosdem, &e. like ¢, eds (§ 19).
The influence of O-stem Adjectives is seen in the Late Latin forms eodem Dat.,
eaedem Dat. (very rare ; see Georges, Lex. Wortf. 8. v.).

§ 22. The Pronominal Gen. and Dat. Sg. In the early Dramatists we
find two scansions of illius, istius, ipsius, &c., (1) illidis, istitis, ipstiis, alteriis, &e.,
(2) ili(u)s, isti(w)s, ipsiiu)s, but not illiis, istiiis, ipsitis, &e., as in classical
poetry. Examples of the second scansion are : isti(u)s modi (4 syll.) in Plaut.
Most. 746; Rud. 321 ; Ter. Heaut. 387, &c. Another O. Lat. form of the Gen.,
fully attested by Priscian (i. 196. 22 H.; i. 226. 16H.; i. 266.3 H.; i. 303.
21 H.; ii. 8 H.) is with -%, ili, isti, ipsi, e.g. illi modi and isti modi (Cato), isti
modi Plaut. Truc. 930 (so the MSS.). This cannot be explained as an O-stem
Adjective formation, for it is found with Fem. Nouns, e.g. toti familiae
(Afranius), isti formae (Terence), and a Dat. Mase. like illo, isto is not found
till Late Latin. It may be a relic of the Locative form which, as we have
seen, was augmented by the Gen. suffix -us (os) to form these Pronoun
Genitives in -ius; but it is also conceivable that it is a doublet which has
arisen out of the contracted form illi(w)s, isti(w)s, &c. in certain combinations.
A word-group like isti(w)s-modi would be pronounced istimodi, as naturally as
O. Lat. dusmus became dimus, or *prismus became primus, or is-dem, idem; and
a large number of Priscian’s examples of this Gen. in 7 show the Pronoun in
combination with modus. Similarly ali(u)s-rei would become ali-rei, as dis-rumpo
became dirumpo; ¢f. Priscian’s examples, alii rei causa (Caelius), nullé rei (Cato).
The byforms ili, nulli, &c. having been produced in such combinations would
push their way into other combinations too, e. g. tam nulli consili (Ter. Andr.
608). They do not however seem to be found before a vowel initial.

The Dative in -% is, as we have seen, undisturbed by the influence of the
Q-stem Noun declension till Late Latin, e.g. ilo, isto, qpso (Apuleius). But
a Dat. Fem. in-ae, attested for O. Lat. by Priscian (i. 197. 12 H.; i. 226, 18 H.», is
not unknown in early authors, e. g. illae rei, Cato (R. R. 153 and 154). (Fora list
of examples of these Gen. and Dat. forms, with references to the passages of
§§ 22, 23.] THE PRONOUNS. RELATIVES, ETC. 443

grammarians dealing with them, and for a fuller discussion of the whole
subject, see Luchs in Studemund’s Studien, i. pp. 319 8qq.)

§ 23. IV. RELATIVE, INDEFINITE, AND INTERRO-
GATIVE PRONOUNS. The I.-Eur. Relative-stem *yo- (O.
Ind. yés, Gk. és) does not supply the Latin Relative, which
shows the stem *q*o-,a stem originally proper (with *q*i-, *q*u-)
to the Interrogative and Indefinite Pronouns (O. Ind. kds, ki-tra,
‘where 2,’ Gk. ris, 703, O. Ir. cia, W. pwy, Goth. hvas, Engl. who 2,
Lith. kas, O. 81. kti-to ; O. Ind. cit Neut., Gk. us, &.), but used
also as a Relative (Engl. who, &c.), I.-Eur. *q%o-, *q4i-, *q2u-
appear in the Latin gues’, si-quis, qui?, ali-cubi, &c, (Umbro-
Oscan pis, Osc. pid Neut.). We may roughly distinguish gu7 as
the Relative, guis as the Interrogative and Indefinite Pronoun,
though the stems frequently overlap, e.g. in the Latin of Cato
and the earliest inscriptions gues is the Nom. Pl. of the Indefinite,
qui (quei) of the Relative ; but in class, Lat. both are gu? In
the Italic languages (as in Celtic and elsewhere) a curious
declension of the Relative (and Interrogative) was in vogue,
a case-form of the Relative-stem being prefixed to a Demon-
strative, like modern Gk. 70d rév for 6v in such a sentence as
aités elve 6 dvdpas mod Tov <ida, that is the man whom I saw.
Thus an Abl. Sg. Fem. of the Relative is in Oscan pillad, which
is compounded of the Relative-stem po- (Lat. guo-) (either the
bare stem or a case-form) and *dllad (Lat. *od/dd), the Abl. Sg.
Fem. of the Demonstrative ollo-; another is poizad, a similar
compound with the Abl. Sg. Fem. of the Demonstrative eiso-.
The old spelling of the Dat. Sg. of Lat. gui, viz. guoiei, shows
it to be a compound of this kind, having for its second element
the Dat, Sg. of is (O. Lat. eiei), *gud-eier ; and Gen. Sg. quovus
(class. cujus) will consequently represent *guo-cius. Whether
this method of declension was used in other cases in Latin does
not appear. Another feature of the Italic Relative is its tendency
to append the Pronominal particle i (cf. Gk. ovros-t), e.g.
Umbr. poi Nom. 8g. M., porsi (*podi) N.; Lat. gui for *q*o-i
(O. Lat. quot). ‘ .

The Latin Interrogative-Relative has a Possessive cujus -a
-um, older quoius -a -um, which is very frequent in Plautus and
Terence, and is found in Republican inscriptions (e.g. in the
444 THE LATIN LANGUAGE. (Chap. VII.

Lex Repetundarum, i. 198. 5, 10, 29 quoium nomen and quoiaue
in fide), in Cicero (eg. Terr, I. 1. 54. 142 cuja res), and
Virgil (4. iti, 1 cujum pecus). Virgil’s use of the word was
objected to by purists, apparently through an idea that cujus
-a -um was a vulgar inflexion of the Gen. Sg. of the Pronoun,
an idea which the occurrence of the word in Cicero and in State
inscriptions disproves ($11). It is rather formed by means of the
Adjectival suffix -yo- (ch. v. § 4) from the stem *q*o-, *q*o-yo-, as
meus (stem *me-yo-) from the stem *me- (§ 1). It is not till
Late Latin that we find the particle -ce added to the Relative-
Interrogative Pronoun in ewjuscemodi (Apuleius, &c.), a word
coined after the type of hajuscemodi.

§ 24. Stems qti- and q¥o-. In 0. Lat. there is a usage of quis, possibly as
a Relative, but rather in the sense of siquis or quicunque ; e. g. in an old treaty
quoted by Festus to illustrate O. Lat nancitor for nanciscitur (170. 25 Th.) :
pecuniam quis nancitor, habeto; in an old plebiscitum (Fest. 322. 11 Th.) :
eum quis uolet magistratus multare, dum minore parti familias taxat, liceto ;
on a public notice affixed to a grove at Luceria (C. I. L. ix. 782): quis uolet
(other examples from Cato and from Cicero's laws are given by Neue, ii®.
p. 430, e.g. Cato, R. R. 147 dominus uino quid uolet faciet’. Cf. O. Lat.
necumquem explained by ‘ nec umquam quemquam’ (Fest. 162. 22 Th. ; Paul.
Fest. 163. 12 Th.). (So Umbr. pisest totar Tarsinater ‘quisquis est civitatis
Tadinatis,’ Ose. pis hafiest ‘qui habebit’). In the Dramatists quis is the
Fem. of the Interrogative, quae of the Relative (cf. Prise. ii. 8. 21 H. quis
etiam communis esse generis putaverunt vetustissimi, sicut apud Graecos tis).
e.g. Plaut. Pers. 200 quis haec est, quae me aduorsum incedit? (other
examples in Neue, p. 441), but the distinction of qui Adj., and quis Pron, in
questions, e.g. gui homo venit? and quis venit? is hardly observed, the habit of
Plautus being rather to use quis before a vowel, qui before a cons. (see Neue,
p. 431, and B. P, W. xiii. 278; similarly Cornificius seems to write siqui before
an initial s, otherwise siquis, v.g. siqui swadebit, iii. 5.8). The I-declension
form of the ‘Abl.’ Sg. qui, e. g. quicum, is Relative as well as Interrogative and
Indefinite ; e.g. Ter. Ad. 477 psaltriam parauit, quicum uiuat; 0. I. L. i. 200
queiue ab eorum quei emit (see Neue, pp. 455 sqq.). But the Nom. Plur. ques,
attested by Charisius (91. 16 K. ut duces, ducibus, mores, moribus, et ‘ques,’
quibus; 158. 21 veteres nominativum pluralem ‘ques’ dixerunt regulam secuti,
unde etiam dativus mansit in consuetudine), Festus (348. 23 Th.), Priscian
(ii. 9. 13H.) &¢:, seems to have been confined to the Interrogative and Indefinite
use. Thus Cato began his Origines with the words: siques homines sunt, quos
delectat populi Romani gesta describere ; on the 8. C. Bacch. (C. I. L. i. 196)
we haye: sei ques esent, quei sibei deicerent necesus ese Bacanal habere,
‘siqui essent, qui sibi dicerent necesse esse Bacchanal habere’; and a line
of Pacuvius (Trag. 221 R.) runs:

ques sunt? ignoti, néscioques ignodbiles ;

(other examples in Neue, p. 466).
§§ 24, 25.) THE PRONOUNS. RELATIVES, ETC. 445

. The I-declension Neuter guia survives only as an Adverb or Conjunction,
e.g. O. Lat. quianam, ‘why ?,’ like quidnam ; the Gen. quium can hardly be
ascribed to Cato on the mere testimony of Servius (ad A. i. 95 : denique Cato
in Originibus ait: si ques sunt populi. Et declinavit ‘ques,’ ‘quium’ ut
‘puppes,’ ‘puppium’), for this remark only implies that Cato’s ques was an
I-stem formation. The I-stem Dat. quibus supplanted the O-stem quis (older
queis) in the Relative, though the O-stem form is by no means uncommon
(see a long list of instances in Neue, ii®. p. 469, e.g. Plaut. Most. ro4o :

quis méd exemplis hédie eludificdtus est).

§ 25. Case-forms. The original Nom. Sg. Mase. of the Relative quo-i (stem
qu6- with affix -i of Gk. obroa-7, &e.) [Ose. pui(?), Umbr. pot; the Umbr. shows
this affix also in Nom. Pl. Mase. pur-i, Acc. Pl. Fem. paf-e, &c.] is probably
intended by the spelling goi of the very ancient Dvenos inscription [Zvet.
I. I, I. 285 qoi med mitat, ‘qui me mittat (mittet)].? The weakened form
quei, due to the unaccented use of the Relative (ch. iii. § 18), is common on
inscriptions of the Republic, from the Scipio epitaph (in Saturnians) of c¢.
200 B.C. (C. I. L. i. 30) :

eénsol cénsor aidilis quei fuit apad uos,

to the Lex Julia Municipalis of 45 B.c. (i. 206), which has quei far more
frequently than qui; also in the MSS. of Plautus, &c. (see instances in
Georges, Lex. Wortf. s. v.). The Fem. of the Indefinite Pronoun in its Adjec-
tival use followed the Adjective Declension in taking usually the suffix -a in
class, Latin, e.g. siqgua causa est (but seiquae causa erit on the Lex Repetun-
darum, C. I. L. i. 198. 37; si quae lex on the Lex Agraria, i. 200. 41; and in
Plautus nwmquae causast guin, &e.; see Neue, ii. p. 445). Gen. Sg. quoius
appears in the Saturnian Scipio epitaph of ¢. 200 B.c. (C. 1. L. i. 30):

quoius forma uirtutei parisuma fuit,

in the Lex Repetundarum (i. 198), and other inscriptions, and in the MSS.
of Plautus, Varro, &c. (see Georges, s. v.; Neue, p. 450). When the of sank
to ui, the gu by a law of Latin phonetics became ¢ (ch. iv. § 137) ; hence cujus
(pronounced ciiy-yus, ch. ii. § 55), a pronunciation indicated by spellings like
cuiius, cuilus (Neue, p. 451). (On monosyllabic quoi(u)s in quot(u)smodi, &c.,
in the Dramatists, and quoi in quoiquoimodi, cuicuimodi in Cic., &e., see
§ 22). The oldest form preserved of the Dative is quoiei of a Scipio epitaph
of ¢. 130 B. c. (C. I. L. i. 34) in a Saturnian line referring to the short life of

the deceased : ; ;
quoiei uita defécit nén honos hondére,

in the Lex Repetundarum of 123-122 Bc. (i. 198. 10, usually quoi), and in the
Lex Agraria of 111 B.c, (i. 200. 68, usually quoi ; cf. quoieique, ll. 3. 6. 32. 45.
99) ; though the dissyllabic Dative seems not to occur in Plautus (it is read by
some editors in Trin. 358. 558, &c. ; see Brix ad locc.), and so cannot have been
used in the ordinary conversation of his time. The common form in use
on Republican inscriptions is quoi (see Index to ¢. J. L. i.) ; the Lex Julia
Municipalis, for example, of 45 B.c. (i. 206) has always quoi, never cui, as
it has always quoius, never cujus, and always, except in two instances, quei for
qui (so quoi in the Comm, Lud. Saec. of Augustus’ reign). And Quintilian (i. 7.
446 THE LATIN LANGUAGE. (Chap. VII.

27) tells us that in his youth quoi was the spelling. ([Cf. Velius Longus
(first cent. a. D.), p. 76. 3 K. itaque audimus quosdam plena oi syllaba
dicere ‘quoi’ et ‘hoic’ pro cui et huic.} In late poetry cui is a dissyllable,
citi, a scansion which is found as early as Seneca, Juvenal, and Martial, e. g.
Mart. i. 104. 22 sed norunt cui serviant leones (examples in Neue, p. 454).
Acc. quem has -em for -im (Ose. pim) by analogy of Noun I-stems which
usually substituted the Cons.-stem -em for the I-stem -im, e. g. turrem (ch. vi.
§ 29). As to the Abl. Sg. we have not sufficient means of deciding, owing to
the absence of sufficiently old inscriptions, whether and in what uses qui,
O. Lat. for quo, was an Ab], (from *quid), a Locative (from *quei), or even an
Instrumental (from *qui). But the spelling quiguam (‘in any way’) on the
S. C. Bacch. (1. 12), an inser. on which original ¢ and original e seem to be
kept distinct, goes against the Locative theory, unless indeed it is a mere
mistake for quisquam (see § 28). The spelling quei on the Lex Agraria of 11x
B. u. (C. I. L. i. 200. 17 queiue ab eorum quei emit) merely expresses the sound
qua (see ch.i. § 9). This qui, byform of qué, qua, occurs in Virgil (A. xi. 822) :
Accam ex aequalibus unam,
quicum partiri curas.

It was not in living use in the time of Servius (fourth cent. a.p.) (én Donat.
p. 411. xz K. nam dicimus ‘a quo’ venisti et ‘a qui’ venisti; sed ‘a qui’ in
usu esse desiit). Already in Terence quicwm is not so frequent as in Plautus.
In the early authors qui is used : (1) as anordinary Abl., e. g. Plaut. Capt. 828
qui hémine nemo uiuit fortundtior ; Bacch. 335 sed qui praesente id aurum
Theotimé datumst?; but especially (2) as an Abl. of the instrument, e. g.
Plaut Men. 391 quis istést Peniculus? qui éxtergentur bdxeae?, ‘what
Mr. Brush do you mean ? one to clean shoes with?’ ; Varro, R. R. ii. praef. 3
frumentum qui saturi fiamus. In this capacity it passes into a mere Adverb
like wt, and is used even with a plural noun, e.g. Plaut. Stich. 292 quadrigas
qui uehar ; Pseud. 487 (uiginti minas) quas meo gnato des, qui amicam libe-
ret ; (3) as an Adverb with the sense of quomodo; e.g. qui fit ut?, a phrase in
common use in classical Latin ; Ter. Adelph. 215 quipotui melius?; (4) as an
enclitic Particle, e. g hercle qui, edepol qui, utinam qui (Plaut.), a usage surviving
in classical at-qui. (For a fuller list of examples, see Neue, ii*. pp. 455 sqq. ;
cf. below, ch. x. § 5.) The Nom. Pl. Mase. is written que on Republican
inscriptions ‘see Index to C. J. L. i.), this being the weakening of an original
*quoi, due to the unaccented nature of the Relative (ch. ili. § 12a. 5), & g. i.
196 sei ques esent, quei sibei deicerent necesus ese Bacanal habere. (On
ques Nom. Pl. of quis in O. Lat., and on queis and quis, Dat.-Abl. Plur., from
*quois, see above, § 24.)

§ 26. The stem q%u-. Corresponding to Umbr. pu-fe, Ose. pu-f, ‘where’
‘0. Ind. ktha, O. Sl. kiide, from I.-Eur. *qtudh-), is Lat. -ciibt of dli-cubi,
‘somewhere,’ si-cubi, ‘if anywhere,’ né-cubi, num-cubi, &c., with cu- instead of quu-
by the same phonetic law of the Latin language that has made quincu-plex out
of *quinquu-plea (ch. iv. § 137) (ef. Vulg. Lat. nescio-cube, Probi App. 199. 16 K.),
and clearly connected with the Interrogative, Indefinite, and Relative Pro-
noun-root (see ch. x. § 1, on the existence of parallel stems of Pronouns in -o,
-i, -u). But apart from compounds the Latin Adverb is ubi. Similarly the
-cunde of cili-cunde (aliquonde is written in Plaut. Pseud. 317 in the Ambr. Palimps.;
so Caesellius ap. Cassiod. 202. 28 K. aliquonde per quon debet scribi), si-
§§ 26-28.] THE PRONOUNS. RELATIVES, ETC. 447

cunde, né-cunde, wants initial c- in the simple form -unde. According to Schmidt
(K. Z. xxxii. 405) this Latin initial u-, internal cu-, represents I.-Eur. *q®u-,a
parallel stem of q¥o-, qii- (cf. § 23). (For another possible explanation of
this loss of the initial guttural, see Brugm. Grundr. i. § 43t¢c) The Latin
Interrogative and Indefinite Pronoun titer, ‘whether of two’ (on the suffix
-tero-, see ch. v. § 18) showsa similar relation to O. Ind. kitra, ‘whither’ (stem
*quutro-, *kutro-), Ose. potro-, ‘ whether of two ’ (stem *qtotro- or *kwotero- ;
both derivatives of a stem *que-t(e\ro- or *kwe-t(e)ro- ?). (On other Latin
Adverbs beginning with u-, derived from the Interrogative root, and wanting
an initial guttural, see ch. ix. § 10.) Ritschl’s proposal [based on Trin. 934,
where there is a manuscript corruption cubitus for ubi tus (gignitur)] to read
cubi, &e., in some passages of Plautus, has not met with approval (Opuse. iii.
135).

§ 27. The Possessive cujus. The case most frequently in use in Plautus
and Terence is the Nom. Sg. Fem., e. g. Plaut. Trin. 45 quoia hie uox prope
me sonat?; cf. Cic. Very, II. iii. 7.16 ne is redimeret, cuja res esset; ib.
27. 68 Apronius certiorem facit istum, cuja res erat. The Plural is very
rare, Plaut. Rud. 745:

argentum ego pro istisce ambabus quoiae erant dominé dedi,

Trin. 533, with quoium for cujorum, Gen. Pl. (?) (cuium A, quoium B ; some read
quorum) :

necinquam quisquamst, quoius ille agér fuit,

quin péssume ei res udrterit, quoitim fuit,

alii éxolatum abiérunt, alii emértui.

With the suffix -ati-, denoting the country of one’s birth (ch. v. § 45), we have
cujds, O. Lat. quoiatis, ‘what countryman ?’ ‘belonging to what country ? *

§ 28. Other derivatives. Ali-quis, some one, is a compound of the stem
ali-. some (connected with dlio-, other)' and the Indefinite Pronoun, like aii-
cubi, somewhere, ali-cunde, &e. An O. Lat. Nom. Pl. aliques is mentioned by
Charisius (159. 7 K.). Ec-quis has been explained as nothing else than et quis,
with c for ¢ by the same phonetic law as reduced *sit-cus (cf. séitis) to sic-cus
(ch. iv. § 159), but it is more likely to come from the pronominal stem é- (used
as a prefix in e-nos, Umbro-Ose. e-tanto-, &c.) with the appended particle -ce
(without this appendage in e-quidem, ch. x. § 6, and in équis, a byform of
ec-quis, in Plaut.). The Neut. ecquid in Plautus often sinks into a mere con-
junction, e. g. ecquid placent? Most. 906, &e.

Quidam for *quis-dam (like idem for is-dem, § 21) appends the particle -dam to
the Indefinite Pronoun. The Nom. Plur. Mase. is not found in Plautus, and
scarcely indeed in any of the older authors. We have quesdam Acc. Pl. in
Accius, Trag. 477 R. Quivis (ef. Umbr. pis-her from the verb heri-, ‘to wish’)
may stand for *quis-vis, ‘any you please,’ ‘whosoever you please’ (cf. O, Lat.
quis nancitor, ‘ whosoever obtains,’ § 24), with the same loss of s before initial
v in a Compound as divello for *dis-vello (ch, iv. § 151), and quilubet, for *quis-lubet,
like diligo for *disligo (ib.), as well as for qui-vis, qui-lubet. The I-declension
Abl. quiuis appears in Plaut. Stich. 627: quicumuis depigno multo facilius

 

1 aliquis alius is not found in Plautus, though it occurs in Terence.
448 THE LATIN LANGUAGE. [Chap. VII.

quam cum fame. Qua-cum-que, O. Lat. quei-quom-que (C. I. L. i. 197. 5 3 200. 50 ;
202, &c.), had probably a byform *guis-quom-que, to judge from Nom. Pl. ques-
cumque un Cato (Orig. ii. fr. 34 J. quescumque Romae regnauissent). The latter
part of the word is probably the Adverb quum (O. Lat. quom), when, with the
enclitic particle -que, so that -cwmque means literally ‘whenever’ (see ch. ix.
§.10. 7). In O. Lat. quisque is used in the sense of quicunque, e.g. quemque
offendero, Plaut. Capt. 798, the particle -que (O. Ind. cain kag-ca, ‘whoever,’ &e.,
Hom. Gk. ve in és ve, Goth. -h in hvo-h) giving to a word the sense of our
‘ever’ in ‘whoever,’ ‘whenever,’ &c. (see ch. x. § 2); but has in classical
Latin the sense of ‘each.’ It is fem. as well as masc. in O. Lat., e.g. Ter.
Hec. 216 quisque uostrarum ; so quemque Acc. Sg. F. in Plaut. Pseud. 185.
Another expression for ‘whosoever ’ is quisquis (Ose. pispis, of which the Neut.
pitpit is mentioned by Paul. Fest. 263. 8 Th.), (riers occurs in an Old Gk.
inser., Mon. Antichi i. 3. p. 594), « doubling of the Indefinite Pronoun’. We
have quisquis Fem. in O, Latin, e.g. Plaut. Cist. 610 mulier quisquis es (cf.
Nonius, 197. 30 M.). The Neuter quicquid, a byform of quidquid, shows the same
assimilation of d (¢) before a guttural as accurro for adcurro (ch. iv. § 160). The
shortened form of the Gen. Sg. of quis, current in ordinary pronunciation
when not specially emphasized, viz. quoi(u)s, cui(w)s, (§ 22), appears without
its final s in Cicero’s cwicuimodi; and the I-stem Ablative quiqui occurs more
than once in Plautus in the phrase cum co, cum quiqui, anyhow, at any cost,
lit. ‘with that thing or with whatsoever thing’ (Poen. 536. 588). A curious
passage, Cas. 523 :

séd facito dum, mérula per uorsts quod cantat, ta colas :
‘eum cibo, cum quiqui’ facito ut uéniant,

suggests that Roman children interpreted the alarm-note of the blackbird
into the words: cum cibo, cum quiqui, sc. veni, ‘come along! food or no
food.’ (Class. Rev. vi. 124.) Quis prefixed to the Adverb quam, quis-quam had
the sense of ‘any,’ and was used especially in negative sentences. Examples
of its use as Fem. in O. Lat. are Ter. Lun. 678 nostrarum numquam quis-
quam uidit ; Plaut. Rud. 406:

neque digniorem cénseo uidisse anum me quémquam,

and of the I-declension Abl., Plaut. Pers. 477 nec satis a quiquam homine
accepi [see § 25 on qui-quam Adv., ‘anyhow’ (?), of 8. C. Bacch.: neue
pro magistratud neque uirum neque mulierem quiquam fecise uelet, like
nequi-quam, in vain, lit. ‘not anyhow,’ always so spelt in the Ambrosian
Palimpsest of Plautus?. Another compound of the Indefinite Pronoun,
quis-piam, some, the formation of which was obscure to the Roman gram-
marians (Festus 338. 28 Th. ‘quispiam’ quin significet aliquis, et ‘quae-
piam ’ aliquae, similiterque alia ejusdem generis, ut dubium non est, ita unde
sequens pars ejus coeperit, inveniri non potest), may be a compound of
*quis-pe, of which quippe for *quipe is an Adverbial case-form (ch. x. § 7), with
Jam, as nunciam of nunc with jam (ch. iv. § 67). Corresponding to -quam, -piam
of Lat. quisquam, quispiam is Oscan -um of pid-um ‘quidquam,’ pieis-wm
‘cujuspiam.’

 

1 So quantus quantu’s, ‘every inch of you,’ Ter. Adelph. 394; quantum quantum,
Plaut. Poen. 738; quaequatis in a poetical inser., 6. I. L. vi. 6314.
§ 29.] THE PRONOUNS. PRONOMINAL ADJECTIVES. 449

§ 29. V. THE PRONOMINAL ADJECTIVES. Theseare:
(1) dius, I.-Eur. *alyo- (Arm. ail, Gk. dAdos, O. Ir. aile,
Goth. aljis, Engl. else), Osc. allo- (on the Bantine Tablet), with
a byform alis in the Nom. Sg. Mase. (e.g. Catull. Ixvi. 28), alid
in the Nom. Sg. Neut. (eg. Luer. i. 263). These byforms,
which are to be compared with Cornelis, &c. for Cornélius (see
ch. vi. § 5), come into special use in the phrase alis alium (e. g.
ad alis alium and cum alis alio in the Vulgate, alis alium, C. I. L.
li. 2633.7), probably owing to the fact that the stress of the
voice in this word-group fell on the antepenultimate syllable (see
ch. ii, § 12). The deriv. Adj. adiénus is for *ali-ino- (ch. iv. § 12).
(2) alter is formed from the root al- (al-?), seen in *alyo-, by
the addition of the suffix -tero- (ch. v. § 18), (Ose. alttro-),
while other I.-Eur. languages show a similar formation from
the root an-, seen in O. Ind. anyd- ‘alius’ (O. Ind. 4ntara-,
Goth. an-par, O. Engl. dper, Germ. ander, Lith. an-tras). It
is often used as an Ordinal Numeral, ‘second.’ A stem altro-
appears in altrin-secus, altro-vorsum (Plaut. Cas. 555) and other
words (cf. Gk. dAAdrpios), but in lines like Plaut. Bacch. 1184
alterum, &e. need not be changed into altrum, &c., but may be
scanned as dactyls, altériim, &e. (see Klotz, Altrém. Metrik, p. 59) ;
even in Pers. 226 altra is not certain,
(3) alas is formed by the LO-suffix (ch. v. § 21) from tinus
(see ch. iil. § 15. 5), which also belongs to the Pronominal Declen-
sion, Gen. unius, Dat. uni, (With quisque appended we have the
word-group or compound wnzs-guisqgue.) The opposite of aus
is willus with the negative prefix ué- of n-usguam, n(e)-utiquam
(pronounced with both first and second syllable short) and the
like (ch. ii. § 149); and *xe-adlus, ‘not a little one, ‘not even
one,’ was probably anterior in formation to wus, in which the
force of the Diminutive suffix is not so apparent. Like nudlus,
but used properly of persons, while zwd/us was used normally
of things, is zémo from né-hemo. (On emo, a byform of homo,
see ch. vi. § 1.) Nudlus is hardly used as a substantive till Late
Latin, but nullius and nudlo take the place of neminis and nemine
in class. Latin. As the Neuter of zdlus, nihit is used, a com-
pound of the negative né- and hilum (see ch. iii. § 52; ch. vi. § 11).
(4) sddus may be connected with the Adverb sé-, séd-, apart

Ge
450 THE LATIN LANGUAGE. (Chap. VII.

(e. g. séd-atio, lit. ‘ a going apart’) (ch. ix. § 51), and be formed of
sd-, a grade of sé- (ch. iv. § 53), with the suffix -lo. Some refer
the word to the stem so//o- of Ose. sollo-, ‘ whole, Lat. soé-ers,
soll-emnis, but the connexion of theideas * alone’ and ‘whole’
or ‘all’ is not apparent.

(5) totus is another word of uncertain etymology. Some
connect it with the Umbro-Oscan word for a community, state
or people, *teuta- (Ose. tovté-, Umbr. tota-), and suppose the
Latin word to exhibit another grade of the root, perhaps *touto- (see
ch. iv. § 41). The word encroached on the sphere of omnis, and
finally supplanted omnts in Vulgar Latin. Of the Romance lan-
guages Italian is the only one which preserves Lat. omuzs (Ital. ogni)
beside Lat. ¢d¢us (Ital. tutto ; on this form, see ch. ii. § 130. p. 116).

(6) wer is one of those Latin Relative (Interrog., Indef.)
forms beginning with w- like #b/, #¢ which are discussed in § 26.
With the addition of -gue, ‘ever’ (see on quis-que, § 28), it
becomes wter-que. The Umbro-Oscan stem is *potro- (Ose.
putards-pid Nom. Pl. ‘ utrique,’ piteref-pid Loc. Sg.; Umbr.
sei-podruh-pet ‘utroque’ Adv. is a formation like sed-utraque
Nom. Sg. Fem., ‘each separately, Plaut. Stich. 106). The
opposite of wter is neuter, a trisyllable (ch. ii. § 32), with the
negative prefix. Alter-uter is a compound of alter and uder,
sometimes with both elements declined, sometimes with the
second only (cf. § 20 on ipse, Fem. ea-psa, ea-pse and i-psa).
A form a/tertra for alterutra is mentioned by Paul. Fest. 6. 2 Th.

All of these take the pronominal Gen. and Dat. Sg. in -cws, and
-2, but only alius takes the Neut. Sg. (Nom.-Acc.) in -d. Still
they admitted more readily than id/e, iste and the other Demonstra-
tive Pronouns the Noun Declension forms in these cases ; e.g. waae
ret (Gen.), Cie. Tull. xv. 36 ; tam nulli consili, Ter. Andr, 608 ; coloris
ulli, Plant. Truc. 293 ; alterae legioni, Caes. Bell. Gall. v. 27. 5 (see
Priscian, i. 196. 18 H. and Neue, ii*. pp. 516 sqq.). For the Gen.
Sing. of alius the Romans discarded a/ius, which was liable to confu-
sion with the Nom.,and used the Gen. Sing. of a/ter instead, alterius
(in dactylie poetry, of course, only a/¢ertus is admissible, but alterins
occurs in other metres, e.g. Ter. And, 628, Seneca, Here. Fur. 21 2).

There are other Adjectives called ‘ Pronominal’ Adjectives,
which are derived from Pronoun-stems, but which do not share
§ 29.] THE PRONOUNS. PRONOMINAL ADJECTIVES. 45!

the Pronominal Declension. From the stem to- (te-) comes Lat.
tdlis (I.-Eur. *tali- of Gk rndi-xos;, cf. O. Sl. toli Adv., ‘so
very, toli-ki Adj., ‘so great’), tan-tus (Osc. e-tanto-, e. g. molto
etanto estud ‘ multa tanta esto’; Umbr. e-tanto-, e.g. etantu mutu
adferture si ‘tanta multa affertori sit’) (¢axtisce pro tantis C. G. L.
v. 155+ 36), ¢6¢, older foti-, preserved in ¢6¢i-dem (I.-Eur. *toti,
O. Ind. tati; ef. Gk. ré0(o)os for *rorios), and (with O-suffix) ‘dtus
(e. g. Manil. iii. 420 detrahitur summae tota pars, quota demitur).
From the Relative (Interrog., Indef.) stem comes Lat. gudlis
(Gk. andi-cos ; ef. O. SI. kolf, koli-ki), guantus (Umbr. panto-),
quot, older quoti (I.-Eur. *q*ti, O. Ind. kati; cf. Gk. aé0(c)os
for *zorios), and (with O-suffix) gudtus (e.g. Hor. tu quotus esse
velis rescribe). (On cotti-die, see ch. ix. § 5.) The Late Latin
use of ¢anti, quanti for tot, quot (e.g. Tertull. nec tamen tantos
inveniunt verba discipulos, quantos Christiani factis docendo),
survives in Romance, e. g. Ital. quanti anni ha ?, ‘ quantos annos
habet?’, ‘how old is he?’ (On cé-ter?, see ch. iv. § 33.)

A Dual, like the Numeral duo (ch. vi. § 59), is ambé -ae -6
(I.-Eur. *ambhd(u), Gk. dud), Gen. amborum, &e., Dat. ambobus,
&e., Ace. ambés and ambé -as -6; in the Acc. the older ambé was
being ousted by ambds even in Plautus’ time, for he uses ambos
always before an initial consonant, and amd with ambos before an
initial vowel as the metre requires, while in Late Lat. we have ambis
for the Dat.-Abl. Plur. e.g. Eph. Hpigr. iv. p. 491 (cf. Caper 107.
14 K. ambobus,non ‘ambis,’ et ambabus ; and see Neue, ii’, p. 279).

The Pronominal Declension has in some languages (e.g.
Lithuanian and the Teutonic languages) extended itself from
these Pronominal Adjectives to all Adjectives!. What is called
in Teutonic the ‘ Strong Declension ’ of Adjectives, in Lithuanian
the ‘ Indefinite’ is really the Pronominal, e. g. Goth. blinds with
Neut. blindata, ‘ blind,’ like Neut. pata, ‘that.’ In Greek, on the
other hand, the Pronominal Declension has lost ground, and
that is why in Greek the declension of the Pronouns does not
appear so unlike the declension of the Nouns as in Latin.

 

1 We have seen ‘ch. vi. § 46) that to Adjectives (e. g. duonoro on a Scipio
in Latin the Gen. Pl. suffix of O-stems, Epitaph of perhaps the end of the
-dsom, then -drom, class. -drum, wasex- _ third cent. B.c.), and in time to Nouns
tended from Pronouns (e.g. olorom (e.g. deorum competed with deum in
‘jNlorum’ on the Columna Rostrata) _Cicero’s time, ch. vi. § 47).

eg 2
452 THE LATIN LANGUAGE.

§ 80. The Pronouns in Romance. The development in the
Romance languages of two series of the Personal and Possessive
Pronouns, (a) the enclitic, e.g. Fr. me, te, mon, ton, (4) the
accented, e.g. Fr. moi, toi, mien, tien, has been already mentioned
(ch. iii. § 12 a, 3, and above, § 12). In Vulgar Latin ego became
eo, whence the Romance forms, e. g. Ital. io, Span. yo, Sard. eo,
O. Fr. eo, io, eu, jo, now je. Spanish nosotros ‘ we,’ vosotros ‘ you’
represent os alteros, vos alteros (cf. Fr. nous autres, &e.). The 2 Pl.
Possessive remained voster in Vulg. Lat., whence Ital. vostro, Fr.
votre, Span. vuestro, &c., while two rival forms competed for the
3 Pl. Possessive, swws (Span., Port.) and id/orwm (Ital. loro, Fr. leur).

The Demonstratives idle, ivse (which takes the place of 2/e in
Sard. as 3 Sg. Pron. and Article, isse and issu, ‘ he’), zs¢e had in
Vulg. Lat. -2 in Nom. Sg., -vi in Dat. Sg. Mase., -aec in Dat.
Sg. Fem.,e. g. di (Ital. egli, Fr. il), dui (C.L. L. x. 2654; Ital.
lui, Fr. lui), 22/aei (Ital. lei, O. Fr. li). Their extension by the
prefix ecce has been noticed in § 15, e.g. Fr. celle from eece-2éla.
Used enclitically, ie has lost its first syllable in Romance, e. g.
Ital. gli (from 777i), lo (from i/wm), Span. le, Sard. li. The
Definite Article in most Romance countries was supplied by 7//e
(but by ‘se in Sardinia and elsewhere), which in Roumanian is
postfixed (e.g. domnu-l for dominus ile), and in all Janguages is
united with a Preposition into one word, e.g. with the Preposition
ad in Ital. al, Fr. au, &. Hic survived only in word-groups,
e.g. Ital. cid from ecce hoe.

Of the Indefinite, Interrogative and Relative Pronouns quae
and guis seem to have been ousted in Vulg. Lat. by gui; quid is
Ital. che, and (accented) Fr. quoi, (unaccented) Fr. que.

For the Pronoun of Identity (ipse had become a Demonstrative)
various phrases were used ; e. g. Ital. medesimo, Prov. medesme,
Fr. méme, Span. mismo are from met ipsimus ; Ital. desso from id
ipsum (or ad ipsum 2); Ital. stesso is a compound of zste and ipse.

Alius survived in a Neuter form *alwm, whence O. Fr. e] and al,
O. Span. al, alter having taken its place (Ital. altro, Fr. autre, Span.
otro). Certus (and certdnus) replaced guidam. Aliquis was joined
with wus into a Compound *alicunus, whence Ital. aleuno, Fr.
aucun, Span. alguno, (See Meyer-Liibke Rom. Gram. ii. pp. 89
8qq-, 595 844.).
CHAPTER VIII.

THE VERB.

$1. I. THE CONJUGATIONS. The I.-Eur. Verb had two
Conjugations, (1) the Thematic, in which the Person-suffixes were
attached to the verb-root augmented by -é- or -b-,e. g. Gk. pép-
o-uev, pép-e-re; (2) the Athematic, in which this vowel, the
Thematic Vowel, as it is called, was absent, e.g. Gk. tord-pev,
toré-re. In the Thematic Conjugation the 1 Sg. Pres. Ind.
Act had -é, e.g. pép-w, Lat. deg-6; the Subjunctive changed the
Thematic Vowel of the Indicative to -é- (and -s-), e.g. Gk. dép-n-re
(pép-w-pev) ; the Optative changed it to -oi-,e. 2. Gk. pép-o1-pwer,
ép-o1-re. In the Athematic Conjugation the 1 Sg. Pres. Ind.
Act. had -mi, e.g. Gk. tory-y; the verb-stem was weakened in
the Dual and Plural Act. and in all Numbers of the Middle, e.g.
Gk. tord-rov, tord-yev, lord-yar beside Sg. torn-; the Subjunc-
tive (with strong stem) showed -é- or -6- between the root and
the Person-suffixes, and the Optative (with weak stem) -yé- in
Sg. Act., -I- elsewhere, e.g. Gk. tord-(y-v, iord-i-pev (ioraiuer).

We find early Derivative Verbs like I.-Eur. tr-4- from the root
ter-, pl-é- from the root pel- (Lat. in-trd-re, im-plé-re), and later
Derivatives from Nouns, e.g. Lat. curd-re from the Noun-stem
cura-, forming the persons of their Present Tense sometimes
thematically with the suffix -yé- (§$ 15), sometimes athematically,
e.g. O. Ind. tra-y4-té and tra-ti 3 Sg. The long vowel, with which
these derivative verb-stems end, is not weakened in the Dual and
Plur, Act., nor in the Middle, e. g. O. Ind. tra-sva 2 Sg. Imperat.

Mid.
454 THE LATIN LANGUAGE. (Chap. VIII.

In Latin almost every athematic verb becomes thematic in
1 Sg. Pres. Ind., and usually in 3 Pl. ; and the declension of the
Pres. Ind. often shows thematic and athematic forms side by
side. Thus I.-Eur. *ei-mf, ‘I go’ (O. Ind. é-mi 1 Sg.,i-més 1 PI.,
Gk. ef-w. 1 Sg., t-pev 1 Pl., Lith. ei-mi) is in Latin eo for *éy-6,
a thematic form, though other Persons, e.g. 2 Sg. 7-s, older ei-s
(*ei-s(i)), are athematic ; I.-Eur. *wél-mi, ‘I wish’ (Lith. pa-
velmi 1 Sg., pa-velt 3 Sg.) is in Latin thematic in 1 Sg. vd/-0,
but athematic in 3 Sg. va/-4. The Latin Substantive Verb sum
has best retained the features of the Athematic Conjugation,
with its 1 Sg. Ind. in -m, its Opt. originally declined s-cé-m, s-ié-s,
s-1é-t, *s-i-mos, *si-tes, and so on.

The four Conjugations of our Latin Grammars, (1) amd-re,
&e., (2) vidé-re, &e., (3) legé-re, &e., (4) audi-re, &e. are, like the
five Declensions (ch. vi. § 1), an unscientific classification, often
bringing forms together which were of dissimilar origin, just as
Modern Italian with its three Conjugations brings together in
the Second forms like vendére (Lat. vendére), potére (Lat. posse),
solére (Lat. sd/ére), and in the Third dire (Lat. dicére), empire
(Lat. implére), apparire (Lat. apparére), seguire (Lat. séqui), and
venire (Lat. vénire). We must substitute for them an enumera-
tion of the various ways in which the Tense-stems are formed,
especially the Present Tense-stem.

Of the Latin Present (i.e. Thematic Present) Tense-stems, the
usual type is that which was also the most prevalent in I.-Eur.,
that namely in which the Present-stem shows the ordinary un-
weakened root (E-grade) of the Verb, e.g. Gk. mev6-o-war from
the root bheudh-, wei@-w from the root bheidh-, wér-o-yar from
the root pet-, while the weak grade of the root is proper to the
Preterite (Aorist) tense, e.g. Gk. é-7t6-d-ynv, &-nl0-0-v, é-mT-6-
pyv. Latin examples (part of the ‘Third Conjugation’) are :
dco, O. Lat. douco, for *deuco (Goth. tiuha) from the root deuk- ;
dico, O. Lat. deico (Goth. ga-teiha, ‘I indicate’) from the root
deik- ; veho for *wegh-6d (O. Ind. véh-d-mi, Lith. vez-u, Goth.
ga-vig-a,‘I move’), Another type shows a Nasal in the Pre-
sent-stem (with weak grade of root) which is omitted in the
other tense-stems, this Nasal being either (1) a nasal infix, e. g.
Lat. ri-m-po (O. Ind. lti-m-pimi) from the root reup-, /%-2-do
§ 2.) THE VERB. CONJUGATIONS. 455

(O. Ind. bhi-na-dmi) from the root bheid-, or (2) a nasal affix,
e.g. Lat. /é-n-o (O. Ind. li-na-mi, O. Scand. li-na) from the root
lei-. Latin meio for *meigh-d, beside mtugo for *mi-n-&h-d, is
a good example of these two modes of forming the Pres.-stem ;
and similarly we seem to have O. Lat. xivo for *(s)neigh?-6 (or
(s)nigh*-6? M. U. iv. 8), beside ninguo for *(s)ni-n-gh*-6, in a line
of Pacuvius (Praet. 4 R.):

sagittis niuit, plimbo et saxis grandinat.

Another affixes -yd- (-yé-) or -iyd- (-tyé-), an affix which often
varied with 1 (ef. ch. iv. § 51); e.g. in Latin (part of the Third
and of the Fourth Conjugations) farcio (Gk. ppdoow for *ppax-ya),
with 2 Sg. /arci-s, from the root bhreq?- (cf. fregueus), Lat. morior
(O. Ind. mr-iy4-té 3 Sg.) from the root mer-. We have also a suffix
-skd- (-sk8-) used to form what are wrongly called ‘ Inceptives’
(Third Conj.), with weak grade of root, e.g. posco for *pore-sco,
from prk-, the weak grade of the root prek-, ‘to ask’; a suffix
-éyé- (with O-grade of root) used to form Causatives (Second
Conjugation), e.g. Lat. mdneo for *mon-éyd, ‘I remind,’ lit.
‘cause to remember,’ from the root men- (cf. Lat. me-min-i), and
soon. A very important class is the class of Verb-stems ending
in a vowel (Vowel-stems), which form their 1 Sg. Pres. Ind.
usually with the help of the suffix -yé-, e.g, in Latin (First and
Second Conjugations) ~o for *(s)na-y6 (O. Ind. sna-ya-té), neo
for *(s)né-yo (O. H. G. nau), but other Persons athematically,
e.g. 2 Sg. nd-s, né-s, many of these Vowel-stems being Deriva-
tives from Nouns and Adjectives (First, Second, and Fourth

Jonjugations), e. g. caro from cura (stem *eurd-), like Gk. tipd-w
from riya (-7), albeo from albus (stem *albo-, *albe-), fimio from
Jinis (stem fini-, finei-). These various modes of forming the
Present Tense-stem will be considered in later sections (§§ 6-33).

§ 2. Traces of the Athematic Conjugation in Latin. Of I.-Eur. verbs
of the Athematic Conjugation, which retain more or less of their athematic
character in Latin, the most important are the roots es-, ‘to be,’ ei-, ‘ to go,’
ed-, ‘to eat,’ wel-, ‘to wish.’ ES- has in 1 Sg. sum (Ose. sum) with the root
in weak form s-, instead of es- of I.-Eur. *és-mi (O, Ind. ds-mi, Arm. em, Gk.
ei-pi for *éo-pi, Lesh. gums, Alb. jam, Goth. im, Lith. es-mi, 0. SI. jes-mi),
though, if we may believe Varro (L. L. ix. 100), the older form was eswm (sum
quod nune dicitur olim dicebatur ‘esum’); 2 Sg. es [older es(s), scanned long
by ‘ position’ in Plautus], I.-Eur. *és-si (Arm. es, Hom. éo-oi) ; 3 Sg. es-t (Ose.
456 THE LATIN LANGUAGE. [Chap. VIII.
ist), L-Eur. *és-ti (0. Ind. ds-ti, Gk. 0-71, 0. Ir. is for *is-t, Goth. is-t, Lith.
és-ti and @s-t, O. Sl. (Russ.) jes-ti]; 1 Pl. simus and simus (ch. ii. § 16),
I -Eur. *s-mes, *s-mos (0. Ind. s-més) ; 2 Pl. es-t’s should be *s-tis, I.-Eur.
*s-té (O. Ind. s-thaé), but cf. Gk. éc-ré, Lith. 6s-te; 3 Pl. sunt (older sont)
from *sont(i) beside I.-Eur. *senti (Goth. s-ind), as Umbr. sent, Ose.
set; the I.-Eur. Optative *s-yé-m (*s-lyé-m) 1 Sg., *s-yé-s (*s-lyé-s) 2 Sg.,
*s-I-mos 1 PL, &¢., is reproduced with some fidelity in O. Lat. s-ié-s, s-i-mus,
though the vowel 7 extended itself in time over Sg. as well as Plur., class.
Lat. sim, sis, sit (§ 55), as in Umbr. sir ‘sis,’ si ‘sit,’ sins ‘sint’; in the
Imperative we should expect *és, the bare stem, in 2 Sg. (and és is probably
the only actual Latin form, on which see § 58), *s-tdd in 3 Sg., but we have
es-to(a@) (ef. Gk. €o-rw) ; the Inf. is es-se, the Loc. Sg. of an S-stem, as the
Umbro-Oscan *es-om (Ose. ezewm, Umbr. erom) is the Acc. Sg. of an O-stem
(ch. v. § 2); EI- is thematic in 1 Sg. in Lat. eo from *éy-6 instead of I.-Eur.
*ei-mi, but athematic in the other persons of the Pres. Ind., 2 Sg. is, older e-is,
3 Sg. it, older eit, r Pl, i-mus, older ei-mus, with strong stem ei- as in Lith.
ei-me, ei-te, instead of weak stem i- (possibly with a bygrade 1 like O. Ind,
j-mahé 1 Pl. Mid.; but Pel. ei-te 2 Pl. Imper. points to an original ei- for Latin
also), 2 Pl. 7-tis older ei-tis (with ei- again for i-), except the 3 Plural eunt from
*ey-o-nt(i), unless -onti was an J.-Eur. byform of -enti in the Athematic
Conjugation (ef. sunt from *sont(i) ) ; the form int of the Philoxenus Glossary
(p. 75. 23 G.: int, mopedovra) is too doubtful to quote as an athematic 3 Pl.,
for it may be a wrong reading for inunt (ef. prod-inunt, red-inunt, ob-inunt),
although indeed the common theory of the origin of these forms presupposes
an old 3 Pl. in-t (see § 73) ; athematic too are the Inf. ire older ei-re from
*ei-si, Imper. i older ei (Lith. ei-k), ito older ei-t0(d) (with ei-, as in Umbr.
ee-tu, for i- of Gk. i-re0(8), § 57); ED- is thematicin 1 Sg. ed-d (Gk. Fut. €5-o0-nar
is a Subjunctive form, and is quite regular), but athematic in the other
persons, which often show a byform éd- (Lith. éd-mi and éd-u z Sg., és-t 3 Sg.)
(thus Donatus ad Ter. Andr. i. 1. 54 distinguishes ut wna ésset, Subj. of edo, from
ut una esset, Subj. of sum); z Pl. edimus, 3 Pl. edunt are like swmus, sunt; the
Imper. es-t is athematic, and the Inf. es-se. (On -st- instead of -ss- from I.-Eur.
-dt- in est 3 Sg. &e., see ch. iv. § 155; on Imperat. es, § 58, below). The by-
forms edit, editis, &c. need not be new Latin types, for there are indications that
in the I.-Eur. period this root wavered between the Thematic and Athematic
Conjugations (cf. Lith. éd-u for *éd-d beside éd-mi ; Goth. ita for *8d-3) ; WEL-
was probably declined : 1 Sg. *wél-mi, 2 Sg. *wél-si, 3 Sg. *wél-ti,z Pl.*wl-més
or -més, &e. ; Lat. volo, vult (volt) are the normal equivalents of *wéld, *wélt(i)
(see ch. iv. § 10 on éliva from Gk. édaia, &e.), so that the 1 Sg. is thematic, as
are possibly also the 1 Pl.'and 3 PI. (ef. swmus, sunt), while the 3 Sg. and 2 Pl.
are athematic ; for the 2 Sg. *wels, which would become *vel(2) (ch. iv. § 146),
and would be probably scanned as a long syllable in Plautus [see ch. ii. § 133
on Plautine ter(r) for *ters from I.-Eur. *tri-s], the Romans substituted the

 

' The -u- in 1 Pl. of these Athematic
Verbs, sumus, volumus, is noticeable.
Volimus, the reading of the Ambrosian
Palimpsest in Plaut. Pseud. 233, Truc.
192 is a Late Lat. form (see Georges,
Lex. Wortf. s. v.), due either to the

Analogy of the Thematic Conj. (so
Late Lat. feris, aufere, on which see
Georges), or to the ordinary weaken-
ing of unaccented % (so possimus, § 97,
simus, ch. ii. § 16). On the spelling
quaesumus see below, § 33. 4.
§ 2.] THE VERB. CONJUGATIONS. 457

28g. of a different Verb-root wei-, ‘to wish’ (O. Ind. vi-, with 2 Sg. vé-xi,
Gk feva) [I.-Eur. wel-, wei-, wek- (Gk. éxov) all mean ‘to wish,’ and are
probably connected]; the athematic Imper. vi (vi in Plautus too) has
become a Conjunction (ch. x. § 4), while the Imper. noli of the Compound,
has been variously referred to a bystem of the Fourth Conjugation (I.-Eur.
*wel-ys-, Goth. vilja; O. Sl. velja), with Imper. *veli, like furci, § 57, and to
the Optative (cf. 2 Pl. Opt. nottis with 2 Pl. Imper. nolite ; also x Pl. nolimus,
used both in an optative and an imperative sense). The root BHER- belonged
to the Thematic Conjugation (Gk. ¢ép-w, O. Iv. -biur from *ber-d, Goth. bair-a ;
ef. 3 Sg. O. Ind. bhar-a-ti for *bher-e-ti, Arm. ber-é, 0. Sl. ber-e-tit; athematic
forms however appear, O. Ind. bhér-ti 3 Sg., bhr-tém 2 Du.; ef. Gk. pép-re
2 Pl., O. Ind. bhar-tam 2 Du.\, but in Latin to the Athematic (with the usual
exceptions of the 1 Sg. and probably x Pl. and 3 Pl. Pres. Ind.), e.g. fer-t 3 Sg.
Pres. Ind., fer-tis 2 Pl. (with strong stem /er- instead of weak stem jor-, I.-Eur.
bhr- ; cf. Gk. gép-re), fer Imper., fer-re (for *fer-se from *fer-si, ch. iv. § 146 ;
ch. iii. § 37) Inf., while 2 Sg. *fer(r) [this is what an original *fer-s() would
become in Latin ; ef. ter(r) for *ters, ch. ii. § 133] has been brought into line
with other 2 Sg. forms by the fresh addition of the 2 Sg. suffix, fer-s. An
Optative *ferim like velim and the other optatives of athematic Verbs is not
found and probably never existed. (On the Pres. Part.-stems sent- and sont-,
ient- and eunt-, volent- and volunt-, see § 90; and on the Optatives sim, O. Lat.
siem, edim, velim, § 56 ; the Compounds possum, malo, nélo, &e. are discussed in
§ 97, ambio in § 46).

Of I.-Eur. athematie Verbs of the type of O. Ind. da-ti 38g. (Gk. 8{8o-p, tory-pt,
&c.) with root dé- (varying with the weak grade), Latin examples are : DO-,
to give, of which 1 Pl., 2 Pl. Pres. Ind. dé-mus, dd-tis are the normal athematic
forms with the weak root dd- (cf. cditus and cds, ch. iv. § 54), while 2 Sg. das,
3 Sg. dat (O. Lat. dat, probably so scanned in Plautus, e. g. Most. 601, Men. 101),
show @ instead of 6, *dd-s, *dd-t; the old athem. 2 Sg. Imper. *do (Lith. di-k)
remains only in cé-dé (shortened under the influence of the preceding short
syllable, like havé, sibi, ch. iii. § 42), for the ordinary form da is coined on the
type of the rst Conjugation ; the 2, 3 Sg. dd-io is however the correct atliematic
form, also 2 PI. dé-te, and Inf. dé-re, older *dd-sit. 1 Sg. Pres. Ind. dé is pro-
bably *dé-yo (cf. O. Sl. da-ja), and the reduplicated Present-stem of Gk.
bi-5wy1, Pelignian dida ‘det’ &c. (§ 9) may appear in reddo, if this stands for
re-d(i)-d6. DHE-, to put, appears in the Latin compounds con-do, cré-do (O. Ind.
Srd-dha-, lit. ‘to put the heart to,’ O. Ir. cretim), which are usually reckoned
as ordinary thematic verbs of the 3rd Conj., like légo, though a great many of
their forms may be explained as athematic, with the weak root, Lat. -dd-
(which at the beginning of the word would be /é-, ch. iv. § 114; ef. fdcio,
I.-Eur. *dhdk-yo, from DHE-K-, Gk. é-6nx-a, an extension of the root by
addition of x); thus, though condis, condit should have *-dés, *-det (O. Lat.
*-dat, class. *-dét, ch. iii. § 49), con-dimus may represent *-ddmus, con-ditis, *-ditis,
and though 2 Sg. Imper. con-dé should be *-dé (Lith. dé-k), 3 Sg. con-dito, 2 Pl.
con-dite may represent *-dé-téd, *-dd-te, and Inf. con-dére, *-dé-si. (On Opt. duim
from the stem *duo, a bystem both of DO- and of DHE-, see § 56.) The roots
STA- and BHA- (Gk. tory-yu 1 Se., iord-pev 1 Pl. ; py-pit Sg., pa-pév rt Pl.)
are treated like the type tra- (a development of the root ter-; see below) and
retain @ throughout, std-s, sta-mus, sta-re, &e., fa-tur, fa-mur, fa-ri, &e. (cf. Gk.
é-ornpev, ¢-ornre), forming the 1 Sg. Pres. Ind. with the suffix -yo, sto from
458 THE LATIN LANGUAGE. (Chap. VIII.

*sta-yo (Umbr. stahu, Lith, sté-jis, ‘I station myself,’ O. SL sta-ja; of. Zend
a-staya, but see Buck, Osk. Spr. p. 24), fo(r) from *bha-yo (Lith. bé-ju, ‘I
enquire,’ O. Sl. ba-ja, ‘I converse’; ef. O. Ind. bha-ya-té Pass., if this be
a genuine form). The weak grade of the two roots appears in stdtus (O. Lat.
also status), fiiteor (cf. Osc. fatium Inf.) ; the reduplicated Present-stem of the
former (Gk. éorn-s for *ot-o7r7-s), in Lat. si-st-o, which usurped the transitive
sense that had in O. Lat. belonged to sto, e.g. med Mano statod, ‘set me as an
offering to Manus,’ on the Dvenos inscription, ‘astasent’ (leg. -int ?) statuerunt
(leg. -int ?) Paul. Fest. 19.32 Th. On these Reduplicated Presents of athematic
roots in Latin, si-sti-mus from the root STA-, se-ri-mus from the root SE- (Gk.
tinue for *o1-onpu\ and perhaps re-d(i)-di-mus, see §9. An athematic 2 Sg. Imper.
from the root BHEU- (of Lat. fui) occurs in the Carmen Arvale, fu (Lith. bu-k,
Umbr. fu-tu), if rightly understood in the sense of ‘be’: satur fu, fere Mars.

Of I.-Eur. athematic verbs of the type of 0. Ind. pra-mi (Gk. wi(w)Aq-pe) with
root plé-, a development of root pel- (the strong grade plé- never varying with
any weak grade), the Latin examples are: PLE- (cf. PEL-), im-plére, ex-plere ;
TRA- (cf. TER-), in-trare, &. These form the 1 Sg. Pres, Ind. thematically with
the thematic suffix -yo-, e.g. im-pleo for *-plé-yo, in-tro for *-tra-yo (thematic
byforms with this suffix perhaps existed in the I.-Eur. period, e. g. O. Ind.
tr4-ya-té beside tra-ti), but the other persons athematically, im-plé-s, in-tra-s,
im-ple-t (O. Lat. -é), in-tra-t (O. Lat. -at), im-plé-mus, in-tra-mus, &¢., as also the
other parts of the verb, Imper. im-piz, im-ple-to, in-tra, in-trd-to, Inf. im-plée-re.
in-tra-re.

Like them were declined other verbs whose stems ended in long vowels
or diphthongs, e.g. I.-Eur. *wid-é- (a stem perhaps originally confined to
Secondary Tenses, *wid-yé- being the stem used in the Present Tense; see
§ 15), Derivative Verbs from A-stems, e.g. cara- from the Noun cira (stem
ciird-, ch. v. § 2), from I-stems, e. g. finei- from the Noun finis (stem finet-, fini-,
ch. v. § 34), but not from U-stems (e. g. stdtuo from status, § 15), nor possibly
(1) those from O-stems (e.g. /ldveo from flavus), (2) Causatives and Intensives
with 18g. Pres. Ind. in -éyd (e. g. mineo, Causative of root men-; f. mémini;

e § 29), though these two last types have a declension which, by reason of the
phonetic changes of Latin, can hardly be discriminated from the athematic
declension [thus monés may represent *moné-s, as well as *mon-dyé-s (cf. trés for
*tréyés, ch. iv. § 66), moné Imper. may come from *moné, a stem in -é, or
*mon-éyé. In the P. P. P. however the two types are distinct, moni-tus, with
I.-Eur. weak stem moni-, im-plé-tus with stem plé-, as in the Perfect Ind. Act.
mon-ut, im-plé-vi; see § 39. 4]. Farcio (Gk. ppacow for *ppaxyw) and facio both
belong to the same I.-Eur. thematic type, a type in which the root has the
suffix -yé-, varying on the one hand with -yé-, on the other with -i- and -i-;
the.divergent roads which they have taken in Latin are perhaps due to the
fact that in the declension of facio the weak suffix -i- asserted itself (Imper.
QO. Lat. face for *fact, ciipe for *capi), in the other the weak suffix -i (Imper.
farci), and this assertion of the long vowel brought with it a transference to
the Athematic type, furci-re beside jface-re from *faci-se, though the original
difference between farcio and a Derivative like jinio, is still maintained in the
Pert. Ind. Act. far-si, for *farc-si; beside fini-vi, and P. P. P. far-tus, older
Jarc-tus, beside fini-tus. (On these stems with suffix -yo-, and on the Derivative
Verbs and stems ending in vowels which form their 1 Sing. Pres. Ind. with
the help of this suffix, see §§ 15, 21.)
§ 3.] THE VERB. AORIST AND S-STEMS. 459

The form cante for canite quoted from the Carmen Saliare by Varro (ZL. L.
vii. 27) can hardly be called an Athematie 2 PI. (Imper.). It is rather an
example of the Latin tendency to syncopate every short unaccented vowel
before a single consonant, which would have destroyed most traces of the
thematic vowel in the Latin Verb, if it had been allowed free play (see ch.
iii. § 23°).

§ 3. II. THE TENSE-STEMS (STRONG AORIST AND
S-FORMATIONS). The Tense-stems are formed by various modi-
fications of the root. From the root gen- for example is formed
by Reduplication (with { as Reduplication-vowel and with weak
grade of root) the Present-stem éi-@n- (Lat. gigno, Gk. yiyvo-yat),
expressive of continued action in Present time, ‘I am producing,’
and by another species of Reduplication [with & as Reduplication-
vowel and in the Singular (see § 39) with the O-grade of the
root] the Perfect-stem ge-gon- (Gk. yéyova) expressive of com-
pleted action, ‘Ihave produced.’ The Aorist-stem, expressive of
action merely, unlimited by the idea of continuance or the idea
of completeness, is in this Verb formed from the root itself gen-
(Gk. éyev-d-ynv), and so the O. Lat. form genunt (Varro, Sat.
Menipp. 35 B. sed quod haee loca aliquid genunt; cf. Luer. iii.
797 durare genique) might be called an Aorist (i.e. unlimited)
tense-form. But. the distinction between a Present-stem and an
Aorist-stem is by no means so clearly marked as between a Pre-
sent-stem and a Perfect-stem, and what is an Aorist-stem in one
language may be used as a Present-stem in another. The stem
&en- (with the thematic vowel Seno-, Sene- as in Gk. é-yevd-unp,
é-yéve-ro) is in O. Ind. used as a Present, jana-ti 3 Sg., and the
exact equivalent of Gk. é-yevo- (with the Augment prefixed) is
in O. Ind. not an Aorist, but the past tense of a Present-stem,
in other words an Imperfect, 4-jana-m, ‘I was producing,’ while
the aorist sense is assigned to the weak grade of root, gn- (8*n-),
a-jfia-ta 3 Pl. Thus in this verb it would be more correct to say
that the Present-stem was both &i-@n- and gen-, than to restrict
the first of these to the Present, the second to the Aorist
signification. 5

 

1 Or is cante, like the other strange (p. 175), merely a trace of the older
syncopated form quoted from the  syllabie writing (ch. ili. § 14)?
Carm. Sal. priuicloes * priviculis ’
460 THE LATIN LANGUAGE. (Chap, VIII.

Some find more exact counterparts of the Greek Strong Aorist
(2nd Aorist) in Latin Perfects like sctdit (O. Ind. 4-chidi-t),
scid- being the weak grade of the root sceid-, a root which forms
its Present-stem in Latin by Nasalization, sciz/-. But, as is
pointed out in § 41, seit is more likely to be a Reduplicated
Perfect (O. Lat. se?-cidi, O. Ind, ci-chidé), and to have lost its
Reduplication syllable in Compounds like disscidit (O. Lat. -7,
-eit) for *iis-sci-cidit, like re-p-pitlit for *re-pe-pulit), the Perfect
having in Latin come to assume Aorist functions, e. g. dédi (1) I
have given, (2) I gave. Where the Aorist-usage most shows traces
of itself in Latin is in phrases like xé attigas (Subjunctive of an
Aorist-stem tag- beside the Present-stem ting-), which suggest
comparison with the Greek use of the Aor. Subj. in pro-
hibitions, &c.

The Present-stem, as it is on the one hand occasionally indis-
tinguishable from the Aorist-stem, so it is on the other from the
Verb-stem. In a verb like Lat. sto the Nasal is clearly part of
the Present-stem, and connected with the idea of continued action
in present time, for it is not found in other parts of the verb
(e.g. st-vi, st-tum) (cf. tango Pres., détigi Perf., tac-tus Verbal
Noun). But in a verb like Lat. jwago the Nasal is extended
throughout the Verb (e. g. junwi, junctum, § 10). Similarly the
“Inceptive’ suffix of cre-sco is properly dropped in eré-vi, eré-tum,
but the same suffix is in posco (for *pore-sco, I.-Eur. *prk-sko-)
extended to Perf. piposci, &e. Some of the stems which are
included in this section in the list of Present-stems are probably
rather to be called Verb-stems, i.e. extensions of the root by
means of a suffix to denote action, whether continued action,
completed action, or momentary action ; for example, the stems
trém-, trés-, formed from the root ter- (O. Ind. tar-alé-, ‘tremb-
ling’) by means of an M-suffix (Gk. rpéu-w, Lat. trém-o) and an
S-suffix (Gk. tpé-(c)-w, Lat. terreo for *ters-eo), are rather Verb-
stems than Present-stems, although, for practical purposes, it is
best with a view to completeness to include them in the list of
Present-stem formations.

We have already spoken of the so-called Aorist (i.e. Strong
Aorist) forms of the Latin Verb. Another series of forms is
better considered here than assigned to any definite Tense, viz.
§ 3.] THE VERB. AORIST AND S-STEMS. 461

the S-formations, which receive further treatment in the sections
dealing with the Perfect, Future, Imperfect and Pluperfect Tenses.
A large number of Verb-forms, whose exact relation to each
other has not yet been clearly explained, show the sibilant s in
some shape or other (-ss-, -s-, -es-, &c.), We find an S-suffix in
the Verb-stem just mentioned, tres-, ‘ to be afraid, tremble’ (O.
Ind. trasa-ti and tar-dsa-ti, Gk. tpé(o)w ; cf. Lat. terreo for *ter-s-);
and in a stem like auk-s- of Gk. atéw (cf. atédvw) from the root
aug- (Lat. augeo), this S-suffix is clearly connected with the
S-suffix (-es-,-os-, -s-) of the Noun *augos, *auges-os Gen. (O. Ind.
djas, ‘strength’; cf. Lat. augus-tus like rdbus-tus, funes-tus, ch.
v. § 71), precisely as in Derivative Verbs like Gk. reAc(c)-w, Pft.
Pass. reréXeo-rau, from the Noun réAos, Gen. réAe(o)-os ; and
this Noun S-suffix we shall find to be the suffix used in Infinitives
like Lat. dgéré for *a&-es-i Loc. Sg., ferré for *bher-s-i Loe. Sg.,
ferri for *bher-s-ai Dat. Sg., Gk. defor, &e., which are nothing
but cases of Verbal Nouns. Side by side with Verb-stems with
an S-suffix stand Verb-stems with a suffix -syo-, eg. O.
Ind. tra-sya-ti, exactly as Present-stems in -yo- like Lat.
fig-io (stem bhtig-yo-) stand side by side with Present-stems
like Gk. gevyo (stem bheug-o-). This suffix -syo- is how-
ever usually the suffix of the Future-stem (e.g. O. Ind.
dék-Sya-mi from I.-Eur. deik-, Lith. bi-siu from I.-Eur, bheu-),
but not of the ordinary Greek Futures, e. g. de(Ew, piow, Tysjoo,
which are now generally regarded as Subjunctives (the L-Eur.
Subjunctive had Future, as well as Subjunctive, force, § 55) of
the S-Aorist, the difference between tijoouer Fut., ryujowper
Aor. Subj., de(fouev Fut., de6£wperv Aor. Subj. being explained
by the fact that originally the Subjunctive of the S-Aorist was
Athematic (§ 1), ryyrjo-o-wev, delé-o-pev (cf. reloouer Hom.), but
afterwards took by analogy of Thematic Subjunctives the long
vowels w, 7, and retained its proper athematic forms only in their
Future Indicative usage. Gk. ryyjow, ériunoa, &e, must have
had at the first double s, *riyzjoow (-dcow), ériunooa (-aroa), for s
between vowels in Greek disappeared when single (e. g. yéve(o)-os),
and was reduced when double (e.g. e#ca from the root eus-, for
*eoca, cf. Lat. ds-si, ch. ii. § 129), as in Latin it became 7 when
single (e.g. génér-is, quaero), and was reduced in the classical
462 THE LATIN LANGUAGE. (Chap. VIII.

period after a long vowel or diphthong (é/.) when double (e.g.
quaeso, older quaesso, haesi, older haessi). Gk. Tyusjnow, older
-dcow, will then correspond to O. Lat. amasso, Gk. delEw to O.
Lat. dizo and similar forms. We have also Latin forms in ss
used as Presents, but always with a peculiar sense (§ 33. 5), eg.
lticesso (cf. lacio, lacto), fuicesso (cf. facio), capesso (cf. capio), incipisso
Plaut. (cf. incipio), pétesso (cf. peto; in O. Lat. petissere, * saepius
petere’ Fest. 250. 19 Th., adpetissis Accius, Trag. 160 R.; cf.
petivi), quaeso (cf. quaero), viso (cf. video); they have sometimes
been called Latin Intensives, and compared with another 8-
formation in which the root is reduplicated, namely, the De-
sideratives (sometimes with Intensive force) of Sanscrit, e.g.
pi-pa-s-ami, ‘I wish to drink, ji-jiva-s-ami, ‘I wish to live,
ip-sami, ‘I wish to acquire, and the Reduplicated Futures of
Celtic, such as O. Iv. gigius ‘ rogabo, gigeste ‘ orabitis’; their
Perfects in -ivi, lacessivi, quaesivi (used as Pft. of guaero), arcessivi,
perhaps point to parallel stems in -ss-yo-, */acessio, *quaessio,
*arcessio (cf. the I.-Eur. Fut. in -syo-, Lith, ba-siu beside Gk.
gv-ow) (on visi, see § 41). Other O, Lat. s-forms (Subjunctive
or rather Optative, § 55), like averruncassis, servassis, famis,
used in prayers, wishes, deprecations, &c. (e.g. deos ut fortunas-
sint precor; Juppiter, prohibessis scelus ; di mactassint), have
been called Latin Precatives, and compared with Sanscrit
Precatives (more usual in the older literature than in classical
‘ Sanscrit ’) such as bhii-ya-s-am from bhi-, ‘ to be.’ These O. Lat.
Fut. and Opt. forms with ss (corresponding to s after a consonant,
faxo, favim, diwo, diwim) are mostly found in Verbs of the first
Conjugation, -asso, -assim, Inf. -assere, but sometimes in Verbs
of the second, e.g. prohibessis, prohibessint. They do not occur
in the Aorist (Preterite) Indicative usage of Gk. ériunoa (-dooa),
epidnoa (-nooa) in Latin; but, if Umbro-Oscan tt is rightly
interpreted as the equivalent of I.-Eur. ss (cf. Att. apdrrw for
mpacow, this oo being a Greek development of xy), they do
occur in this usage in the other Italic languages, Osc. terem-
nattens ‘ terminaverunt,’ quasi *terminassunt, préfatted ‘ pro-
bavit, quasi *probassit, duunated ‘donavit’ quasi *donassit,
Pel. coisatens ‘ curaverunt’ and so on (all the examples preserved
belong to the first Conjugation).
§ 3.] THE VERB, AORIST AND S-STEMS. 463

So far we have found evidence of Verb-forms with ss after
a vowel,s after a consonant, used as Aorists, Futures [in the
Future use often with -(s)syo- for -(s)so-], Precatives, &c., as well
as of Verb-stems with a suffix consisting of a single s (e.g.
*tres- the stem of Gk. rpéw), a suffix perhaps identical with the
S-suffix of Verbal Nouns. It is therefore a natural inference to
explain the formations with double s as due to the addition of
some S-suffix to a Verb-stem already composed by means of an
S-suffix, so that Latin amasso would be resolved into ama-s-so,
and possibly dizo into deic-s-so. (On diwo however, see § 55;
Greek Locatives Plural also like Ovpdou offer an original -ss-,
which after a consonant appears as -s-, e. g. pvAaét).

This additional S-suffix may be the same as that which seems
to show the form -és- in Latin Future Perfects like véd-ero,
amav-ero, dix-ero, and Pluperfects like vid-eram, amav-eram,
dixz-eram, and which has been identified, plausibly enough, with
the Substantive verb, es-,‘ to be’ (it appears in the form s in
ama-rem, fer-rem, age-rem, &c.), though some regard it as the
suffix -es- of Noun-stems, comparing vid-ero to Gk. cidé(r)u,
viderimus Subj. to Gk. eide(o)iwev Opt., and these to Gk. cidos,
-e(o)-os (similarly ferrem, agerem, &c. to ferre, agere). Others
make it not only -es- but -is- or -as- (ch. iv. § 3), comparing
vid-eram to O. Ind. 4-véd-igam, dieram to O. Ind. Aorists with
-sig- such as ak&igur 3 Pl., and identify this -is- or -as-, as well as
-es-, with the suffix of Noun S-stems (e.g. O. Ind. roci8- N.,
Gk. yiipas, O€us-, Lat. cinis), and further with Latin -is- of
amav-is-tis, amav-is-sem, amav-is-se, although the 7 in these Latin
Perfect-forms may be merely an example of the continuation of
the vowel of the 1 Sg. Ind., &. of the Perfect throughout the
declension of this Tense, as Gk. deffayu, der€drw, &c. continue
the a of géeua (see §§ 67, 52, 39). The Umbro-Oscan Future
forms, e.g. Osc. didest ‘dabit,’ Umbr. ferest ‘feret’ point to
the vowel having been originally e, for the Latin weakening of
unaccented vowels is almost unknown in Umbro-Oscan.

Another moot point in the analysis of the Latin Verb is the
explanation of what the Roman grammarians regarded as con-
tracted forms, such as amassem beside amavissem, amastis beside
amavistis, amarunt, amaro, and amarim beside amavérunt, amavero,
464 THE LATIN LANGUAGE.

and amaverim, and also such as invassem beside invasissem, dixem
beside dixissem, dzati and dixtis beside diwisti and dixistis. There
is nothing in the laws of Latin Phonetics to prevent the Con-
traction Theory from being right; aaudivissem would become
audissem as naturally as st vis became sis, oblivisci became oblisci
(Plaut.), dévinus, dinus (Plaut.) (see ch. iv. § 70) ; euvasissem would
become invassem by that Roman practice of discarding one of
two similar neighbouring syllables which reduced *areieubii to
arcibii, Restitutus to Restitus (see ch. ui. § 13. p. 176). Thus
although the comparison of amassem (E-Subj. like amem) with
amassim (Opt. like sim, O. Lat. stem), diwem with dixo and dixim,
amarim with amarem is a very natural one, it cannot be said that
the evidence is at present strong enough to warrant us in
relinquishing the old explanation.

Indeed the evidence to be derived from the usage of Plautus
and the other Dramatists is all the other way, for we find that
Plautus and Terence treat these shorter forms exactly as they
treat forms that are indubitably contracted, like jurgo from jirigo
(ch. ui. § 13); in the older poet the two are used side by side, in
the later the contracted have ousted the uncontracted, the latter
being used only at the end of a line, i.e. only through metrical
necessity. (For statistics, see §§ 48, 49.)

§ 4. ‘Strong Aorist’ forms in Latin. Beside the Nasalized Present-stem
tting- of tango, attingo we find a stem tdg- in O. Lat. In the Dramatists attigas,
attigatis (only in prohibitions) are not uncommon, e.g. Plaut. Bacch. 445 ne
attigas puerum istac causa (cf. Non. 75. 26 M.). A lamp discovered in the
very ancient Esquiline burying-ground bears the inscription : ne atigas. non
sum tua. M. sum (Ann. Inst. 1880, p. 260). In the simple verb we have, e. g.

si tagit, nisi tagam in Pacuvius (Lrag. 344 and 165 K.) (forms compared to
con-tigit, at-tigit by Festus 540. 27 M.), and probably tdgo in Plaut. Mil. 192:

remordre: abeo. Neque té remoror neque té tago neque te—tdceo,

Similarly beside the Nasalized Present-stem toln- of tollo, attollo (ef. affero),
abstollo (ef. aufero) we have attulas, gbstulas, &c. (or in the older spelling attolas,
abstolas) in prohibitions in the Dramatists, e.g. Novius, Com. 87 R. dotem ad
nos nullam attulas; Pacuv. Trag. 228 R. :

custodite istine uos: ne uim qui dttolat, neu qt Attigat,

and in the simple verb nisi tulat (Accius, Trag. 102 R.).

Beside the Present-stem formed with the suffix -yo- (§ 15), vinto-, we find
a stem vén- without this suffix in the compounds evenat, advenat, pervenal, &c.
(e. g. Plaut. metuo ne aduenat, priusquam peruenat, utinam euenat, quomodo
§§ 4, 5.] THE VERB. AORIST AND S-STEMS. 465

euenat, &c.) These Aor. forms seem to occur only at the end of a line (the
same indeed is generally, but not always, true of attigas, attulus, &c.), and so
are licences of which the Dramatic Poets availed themselves under metrical
necessity. They have been also referred to a suppression of the 7 (y) of
eveniat, &c., like that of i (y) in dbicio (the scansion of the Dramatists) and
possibly augiir(i)a of Accius (see ch. ii. § 50), on the plea that the Aor.-stem
would be vem-, for the n is produced from m under the influence of the
following consonantal ¢ (y), *vem-yo (I.-Eur. *gum-yo-, Gk. Baivw for *Bapyw)
becoming tren-yo, vénio, as quom-jam became quéniam (ch. iv. § 73). The root is

em- (cf. Goth. qiman, ‘to come’). Similarly beside pério, to give birth to,
we have parentes. Beside the Reduplicated Present gi-gn-o from the root gen-,
we have an O. Lat. Present geno (cf. Priscian, i. 528. 25 H.) of the common type
of fero (root bher-), veho (root wegh-), sequor (root seq’-) (see § 6). This form of
the Present is frequent in Varro (e.g. R. R. ii. 2. 19 nam et pingues facit
facillime et genit lacte; ib. i. 31. 4 antequam genat; Men. 35 B. quod
genunt), and occurs in the testamentary formula ‘si mihi filius genitur ’ (Cic.
De Orat. ii. 42. 141) (see Georges, Lex. Wortf. s. v. for examples of the word).
It was a moot point among Latin grammarians whether in the phrase of the
XII Tables nr rra pacvnt, the last word was 3 Plur. of a verb *paco (whence
paciscor ; but ef. § 28), or with the old usage of writing c both for c and for g,
of a verb pago (an unnasalized form of pango ; ef. pépigt) (see ch. i. § 6). (On
rudentes, the ropes of a ship, see below, § 6.)

§ 5. O. Lat. forms with -ss- (-s-). The forms in -(s)so 1 Sg. may be called
Future-Perfects ; thus faxo corresponds to féecéro in Plaut. fr. 62 W. :

peribo si non fécero, si faxo uapulabo,

Capt. 695 pol si istuc faxis, haud sine poena feceris ; but they are often used
in the Dramatists, as the ordinary Fut. Perf. is also used, in the sense of
a Future, e.g. Plaut. Poen. 888 nisi ero uni meo indicasso, ‘I will tell my
master only’ (see Neue, ii. p. 548). They have an Inf. in -ssere, e.g. hoc
credo me impetrassere, illum confido me reconciliassere, and occur sometimes
in the Passive Voice in laws, e.g. mercassitur (C. I. L. i. 200. 71); faxitur in an
old Rogatio mentioned by Livy, xxii. ro. 6: si antidea senatus populusque
iusserit fieri, ac faxitur ; turbassitur in a law in Cicero, De Legg. iii. 4. 11; cf.
uti iussitur, ‘as shall be ordered,’ Cato R. R. xiv. 1.

The forms in -(s)sim 1 Sg. have never a reference to past time like the
ordinary Perfect Subjunctive ; contrast, for example, Plaut. Capt. 127 uisam
ne nocte hac (‘last night’) quippiam turbauerint, with Pacuvius, Trag. 297 R.
precor uenidm petens Ut quae égi, ago, vel Axim uerruncént bene (see A. L. L.
ii. 223); they are frequent in Plautus, who normally uses cave dixis, cave
faxis, &e., and not (except at the end of a line, i.e. for metrical convenience)
cave dixeris, cave feceris, &e. (after ne only dixeris, &c.), but are much less fre-
quent in Terence. They are generally used by Plautus in the protasis of
a conditional sentence (except ausim, faxim, which are found in the main
clause), v.g. Awl. 228 si locassim, and similarly in old laws, e.g. the Lex
Numae (ap. Fest. 194. 21 Th.) : si hominem fulminibus occisit ; we find them
also in wishes, e.g. Plaut. Aul. 50 utindm me diui adaxint ad suspéndium, in
prayers, asin the Augural Prayer (quoted by Festus, 526. 11 Th.): bene sponsis
beneque uolueris, and with ne in deprecations, e.g. Plaut. Most. 1097 ne
occupassis opsecro aram, and expressions of anxiety, e.g. Plaut. Bacch. 598 :

Hh
466 THE LATIN LANGUAGE. (Chap, VIII.

mihi catitiost
ne nucifrangibula exctissit ex malis meis.

These usages mark the forms in -(s)sim as the Optative Mood of the forms in
-(s)so. This -(s)s- formation appears in Vowel Verbs almost only in the first
Conjugation, though we find in the second prohibessit, prohibessint and prohibessis
occasionally, cohibessit (Lucy. iii, 444), licessit (Plaut. Asin. 603), &e., while for
the fourth ambissit (MSS. ambisset) in the (un-Plautine?) prologue of the Am-
phitruo of Plautus, 1. 71, is quoted. In Terence these forms of Vowel Verbs
ave very rare, v. g. appellassis, Phorm. 742. In Consonant-stems of the third
Conjugation they are found in poetry of all periods (fuxim and ausim even in
prose); examples are (Labials) capso, accepso, incepsit, (Gutturals) axim, taxim,
insexit, novit, (Dentals) baesis (C.G.L. ii. 27. 55), incensit, (Nasals) empsim,
surempsit (Fest.). [For other instances, see Neue, ii’, 539 saqq.; essis of
the MSS. of Nonius (200. 30 M.) in a line of Accius [Trag. (Pract.) 16 R.],
quoted as an example of castra F.: castra haec uestra est : éptime essis méri-
tus a nobis, seems a miswriting of escis (cf. below, § 33. 5); adessint of the Lex
Repetundarum (C. I. L. i. 198. 1. 63) may stand for ad-essent (ef. ch. vi. § 33);
uiolasit (0. I. L. xi. 4766, with anua) and similar forms with s for ss belong to
the period when double consonants were written single (ch. i. § 8)].

§6. A. Present!. (1) With E-grade of root and Thematic
Vowel. Examples are of E-roots: I.-Eur. *séq*5-, *stq*é-,
Mid., ‘to be following’ (O. Ind. séca-té 3 Sg., Gk. éro-ua,
O. Ir. sechur, Lith. sekt: Act.), Lat. seguor ; L-Eur. *weghi-,
*wéshé-, ‘ to be carrying ’ (O. Ind. véha-mi, Pamphyl. Gk. Fexa(?),
Lith. vezt, O. Sl. veza, Goth. ga-viga, ‘I move’), Lat. eho.
Of EI-roots: I.-Eur. *bheidhi-, *bheidhé-, ‘to be trusting’
(Gk. ei@o-yar, Goth. beida, ‘I abide’), Lat. fado from O. Lat.
feido; T.-KHur. *deiki-, *deiké-, ‘to be showing, indicating’
(Goth, ga-teiha, Germ. zeige), Lat. dico from O. Lat. deico ;
I.-Eur. *meighé-, *meighé- (O. Ind. méha-mi, O. Engl. mige),
Lat. mezo. Of. EU-roots: I.-Eur. *deusi-, *deuté-, ‘ to be lead-
ing, drawing’ (Goth. tiuha, Germ. ziehe), Lat. dico, O. Lat.
douco for *deukd (ch. iv. § 37); I.-Eur. *eusid-, *eusé-, ‘ to be
burning, singeing’ (O. Ind. 68a-mi, Gk. cw for *ethw), Lat. iro,
O. Lat. *ouro from *eusd ; I.-Eur. *plews-, *plewé-, ‘ to be flow-
ing, sailing, swimming’ (O. Ind. pléva-té 3 Sg. Mid., Gk. mA€(F)w,
O. SI. plova for *plew-), Lat. plwo from O. Lat. plovo (cf. Fest.
330. 29 Th. ‘ pateram perplovere ’ in sacris cum dicitur, significat
pertusam esse; so plovebat, Petron. 44. p. 30.1 B.). (The form
pluo is proper to Compounds, e. g. perpluo, ch. ili. § 24.)

 

1 For a fuller list of examples of see Joh, le Présent et ses dériv’s dans la
the various Present-stem formations, — conjugaison latine, Paris, 1893.
§§ 6-8.] THE VERB. PRESENT-STEMS. 467

The weak grade of the root, which in Greek appears with the
accent on the suffix, and with the Aorist sense (e.g. rpameiy Aor.
but rpémew Pres., miOécOa. Aor. but welOecOa. Pres., mvdécOar
Aor. but red#er8a: Pres.), has sometimes encroached on the Pres.-
stem, e.g. Dor. Gk. rpdiw beside Att. rpémw, Gk. yAtiw beside
L-Eur. *gleubhd- (Lat. glabo, O. H. G. chliubu, Engl. cleave).
Similarly Lat. vido (O. Ind. riida-mi) beside r#do (O. H. G. riuzu)
seems to show *rtidé- beside *retidd-, so that r%dentes, the ropes
of a ship, lit. ‘the rattlers,’ ‘ roarers, might be called an‘ Aorist’
participle (§ 4). [In Plautus we have ridentes, Rud. 1015 mitte
ridentém, sceleste, as r#do in Persius, ii. g Arcadiae pecuaria
riidere credas. The Pft. rudivi (Apuleius) and Verbal Noun rv-
ditus point to a Pres.-stem *rtid-yd-, like fligio, § 15.]

Examples of Verbs which have not an E-root are: rado, I am
going (the weak grade vdd-is seen in védum, a ford); caedo, from
older caido, I.-Eur, (s)kaidh- (Goth, skaida, cf. Engl. water-
shed); dgo, I.-Eur. *4g6 (O. Ind. aja-mi, Gk. dyo, M. It.
agaim, &c.; the strong stem 4&- is seen in Lat. amb-dges, &c.) ;
scibo, to scrape, I.-Eur. skabh- (Goth. skaba, ‘I shave’; the
strong stem skabh- perhaps appears in Perf. scdbi ; see § 39).

§ 7. Other examples. Lat. teyo (Gk. oréyw); rego (Gk. 6-péyw); lego (Gk.
Aéyw). Like O. Lat. nivit, beside ninguit, for *(s)neigh®é- (Zend snaézaiti, Gk.
veipet, Lith. dial. snéga) is O. Lat. fivo for *dheigh%-(?) (Lith. dégia Intr.),
for which sigo was afterwards substituted by Analogy of fix’, &. The form
with v which, we are told, was used by Cato [Paul. Fest. 65. 19 Th. ‘fivere’
(apud Catonem) pro figere], reappears in the derivative fibula for *fivi-bula (cf.
fixul-ae, Paul, Fest. 64. 7 Th. ‘fixulas,’ fibulas). (Lith. dygiis however points
to *dhights-, and we have figier, not fei- on the S. C. Bacch.).

O. Lat. amploctor (veteres immutaverunt ‘amploctor’ crebro dictitantes,
Diom. 384. 8 K.), e.g. Liv. Andr. Odyss.:

utrum génua amploctens uirginem ordret,

may take its o from a Derivative Noun (cf. toga from tego, ch. iv. § 52, and see
below, § 33.3). The o of cdgué for *qué-quo (O. Ind. pac-, Gk. wéoow, O. SI.
pekg) has been similarly referred to the influence of ciquus.

§ 8. Weak grade of root. Verto is in O. Lat. vorto (Plaut., &c.), so that
instead of showing the vowel é of I.-Eur. *wért-6 (0. Ind. vartaté 3 Sg. Mid.,
Goth. vairpa, ‘I become’), it seems to show the o of the Perfect, O. Lat.
vorti (with weak-grade of stem, wrt-, as in O. Ind. va-vrt-é Perf. Mid.) and the
P.P.P., 0. Lat. vorsus for *wrt-to (O. Ind. vrt-té-). But in Umbr. we find
vert- in the Present, vort- in the Perfect and Verbal Noun, ku-vertu, co-vertu
‘gonvertito,’ ku-vurtus, ‘converteris,’ co-vortus, vorsum Acc. tAé@pov. The O.
Lat. spelling vorto is in all probability a mere matter of spelling; the Present

Hh2
468 THE LATIN LANGUAGE. [Chap. VIII.

was always pronounced with e¢, verto, but at the time when vo- had come to be
pronounced ve- (see ch. iv. § 10), it was occasionally spelt (not pronounced)
vorto. O. Lat. vorro for *verso (O. H. G. wirru, ‘verwirre’) may be similarly
explained. In Gk. yi-yv-o-var Pres., é-yev-d-unv Aor. we find the E-root
relegated to the Preterite, while a modification of the root by Reduplication
is assigned to the Present (see next paragraph). In O. Lat. and O. Ind.,
however, the E-form, gen-, appears also as a Present-stem (O. Lat. genunt,
O. Ind. jan-a-ti 3 Sg.) (see §§ 3-4). Similarly, Lat. peto is regarded by some as
originally an Aorist-stem, the Present-stem being formed with the -YO-suffix
(§ 15), *petio 1 Sg., whence the Perfect petivi.

Other examples of the unaccented verb-form of a Latin Compound asserting
itself in the uncompounded verb are : cliido for claudo (see ch. ii. § 36) ; luo for
lavo (Sil. Ital. xi. 22); spicio, sico for spécio, séco (see ch. ii. § 12). (For additional
instances, e. g. plico, see Solmsen, Stud. Lauty. p. 130.)

§ 9. (2) With reduplicated root. Latin examples are:
gigno [1.-Eur. *&i-8nd-,. *Si-6n-é from root Sen-, which also
occurs with Them. Vow. as a Pres.-stem in O. Ind. and O. Lat.
($$ 3-4), Gk. yéyvopat], bib0 for *pibo (ch. iv. § 163) (I.-Eur. *pi-
bo-, *pi-bé-, O. Ind. piba-ti 3 Sg.,O. Ir. ibi-d; cf. Faliscan pipa-
fo Fut.), sisto (I.-Hur. *si-std-, *si-sté- from root sta-, O. Ind.
tistha-ti 3 Sg.), sido for *st-sdo (ch. iv. § 151) (I.-Eur. *si-zdi-,
*s{-zdé- from root sed-, O. Ind. sida-ti 3 Sg. for *sisd-, Umbr.
ander-sistu ‘intersidito’ Imper., for *sisd(e)-tdd), Often these
Reduplicated Present-stems belong to the Athematic Conjugation,
e.g. Gk, torn-yu for *si-sta-mi (O. H. G. sestd-m), beside Lat. sisto,
O. Ind. tisthi-mi; and Latin sistimus, sistitis correspond as well
with the Greek ford-yev, tord-re, as with the thematic forms.
Similarly Gk. inws (I.-Eur. *si-sé-, athematic) is in Latin thematic
in séro for *si-so, though seri-mus, seri-tis may be equally referred
to athematic *sisa- as to thematic *sisd-, *sist- ; I.-Eur. *di-dé-,
athematic (Gk. 6/3w-u1, O. Ind. déda-mi) has in Latin lost its
reduplication, except in reddo, if this stands for re-d(i)do as rep-
pult, repperi for re-p(e)puli, re-p(e)peri, but not in Umbro-Oscan
(Pelign. dida ‘det,’ Umbr. dirsa pronounced *dida?). All these
Latin examples reduplicate with the vowel i, and most belong to
roots ending in a long vowel. Some Greek Aorist-stems show
this reduplication with the vowel é, e. g. é-edvo-v from root dev-,
é-xéxhe-To from root KeA-, wentO-dv from root Tei6-, wemvO-o1T0
from root 7ev6-, by analogy of which the spurious Presents répve,
xéxAowat have been formed. Short ¢ is also the Reduplication-
vowel of the Perfect-stem (see § 39).

q
§§ 9, 10.] THE VERB. PRESENT-STEMS. 469

§ 10. (3) With root nasalized. Of the ten conjugations under
which the Hindu grammarians have classified the Sanscrit verb,
three are assigned to these nasalized Present-stems, one (the
seventh conjugation) showing a nasal infix, I.-Eur, -né-, varying
with -n-, e.g. yit-ndé-j-mi 1 Sg., yii-fi-j-més 1 Pl, yunk-té
3 Sg. Mid. (Lat. jz#-n-go), from the root yuj- (I.-Eur. yeug-),
the other two showing a nasal affix, viz. the ninth conjugation
with -na- varying with a weak grade (O. Ind. -ni-), e.g.
str-na-mi 1 Sg., str-ni-m4s 1 Pl., str-ni-té 3 Sg. Mid. (Lat.
ster-no), and the fifth conjugation with I.-Eur. -neu- (O. Ind.
-no-) varying with -nt-, e.g. r-nd-mi 1 Sg., r-nti-mds 1 PL,
r-nii-té 3 Sg Mid. (Gk. ép-vi-y). In Greek the type of Present
corresponding to the Sanscrit seventh conjugation has only -n-,
never -né-, and has been usually modified by the addition either
of a nasal affix, e.g. not *Ai-p-me (Lat. /i-n-guo) but Atyu-m-dvw (so
Tu-y-x-dva, da-v-0-dvw, &e.), or of the -YO-, -YE- suffix, e. g.
xAd@w for *xdayy-yo (cf. é-kAay&a) (Lat. clango); the Sanscrit
ninth conjugation is represented by oxid-vy-ms 1 Sg., cxid-vd-pev
I PL., oxid-vé-yar Mid., air-vy-ye 1 Sg, wér-vd-yev 1 Pl., air-vd-
pat Mid., &c.; the Sanscrit fifth conjugation by dy-vi-w. 1 Sg.,
dy-vd pev 1 PL, dy-vi-yar Mid., cxeddvvipn, merdvvdps, &e. In
Sanscrit all these nasalized stems belong to the Athematic Con-
jugation, though we have thematic byforms like 3 Sg. yuijati,
rnv4ti; but in Greek the first type mentioned is always thematic,
e. g. Aysmdve, the others occasionally, e.g. murvdw, dayydw (byforms
of airvnpt, ddprvnpe), ioxavde, &e., ctpwvvtw, tpwvvio, dpive for
épwwFw, &ec., while we have another type with -vew, e.g. ixvéouar,
bmucxvéopar, kiygo. The discrepancy between Greek and Sanscrit,
the two languages in which these nasalized stems have been
most fully preserved, makes it difficult to determine the original
I.-Eur. types of nasalization (see J. /. tl. pp. 285 sqq.). In
default of a better classification, we may arrange the Latin
nasalized Presents in two classes, according as the nasal presents
the appearance of a nasal infix or a nasal affix.

i. With nasal infix, e.g. i-n-quo (O. Ind. ri-né-c-mi 1 S¢g.,
ri-fi-c-més 1 Pl., Pruss. po-linka, ‘he remains’; ef. Gk. Aiurdve),
from root leig?- (Gk. Aeémw, Lith. léki, Goth. leihva, ‘I lend,
Germ. leihe) ; /%-a-do (O. Ind. bhi-né-d-mi) from root bheid-
470 THE LATIN LANGUAGE. [Chap. VIII.

(Goth. beita, ‘I bite’). The variation of -né- and -n- seen in
O. Ind. ri-né-c-mi 1 Sg., ri-fi-c-més 1 Pl., is not seen in other
languages, where the weak grade -n- is used throughout. The
Latin Presents conguiniscor (Perf. conguexus), to stoop, and
possibly frziniscor (cf. fructus), to enjoy, do however perhaps show
the fuller suffix-né- combined with the Inceptive suffix -sko- (§ 22),
if -niscor stands for -vée-scor, with loss of ¢ (g) in the group
-esc- asin disco for *dic-sco (ch. iv. § 157), and with 7 by analogy
of other Inceptives (see § 28); and Gk. xv-vé-(c)-w, Aor. éxvo-a,
may do the same. These forms with nasal infix are often
augmented by the YO-suffix (§ 15), e.g. Gk. rrloow for *rrivo-ya,
mAaCw, to beat, for *rrayy-yo (ef. &-rAayéa), rAdCw for *krayy-yo
(cf. é-«Aayéa), Lith. jimg-iu, beside Lat. pinso, plango, clango,
and jango; and so in Lat. pinsio, sancio (cf. siic-er), vincio (from
the root vyek-). English examples of nasal infix are: ‘I spring,’
from *spr-n-&ho, from the root spergh- (Gk. omépyopar); ‘I
wring,’ from the root wergh- (Lith. verz-it, ‘I squeeze’).
(‘I stand’ belongs to a rare type of JI.-Eur. Present-stem in
-NT, on which see Osthoff in Versamml. Philolog. xxi. p. 300.)
ii. With nasal affix. I.-Hur. li-na- (O. Ind. li-na-mi, Gk.
Ai-vd-pau' tpéroua, Hesych,, O. Ir. énim, ‘I cling to, follow, O.
Scand. lina, ‘I grow weak’) is Lat. /¢-no, from a root lei-, so that
the J.-Eur. affix -na- has been lost in Latin (unless Zino represents
*li-nd-6), and only its weak grade (O. Ind. -ni-, Gk. -vd-) remains,
e.g. li-ni-mus (Gk. *di-vd-pev). The -nd- of aspernari (beside
sperno), consternare, to terrify (beside consterno, to strew) (but cf.
ch. iil. § 19), declindre and inclindre (beside Gk. xdive), destindre
(beside Gk. ordvo and cravdw) cannot quite be identified with
I.-Eur. -na- (O. H. G. spor-né-n, O. Ind. str-na-ti 3 Sg.,O. Sax.
hli-né-n; cf. O. H. G. stornén, ‘to be astonished,’ hlinén), for
the same d@ appears in compounds of other than nasal-stems, e. g.
occtipdre (beside ciipere), profligare (beside flere) (see § 32), and
the -ind- of la-n-c-ind-re (cf. liie-er), cogu-ind-re (cf. cdquo), which
indeed suggests comparison rather with Gk. -dvo- of du-u-1-dve,
éyapt-dve, &e., than with Gk. -vd- of o«(d-vn-, alr-vn-u1, seems
to show the -d- (I.-Eur. -ayd-) of Derivative Verbs (§ 32), like
sarcinatus from sarcina (a Derivative with nasal suffix from sarcio,
as facinus from facio),runcinare from runcina, paginare from pagina,
§§ 11, 12.] THE VERB. PRESENT-STEMS. 471

or nominare from nomen (cf. destina, a prop). The I.-Eur. affix
-neu-, -nti- has left very few traces in Latin: ster-nu-o (Gk.
mTdp-vi-yar), mi-nu-o (O. Ind. mi-né-mi; ef. Gk. pi-vi-0o),
probably for *mi-new-6, &c. (cf. dénwo for de *newéd, ch. iii.
§ 24). English examples of nasal affix are ‘I spurn, ‘I shine’
(Goth. skei-na), ‘TI fill’ (with 11 for In).

By rule the Nasalization should be confined to the Present
Tense, and not extended to other than Present forms: e.g.
hi-n-quo, re-liqui, re-lic-tus ; st-n-o, si-vi, si-tus. But it pervades
the whole verb in some cases, e. g. jungo, junxi, junctus.

The weak grade of the root is proper to all these Nasalized
Present-stems, e.g. jungo from root yeug-, /ixo from root lei-,

mi-nu-o from root mei-.

§ 11. Other examples of nasalinfix. Lat. rii-m-po (O. Ind. li-m-pami), from
root reup- (O. Engl. beréofe, Engl. I bereave) ; pi-n-so (O. Ind. pi-na-8-mi; ef.
Gk. wricow for *rrwayw, Lat. pinsio) ; sct-n-do (O. Ind. chi-nd-d-mi) ; fungor
(0. Ind. bhu-na-j-mi) ; li-n-go, from root leigh- (Gk. Acixw) (ef. O. H. G. lecchom,
from *ligh-na-mi) ; vi-n-co, from root weik- (Goth. veiha, ‘I fight’); cla-n-go
(O. Seand. hlakka; cf. Gk. «Aayydvw, xAdfopa), from root klag- (Lith. klagéti,
‘to cackle’) ; pre-he-n-do (Alb. gendem, ‘I am found,’ Lett. gidu, ‘I appre-
hend, perceive,’ for *gendu ; cf. Gk. xavddvw, yeicopat Fut. for *xevd-copar),
from root ghed- (Engl. get, Lat. praeda for *prae-hed-a) ; di-sti-n-yuo (Goth.
stigqa, ‘I thrust’ ; cf. Lith. sténgiu), from the root steig}- (O. Ind. téjaté, ‘is
sharp’; cf. Lat. in-stigare) ; e-mii-n-go (O. Ind. muiicati, ‘he releases,’ Lett.
muku, ‘I escape,’ for *munku) ; fi-n-go (O. Ir. dengaim, ‘I fasten’), from
root dheigh- (Goth. deiga) ; tii-n-do (Pft. tti-tii-di) ; pti-n-go (Pft. pii-piig-i) ; ac-
ctiem-bo (cf. clibare, Pft. ac-ciib-ui) ; té-n-go (Pit. te-tig-i, for *te-tiig-i, Gk. reraydv) ;
lé-m-bo (cf. léb-ium) ; ri-n-gor (O. Sl. reg-na, augmented by -no-), beside rictus.

§12. Retention of Nasal throughout the Tenses. Like jungo, junxi, junctum
(with possibly a Neuter Noun jungus, -eris, ‘a team,’ like Gk. (edyos, in Plaut.
Men. 913: n6n potest haec rés ellebori iingere optinérier),
we have pungo, punctum (but piptigi; Priscian says the Perf. of repungo is
repunxi, like expunxt, or repupugi, i. 524. 13 H.); distinguo, distinai, distinctum ;
Jingo, finxi (but fictum ; finctum, Ter. Eun. 104) ; plango, planaxi, planctum ; emungo,
emuncxi, emunctum ; lingo, lina, linctum ; fungor, functus [but on plebeian inscrip-
tions defuctus (C. I. L. ii. 4173), like sactus, e. g. sactissimae (vi. 15511, V. 6580),
whence the Welsh loanword saith beside sant ; nactus and nanctus are equally
good spellings, see Georges, Lex. Wortf. s.v.]. Spellings in MSS. like corrwmp-
tus (Neue, ii’. 560), relingui, Perf., are due to the same confusion. When the
stem is extended by the YO-suffix, the n is retained, e. g. vincio, vinxi, vinctum,
from root vyek-, beside vinco, vici, victum from root weik-, sanctus from sancio.
Of roots ending in a dental we have e. g. from tundo, tunsus and (post-Aug.)
tusus, and the grammarians speak of a Perf. twnsi (Georges, ». v.); but the
dropping of » before s in pronunciation makes it doubtful how far the nasal
was really present in such forms (see ch. ii. § 66 on thensaurus for Onoaupds ;

ef. mensus from métior).
472 THE LATIN LANGUAGE. (Chap. VIII.

§ 13. Other examples of nasalaffix. Lat. sperno (0. H. G. fir-spirni-t 3 Sg.,
spurnu), with Perf. spre-vi, as cerno Perf. cré-vi (cf. Gk. Tép-v-w, &-Tyj-Onv) 5
O. Lat. degunere (degustare, Paul. Fest. 50. 36 Th. ; cf. Gl. Philox. degunere :
dmoyetcacda Kat ovyyvaoa) for *de-gus-nere from root geus- (Gk. yev(c)w,
Goth. kiusa, Engl. I choose); ap-pell@re and com-pelldre for *-pel-nd-re (Gk, mid-
va-por, I approach, O. Ir. ad-ellaim), beside pellere, to strike ; the Compound-
stem with -nd- has a peculiar sense also in de-sti-nd-re, prae-sti-nd-re, which in
Plautus are used for ‘to buy,’ e.g. Most. 646 quid, eas quanti destinat? ;
Capt. 848 alium piscis praestinatum abire (cf. Arm. sta-na-m, ‘I possess, buy’),
ob-sti-nd-re, to stickle for, Plaut. Aul. 267:

id inhiat, ea affinitatem hanc dbstinauit gratia,

whence obstinatus ; there is a gloss, gredinunda Badifovea, C. G. L. ii. 36. ro.

§ 14. Other Verb-stems with n. From Nasalized Present-stems we must
distinguish (1) O. Lat. forms of the 3 Plur. Pres. Ind. like dénunt, explénunt,
prodinunt, on which see § 73; (2) Derivative A-Verbs from Noun and Adjec-
tive Nasal-stems : e. g. opinor, -dri from a Noun *opion-, connected with praed-
opiont (MSS. praedotiont) ‘praeoptant’ of the Carmen Saliare (Fest. 244.
13 Th.), optio, optare, &c.; festino, -dre ; vulpinor, -Gri, to use the wiles of a fox
(Varro, Men. 327 B.); auctidnor, -dri ; contionor, -Gri; sarcino, -dre ; nomino, -dre
(see above); (3) Verbs in which the nasal belongs to the root, e.g. tendo,
formed from the root ten- by means of the suffix d (§ 33) ; frendo similarly
for frem-d-o ; offendo, defendo from the root ghten-, ‘to strike’ (Gk. deivw for
*Oevyw).

The verb pando is of doubtful origin. Some make it a nasalized form like
unda (cf. Lith. vandti beside Goth. vato, Engl. water; O. Ind. uddn- beside
Gk. #5wp); others make it a word-group, *patem-do, lit. ‘I make opening,’
like vendo and vénum do (Osc. patensins ‘ aperirent’ (?) has also been variously
explained]. Mando, to chew, if connected with Gk. pacdopa: for *yarraopat (?),
will be a parallel formation.

§ 15. With suffix -YO-,-Iy0O-. Like the Noun-suffix -yé-,
-ly8- (ch. v. § 4) this varies with -yé-, -iyé- or with 1-. Owing
to the weakening of vowels in unaccented syllables in Latin, it
is difficult to ascertain the exact form of the suffix in the
various persons of the Present Tense; but the analogy of other
languages points to a declension like this of those Presents in
which -y- varied with -1-: 1 Sg. *eup-yd, 2 Sg. eup-i-s, 3 Sg.
cup-i-t, 1 Pl. *ceup-yo-mos, 2 Pl. *eup-d-tes, 3 Pl. *cup-yo-nt (see
Brugmann, Grundriss, il. § 702).

Two classes of Present-stems with the YO-suffix stand out
very clearly, though they occasionally overlap :

i. With E-grade of root and accent on the root, e.g. Lat.
spéc-io (O. Ind. pas-ya-ti, Zend spas-ye-iti, Gk. oxénrowar for
*oKem-Yo-j.a1).
§§ 13-15.] THE VERB. PRESENT-STEMS. 473

ii. With weak grade of root and accent on the suffix, e.g.
Lat. mér-ior for *my-yor (O. Ind. mr-iyd-té 3 Sg.), vénio for
*g*m-yé (O. Ind.gam-yé-té, Gk. Baivw for*Bm-yw). To the second
belong intransitive verbs (e.g. O. SI. sto-ja, ‘I stand’); and so
intimately connected is this type of the suffix with intransitive
sense, that in Sanscrit its Middle is used as the Passive of all
verbs, e.g. kriyé, ‘I am made,’ kriyaté,‘ he is made,’ the Passive
of karémi, ‘I make,’ karoti, ‘he makes.’ In the Balto-Slavic
family of languages these verbs show in the other tenses an
E-suffix, clearly the same as the Greek Passive -7- of é-ydv-n-v
beside pafvopae for *un-yo-wat, a suffix likewise identified with
the intransitive or passive sense. This conjugation of intransi-
tive verbs is not found in Latin, but it has perhaps left its
mark in the coexistence of Presents in -2o and -eo, e.g. jdcio and
jiceo, pivio and paveo. In most cases however the E-suffix ousted
the YO-suffix altogether, e. g. sédeo, sed-6-s, sed-é-mus, &e. from an
I.-Eur. Present-stem *sed-yd- (Gk. é(ouarfor *oed-yo-ar,O. H.G.
sizzu for *sed-y6), video, vid-é-s, vid-é-mus, &c. from an I.-Eur.
Present-stem *wid-yd- (O. Ind. vid-y4-té, ‘ he is perceived, Lith.
pa-vydziu 1 Sg. Pres.) with another (originally not a Present)
stem in -é- (Lith. pa-vydéti Inf.). [In Goth. vitais for *wid-
éyé-s, vitaip for *wid-éyé-t, and in other Teutonic verbs, the same
intrusion of -é- (-éyd-) into the Pres.-stem is seen as in Latin.]
These Intransitive Verbs with Inf. -ére constitute an important
part of the second Conjugation in Latin, e.g. cilére, riibére,
patere; they acquire a Transitive sense by appending ficio to
a Verb-stem in -é (ch. v. § 51), e.g, caléfacio, and often take as
their Present-stem an ‘ Inceptive’ formation in -sco, e. g. incilesco,
éritbesco (see § 28), The association of this type with the In-
transitive functions is seen in pexdeo Intrans. beside pendo Trans.,
mideo (Gk. paddw), véreor (Gk. dpdw, ch. iv. § 10), edueo and cluo
(Gk. KAdo). f

Another class of Presents ‘which show the YO-suffix is—

iii. With -a, -6, -d after the root. Beside the root pel-, for
example, we have the root plé-, ‘ to fill,’ with a Present-stem *plé-
y6-, ‘ to be filling’ (Lat. im-pleo) ; beside the root ter- we have
the root tré-, ‘to penetrate, (cf. Gk. rpjj-ua) with a Present-stem
*tré-yo- (O. H.G. drau, Germ. drehe), as well as the root tra-,
474 THE LATIN LANGUAGE. [Chap. VIII.

with a Present-stem *tra-yb- (O. Ind. tra-ya-té, Lat. zv-tro for
*_trayo) (see ch. iv. § 66). Unlike the second Conj. verbs just
mentioned, video, sédeo, &c. with Perfects vidi, sédi, Supines visum,
sessum, and ciéleo, rlibeo, &c. with Perfects cdlwi, riibui and with
Supines wanting, these é-verbs retain their é throughout the con-
jugation, -plévi, -plétum, &c. In addition to monosyllabic Verb-
stems we have such dissyllabie stems as Lat. *domd-yé, démo (O.
Ind. dama-yé-ti), and a group of onomatopoetic words, e. g. Lat.
*ul-uld-yo, tidilo (Lith. ul-até-ju, and unreduplicated utd-ju, Gk.
brAdw for *iAd-yw). Beside the Present-stems with the YO-suffix
we find athematic Presents from these roots with -a, -é, -6 (e. g.
O. Ind. tra-ti beside tra-ya-té, Gk. aiy-mdn-pas, Ti-Tpy-pe, Ki-xn-pL)
which seem to have originally retained the long vowel throughout,
and not to have variation with the weak grade (e. g. O, Ind. tra-sva
2 Sg. Imper. Mid., Gk. «i-yn-wev Pl.); and in Latin this athematic
formation appears to be used in all persons but the first, 2-¢rd-s,
im-tra-mus, &c., though this cannot be proved, seeing that, e.g.
im-plés is equally derivable from thematic *-plé-yé-s (cf. trés for
*tréy-és, ch. iv. § 66) as from athematic *-plé-s (O. Ind. pra-si).
The YO-suffix played a great part in the I.-Eur. languages as
a secondary suffix, added to Verb-stems, e.g. Lat. pinsio beside
pinso (an already-formed Pres.-stem, § 10), Gk. éc@fw beside
écOw, or to Noun-stems, &c. to form derivative verbs, e. g. Lat.
cio, for *curd-yo (Umbr. kuraia, “curet *) from the Noun-stem
*curd-, claudeo for *claudé-yo from the Adjective-stem *claudé-,
*claudo-, finio for *fini-yo from the Noun-stem */ini-, stdtuo for
*statu-yd from the Noun-stem *statu-, custédio for *eustod-yo
from the Noun-stem *custéd-,and so on. But since the suffix is
in these derivatives usually maintained throughout the Latin
conjugation, pinsttus, custodivi, custoditus, &e., they are better
reserved for discussion among the Verb-suffixes in § 26 (cf. Gk.
dai-ow, dat-vi-u. beside daiw for *da-yw). Derivatives from
A-stems follow the analogy of roots with -a (e.g. Lat. in-tro,
in-tras, in-trdmus, see above) in using the YO-suffix only to form
the thematic ist Pers. Sg. of the Present Tense, while those
from U-stems use it in all persons, e. g. stdtuo, statuis, statwimus.
For Latin athematie forms like cwrd-mus derived from stem
eurd-, &e. we may compare the athematic flexion of similar
§§ 16, 17.] THE VERB. PRESENT-STEMS. 475

Derivative Verbs in the Lesbian dialect, e. g@. tiva-pev, derived
from stem tima-, pidn-m, derived from stem uAc-, orepdvayt,
derived from stem oredavo-.

English examples of Present-stems with the YO-suffix are
‘I lie’ [O. H.G, ligg(iju, but Pret. lag, «I lay’], and the two
Pres. Participles which have become Nouns, ‘a fiend’ (Goth.
fijands, lit. ‘hating’),* a friend’ (Goth. frijonds, lit. «loving y.

§ 16. i in the third Conj. Presents with YO-suffix. We have i often in
O. Lat. ; cupis, Plaut. Cure. 364:

laudo. Laudaté, quando illud, quod cupis, effécero,

Sacis, Amph. 555 (so the MSS.) ; facit, Cure. 258 (2) ; inlicite is the scansion required
by the metre in a line of Naevius (Trag. 30 R.):

sublimen altos sdltus inlicite, ubi
bipedés uolantes (MSS, uolucres) lino linquant lumina ;

in 1 Pers, Plur., morimur is attested by Priscian (i. p. 501. 16 H.) in a couplet
of Ennius (Ann. 415 M.):

nune est ille dies, cum gloria maxima sese
nobis ostendat, si uiujimus siue morimur ;

we have adgredimur, Plaut. Asin. 680, Rud. 299; and in 3 Sg. Dep. adoritur is
attested by Prise. (J.c.)in a line of Lucilius (or Lueretius iii. 515 ?). The long
vowel is especially common in the Inf. of the Deponent in Plautus, e. g.
adgrediri, moriri, effodiri (ef. partre). (For other examples, see Neue, ii?. p. 415.)
These forms can hardly be due to the false Analogy of verbs like finio, -ivi,
-itum, -ire, such as is seen in Late Lat. farci-tus (coined on the type of
Jini-tus), for they are a feature of the older language. They rather indicate
that in the period of the early literature the suffix might appear as 7 or as ?,
whereas in the classical period the usage became restricted to one or other of
these forms The best explanation then of Verbs in -io 1 Sg. Pres. Ind.
which belong to the third Conjugation is that they are YO-stems in which -i-
asserted itself, rather than -i-, as the weak grade of -y- (-yé-) ; capére will then
stand for *capisi, cape for *capi. This does not however preclude the possi-
bility of other explanations being right in particular cases, such as that
a bystem without -yd- existed, say */ac- beside *fac-ys- (cf. bene-ficent-ior beside
faciens), *rap- beside *rap-y6 (ef. rapo, ‘a robber,’ Varro, Men. 378 B.), that -7-
belongs to a stem in -iyé-, -i- toa stemin -yé-, e.g. spécio from stem *spek-yo-
(ef. Gk. oxérrw) with the Y0-suffix immediately following on a consonant.

§17. Other examples of E-grade roots. Lat. *vério, ‘to close,’ seen in
op-(w)erio, ap-(w)erio (Lith. uz-veriu, ‘I shut,’ at-veriu, ‘I open’) (on the loss
of w, see ch. iv. § 71); id-uo for *indovo from *-ew-y6, to judge from Umbr.
an-ovihi-mu for *and-ov-i-mu, ‘induimino’ (ch. iii. § 24) from the root ew-,
‘toputon.’ Similarly hawrio for aurio (ecaurio is the almost invariable spelling
476 THE LATIN LANGUAGE. [Chap. VIII.

of Latin Glossaries, Liéwe, Prodr. p. 371 ».) from root aus- (ef. Gk. &atoa, ‘to
take out,’ égavornp, ‘a flesh-hook’); crdc-io (Gk. xpw(w from *xpwy-yo, Lith,
krok-it and krog-iu).

§ 18. Of weak grade roots. i. With -io. Lat. ctip-io (O. Ind. kup-ya-ti, ‘ is in
agitation’) ; frgio (ef. Hom. we-pu(dres for -puyyo-) from root bheug- (Gk. pedyw) ;
grédior from the root ghredh- (Goth. gridi- F., ‘a step,’ O. Sl. greda, ‘I come,’
O. Ir. in-grennim, ‘I pursue, attack,’ the last two with Nasalized stem) shows
the weakening of -ré- to -rd- mentioned in ch. iv. § 51; similarly farcio for
*fracio (Gk. ppaoow) from the root bhreq?- of frequens (and for the connexion
of ‘cramming’ and ‘frequency,’ cf. saepe and O. Lat. saepissumus, ‘ closely
packed,’ ch. ix. § 4); likewise rdpio, if connected with répens, ‘ sudden.’

ii. Intransitive with -co. Lat. riibeo from an I.-Eur. *riidh-yo (0. Sl. rii%da,
with the 6-suffix in Inf. riidé-ti) ; torpeo with trp- the weak grade of the root
terp-; stiideo (cf. Gk. omediu?) ; méineo (cf. Gk. wévw), and pateo (cf. Gk. wer-avvupt),
both seem to show Lat. dé as a weak grade of é (ch. iv. § 3). Like ribeo with
Noun riibor, torpeo with Noun torpor and Adj. torpidus are a large number of In~
transitive Verbs, céleo (calor, calidus), pliceo (placidus; but Transitive pldco),
liqueo (liquor, liquidus ; ef. Nquor, third Conj.), and so on (ch. v. § 74). The téneo
of per-tineo, lit. ‘to reach through,’ trans-tineo, ‘to reach across’ (Plaut. Mil.
468 commeatus transtinet trans parietem) is the Neuter of tendo,

§ 19. Alternative forms in-o and-eo. Lat. tweor and tuor, ‘to look’ ; fulgeo
and fulgo ; ferveo and fervo; sciteo and scato; abnuo and O. Lat. abnueo (Diom.
382.11 K.). In all of these the form in -o is the older (e. g. contuor, intwor Plaut.,
scato Plaut., Enn., Luer. ; fervit and fervére are common in the early Dramatists,
the latter often in Virgil, but to Quintilian a third Conjugation form of this
verb is ‘ inauditum,’ Quint. i. 6. 8), while the form in -eo is a new formation
on the Analogy of the numerous Intransitives in -co (ef. Caper 109. 16 K. fido
non ‘ fideo’). We have sordtre in Plaut. Poen. 1179. (See also below, § 33.)

§ 20. Of roots with -a, -é, -6. Latin no for *snd-yo (O. Ind. sna-ya-té,
and athem. sna-ti, Lat. nat) ; neo for *(s)né-yo (Gk. véw, O. H.G. nau, Germ.
nihe), téceo for *tacé-yo (Goth. pahaip 3 Sg. from *také-yé-ti, and athem. O. H.G.
dagé-s, Lat. tacés) ; flo for *fla-yo (cf. O. H. G. blau from *bhlé-yo, perhaps the
same word as Lat. fleo for */le-yo, Gk. préw, to overflow) ; hio for *hid-yd (Lith.
2i6-ju) from the root ghei-; jitvo for *juvd-yo, I.-Eur. *dytiga-yo, from the root
dyeug}- (ef. Lith. dziung-ii-s, ‘I rejoice,’ a nasalized Present) (butcf. ch. iv. § 64) ;
ciibo for *cub-dyd from the root keubh- (cf. -cumbo, a nasalized Present). Of
onomatopoetie words with 1 Sg. Pres. in -dy, we have murmuro, tintinno, unco,
to bray (Gk. éyedopar).

§ 21. Inceptives, and other Verb-stems. Though Inceptive
verbs by virtue of their meaning restrict, as a rule, the in-
ceptive suffix -skd- (-ski-) to the Present sense, e.g. cresco, Pft.
crévi; they differ from Present-stem formations like s%wo, Pft.
sivi, cayio, Pit. cépi, in this respect, that the meaning, which
they express, is something more than the mere sense of action in
present time, e.g. séuesco means, not ‘I am old’ (séneo), but ‘1
become old.’ They are therefore better considered in a separate
§§ 18-23.] THE VERB. PRESENT-STEMS. 477

section, along with some verb-formations which are more than
mere Tense-stems, such as Causatives, Intensives, Desideratives,
and the like.

§ 22. Inceptives in -sk6- (-sk6-). The root shows, as a rule,
the weak grade, e.g. I.-Eur. *prk-skd- (O. Ind. pr-cha-mi, with
*-skh- for *-sk-), Lat. posco for *pore-sco, from the root prek-.
It is sometimes reduplicated in Greek, e. g. 3u3doKw for *81-Sax-cKw,
but not in Latin unless disco, from root deik-, stands for *di-de-sco,
(cf. di-dic-i) and not for *di%e-sco. An English Inceptive is
‘I wash’ (O. Engl. wesce, from a Teut. *wat-sko ist Se.,
derived from the same root as ‘ wet,’ ‘ water, Lat. unda, &c.).

§ 23. Causatives and Intensives in -eyo-. The root has the
O-grade, and the accent rests on the first syllable of the suffix.
Causatives of this type are a regular feature of the Sanscrit
conjugation, and may be formed from any verb, e. g. man-dya-mi
(Lat. ménxeo for *mon-éyd) from the root man- (I. Eur. men- ;
cf. Lat. mémtni for *me-men-i); tar$-dya-mi (Lat. zorreo, O. H. G.
derr(i)u, for *trs-éy6) from the root trs- (1.-Eur. ters-), so that
Lat. moneo was literally ‘to cause to remember’ [cf. Plaut. M7.
49 Edepé] memoria’s ‘éptima. Offaé monent; Paul. Fest.
115. 6 Th. ‘monitores’ qui in scaena monent histriones (our
‘ prompters’)], forreo ‘to cause to be dry.’ These Causatives of
Sanscrit have a different accent from Derivatives in -éyd- from
O-stems, in which the accent falls on the suffix -yd-, e.g. déva-
ya-mi, ‘I honour the gods,’ from déva-, ‘a god’ (an O-stem,
*deiwd-). The same formation often has the Intensive or the
Iterative sense, e.g. Gk. dopéw from [.-Eur. *bhér-éyo, ‘I carry
frequently,’ from the I.-Eur. root bher-, ‘to carry’ (Gk. ¢épw,
Lat. fero); Gk. row for *podew from I.-Eur. *gh*odh-éy6 (O. Ir.
guidiu), ‘I ask or desire earnestly, from the I.-Eur. root gh*edh-,
‘to ask’ (Gk. Oéccer0a). The P.P. P. of these verbs shows -3-
in some languages (O. Ind. varti-té-, Goth fra-vardi-ps), -i- in
others (cf. Lith. varty-tiInf.,O.S1. vrati-ti) ; in Lat.?,e.g¢. monitus,
nocitns ; and there are indications that the J.-Eur, declension of
the Present Ind. was *wortéyd 1 Sg., *wortimds 1 Pl., &e. (P. B.
Beitr, xviii.p.519). An English example is ‘I lay ’(Pres.), inGoth.
lagja, from I.-Eur. *loghéyo; ‘I Jay’ is the Causative of ‘I lie.’
478 THE LATIN LANGUAGE. [Chap. VIII.

§ 24, Latin Desideratives in -tirio. These are formed
with the YO-suflix (§ 15), probably from Verbal Noun-stems in
-tor-, e.g. parturio from partor, scripturio from seriptor, with the
same change of unaccented ¢ to # as in fulgitro, O. Lat. fulgdrio
(ch, iii. § 26).

§ 25. Latin Iteratives or Frequentatives in -*tayd- are
formed from Perf. Part. Pass.-stems, or rather from the Fem. of
these used as a Noun (ef. offensa beside offensus, répulsa beside
repulsus), with the help of the YO-suffix, e. g. pulso, older pulto,
for *pultd-yd, from pulsus, older pultus, P. P. P. of pello. Some-
times the TO-suffix is doubled, e. g. factito, ventito.

§ 26. Other Derivative Verbs with the YO-suffix. The
ending -ayé-, which properly belongs to Derivatives from A-stem
Nouns (e.g. from planta, a plant, a shoot, plantare, to plant, lit.
‘to make or turn something into a plant’), acquired a transitive
sense, and was used in Latin, as in other J.-Eur. languages, with
any Noun- or Adj.-stem, e.g. clarare, ‘to make clear,’ from the
Adj.-stem claro-, pulverare, ‘ to turn something into dust,’ or ‘ to
cover with dust, from the Noun pulvis, a Consonant-stem. The
ending -éyo-, which properly belongs partly to Derivatives from
O-stem Nouns or Adjectives (e.g. clandeo,‘ to be lame,’ from
clandus, albeo,‘ to be white, from albus), partly, as a Primary
suffix, to Intransitive Verbs like radeo, sedeo (see § 32), is the
corresponding intransitive formation, e.g. clarere, ‘to be clear.’
Latin Verbs in -io include Derivative Verbs from Consonant-
stems, e.g. custéd-io from the stem custod-, and from I-stems,
e.g. iudnio from the Adj. I-stem imani-. This ending acquired
to some extent an intransitive sense, expressing a state of body
or of mind, and was in this capacity applied to other stems too,
e. g. insdnio from the Adj. O-stem insanus, to be mad, saevio from
the Adj. O-stem saevus, to be fierce,

§ 27. Other suffixes commonly used in forming Verb-stems,
primitive suffixes, the sense conveyed by which cannot now be
detected, were (1) -dh-,e. g. Gk. xv7}-0-w beside xvdw, (2) -d-, e.g.
Gk. €d-60-yar from root wel- (Lat. volo), which may be nothing
§§ 24-28.) THE VERB. INCEPTIVES. 479

but the Verb-stems dhé-, * to put,’ and dé-,‘ to give’ (cf. Lat.
eredo, O. Ir. cretim with O. Ind. Sréd dadhami, lit. ‘I set heart
to’), as the common Latin ending ~2go, -dre, e.g. ndvigo, seems to
be nothing else than the Verb digo (from *navigus; ef. ch. v.
§ 80); (3) -t-,eg. Gk. aéx-r-w (Lat. pecto) and mexréw, beside
méxw, which seems connected with the P. P. P. suffix -to- (ch. v.
$27); (4) -s-,e o. Gk. d€yw beside dépu, tpé-(o)-w beside rpé-p.-w,
which seems the same as the ES-suffix of Nouns, e.g. Gk. réos,
stem tedeo- (Gk. red€(c)w, re-réeo-rar) (§ 3). Latin examples
are: gau-d-eo (Gk. yjOouar and ynbéw) for *gavi-d-eo (cf. gavisus),
sallo for *sal-do (Goth. salta), plecto (O. H. G. flih-tu, Germ.
flechte) beside plico, -dre (Gk. mAdkw), viso, older visso, veisso for
*weid-so (Goth. ga-veisd ; cf. the O. Ind. Desiderative vi-vit-sa-
mi), guaeso, older quaes-so for *guais-so, beside guaero for *quaiso,
in-cesso for *in-ced-so (cf. cédo). Other Latin endings are: (5)
-880, e.g. cdipesso, incipisso (Plaut.), pétesso, O. Lat. petisso (Fest.
250. 19 Th.) from cépio, péto (or a bystem *petio, whence petivi,
§ 47); on these see § 3; (6) -Zo and -id/o of Diminutive Verbs,
e.g. conscrihillo Catull.; (7) -co of albico, fodico, &e. ; the last
two are like Derivative A-verbs and belong to the first Conju-
gation, conscribillare, fodicare, albicare, like médicari from medicus.
Similarly, (8) -r0 of damb-éro (§ 41) is like -ro of the Derivative
tempérare from tempus.

§ 28. Other examples of Latin Inceptives. Misc-eo has added the Causa-
tive ending to a lost *misco [cf. misc sane for misc? sane on an old Praenestine
cista (§ 58)] for *mik-sko [the shortness of the 7 is seen in the Romance forms,
such as Ital. (Tusc.) mesci, ‘give me a drink,’ the Latin misce mi] with the
weak grade of the root meik-. But the E-grade is retained in O. Lat. esco for
*es-sco, the Inceptive of the root es-, ‘to be,’ used for ero (or rather for swm) in
the Laws of the XII Tables : si morbus aewitasue uitium escit, and ast et custos nec
escit, &c., and even by Lucr. i. 619:

ergo rerum inter summam minimamve quid escit ?,

as by Ennius, A. 322 M.:

dum quidém unus homo Romanus toga superescit.

Roots extended by -4, -6, -like @nd- from gen-, keep this vowel long, as is their
custom in such eases (§ 2); hence (g)né-sco (Gk. y-yvdonw, Epir. yrwouw),
(g)nd-scor, cré-sco, vit-sco, hid-sco ; and similarly Latin Intransitives in -eo (§ 32:
and Derivatives in -o (for *-Ay5), -¢0, -io (ib.), e.g. rubé-sco, con-tice-sco, tra-
scor, fldvé-sco, ob-dormi-sco, erci-sco [erceiscunda on the Lex Rubria, C. I. L. i.
205. (2). 55], descisco (with tall form of I on Mon. Ane. v. 28, which also offers
480 THE LATIN LANGUAGE. [Chap. VIII.

nascerer With an apex over the a); though ata later time, when the difference
of quantity between vowels had become less marked, we find some uncertainty
about the e of quiesco (see Gellius, vii. 15, who decides in favour of quiésco, on
the strength of ccilésco, nitésco, stiipésco and other Inceptives ; ef. ch. ii. § 144).

The name ‘Inceptive’ is unsuitable. It is only verbs of the second Con-
jugation uncompounded with a Preposition, such as calesco, léquesco, to which
a notion of ‘beginning’ can be attached, and even there the notion conveyed
by the suffix is rather that of passing into a state or condition, of ‘becoming’
than of ‘beginning,’ e. g. liquesco, ‘to pass into a liquid state,’ ‘to become
liquid.’ The suffix is closely associated with Intransitive Verbs of the second
Conjugation, so closely indeed that these, when compounded with the
Prepositions cum, e2, in (Prepositions which convey the idea of ‘becoming’),
always form their Present-stem with this suffix in good authors, e. g. érvibesco
(not erubeo), conréilesco (not convaleo), inardesco (not inardeo), unless the Preposi-
tion retains its separate force, e. g. é-luceo, ‘to shine out,’ co-haereo, ‘to be united
with,’ which have the force of luceo ex, haereo cum. Intransitive Derivatives
from stems like dulci-, igni-, grivi-, &c. are formed on this type, and take -esco
instead of -isco, dulcesco, ignesco, gravesco, mitesco, pinguesco, &c.; and -dsco of
Intransitive Derivatives from A-stems, &c., is in Late Latin often changed to
-esco, e. g. gemmesco for yemmasco. The spelling -isco for -esco in Late Latin, ve. g.
erubisco, may often be a mere interchange of the similarly sounding vowels ¢
and @ (see ch. ii. § 14), but it may also be referred to the Late Latin importa-
tion of Verbs of the second Conjugation into the fourth (e.g. floriet, flurient in
the Itala), which has left its mark on the Romance languages, e. g. Ital. appa-
rire (apparisco Pres.) from Lat. apparére (§ 33.4). The -iscor of dpiscor, nanciscor,
may be referred to the old forms apio, nancio (whence coepiam, Paul. Fest. 41.
34 Th., nanciam, Prise. i. 513. 17 H.), as the -isso of O. Lat. pitisso to a lost *petio
(whence petivi, § 47). ‘Inceptives’ from fourth Conjugation Verbs are for the
most part ante-classical, e.g. condormisco (Plaut.), édormisco (Plaut., Ter.), perpri-
risco (Plaut. Stich. 761), persentisco (Plaut., Ter.), but obdormisco, ‘ to fall asleep,’ is
used by Cicero (Tuse. i. 49. 117). They are mostly Compounds (except scisco),
and the same is true of the ‘ Inceptives’ from third Conjugation Verbs in the
Republican writers, e.g. résipisco (Plaut., &c.), préficiscor (cf. ficessere, ‘to take
oneself off,’ and Late Lat. se fucere, ‘to betake oneself,’ e.g. intra limen sese
facit, Apul.), conciipisco (Cic., Sall., &c.), impliciscier, to become affected (by a
disease), Plaut. Amph. 729:

ubi primum tibi sensisti, mUlier, inpliciscier ?,

though in the poets and later prose writers we have trémesco, gémesco, &c.
‘Inceptives’ from first Conjugation Verbs found in the early writers are
cmasco (Naev.), hiasco (Cat.), ldbasco (Plaut., Ter., Lucr.), collabasco, permanasco,
destidasco, and a few others; but this formation was not continued in the
classical period, though we find Derivatives in -asco, derived from Noun- and
Adj.-stems, e.g. veterasco, vesperasco, gemmasco, in which the suffix -sco seems to
be added to change the transitive sense attaching to these Derivative A-verbs
(e.g. clarare, to make clear, to clarify, § 32) into an intransitive.

It thus appears that an intransitive sense attached to the suffix -sco- in
Latin, and that this was the reason of its close association with the Intransi-
tive Second Conjugation. Its sense of ‘passing into a state or condition’
suited it for acting as the Present Tense-stem of Intransitive Verbs. A
§ 29.) ‘ THE VERB. CAUSATIVES. 481

verb with this suffix did not govern an Accusative, unless the simple verb
from which it was formed governed an Accusative, e. g. perhorrescere aliquid,
Cic., like horrere aliquid (a construction of horreo, piveo, &e., not found before
Cicero’s time). But in the fifth cent. a. p. the termination acquired a causa-
tive sense, e. g. inndtescere, to make known, inform ; mollescere, to make soft, not
“to become soft,’ asense which was properly expressed bf the Auxiliary fucio,
e. g. calé-facio, rubé-facio. Assuesco, insuesco, mansucsco, and other compounds of
suesco had at an earlier period assumed the sense of assué-fucio, mansué-facio,
and perhaps supplied the type for this new formation, which was widely
extended in the Romance languages (cf. § 33 «).

Of individual ‘Inceptive’ Verbs may be noticed : callesco, in whose Perfect
Cato retained the Inceptive suffix, callescerunt 3 Pl. [Nonius 89. 26 M. quotes
this form (MSS. calliscerunt) from Cato’s speech on the Punic War: aures
nobis callescerunt ad iniurias] ; obsdlesco and exélesco from séleo with P.P. P.
obsolétus, exolétus ; uddlesco and codlesco (cdl-) from the root al- (ol-), ‘to grow,
nourish’ (whence indéles, subdles, priles), with P. P. P. adultus, coalitus.

For a list of Latin ‘ Inceptives,’ and full details of their history, see A. L. L.
i. 465 sqq. Umbro-Oscan examples are Umbr. pepurkurent ‘ rogaverint,’
Ose. comparascuster ‘ consulta erit.’

§ 29. Of Latin Causatives, &c. O. Lat. liiceo, ‘ to cause to shine,’ to light
or kindle (e.g. Enn. 4. 158 M.: prodinunt famuli, tum candida lumina
lucent; Plaut. Curc. g tuté tibi puer es: lattus luces céreum; .Cus. 118
primum é6mnium huic lucébis nouae nuptaé facem) may be a Causative form,
I.-Eur. *louk-éyo6 (O. Ind. réciya-mi), and different from liiceo, to shine, which
seems to be an intransitive form like sédeo, with -eo instead of -io, I.-Eur.
*leuk-yo (Gk. Aevoow); néceo, I.-Eur. *ndk-éyd (O. Ind. naSiyi-mi) is the
Causative of the root nek- (Lat. nex’, and has in Late and Vulgar Latin the
construction which we should expect, viz. with the Accusative case ; its use
with the Dat. in class. Latin must be due to the analogy of obesse, officere, &e.
The rivalry between Transitive A-stems and these Causative-stems, best seen
in Lithuanian, where the Causatives (e. g. varty-ti Inf.) form their Present-
stems with -a- (e.g. varto- for I.-Eur. *worta-), appears in Lat. nécare beside
nocére ; domare may be I.-Eur. *déma- (0. H.G. zamé-, ‘to tame’), « byform
of I.-Eur. *déméyo- (Goth. tamja, O. H.G. zemm(i)u). In O. Ind. we have
examples of verbs with this suffix which have not the O-grade of root, but the
weak grade, e. g. grbh-dya-ti, ‘he seizes.’ Perhaps Latin ci-eo (beside cio) belongs
to this type. But the ending -co is sometimes added to other Verb-stems
which have a Causative sense. Thus in Lat. misc-eo it is added to a stem
formed already with the inceptive suffix, so that misceo for *mic-sc-éyo has
really two suffixes (cf. O, Ind. dhinaya-ti, ‘he shakes, shatters,’ beside dhu-
na-ti and dhii-né-ti, Gk. eiAéw beside e/Aw for *2A-v-w), And this may be the
true explanation of cieo also. Jtibeo for *yii-dh-eyo [from the root yeu- with
the formative suffix -dh- (see § 27), ef. O. Ind. yé-dha-ti, ‘is set in motion,’
Lith. j-un-di, ‘I am set in motion’] is spelt in the 8. C. Bacch. with -ou-, the
diphthong always found in the perfect (C. I. L. i. 196, 1, 27 ioubeatis, 1. 9, 1. 18
tousiset ; cf. iousit 5474, 1166, touserunt 199, 1. 4, iouserit 198, 1. 12) ; this, if not
a misspelling (ch. iv. § 37), will exactly correspond with the O. Ind. causative
yodhaya-mi (I.-Eur. *youdhéys). Terreo, for which we should expect
*torreo, has in Umbrian the O-grade of root (Umbr. tursitu, O. Umbr. tusetu,
‘ terreto’).

11
482 THE LATIN LANGUAGE. [Chap. VIII.

Other Causatives, or Transitive Verbs with the Causative -co appended, are
végeo (older rigeo? , to rouse to life, c.g. Pomponius, Com. 78 R. animos Venus
veget voluptatibus, déceo, suddco, urgeo, tondeo, torqueo, mor deo, spondeo, augeo, &e.

§ 80. Of Latin Desideratives. These were called by the Latin gram-
marians ‘ Meditativa.’ They were avoided in the higher literature and went
out of use in Late Latin. They are not found in the Romance languages.
Examples of Desiderative Verbs are éstirio, partitrio, emptiirio, céndtiirio (see
A.L.L. i. 408). Verbs in -wrrio (-iirio), e. g. ligurrio, sedturio, are a quite distinct
class, being apparently Derivatives from Verbal Nouns in -wris (€. g. sécuris)
or -iira (e.g. figura).

§ 31. Of Latin Iteratives. The distinction of (1) ‘Iteratives’ in -tito, (2)
‘Intensives’ in -fo, -so is untenable. The suffix in all its forms denotes
repeated action ; the usual type is that of a Derivative A-Verb from a Perfect
Participle Passive, c. g. détare, dormitare, though from Verbs of the first Conju-
gation we have sometimes forms in -io like cldmito, vicito, vilito, pointing to
P.P.P. vocitus like crtpttus (§ 92). As the to-suffix of the P.P. P. became in
time so- (ch. iv. § 155), we have Iteratives in an older form, merto, pulto
(‘mertare’ atque ‘pultare’ dicebant Quint. i. 4. 14 ; Plautus puns on puliem
Subj. and pultem Acc. of puls in Poen. 729), and in the class. form merso, pulso.
Iteratives which add the suffix to a Present-stem, e. g. sciscito Plaut., noscito
Plaut., dgito, are especially frequent in Late Latin, e. g. mergito Tertull., misctto
Script. Gromat., while to the class of Iteratives with double suffix belong actito,
lectito, cantito, dictito, haesito, jactito, ventito, victito, curstto, factito, Vulg.-Lat. *taxitare,
whence Ital. tastare, Fr. tater, ‘to taste, try.’

Iteratives are especially used in anteclass. and postclass. Latin. They are
avoided by Terence, and not much used by Cicero and Caesar, hardly at all
by the Augustan writers; in fact they seem to have been regarded as a part
of the uncultured speech. In the Romance languages they have often taken
the place of the parent verb, e. g. Fr. jeter (Lat. jactare), to throw (Lat. jécio),
mériter (Lat. méritare), to deserve (Lat. méreo), chanter (Lat. cantare), to sing
(Lat. céino). (See A. L.L. iv. 197.) Diiditare ig the Iterative of an O. Lat. verb
dubare (Paul. Fest. 47. 18 Th. ‘dubat,’ dubitat) ; hortart of an O. Lat. *horior
[attested in 3 Sg. horitur by Diomedes (p. 382. 23 K.) for Ennius (A. 465 M.):
prandere iubét horiturque], which seems to be a Deponent of I.-Eur. *ghr-yd
(Gk. yxaipw), a byform of *gher-yo (O. Ind. har-ya-mi, ‘I delight in’; *heriu,
the Umbro-Osean word corresponding to Lat. volo, e.g. Umbr, heris, ‘ vis,’
Ose. heriiad, ‘velit,’ whence Herentas, the Oscan name of Venus) from the
root Sher-; giistare is apparently an I.-Eur. Iterative of this type (0. H.G.
costén) from *étis-to-, P. P. P. of Zeus-, ‘to taste’ (Gk. yed(c)ev, Goth. kiusan,
‘to approve,’ Engl. choose : ef. Germ. Kur-fiirst) ; ito, -are (Gk. irn-réov) from
*itus P. P. P. of co, for *itayo (Umbr. etaians, ‘itent’) ; piito, -are, to prune, to
think, lit. ‘to sift or cleanse often’ (in Romance ‘ to prune,’ c. 8. Ital. potare),
from a P. P. P. stem *pti-to-, ‘cleansed’ (Lat. piitus, clean, in the phrase piirus
piitus, e.g. Plaut. Pseud. 1200), from the root of Lat. pi-rus, for *pitayd (ef.
0.8). pytaja, ‘I investigate,’ with ti) ; dimito from domitus ; crépito from crepitus ;
habito for habitus, which monopolized in class. Lat. the sense of ‘to dwell,’
‘inhabit’ (cf. archaic Engl. ‘to keep,’ as in the Merchant of Venice, iii. 3: it is
the most impenetrable cur That ever kept with man), a sense which it shared
§§ 30-32.] THE VERB. DERIVATIVES. 483

in 0, Lat. with nébeo, e.g. Plaut. Men. 308: nén tu in illisce aédibus Hahes ?
Di illos homines, qui illic habitant, pérduint.

§ 32, Of Latin Derivative verbs with YO-suffix. The onomatopocetic
verbs tintinnio (also tinnio and tintinnare), gingrio, to cackle, of geese (whence
gingrina, the name of asmall size of fife : genus quoddam tibiarum exiguarum,
Paul. Fest. 67. 23 Th.; cf. Gl. Philox.) have a formation analogous to the
Sanscrit Intensives (c.g, nan-nam-ya-té from the root nam-, ‘to bend’) and
to Greek rappaivw for *rav-pav-yw, pappaipw for *pap-pap-yw, &c., that is to. say
with the suffix -yé- appended to the fully reduplicated root.

The suffix -yé-, as was remarked before (§ 10), is often added to nasalized
Present-stems, especially in Greek, e. g. cAtvw for *At-v-yw from the root Klei-,
andso we have : linio, a Late Lat. derivative from lino, the Present tense of the
root lei- ; pinsio (Gk. mricow for *rrwo-yw) beside pinso; vincio from *vi-n-co, the
nasalized Present of the root vyek- (O. Ind. vi-vyak-ti, ‘he encompasses’) ;.
sancio beside scicer. These derivatives naturally retain the nasal throughout
the verb, e.g. P.P.P. pinsitus (but from pinso, pistus), sunctus (in Vulg. Lat.
sactus, ch. ii. § 70).

Examples of Verbs in -ayi- from Noun A-stems are: scintillo, -ure from scintilla ;
ldcrimo, -are from lacrima ; multo, -are (Osc. moltaum Inf.) from muita; insidior,
-ari, insidiae ; mdciilo, -are from macula; praedor, -ari from praeda; mérari from
mora [in O. Lat. always transitive, ‘to cause delay,’ ‘to detain,’ whence nil
moror (hane rem), I do not care for, lit. ‘I do not (care to) detain’]. Lat.
poenio, piinio, from poena, may exhibit an alternative method of forming deriva-
tive verbs from Noun A-stems, viz. with the mere suffix -yé-, the final vowel
of the Noun-stem being suppressed, as in derivative Adjectives like Gk. rip-tos
from tip} ; or may follow the analogy of derivatives from I-stems, or deriva-
tives from Consonant-stems.

Of Transitive Verbs from O-stems : amplunt, ‘pro amplificant,’ Pacuv. Trag.
339 R. (ap. Non. 506. 30 M.; ; novo, -are from novus ; scno, -are from sanus ; narro,
-are from gndrus (?) (ch. ii, § 132); armo, -are from arma Pl. ; spétior, -ari from
spolium ; numéro, -are from numerus ; lico, -are from locus ; dino, -are from donum ;
ciimiilo, -are from cumulus ; damno, -are from damnum. This use of the -AYO-
suffix for Derivative Verbs from O-stem Nouns and Adjectives is common in
all L.-Eur. languages, e.g. Goth. frijo, ‘I love’ (of which Engl. ‘friend’ is a
Pres, Part.), 0. Sl. prija-ja, O. Ind. priya-yd-té 3 Sg., all from an I.-Eur, O-stem,
*priyo- (0.-Ind. priyd-, ‘ dear’), O. Iv. caraid, ‘he loves,’ from I.-Eur. karo-,
‘dear’ (Lat. carus). On the use of -dis- as a Participial Adjective suffix, mean-
ing ‘provided with,’ ‘clothed in,’ &., e.g. armatus, dentatus, pilatus, from
pilum, cordatus in Ennius’ egregie cordatus homo (ef. re-corddri), see ch. v. § 28.
It is probably seen in Gaulish ya:cara: [gaesati, ‘Gaulish mercenaries,’ C. G. L.
v. 71. 23 (?)], from Gaulo-Latin gaesum, a spear.

The natural formation from O-stems is sometimes in -éyé-, e.g. Gk. piréw,
to love, for *piAc-yw, from pidos, dear (stem giAo- or piAc-, ch. v. § 2), ddvvaréw,
to be unable, from dévvaros, unable ; sometimes in -yé- merely, e.g. GK. pe-
Aicow for *peAcx-yw from peiAcxos. With the latter Lat. anio from unus, bland-
ior from blandus may be compared, as in Noun derivative YO-stems we have
somn-tum from somnus, Octévius from Octavius, &e. (ch. v. § 4) ; with the former,
intransitive Verbs from second Declension Adjectives, like claudeo from
claudus, albeo from albus, cléreo from clarus, fldveo from flavus [as in Adjective
derivative YO-stems like aurens from awrum (ib.)], unless these follow the

112
484 THE LATIN LANGUAGE. [Chap. VIII.
analogy of Intransitives like sédeo for I.-Eur. *sed-yo (see above, § 15),
in which case their ending will be not -8yd but -éyo. (The correspond-
ing verbs in Balto-Slavie have -éyé, e.g. Lith. kété-ju, ‘ I grow hard,’ from
kétas, ‘hard’), The same distinction between transitive -0, -dre and
intransitive -¢o, -2re is seen in primary verbs like Niquare and liquere. (On the
proneness of Intransitive Verbs to take -eo, e.g. ferv-co, fulg-e0, O. Lat. jfervo,
Sul-go, see § 19.)

We have also Participial Adjectives in -dtus, e.g. aegrdtus from aeger (stem
aegro-), like Gk. proOw-rds from pio6ds, Lith. ragii-tas from riigas, ‘a horn,’ with
corresponding Verb-stems in Greek and Lithuanian, ve. g. Gk. pucdéw, Lith.
jakit-ju, but no *aegrdd or *aegré, -ds, -Omusin Latin. Derivatives from I-stems
have -iyé-, e.g. O. Ind. kaviyé-té, ‘he acts like a seer,’ from kavi-, ‘a seer,’
janiya-ti, ‘he desires a wife,’ from jani-,‘a wife’; Gk. pyriopa (¢) from pars,
pnvio (¢) from pis, xovia (i) from «dvs; Latin examples are : finio from finis, lento
from lenis; they show -i- in the Perfect Participle Pass., &c., v. g. Gk. d-dipi-
tos, Lat. fini-tus, leni-tus, molli-tus, insigni-tus, vesti-tus, mini-tus, stabili-tus, @-rlidi-
tus ; from sors, O. Lat. sortis, comes sortior ; from pars, stem parti-, comes partior ;
from potis (O. Ind. pati-, ‘lord,’ Gk. mais) comes pitior, ‘to become master of,’
with an Active potio (e. g. Plaut. Rud. g11 piscdtu noué me uberi conpotiuit),
which was used in Oscan as the equivalent of the Latin posswm.(Ose. putiiad
‘ possit,’ putiians ‘ possint ’) (§ 97). U-stem derivatives have -iiyé-, e.g. O. Ind.
fatrii-yé-ti, ‘acts like an enemy,’ from Satru-, ‘an enemy,’ Gk. Saxpiw (é) from
ddxpu (from U-stems, e. g. ioxtw in the Dramatists), Lat. stdtwo from status,
métuo from metus, with -ui- in the Perf. Part. Pass., e.g. Gk. d-5dxpi-ros, Lat.
stati-tus, argi-tus (from the stem argu- of O. Ind. arju-na-, ‘ white,’ Gk. dpyv-
pos), acit-tus (ef. acus, aneedle)'. From the analogy of « number of verbs of
similar meaning, which happened to be formed with one or other of these
types of YO-suffix, a definite meaning came to attach itself in the various lan-
guages to certain suffixes. Thus in Sanscrit the ending -Iyd- came to acquire
a desiderative sense and was used to convey this notion, not merely in deriva-
tives from I-stems, e. g. jantyd-ti, ‘he desires a wife,-from jani-, ‘a wife,’ but
in derivatives from other stems too, e.g. putriyd-ti, ‘he desires a son,’ from
putrd-, ‘a son’ (an O-stem). In Latin, as we have seen, the desiderative
ending is -tirid, e. g. parturio, but it is possible that the ending -i6 conveys this
sense in cdtilio from the O-stem catulus, équio from equus.

A fact of more certainty is that Latin -d for *-ayd was used to give a transi-
tive sense’, e. g. claro, -are, to make clear, from the O-stem clarus, nové, -dre from
novus (and so in other languages, § 29, e.g. 0. H.G. niuwén, though in Greek
the ending -ow ousted -aw from this usage, e.g. vedw, to make new, from vées,
tyidw, to make healthy, from bys), and Latin -co, to give an intransitive
sense, e. g. clareo, -ére, to be clear, from the same stem, claro-, while -io possibly
had attached to it the notion of a state of body or mind, e. g. férdcio, -ire, to be

 

1 These Derivatives in -tyé- must
be distinguished from Presents end-
ing in -nuo, like minuo (O, Ind. mi-
no-mi, from I.-Eur, *mi-neu-mi, § 10),
as well as from a Present like pluo
(O. Lat. plévo, from I.-Eur. *pléwa,
§ 6), and from fluo, fruor, &c., whose

root has a Guttural (ef. fluai, fructus,
§ 39. 3). Suo represents an I.-Eur,
*syi-yo (Goth. siuja, Lett. schuju,
Gk. nagoiw).

2 On First Conjugation Deponents
with intransitive sense, e.g. aemiilari,
Sluctuari (and fluctuare), see §§ 62, 64.
§ 32.) THE VERB. DERIVATIVES. 485

haughty (Cato, &c.), saevio, -ire, to be fierce, insdnio, -ire, to be insane, dentio,
-ire, to be teething, as -aw, -1aw were used in bodily ailments in Greek, e.g.
opOarpiiw, ddSovridw, or desideratively, ©. g. roudw.

The -YO- suffix was theI.-Eur. suffix by which verbs were formed from Nouns
and Adjectives, v.g. O. Ind. apas-yé-ti from apas-, apas- (Lat. opus), Gk. dvo-
paiva for *svopn-yw from dvopa (Lat. nomen’. But in Latin, denominative -yé-
has been to a great extent supplanted by -i-yé- (see below). Thus the Latin
equivalents of the Sanscrit and Greek verbs, just quoted, are dpéruri and nomi-
nare. This process of extending the AYO-suffix at the expense of the YO-suffix
went on as late as the literary period. Many O. Lat. verbs of the fourth Conj.
are in class. Lat. verbs of the first Conj.; e.g. O. Lat. fulgorio (fulgur-to), from
fulgur, used by Naevius (Trag. 13 R.):

suo sénitu claro falgoriuit Iuppiter,

is in class Lat. fulgiiro, -are ; impetrire is the old form of impetrare, consecrated
to religious usage, like porricere the old form of projicere (ch. ix. § 44) ; artire,
Cato, &c. is in class. Lat. artare ; ef. dolitus, Varro, Men. 7 B. for dilatus ; atritior,
Compar. of atritus, Plaut. Poen. 1290 (if this is the right reading) ; gnarivisse
quoted (apparently from Livius Andronicus) by Paul. Fest. 68. 5 Th., for
narrasse (cf. Gl. Philox. gnaritur: yvwplfera; C. G. L.v. 72. 9 gnoritur: cognitum
sive compertum est).

The old formation remains in custddio from custos ; dentio, to grow teeth
(used by Plautus for ‘to suffer through lack of food’ in Mil. 34, where the
parasite apologizes to the audience for his complaisance in listening to the
soldier’s bragging: auribus Peratidienda sunt, ne dentes déntiant), though
dentatus, and not dentitus, is the Participial Adjective. In dentio we have the
same -YO- or -[YO-suffix as in the Verbs indicating disease, state of body, &c.
(see below), like insanio from insanus.

When a Verb is compounded with anything but a Preposition the Com-
pound assumes the form of an A-Derivative, e.g. aedificare from aedes and
facio, sacrificare from sacrum and facio (cf. sacrificus), as in Greek we have -ew of
oixodopéw, &c. Mandédre seems to be a similar formation, as if a Derivative
Verb from *mani-dus, ‘giving into the hand’; and the Derivative Verbs in
-igo, -are, e.g. ndvigo, -are, pir(2)go, -are, ji i)go, -are (on O. Lat. purigo, jurigo,
see ch. iii. § 13) point to *navigus, &c. from navis and ago. Morigérari, to
humour, devote oneself to (also morem gerere), comes from the Adj. morigero-
(Plaut. Amph. 1004 meo me aequomst morigerum patri esse) ; dpitilari from
a stem *opi-tulo-, &. (cf. ch. v. § 80, p. 363). (On the predominance of the A-
type of Verb in Latin, see § 33 a.)

Examples of these endings are : (1) -0, -dre: nimino, -are from nomen ; céloro,
-are from color ; énéro, -are from onus ; scéléro, -are from scelus ; exdmino, -are from
examen ; privo, -are from privus ; ignoro, -are to make unrecognizable, Plaut.
Men. 468; pio, -are from pius ; prébo, -are from probus; sécro, -are from sacer ;
grévo, -are, to make heavy, from gravis ; cicuro, -are from cicur (not t, Rev. Philologie,
Xv. 64); levo, -are, to make light, or to lift, from levis ; péiro, -are, to make equal,
from par, Plaut. Cure. 506 :

eodem hércle uos pono ét paro: parissumi estis hibus.

Ampliare, to adjourn a case, is a rough-and-ready Derivative from amplius, the
judge’s phrase in giving notice of adjournment ; similarly compirendinare, to
remand for two days, from (com)perendinus (dies), Vulg. Lat. *hticare, to call
486 THE: LATIN LANGUAGE. [Chap. VIII.

hither (Fr. hucher), from huc ; some explain nigare as a Derivative of this sort
from nec (cf. Germ. verneinen from nein), or rather from its byform neg-
(ch. x. § 18), the byform being chosen to avoid confusion with nécare, to kill.
Whether the @ so cften seen in Verbs compounded with a Preposition, e.g.
profligare (from fligo), occtipare (from cpio), aspernari (from sperno) is due to the
transitive sense of the Compound, or to derivation from lost Adjective-stems
*profligo-, *occupo-, *asperno-, is not clear. (On amplio see K. Z. xxxiii. 55.)

(2) -io, -tre: raucio, to be hoarse, from raucus ; singultio from singultus (U-stem),
blandior from blandus, largior from largus, prae-sdgio from sdgus (an Adj.
especially found in the Fem, saga, a go-between, e.g. Lucil. vii. 6 M. saga et
bona conciliatrix).

(3) -c0, -@re : ardeo from @ridus [O. Lat. ardus, ve. g. C. 1. L. i. 577. (2). 213 see
ch, iii. § 13], which is the Adjective corresponding to deo, as célidus to cdleo,
nitidus to niteo, &e. (ch. v. § 74) ; audeo, from fridus, the Adj. corresponding to
dveo, had originally the sense of ‘ to be eager,’ ‘to have a mind to,’ e.g. Plaut.
Mil. 232 auden participare me quod commentu’s, whence the colloquial sides
(Terence, &c.), ‘if you please,’ for si audes (Plaut., e.g. Trin. 244 d& mihi
hoc, mél meum, si me amas, si atdes) (ch. iv. § 67’. These two classes of
verbs in -co must be kept distinct, the Derivatives from O-stem Adjectives
like fldveo, ardeo, audeo, and the Intransitives with Nouns in -or and Adjectives
in -idus, e.g. caleo (calor, calidus), areo (aridus), aveo (avidus), niteo \nitor, nitidus)
(see ch. v. §§ 67, 74).

§ 33. Of other Verb-suffixes :—(1) -dh- : on jiibeo with jub-, for *dyu-dh-, lit.
‘to set in motion,’ ‘rouse to action,’ see § 29.

(2) -d- : tondeo seems to be Causative of a lost “tendo (Gk. révdw, to gnaw), for
*tem-do from the root tem-, ‘to cut’ (Gk. réu-vw) ; per-cello for *cel-do (ef. cla-d-es,
Gk. «Aa-8-doa ceitoat, Hesych.) from a root kel- (ef. Gk. dao-«d-d-s) 3 triido
(Goth. us-priuta, ‘I trouble’).

Since -dh- would become d in Latin (ch. iv. § 114), it is impossible to
determine whether the suffix -dh- or -d- appears in cii-do (cf. Lith. kéu-ju),
Fren-do (and frendeo)! for *frem-do (cf. fremo), ten-do from the root ten-, ‘to
stretch’ (ef. Gk. reivw for *revyw), of-fen-do and de-fen-do from the root ghven-
(ef. Gk, Oeivw for *6ev-yw), On pando, mando (third Conj.), vendo, see §§ 14,
95, and on mando ‘first Conj.) § 32, above.

(3) -t-: mé-to (cf. Mid. Ir. methel, ‘a reaper’; so some explained the name
Métellus, Léwe, Prodr. s.v.) cannot be dissociated from Gk. dudw; nor flec-to
from falx. But this formation is not so common in Latin as in Greek, though
all of the numerous Gk. Verbs in -mra, ve. g. rumTw, xadénrw may be formations
with the YO-suffix, for -py- seems to have become -nr- in Greek, *ru7-yw,
*yaden-yw, &c. (ch. iv. § 65). The E-grade of root seems to he used with the
T- as with the D-suffix, so that plecto (from root plek-, Gk. wAéw) is the correct
form, O. Lat. -plocto being due to false Analogy (ef. pondus with o by Analogy of
pondo-, ch. v. § 72, and see above, § 7).

(4) -8+: pri-r-io shows the ending -io of verbs indicating bodily ailments, &c.
(§ 32) attached to a lost *priiro for *preu-so (O. H. G. friusu, Engl. I freeze) from

 

1 The two forms frendo and frendeo Eur. *-dhés) 2 Sg., fren-ditis (I.-Eur.
have been explained by a supposed = *-dhvites) 2 Pl.
original declension like fren-dés (I.-
§ 33.] THE VERB. DERIVATIVES. 487

the root preu- ; with Gk. avgw, a development of *adyw (ef. Lat. aug-eo, © Causa-
tive formation, § 29) with the suffix -s- (ef. I.-Eur. *auges-, *augos-, a neuter noun,
0. Ind. djas, ‘ strength’ ; cf. Lat. augus-tus), we may compare Lat. aux-ilium, and
the Umbrian formula in invocations of deities orer ose, if this means ‘ his (se.
donis) macte,’ ose being Voc. of a stem *aukso- with the sense of Lat. auctus.

If arcesso is connected with accédo as incesso with incédo, the suffix must in
this Verb have Causative force (cf. Lib. Gloss. ‘arcesserat,’ advenire compu-
lerat ; Porphyr. ad Hor. Epp. i. 17. 50 corvus cum accedit ad cibum, strepitu
vocis alias aves arcessit), but the use of 7 for d before ¢ is peculiar (ch. iv.
§ 112). Accerso is a metathesis of arcesso and belongs to colloquial Latin (e. g.
Terence, Petronius), as arcesso to legal phraseology (see A. L.L. viii. 279).
Quaeso, in older spelling quaesso (ch. ii. § 129), from *quai-s-s-o, is in the earlier
writers used along with quaero (older quairo, C.1.L. i. 34, from * quat-s-o),
e.g. Plaut. Bacch. 178:

mirimst me, ut redeam, te dpere tanto quaésere,

Enn. Ann. 143 M. (a description of Ostia) :

Ostia munitast. idem loca nauibus celsis

munda facit nautisque mari quaesentibus uitam,
and Trag. 97 R. liberum quaesendum causa‘; in classical Latin it is found in
1 Sg. quaeso, t Pl. quaesumus. (On the spelling quacsimus in MSS. of Cic., see
Neue, ii?. p. 437. The spelling with u seems to be an affectation due to the
archaic character of the word.)

(5) -ss-. These bear the same relation to the stems just mentioned as Lat.
cmasso, &c. to dixo, capso. They are called ‘ Desiderativa’ by Priscian (i. 431.
18 H.), who explains capesso as ‘desidero capere’ (i. 535. 10 H.). In MSS.
they are often confused with ‘ Inceptive’ forms, e.g. lacescentem for ldécessentem,
capescit for céipessit (for a list of examples, see A. L. L. i. 515).

(6) Verbs in -illo, These are hardly to be separated from A-Derivatives
from Diminutive Nouns in -illd-, Adjectives in -illo-, &., such as scintillare from
scintilla, stillare from. stilla (cf. stiria), tranquillo from tranquillus. They are evi-
dently Diminutive Verbs derived in the same style from Verbs instead of
Nouns. Examples are: conscribillare from conscribo, e.g. Catull. xxv. 11:

ne lianeum lJatusculum mantsque mollicéllas
inusta turpitér tibi flagélla conscribillent ;
Varro, Men. 76 B. itaque eas inceravi et conscribillavi Herculis athlis ; Varro,
Men. 280 B. astrologi non sunt? qui conscribillarunt pingentes caelum ;
occillare From occo, Plaut. Amph. 183; sorbillo from sorbeo; obstringillare from
obstringo, e.g. Enn. Sat. ii. z M.:
. restitint, occurrunt, débstringillant, ébagitant.
(See 4, L. L. iv. 68, 223.)
We have -lo in vapiilo, vertild-bundus, ventilo, ustiilo, &e.
(7) -co in albico, fodico, vellico seems to have the same Diminutive force as the
preceding suffix. As the Adj. suffix -co- is often combined with -lo- in Diminu-

1 quaesere, quaesentibus, quaesendum, vocalic s had become r long before
cannot be the O. Lat. forms of quae- the time of Plautus and Ennius
rere, quaerentibus, quaerendorum, as (ch. iv. § 148).

Valesius of Valerius, &¢., for inter-
488 THE LATIN LANGUAGE. (Chap. VIII.

tives, e. g. puer-cu-lus (ch. v. § 31), 80 we have in Verbs pandictilans, stretching
oneself (in yawning) Plaut. Men. 834 (cf. gesticiilari from gesticulus, Dim. of gestus).

(8) -r0. Considéro can hardly come from sidus, a star, but must with désidero
be an extension of a Verb-stem -sid-. Other examples of this formation are
macéro, réciipéro, toléro.

(9) Other formations. Verbs in -igo, -are, e. g. O. Lat. gnarigavit, used by
Liv. Andr. for narravit (Paul. Fest. 68. 5 Th:), with an Auxiliary dgo giving
a Causative force, have been already mentioned (§ 27), and verbs in -fico, -are
(§ 32), e.g. amplifico, -dve, in which an Auxiliary fécio plays the same part.
Facio does not enter into so close composition with the Verb in béné facio, ave
facio (Luer. vi. 962 sol excoquit et facit are), cdile facio later caléfacio with the same
shortening ‘of a final long vowel as in havi (ch, iii. § 42), consué fucio (Varro,
R.R ii. 9. 13 consue quoque faciunt) (on this -6 sew § 34), compendi facto, to cut
short (Plaut. orationis operam compendiface) ; and we have an Accusative
vase-form as the first element of vénwm-do (vendo), venwm eo (veneo), pessum-do,
pessimm eo (on credo, see § 27; on mando, -dre, § 32; on pando, mando-, -ére, § 14).
Other Latin Verb-suffixes are -ut(t)io of balbitio, to stammer, /riguttio, to
chirp ; -urrio (-ario) of ligurrio, scdturio (on which see above, § 30) ; -cinor, -ari
of patrocinor, léndcinor (cf. patrociniwm, lenocinium), riitidcinor, aliicinor, tuburcinor,
&c. ; -isso, -@re, which is borrowed from Gk. -1(w, e. g. atticisso (drTuxi¢w), graecisso,
patrisso (cf. § 33 a). i

§ 33a. The Conjugations in Romance. In the Romance
languages the Latin Conjugations are much better preserved than
the Latin Declensions (ch. vi. § 1). The first Conjugation is the
prevailing type. Its encroachment on the others even in the Latin
period is shown by Vulg. Lat.1 forms like fiddre (Fr. fier, Span.
fiar), a Derivative from fidus, which supplanted class. /idére, as the
Derivative pectinare supplanted pectere (Caper 93. 8 K. pecto caput
non ‘ pectino,’ et pexum non ‘pectinatum’), as well as from the
fact that Greek loanverbs appear naturally to drift into it,
épavety becoming obsdnare, mpotivew propinare, &c., just as
loanwords in French take -er (Lat. -dre), e.g. trinquer, and in
German, -ieren, e.g. marschieren, amusieren. The freaks of
false Analogy appear in Vulg. Lat. fugire, cupire, &c: (Ital.
fuggire, Fr. fuir, Span. huir; Sard. kubire, Prov. cobir), with
transference to the fourth Conjugation by the analogy of their
1 Sing. Pres. Ind., fugio, cupio; in Vulg. Lat. florire, complire
(Ital. fiorire, Fr. fleurir; Ital. compire; cf. Fr. emplir), with
a similar transference, due to the identity in Vulg. Lat. of -eo
and -io (both pronounced -yo, ch. ii. § 149); and the confusion

 

' The same tendency appears at an _ planted *ctlo (ef. occiilo), and the samo
early period. Célare, a Derivative explanation should perhaps be given
from *céla, concealment, has sup- of sinare (Perf. sonut), &e. (See § 32.)
§§ 38a, 34.1 THE VERB. IMPERFECT. 489

of second and third Conjugation Verbs, which we have already
remarked in the Latin of Plautus in the case of Serveo, sordeo, &e.
(§ 19), is intensified in Romance through the approximation of
the é- and é-sounds (ch. ii. § 141), so that we have Vulg. Lat.
forms like respondére (Ital. rispondere, Fr. répondre), /ondére
(Ital. tondere, Fr. tondre), and (by Analogy of the Perfect
tense saput) sapére (Ital. sapere, Fr. savoir); so Anal. of potui,
posse became Vulg. Lat. potére (Ital. potere, Span. poder), and
by Anal. of volui, velle became volére (Ital. volere, Fr. vouloir).
The Perfects struwi, traxi, prostravi, contrivi have similarly pro-
duced the Vulg. Latin forms ¢ragere (Ital. trarre, Fr. traire),
strugere (Ital. struggere, Fr. dé-truire), prostrare (Span., Port.
prostrar), contrive (Span., Port. eurtir), Verbs of the fourth
Conjugation often show the ‘ Inceptive ’ suffix in their Present,
e.g. Ital. unisco Pres. Ind., unire Inf., on which see § 28. (For
a fuller account of the Romance Conjugations, consult Meyer-
Liibke, Rom, Gram. ii. pp. 137 sqq.).

Of Derivative Verb endings may be noticed (1) -icare, a fre-
quent formation in Vulg. Lat., e.g. *nivicare, to snow (Ital.
nevicare, Fr. neiger), (2) ~idiare (Gk. -iCew), which appears in
Ital. as -eggiare, in Fr. as -oyer, in Span. as -ear, e.g. Ital.
biancheggiare, O. Fr. blanchoyer, Span. blanquear, (3) -aztare,
-entare (cf. Lats praesentare), used for Factitive Verbs, e. g. *expa-
ventare, to terrify (Ital.+ spaventare, Fr. épouvanter, Span.

espantar) (2d. il. pp. 604 sqq.).

§ 84. B. Imperfect. In Slavonic the Imperfect is formed
by appending a Preterite of the Substantive Verb (1 Sg. *jachii
from original *és-o-m, ‘I was’) toa case form (usually called an
Instrumental) of a Verbal Noun. Thus of the verb ‘to see,’ of
which the Inf. is vidé-ti (Lat. vidé-re), the Imperfect (1 Sg.) is
vidé-achti, lit. ‘I was a-seeing’; of the verb ‘to carry,’ Inf.
nes-ti (from the I.-Eur, root nek- of Gk. éveyxeiy, &e.), the
Imperfect (1 Sg.) is nesé-achtii; of glagola-ti, ‘to speak,’ the
Impft. (1 Sg.) is glagola-achii. The same method of forming
an Imperfect was followed in Latin, only the Auxiliary verb
chosen was not I-Eur. es-. Lat. vidé-bam, feré-bam, amad-bam,
are formations consisting of a Verbal Noun-stem (probably in
490 THE LATIN LANGUAGE. [Chap. VIII.

some case form such as the Instrumental) followed by the
Preterite of an Auxiliary verb. The 4 of the Latin Imperfect
is in Umbro-Oscan f (Osc. fu-fans, ‘they were’), and must
represent an I.-Eur. bh or dh (e.g. Lat. #1, Umbr. pufe, Ose.
puf, O. Sl. ktide, O. Ind. kitha from an I.-Eur. DH-suffix)
(ch. iv. § 114). It can hardly be separated from the J of the
Latin Future vidé-bo, ama-bo, &e. (see § 36), which is similarly
in the Italic languages /, e.g. Fal. kare-fo ‘ carebo,’ but whose
equivalent in O, Ir., b, e. g. no charub, ‘ amabo’ [quasi *nu (nunc)
carabo] (ef. Lat. edrus), shows it to represent I.-Eur. bh, not dh
(which would be d in O.-Ixv.), The Auxiliary verb used must
then have been the I.-Eur. root bheu-, whence Lat. fui, O. Lat.
Subj. fuam, &e.; and Lat. -dam, -bas, -bat, 3 Pl. -bant (Ose.
-fans) may represent an J.-Eur. Preterite *bhwim, -as, &e.
(on the loss of postconsonantal w, see ch. iv. § 71), seen in
O. Iv. ba (from *bam), and corresponding to the Lat. Preterite
of the verb es-, evam (from *esam). The Verbal Noun-stem
employed recurs in such formations as the Fut. widé-bo, aré-bo,
sci-Lo, as well as in Verbs compounded with fdcvo, e. g. aré-facio
(§ 33. 9), and in Adverbs compounded with Jicet, e.g. vide-licet,
sci-licet, 1-licet (ch. ix.§ 7) It appears in the Ace. case in O. Ind.
Perfects like vida-cakara, compounded of vidim, the Acc. Sg.
of a Verbal Noun (cf. Lat. vidé-) and cakara, the Perfect of kr-,
‘to make’ (I.-Eur. ker-, Gk. xp-aivw, Lat. creo). Lat. vidé-, ama-,
jini- might be regarded as the bare stem of the Verb without any
Case-suffix, but- this explanation does not suit with legé-lam, for
the Verb-stem would here be legé- (legd-), and the only way of
avoiding this difficulty’ would be to suppose that Verbs of the
third Conjugation followed the Analogy of Verbs of the second
Conjugation, *legé-bam becoming legé-bam after the fashion of
vidé-bam, as in the third Declension of Nouns Consonant-stems
followed the analogy of I-stems in their Nom. Pl., *milit-ds
becoming milit-és like part-és (originally -8yés, ch. vi. § 40).
A similar change of their Imperfect formation was made by
Verbs of the fourth Conjugation in the second century B.c., for

 

' Some prefer to regard the Latin case legzbam will represent *legés-bam
Imperfect as formed of a Verbal Noun (ef. ch. iv. § 152).
in -és with an Auxiliary. In that
§§ 35, 36.] THE VERB. FUTURE. 491

while these have -2bam (or -iébam) in Plautus and the earlier
writers, e. g. audi-bam (and audié-bam), they follow exclusively in
classical Latin the analogy of Verbs like facio in their Imperfect
audié-bam like facié-bam, as in their Future audiam (O. Lat.
audvbo and audiam) like faciam. The Romance languages point
toa Vulgar Latin Imperfect of the second and third Conjugations
in -éam, of the fourth in -7am (e. g. Sard. timia, finia, Span. vendia,
Port. dormia, O. Fr. diseie * disais,’ senteie ‘ sentais ’), while the
first Conj. Impft. had -dbam (e.g. Sard. istava from Lat. stalam),
but whether this may be taken as evidence that byforms *¢imé-am,
*fini-am, &c. formed like er-am, *bhw-am, existed in Latin from
the earliest times is doubtful (see Meyer-Liibke, Rom. Gram.
i, p. 282; and ep. below, § 37).

The Imperfect Subjunctive is formed with an S-suffix which
is usually referred to the Substantive Verb es-, appended as an
auxiliary, though some explain it as a Noun-stem suffix, com-
paring dger-em (on Subjunctive e, see § 55) to agere for *ages-i,
Loc. Sg. of a Neuter S-stem *agos, *aSes-os Gen., ferr-em for
*fers-em to ferre for *fersé, dmarem, mérerem, awlirem to amare,
merere, audire (see § 83).

§ 35. Fourth Conj. Impft. in -ibam. These forms are very common indeed
in the Dramatists, e.g. Plaut. Awl. 178:

praésagibat mi dnimus frustra me ire, quom exibim domo,

and the Republican poets, e.g. Luer. v. 934:
nec scibat ferro molirier arva,

and are often used by the Augustan poets and their followers, where the
ordinary form would not come into a dactylic line, e.g. lenibat in Virg. A.
vi. 468:

lenibat dictis animum lacrimasque ciebat.
From ¢0, the classical Imperfect remained ibam, pér-ibam, vén-tbam, and from
queo, quibam, né-quibam.

Aibam, a dissyllable, is the usual form in Plautus, but we have also aiebam,
ev. g. Rud. 1080 quam esse aiebas (MSS. alebas).

The question whether -ibam ‘or -izbum was the correct ending for these
fourth Conjugation Imperfects was a matter of discussion among Roman
grammarians. We hear of a certain Aufustius who wrote a treatise, dedi-
cated to Asinius Pollio, to show that veniebam and similar forms were preferable
to venibam, &e.

§ 36. C. Future. (1) In-dé. (I.-Kur.*bhwo). This forma-
tion is shared by the Celtic languages [e.g. O. Ir, no charub
492 THE LATIN LANGUAGE. (Chap. VIII.

‘amabo,’ quasi ‘*nu (nunc) carabo’ (cf. Lat. eérws)], and appears
in Faliscan, karefo ‘ carebo,’ pipafo ‘ bibam’ in the inscription on
a drinking bowl: foied vino pipafo, kra karefo ‘hodie vinum
bibam, cras carebo’ (Not. Scav. 1887, p. 262). In class. Lat. it
is confined to the first and second Conjugations, dmdbo, vidé-bo,
but in the early Dramatists the fourth Conjugation Verbs show
-ibo as well as -2am in the Future, and in the plebeian Latin of
Novius’ Ated/anae, or rustic farces, we find virebo, for vivam,
dicebo for dicam (as in the Atellanae of Pomponius paribis for
paries, Com. 20 R.). These last forms are doubtless due to the
influence of the Imperfect in -bam (see the preceding section),
a formation shared by the third Conjugation, as well as the
first, second, and fourth ; vivébam, dicébam called into existence
rivébo, dicébo, on the analogy of ridébam, carébam,which had Future
forms vidébo, carébo. This -b4 of the Future Tense, Falisc. -fo,
O. Ix. -b(6) is clearly some part of the Verb bheu- (Lat. fui, &.),
of which we have seen -4am of the Imperfect Tense to be
a Preterite. The Future of Latin sum, ero, is a Subjunctive
form, *es-6, with Future meaning ; a meaning which seems to
have attached itself to the I.-Eur. Subjunctive (see § 55).

(2) In -am. For verbs of the third and fourth Conjugations
in Latin ther Sg. of the A-Subjunctive (see § 55) is used for
the 1 Sg. Future, though, as we have seen, in the case of
Verb-stems of the fourth Conjugation, this Subjunctive did
not succeed in entirely ousting the formation in -do till the
classical period, e.g. dég-am 1 Sg. Fut. and 1 Sg. Subj.
audi-am 1 Sg. Fut. and 1 Sg. Subj. (but im the older
literature, also audt-lo), For the other Persons of the Future
the E-Subjunctive forms (see § 55) are used, /egés, leget, legémus,
&e., audiés, audiet, audiémus, &e. The reason of this distinction
between the 1 Sg. and the other Persons of the Future Tense
is not clear. It may be that the 1 Sg. Subj. had already for
along time played the part of the 1 Sg. Fut. of these verbs,
as it played the part of the 1 Sg. Imperative at all periods of
Latin, and still retained its place when the new Future forms,
which supplanted an older diao, &e. (see below), were introduced.
The spellings affected by Cato dicae, faciae for dicam, faciam
seem to have had nothing to do with this variation of @ and é
§§ 37, 38.] THE VERB. FUTURE. 493

in the Future of these verbs, but to be merely an attempt to
express by a written symbol the weak sound of final m in Latin
(see ch. ii. § 61)1,

(3) In -so, This formation belongs to O. Lat., e.g. dixo, faxo.
It is discussed in §§ 3 and 5. The similarity in use between dzwo,
7axo, capso on the one hand, and dmasso, &e. on the other, suggests
that diro stands for *dic-sso, faxo for *fac-sso, capso for *cap-sso.
But the Umbro-Oscan Future, e.g. Osc. deivast ‘ jurabit,’
censazet ‘censebunt, Umbr. fust ‘erit, furent ‘erunt’ had only
one s. In Romance the Future has been lost (but O. Fr. ier
from ero), its place being supplied by periphrastic formations
with Aabeo (the usual type), volo, &e. Thus Ital. canterd, Fr.
chanterai represent cantare habeo (reduced to hayo) (see Meyer-
Liibke, Rom. Gram, ii. pp. 138 and 354 sqq.).

§ 37. Fourth Conj. Fut. in -ibo. This formation is extremely common in
the Dramatists, e.g. Plaut. Asin. 28 ut ipse scibo, te faciam ut scias, but was
not adopted (as the Impft. in -ibam was) by the Augustan poets, though
Propertius ventures on lénibunt, iii. 21. 32. In Late Latin poetry it reappears,
e. g. largibor, Juvene. ii. 562. The existence of audiam beside audibo, and of
dicébo beside dicam may have led in Vulgar Latin to the coinage of an Imper-
fect without b, whence Romance Imperfects like Sard. timia, finia (but
istava, Lat. stabum) (see § 34). Red-dibo, an O. Lat. Fut. of reddo, shows the
Fut. of do, débo, e.g. Plaut. Men. 1038 (perhaps reddébo is the spelling indicated
here and elsewhere by the MSS. :

sdluom tibi ita, ut mihi dedisti, réddibo: hic me mane modo.

The Fut. of eo and its Compounds, as of néqueo, has -tbo in class. Latin as
well as in the earlier period ; and it is not till Late Latin that forms like
rediet, transient came into use (eriet in Hor. C. iv. 4. 65: merses profundo,
pulcrior evenit, is the reading of inferior MSS., and ‘like mollibit of ¢. iii. 23.
19 has no probability).

Whether third Conjugation YO-stems took this form of Future (and the
Imperfect in -ibam) is very doubtful. Paribis for pédries, quoted from an
Atellana of Pomponius (Com. 20 R.) may be a vulgarism. Adgrédibor is the
reading of the Palimpsest in Plaut. Pers. 15, as against adgredior (aggredior) of
the other MSS., and seems to be required by the metre; congrediar of all the
MSS. (AP) in Plaut. Most. 783 is changed by editors to congredibor.

§ 38. Third Conj. Fut.in-ébo. Vivebo is quoted by Nonius (509.3 M.) from
Novius (Cum. 10 R.) :
tibi dum uiuebé, fidelis éro,

 

1’ Still all the instances of this spel-  _p. 447; and cf. above, ch. ii. §§ 1, 3,
ling of Cato’s seem to be 1 Sg. Fut. on the Vulg. Lat. change of a to ¢
In MSS. of Plautus we find -em occa- after j), e.g. faciem, accipiem; also
sionally in 1 Sg. Fut. of Verbs in -io simem, Truc. 963 (in the Palatine
(for the references, see Neue, ii MSS.).
494 THE LATIN LANGUAGE. [Chap. VIII.

and dicebo by Nonius (507. 1 M.) from the same play of the same author
(Com. 8 R.) primum quod dicebo. Plautus (Zpid. 188) uses the phrase exsugébo
sanguinem :

iam ego mé conuortam in hiridinem atque eorum éxsugebo sanguinem,
though in another passage he, has the usual exsugam in this phrase, Poen. 614:
idm nunc ego illic égredienti singuinem exsugim procul

(in Rud. 1009 the correct reading is: exurgebo quidquid umoris tibist), (On
Late Latin inferebis. tremebit, see Georges, Lex. Wortf. s. Vv.)

§ 39. D. Perfect. The I.-Eur. Perfect-stem, denoting com-
pleted action, was either (a) Reduplicated, with O-grade of an
E-root in the Singular Active, and elsewhere the weak grade,
é being the usual Reduplication-vowel : e.g. from the root Sen-
(Lat. gi-gu-o) the Perfect-stem was *Ge-Gon-, *Ge-Gn- (Gk.
yé-you-e, yé-ya-ev ; cf. O. Ind. ja-jan-a, ja-ji-ar), or (6) Un-
reduplicated, usually with a high grade of the Stem, and that
either confined to the Singular Active, or extended over all the
declension: e.g. from the root weid-, ‘to see, know’ (Lat.
vid-eo) the Perfect-stem was *woid-, *wid- (Gk. ofd-e, t5-pev,
O. Ind. véd-a, vid-m4) ; from the root séd-,‘ to sit’ (Lat. séd-eo),
the Perfect-stem was séd- (Goth. sét-um, Lith. séd-és, Partic.).
Verbs beginning with vowels, e.g. root ed-, ‘to eat’ (Perfect-
stem éd-, O. Ind. ad-a, Goth. fr-ét), may have been the stepping-
stone between these two formations. Other Unreduplicated
Perfects occur in various languages side by side with Redupli-
cated forms, and show the same vocalism, e.g. O. Ind. sarpa (and
sa-sarpa), from the J.-Eur. root serp-, ‘to creep,’ Hom. 8€x-aTar
(beside 8€-5ex-rar) from dSéxouar, O, Ind. skambh-ur (and c¢a-
skambh-a) from the root ska(m)bh-, ‘to support’ (cf. Lat. seam-
num). In Latin both the Reduplicated and the Unreduplicated
type of Perfect are found, and it is not always easy to decide
whether on the one hand the Reduplication has been lost in
what we may call the ‘ Latin period’ (e. g. ¢#i, which is in the
time of Plautus ¢é-li/-i; of. Gk. ré-rAapev), either by that
Latin habit of discarding one of two neighbouring syllables
with similar form [ar(c7)-ciibii, &e.; see ch. iti. § 13. p. 176|
re-t(e)-tuli (so dé-tondi, O. Lat. de-td-tondi beside to-tondi), ov
by adaptation to other Unreduplicated forms, or whether on the
§ 39.) THE VERB. PERFECT. 495

other the want of Reduplication dates from what we may call
the ‘ Indo-European period.’

The Latin Perfect had usurped the functions of the Aorist or
Preterite, and denoted action in past time, as well as completed
action in present time, e.g, dédi (1) I gave, (2) I have given.
It thus drew within its sphere Aorist or Preterite formations,
such as what is called the S-aorist, e.g. Gk. &de€-a, O. Ind.
a-dikg-am, and gave to these its own peculiar person-endings
(1 Sg. -7 from L-Eur. -ai or -ai, O. Ind. -é, the I.-Eur. Middle
t Sg. suffix, so that Lat. ded-i corresponds to O. Ind. da-d-é,
a Middle, not an Active; 2 Sg. -is/i, and so on; see §§ 65 sqq.).
Thus Lat. dix-i from the S-Aorist stem *deik-s- (Gk. é-deré-a) is
in Latin called a Perfect, not an Aorist, and is declined like
a true Perfect, dix-1 (O. Lat. deixei, *deik-s-ai), 3 Sg. diwit
(O. Lat. -i, -eit) like mémint, 3 Sg. meminit (O. Lat. -a, -cit).
Strong Aorists, both Reduplicated (e.g. Gk. é-7e-vo-v from
gh*en-, ‘to kill’) and Unreduplicated (e.g. Gk. é-yevd-ynv from
gen-), may have been absorbed into the Latin Perfect with the
others, but decisive evidence that will prove their existence
separate from true Perfects (Reduplicated and Unreduplicated)
is hard to find. Thus Lat. fidi of dif-fid-i from fi-n-d-o, root
bheid-, may perfectly well be a true Perfect, with Reduplication
lost in the ‘ Latin period, the weak grade of the stem being
appropriate to the whole of the I.-Eur Middle (e.g. O. Ind.
bi-bhid-é, 1 Sg.) as well as to the Plural (and Dual) Persons of
the Active Perfect (e.g. O. Ind. bi-bhid-ur, 3 Pl.). It may
however be claimed for the Strong Aorist class, like O. Ind.
4-bhida-m, Aor.; and similarly sczd-2 has been called a Strong
Aorist (O. Ind. 4-chida-m) with the Latin Perfect-ending, while
set-cid-i is the true Perfect-form ; even sci-cid-i might be itself
ascribed to a Reduplicated Aorist-class, as ¢éigi has been
identified with Gk. re-ray-év, pépiga with Gk. we-mayo-(nv. It
seems best to class these Latin forms merely as ‘ Reduplicated ’
and ‘ Unreduplicated,’ without the more definite title of ‘ Re-
duplicated Perfect,’ ‘Unreduplicated Perfect.’ To attempt
a further subdivision into Reduplicated Aorist (or Preterite),
Unreduplicated Aorist (or Preterite) would be to go beyond the
evidence at our disposal.
496 THE LATIN LANGUAGE. (Chap. VIII.

There is still another formation which in Latin has the name
and the person-endings of the Perfect-stem, viz. the formation
adopted for Vowel-verbs (first, second, fourth Conjugations),
which appends v (after a Consonant w), e.g. dmd-v-1, re-plé-v-i,
audi-v-i, mon-u-i. This formation has not yet been satisfactorily
explained (see below). The disappears in audit (pronounced
audit in ordinary speech, audiit in Poetry), audit, abét, as in
ditior from divttior, O. Lat. dinus from divimus (ch. i. § 53), sis
for st vis.

Examples of the various types of Latin Perfect are :

(1) Reduplicated: dé-d-i (O. Ind. di-d-é, Perf. Mid.; cf.
O. Ind. 4-da-da-m, Impf.), Osc. de-d-ed ‘dedit,’? Umbr. dede (on
the Umbro-Oscan 3 Sg. Person-ending, see § 69), cré-di-d-i
(O. Ind. dii-dh-é, Perf. Mid. On crédo, O. Ind. Sréd-dhi.,
‘T believe,’ lit. ‘put the heart to,’ O. Ir. cretim, see § 27). The
I.-Eur. Reduplication Vowel @ (in O. Ind. the vowel of the -
stem is often substituted) remains in class. Latin in forms like
mé-min-i (older *me-men-2), pé-pér-i, but came to be assimilated
in Latin fashion (see ch. iv. § 163) to the vowel of the second
syllable in m6é-mord-i, ti-tid-i (cf. O. Ind. tu-tud-é), &e., although
in O. Lat. to the end of the Republic forms like mé-mord-i were
in use.

What the Latin Stem-vowel originally was is not always to
be discovered, owing to the Latin habit of weakening every
vowel in a short second syllable to 7 (2) (ch. ui. § 18). Zu-ttid-i,
which has retained the u-vowel unweakened (ch. iii. § 28),
shows the weak stem tiid- of the I-Eur. Middle and Plural
Active (O. Ind. tu-tiid-e 1 Sg. Mid., tt-tiid-ar 3 Pl. Act.),
while con-dad-it, quoted by Priscian (i. p. 518. 13 H.) along with
con-téd-it from the Annads of Ennius (Il. 515 and 418 M.), may
be the trace of a Latin *ti-tad-, *te-taud- (O. Ind. ti-tod-a
1 Se. Act., Goth. stai-staut 1 Sg. Act., stai-staut-un 3 PI.
Act., apparently from a root (s)taud-), like ce-ca/-i from *ce-cairl-,
root (s)kaidh-. But to-tond-i, spo-(s)pond-i (O. Latin spe-pond-i),
from the roots tend- (Gk. révdw), spend- (Gk. oréviw), are not
necessarily derived from that form of the Reduplicated Perfect-
stem which was reserved for the Singular Active in I.-Eur., the
form namely with O-grade of an E-stem vowel, *te-tond-, &c.,
§ 39.] THE VERB. PERFECT. 497

for their 0 may be due to the Causative Present-stems (§ 29)
with which they were associated, tondeo (I.-Eur. *tondéyo-),
spondeo (I.-Eur. *spondéyo-). The or of mo-mord-i (O. Lat.
me-mord-2) is most naturally explained as the Latin equivalent
of I.-Eur. r, so that me-mord- is the weak stem, *me-mrd-, from
root (s)merd-, and me-mord-i is exactly O. Ind. ma-mrd-é 1 Sg.
Mid.; similarly the «7 (older of) of teti#/i will be I.-Eur. | (1)
(cf. Lat. te-¢ul-i-mus, Gk. ré-rh-a-pev) from the root tel-.

(2) Unreduplicated: «id-i, I saw, have seen, if vid- repre-
sents I.-Eur. *woid- (ch. iv. § 10, p. 228), exactly corresponds to
O. SI. véd-8, ‘I know, a Middle form, and perhaps the only
trace of the Perfect Tense to be found in the Balto-Slavic family
of languages, with the exception of the Participle, while in Latin,
Celtic, and Teutonic it is the Perfect Participle which has been
lost, and the other parts of the Tense retained ; the Active form,
I.-Eur. *woid-i, appears in O. Ind. véd-a, ‘I know,’ Gk. ofd-a,
Goth. vait ; séd-i, unless it stands for *sé-sd-i, as sido for *si-sd-o
(§ 9)1, shows the I.-Eur. Perfect-stem séd- of Goth. sét-um 1 Pl.,
Lith. séd-és Partic. ; éd-2 shows I.-Eur, éd- of O. Ind. ad-a Act.,
Goth. fr-ét ; @i, the Perfect of ago, does not show I.-Eur. a&-
(O. Ind. aj-a, Gk. jjy-war, O. Scand. ok ; cf. Lat. amb-dges), but
adopts a Latin raising of & to é, seen in other verbs whose root-
vowel is a, e.g. pégi from root pag- (Lat. pango) [cf. ch. iv.
§ 51, where it is shown that I.-Eur. d, Lat. @, is the weak
grade of JI.-Eur. 6, Lat. é, e.g. I-Eur. *sdto- from root sé-,
Lat. satus beside sé-men, fac-io beside féc-i (Gk. €0yxa) from
root dhé-].

A similar absence of Reduplication appears in Umbro-Oscan,
e. g. Osc. dicust ‘ dixerit’ (beside Umbr. dersicust), Umbr. fakust
‘fecerit’ (beside Osc. fefacust), Umbr. vurtus ‘ verterit’ and
ku-vurtus ‘ converteris.’

(3) S-Aorist (Preterite): diw-1, older detw-ci (Gk. edeg-a;
ef. O, Ind. 4-dikSa-m) from the root deik-, a Reduplicated Per-
fect (Aorist ?) of which appears in Umbr. de-rsic-ust ‘ dixerit’
(Umbr. 7s expresses an intervocalic d-sound, ch. ii. § 88), and
apparently an Unreduplicated in Ose. dic-ust ‘ dixerit’ ; wssi (Gk.

 

1 Umbr. pru-sik-urent '‘ pronuntiaverint’ has certainly I.-Eur. 6 (Umbr. i),

Kk
498 THE LATIN LANGUAGE. [Chap. VIII.

eica for *e¥ooa) has a short vowel according to Priscian, 1. 466.
7 H., like %s-tvs which has the weak grade tis- of the root eus-
(cf. Late Lat. ostile for iistile, &e., A.D. L.ii.607). The S-Aorist
forms show sometimes the weak stem, sometimes the E-grade in
I.-Eur. languages, e.g. O. Ind. 4-dikgam just mentioned, Gk.
é-pe€-a, sometimes a still higher grade; e.g. from the root
leiq?-, O. Ind. 4-raikSam Act. [Gk. é-Aeww-a may represent
-Iéig®- or -léiq®- (ch. iv. § 45), O. Ind. 4-rik§-i Mid. has
the weak grade]; and the same variety appears in| Latin
S-Perfects, e.g. di-vido, di-visi, but tro, asst. But it is not
always possible to ascertain the quantity of the vowel in Latin,
for the grammarians of the Empire are uncertain guides about
the natural quantity of any vowels long by position, for which
they could get no clue from the classical poets. One requires
further evidence before one can fully believe Priscian (sixth cent.)
(i. p.466.17 H.), when he posits a naturally long penult for all Per-
fects in -a which have the vowel e before this ending, e. g. i¢-Zeai
from /écio, and for no others, e.g. dux-i from dico (so perdiictus,
Audacis exe. 359. 15 K., but deddaerunt in the Mon. Ane. ili. 26).
The use of the apex, or accent-mark, to indicate a long vowel, on
inscriptions, especially on inscriptions later than 150 A. D., is also
evidence of a more or less doubtful character ; and even when the
length of a vowel seems fairly established, e.g. réxi (with apex
over the ¢ on an inser. of 105 a.v., C. 1. L. v. 875; also declared
to have long e by Priscian, /.c.; contrast Gk. &-pefa), there
remains a further question whether the long vowel is not due to
a similar phonetic law for the group g-s, as that which gramma-
rians mention for the group g-¢ (in rectus, lectus, tectus, &e.), viz.
that a naturally short vowel is lengthened before this Consonant-
group. (On the question of the quantity of the Stem-vowel in
these S-Perfects, see ch. 1. § 144, and cf. ch. iv. § 51, p. 254,
where the high grade of the root is accounted for by the syncope
of a short vowel, réx- for rég-(8)s-.) The vocalism of the Perfect
of y#eo may have been altered before the end of the Republic,
for the spelling jows- on old inscriptions (see C. I. L. i. Index,
s.v.) proves an original diphthong ow (probably I.-Eur. eu;
the root is yeudh-, O, Ind. yodhati), and O. Lat. joussei (with s
for ss before the practice of writing a double consonant came in,
§ 39.] THE VERB. PERFECT. 499

ch. i. § 8) would become naturally in classical Latin jasi, as in
the Perfect of mitto O. Lat. meissei became class. Lat. misi (see
ch. ii, § 129). Cédo has céssi (Prise. i. 466. 6 H.). Perfects
in -ssi often come from Verb-stems ending in -s, e.g. from
the root eus-, us-si; thus prémo probably takes its Perfect
nres-si from a lost Present *preso [ef. Gk. rpé(o)w and rpéuw ;
Lat, ¢er(s)-eo and trém-o, § 3]. Fluxi comes from the stem
*bhlug*- (Gk. oivd-pav€), the guttural becoming v in fluvo, class.
Jluo ; co-nivi from the root kneigh®- (cf. nicto) ; fai from the stem
dheig*-, the O. Lat. Present being fivo (§ 7); veai from the
root wegh-, Pres, veho, and so on.

The S-Aorist is not found with Vowel Verbs in Latin as it is
in Greek, e. g. é-rijun-oa (Dor. -doa), é-pidn-oa (with o apparently
for oo as in etoa for *eto-ca, quoted above); at least it is not
found in the Indicative. But in Umbro-Oscan, if tt in these
languages represents an original ss (as Att. mparrw for apdcow),
we have this formation in Perfects like Pelignian coisatens
‘curaverunt’ quasi *curassunt, Osc. duunated ‘donavit’ quasi
*donassit, dadikatted ‘dedicavit,’ prafatted, prafattens ‘ probavit,
probaverunt, teremnattens ‘terminaverunt’ [all the examples
preserved belong to the first Conj., and so do almost all of the
Latin examples of -ss- forms, e. g. amasso, amassim, amassere (see
§ 5)].

(4) With v (w). This is the Perfect-stem formation of Vowel-
Verbs, as of the first Conjugation, dmdvi, nécd-vi, and with the
v (uw) added to a stem not ending in -d, méc-ui (so erépavi and
créput, cibavi and ctbui, dimicavi and dimicui, &c.), of the second
replévi, monui, of the third strd-vi, sé-vi, cré-vi, si-vi, of the fourth
audwi, déstlii, and déstlivi. It isalso found with some Consonant
Verb-stems, e.g. cdlui, dlui, gémui, and is sometimes added to
a Perfect-stem already formed; e.g. mess-ui (but see § 51).
When the stem ends in v (z), two v’s are not written, but the
preceding vowel is lengthened, e. g. cdvi, Javi (see § 47 on fort).
Statui, fui, &e. were in O. Lat. statu (stativi), fuvi. Pdsut is
a form introduced by the false apprehension of pd-stéus [from
the Preposition pd- (Gk. and, Lat. ad, ch. ix. § 12) and situs
P,P. P. of sino (pono for pd-s(z)no, ch. iv. § 151)], as if it were
posi-tus like mdni-tus. The true Perfect, used by the older

Kk2
500 THE LATIN LANGUAGE. (Chap. VIII.

writers, is yo-sivi, but this could not become posui (imposti, Lueil.
xxviii. 26 M.) by ordinary phonetic development.

Both the Perfect with v and the S-Perfect occasionally present
shortened forms, e.g. amasti, misti, which are best regarded as
phonetic developments of the full forms dmdavisti, misisti, due in
the one case to the Roman tendency to drop intervocalic v (see
ch. ii. § 53), and in the other to the practice of discarding one
of two neighbouring similar syllables (see ch. iii. § 13, p. 176).
The same shortening occurs in various parts of the Perfect-stem
conjugation, e.g. amdro Fut. Pft., amassem Pluperf. Subj., and
so on (see $ 3).

The Perfect-stem was formed in I.-Eur. from the root of the
verb, not from the Present-stem. This is the reason why a
Latin Perfect often presents so different an appearance from
a Latin Present. Thus sé-v is formed from the root sé- (cf.
sé-men), but séro is a Reduplicated Present, for *si-s-6 (cf. Gk.
tne for *ou-on-pe, § 9); eré-vi, I determined, si-v2, stra-ve and
others exhibit Nasalized Present-stems, cer-no, si-no, ster-no
(see § 10); cré-vi, I grew, 2d-vi, dbdlé-vi and abol-ui have as
Presents ‘Inceptive’-stems ecre-sco, no-sco, abole-sco (see § 28).
But it often happened that the Perfect was influenced by the
form of the Present-stem. Thus véai from root g"em- (Goth.
qgémum 1 PI.) should be *vémi, but takes its 2 from the Present,
a YO-stem, which by the Latin Phonetic law changes its m to
n before i (y), vénio.

In Oscan the Perfect Subjunctive differs from the Perf. Ind.
only in its use of the Subjunctive vowel é (§ 55), e. g. Ose.
fefacid ‘ fecerit. But in Latin the Perfect Subjunctive adds to
the Perfect-stem -érim, -éris, -érit, &e. (3 Pl. -erit), which is
generally regarded as an Optative form from the root es- (cf.
sim) used as an Auxiliary. Others explain vidérim, with vidéro,
&e., as containing a Noun-stem vidis- (cf. cints-), which is seen in
its proper form in vidis-sem (see § 52). In the Perfect Sub-
junctive person-endings 2, not 7, is correct; scansions with 7 are
due to confusion with the Fut. Perf. (see Neue, ii”, p. 510). In
its Potential use, e.g. dixerim, I would say, affirmaverim, &c. the
first Pers. Plur. is rare (e. g. dimerimus, Cic. Tuse. iti. 4.7; Nat.
Deor.i. 20. 52), and in the Deponent Conjugation this use is rare
§§ 40, 41.] THE VERB, PERFECT. 5Or

even in the Singular (e.g. passus sim, Ter. Andr. 20 3; cf. sit passus,
Virg. G. iti. 141) (see 4. Z. L. i. 347).

§ 40. Other examples of Reduplicated forms : sté-i-i (O. Ind. ta-sth-8),
ste-ti-mus, older *ste-t’-mos (Gk. €-ora-pev) ; pe-ped-i from pedo for *pezdo (0. Engl.
fist) ; ce-cin-d, older ¥*ce-cen-ei (O. In. ce-chan) ; but with the Reduplication-
vowel changed, di-dic-i, Pres. disco for *dic-sco ; po-posc-7, O. Lat. pe-pose-i for
*pe-prk (root prek-) with ‘Inceptive’-stem suffix -sk- (§ 22) (ef. O. Ind. pa-
prach-a, Act.), Umbr. pe-purk-urent ‘rogaverint.’ Peculiar to O. Lat. are:
te-tin-1 (O. Ind, ta-tan-é from tan-, I.-Eur. ten-,‘to stretch’), which was used asa
Perfect of ténco (apud veteres ‘tetini’ dicitur Diom. 372. 18 K.) ; Nonius (178.
7 M.) quotes tetinisse and tetinerim from Pacuvius, tetinerit from Accius; in
Plaut. Amph. 926 we should probably read abstinei for abstines of the MSS. :

nune quando factis me inpudicis dbstinei,
ab inpudicis dictis auorti uolo,

and tetini, -tini should perhaps be read for tenwi, -tinwi in other passages of
Plautus (Studem, Stud. ii. r2e7.) ; Paul. Fest. (335. 7 Th.) has preserved for
us an old augural phrase, discussed by Messala in his Treatise on Augury,
purime tetinero ‘purissime tenuero’ ; scicidi (sciscidi?), illustrated by Priscian
Gi. 517. 3 H.) with several passages from the Dramatists.

§ 41. Unreduplicated. The following, with short vowel, which are often
referred to I.-Eur. strong (unreduplicated) Aorists, are more probably redu-
plicated forms which have lost their Reduplication, first in Compounds, then
in the Simple Verb : tii (see below) ; sci-di (ef. O. Ind. ci-chid-é Perf. Mid.,
a-chida-m Aor.), in O, Lat. sci-cid-1 (see above), cf. ab-scidi, di-scidi, re-scidi, &e. ;
Jidi (cf. O. Ind. bi-bhid-é Perf. Mid., é-bhida-m Aor.), rare in the Simple Verb,
but more frequent in the Compound dif-jidi (e.g. Virg. A. ix. 588 tempora
plumbo diffidit ; Hor. C. ili. 16. 13 diffidit urbium Portas vir Macedo) ; -ciiti
only occurs in the Compound per-culi, just as -piili (in the Simple Verb pépiii)
is only found in Compounds, dispuli, impuli, &e. (cf. rep(e)puli).

And a large number of the forms usually quoted as Unreduplicated Perfects
with stem-vowel unchanged rest on very insecure foundation : -cendi occurs
only in Compounds uac-cendi, in-cendi, &e. ; the same is true of -fendi of de-fendi,
of-fendi, and -hendi of pre-hendi, nor can we be certain whether the e¢ in these
Perfects was short or long ; */ambi is attested by Priscian (i. 506. 25 H.) with
the single example of a line of Lucilius (xiii. 11 M.):

iucundasque puer qui lamberat ore placentas,
where lamberat is evidently Pres. of lambero (Plaut. Pseud. 743 :
etigepae : lepidé, Charine, med me ludo lamberas),

and is probably the very word which is quoted by Paul. Fest. 84. 30 Th.
‘lamberat’ scindit ac laniat ; for *psalli Priscian quotes only a line of Caesius
Bassus, the friend of Persius, with psallerat ; of mando he says (i. 419. 13 H.)
ejus praeteritum perfectum quidem alii mandui, alii mandidi esse voluerunt ;
Livius tamen in Odyssia (a later hexameter version) :

cum socios nostros mandisset impius Cyclops;
502 THE LATIN LANGUAGE. [Chap. VIII.

-vérri (cf, Prise. i, 532. 22 H.) is easily attested for Compounds, but not for the
Simple Verb, and the same is true of -c#di (Prise. i. 515. 16 H.), and to a great
extent of velli (Virg. Ecl. vi. 4 vellit et admonuit).

Of the remaining examples of Unreduplicated Perfect with Stem-vowel
unchanged, ici (Pft. of ico? see Wharton, Etyma Lat. s. y.), stridi, vist (an S-
Aorist? ; the usual Perfect of these Present S-stems is in -ii, e. g arcess-ti,.
quaes-wi, from arcesso, quaes(s)o, § 3), and better attested than these, pdndi,
prandi, sctindi, vérti (on sidi, bibi, see below) ; the Verbs with Present in -ndo
are sometimes credited with a Perfect in -didi instead of -di (ef. condédt, but in
composition abscondi, though abscondidi is quoted from the Republican Dra-
matists by Nonius, 75. 22 M.), v. g. descendidé, quoted from Valerius Antias and
Laberius by Gell. vi. 9. 17, mandidi, a byform of mandi (Prise. i. 419.13 HL,
just cited), prandidi, censured by Diomedes, 367.17 K. This -didi is proper to
verbs compounded with do, (1) to give, (2) to put, so is applied by false analogy
to de-scendo, from the I.-Eur. root skand- (O. Ind. skand-, Gk. oxavddAnOpov,
O. Iv. ro-seseaind) ; the formation of pdndo (cf. Ose. patensins ‘aperirent’ (?)),
mando, prandeo is not clear (see § 14). Verti, O. Lat. vorti (vo- became ve- in the
course of the second cent. 8. c., ch. iv. § 10), has probably the weak stem *wrt-,
with which we may compare either the O. Ind. Reduplicated Perf. Mid. va-
vité or the O. Ind. Unreduplicated Aorist d-vrta-m) ; but while the Umbrian
forms ku-vurtus, co-vortus, covrtust (or cowrtust, for covurtust?), ‘ converteris’
“eonverterit,’ ku-vertu, co-vertu ‘convertito, point to a different stem for the
Present (wert- ; cf. O. Ind. virtaté) and the Perfect (wrt-), the O. Lat. spelling
shows vort-(wrt-) both in the Present and in the Perfect, as well as in the Per-
fect Participle Passive vorsus (*wrt-to-, 0. Ind. vrtta-), so that the weak stem has
the appearance of having forced its way in Latin into the Present Tense-
system, leaving an identity of Stem-vowel between the Present and the Per-
fect. The spelling vorio may, however, have represented the pronunciation
verto (see ch. iv. § 10).

The alternation of @ in Present- with @ in Perfect-stem is seen in ftic-io
(Umbr. fagia, Ose. fakiiad ‘faciat’) and fec-i (I.-Eur. dhé-k- of Gk. &-Onx-a),
Sréngo (root bhreg-, Goth. brikan ; see ch. iv. § 51), and /rég-i, jacio and ject,
ctpio and cépi, pango and pégi, digo and égi, cofpi (Lucr.) and apere (ch. ii. § 150), it
is seen also in Oscan hipid ‘habuerit’ (Osc. i is I.-Eur. é), though *hébi is
unknown in Latin; that of 6 with 6 in féd-io and féd-i [the I.-Eur. root is
bhedh-, Lith. bedu, ‘I dig,’ W. bedd, ‘a grave,’ varying with bhodh-, Gk.
Ad6pos (with @ for 7), Lith. badaii, ‘I prick,’ O. Sl. boda] ; of d with @ there is
one doubtful example, sccibo (with a in sctiberet, Hor. S. i. 10. 71) and scdberat,
quoted as a Plupft. (with lamberat, on which see above) by Priscian from Lucil.
ix. 77 M. scaberat ut porcus contritis arbore costis ; that of ¢ with 2 is seen in
édo and édi, émo and émi, légo and legi, sédeo and sédi, vénio (en for ¢m, m, like Gk.
Baivw) and véni for *vémi (Goth. qgémum, 1 Pl.); the weak stem with short
vowel is seen in the Present, but not in the Perfect, in fiigio and fugi (root
bheug-), linguo and liqui (root leiq’-), riimpo and ripi (root reup-), video and
vidi (root weid-), vinco and vici (root weik-), frindo and fidi.

§ 42. Form of Reduplication. When the root of the Verb began with
a group of consonants the practice in I.-Eur. languages is to use only the first
consonant of the group in the Reduplication-syllable, e.g. I.-Eur. *ke-klow-,
*ke-klu-, the Perfect-stems from the root kleu-, ‘to hear’ (Gk. «é-«Aure ; ef.
§§ 42-44.] THE VERB. PERFECT. 503

O. Ind. bu-Srdva, 0. Ir. ru chuala from *cu-clowa, both of these last having
the weak stem vowel ii instead of the Reduplication-vowel é). This form of
Reduplication appears in Latin Reduplicated Presents (see § 9) like si-s-to
(L-Eur. *si-st-, Zend hi-Staiti, Gk. torn, Umbr. se-stu). But in Latin
Perfects the whole group appears in the Reduplication-syllable when
the group consists of s followed by a mute, while in the stem-syllable
the s is dropped (possibly sometimes retained, as in Teutonic), whereas
in other I.-Eur, languages (e. g. 0.-Ind., Gk.) only the mute appears in the
Reduplication-syllable, e.g. Lat. sps-pond-i (spo-spondi is oceasionally found in
MSS.), sci-cid-i (v. 1. sciscidi), Goth. stai-staut, O. Ind. ta-sthaad (from O. Ind.
stha-, ‘to stand’). Bibi shows the i of bi-b-o, the Reduplication-vowel of
Present-stems, as 8 of Perfect-stems. Similarly di-dic-i may owe its % to the
fact that disco is a Reduplicated Present for *di-de-sco (ef. Gk. 8:-3d0nw for *5:-
Sax-cxw) (but see § 22). The Roman grammarians were in doubt whether the
correct Perfect of sisto was sti-t-i or st8-t-i (see Georges, x. v.), Gellius (ii. 14)
speaks of an old MS. of Cato’s speeches which had the reading vadimonium
stitisses, 1 reading changed by ‘ emendatores ’ to vadimonium stetisses; the same
doubt existed whether stdi or sedi was the Perfect of sido (*si-sd-o).

§ 43. Assimilation of Reduplication-vowel to Stem-vowel. Aulus
Gellius, in a chapter dealing with peculiar forms of the Latin Perfect (M.A,
vi.g; cf. Nonius 140. 19 M.), says that although poposct, momordi, pupugi, cucurri
were the forms used in his time by almost all educated men (omnes ferme
doctiores), the older writers used ¢ in the Reduplication-syllable. He quotes
memordi from Ennius (from his poem to Scipio, and referring to the rivals of
his hero ?) (Sat. 20 M.):

meum non est, ut (v. 1. at) si mé canis memorderit,

from Laberius, Nigidius, Atta, and Plautus (Awl. fr. 2 ut admemordit homi-
nem), remarking that the last author used also prae-morsisset (the S-Aorist
form), peposci from Valerius Antias, pepugero from Atta, occecurrit from Aelius
Tubero, speponderant from Valerius Antias. He even assigns similar forms to
Cicero and Caesar (sic M. Tullius et C. Caesar mordeo ‘memordi,’ pungo
‘pepugi,’ spondeo ‘spepondi’ dixerunt), although our MSS. of these authors,
as of Plautus and the older writers, hardly preserve a trace of them [see
Neue, ii?. 465. In Plaut. Poen. 1074 one of the Palatine MSS. has memordit
(D), another me mordit (C), another momordit (B)].

§ 44. Loss of Reduplication. This assimilation may have in some cases
facilitated the loss of the Reduplication-syllable in Compounds, which was
ascribed above (§ 39) to the Latin practice of discarding one of two neigh-
bouring and similar syllables [e. g. ar(ct)-ciibii, ch. iii. § 13. p.176]. This
practice of Latin [ef. Gk. du(t)-popeds), along with the liability of every short
second syllable to Syncope under the older law of Accentuation (ch. iii. § 13),
must have operated most powerfully to the detriment of the Reduplication-
syllable in Compounds, so that it is wrong to refer all Latin Unreduplicated
Perfects to I.-Eur. Unreduplicated Perfect and Aorist forms. A Perfect like
tii, which has in Plautus the form téili, in Terence usually the Unredupli-
cated form, tli, as always in classical Latin, is most naturally explained as a
Reduplicated form which lost its Reduplication in the second cent. B. c. The
Compound rettulé (not retuli, see Georges, Lex, Wortf.s.v.) has a double ¢, which
504 THE LATIN LANGUAGE. (Chap. VIII.

is clearly not due to the Preposition having had the form red-, but like the
double p of reppéri (beside réptrio), reppiili (beside répello), to the Syncope of the
Reduplication-vowel, re-t(e)-tuli, like re-p(e)-peri, re-p(e)-puli. Other compounds,
like con-tuli, ob-tuli, at-tuli, sus-tuli have precisely the form which they would have
had if they, like rettuli, had suffered Syncope of their second syllable ; only,
while Syncope has left a trace of its operation in the double ¢ of rettuli, this
trace has been obliterated by the phonetic conditions of these other Com-
pounds. *Conttuli, *obttuld, *atttuli, *susttuli are impossible forms; they must in
Latin orthography be written contuli, obtuli, attuli, sustuli, so as to afford no
criterion of whether the Preposition was originally compounded with tetuli or
tuli. The one Compound however which does afford such a criterion pro-
nounces for the first of these alternatives, and warrants us in asserting that
an original tetuli has been reduced to tuli in the Perfect of all the Compounds
of fero (tollo), and in denying the theory that they show an original tuli, an
Aorist formation like Subj. attulas, &. (see § 4). The shortened form tuli
having thus established itself in the numerous Compounds of féo, words in
constant use in the speech of every-day life, it may well have spread from
them to the simple verb. The habitual use of con-tuli, at-tuli, ob-tali, &e. would
naturally lead to the preference of tuli for the more cumbersome tetuli, although
by a freak of language, by the caprice of that ‘ usus,
quem penes arbitrium est et jus et norma loquendi,’

the shortened form -puli of the Compounds re-p(e)puli, com-(pe)puli, im-(pe)pult,
&e, did not supplant the full form pepuli at the same time. It is the Compound
Verbs which as a rule in Latin show the short form, the Simple Verbs which
are most retentive of the full Reduplicated form; e.g. dé-tondi (detotondi is
quoted from Varro by Priscian, i. 482. 7 H., but is declared impermissible by
Servius, who lays down the law that no Compound can have a Reduplicated
Perfect : ‘verba quae in praeterito perfecto primam syllabam geminant, cum
composita fuerint, geminare non possunt), atlondi (in Plaut. &c.), despondi
(Plaut., Ter., Cic.), praecidi (Plaut., Cic.) beside totondi, spopondi, cectdi; and this
fact is additional evidence of Latin Unreduplicated forms having lost the
Reduplication-syllable in Composition.

§ 45. Co-existent Reduplicated and Unreduplicated forms. A more certain
example of an Unreduplicated form existing side by side with a Reduplicated,
and not a phonetic development from the latter due to the loss of the
Reduplication-syllable, is the Perfect of pango, pégi beside pépigi. Both forms
survived in classical Latin, the Reduplicated pepigi being reserved for the
sense ‘T have agreed,’ ‘ stipulated,’ so that it is often called by the Roman
grammarians the Perfect of péciscor; and in addition a third Perfect is men-
tioned by the grammarians for the Simple Verb (not the Compounds), an
S-Aorist form panzi, e.g. in the elegiac epitaph of Ennius (ap. Cie. Tuse. i. 15.34) :

hic uestrum panxit maxima facta patrum.

In the same way O. Lat. fefuked of the very ancient inscription on a Praenes-
tine brooch (C. I. L. xiv. 4123): Manios med fefaked Numasioi, seems to be
from a Perfect-stem *fe-fic- (and similarly Ose. fefacust ‘ fecerit’ Fut. Pft., and
Jefacia ‘fecerit’ Pft. Subj., will have 4 not 4), anterior to the weakening of
unaccented vowels (see ch. iii. § 18). It would have been in class. Latin, had
the form survived, *jéfici, related to feci as pépigi to pogi.

§ 46. S-Preterite. Examples of parallel Latin and Greek forms are dizi.
§§ 45-47.] THE VERB. PERFECT. 505

O. Lat. dei- (Gk. é5e¢a) (both Lat. dei- and Gk. 3e- may before a consonant repre-
sent I.-Eur. déi-, ch, iv. § 45), réxi (Gk. dpega), texi (Gk. Zorega), clepsi (Gk. ZuAepa),
mansi (Gk. gueva for *épevoa), tissi (Gk. eboa for *ebaca), pexi (Gk. évegaunv). The
substitution of -si for -xi (of sanai, tinxi, &c.) in mulsi from (1) malceo, (2) mulgeo,
Julsi from (1) fuleto, (2) fulgeo, parsi (and peperci) from parco, sparsi from spargo, &c.
isdue to the preceding consonants J, r (see ch. iv.§ 157). The same Verb may
use as its Perfect both the S-formation and the Reduplicated Perfect : thus
parco has parsi and peperci ; praemordeo has praemordi and praemorsi (Plaut. fr.
120 G.) : surgo has surrexi, but in Livius Andronicus often surregi (in the old
spelling suregi; Paul. Fest. 423. 1 Th. ‘suregit’ et ‘sortus’ pro surrexit, et
quasi possit fieri surrectus, frequenter posuit Livius) ; the difference between
di-lexi, neg-lexi (cf. intel-lexi) and dé-legi, 2-legi, &c. has been explained by referring
the Perfects with x to a lost Present *lego (Gk. dAéyu, to care for) ; emo, to buy
(cf. redimo, coémo), in O. Lat. to take (Paul. Fest. 53. 26 Th. emere, quod nune
est mercari, antiqui accipiebant pro accipere) as in ad-imo, to take away,
O. Lat. ab-emo with the same sense (Paul. Fest. 4. 11 Th. ‘abemito ’ significat
demito vel auferto. ‘emere’ enim antiqui dicebant pro accipere), dir-imo,
intér-imo, ex-imo (cf. M. Ir. fo-emaim, ‘I take,’ Lith, imu, O. Sl. ima),
retains in these Compounds the Perfect-formation of the Simple Verb ai,
redemi, coemi, adémi, &c., but in the Compounds cdmo, demo, prémo, simo, in
which its connexion with emo is obscured, it takes an S-Preterite, compsi,
dempsi, prompsi, sumpsi (O. Lat. suremit however for sumpsit is quoted by Paul. Fest.
425. 3 Th.) ; similarly dmicio, a Compound of jdcio, is conjugated inconsistently
with its forgotten origin amicui and amixi, amictum, amicire (cf. ambitum Supine
of ambio, a Compound of eo, but amb-itus, a going round, canvassing, &c.); why
élicio should have as Perfect elicui, but ailicio, allext is not clear.

§ 47. Origin of the Perfect in -vi (-ui). The v (w) of Perfects like ama-vi,
mon-ti has been variously explained as a case-ending of a Verbal Noun U-stem
followed by the Auxiliary Verb ei-, ‘to go,’ as a formation on the analogy of
Verb-stems which end in v (wv) such as fov-eo (ch. iv. § 144) from the root dheght-
(fort being regarded as a Perfect formed in the same way as /fodi from fodio,
and not in the manner stated in § 39. 4) and other even less satisfactory
hypotheses. A very plausible theory supposes I.-Eur. -wi to have been the ending
of the first and third persons singular of the Perfect of roots ending in a long
vowel and compares O. Ind. ja-jiau (with final unaccented i dropped) with Lat.
(g)novt (with the middle 1 Sg. ending -ai), O. Ind. pa-prat with Lat. plévi.
These O. Ind. forms are confined to the 1 and 3 Sg. of the Perfect of roots
ending in a long vowel, e. g. da-dau, ‘I have given,’ ‘ he has given,’ but da-da-
tha, ‘thou hast given,’ &c., so that the original Latin paradigm may have been
plevi, plesti, plevit, &e. There is also an explanation possible which connects
these Latin Perfects with Oscan (and probably Umbrian) Perfects in -f- (-ff-),

“such as Osc. aa-mana-ffed ‘amandavit’ ‘faciendum curavit,’ aikda-fed
‘*aequidavit’ ‘fines ad normam derexit,’ prif-fed ‘ probavit,’ fufens ‘fuerunt.’
This Oscan f has been naturally explained, like f of Oscan fu-fans ‘ erant,’
Falisc. kare-fo ‘earebo,’ as representing I.-Eur. bh, seen in O. Ir. charub
‘amabo’ (quasi ‘*carabo’). But Italic f may also represent I.-Eur. dh (see
eh. iv. § 114), so that these Oscan Perfects, if separable from the Impft.
fufans, may contain an Auxiliary Verb connected with the root dhé-, ‘to
put’ (Gk. ri@n, Lat. cré-do, § 27 and p. 363). Latin v (w) in the middle of a word
506 THE LATIN LANGUAGE. (Chap. VIII.

may represent dv (1.-Eur. dw, as in swdvis, I.-Eur, *swadu-, and probably also
I.-Eur, dhw), while at the beginning of a word I.-Eur. dw- is represented by
Latin }, e.g. bis from I.-Eur. *dwis (Gk. 8(F)is), so that the v of the Latin Perfect
and the f (ff) of the Oscan may both represent the dhw- of a weak stem from
the root dhé-, ‘to put, place,’ a stem which appears in u fuller form (the
E-grade ? ch. iv. § 51) in Lat. cré-duas, as the numeral duo (see ch. vi. § 59)
appears beside bis. If this explanation, which requires a good many possi-
bilities, be right, Oscan -manaffed will correspond to Latin cubavit, &c., Osean
pruaffed to Latin cubuit, &e., and the -ss- of messui, &e. will be due to the com-
bination of the final dental of the Verb-stem with the initial dental of the
Auxiliary (see below § 51).

In citbui beside cubari, crépui beside -crepavi, the u-forms may be referred to
the parallel Consonant-stems of cumbere, *crepere (percrepis, Varro, Men. 124 B.),
and similarly lévi (for *lavvi) to O. Lat. ldvere, sinui to O. Lat. sdnere (cf. § 92).
But Perfects in -ui may also come from forms in which a short vowel?
preceded the Auxiliary *dhwai, just as duo comes from Zéldvo, dénwo from dé névo
(*néwod), errum from *érégto- (ef. Gk. dpoBos, épéB-tvOos) (see ch. iii. § 24).

Perfects in -tvi from Consonantal Verbs, such as petivi from péto, rudivi from
rudo, lacessivi from ldcesso, arcessivi from arcesso (and similarly quaesivi for
*quaessivi from quaeso, older quaesso, used as Perf. of quaerc), probably come
from parallel YO-stems, *petio, *rudio, *lacessio, &c. (ef. sallo and sallio, Perf.
sallivi, lino and linio, Perf. linivi, &c., § 15).

§ 48. Shortened forms of the Perfect in -vi. Latin v was often dropped
when it stood between two vowels, oftenest between two ?’s, e.g. O. Lat. dinus
for divinus, oblisci for oblivisci, ditem for divitem (so dis for dives) (see ch. ii. § 53),
but the dropping of w after a consonant is not practised in the same way
(ch. iv. § 71). The only Latin Perfect form without u is the curious monerim
of Pacuvius, quoted as an irregular formation by Nonius (507. 23 M.)
Tray. 30 R.: ;
. .. dic quid ficiam ; quod me méneris effectim dabo,
and Trag. 112 R. di monerint meliora dtque amentiam duerruncassint tuam
(parodied by Lucil. xxvi. 35 M.), which seems to be a formation from a stem
mon- (cf. Gk. pé-pova?) instead of the ordinary Perfect-stem monu-, unless it
indicates a pronunciation min(w)erim like db(yvicio of the old Dramatists
(ch. ii. § 50), and possibly augiiry)a in a line of Accius (rag. 624 R.) (quoted
by Nonius 488. 2 M., who makes the word augura, by some regarded as a Neut.
Pl. like rabira ; ef. O. Ind. djas- N., ‘ strength’) :

pré certo arbitrabor sortis, 6racla, adytus, atigura,
and progen(y)em (? prégéniém) in a pentameter line of a Scipio epitaph of this
period (C. I. L. i. 38, v. 130 B.c.):

progenie mi genui: facta patris petiei

(probably a graver’s mistake for progentem genui). (Posisse, stipisse are shortened
forms, not of posuisse, supuisse, but of posivisse, sapivisse, § 39.4). But Perfect-forms

 

 Démui, which is usually explained domitus). An example of a stem in
as Perf. of *diméyo-, a Causative by-I.-Eur. -@ is O. Ind. vami-mi (3 Sg.
form of *domayo- (§ 29), is by some vami-ti, but also with the Thematic
referred to an I.-Eur. stem ending in Vowel, vama-ti), the equivalent of
-A or -d (ef. Gk. d-Sdudros with Lat. Lat. vimo, -cre.
§ 48.] THE VERB. PERFECT. 507

without v were extremely common ; in Cicero’s time nosse, jidicasse, &e. were
usual (Orat. xlvii. 157 quid quod sic loqui, ‘nosse, judicasse’ vetant, ‘novisse’
jubent et ‘judicavisse ’? quasi vero nesciamus in hoe genere et plenum verbum
recte dici et imminutum usitate) ; and in Quintilian’s time the forms audivisse,
scivisse, conservavisse were scarcely heard even in public speaking [Quint. i. 6.
17-21 inhaerent tamen ei (sc. analogiae) quidam molestissima diligentiae
perversitate ut ‘audaciter’ potius dicant quam ‘audacter,’ licet omnes
oratores aliud sequantur... his permittamus et ‘audivisse’ et ‘scivisse’ et
‘tribunale ’ et ‘faciliter’ dicere .. . sed abolita atque abrogata retinere
insolentiae cujusdam est et frivolae in parvis jactantiae. multum enim
litteratus, qui sine adspiratione ct producta secunda syllaba salutarit (‘avere*
est enim), et ‘calefacere’ dixerit potius quam quod dicimus et ‘conservavisse,’
his adiciat ‘face’ et ‘dice’ et similia. recta est haec via: quis negat? sed
adjacet et mollior et magis trita]. Servius, in a note on the form leniit in
Virg. Aen. i. 451, tells us that the pronunciation of every-day life was leniit, the
form with the short penult being confined to poetry (sed hoc in metro ubi
necessitas cogit : nam in prosa et naturam suam et accentum retentat).
Varro (ZL. L. iii. fr. p. 148 W.) mentions amasti, nosti, abiit as the favourite forms
of his time ; and in Terence the usual formsof Perfects in -évi, -wi and of névi
are the contracted (in Plautus the uncontracted, though in the middle, not
the end, of a line siris, not swvéris, is used), while the uncontracted forms of
Perfects in -dvi are used only at the end of a line (in Plautus equally with
the contracted). Eo and its compounds have even in Plautus usually the
form -ii- in Perfect forms, except iwi, exivi, ambwi, &c. All this indicates the
forms without v to have been phonetic developments of the others, and to
have gradually established themselves in exclusive use, the v being dropped
e@rliest and most persistently in forms where it came between two ?’s. The
shortened form of -Gvit, if it did not push itself into literary usage, must
however have become, sooner or later, a part of every-day pronunciation, for
the Romance forms point to a Vulgar Latin -aut for -avit, e. g. Ital. compro
from a Vulg. Lat. *comp(d)raut. (On audit, &e. for audivit, see ch. iii. § 10. 2.)
The v of the Perfect of verbs whose stem ends in v (u) was not dropped to
the same extent ; a fact which may be explained by the different sound of -ar-
for -avv- in céveram (pronounced *cavveram?) and -dv- in amaveram. But even
these Perfects when 0, u precede v, are contracted almost as readily as the
others : e. g. dévoro for dévdvéro in the Praetextata of Accius called ‘ Decius’
(referring to Decius Mus) (Trag. (Praet.) 15 R.]:
pitrio exemplo et mé dicabo atque dnimam deuoro héstibus ;

commérat, Ter. Phorm, 101 ; comméorunt, remosse, Lucr.; adjiro (or adjiiéro ?), in
a passage of the eleventh book of Ennius’ Annals (1. 386 M.) describing the
Macedonian campaign of Flamininus (the words are addressed by a shepherd-
guide to the Roman general) :

O Tite, siquid ego adiuro curamue leuasso,

quae nune te coquit et uersat in pectore fixa,

ecquid erit praemi ?
and the same form (3 8g.) in Plaut. Rud. 305 and Ter. Phorm. 537, &c. Morunt,
méram, commorunt, commoram, commossem, &c. are not uncommon in tlre Augustan

poets, e. g. Hor. 8. i.9. 48: :
dispeream ni

summosses omnis,
508 THE LATIN LANGUAGE. (Chap. VIII.

and jrcrint occurs in Catullus, Ixvi. 18:

non, ita me diui, uera gemunt, iuerint
(for other examples, see Neue, ii’. 533). The usage on inscriptions of the
3 Pl. Pft. forms of cio, prébo similarly points to the forms with « being the
older [e. g. coraueront, C.I. L. i. 73, cowrauerunt 1419, cotrauerunt 565, &c. (once
coirarunt), but coeraverunt and coerarunt, curarunt].

§ 49. Shortened forms of the Perfect in -si. A better case might be
presented for the theory that the shorter forms of Perfects in -si (-xi) are
really ancient forms of different origin from the full forms, for dixti, dixem,
dixe, &c. are most found in the early Dramatists, though they are by no
means uncommon in Virgil and later poets. But the explanation of this
fact is rather that they were felt to belong more or less to colloquial Latin,
and so were freely admitted into Comedy and Tragedy (in both of which
the every-day language of cultured society was employed), but not so freely
into other literature. (Quintilian ix. 3. 22 remarks on Cicero’s use of dixti
“excussa syllaba’ in the pro Caecina, 29. 82.) A thing that is almost con-
clusive proof that dixem, &c. are not parallel forms to dixim, dixo, &e. is that
these shortened forms are only found with Verbs which form the Perfect in
-si (-wi); we have no *capsem, answering to capsim, no *axem answering to
axim. [On the wrongness of the readings subaxet (for subaxit) in Pacuy. Trag.
163 R., faxet in Plaut. Cap. 712, accepsti, Trin. 420, 964, &e., see Neue, ii”.
p. 539, and for examples of the shortened forms of Perfects in -si (-ai), e.g.
vize and vixem in Varro (Men. 321 B. non eos optime vixe qui diutissime vixent
sed qui modestissime), misti, misse, scripsti, rescripsti (Cic. ad Att. v. 9. 2),
-sumpsti, -sumpse, -cesti, -cessem, -cesse, duxti, duxe, -spexti, -spexe, veuti, traxe,
emunati, &e. see Neue, ii’. pp. 536 sqq. ] f

§ 50. O. Lat. Perfects in -a(v)i. Varro declares that in the correct pro-
nunciation of his time the « of verbs like pluo, lwo was short in the Present
but long in the Perfect Indicative, as in the Perfect Participle Passive of
similar verbs (Z. L. ix. 104 quidam reprehendunt, quod ‘ pluit’ et ‘luit’ dica-
mus in praeterito et praesenti tempore, cum analogiae sui cujusque temporis
verba debeant discriminare. falluntur; nam est ac putant aliter, quod in
praeteritis u dicimus longum ‘ pluit,’ ‘liit,’ in praesenti breve ‘ pliiit’ ‘luit’;
ideoque in venditionis lege fundi ‘rita caesa’ ita dicimus ut u producamus)
(but ef. dirtitus). Priscian (i. 504. 22 H.) makes this pronunciation, -a, in the
Perfects of Verbs in -wo, a feature of Old Latin, and quotes Ennius (Ann. 135 M):

adntit sese mecum decernere ferro,
while in another passage (i. 503. 14 H.) he quotes pluvisse, pluvit with v (the
quantity of the uv is not mentioned) from Livy.

Institai is the reading of the MSS. in Plaut. Most. 86 (in bacchiac metre,
a metre in which the long quantity of a vowel preceding another vowel is
often retained) :

argtimentaque in pectus miulta institii.
The commonest instance of -i#i or -divi is in the Perfect of the old verb /uo, viz.
fui, used as the Perfect of swn, as fiimus in the boast of Ennius on being
made a Roman citizen (Ann. 431 M.):

nos sumus Romani qui faiimus ante Rudini,
Siit (along with profitit) in Plaut. Capt. 555 :

quibus insputari saluti fait atque is profitit,
§§ 49-52.) THE VERB. PLUPERFECT. 509

Sierim, Mil. 1364, &c., fdisset in Ennius (Ann. 297 M.):

magnam cum lassus diei
partem fiisset de summis rebus regundis.

We have the spelling rvverr on an epitaph (C. I. L. i. 1051). (See ch. iv. § 70.)

§ 51. Some Irregular Perfects. Besides the byforms in s of the Perfect
of parco, -légo, &e. which are found in the older as well as the later literature
(§ 46), we have some which are more recent coinages on the type of other
verbs. Velius Longus (74. 4 K.) spéaking of sorpsi, a byform of sorbui, says :
cum recens haec declinatio a sordidi sermonis viris coeperit (ef. Caper 94.
14 K. non est... ‘sorbsi,’ sed sorbui), and similarly describes terui (for trir’) as
a ‘consuetudo nova’ (absorpsi is used by Lucan, iv. 100), The Perfect vulsi,
for velli, seems also to belong to this class of later coinages (dvulsit in Lucan,
v. 594; cf. Neue, ii?. 503), and -céimui of concinui (Ov., Tibull.), occanwi (Sall.),
ineinui (Varro), &e. may be a remodelling of a Perfect with lost Reduplication
-cini (con-cini, oc-cini, in-cint) after the pattern of sdnui, consonui, &e. Messui
(found in the older writers) is generally regarded as due to a similar
addition of the ending -wi to an already existing Perfect *messi, and also
nexut (with nexi), pexui (with pexi), although, if the theory of the origin of
the V-perfect stated in § 47 be correct, the sibilant might be referred to the
influence of the dental in the auxiliary *dhwai (cf. asus for *ut-tus, *oit-to-,
fisus for *fud-tus, &c., ch. iv. § 95. p. 281). In late inscriptions we have e. g.
reguit (C. I. L. v. 923, from Aquileia: septimae qui cohortis centuriam reguit,
a pentameter line), convertuit (viii. 2532 D b 1); Apuleius uses conterui (see
Georges Lex Wortf. s.v.). On Heteroclite Perfects like fui, Perfect of sum, tili
(older tétili) of féro, see § 97.

In Vulgar Latin, as reflected in the Romance languages, the ending -ui
encroached still more. Cognovit became cognovuit (Ital. conobbe), movit became
movuit (Ital. movve), cecidit was replaced by caduwit (Ital. cadde), stetit by stetuit
(Ital. stette), venit by venwit (Ital. venne), &c. (ef. § 92 on the Vulg. Lat.
P.P. P. in -dtus, e.g. caditus, fallen (Ital. caduto)]. The ending -si also
extended its sphere, respondit passing to responsit (Ital. rispose), prendit to
pre(n)sit (Ital. prese), cucurrit to cursit (Ital. corse), &e. (See Meyer-Liibke,
Rom. Gramm. ii. pp. 297 8qq-)

§ 52. E. Pluperfect. The Pluperfect-stem is formed by
adding to the Perfect-stem an S-suffix, probably a Preterite
of the Substantive Verb es-, as the Imperfect in -dam, e.g.
dmda-bam, uses a Preterite of the Verb bheu-. The -éram, -éras,
-érat of dimdv-eram, méru-eram, replév-eram, dia-eram, vid-eram,
pépig-eram, audiv-eram is most naturally referred to the Im-
perfect of sum, though videram has been, with the O. Ind,
4-védig-am, analyzed into a Noun-stem in -is (cf. vedis-sem) as
Gk. 7j5ea into a Noun-stem in -es (see § 3). The -is- of the
Pluperfect Subjunctive dmdvissem, minwissem, vidissem, audivissem
must, if the old view be correct which sees in these forms an
510 THE LATIN LANGUAGE. (Chap. VITI.

Auxiliary essem appended to the Perfect-stem, be a change of
-es- to the I-vocalism of the Perfect, just as the a of Gk.
derédrw, &e. is due to the apprehension of a as the appropriate
vowel of the Aorist Tense.

§ 53. F. ‘Future-Perfect. The Future-Perfect adds to the
Perfect-stem -é7d, -éris, -érit, &e. (3 Pl. -érunt), which seems to
be nothing else than the Future (in form a Subjunctive, § 55)
of sum, appended as an Auxiliary Verb, though those who
explain the -er- of vidéram as I.-Eur. -is-, the suffix of a Noun-
stem, see the same -is- in vad-éro (cf. O. Ind. védigam) (see 3).
Scansions like fecertmus (Catull. v. 10), 1 Pl. of fécéro, are due
to the confusion of the Future-Perfect forms with Perfect
Subjunctive forms (see Neue, ii?. p. 510); aderiut is similarly
used for ddérunt throughout the Lex Col. Jul. Genetivae
Urbanorum of 44 B.C. (Aph. Epigr, ti. p. 122), The use of the
Fut. Perf. in the Republican Dramatists in sentences like mox
ivero, where it hardly differs from the Future, suggests that in

- its original usage the Tense was more of a Future than of
a Future Perfect (see 4. ZL. L. iv. 594).

In the Umbro-Oscan languages we have a suffix -us- (Umbro-
Oscan u may represent I.-Eur. 6 as well as I.-Eur. u), which is
supposed to be the termination of the Perfect Participle. Thus
Osc. fefacust * fecerit’ will be like Gk. dedpaxds éoouar, Umbr.-
fakurent like dedpaxdres eoovtat. (On the Umbro-Oscan Perfect
Participle in -us, see § 89.)

§ 54. G. Tenses formed with Auxiliary Verbs. We have seen
that the suffix used to form the Imperfect Tense of all Verbs,
-ba-, and the suffix used for the Future of Vowel Verbs, -dd-
(-4é-), are nothing but parts of the Auxiliary Verb bheu- (Lat.
fut), that the S-suffixes of the Pluperfect and Future-Perfect
Indicative, and of the Imperfect, Perfect and Pluperfect Sub-
junctive are probably to be referred to the Auxiliary es- (Lat.
sum), and we have seen the possibility of a similar explanation
of the V-suffix of the Perfect Tense as a form of the root dhé-,
‘to set, put,’ used as an Auxiliary Verb. These theories receive
a sanction from the proneness of the Latin and other Italic
languages for Periphrastic Tense-forms. Not only was the
§§ 53-55.] THE VERB. SUBJUNCTIVE. 51il

Italic Perfect Indicative Passive a form of this kind [Lat.
dmatus sum, (1) I was loved, (2) I am in a state of being loved,
amatus fui, | was in a state of being loved, Osc. prifti-set
‘probata sunt,’ teremnatust ‘ terminatus est,’ Umbr. serehto est
‘scriptum est’ (the Auxiliary is joined in writing with the
Participle in frosetomest)|, and other Tenses of the Passive
(Pluperfect, Lat. amatus éram, amatus fuéram, Future-Perfect
amatus éro, amatus fuéro, Perfect Subjunctive amatus sim, amatus
Suérim, Pluperfect Subjunctive amatus essem, amatus fuissem),
but we have many other examples of Periphrastic formation.
The 2 Pl. Pass. ending -mini, e.g. férimini, is explained (§ 81)
as the ending of a Pres. Part. Pass. (Gk. gepduevor) with
suppression of estis, just as esse is commonly suppressed in the
Fut. Inf. dictirus, -a, -um (esse) (§ 86). Osc. manafum ‘mando,’
if it correspond to ‘mandans sum,’ is another example, showing
a Pres. Part. with the Substantive Verb (cf. C.L.L. i. 196
senatuosque sententiam utei scientes esetis). The Auxiliary co
appears in the Latin Fut. Inf. Pass., e.g. ditwm (1 Sup.) wi
(§ 87), the Auxiliary Addeo in such phrases as missum habeo for
misi or dimisi (e.g. Plaut. Pseud. 602 illa omnia missa habeo,
quae ante agere occepi), the Auxiliary DHE- (reddo, do, fucio, &c.)
in phrases like missum facio for mitto or dimitto (e.g. Plaut.
Amph. 1145), perfectum reddo for perficio (e.g. Plaut. Asin. 122),
factum dabo for faciam (e.g. Ter. Bun. 212). In the Romance
languages these expressions have supplanted many of the Latin
tenses; e.g. cantare habeo (Fr. chanter-ai, Ital. canter-d, with
habeo, habes, &c., reduced to hayo, has, hat, haunt, &c.) has sup-
planted cantabo, &c., so that the Latin Future survives in Romance
only in a few isolated forms, Ital. fia, O. Fr. ier from Lat. jiam,
ero; we find a Periphrastic Perfect consisting of the Perf. Part.
Pass. with swm, sto (Intrans.) or with habeo, teneo (Trans.); and
for the Passive the same Participle with sum, fio, venio, &c. (See
Meyer-Liibke, Rom. Gram. ii. pp. 138 sqq-)

§ 55. III. THE MOODS. A. Subjunctive. (Relics of the
I.-Eur. Optative Mood in Latin.) The I.-Eur. Subjunctive had
the functions of a Future (cf. Hom. J/. i. 262 08 yap mw totovs idov
dvépas od% Twpya), as well as of a true Subjunctive ; it had
512 THE LATIN LANGUAGE, [Chap. VIII.

also Imperatival functions, and in O. Ind. while Subjunctive
forms are frequent in the oldest literature, the Vedas, they
survive in classical ‘ Sanscrit’ only in Imperatival use, e.g. 1 Sg.
Imper. bhar-d-ni, ‘let me carry.’ To the unweakened stem
(E-grade) was appended in the Athematie Conjugation the short
vowels % and 6, followed by either the Primary or the Secondary
Person-endings (§ 65): e.g. from the Athematic Verb es-, ‘to
be, we have 3 Sg. Subj. *es-e-t(i) (O. Ind. dsat(i) ). In Greek
these athematic Subjunctive forms are still seen in Homer (e.g.
Bihoopev, reloere), and later ina few so-called Futures like €-0-par
from the Athematic Verb ed-, ‘ to eat,’ as in Latin we have éo,
érit 3 Sg. similarly retained as a Future Tense. But in Greek,
as in Latin, the athematic forms have been almost wholly
supplanted by the thematic forms, which show a long vowel,
sometimes 4, as in the Italic, Celtic and other languages,
sometimes ¢ (which in Greek varies with 6, e.g. pép-n-Te, pép-
w-pev, like athematic e-o in relo-ere, Bijo-o-yev), as in the Italic
and Greek languages; and this process must have begun very
early, for thematic Subjunctive forms of verbs belonging to the
Athematic Conjugation appear in many instances to have been
I.-Eur. forms. Thus *ed-a-, *ed-é- must have been an I.-Eur.
Subjunctive stem of ed-, ‘to eat, for we find in Lat. edd-mus,
edda-tis, in Greek éw-pev, edy-re, and in O. Ind. dda-n 3 PI.
(O. Ind. 4 may represent I-Eur. a, é, 6); *es-é- from es-,
‘to be,’ appears in Gk. éy-re, @w-yev (cf. O. Ind. 4sa-t 3 Sg.,
asa-tha 2 Pl.); and ey-a- from ei-, ‘to go, in Lat, ed-s, ed-tis
(cf. O. Ind, aya-s 2 Sg., aya-t 3 Sg.). The discarded athematic
forms are supposed to have been utilized in Greek as Futures
[the Future in -co- (-ce-)]; thus Att. relcouer, retoere, de(Fouer,
delLere, Tyunoowev, tyunoere, &c. will not be of the same class
as the ordinary I.-Eur. Future in -syo-, e.g. O. Ind. dék-Sya-mi
corresponding to Gk. deffo, Lith. ba-siu to Gk. dice, but will be
Subjunctives of S-Aorists ; and Latin Futures in -so-, like dima,
faxo, &c., have been referred to the same source. (On this
theory, see § 3.)

In Latin the 4-forms were not used for the first Con-
jugation, probably because a Subjunctive like *amay-a-s(i) 2 Sg.
must have become amdés, and so been merged in the Indicative
§ 55.] THE VERB. SUBJUNCTIVE. 513

2 Sg. amas. The E-forms only were used for this Conjugation
in Latin, e.g. amés from *amay-é-s(i) (cf. Ose. devaid ‘ juret,
tadait ‘ censeat, sakahiter ‘sacretur’; Oscan {, in the Latin
alphabet 7, represents I.-Eur. 6, e. g. ligato- ‘ légatus,’ zzcolo- M.
‘diécula’), but in Umbrian we have A-forms in kuraia ‘ curet,’
etaians ‘itent. Similarly the E-forms were not used for the
second Conjugation in Latin, to avoid confusion between *widéy-
é-s(i), *vidés 2 Sg. Subj., and vidés 2 Sg. Ind., only the A-forms
being allowed, e.g. *widéy-a-s(i), videds. But in the other
Conjugations the A- and E-forms probably existed side by side,
until the latter were appropriated for the Future functions
(see § 36), e.g. ferds 2 Sg. Subj. from *bher-a-s(i), ferés 2 Sg.
Fut. from *bher-é-s(i), faciat 3 Sg. Subj. (Osc. fakiiad, Volse.
fagia, Umbr. facia), faciet 3 Sg. Fut. These A- and E-Sub-
junctives show in Umbro-Oscan the Secondary Person-endings
(-d in 3 Sg., -ns in 3 Pl.).

The Optative must have entered into competition with the
Subjunctive at an early time, for it has almost entirely ousted
the Subjunctive forms in the Teutonic and Balto-Slavic families
of languages, and in Latin we see the struggle still going on
between the Optative-forms édi-mus, edi-tis, and the Subjunctive-
forms edd-mus, eda-tis from the root ed-, ‘to eat,’ while in the
case of the root es-, ‘to be,’ the Optative si-mus, si-tis have
driven out of the field the proper athematic Subjunctive forms
ert-mus, ert-tis (from *eso-, *ese-), which have been relegated to
the Future function, as an *(e)sé-mus, *(¢)sé-tis may lurk in the
Imperfect Subjunctive ending (see § 34).

The I.-Eur. Optative had in the Athematic Conjugation the
weak grade of stem, with a suffix which was in the Singular
Active -yé- (-tyé-) and elsewhere -i-, and with the Secondary
person-endings. Thus from the root es-, the I.-Eur. Optative
forms were: 2 Sg. *syé-s (O. Ind. sya-s) or styé-s (O. Ind.
s-lyd-s, O. Lat. siés), 2 Pl. *s-i-té (cf. Lat. s-z-dis). In the
Thematic Conjugation the suffix was -oi- in Singular and
Plural alike, e.g. from the root bher-, to carry, 2 Sg. *bher-oi-s
(Gk. pép-or-s), 2 Pl. *bher-oi-te (Gk. ép-o.-re). L-Eur. 01 of
the Thematic Optative would in the unaccented syllable in
Latin become -e?-, then -7- (ch. iii, § 18), and I.-Eur. -yé- (-lyé-)

ul
514 THE LATIN LANGUAGE. [Chap. VIII.

of the Singular Active of the Athematic Optative has been
mostly replaced by -i-, the weak form of the suffix (e. g. class. stm
for O. Lat. siem, like stmus, sitis), so that a Latin Optative form
like edis might equally well represent an I.-Eur. athematic
*éd-i-s (with E-grade of stem and with i transferred to the
Singular from the Plural, as -yé- is transferred to the Plural
from the Singular in Gk. ord-fy-yev, a byform of craiper), and
an I.-Eur. thematic *éd-oi-s. The probability however is that
the Optative was confined to the Athematic Conjugation in
Latin, and represents in every case I.-Eur. -yé-, -i-. Again,
it would be possible to argue that stés, stémus, &c. were
representatives of I.-Eur, *stit-yé- (Gk. ord-ins, ord-(y-pev), and
not examples of E-Subjunctives. The resolution of these doubts
must come from the Umbro-Oscan languages, of which the
remains hitherto discovered offer too scanty material to enable us
to separate with certainty Latin thematic and athematic Optatives,
and Latin Subjunctives in -é-. The use of the weak grade of
the suffix -yé- in the Singular of the Optative.of es- appears in
these languages too, e.g. Umbr. sé ‘sis,’ si‘ sit, like szws ‘sint,’
Marrue. -si ‘sis’ or ‘sit,’ and has been referred to an Italic
weakening of unaccented yé to i. How far it is possible to
assign an Optative force to forms like créduim in the older
literature, and a Subjunctive force to forms like creduam is
doubtful. The Optative origin of the old forms in -(s)sim comes
out clearly in their use in prayers (e.g. Juppiter prohibessis
scelus, di mactassint, and the formula of the ancient Augural
prayer: bene sponsis beneque uolueris; see § 5), but in
process of time all distinctions between Optatives in -im, Sub-
junctives in -am and Subjunctives (possibly Optatives) in -em
came to be effaced.

§ 56. Some O. Lat. Subj. and Opt. forms. In one of the oldest Latin
inscriptions preserved, the Dyenos inscription, we have an A-Subjunctive
mitat ‘mittat’ used as a Future, with the Optative of the root es-, sied ‘sit,’
used in the true Optative sense: qoi med mitat, nei ted endo cosmis uirco
sied ‘qui me mittet, ne erga te comis Nange sit’ (asted on the same inser.
is variously interpreted as ‘adstet’ and as ‘ast’; see ch. x. § 5). Siem, sies,
siet (on sient, see § 73; *siemus, *sietis have not found their way into Latin
as syama, syata have into O. Ind.), possiem, &e. are by Terence used almost
only at the end of a line or hemistich, i.e. through metrical necessity, but
§ 56.) THE VERB. SUBJUNCTIVE. 515

siem, sitet are almost invariably used in old laws (e.g. C. 1. L. i. 196. 303 197.
21; 198 passim ; 199. 6; 200 passim), and Cicero (Orat. xlvii. 157) says, ‘siet’
plenum est, ‘sit’ imminutum: licet utare utroque. [For statistics, see
Zander, Vers. Ital. p. exx, who makes -i- of sit, &c. not the I.-Eur. i of the
Plural, but a Latin weakening of -ie- as in Corneli Voc. for Cornelié (? ch. vi.
§ 31)]. Besides the class. Lat. ‘Subjunctives’ (Optatives) in -im, sim, édim,
rélim (in Plaut. velis and vis are used as the metre requires, without difference
of meaning), with its Compounds ndlim and malim (on noli, see § 58), we find
an O. Lat. Optative duim. It comes from duo, a bystem both of do, to give
(root di-) and of -do, to put (root dhé- of ri-On-w, &e.), &. g. duitur (v. 1. ardu-
uitur) XII Tab. x. 7, interduo, Plaut. fr. inc. 2G. ciccum non interduo, concreduo,
Aul. 585, concredui, Perf., Cas. 479. We have in Plautus the Optative forms
duim, perduim, especially in prayers and execrations (e.g. Most. 668 di istum
perduint), interduim (e. g. Rud. 580 ciccum non interduim), creduim (in phrases
like Amph. 672 si situlam cépero, Ninquam mihi diuini quicquam créduis
post hune diem, ‘may you never trust me again’), as also the A-Subjunctive
forms creduam .e. g. Bacch. 504 nam mihi diuini nimquam quisquam créduat,
Ni ego, &.), and accreduam (Asin. 854 néque diuini néque mi humani pésthac

quicquam adcréduas . si, &e.), while an E-Subjunctive form (in Future
sense ?) is quoted by Paul. Fest. 20. 22 Th. addues, addideris (cf. ib. 47. 6 Th.
duis... pro dederis). We have in Conditional use, e. g. duit in a Law of Numa

(ap. Paul. Fest. 278. 9 Th.) si qui hominem liberum dolo sciens morti duit,
paricidas esto; adduit in a Plebiscitum de Ponderibus Publicis (ap. Fest.
322. 11 Th.) siquis .. . faxit iussitue ... dolumue adduit, &. Festus also
quotes an old form produit which he explains by ‘ porro dederit’ (284. 16 Th.).
Duim was the form appropriate to Early Latin prayers, as in the prayer at
the ‘agri lustratio,’ preserved by Cato (R. R. exli. 3) : pastores pecuaque salua
seruassis duisque bonam salutem ualetudinemque mihi domo familiaeque
nostrae; it is used even in Tiberius’ letter to the Senate (Tac. Ann.
iv. 38). There are also uncertain traces of Optative forms from other verbs,
such as coguint, the reading of the Palatine MSS. in Plaut. Pseud. 819 (but
cocunt in the Ambrosian Palimpsest), ¢emperint, the reading of the same family
of MSS. in Truc. 60, a line for which the evidence of the Palimpsest is not
available (other examples, see in Neue, ii’. 442: carint of the Palatine MSS. in
Most. 858 is carent in the Palimpsest, and the corrupt reading of the Palatine
MSS. in Men. 984 u, where this line of the Mostellaria is wrongly inserted,
culparent for culpa carent, shows carent to be the right form ; verberit of the law
of the XII Tables, ap. Fest. 290. 15 Th.: si parentem puer uerberit, ast olle
plorassit, is a corruption due to the fact that the words were wrongly divided
in the archetype verberetas tolle, and the first word, being mistaken for a fre-
quentative Verb, was changed to verberitas).

An O. Lat. A-Subjunctive is fuam used in the sense of fiam, e.g. Bacch. 156,
in the amusing conversation between young Pistoclerus and his ‘paeda-
gogus’ Lydus:

Pisr. fiam, ut ego opinor Hércules, tu autém Linus.
Lyp. pol métuo magis, ne Phoénix tuis factis fuam,
teque dd patrem esse mértuom rentntiem,

sometimes merely in the sense of sim, e.g. Virg. A. x. 108:

Tyros Rutulusve fuat nullo discrimine habebo.
Ll2
516 THE LATIN LANGUAGE. (Chap. VIII.

The equivalent of forsitan (which is not used by Plautus, and only seldom, if
ever, by Terence, ch. ix. § 5) is in Plautus fors fuat an, e.g. Pseud. 432:

fors faat an istaec dicta sint mendacia.

(Cf. Ter. Hec. 610 fors fuat pol! ‘heaven grant it may !’)
Nonius (478. 26 M.) quotes volam for velim from Lucil. (xxviii. 15 M.):

eiddla atque atomus uincere Epicuri uolam,

and Plaut, Asin. 109 siquid té uolam, Ubi eris?; but in the second at least
of these passages it seems to be used in the Future sense.

§ 57. B. Imperative. In the 2 Sg. Act. of the Present Im-
perative the bare stem of the Verb is used, e.g. Athematic *ei,
from the root ei-, ‘to go’ (Gk. é-e., Lat. ew-i from *ew-ei),
Thematic *bhere, from the root bher-, ‘to carry ’ (O. Ind. bhara,
Arm. ber, Gk. dépe, O. Iv. beir, Goth. bair; Lat. age), But
a particle was often added: the particle -dhi to the 2 Sg. in the
Athematic Conjugation (e.g. O. Ind. i-hi, Gk. t-@.; O. Ind.
vid-dhi, Gk. fo61 for *F.3-61, O. Lith. veiz-di, O. SI. viz-di); the
particle -/i or -4e in Lithuanian, e. g. ei-k, ‘ go,’ du-k, ‘ give,’ ba-k,
‘be’; the particle -u (cf. O. Ind. sé for *sou, Gk. of-ros for *cov-
tos) in Sanscrit to the 3 Se. and Pl., e.g. bhdratu, bhérantu ;
the particle -tad to various persons both in the Athematie and
Thematic Conjugations [e. g. O. Ind. vit-tat 2 Sg.; Gk. torw for
*Fir-rwd 3 Se.; Gk. éo-rw(d), Lat. es-td(d)]. In Latin the 2
Sg. Imper. in -20(7) is called the Future Imperative, because it
expresses a command, not for immediate performance, but for
performance after something shall have happened (e. g. Plaut.
Mere. 770 eras petito, dabitur; nunc abi; Hor. C. iii. 14. 23
sl per invisum mora janitorem Fiet, abito), and the same sense
is attached to the O. Ind. 2 Sg. Imper. in -tat (see Delbriick,
Altind, Syntax, p. 363), so that it is not unlikely that this
particle -téd is nothing but the Abl. Sg. of the Demonstrative
Pronoun-stem to- (ch. vii. § 13. 1), and means ‘ from this,’ ‘ there-
upon. It is also found with the 3 Sg. Dual and Plur., and
perhaps took with it originally the weak grade of the Verb-
stem (e.g. Gk. t-rw 3 Sg., du-vi-ro 3 Sg., O. Ind. ky-nii-tat
2 Sg.,‘do thou’; but not in Latin zo, Umbr. etu, eetu, for
*ei-tod). Inthe 3 Pl. it is added to what is called the ‘ Injunc-
tive’ 3 Pl., viz. a form resembling an augmentless Imperfect
§§ 57, 58.] THE VERB. IMPERATIVE. 517

(I.-Eur. *bheront like Impft. ¥*e-bheront), e.g. Gk. depdvrw for
*pepovt-rwd, Lat. ferunto for *ferdnt-tod. For the 1 Sg. and
Plur. the Subjunctive was used, e. g. Lat. feram, feramus, In
the 2 Pl. we have the ‘Injunctive’ form (e g. *bherété, O. Ind.
bharata, Gk. pépere, O. Ir. berid, Goth. bairip ; Lat. dgite from
*agete), beside which we find in Latin a ‘Future’ form with
-tote, e.g. estote, apparently the addition of the 2 Plur. suffix -¢¢
(§ 72) to the 2 Sg. ‘Future’ Imperative (e.g. estd.)

In the Latin Passive, the Injunctive form in -sd (e. g. I.-Eur.
*bhere-si, Zend bara-rwha, Gk. pépeo, contr. pépov, é-pépeo, contr.
é--€pov) is used, e.g. dyé-ré for *age-sd (see ch. iii. § 38), in the
2 Sg., and in the 2 Pl. the old Passive Infinitive! (Dat. Sg. of
a MEN-stem), e.g. dgi-mini for *age-menai (Hom. Gk. dyé-
evar), da-mini (O. Ind, di-mané Inf, Hom. Gk. 86-year); in
the 2 Sg. ‘Future’ Imper. and in the other persons the final
of -téd is changed to -r, e.g. dgttor, ditor, aguntor (cf. Umbr.
emantur, emantu, twrsiandu). A byform for the 2, 3 Sg. is in
-mind, e.g. prae-faimino, formed apparently by Anal. of 2 PI.
-mini. There is no Perfect Imper. in Latin, though mémini,
a Perfect used for a Present, has mémento for *mement-tod
(Gk. weudrw for *ueun-rwd). Memento, like other 2 Sg. Impera-
tives in -¢6, expresses a command that usually has reference to
the future, a reference naturally suggested by the command
* remember,’

§ 58. Other examples of 2 Sg. Imper. with bare stem. (1) Athematic:
like athematic Ind. im-plé-s, vidé-s, curd-s, fini-s are Imper. im-plé, vide (Lith.
pa-vydé-k), curd, fini, &e. Fer, vel, es (from sum) are then likely to be athe-
matic too, since their Ind. is athematic (e.g. 3 Sg. fer-t, vul-t, es-t; see § 2).
Fer cannot be an Injunctive form *fer-s (like ter for *ter-s from I.-Eur. *tri-s,
ch. vi. § 61), if, as is probable, it is in Plautus a thoroughly short syllable,
capable of acting as a brevis brevians (see ch. iii. § 42), in Cure. 245 aufér
istaee quaeso, whereas ter is a long syllable in Plautus, e.g. Bacch. 1127. In
Mil. 1343 however one family of MSS. reads: fér aequo animo, the evidence
of the Ambrosian Palimpsest being unfortunately wanting, a reading which
may easily be changed to fér animo aequo (as Cure. 245, for which there is
similar MS, evidence, is changed by some editors to aufer quaeso istaec). The

 

1 Or the Plural of the old Pres. ellipse of estis (§ 82). The Inf. is
Part. Passive, agimini for *aSo-menoi used for the Imper. in Italian é&e.
(Gk. dyd-pevor), with ellipse of este, as in phrases like non parlare ‘do not
2 Pl. Ind. agimint for the same, with speak.’
518 THE LATIN LANGUAGE. (Chap. VIII.

small number of lines with decisive evidence on the quantity of fer in Plautus
makes it difficult to speak with certainty. Fer is short in Asin. 672 fér amanti
ero salitem. Vel isa short syllable, capable of acting us a brevis brevians in
Plautus, e. g. Amph. 917 Vél hiine rogato; es, ‘be,’ cannot be shown (like es,
‘art’) to be long by position in Plautus (see Solmsen, Stud. Laut. p. 185) ;
es, ‘eat,’ for which we should expect *ed, beside 3 Sg. esto, may be coined on
the type of es, ‘be,’ beside 3 Sg. esto, although both es, ‘be,’ and es, ‘eat,’ can also
be explained as Injunctive forms (like Gk. émi-oxe-s, &c.) for *es-s and *ed-s'.

Cé-dd (with Plur. ce-tte for *ce-dite?), ‘give me’ or ‘tell me,’ seems to contain
the Pronoun *ke (perhaps Lith. szé, ‘hither’; see ch. vii. § 15), prefixed as
an Adverb or Preposition, ‘here,’ ‘hither’ (cf. Osc. ce-bnust ‘hue venerit,’
composed of *ke and a tense of the I.-Eur. root gtem-, ‘to come’) to an
athematic Imperative *do (Lith. dt-k; cf. Gk. 6-6), the final vowel,
shortened by the Law of Breves Breviantes after the short syllable cé-, being
invariably short owing to the rapid utterance of the word in every-day talk
(so havé for avé in the pronunciation of Quintilian’s time, ch. iii. § 42).

Dé for *dd shows the same transference to the A-Conjugation as Pres. Ind.
dds, dat (O. Lat. dat) (see § 2). ,

Another example of an athematic Imperative may be fu in the Carmen
Arvale, if the words satur fu, fere Mars, are rightly interpreted ‘satur esto, fere
Mars.’ Fu will be Imper. of *fwo, like Lith. bu-k. Nod may come from
a bystem of the fourth Conjugation, I-Eur. *wel-yo- (Goth. vilja, O. SL
velja) (see § 2). (On Late Lat. aufere, see Georges, Lex. Wortf. s. v.)

(2) Thematic. The thematic Imperatives 0. Lat. dice, dice, fice (for *fact,
from stem fac-yé-, § 16) drop their final -¢ in classical Latin owing to their
frequent use in word-groups, i.e. in close connexion with a following word
(like atg(ue), neq(ue) before consonants ; see ch. iii. § 95). Dic mihi, fac sciam
are regularly used even in O. Latin authors, and in Plautus we find dic in
questions when the next word begins with a consonant, e. g. dic quid est, but
dice is the form employed where there is anything of a pause after the word;
cf. Mil. 256 dice, monstra, praecipe, and especially Rud. 124:

tu, siquid opus est dice. Dic quod té rogo.

Abdiice, adduice and other Compounds of duco are still found in Terence before
a vowel, abdiic, &c. before a consonant, while face is the form employed at the
end of a line; édice in Virgil (‘antiquitatis amans’ Diom. p. 349. 30 K.) in
Aen. xi. 463 :

tu, Voluse, armari Volscorum edice maniplis.

The Compounds however of fucio, whose short penult was not so favourable to
Syncope (ch. iii. § 1g. p. 173) retain the -e, confice, affice, infice, &e. Catullus
has inger mi (xxvii. 2) : inger micalices amariores, from which we may perhaps
explain misc sane on an old Praenestine cista with a kitchen-scene (Mél. Arch.
1890, p. 303) as misc(é) sune from *misco, an earlier form of misceo (see § 28).
The Interjection em (ch. x. § 19) [e.g. em tibi, ‘take that’ (with a blow), em
ergo hoc tibi, &e., in the Comedians] may have been originally Imperative of
émo, which in O. Lat. (§ 46) meant ‘to take’ (thematic *eme, to judge from
the Ind. emis, emit, &c.), just as the Conjunction vel (ch. x. § 4) was the

 

' Sins of the Carmen Arvale (quoted in ch. vi. § 55) is a very doubtful form.
§§ 59-62. ] THE VERB. DEPONENT. 519

Imperative (athematic) of vélo (I.-Eur. *wél-mi) (cf. Umbr. heris .. _heris,
‘either... or,’ lit. ‘do you wish .. . do you wish’ ?).

§ 59, Other examples of Imper. in -t6d. The final -d is retained in Oscan,
e.g. 3 Sg. deivatud ‘jurato,’ estud, likitud ‘licéto,’ and in Early Latin inscrip-
tions, ve. g. 2 Sg. stutod, ‘sistito’? on the Dvenos bowl [a Future Imperative if
we are right in interpreting dienoine med Mano statod ‘on the ninth day set me
(with an offering) for Manus’); 3Sg. wiolatod, licetod, datod with exuehito, eaferto,
cedito ‘caedito,’ on the Spoletium inscription (C.I. L. ii. 4766), estod, licetod with
Sundatid, proiecitad, parentatid on the Luceria inser. (C.I.L. ix. 782);
but in class. Latin, as in Umbrian, -d was by the phonetic laws of the
language lost after a long vowel (see ch. ii. § 137), e.g. Lat. esti, licétd, fertd,
habeto, Umbr. fertu, futu ‘esto,’ habetu. The curious forms in -tid and -tad
on the Luceria inscr. (in hoce loucarid stireus ne[qu]is fundatid neue cadauer
proiecitad neue parentatid) are dialectal (Subj., with i for é in -tid?), A Third
Pl. form with -d, suntod, occurs on the Spoletium inscription. (On Umbrian
*-t0-ti in 2, 3 Pl., see § 73.)

Deponents sometimes show -to for -tor, e.g. nitito (Cic. ap. Diom. 340. : K.),
utunto [C. I. L. i. 204. (1). 8], and on the Lex Repetundarum we have the Passive
censento (i. 198. 77).

§ 60. Imper. Pass. 2, 3 Sg.in-mino. This is an O. Lat. form, found as

2 Sg. in Plautus (e. g. progrédimino, Pseud. 859: tu spectaté simul,

si quo hic gradietur, pdriter progredimino),

and Cato (praefamino, R. R. cxli. 2 Ianum Iouemque uino praefamino, sic dicito)
(ef. Paul. Fest. 62. 10 Th. ‘famino’ dicito), and as 3 Sg. in early legal Latin ;
antestamino in XII Tables: si in ius uocat, ni it, antestamino, igitur em
eapito; fruimino in the Sententia Minuciorum of 117 B.c. (C. I. L. i. 199. 32
quei ... non parebit, is eum agrum nei habeto niue fruimino) ; profitemino in
the Lex Julia Municipalis of 45 B.c. (i. 206, 11. 3, 5, 8, Ir).

The corresponding suffix in Umbrian is -mu 3 Sg. (i.e. -md, from -*mnid ?,
ch. v. § 13), e. g. persnihi-mu ‘ precamino,’ with 3 Pl. persnihi-mumo. In Osean
the Passive ending -r appears in 3 Sg. censamur (i.e. -mor) ‘ censemino,’ but
the so-called Latin 2 Pl. Imper. Pass. in -minor is a fiction of the grammarians
(see Madvig, Opusc. p. 239.) [Cicero in the archaic language of his laws
employs appellamino (MSS. -minor) as 3 Pl. Pass. (Legg. iii. 3. 8), but that this
is a genuine old usage may be doubted.

§ 61. 3 Pl. Imperat. The grammarians occasionally offer curious forms,
amento (Sacerdos) ; probunto, docewnto (Probus; cf.3 Pl. Ind. mereunt, Commod.,
neunt, Tibull.).

§ 62. IV. THE VOICES. Deponent Verbs. Passive. Since
the Passive in Latin does not differ from the Active in the
Tense-stems so muchas in the Person-endings, it is best discussed
here, immediately before we proceed to the consideration of the
suffixes used to denote the different Persons of the Verb. In
close connexion with it goes the Middle, which had originally
a Reflexive force, e.g. Gk. réarouar like rémrm éuavtdv, or
520 THE LATIN LANGUAGE. (Chap. VIII.

a Neuter force, e.g. Gk. O¢poma, pxouar. In Latin Middle
Verbs are called ‘ Deponents,’ e.g. cingor like cingo me Thus
the I.-Eur. Middle *seq®-, ‘to follow, accompany’ (O. Ind.
sica-té 3 Sg., Gk. ére-rat) is in Lat. the Deponent seguor, sequitur
3 Sg. (O. Ir sechur, sechethar 3 Sg.). The R-endings of the
Passive and Middle or ‘ Deponent’ in the Italic and Celtic lan-
guages are discussed in § 65, where it is suggested that the Passive
R-forms may originally have been restricted to an Impersonal
use in Latin, as in Umbro-Oscan and Celtic the Impersonal
Passive with -r, e. g. Umbr. fera-r Subj., ‘ there may be carrying,’
O. Ir. do-bera-r, ‘there may be giving’ (I-Eur. *bhera-r), is by
its absence of person-ending distinguished from the Deponent
with -r, e.g. Osc. karanter ‘vescuntur,’ O. Ir. sechethar
‘sequitur.’ The Personal Passive of O. Iv. by its restriction to the
Third Person indicates its Impersonal origin. A Latin phrase like
itur in antiquam silvam, the peculiarity of which is commented on
by Quintilian (i. 4. 28: jam ‘iturin antiquam silvam’ nonne
propriae cujusdam rationis est ? nam quod initium ejus invenias ?
cui simile ‘fletur’) may then exemplify the oldest use of the
R-Passive, except that the original form would be *ir (*ei-r)
without the 3 Sg. Person-suffix of tur (*ei-td-r); and the
change from the Impersonal vitam vivitur of O. Lat. to the
Personal vita vivitur of class, Lat. is parallel to Horace’s use
of invzdeor for invidetur mihi, or Cato’s change of contumeliam
factum itur to contumelia factum itur (§ 87).

The Latin Perfect, as we have seen (§ 39), represents the
I.-Eur, Perfect Middle, its 1 Sg. -7, older -e, being I.-Eur. -ai
or -ai (O. Ind. -é), so that a Perfect like réverti (older -vorti,
-vortei; ef. O, Ind. va-vrté) goes naturally with a Present
revertor. On the other hand, the Participle in -to- (§ 92), which
properly belonged to the Preterite Passive, was often used in an
Active (or Middle) sense, e.g. Lat. céndtus, pransus, potus, so
that reversus (older -vorsus ; cf. O. Ind. vrtt4-), veversus sum are
also admissible. More questionable is vertens as the Pres. Part.
of verfor in the phrase: intra finem anni vertentis, ‘within the
current year,’ vehens of vehor, &c. In the older literary period
we find Active and Middle forms of the same verb side by
side, but by the time of the classical writers there is less freedom
§§ 63, 64.] THE VERB. PASSIVE. 521

of choice ; assentior, for example, had almost wholly supplanted
assentio in Varro's time, who tells us that Sisenna the historian
still clung to the old-fashioned assentio in giving his vote in the
senate [Z. L. fr. ap. Gell. ii. 25.9: ‘sentior’ nemo dicit, et id
per se nihil est; adsentior tamen fere omnes dicunt. Sisenna
unus ‘adsentio’ in senatu dicebat, et eum postea multi secuti,
neque tamen vincere consuetudinem potuerunt ; cf, Quint. i. 5,
13. Quintilian however (ix. 3. 7) allows both assentior and
assentio, and the latter form is found in Cicero’s Letters and in
the De Inventione ; see Georges, Lex. Wortf. s.v.]. The Middle
form of assentior (as distinguished from sentio) is justified by the
use of the Middle in verbs which express a state of feeling, e.g.
reor, véreor, wrascor, For the Greek Verb-ending -eve (e.g.
kohak-evo), used in Derivative verbs indicating ‘to play a part,’
‘to act like, we have in Latin an A-Middle, e.g. aemiil-or, -ari
from aemulus, augitror, from augur, dim%nor from dominus, poetor
of Ennius’ frank confession (Sat. 1 M.):

nunqudm poetor nisi sim podager,

from poeta and so on. Examples of Frequentative Middles are
hortor from O. Lat. *horior (3 Sg. horttur Enn.), méditor from
a lost *medor (Gk. pedopa1), tmitor (cf. tmdgo), nitor for *nivitor
from a root with a Guttural (cf. miaws, and see ch. iv. § 116).

§ 63. Impersonal use of Latin Passive. An example of an Acc. governed
by an Impers. Pass. is Ennius, Trag. 190 R. :

incerte errat dnimus, praeterprépter uitam uiuitur,

quoted by Gellius (xix. ro) in illustration of the word praeterpropter, ‘inexactly,’
‘so so,’ a word which was in his time only used in plebeian Latin (nescioquid
hoc praenimis plebeium est et in opificum sermonibus quam in hominum
doctorum disputationibus notius) ; the same construction has been seen in
Plaut. Mil. 24 (epityra estur), Pseud. 817 (teritur sinapis scelera), (but see edd.
ad loce.), Pers. 577 (uéniri hance uolo) (but cf. Plaut. fr. inc. 1. 64 G. ego illi
uenear). Without an Acc. the Impersonal Passive is very common in the older
writers, e.g. Plaut. Pseud. 273:
Quid agitur, Calidére? Amatur dtque egetur dcriter ;

Pers. 309 ut ualetur ?, 386 fucile nubitur, ‘marriage is easy’ ; Trin. 580 ibitur ; Capt.
80 quom caletur, ‘in hot weather’; Rud. 1018, &c. The Latin for ‘No
admittance’ is PRIVATVM. PRECARIO ADEITVR, a notice preserved inj

an inscription (¢. I. L, i. 1215).

§ 64. Active and Middle. Gellius (xviii. 12) remarks on the O. Lat. use
522 THE LATIN LANGUAGE. (Chap. VIII.

of Active Verbs like augeo, mito in a Neuter sense, instead of augeor, mutor, and
similarly of contemplo for contemplor, &c., and the seventh book of Nonius
contains a host of examples from the older writers, aucupo for auctipor, vago for
vagor, and soon. Quintilian (ix. 3. 6-7) remarks on the inconsistency of the
Middle form of Transitive Verbs, fabricor, piinior, arbitror, suspicor with the
Active form of a Neuter or Passive Verb, vépiilo, and mentions as parallel
forms luxiriatur and luxuriat, fluctuatur and fluctuat, adsentior and adsentio (see
§ 62). The same uncertainty with Active and Middle forms is shown for
a later period by the precepts of the grammarians, e. g. Caper (93.10 K. ructo
et nausio dicendum, quamvis quidam veteres ‘ructor’ et ‘nausior’ dixerunt.
non ‘egeo’ ... dicendum, non autem ‘egeor’ ; 93. 21 K.suffragor non ‘ suffrago’;
95. 1 K. somnio dicendum, non ‘somnior’). The Passive use of Deponents (e.g.
vereor abs te) is discussed by Gellius (xv. 13) with examples from the older
writers, and a fuller list is given by Priscian (i. pp. 379 sqq. H.). A curious
instance of Attraction is seen in the use of coeptus sum, instead of coepi, desitus
sum instead of desii, with a Pass. Inf., e.g. urbs coepta est aedificari (ef.
mitescere discordiae coeptae, Liv.), and in O. Lat. we find pitestur, poleratur,
possetur similarly used (instances in Nonius p. 508 M.), néquitur, nequitum (see
Georges, s.v.). The Neuter sense of fio (cf. Osc. filet ‘fiunt’), and its use as
Passive of fdcio (on the occasional use of facior, apparently a vulgarism, see
Georges), have changed its old Inf. fere to fiert. Ficre was used by Ennius (see
Gram, Lat. v. p. 645. 9 K.), perhaps in the line in which the spirit of Homer
related his experience of metempsychosis (Ann. 8 M.):

memini me fiere pauom (MSS. fieri),

while Cato (ap. Prise. i. p. 377.11 H.) used fitur for sit, febantur for fiebant. (On
Imperat. f7, e.g. Plaut. Pers. 38, Curc. 87, fite Curc. 89, 150, &., see Georges,
Lex. Wortf. sv.) ; so venirt is used for vénire (vénum ire) in Plaut. Pers. 577 (cf.
bereunda, Epid. 74, plicenda, Trin, 1159). As coeptus sum and desitus sum replaced
coept and desii (originally Middle formations), when these verbs were used
intransitively, so Intransitive Verbs like sdleo, gaudeo, wudeo took a Perfect of
Passive form solitus sum, gavisus sum, wusus sum, in O, Lat. also solui, gavisi, ausi
(Prise. i. pp. 420, 482 H. ; Non. 508. 27 M.), e.g. Liv. Andr.: quoniam audiui,
paucis gauisi. On the Deponent Imperatives nitito, utunto, &c., see § 59.

 § 65. V. THE PERSON-ENDINGS. The I.-Eur. person-
endings were slightly different in Primary Tenses (the Present
Ind., Future Ind., &.) and in Secondary Tenses (the Preterites
Ind., the Tenses of the Optative Mood, &c.), e. g. I.-Eur. *bhéré-ti,
‘he is carrying, *é-bhéré-t, “he was carrying,’ ‘he carried,’
*bhéroi-t 3 Sg. Opt. The Tenses of the Subjunctive Mood
seem to have taken sometimes the Primary, sometimes (in
Umbro-Oscan perhaps always) the Secondary person-endings.
In the Perfect Tense an entirely different set of endings was in
use, e.g. I.-Eur. 1 Sg. *woidit, ‘I know,’ 2 Sg. *woit-thi, ‘thou
knowest,’ and so on ; and in the Imperative the persons are often
§ 65.] THE VERB. PERSON-ENDINGS. 523

distinguished by the addition of particles, e.g. I-Eur. *bhéré-
tod, with the particle *tad, Abl. Sg. of the Pronoun *to-, meaning
‘from this” or ‘thereupon ’ (§ 57).

In Passive and Deponent Verbs, Latin departs widely from
the I-Eur. scheme of Passive (or rather Middle) person-endings
(contrast Lat. feror, sequor with Gk. pépo-pat, éo-par, fertmur,
sequimur with Gk. $epd-peda, émd-pea, and with O. Ind. bhar-é,
bhéra-mahé). Latin, as well as the Umbro-Oscan languages,
and the Celtic family, uses as the characteristic mark of its pas-
sive and deponent flexion the letter r (cf. O. Ir. sechur 1 Se.,
sechethar 3 Sg., sechemmar 1 Pl, sechetar 3 Pl., with Lat. sequor,
sequitur, sequimur, sequuntur). [In Celtic r is used in all persons of
Deponents (except 2 Pl.), but only in 3Sg., Pl. of Passives.] This r
cannot be connected with the Reflexive Pronoun *swé- (Lat. sé
Ace.), seeing that s between vowels does not become r in Oscan or
in the Celtic languages (e. g. Lat. seyuor might conceivably stand
for *seq*o-sé, but O. Ir. sechur could not). Nor does it go well
with the r of 3 Pl. suffixes in the Sanscrit Verb, and the -runt,
-re of the 3 Pl. Pft. Act. in Latin (e. g. O. Ind. 4-duh-ra, 4-vavrt-
ranta, Lat. dédérunt, dedéré), since the Sanscrit r is confined to
3 Pl. suffixes, and is used in Active as well as Passive Verbs.
Its original sense may have been impersonal [cf. Lat. itur in
antiquam silvam, ‘they go,’ ‘one goes’; originally without
any person-ending, as in Umbro-Oscan and Celtic, e.g. Umbr.
pone esonom-e ferar, ‘when there is carrying to the sacrifice,
‘when the carrying to the sacrifice takes place,’ which would
be in Lat. feratur or feretur, with person-ending -tu- (-t5-) ;
O. Ir. doberr or doberar, ‘they give, ‘one gives, Bret.
gweler, ‘they see,’ ‘one sees’] ; and the original construction of
these Impersonals Passive seems to have been with an Acc. of
the object (e. g. O. Lat. vitam vivitur; Osc., tavilas . . . sakriiss sa-
krafir avt dltiumam kerssnafs, ‘ let the jovi/ae be consecrated with
victims, but the last with banquets’ ; Welsh, Etlym gledyf coch
ym gelwir, ‘ they call me Etlym of the red glaive,’ like Lat. me
appellatur (see § 62).

This Impersonal (Passive?) governing an Acc. has been ex-
plained as the Verb-stem with the Locative suffix r (seen in Engl.
‘where,’ ‘there’) used predicatively like the i-Locative of the
524 THE LATIN LANGUAGE. [Chap. VIII.

Verbal S-stem (the Latin Inf. Act.; see § 83) in such a phrase
as hostes apparere, ‘the enemy appeared’ (Historical Inf.), lit.
‘the enemy in the action of appearing.’ On the change from,
e.g. *amd-7 amicos to amantur amici, like Horace’s invideor for
invidetur mihi, see above, § 62.

§ 66. (1) Active. 1 Sing. I.-Eur. Athematic Verbs ended
in Primary Tenses in -mi (e.g. I.-Eur, *es-mi, O. Ind. 4s-mi,
Arm. em, Gk. eiué for *éo-us, Alb, jam, Goth. im, Lith. es-mi,
O. SI. jes-mi), Thematic in -6 (e. g. L.-Eur. *bhér-6, Gk. pépa,
O. Irv. -biur for *berd, Goth. baira for *berd ; Lith. vez, * veho’).
In Secondary tenses the ending was -rh (e. g. I.-Eur. *é-bhér5-m,
O. Ind. ébharam, Gk. éepor; O. SI. nesit, ‘I carried, for *nesom),
after a consonant, -m, e.g. Gk. éde.€a for *edeg-m. The Perfect
had -% (e.g. I-Eur. *woid-i, O, Ind. véd-a, Gk. of8-a, Goth.
vait for *vaiti ; O. Ir. ro cechan ‘ cecini’).

In Latin, as we have seen, Athematic Verbs form the 1 Sg.
according to the Thematic Conjugation, e. g. Lat. é/-d for I.-Eur.
*éd-mi, vd/-6 for I.-Eur. *wél-mi (§ 2), just as in Sanscrit
Thematic Verbs take the Athematic -mi, e. g. bhéra-mi for I.-Eur.
*bherd. I.-Eur. *es-mi, however, is in Latin swm, whether
originally *sém% with Syncope of final i (ch. iii. § 37) or merely
*sdm, it is impossible to determine (cf. Osc. sum). The curious
1 Sg. inguam 1, with the other Persons formed froma Sg. zaquio,
and with Imper. ingué, resembles a Subjunctive in form, ‘I
should say,’ or ‘I shall say, or ‘let me say’ (see § 55). The
Secondary Tense-ending -m appears in Imperfects and Pluperfects,
e.g. éra-m (cf. O. Ind. 4-ya-m, ‘I went, Gk. é-dpd-v, Goth, id-
dja, ‘I went’), dma-ba-m (cf. O. Ir. ba for *bam, ‘I was’), dmav-
éra-m, in Optative forms, e.g. si-m, O. Lat. sie-m (O. Ind. siyaim,
syi-m, Gk. ein-v), dmdv-eri-m, tima-ssi-m, faxi-m, and in Sub-
junctive, eo. dme-m, dmdvisse-m, videa-m, féra-m (also used as
Fut., § 36), fauia-m (cf. O. Ir. do-ber for *-ram, O. SI. berg,
used as Pres. Ind.). But the ‘ Future-Subjunctives’ (see § 53)

take the Thematic ending, e.g. é-o for *és-d (Zend. aha,

 

‘ Explained as *ind-(s)quam from the root seqt- (ch. iv. § 158), or as
*ind-(ve)quam from the root weq?-.
r

§§ 66, 67.] THE VERB. PERSON-ENDINGS. 525

Hom. Gk. & for *éow), cimdv-ér-6, tima-ss-0, fax-d, like fér-d, std
for *stay-o (Umbr. stahu).

In the Perfect, the ending of the Active Voice has been re-
placed in Latin by the Middle ending -ai (-ai), which became in
the unaccented syllable -ez, then -2 (ch. iii. § 18), ¢#tiud-2 (O. Ind.
tutud-é), déd-i (O. Ind. dad-é), vid-7 from *veidei, a Middle form
which survives in the O. Sl]. védé, ‘ I know, for *woidai (O. Lat.
Sect, poseinei, conquaeisiuet, redidet, all on the milestone of Popi-
lius of 132 Bo, C. 2 ZL. i. 551). Lat. reverti, assensi, &c. are
thus really Middle forms, and go suitably with Pres. revertor,
assentior.

§ 67. 2Sg. The J.-Eur. endings are -si (e.g. [.-Eur. *é-si
and *és-si, O. Ind. 4-si, Gk. e? for *é-o1, which with the addition
again of the suffix -s became «fs, Hom. éo-oi; I.-Eur. *ei-si from
the root ei-, ‘to go,’ O. Ind. é-8i, Gk. ef for *ei-o.; O. Ind.
bhara-si, O. Ir. beri. Goth. bairi-s), -s (e.g. I.-Eur. *é-bhéré-s,
O. Ind. 4-bhara-s, Gk. €-epe-s, O. Ir. do-bir; O. SI. veze
‘vexisti’; and in the Present Tense, Gk. riéy-s, Dor. pépe-s),
in the Perfect -thi (e.g. I.-Eur. *woit-tha, O. Ind. vét-tha, Gk.
oicda; cf. O. H.G. gi-tars-t). In Latin we have -s in és, ‘ thou
art,’ for *éss (scanned as a long syllable in Plautus), whether
from older *éssi (ch. iii. § 37) or not, it is impossible to say, dgi-s for
*agé-s (if from an original *age-si, like O. Ir. beri from *bhere-
si, the final -i must have been dropped before the fourth cent.
B.c. when s between vowels became 7; see ch. iv. § 146), s-s
(O. Lat. sié-s), agd-s, while in the 2 Sg. of the Perfect Tense we
have -st2, e. g. dédi-sti, the final vowel of which (O. Lat. -e7, e. g.
gesistei on a Scipio. Epitaph of c. 180 B.c., C. I. L. i. 33), like the
i of 1 Sg. déd-i (O. Ind. dad-é), must represent an original -ai
(-ai), while the -s reminds us of that -s- which so often appears
in the endings of the second Person (e.g. Gk. épn-o8a, ridn-o8a,
20€dy-09a, Bddou-o8a). (For another theory which regards vidis-
of vidisti as a Verbal Noun-stem, see § 52.)

In the Imperative, the bare stem is used, as the bare stem is
used in Vocatives Sg. of Nouns, e.g. (Thematic) age from the
Thematic Verb-stem *48%8-, *40-, ‘to lead,’ as *4gé is Voc. Sg.
of the Thematic Noun-stem *#é-, *iS5-, Gk. ayds, a leader,
526 THE LATIN LANGUAGE. [Chap. VIII.

Lat. prad-tgus; (Athematic) 2 for *e: from the Athematic Verb-
stem ei-, ‘to go’ (see § 2). Sometimes the particle *tod (Abl.
Se. of the Pronoun-stem té-, ‘from this,’ ‘ thereupon’) is added,
when the command refers not to immediate action, but to action
after something shall have happened, e. g. quum venerit, scribito,
‘when he comes, write,’ lit. ‘write thereupon.’ (So in O. Ind.,
e.g. 2 Sg. vit-tat for I.-Eur. *wit-tod from the root weid-, ‘to
know’; see § 57.) The final -2 of O. Lat. diice, dice, face [for
*fact (ch. iii. § 37), from the stem facyo-, faci-, § 16] is dropped
in classical Latin, through its frequent use in word-groups like
dic(e) mihi, &e. So in Catullus (xxvii. 2) inger mi for ingere mi
(see § 555 ch. iii. § 36).

§ 68. The athematic 8g. of féro, vilo would be *fers, *vels, which would become
*fer(r), *vel(1) (ch. iv. § 153). The former word was changed to fer-s by the addi-
tion of the 2 Sg. suffix -s, as in Greek e? for *éor was made és, thou art ; for the
latter was substituted the word vis, apparently 2 Sg. of a root wei-, ‘to wish’

O. Ind. vi-, 2 Sg. vé-3i, Gk. fepar).

Whether dices of the MSS. in Plaut. Trin. 606 non credibile dices, is a relic of
the old spelling of dicis is doubtful. It may be Future.

§ 69. 88g. The Primary Tense-ending in I.-Eur. was -ti (e. g.
*és-ti, O. Ind. As-ti, Gk. éo-r1, O. Ir. is, Goth. is-t, Lith. és-ti and
és-t, O. SI. (Russ.) jes-ti; Dor. Gk. di30-r1, Att. 5/3e-c1, riOn-or,
&c.), the Secondary Tense-ending was -t (e. g. *é-bhéré-t, O. Ind.
a-bhara-t, Gk. €qepe for *é-pepe-r), and the ending of the
Perfect -& (e.g. *woidé, O. Ind. véd-a, Gk. of8-e, Goth. vait;
O. Ir. ro cechuin ‘cecinit’). In Latin all trace of the -i of
I.-Eur. -ti has been lost (e.g. Lat. es-t, fer-t, digi-t for *agé-t),
but in Osean we find -t for L.-Eur. -ti, and -d for I.-Eur. -t.
On the very oldest Latin inscriptions we find -d for the
Secondary ending, but certain instances of Primary endings are
unfortunately wanting. Thus the Praenestine fibula has fe/aced,
‘fecit,’ the Dvenos inscription has feced, sied (but mitat, ap-
parently 3 Sg. Subj. used as Fut.), and in Oscan we have deded
‘dedit,’ kim-bened ‘con-vénit,’ fusi-d ‘foret,’ deivaid ‘ juret,’
heriiad ‘velit,’ while Primary Tenses show -t, e.g. faamat
‘habitat. These endings had been ‘levelled’ to ¢ in Latin
before the second century B.c.; for a Praenestine cista, not of
the same antiquity as the fibula just mentioned, has dedit beside
feel (C.I.L. i, 54 Dindia Macolnia fileai dedit. Nouios
§§ 68-70.] THE VERB. PERSON-ENDINGS. 527

Plautios med Romai fecid), and in all other old inscriptions we
have invariably -d, e. g. zousit (ii. 5041, of 189 B.c.), wedet ‘ vellet,’
eset, fuit, censuit (all on the 8. C. Bacch. of 186 B.c., i. 196),
cepet on the (restored) Columna Rostrata (i. 195), fwet, cepit,
dedet on one of the oldest Scipio Epitaphs (i. 32.).

In the Latin Perfect the original ending seems to have been
-eit, written in the very oldest inscriptions -ed (with that use of ¢
to express the e7-sound which we have seen in Nom. Pl. plotrume
on a Scipio Epitaph, Dat. Sg. Diove Victore, ap. Quint. i. 4. 17 ;
see ch. iv. § 34), then -e¢, -eit, and in class. Lat. -i¢. Before
a final -t a long vowel was shortened in the course of the
second cent. B.c. (ch. ili. § 49), so that in class. poetry this -d¢
of the Perfect is a short syllable. But in Plautusit is invariably
scanned long, unless shortened in iambic words, &e. by the Law
of Breves Breviantes (e. g. dédit may be scanned dedét, but only
vidi, fecit, &c.). The long quantity is found in every type of
Perfect in Plautus (see § 39), in viwtt, habuit, adnumeravit, as
well as in Perfects proper like vic#t. This -ei¢ can hardly have
been anything else than the I.-Eur. ending of the 3 Sg. Mid. of
the Perfect Tense (which was, like the 1 Sg. Mid. ending -ai,
adiphthong weakened in the unaccented syllable in Latin to -e7,
class. -7, ch. iii. § 18), augmented by the 3 Sg. Act. Secondary
suffix -f. So that all trace of the I.-Hur. 3 Sg. Act. ending -é is
lost in Latin. In the Oscan Perfect, however, the ending -ed,
e. g. deded ‘ dedit, prafatted ‘ probavit’ (quasi *probassit, § 3),
aamanaffed ‘-mandavit’ (cf. d<der, avaFaxer, Pel. afded < abiit’)
is usually referred to the I.-Eur. Preterite ending -ét, or to the
Perfect. Active -% augmented by -t, though whether Umbrian -d
(fefure for *fefured, dede for *deded) was dropped after a short
vowel, so readily as after a long vowel (like Latin -/) is doubtful.

In the Imperative the same form is used as the 2 Sg. ‘Future’
Imper. e. g. es-¢0(d), on which see above, § 67.

§ 70. The 3 Sg. Pft. in Latin. We have found (ch. ili. § 49) that the
shortening influence of final -¢ on a preceding long vowel is already seen in
the poetry of Ennius (239-169 B. c.), who scans, ¢. g. mandebdéit beside ponebddt,
splendét beside jubét, potesstt beside esset, though it probably does not appear in
Plautus (c. 254-184 8. c.), that Terence (195-159 B. c.) follows the same usage

as Ennius, while Lucilius (148-103 B. c.) scarcely ever allows a vowel before
final -¢ to retain its original length. It is thus the versification of Plautus
528 THE LATIN LANGUAGE. [Chap. VIII.

which must decide the original quantity of the vowel in the 3 Sg. Pft. ending
-it, Indubitable instances of -it in Plautus are émit, Poen. 1059:
emit, et is me sibi adoptauit filium ;
vintt, Pseud. 311:
ilico uixit amator, abi lenoni supplicat ;
which are the readings of both families of MSS., while we have vicit, Amph. 643 :
uicit et domtim laudis cémpos reuénit (a bacchiac line) ;
adniméravit, Asin 501:
adnimerauit et crédidit mihi néque deceptust in eo,

where the evidence of the Ambrosian Palimpsest is wanting. (For other
examples, see Miiller, Plwut. Pros. p. 71.) The instances of -i in Plautus,
except where the Law of Breves Breviantes operates (e. g. dédit like dédi, abi,
‘bit, dbis, &e.), are very few and uncertain (e. g. dixit, Pers, 260); and even if
it were true, as it almost certainly is not, that this seansion occurred now
and then in Plautus, it would only prove that the shortening influence of
final -t showed itself even earlier than Ennius, for there is no indication of
one type of Perfect having had -it and another type -t. In other writers we
have, e.g. excidit Naev. Tray. 5 R., dédit Ter. Eun. jor, stétit, Phorm. prol. 9,
crissavit Lucil. ix. 70 M., and the long quantity is found after 7 in the com-
pounds of eo in Ovid, e. g. interiit, abiit, rediit (see Lachmann and Munro on
Luer. iii. rog2 ; and ef. interieisti, C. I. L. i. 1202).

The spelling in the very oldest inscriptions is -ed : fefaced on the Praenestine
fibula (C. I. L. xiv. 4123 Manios med fefaced Numasioi, ‘Manius me fecit Num-
erio’), feked (or feced ?) on the Dvenos bowl (Zvet. I. I. 1. 285 Duenos med feked).
This can hardly be equated with Osc. -ed, since the other spellings -eit, -it point
to this earlye being merely that symbol of the e/-sound which is often found in
old inscriptions (ch. iv. § 34), derivable from an I.-Eur. -ai, for which in Oscan
we should expect to find a diphthong rather than the simple vowele. Other
old spellings are -et, -cit, -id, and -it. Examples of -et are fuet and dedet (beside
cepit) on a Scipio Epitaph, perhaps of the end of the third cent. B.c. (C.I. L.
i. 32), dedet in two old inscriptions with Ablatival -d (i. 63 de praidad Maurte
dedet ; i. 64 de praidad Fortune dedet) ; and this is the spelling adopted in
the (restored) Columna Rostrata (i. 195 exemet...cepet.. . ornauet). The classi-
cal spelling -it occurs as early as the Scipio Epitaph, just mentioned, with
cepit ; the dedicatory tablet of Minucius, 217 B.u., with vovit (i. 1503 Hercolei
sacrom . M. Minuci C. f. Dictator uouit) ; aScipio Epitaph of c. 200 B. c. (i. 30)
with fuit, cepit; the dedication of Aurelius, 200 B.c. (Not. Scav. 1887, p. 195),
with didit, probauit ; the decree of Aemilius Paulus of 189 B. c. (C. I. L. ii. 5041)
with decreiuit, iousit ; and the contemporary decree of Fulvius Nobilior 189 B.c.
with cepit (i. 534 Aetolia cepit, ‘took from Aetolia’); the S.C. Bacch.
186 B.C. (i. 196) with censwit; while we have both -it and -id on the old
Praenestine cista (end of third cent. ?) quoted above, with dedit, fecid, But -eit
is not common, e. g. probaveit (with coeravit) (i. 600, of 62 B. c.), fuueit (i. 1051),
redieit (i. 541, of 145 B.C.), venieit (i, 200, 58, &e., of 111 B. c.), so that if we had
only the spellings of inscriptions to guide us, and not the versification of the
early poets, we should be inclined to suppose the original form of the 3 Sg. Pft.
suffix to have been -ét, -é¢, which, with the usual change of unaccented é to i,
became about the end of the third cent. -it (-id). It has been suggested that
in some types of Perfect, e. g. dimit (ef. Gk. eege(7)), fidit (cf. O. Ind. d-bhid-
§§ 71-73.) THE VERB. PERSON-ENDINGS. 529

At; but see § 39), the final syllable was originally short, while in other types
it was long. But the versification of Plautus points-to no distinction having
been made between the different types in his time at least. The pronuncia-
tion of his age must have been diatt, fidit, amarit as well as tutudit, vidit, fecit,
and this 7-sound can hardly be dissociated from the final -% of 1 Sg. tutudi, &c.
This -7 of tutudi we have seen to be the I.-Eur. -ai, the ending of the r Sg. Pft.
Mid. (O. Ind. tutud-8). In the 3 Sg. of the Perfect Middle, Sanscrit shows
a similar form to the 1 Sg., viz. tutud-é. We are therefore led to suppose
that in Latin as in Sanscrit *tutudai was the original form both in 1 Sg. and
38g. This *tutudai, which would become in Latin *tutudei, then tutudi, was
in the 3 Sg. discriminated by the addition of the 3Sg. suffix used in Secondary
Tenses in the Active Voice, -t, and became tutudei-t (written in the old ortho-
graphy tutudé-t), then tutudi-t, then in the second cent. B.c. tutudit. The other
types of Perfect followed in the 3 Sg., as in all other persons, the type of
I.-Eur. Perfects like tutudi.

§ 71. 1 Plur. In Lat. we have in all tenses the ending *-més
(class. -mis), while in the other I.-Eur. languages we have
a variety of endings, e.g. Gk. gepo-pev, épépo-uev, Dor. pépo-pes
(this I.-Eur. *-més varied with *-mds, the Latin person-suffix),
O. Ind. 4-bhara-ma, Vedic vid-ma. The scansion -mds in Plaut.
is illusory (see Miller, Plaut. Pros. p. 57).

§ 72. 2Plur. The ending -té of Gk. pépe-re, &c. appears in
Latin only in the Imperative, e.g. fer-te, agi-te for *ayé-te, &e.
Elsewhere it was replaced by -¢7s (older *-¢és), an ending like
the 2 Dual ending (with th- apparently) of O. Ind. bhara-thas,
Goth. baira-ts, the use of which discriminated agitis Ind. from
agite Imper., as agis Ind. differed from age Imper. Corresponding
to the 2 Sg. ‘Future’ Imperative in -/6 (older -t07), we have
a 2 Pl. ‘Future’ Imperative in -d¢e, apparently composed by
adding to the Sg. form the 2 Pl. suffix -té, -td¢e (e.g. estdte),
for -tat-te (ch ii. § 127).

§ 78. 8 Plur. The I.-Ewr. suffixes end with - in Primary but
notin Secondary Tenses. Thus -enti, -nti, -ntiare Primary, -ent,
-nt, -nt are Secondary, e.g. *s-énti (O. Ind. s-dnti, Gk. fot for
Dor. é-vri, O. Ir. it, O. W. int, Goth. s-ind; so Umbr. s-ent,
Ose. s-et for *s-ent), *bhérd-nti [O. Ind. bhéra-nti, Arm. beren,
Dor. Gk. $épo-vrt, O. Ir. berit, Goth. baira-nd, O, Sl. (Russ.)
beratY], *é-bhérd-nt (O. Ind. 4-bhara-n, Gk. é-epo-v). In class.
Lat. the ending both for Primary and Secondary Tenses is -z#,
e.g. feru-nt (older *fero-nt, e. g. cosentiont on a Scipio Epitaph,

Mm
530 THE LATIN LANGUAGE. (Chap. VIII.

O.L.L. i. 32), fereba-nt, fera-nt, si-nt [Umbr. sims; but O. Lat.
sient, either by analogy of Sg. siem, sies, siet (see § 55), or with
-ent for -nt, *siynt], s-wnt (older sont, C.D. i. 1166). But in
Umbro-Oscan there is a distinction. We have -nt for I.-Eur.
-nti, but -ns for I.-Eur. -nt, e.g. Umbr. sezt, Osc. set for *sent in
the Ind. of the Substantive Verb, but Umbr. szzs, in the Opt.,
Ose. préfattens ‘ probaverunt,’so that itis likely that at some early
period Latin, like the other languages of Italy, distinguished
Primary -nti and Secondary -nt. The 3 Pl. ¢remont: in the
fragment of the Carmen Saliare, ascribed to Numa’s time, is
a doubtful reading [Ter. Scaur. 28.9 K. Cum .. . quoniam antiqui
pro hoe adverbio ‘ cuine’ dicebant, ut Numa in Saliari carmine :

cuine tonas (MSS. ponas), Leucesie, prae tet tremonti (MSS. praetexere
monti)],

though it is contirmed by another corrupt passage of Festus
(244.17 Th.; he is quoting from the Carmen Saliare): ‘prae tet
tremonti’ (MSS. pretet t.) praetremunt te (MSS. praetemunt pe).
In Old Latin we have a curious form in -zent (older -nont) in
the Pres. Ind. only, e.g. dd-nunt, eaplé-nunt, prodi-nunt, néqui-
nont, which has been explained on the theory that the 3 Pl.
of the Pres. Ind. had once ended in -n, *dién, *eaplén, *prodin,
*nequin, and that these forms were expanded by the subsequent
addition of the Thematic Secondary ending -dzzd, later -#¢, much
as Gk. ef for *éo., 2 Sg. of edui, was by the addition of the
Secondary suffix -s expanded to «f-s, or O. Engl. sind, ‘ they are,’
to sind-un. If it were a mere case of Nasalization like tu-n-do,
li-n-o, *std-n-o in destino (?), &c., Lith. eint, ‘I go’ (§ 10), it is
difficult to see why it should be confined to this single person,
the third person plural of the Present Indicative Active. But
how *danti could become *dan, *elcs-plenti become *eaplen, has
not yet been satisfactorily shown (see J. F. ii. 302).

L-Eur. -nti, -nt (e.g. O. Ind. déd-ati, Gk. AeAdyy-acr for -nti)
does not appear in Latin, unless possibly in O. Lat. scent (see
above), But in Umbro-Oscan -ent, -ens [I.-Eur. -ent(i) or
-nt(i) | is as universal as -wut [I.-Eur. -ont(i) ] in Latin, e. g. Ose.
fiiet ‘ fiunt,’ prifattens ‘ probavernnt, Umbr. furent ‘ erunt,’ Ose.
censazet for -ent ‘ censebunt.’
§§ 74, 75.] THE VERB. PERSON-ENDINGS. 531

The ending of the 3 Pl. of the Latin Perfect -2rwnt is to be
compared with the O. Ind. 3 Pl. endings with -r- of various
Tenses (e.g. 4-duh-ra Pret., duh-raté Pres., bharé-rata Opt.,
duh-rim and duh-ratim ee 5 very rarely -ranta, in 4-vavrt-
ranta. The O. Ind. Perfect has in 3 Pl. Act. -ur, e.g. dadtr,
‘they have given,’ and in 3 Pl. Mid. -ré, e.g. dadiré, represent-
ing probably I.-Eur. -r and -rai). The byform -éré is in O. Lat.
-ert (dederi, C. I. L. i. 187, probably from Praeneste : M. Mindios
L. fi. P. Condetios Ua. fi. aidiles uicesma parti Apolones dederi).

In the Imperative we have in Latin, as in Greek, the particle
*tod, ‘from this, ‘thereupon,’ added to a 3 Pl. form, e.g. Lat.
Serunto for *feront-tod, Gk. pepdvtw for *pepovt-tws. The final
-d is seen in suntod on the Spoletium inscription (C. I. L. xi. 4766).
The Umbrian ending appears to have been -tdta, e.g. etuta and
etuto ‘ eunto.’

§ 74. 3 Pl. Pres. in -nunt. Ddnunt for dant is quoted from the older poets
by Nonius, 97. 13 M., e. g. Caecilius, Com. 176 R. patiére quod dant, quando
optata non danunt (ef. Paul. Fest. 48. 18Th. ‘danunt’ dant) ; itis often used by
Plautus (the references are given in Neue, Formeni. ii. p. 412), and is found on
a Saturnian dedicatory inscription of two brothers, money-lenders, called
Vertuleius (C. I. L. i. 1175, Sora:

dénu danunt Hércolei maxsume méreto) ;

prodinunt for prodeunt is quoted by Festus (284. 22 Th.) from Ennius (A.
158 M.):
prodinunt famuli; tum candida lumina lucent ;

so obinunt for obeunt (id. 214. 4 Th.), redinunt for redeunt (id. 400. 12, a passage
badly preserved in the MS8.), quoted from Ennius [possibly with mention of
inunt for eunt, so that the Philoxenus Gloss (p. 75. 23 G.), int : wopevovra: may
be a corruption of inunt: mopevovra (see § 2)]; nequinont for nequeunt (Fest.
162. 24 Th.), quoted from the Odyssea of Livius Andronicus :

partim érrant, nequinont Graéciam redire,

and paralleled with ferinunt (MS. fernunt) for feriunt (MS. fereunt; cf. 400.
14 Th.), and solinunt for solent [rather for solunt, the obsolete verb of which
consulo is a compound; ef. Fest. 526. 14 Th. ‘solino’ idem (Messala) ait esse
consulo]; explenunt for explent by Paul. Fest. (56. 14 Th.); inserinuntur for
inseruntur is used by Liv. Andr. (ap. Fest. 532. 24 Th.): millia alia in isdem
inserinuntur.

§ 75. 3 Pl. Perf. The isolated forms dedro on an old inscription of Pisaurum
(C. I. L. i. 177 Matre Matuta dono dedro matrona; beside dedrot on another
inscription from the same place, i. 173 Iunone re. matrona Pisaurese dono
dedrot) and emeru on an inscription of Cora (i. 1148 Q. Pomponius Q. f. L.
Tulius Ser. f. praitores aere Martio emeru) cannot be taken as a proof that -16

Mm 2
532 THE LATIN LANGUAGE. (Chap. VIII.

was a byform of -7i in the Latin 3 Pl. Perfect. For the dropping of final
consonants is a feature of the Latin of Pisaurum (see the inscriptions quoted
above with -s, -r dropped ; and cf. the loss of -t in dede, i. 169, and apparently
of -nt in i. 177 M’. Curia, Pola Liuia deda), and emeru on the Cora inseription
may be a similar dialectal variety, or merely a graphic contraction for eme-
runt. Final -nt often loses the dental on late inscriptions and is written -n or
-m; thus we have fecerun and fecerum (also feceru, vi. 24649) in plebeian inscrip-
tions of the Empire (see ch. ii. § 137). Some Roman grammarians called the
form with -re the ‘Dual’ form, a theory which is rightly rejected by Quin-
tilian, though his own explanation, that it is a weakening of -runt, cannot
stand (i. 5. 43 quanquam fuerunt qui nobis quoque adicerent dualem ‘scrip-
sere’ ‘legere’: quod evitandae asperitatis gratia mollitum est, ut apud
veteres pro male mereris ‘ male merere’) (cf. Serv. ad A, ii. 1). In the older
writers -runt and -re seem to be used at will, e. g. Plaut. Trin. 535 alii éxolatum
abiérunt, alii emértui, Alii se suspendére, and Cicero (Orat. xlvii. 157), quoting
a line of Ennius, says : nee vero reprehenderim ‘scripsere alii rem’ ; ‘ scrip-
serunt’ esse verius censeo, sed consuetudini auribus indulgenti libenter obse-
quor. (Ennius, however, seems to prefer -erunt to -ere in his Annals ; Terence
prefers -ere.) The older spelling -ront, mentioned by Quintilian (i. 4. 16), who
quotes dederont and probaveront, is found on early inscriptions (see Index to
CLL...

The ending -ré (older -r) seems to come from an I.-Eur. -ri, cognate with the
I-Eur. -rai of the O. Ind. 3 Sg. Pft. Mid. -ré, e. g. dadiré, ‘dedere’; the ending
-runt (older -ront), either from an I.-Eur. -ront (ef. O. Ind. 4-vavrt-ranta), or
from a subsequent addition of the usual 3 Pl. Thematic suffix -wnt (-oné) to
a3 Pl. Pft. in-r. Thus *dedér (a ‘doublet’ of dedére ?) would become dedér-unt
by the same process as we have supposed an earlier *explén to have become
explén-unt. In view of the presence of r in these O. Ind. third persons plural
this is a more likely explanation than to suppose that* is the Latin substitute
for intervocalic s. ,

With regard to the quantity of the ¢ in -erunt, the short quantity, though it
is in the classical and later period more prominent in Dactylic Poetry than in
other verse, owing to its suitableness for the dactylic metre, is not by any
means unknown in the older (and later) dramatists, e.g. in Plautus subegérunt,
Plaut. Bacch. 928, fecérunt, Amph. 184 locavirunt, Pers. 160, cessarunt, Mil. 1432,
&e. ; emérunt, Ter. Eun. prol. 20, conlocarunt, ib. 593). Plautus appears, how-
ever, to use it only at the end of a line or hemistich, so must have regarded
its use as a licence to be resorted to under metrical necessity. It does not
appear to have been used in Tragedy, nor by the earlier Epic writers, like
Ennius, and not very frequently by Lucretius (e. g. institérunt, i. 406; see
Munro’s note), which points to its having been a pronunciation of colloquial
Latin that won its way only gradually into the higher literature. It is gene-
rally explained as « 3 Plur. of the Auxiliary stem es-, ‘to be,’ and is com-
pared with Gk. -ecav of 3 Plur. Plupft., so that dedérunt from *dedesunt would
be a quite different formation from dedérunt and dedére, with I.-Eur. 1. An-
other theory makes it *dedis-ont, the first part being a Verbal Noun-stem
*dedis- (see § 52). [On curarunt for curavérunt beside the older coirawerunt (é?)
on inscriptions, see § 48.]

§ 76. (2) Passive (Deponent). 1 Sing. The Italo-Celtic
§§ 76-78.] THE VERB. PERSON-ENDINGS. 533

ending was -dr in the Pres. Ind. [e. g. O. Lat. sequdr, class. seguir
(ch. iii. § 49), O. Ir. sechur], apparently an addition of Passive
-r to the Active ending -6. Those Tenses and Moods which in
Latin formed their 1 Sg. Active in -m substitute in the Passive
-r for -m,e. g. fera-r Fut. and Subj., fereba-r Impft., except in the
Perfect group, where a periphrastic form is used, e. g. /dtus sim,
latus essem, not *tulerir, *tulisser, also datus ero, not *tuleror (see
§ 54), though in Oscan we do find this adaptation of the Active
forms in comparascuster, Fut. Pft. Pass. (pon ioc egmo comparas-
custer ‘cum ea res consulta erit’’), the Active of which would be
*comparascust (§ 89); but in Umbr. we have pihaz fust ‘piatus
erit,’ &c.

$77. 2 Sing. Since O. Iv. sechther ‘sequeris’ shows the
I.-Eur. (athematic) ending -thés (e.g. O. Ind. 4-di-thas, Gk.
é-56-Ons), which is retained in the O. Ir. Imperative (e. g. cluin-
te,‘do thou hear, Dep.), with the usual ending -r, we might
expect to find in Latin the I.-Eur. (thematic) ending -si (e. g.
Zend. bara-rwha, Gk. épeo for *-co, é-pépe-o), which is retained
in the Latin Imperative (e. g. séguéré for *sequesd, § 57), with an
appended -r, *seguesdr, in class. Lat. *sequertir, But the forms
actually found are (1) sequeré, the same as the Imperative 2 Sg.
(this is the usual form in O. Lat., and even in Cicero); (2)
sequerts, which adds to this the ending -s of the 2 Sg. Act.,
*sequeré-s becoming sequerts (ch. ii. § 18). The addition of this
-s discriminates the Ind. from the Imper. form as dys differs
from dgé, dgitis from dgité. The ‘ Future’ Imperative changes to
-r the -d of its 2 Sg. Act., e.g. fer-tor 2 Sg. Pass., ferto from fer-
tod 2 Sg. Act. [On the O. Lat. ending for the 2, 3 Sg. Imper.
-mino, formed apparently from 2 Pl. -mtnt on the type of 2, 3 Sg.
Act. -t6(d), see § 60, and on forms like atito for utitor, § 59.]

§ 78. For statistics of the use of -re and -ris, see Neue, ii, pp. 393 sqq.
Terence uses -re only, Plautus both -re and -ris, but -re far more frequently.
(He puns on obloquere Ind. and obloquere Imperat. in Cure. 41.) Cicero prefers
-rein all other tenses than the Pres. Ind., and in this tense too in Deponent
Verbs. Quintilian is wrong in supposing -re to be a weakening of -ris (i. 5. 42
quod evitandae asperitatis gratia mollitum est, ut apud veteres pro male
mereris ‘male merere’), for -is did not become -e in Latin (ch. ii. § 137), nor
can an isolated spelling like tribunos [milita]re for militaris on an old inser. (C. I. L.
534 THE LATIN LANGUAGE. [Chap. VIIL.

i. 64) be quoted as a proof of this change. The form -rus on a few inscriptions
(spatiarus, C. I. L, i, 1220, Beneventum ; utarus, i, 1267, Venusia ; figarus, iv.
2082, Pompeii), none of them old, may be a mere dialectal or vulgar variety,
but it may also (like -us in the Gen. Sg. of the grd Decl., ch. vi. § 22) be a
genuine tradition of an older form, which arose from the addition of -s to
*sequesd, &c., at a stage prior to its weakening to sequeré. Thus *spatid-sé would
be expanded to *spatidsé-s, which would become *spatiaris, spatiarus,

§ 79. 8 Sg. The Italo-Celtic ending is -tér [e. g. Lat. ségui-
tur from *seque-tor, O. Ir. sechethar ; cf. Osc. sakarater ‘ sacra-
tur, with -ter from syncopated -t(5)r as Umbr. ager from *agr(5s)],
formed by adding Passive -r to the I-Eur. Secondary ending -té
(e.g. O. Ind. 4-di-ta, Gk. é-b0-r0; O. Ind. 4-bhara-ta, Gk.
é-pépe-ro). The Imperative changes to -r the -d of the Particle
-téd which it appends to the bare stem in the Act., e. g. fer-dor
Pass., fer-to(d) Act., dgi-tor Pass., agi-ta(d) Act. (On O. Lat.
-mind, e.g. antestamino, ‘let him take to witness,’ see § 60, and
on -to for -¢or in atito, &e., § 59.)

§ 80. 1 Plur. The Italo-Celtic ending is -mér (e.g. Lat.
séqguimitr for *sequo-mor, O. Ir. seche-mmar with a curious
doubling of the m), formed by changing to 7 the s of the Active
-mbs (Lat. féri-miis for *fere-mds, O. Ir. do-beram for -mds ?),

§ 81. 2 Plur. Both the Celtic and the Italic languages
depart from the ordinary procedure in this person. In O. Ir.
Deponents we have the Active ending; in Latin we have the
Nom. Plur. of the old Pres. Part. Passive with ellipse of estis,
eg. férimini from *fero-menoi (Gk. epdpevor) in the Present
Tense, and analogical formations in the others, e.g. férébd-mini,
fera-mini, ferré-mint. The 2 Plur. Imper., though similar in
form to the 2 Plur. Pres. Ind., e. g. ferimini, is usually explained
as an old Infinitive (Dat. of a MEN-stem), for I.-Eur. *bhere-
menai (Gk. pepé-pevar) (see § 57).

§ 82. 8 Plur. The Italo-Celtic ending is -ntér (e.g. Lat.
séquuntir from *sequo-ntér, O. Ir. sechetar ; ef. Osc. karanter
‘pascuntur’ with -nter from syncopated -nt(d)r, like 3 Sg. -ter
for -t(6)r ; see above), formed by adding Passive -r to the I.-Eur.
Secondary ending -ntd (e.g. O. Ind. 4-bhara-nta, Gk. é-pépo-rto.
Cf. Gk. éu-mAn-vro with Lat. im-ple-ntur). In the Imperative the
§§ 79-83.] THE VERB. INFINITIVE. 535

-d of the particle -téd appended in the 3 Pl. Act., is changed to
r, eg. féruntor Pass., ferunto from *feront-tod Act. (see § 57,
and on O. Lat, censento for censentor, § 59).

§ 83. VI. THE INFINITIVE. The I.-Eur. Infinitive was
merely a Case (usually Dat. or Loc. Sg.) of a Verbal Noun, and has
best retained its character in the Celtic languages, where its object
stands not in the Acc., as after a verb, but in the Gen., as after
a Noun. The form of the Inf. varied not merely according to
the case employed, but also according to the Noun-stem which
was chosen. In O. Ind. we have a great variety of Infinitives,
e.g. (1) Dat. of a Root-stem, a stem which was the same as the
root of the Verb, e, g. -Ajé, ‘to drive, lit. ‘for driving ’ (Lat. a ;
cf. Gk. xed-av ?); (2) Dat. of an S-stem, e.g. ji-8é, ‘ to conquer ’
[ef. Lat. da-r2, O. Lat. da-sei (§ 85), Gk. wet-cat] ; (3) Dat. of
a MEN-stem, e.g, da-mané, ‘to give’ (Gk. dduevar; Lat. da-
mint 2 Pl. Imper. Pass., see § 57), vid-mané, ‘to know’ [Gk. (F)ié-
pevat|; (4) Loc. of a MEN-stem (without -i, see ch. vi. § 37), e.g.
dhar-man, ‘ to keep up’ (cf. Gk. 6d-uev) ; (5) Dat. of a TU-stem,
e.g. dha-tavé, ‘ to set’ (cf. Pruss. di-twei, ‘to give’); (6) Acc. of
a TU-stem, e.g. dha-tum (this is the classical or Sanscrit form of
the Infinitive of every verb) (Lat. con-ditum 1st Sup., Lith. dé-ty
Sup., O. Sl. dé-tti Sup.; the Balto-Slavic Supine in -tum is fused
with the auxiliary verb of the same root as Lat. fui to form
a Compound Tense, e.g. Lith. détum-bime, 1 Pl. Opt., as- the
Lat. 1st Supine is joined with impersonal #7 to form the Fut.
Inf. Pass., e. g. sublatum iri or sublatuiri, see below); (7) Dat. of
an I-stem, e. g. dré-4yé, ‘to see,” and many others. The Teutonic
Inf. is Ace. of an ONO-stem, e. g. Goth. itan, Germ. essen from
*éd-ond-m (cf. O. Ind. 4danam, a Neut. Noun). The form
chosen for the Latin Inf. Act. was a Loc. Sg. of an S-stem, e. g.
digé-ré from *age-si, ama-ré from *ama-si, vidé-ré from *vide-si,

fiu-ré from *fini-si, es-sé from *es-si, dédis-sé (see § 52), fer-ré
from *fer-si, vel-lé from *ve/-s7, the last two showing the regular
change of rs to 77 (cf. torreo from *torseyd, ch, iv. § 153), 2s to Ud
(cf. collum from *colso-, Germ. Hals, ch. iv. § 146). For the
Inf. Pass. a Dat. Sg. was chosen, either (1) of a Root-stem, e.g.
dg-t (O. Lat. ag-ei) from *8g-ai, mor-i (with the diphthong ai
536 THE LATIN LANGUAGE. [Chap. VIII.

weakened in the unaccented syllable, first to ¢2, then to 2, as in
*6c-caido, oc-ceido, oc-cido, ch. iii. § 18), or (2) of an S-stem, e.g.
cima-rt (O. Lat. ama-rei) from *ama-sai, vidé-ri from *wide-sai,
fini-rt, O. Lat. méri-ri, so that the Lat. Inf. Pass. differed from
the Inf, Act. only conventionally, and had no distinctive Passive
suffix. This however seems to be present in the byforms ager,
amarier, morwier, though the exact origin of this -zer, -rier is
doubtful!. For the Perfect Inf. Passive the Perf. Part. Pass.
was used with the auxiliary verb esse, e.g. constat id factum esse,
constat ea facta esse; for the Fut. Pass. the 1st Supine with 77,
Inf. Pass. of e0,‘to go,’ e.g. constat id factum iri, constat ea fac-
tum iri. The Fut. Act., e.g. constat id eventurum (esse), is most
naturally explained as a combination of the Fut. Part. Act. with
esse, though its Old Latin indeclinable use, e.g. credo inimicos
meos dicturum (from a speech of C. Gracchus) has suggested the
theory that it is a compound of the and Supine in -## (e. g. even-
tit, dict, Locs. of TU-stems, ch. vi. § 37) with an old byform of
esse, viz. *esom, later *erum (Umbr. erom, Ose. ezum, Acc. of O-
stem), dictirum for *dictit-erum being in time made personal
dicturus -a -um (the Fut. Part. Act.) in the same way as O. Lat.
‘dicendum est orationem’ changed to class. Lat. ‘dicenda <est
oratio’ (Postgate in Class. Rev. v. p. 301). The Umbro-Oscan Pres.
Inf. Act. is the Acc. Sg. of a Verbal O-stem (e.g. Umbr. er-om,
Ose. ez-um, derce-wm, molt-aum * multare’) (ch. v. § 2).

On Lat. aré in aré-facio and similar Verb-stems, see § 34.

In Vulg. Lat., as reflected in the Romance languages, the
Perf. Inf. has been lost; esse has become essere (Ital. essere,
Span. ser, Fr. étre); vedle, volére (Ital. volere, Fr. vouloir), this
verb having been transferred (by the analogy of its Perf.
volut like monut, habui, &e.) to the second Conjugation, voleo
1 Sg. Pres. Ind. (Ital. voglio), voleat, 3 Sg. Pres. Subj. (Ital.
voglia), as posse became potére (Ital. potere, Sp. poder) through
the likeness of its Perf. potwi to the second Conjugation type.

(cL. § 33 a).

 

* Some make it an addition tozof (A. LL. vii, 132). Similarly in
the Active Inf. ending, with Syncope Vulg. Lat. esse-re replaced esse (see
of the final 2, as in biber for bibére below).
§§ 84-86.] THE VERB. INFINITIVE. 537

§ 84. Pres. Inf. Act. Biber for bibéré in the phrase biber dure (quoted by
Charisius 124. 1K. from various early authors, and censured as a mispronuncia-
tion by Caper, 108, 10 K, bibere non ‘biber’), seems to be a case of syncope of
final -¢ (like nec for néque, animal for dniméile, calear for caleare, ch. iii. § 36),
though it has been also regarded as a veritably old form, a Locative without i
(ch. vi. § 37), like Gk. 36-pev Inf. (see above). Instar may be a similar synco-
pated Inf. (for instare) used as a Noun (cf. bustar, an oxstall ; see ch. iii. § 36).
On late inscriptions we have, e. g. haber (C. I. L. viii. 8369, of 128 a.D.); and
on a lamp found in the oldest Esquiline cemetery (Ann. Inst, 1880, p. 260),
Sotae sum. noli me tanger.

§ 85. Pres. Inf. Pass. The form in -éer, -rier belongs to O. Lat. and is
employed as an archaism by the Augustan poets and their imitators (see
statistics in Neue, ii*. p. 409). Even in the time of Plautus it can hardly have
been so current as the form in -i, -ri, for it is confined to the end of iambic
and trochaic lines, e.g. percontarier, Most. 963 (see Lorenz, ad loc.), and is never
found with a short antepaenultima (except deripier, Men. 1006), restrictions
which indicate that it was a form used only for the sake of the metre.

That the -r of -ri was originally s we see from the O. Lat. form dasi men-
tioned by Paul. Fest. [48.19 Th. ‘dasi’ dari (should we read ‘dasei’ or else
dare ? )].

The occasional scansion of the Pres. Inf. Act. with -2 in Plaut. (e. g. Psewd.
355, 1003) has been explained as a relic of the use of -ai (O. Lat.-ei or -2, ch.
iv. § 34) as Active suffix (. F.iv. 240). But it may be otherwise explained,
as syllaba anceps before final dipody (see Miiller, Plaut. Pros. p. 22). The theory
that the i- and ai- suffixes were in the O. Lat. period used indifferently as
Act. or Pass. is plausible enough but lacks proof.

§ 86. Fut. Inf. Act. Gellius in the seventh chapter of the first book of
his Noctes Atticae quotes several instances of the indeclinable use of the Fut.
Inf. Act. from the older authors, in connexion with the reading: hane sibi
rem praesidio sperant futurum (Cie. Verr. II. v. 65. 167), found in a copy
of Tiro’s edition (libro spectatae fidei, Tironiana cura atque disciplina
facto). This reading was defended by such examples as: credo ego inimicos
meos hoe dicturum (from a speech of C. Gracchus) ; hostium copias ibi occu-
patas futurum, and again: deos bonis bene facturum (from the Annals of
Claudius Quadrigarius) ; omnia ex sententia processurum esse (from Valerius
Antias ; the use of esse is irregular) ; altero te occisurum ait (sc. Casina) (from
Plaut. Cas. 693, where our MSS. are almost unanimous for occisuram !) ; non
putavi hoc eam facturum (from Laberius, Com. 51 R.). Priscian (i. p. 475.
23 H.) quotes from Cato: illi polliciti sese facturum omnia ; from Lucilius
(xvii. 8 M.): nupturum te (sc. Penelope) nupta negas.

The existence of nouns like scriptira, versiira, pictiira points to the Fut. Part.
seriptiivrus -a -um, versitrus -a -wm, pictiirus -a -um, &c. (formed from U-stems,
as Gk. iosxipé-, from a U-stem; cf. ch. v. § 16, ch. iv. § 60) having been
an old formation (ef. offensa beside offensus, répulsa beside repulsus), and makes
it unlikely that the declinable Fut. Part. arose from this indeclinable Inf.
So it may be better to regard O. Lat. dicturwm as the Neut. Sg. of this
participle used (without esse) impersonally, just as the Impersonal constat,
e.g. ‘constat inter omnes haec ita esse,’ becomes in the Fut. Inf. Pass,
538 THE LATIN LANGUAGE. {Chap. VIII.

constaturum, e.g. ‘spero constaturum inter omnes haec ita esse.’ Similarly
the Gerundive in -ndus -a -wm seems the older form and the Gerund in -ndum
an Impersonal use, cundum est being analogous to itur (see §62). On the other
hand the preference shown by Plautus and Terence for the omission of esse
(for statistics, see Postgate in J. F. iv. 252; cf. Plaut. Bacch. 592 negat se
iturum, with Truc. 85 is nune dicitur Uenturus peregre), and the rarity of the
use of the Fut. Part. in apposition (e.g. Enn. Ann. 412 M. carbasus alta
uolat pandam ductura carinam) are quoted in support of Prof. Postgate’s

explanation.

§ 87. Fut. Inf. Pass. Jvi is impersonal, like itur in Virgil’s ttur in antiquam
silvam, so that the line of Terence (Hec. prol.): rumor uenit datum iri gladia-
tores, should be translated ‘that they are going to exhibit gladiators,’ ‘that
there is going to be a gladiatorial show.’ Gellius (x. 14) quotes a curious
extension of this usage from a speech of Cato, contuwmelia mihi factum itur:
atque euenit ita, Quirites, uti in hac contumelia, quae mihi per huiusce
petulantiam factum itur, rei quoque publicae medius fidius miserear,
Quirites. This throws some light on the development of the Italo-Celtic
passive from an Impersonal R-form (§ 62). (Cf. Plaut. Rud. 1242 mihi istaée
uidetur praéda praedatum irier.) The word-group factum iri, &c. seems to have
become a single word in ordinary language, for we often find the Fut. Inf. Pass.
written with -twiri in MSS., which is frequently corrupted in later copies to
-turi as if Nom. Pl. Mase. of Fut. Part. Act. (e.g. sublatuirt in Cod. A of Bell.
Alex, xix. 2, but in other MSS. sublatwri; for a list of examples from
Lactantius, see A, ZL. L. ii. 349). The suppression of -m of sublatum, &e. is
like the suppression of m of circum in cirewit for circwm it (see ch. ii. § 153).

§ 88. VII. THE SUPINES. The First Supine, used after
a Verb of motion, is the Acc. Sg. of a Verbal Noun, a TU-stem
(e.g. wre spectatum, lit. ‘to go to the seeing,’ like ire ddmum, to go to
the house, ive Rémam, to go to Rome), the same form as is in San-
serit the regular Infinitive of the verb (see § 83). An O. Ind. usage
like dr4Stum i gachanti, ‘they come to-see,’ hétum éti, ‘he goes
to sacrifice ’ (cf. O. Sl. vidétti ideti, ‘he goes to see’) (Delbriick,
Altind. Syntax, p. 428) is what the Latin First Supine has
developed from.

The Second Supine, used after an Adjective, is the Loe. Sg. of
the same Verbal Noun (e.g. dgilis cursi, nimble in running).
The Loc. Sg. in -# of U-stems often played the part of a Dat.
(e.g. curru for currui in Virgil, see ch. vi. §§ 37, 23, 27); and we
find the Second Supine used not only as a Loe., but as a Dative,
e.g. (fabula) lida mémbréti, pleasant for telling, where in the
older language the Dative proper in -wi is used, e.g. lepida
memoratui (Plaut.), as well as the Locative, e.g. ridiciila auditu
§§ 87-89.) THE VERB. SUPINES. PARTICIPLES. 539

(Plaut.), (in Plaut. Rud. 294 sunt nobis quaestu et cultu, this Loe.
plays the part of a Predicative Dative).

This TU-stem bulks largely in the language of Plautus and
the older Dramatists, e.g. opsonatu redeo (Plaut. Men. 288),
essum vocare, ‘to invite to dinner’ (Men. 458), nuptum dare, ‘ to
give a girl in marriage’ (Pers. 383). In Aud. 736 perditum ire is
used almost like perdere :

quam 6b rem ita faceres méque meosque pérditum ires liberos,

(cf. Bacch. 565 mi ires consultum male), and the use of the
Accusative without a Preposition is paralleled by phrases like
i malam criicem (Plaut.), suppétias, infitias, exséquias ire, &e. Like
nuptum dare and nuptum ire are vénumdire or venundare (vendere)
and venum ire (vénire; but cf. O. Ind. vasna-yé-ti, 3 Sg., Gk.
@véopat) ; pessumdire or pessumdare and pessum ire. A similar
Ace. of a Verbal Noun TU-stem is asom (class. assum, 1st Supine
of ardeo) in the phrase asom fero on an old Praenestine cista with
the representation of a kitchen scene (Mé. Arch. 1890, p. 303),
a phrase which recurs on a Marrucine inscription, asum . . . feret
(Zvet. I. LT. 8). The 1st Supine is also found in Umbrian, e. g.
aseriato etu ‘observatum ito.’ In the Romance languages the
Supines have been lost.

§ 89. VIII. THE PARTICIPLES. The I.-Eur. Participles
were merely Verbal Adjectives formed with the various suffixes _
already mentioned in the chapter on Noun- and Adjective-stems
(chap. v). ,Thus for the Perf. Part. Pass. the TO-suffix was used
(e.g. O. Ind. -dhi-té-, Gk. 6¢-rés, Lat. cré-di-tus, Lith. dé-tas ;
Goth. vaurh-ts, ‘ wrought’), or the NO-suffix (e. g. O. Ind. pir-na-,
‘filled, O. Engl. bund-en, ‘ bound-en,’ O, S1. nes-ent, ‘ carried’; ef.
Lat. plé-nus); for the Gerundive, the YO-suffix(e.g. O.Ind. agar,
‘seeable, worth seeing,’ O.Sax. un-féd-i, ‘insatiable’; cf. Gk. &y-tos,
venerable, Lat. ewim-ius), or -TWO-, -TEWO- (e.g. O. Ind. kér-
tva-, ‘ worth doing, Gk. d:wx-ré(F)os, worth pursuing), and so on.
With the LO-suffix is formed the second Past Participle Act. in
O. SL, e.g. nes-lti, used in the periphrastic neslti jesmi, ‘I have
carried,’ and the Aor. Part. Act. or Pass, in Armenian, e. g. gereal
‘ capiens, captus.’

The Pres. Part. Act. (and all Active Participles, except the
540 THE LATIN LANGUAGE. [Chap. VIII.

Perfect) took the suffix -ént-, -nt-, -ont- (see ch. v. § 63) (e.g.
O. Ind. bhérant-, Gk. @épwv, -ovros, Goth. bairands, O. SI. bery),
Lat. fér-ens, -entis, scdens (Umbr. zedef), The Perf. Part. Act.
took -wés- (e.g. O. Ind. ririk-vés-, Gk. Aedour-(F)ds, Lith. likés ;
ef, O, SI. mltiz-ti, ‘having milked’); and this formation appears
in Umbro-Qscan in the Compound Tense, which corresponds to
the Lat. Fut. Perf. (e.g. Osc. fefacust ‘ fecerit:’), and probably in
Ose. sipus, knowing, with full knowledge [from *sép-wes- (?).
That the w is long (L.-Eur. i or 6) is inferred from the absence
of Syncope, for *sipis, *sipds would become in Oscan *sips; but
see ch. iii. § 16]; but in Latin the Perf. Part. Act. is not used.
Neither is the Pres. Part. Middle (or Pass.), which was formed in
-méno- (-mono-, -mno-) (e.g. O. Ind. bhaéra-mana-, Gk. pepd-pevos ;
ef. Pruss. po-klausi-manas, ‘being heard’), though it is found
in the 2 Pl. Ind. Pass., e. g. ferimini (se. estis, see § 81), and in
nouns like alumnus (cf. Gk. 6 tpepduevos, see ch. v. § 13). The
Perf. Part, shows the suffix -to-, e. g. scriptus (Umbr. screihto-,
Ose. serifto-); the Fut. Part. Act. im -¢dus is probably a form-
ation with the suffix -ro- from a TU-stem Verbal Noun, e.g.
scriptirus, stem *scripti-ro- from the stem *scriptu- of scriptus,
-us, pictirus from the stem *pictu- of pictus, -us, &c., like Gk.
ioxi-pdés from icxis. [On its relation to the Fut. Inf. Act. in
O. Lat. seripturum, in class. Lat. seripturus (-a -um) (esse), see
§ 86. For a Participle the Latin writers, especially the poets,
often substituted an Adjective, e.g. dicér for lactrdtus (Virg.
lacerum crudeliter ora), and these Adjectives or ‘ truncated Parti-
ciples’ have to some extent encroached on the Perf. Part, Pass. in
the Romance languages, e. g. Ital. trovo beside trovato. (Meyer-
Liibke, Rom, Gram. ii. p. 375.)

§ 90. Pres. Part. Act. The Pres. Part. Act. was liable to become in all
languages an ordinary Adjective or a Noun ; e.g. Goth. frijond-s, lit. ‘loving,’
assumed the sense of ‘friend’; Lat. riidens (see § 6), lit. ‘rattling,’ assumed
the sense of ‘a rope, tackling’; bénévdlens is a noun in Plautus, e. g. Trin. 46,
1148, &. Very early examples of this seem to be Lat. dens (Ose. dont-?),
a Pres. Part. of the I.-Eur. root ed-, ‘to eat,’ sons a Pres. Part. of I.-Eur. es-,
‘to be’ (cf. O. Scand. sannr, ‘sooth, true,’ which acquired the sense of ‘ truly
charged.’ ‘guilty’). Of Pres. Parts. becoming Adjectives in Latin examples
are congruens (beside congrwus), benevolens (beside benevolus), bénémtrens (ef. -mérus
in Lucilius’ mercédimérae légidnes) ; indigens, not indigus, and insciens, not inscius
(but nescius), are used by Plautus. The result of this close connexion of
§§ 90-92.) THE VERB. PARTICIPLES. 541

Verbal Adjectives (especially Compounds) in -us with Participles in -ens, is
a Comparison like b Ls, l » lenti. , gnificus, -entior,
-entissimus (see ch. vi. § 55).

The O-grade of the suffix which predominates in the Greek declension,
~€pwy, -ovTos, -ovte appears in Lat. sons (beside prae-sens, prae-sentia, ab-sens,
&e.), ewntis, -i, -em, -es, &e. (beside Nom. Sg. tens), voluntas (beside volens). It
can hardly be due to a mere accident that all these traces of the suffix -ont-
are found in verbs belonging to the Athematic Conjugation, I.-Eur. *es-mi,
*ei-mi, *wel-mi (see § 2). It almost seems as if the declension of the Pres.
Part. of Athematic Verbs in Latin had originally exhibited the suffix -ont-,
perhaps varying with -ent- (I.-Eur -nt- or even -ent-), v. g. ab-dens Nom. Sg.,
euntis Gen., eunti Dat., *ientem Acc. (see ch. vi. § 1, p. 367), or sons Nom. Sg., sentis
Gen. from an L.-Eur. *sénts Nom., *sntdés Gen. (cf. O. Ind. sint-, satds Gen.),
*sentia Fem. (ef. prae-sentia) from snt- like Dor. Gk. éacoa for *éonria, the
equivalent of Att. ota (ch. iv. § 81). The use of euntis, eunti, &, cannot well
have been due to the dislike of the combination -ié- (iens has iz-, the vowel
being lengthened before ns, ch. ii. § 144), for this combination is not
objected to in other Participles, facientis, capientis, &. The survival of the
suffix -ent- in the struggle for existence in the Latin Present Participle,
athematic and thematic, was probably aided by the Latin tendency to turn
every short unaccented vowel before a consonant-group to @ (see ch. iii. § 18;
and ef. below, § 94, on -undus and -endus in the Gerundive). _

 

§ 91. Perf. Part. Act. This has been lost also in Celtic and Teutonic,
but in Balto-Slavie is the only part of the Perfect Active retained. Some
find traces of the formation in Lat. céiddver, pdpdver, others in O. Lat. gndriires
Plur., ‘knowing,’ e.g. Plaut. Most. 100:

simul gnaruris uos uolo ésse hance rem mécum.

(Cf. Gloss. Placid. ‘gnaruris’ gnarus, sciens: Gloss. Philox. ‘gnarurem’
ywohpmov: ‘ignarures’ d-yvootyres: ‘gnarurat’ ‘yvopifer). Mémor seems to be
not a Perfect Participle, but an Adjective derived from a Perfect Participle-
stem, as Gk. xexparyds (Eurip.) is a Noun derived from xéxpaya.

§ 92. Perf. Part. Pass. This participle too became often an ordinary
Adjective (Engl. ‘cold,’ ‘dead,’ &e. are TO-stem Participles), v. g. citws, swift,
lit. ‘bestirred’ (O. Ind. $i-t4-) ; cditus, sharp (this was the meaning of the
word in the Sabine district, Varro’s home, Varro, L. L. vii. 46), then (meta-
phorically) (1) piercing, of sounds, v. g. Enn. 4. 538 M. cata signa, the shrill
clarions; (2) shrewd, of persons, literally ‘sharpened,’ from I.-Eur. ko- (ef.
Lat. cos, a whetstone) (ch. iv. § 54) ; latus, O. Lat. stldtus, broad, lit. ‘extended ’
(ef. O. SI. stel-ja, ‘I spread, extend’) !, or a Noun, e.g. ndtus, a son (in Plautus
and Terence we have as a rule gnatus, a son, natus, born), légdtus, a lieutenant,
deputy. When used as an Adj. it sometimes passes into the I-declension,
the favourite Adjective declension (ch. v. § 34), e.g. fortis, O. Lat. forctus,
originally P. P. P. of the I.-Eur. root dhergh-, ‘to establish’ (0. Ind. drdhé.,

 

1 Another example is cunctus 1168: fac istam cunctam gratiam.
(for co-vinctus ?), which has still its Acciératus is always a Participle in
participial sense in Plaut. Most. O. Lat.
542 THE LATIN LANGUAGE: [Chap. VIII.

‘established, firm’), in-gens, lit. ‘unknown,’ from the root Sen-, ‘to know’
(O. Engl. un-cid, ‘uncouth’). When used as a Noun the Neuter often
appears, e.g. lectum (also Mase. lectus, of the fourth or second decl.), tectum,
fatum (the sense of ‘destiny’ probably originated in the phrase ‘fari fatum
alicui,’ to lay a doom or spell on one, like the Welsh tynghu tynghed ; see
Rhfs, Proc. of Internat. Folklore Congr. 1891, p. 150), and (especially in the case
of Abstract Nouns) the Feminine, e.g. offensa, répulsa.

The Participle in -to- of Intransitive Verbs has the sense of a Perfect Part,
Active, e. g. céndtus, having dined, pransus, potus, like our ‘learned’ in such
a phrase as ‘a learned man,’ ‘a learned judge.’ Hence its use as the Perfect
Participle of Deponent Verbs, e.g. asperndtus, sécitus, and the coexistence of
Deponent Perf. Part. and Act. Verb, e.g. fisus beside fido, maesius, sad, beside
maereo, These participial TO-stems from roots which have developed in
Latin into Deponent Verbs often retain their true passive sense, e.g. dbizsa
Pass. (Plaut. Asin. 196; cf. Pelign. oisa aetate ‘confecta aetate’), so that
there is a justification for Virgil’s oblita carmina, Horace’s détestdta bella, and
the like.

It seems to have taken in I.-Eur. the weak grade of the Verb-root, e. g.
I.-Eur, *wid-to- from the root weid-, ‘to know’ (0. Ind. vitté-, Gk. d-coros,
Goth. un-vis), *klii-to- from Heu- [0. Ind. Sriité-, Gk. «atrés, Lat. in-chiitus,
O. Ir. cloth for *eltito-, O. H. G. Hlot-hari (the equivalent of Gk. Kaurd-
orparos, cf. Germ. Heer, an army), the name ‘Lothair’]. In Latin this is
also the rule, e.g. diic-tus from diico, tis-tus from tio, tentus for *tn-tos (Gk.
Tarés) from root ten-, piitus (Plaut. Pseud. 1200) used with its equivalent
pii-rus, See ch. ii. § 144, where it is suggested that the long vowel of réctus,
&c. may be due to a Latin tendency to lengthen a vowel before the group
g-t. The Perfect Ind. Act. too has often influenced its vocalism. The
same influence caused the substitution of -sus for -tus, which spread from
Dental Verb-stems, where it was due to a Latin phonetic law (ch. iv. § 155),
e. g. tensus for *tend-tus, iisus, older tssus, for *it-tus, *oit-to-, flecus for *flecttus,
salsus for *sald-to- (ef. Goth. saltan), perculsus for *kld-to- (cf. clades), to others
where the Perfect has s (a), e.g. farsus (cf. farsi Perf.), sixus (ef. fixi; but
Jictus from fingo, finwi), just as the analogy of haesi has produced the late form
haes-trus, and hausi (P. P. P. haustus), haus-arus beside hausturus, or as the
analogy of the Present Tense is followed in O. Lat. sortus (surtus?) with -rt-
for -rct-, ch. iv. § 157) for *swrrectus (Paul. Fest. 423. 1 Th. ‘suregit’ et
‘sortus’ pro surrexit, et quasi possit fieri surrectus, frequenter posuit Livius),
expergitus for experrectus (Lucil. iii. 56 M. ; Luer. iii. g29, &c.).

On Osean prafto- ‘ probatus,’ Umbr. vaseto- (from stem vaka-), &e., which
have been compared with Lat. cripitus from crepare, implicitus from implicare,
see von Planta, i. p. 214. The difficulties which they offer have not yet been
satisfactorily removed.

In Italian, French, and Roumanian the ending -dtus (like statatus, minitus,
&c.) has become the normal ending of the P. P. P. of the third Conjugation ;
e.g. Ital. venduto, Fr. vendu, Roum. vindut, point to a Vulg. Lat. venditus
for venditus (see Meyer-Liibke, Rom. Gram. ii. p. 370). (Cf. § 51, above, on the
encroachment of the Perfect in -ui in Vulg. Lat.) For the Past Part. of sum
the Italian and French languages use status (Ital. stato, Fr. été, &c.) (ib. p. 385).
A Romance example of a P.P. P. which has become an Adj. is strictus with
the sense of ‘ narrow ’ (Ital. stretto, Fr. étroit, Span. estrecho).
§§ 98,94.) THE VERB. GERUND AND GERUNDIVE. 543

§ 93. ‘Truncated’ Participles. Priscian (i. 534.6 H.) says: ‘retus’ pro
retitus dicebant, quomodo ‘saucius’ pro sauciatus, et ‘lassus’ pro lassatus,
et ‘lacerus’ pro laceratus et ‘potus’ pro potatus; Gellius (xix. 7) quotes
from Laevius ‘oblitteram’ gentem for oblitteratam, with other novelties
such as ‘accipitret’ for laceret, ‘pudoricolorem’ auroram, curis ‘intoler-
antibus’ for intolerandis.

§ 94. IX. THE GERUND AND GERUNDIVE. The Ger-
undive (Adj.) in -ndo-, Umbro-Oscan -nno- (e. g. Ose. trifbim .. .
Gpsannam deded ‘domum ...operandam dedit,’ Umbr. esonzr...
popler anferener et ocrer pihaner ‘sacris ...populi cireumferendi et
arcis piandae,’ Gen. of ‘Purpose’) has beside it in Latin, but not, so
far as we can tell,in Umbro-Oscan, a Gerund (Neut. Noun) in -ndo-,
which seems to stand to the Gerundive in the same relation as
an Impersonal to a Personal Verb, ewndum est in antiquam silvam
being Impersonal like itur 7 antiquam silvam. Until more
Umbro-Oscan inscriptions with this formation have been dis-
covered, it is rash to attempt a history of the Gerundive and
Gerund on Italian soil. The evidence at present at our disposal
points to the Adjectival use as having been the original one, shared
by all the Italic languages, e.g. domus aedificanda, iter cognoscen-
dae antiquitatis (ch. vi. § 20, p. 383 ~.). In the older Latin writers
when this formation is turned into finite form, i.e. when a state-
ment is made by means of it, the usual method is to employ the
Gerund with est governing an object, e. g. agitandum est vigilias,
imperandum est servis, carendum est urbe; but in classical Latin
the Gerundive is preferred if the Verb is one which governs the
Accusative, e.g. agitandae sunt vigiliae, but still as before,
imperandum est servis, carendum est urbe (see Roby, Lat. Gram.
ii2, Pref. pp. Ixi sqq.); and the transition from the impersonal to
the personal mode of expression, marked by a construction like
Plautus’ nominandi istorum copia (a construction allowed by
Cic. with a Gen. Pl. for the sake of euphony, e.g. facultas
agrorum condonandi) is like the transition from ‘factum itur
contumeliam’ to Cato’s contumelia factum itur ($ 87).

The origin of the Gerundive suffix still remains doubtful, after
all the theories that have been started to account for it (see
especially Brugmann, Grandriss, ii. §§ 69 and 1103; Thurneysen,
K.Z. xxx. 493; Conway, Class, Rev. v. 296). In the third and
fourth Conjugations the form -éudo- cannot be a phonetic
544 THE LATIN LANGUAGE. [Chap. VIII.

development of -dzdo- through weakening of the vowel 6 in the
unaccented syllable, for we have anfereno- for *imbhi-bhérénds-
in Umbrian, where a weakening of this kind would not be
found. Although the form -endo- became the approved form in
classical Latin, while -é%ndo- (from an earlier -dudo-) was relegated
to the legal and archaistic style, e.g. rés répétundae, yet the
classical form is found on the earliest inscriptions. On the
Senatus Consultum de Bacchanalibus of 186 B.c. (C. 1. LZ. 1. 196)
we have eadeicendum * edicendum’ and faciendam, and on the Lex
Repetundarum of 122-123 B.c. (i. 198) tribuender, fruenders,
Jruendum, &c. stand side by side with legundis, scribundi, defer-
undo, quaerundai, &e. The two are rather parallel endings like
-Ont- and -ént- in the Pres. Part. Act. (see § go), although the
ultimate acceptance of the E-form in both these cases may have
been helped by the Latin tendency to turn a short unaccented
vowel before a consonant-group into é (ch. ili, § 18).

With the Gerundive suifix are evidently connected the suffixes
of Adjectives in -Lundo-, -cundo-, &c., e. g. erra-bundus, wa-cundus,
ribi-cundus, rotundus (cf. O. Ir. cruind, ‘round, alaind, ‘ beauti-
ful’?). The d of the first of these has been referred to the root
bheu- of fui, &e., the c of the second to the suffix seen in ra7-care,
alli-care, &e. (above § 27, ch. v. § 31).

§ 95. Origin of the suffix -ndo-. The suffix -do- of lici-dus, &e. has been
referred to the verb dare, so that luci-dus would really mean ‘ giving light’
(ch. v. § 67). In the Gerundive this same suffix (cf. Aius from aio, Panda
from pando, &., ch. v. §§ 2,3) may perhaps be joined not with a stem, as in luci-
dus, but with an Accusative case, as -dex (from dico) is in vin-dew (ch. v. § 80), or as
the finite verb is joined in composition with an Accusative in venun-do, vendo,
pessum-do. These may be relics of a mode of expression that prevailed much
more widely in the pre-literary period of Latin. The combination of Active
and Passive sense, which is so peculiar a feature of the Latin Gerundive
forms, e.g. ‘agitandum est vigilias’ Act., and ‘agitandae sunt vigiliae’ Pass.,
and ‘anulus in digito subtertenuatur habendo’ Act.-Pass., may then be
explained from the double sense that can be attached to an expression like
ruborem dare, (1) to blush, Neut., (2) to cause to blush, Act. The Accusatives
will be Accusatives of Verbal Noun-stems *rotam-dus, *laudam-dus (lauddndus,
Audacis exe. 359. 15 K.), *rubem-dus, *habem-dus (Ital. provienda points to habén-
dus) ,*ferim-dus (with *ferém-dus; cf. fere-bam, § 34) (see § 34 on rube-facio, &e. , i-Licet,
&e., and cf. the Zend Inf. dam, ‘to set, to give’). That -md- became -nd- in

Latin, but -nn- (often written x) in Umbro-Oscan, we see from Lat. quon-dam,
quan-do, Umbr. ponne and pone, Osc. pon from *qtom-dé.

§ 96. Adjectives in -bundo-, -cundo-, &c. Examples of -bundo- from Verbs
§§ 95-97.] THE VERB. IRREGULAR VERBS. 545

of the first Conjugation : errd-bundus, praeda-bundus (Sall., &c.), populd-bundus
(Liv., &e.), vertila-bundus (Varr. Men. 108 B.); of the second: ptidi-bundus (Aug.
poets) ; of the third : fiiri-bundus, miri-bundus, quéri-bundus, trémé-bundus (cf.
treme-facio, treme-sco), frémé-bundus (Accius, with fremi-?, and Aug. poets); of
the fourth : luscivi-bundus (Plaut. Stich. 288 ; the quantity of the third vowel
is not decided by the metre, but must belong). Examples of -cundo- are from
the first : fa-cundus (cf. fa-tus), wra-cundus (cf. ird-tus), jii-cundus (ef. ~jatus) ; and
from the second: fé-cundus (ef. fé-tus*, véré-cundus (from Plaut. onwards),
ribi-cundus (Ter., &e. ; cf. rubé-facio), Rotiindus has -tindo- (ef. Ital. rotondo,
Span. redondo, &c.) from -dndo- (on the spelling rutundus, see ch. iii. § 33).
Séciennus, a Plautine word for sdcius (Aul. 659), may be a dialectal form
(Plautus was an Umbrian), of which the true Latin equivalent would be
*sociendus.

$97. Some Irregular Verbs. The irregularity of many
verbs consists in their use of different stems for different tenses.
The Substantive Verb shows the root ES- in some tenses (§ 2),
the root BHEU- in others, e.g. Pft. fwi1; and the heteroclite
conjugation of verbs like ‘ to be, ‘to go’ seems to date from the
J.-Eur. period. Other Latin examples are jio (for *fwio from
bhw-, a weakened form of the root bheu-; cf. Osc. fiiet, fiet
‘fient’) and factus sum; fério and percussi; féro and til (O.
Lat. ¢éili, § 39), the Perfect bemg taken from the root of
O. Ind. tul-, ‘to lift,’ Gk. rerAdvat, Ir. tallaim, ‘I take away,’
Goth. pulan, ‘to endure,’ Scotch thole ; ¢oldo for tol-no (Ir. tallaim),
with the Nasal Present-stem and sustdlz for sustétili (§ 44), with
Preposition and Aorist- or Weak stem (§ 3); vescor and pastus sum ;
arguor and couvictus sum; réminiscor and recordatus sum ; médeor
and médicatus sum; surgo differs from swrreaa and surrectus in
being syncopated (cf. porgo and porrigo) ; a contracted form of
the P. P. P., sortus, is said by Fest. 422. 5 Th. (ef. Paul. Fest. 423.
1) to have been frequently used by Livius Andronicus ; Virgil
(4. iv. 183) uses the full form swd-7%go in an active sense: tot
subrigit aures; on vs beside vd/o, see § 3. Other verbs, classed
as Irregular, are the Defective Verbs: coepi (the Present coepio
is found in O. Lat., see Georges, Lex. Wortf. s.v., e.g. Plaut.
Men. 960 neque ego litis coepio; Pers. 121 coepére Inf.; on
coepi dissyll. and coépi trisyll., see ch. ti, § 1 50); inguam for

 

1 foret beside esset seems tobeaform — -puer ; the Oscan equivalent is fusid,
of *fueret, *bhwését, like soror- from ef. Ose. fust Fut.
*swésor (ch. iv. § 10), or Marci-por for
Nn
546 THE LATIN LANGUAGE. (Chap. VIII.

*ind-squam (ch. iv. § 158) from the root seq*-, ‘to speak’ (W.
heb, O. W. hepp, ‘quoth he, &c.), the unweakened form of
which appears in O. Lat. inségue Imperat., imsequis, ‘ narras,
refers, et interdum pergis, C. G. LZ. v. 78. 10 [or *im(d)-ve-
quam from the root weq!-?]; imque, an Imperat. of imguam,
is used by Plaut., and the 2 Sg., &c. of Pres., inquis, in-
quit, inquiunt and of Fut., inguies, inguiet are frequent, but
inquio Pres. Ind. is not found till Late Lat. (ef. inguzo ‘dico,’
C. GL. iv. 250. 27). [On the parts in use of inquam, dito for
*~igh-io (ch. iv. § 116), infit, see Neue ii”. p. 612; wfio is quoted
from Varro by Priscian, i. p. 450. 17 H.; ai Imperat. (a dissyll.)
is used by Naevius, Com. 125 R. uel ai uel nega (but ae ‘incipe,
dic” C. G. L. v. 165. 7, like infe ‘incipe, dic, narra,’ id. 211. 10,
infens ‘ dicens,’ 7b, 211. 37); the two vowels of azo are sometimes
united in a diphthong, e.e. adam (dissyll.), aiz (monosyll.) in a
phrase like aim vero? ds you really mean that?’; on the
spelling aiio, see ch. i. § 7.]

Possum is found in the older writers in its uncompounded form
potis and pote sum (examples in Georges, Lex. Wortf. s.v., Neue,
17, p. 600); pote is properly the Neuter of potis, e.g. pote est
(class. potest), it is possible, but the Mase. (and Fem.) and Neut.
forms are used of any gender and of any number, e.g. potis est, it
is possible, Ter. Phorm. 379, credo equidem potis esse te, scelus,
Plaut. Pseud. 1302; so with potis and pote, without sum, just
as O. Ind. iSvaré-,‘ able, which is used with the sense of ‘can,
though generally agreeing with the subject, sometimes appears in
the stereotyped form iSvards (Delbriick, Altind., Syntax, p. 88). In
Oscan we have a derivative verb used in this sense, putiians and
putians ‘possint,’ putiiad and pitiad ‘possit’ (cf. Lat. pdtui,
like ménui from moneo), but the Latin derivative potio is used in
the sense of ‘ potem facere, potior (fourth and third Conj.) of
‘potis fieri’ [ef potior fieri used in the sense of potiri, Plaut.
Cas. 112 quam tu eius potior fias, like certior (and certus)
fieri], e.g. eum nunc potiuit pater seruitutis, Plaut. dmpf.
178; postquam meus rex est potitus hostium, Capt. 92; regni
potiri, Cic. On possimus for possumus in MSS. of Virgil,
Fel. vii. 23 and perhaps viii. 63 : non omnia possumus omnes, also
in the Verona Palimpsest of Gaius (g9, 14; 101, 2), &c., see ch. ii.
§ 98.] THE VERB. IRREGULAR VERBS. 547

§ 16. Possem for potessem has been explained by the Analogy of
possum (for pot(e)-sum) beside potis-sum, possim beside potis-sim.

Malo, a contraction of mavélo}, seems to come similarly from
magé (a byform of magis, ch. ix. § 4) united into a word-group
with vio, and nédlo from the Negative xé- and volo, though
the exact process of phonetic or analogical change by which
the various parts of these verbs arose is not easy to trace.
Perhaps mag(é)-vdlo became *mavvdlo (written mdvolo, as *cavvi
was written cdvi, § 39. 4), much as I.-Eur. g* or gw became
Lat. v (ch. iv. § 139). And xé-vdlo (for the word-group cf.
ne-scio, né-queo, Engl. cannot) may have become *ndvdlo as
I.-Eur. *néwos became Lat. névus (ch. iv. § 10), and similarly
*novis, *novolt, *novolumus, *novoltis, *novolunt. By loss of
intervocalic v (ch. ii. § 53) arose d/o (the only form known to
Plautus), ndlumus, noltis (ap. Diom. 386. 19 K.), nolunt, while
O. Lat. né-vis, né-volt, as well as class. non-vis, non-vult, non-
vultis are re-formations. On Imperat. 2d/i see § 2, and on the
spellings mallo, nollo, ch. ii. § 129. Another account of md-volo
makes its first element the Comparative Adverb *ma, *mé (see
ch. x. § 6 on im-mo), and not mage.

§ 98. Irregular Verbs in Romance. Of Lat. swm the Vulg. Lat. Inf. essée
(Ital. essere, Fr. étre, Span. ser), and P. P. P. statws (Ital. stato, Fr. été), have been
already mentioned (§§ 83, 92); the Pres. Subj. seems to have been siam (Ital.
sia, Span. sea, O. Fr. soie). For Lat. eo various verbs were united in the Vulg.
Lat. paradigm, e.g. Fr. je vais (from Lat. védo), nous allons (from Lat. am-
bulo?). Lat. hdibeo was, owing to its Auxiliary use, shortened to a declension
like hayo, has, hat... haunt (Ital. ho, hai, ha... hanno; Fr. ai, as, a... ont;
Span. he, has, ha... han). Vulg. Lat. voleo, volére Inf. (Ital. voglio, volere ;
Fr. veux, vouloir) by Analogy of Pft. volui, as Vulg. Lat. potére (Ital. potere,
Span. poder) by analogy of potui, have been cited in §§ 33 a, 83.

 

1 Both mavolo and malo, mavelim end of a line in Terence, Hec. 540.
and malim occur in Plautus, but malo, Mallem is not found in Plautus,
malim, &c. are predominant later, only mavellem (Solmsen, Stud. Lautg.
though we have mavolo once at the  p. 55.)

Nn 2
CHAPTER IX.
ADVERBS AND PREPOSITIONS.

§1 ADVERBS. Latin Adverbs are for the most part cases of
Nouns, Adjectives (or Participles), and Pronouns, the cases most
frequently found being the Accusative (cf. O. Ind. néktam,
‘by night,’ satydm, ‘truly,’ Acc. Se. Neut., pratarim Acc, Sg.
Fem., Gk. apdrepov Acc. Sg. Neut., axpdv Acc. Sg. Fem.,
xpvpa Acc. Pl. Neut.), Ablative (cf. O. Ind. dirt, ‘afar ’) [also
Instrumental (2), cf. O. Ind. diva, ‘by day,’ Gk. a&Ay]}, and
Locative (cf. O. Ind. diiré, ‘afar, Gk. otkor, dvaipwrel, éxovri),
and often retain case-forms which have become obsolete in the
ordinary declension. Thus, -2m, the original form of the Acc.
Sg. suffix of I-stems (ch. vi. § 29), which in classical Latin was re-
placed by the -e of Consonant-stems, e. g. partem from the stem
parti-, like militem from the stem mzéit-, is retained in Adverbs
like partim, and in the Adverbial Accusatives of Verbal Noun
I-stems, e.g. raptim from the stem rapti- (class. raption-, ch.v.§ 42),
sensim from the stem sensi- (class. sensidn-), wni-versim (Ose. bini-
veresim) &e. ; i an old equivalent of noctu, seems to be an early
byform of xoctis Gen.; -é (older -éd), the suffix by which
Adverbs derived from Adjective O-stems are formed in the
Italic languages, is an Abl. Sg. suffix (parallel with -d, older
-6d, ch. vi. § 33), which has been reserved for Adverbs alone. The
Adverbial suffix -¢us (I.-Eur. *-tés) of fundi-tus, divini-tus, &e.,
is in O. Ind. occasionally used as an Ablative suffix (Sing. or
Plur.) of Nouns, e. g. mity-tas, Abl. Sg. of matér-, ‘a mother,’

 

"In O. Ind. the Instr. Case is in the sphere of motion, ‘by the route
the Noun Declension used to denote of.’ (Delb. Altind. Synt. p. 129.)
§ 1.) ADVERBS. 549

Satru-tas, Abl. Pl. of Sétru-, «an enemy’; and this close relation
of Noun and Adverb suffixes makes it difficult to draw any hard
and fast line between the suffixes used in the formation of
Adverbs, and the suffixes used in the declension of Nouns. An
example of a purely Adverbial suffix is -iés (older -iens), by
which most Numerals form their Adverbs, e. g. sea-ies (cf. t0t-ies,
quét-ies), or the *-mente of the Romance languages, e.g. Fr.
facile-ment, Span. facil-mente, which is nothing but the Abl.
Sg. of Latin mens, just as the -versus, -versum (older -vorsus,
~vorsum) of quaqudversus (-m), alidvorsum (contracted aliorswm),
retrovorsum (contracted retrorsum, retrosum), rursus (-m) for
reversus (-m), &c. are nothing but the Nom. Sg. Mase. (see below)
and Acc. Sg. Neut. of the P.P.P. of verdto, to turn, or as the
-ténus of hactenus, dliquatenus, &c. is the Acc. Sg., employed
adverbially, of the old neuter noun fenus, a stretching, used by
Plautus in the sense of a string or snare (§ 54). The Adverbs
derived from Pronouns, e. g. 742, inde, illine, offer special difficulty,
because of our ignorance of the full number of case-suffixes used
in the I.-Eur. declension of the Pronoun. Thus -am of quam,
tam, nam, jam is naturally taken as Acc. Sg. Fem., but it has
also been referred to an Instrumental formation with the suffix
-m, -mi, which appears in the Instrumental case of Nouns and
Pronouns in Balto-Slavic (ch. vi. § 36); and various other Latin
Adverbs in -i, usually called Accusatives, have been referred to
the same source.

Adverbs in -ter, e. ¢. bréviter, are best explained as Nominatives
Singular Mase. of stems in -téro-, a Stem-suffix which occurs in
various pronominal and locative Adjectives, e.g. al-ter, dex-ter, &c.
(ch. v. § 16) (%-¢erwm isan Acc. Sg. Neut. of a similar formation),
though they have been also referred to the noun ‘er, brev-iter
corresponding to German kurz-weg. Other examples of a Nom.
Sg. Mase. Adjective stereotyped as an Adverb are probably
deinceps, which was declined in O. Lat. like princeps, though it
might be regarded as Acc. Sg. Neut. (cf. exordium princeps),
demus an O. Lat. byform of démum; more certainly adversus,
rursus, &¢., the byforms of adversum, rursum,and other compounds
of versus (-m), &e.

Examples of Acc. Sg. Neut. forms are, beside the Adverbs in
550 THE LATIN LANGUAGE. [Chap. IX.

-tim (-sim) just mentioned, Adverbs in -wm like 7-terwm, another
time (Ace. of *itero-, O. Ind. itara-, ‘ other,’ from the Pronoun-
stem i-, ch. vil. § 13, with the sutfix -tero-, ch. v. § 16), cé-terwm
(ch. x. § 5), comméddum (beside commode), multum, pirum (§ 7),
plérum-que (p. 559), vérum, tantum, along with some Superlatives
like minimum (usually minime), potissimum, insdnum (not insane in
Plaut.) ‘very, and occasional Adverbs from I-stem Adjectives,
e.g. fucilé, difficile, sublimé (though -i¢er is the usual suffix, e. g,
breviter, fidéliter, the formation in -é being reserved for poetry,
e. g. dulce ridentem . . . dulce loquentem, Hor.). The comparative
degree of the Adverb is always the Acc. Sg. Neut. of the Com-
parative Adjective, e.g. longius, facilius, brevius. Accusatives
Sing. of Nouns used Adverbially are vicem, id génus, &e., to
which we may add an Acc. Pl. fords, which is only used with
verbs of motion towards (foras ire, like rus ire, domum ire), and
seems to be Acc. Pl. of *fora (Gk. @¥pa), a door, Of Ace. Pl.
Neut. forms we have ceteré (Virg. A. ix. 656 cetera parce puer
bello), omuia (Virg. A. iv. 558 omnia Mercurio similis; C.J. L.
vi. 1144 omnia magno Constantino), possibly the O. Lat. forms
contra [for which contrd, an Abl. Sg. Fem., was used in class.
Lat., and contrud, an Abl. Sg. Neut. (cf. Lat. contré-versia), in
Oscan], and frustré (class, frustra).

Of Ablatives we have for O-stems, beside the éd-suffix already
quoted (a suffix used to form the Superlative Degree of Adverbs,
e.g. facillime, O. Lat. facilumed, brevissime), the ordinary dd-
suffix (ch. vi.§ 33), e.g. certo (beside certe), vero (beside vere), asstduo
(usually assidue), eaplordto (and explorate), mérito and immerito
(neither of these are Adverbs in Plautus, for they are used with
meo, tuo, &e.), festindto (also festinatim), fortuito (also fortwitu ;
see ch. v.§ 49). Examples of Ablatives Sg. Fem. are dewtéra (scil.
parte), recta (scil. via), eddem (scil. opéra), extra, supra. As was
pointed out before (ch. vi. § 36), the loss of final d after a long vowel
at the beginning of the 2nd cent. B. c. makes it impossible to be
certain that some of these ‘ Ablatives’ (e.g. ina, omnino) are
not really Instrumentals (e.g. Gk. Xd0pd, cow ?); for it is
natural to suppose that the Instrumental suffix, whose existence
in the Latin declension of Nouns is difficult to establish (ch. vi.
§ 36), may have survived in Latin as an Adverbial suffix. The
§ 1] ADVERBS. 551

shortening of -d (by the Law of Breves Breviantes, ch. iii. § 42)
in cité, mbdd, and of -2 in béné, maz, is no proof that these words
ended originally in -6, -é (Instr.) and not in -dd, -éd (Abl.), but
should be referred to their greater use in every-day speech (cf. Aavé
beside méné, &e. in Quintilian’s time, ch. iii. § 42), and their more
frequent occurrence in word-groups, e.g. dene-rem-géras, mate-
sdnus (cf. diéquinti beside fide, ch. iil. § 44). It is only rarely that
the occurrence of an Adverb on an old Latin inscription, or on
inscriptions written in Oscan, or some other dialect which retained
-d, enables us to decide, e.g. O. Lat. meritod (meretod), porod,
eatrad, suprad, Osc. contrud, Falise. rected, which are Abl. forms,
Ose. suluh,‘ wholly,’ from the stem sod/e-, all, whole, which is called
an Instrumental form (but see ch. vi. § 36). The third Declension
‘Abl. (originally a Consonant-stem Locative, ch.vi.§ 37) appears in
forté, sponté, répenté (O. Lat. dérepente), &e., and the Plural Abl.
(Instr. Loc.) in gratis (older gratiis), ‘for mere thanks’ (cf. Ter.
si non pretio, at gratiis), mgratiis (ingratis), foris (used with verbs
of rest, e. g. foris manere, occasionally with verbs of motion from,
e. g. foris venire), adternis. Examples of Locatives of U-stems are
hiimi (first found in Terence, Andy, 726), postri-dié, die crastini,
nocti (cf. O. Ind. aktazi) (see ch. vi. §§ 37-38).

Pronominal Adverbs show various suffixes: (1) -62 (older -de2)
with locative sense, e.g. 727, u07, also 70%, wb?, with final vowel
either shortened by the Breves Breviantes Law (ch. i. § 42), or
originally short (cf. Osc. puf); (2) -7 (older -e7), the Locative
O-stem suffix, e.g. O. Lat. i//i, isti, there, which in classical Latin
always have the particle -c(e) appended, l-a-c, ist-2-c ; (3) -0, to
indicate motion to a place, e.g. ¢, gud, isté, alid, apparently Abl.
forms like porrd (O. Lat. porod); O. Lat. hoc, istoc, iloc may
represent *Add-c(e), *istéd-c(e), *ilod-ce, Acc. Sg. Neut., for in
class. Lat. we have huc, istuc, luc ; (4) -d, to indicate direction,
manner, &c., e.g. qua, ea, which like quo, eo are Abl. forms
(O. Lat. arvorsum ead), not Instrumental like Greek 77, ratrn,
navrn (afterwards confused with Dative -n); (5) -ém, to indicate
motion from a place, e.g. iddim, istim, which in class. Lat. always
append the particle -ce, wlinc, istine (ck. hinc, dehinc) ; (6) -nde,
with similar sense, in wade, inde; this inde is shortened by
syncope of the final vowel to -i in the Compounds proin, dein,
552 THE LATIN LANGUAGE. (Chap. IX.

ewin, &c. (ch. iii. § 36). Other endings like -dam of quondam (cf.
quidam), -lem of qui-dem, tan-dem, with the sense of ‘ exactly,’
‘ precisely ’ in 2i-dem, tanti-dem (cf. idem, ch. vii. § 21), -tem of
t-tem, -ta of i-ta, are apparently case-forms of pronominal stems,
as -guam of un-quam, us-quam (cf. yuis-quam), appears to be Acc.
Se. Fem. of the stem *q*o-. (See ch. x. on the Conjunctions.)

Other Adverb formations are (1) in -/ariam, indicating division,
e.g. Ui-fariam, quadri-fariam (cf. Gk. -pdovos from -dartos, eg.
tpi-pdcwos); (2) in -séeus, indicating motion from a place; this
secus is an Adverbial Noun, and is appended to Adverbial forms
in -d, e.g. eatrin-secus, intrin-secus, altrin-secus, as -tenus to Abl.
Sg. Fem. forms, eg. gud-tenus, aliqua-tenus, ed-tenus; it is
derived from the root seq?-, ‘to follow,’ and must be distinguished
from (1) the Adverb sécews, otherwise (O. Ir. sech, ‘ beyond,’
W. heb, ‘ without’); (2) the Adverbial Noun sécus, a Neuter
byform of seas, used in phrases like: trecenti occisi sunt virile
secus, ‘ 300 were killed of the male sex’ (see § 50 on the Prepo-
sition secus). The Abl. fini (finé) occurs in O. Lat., like zenus,
after an Abl., e.g. senem osse fini dedolabo, Plaut., oleas operito
terra radicibus fini, Cato; but came to take a Genitive, e.g.
amphoras nolito implere nimium, ansarum infimarum fini, Cato;
fine inguinum ingrediuntur mare, Sall.

Instances of Adverbial word-groups are ad-fatim, sufficiently,
lit. ‘to weariness,’ ad-mddum, quem-ad-modum, qua-ré, quam-ob-
rem, dé-uud for de ndvo, sé-dilé for se dolo, 7-licd perhaps for im
sloco (old form of Jdcv), 1-licet, sci-licet, vide-licet, im-primis, dum-
taxat, &e.

In the Romance languages Lat. mente (p. 549) is the favourite
Adverbial suffix, though Lat. -d(-wm) is not uncommon, e. g. Vulg.
Lat. alto or alti (Ital. alto, Fr. haut, Span. alto) (-i/er has been
lost). Mente (Abl. of mens) retains a trace of its independence
in usages like Span. temeraria y locamente, O. Fr. humle e
dulcemente. So firmly has it established its footing as Adverbial
suffix that it is even added to already formed Adverbs, e.g. Ital.
quasimente, O. Fr. ausiment. (For fuller details of the formation
of Adverbs in Romance, see Meyer-Liibke, Rom. Gram. ii.
pp. 637 sqq.).
§ 2.) ADVERBS. 553

§ 2. Nominative Adverb-forms. Breviter, &c. are better considered Nom.
Sg. Mase. than Acc. Sg. Neut. for *brevi-terwm, since the loss of -wn seems only
to oceur in a few words of constant use, such as noenum, ni(hi)lum, which came
to be employed exclusively in that doublet-form which the words assumed
before a vowel, e. g. no(e)n(um) est, ni(hi)l(um) habeo (ch. iii. § 52). Iterum did
not become *iter, nor ceterum, *ceter. The crystallizing of the Nom. Sg. Mase. of
the Adjective as an Adverb, used with any number or gender, finds a parallel
in the extension of pitis, properly Nom. Sg. Masc., Fem. only, e. g. potis sum, potis
es, potis est ille, polis est illa, to all persons and numbers of the Verb, e. g. qui istuc
potis est fieri?, Plaut., with Neuter subject (see ch. viii. § 97). The Adverbs
(Prepositions) praeter, propter, &c. might be similarly explained ; practer it ile,
*praetera it ila, *praeteri eunt ili became praeterit, praetercunt without distinction
of gender or number (but see p. 554). Naper (Superl. nuperrime) appears to be
Nom. Sing. Mase. of the Adjective nupéro- (Plaut. Capt. 718 recéns captum
hominem, nuperum, nouicium) for *névi-péro-, ‘newly acquired,’ unless it is
Acc. Sg. Neut. of an I-stem Adj. (ch. v. § 34) for *nupere. Fdcul (Paul. Fest. 61.
32 Th. ‘facul’ antiqui dicebant et ‘faculter’ pro facile ; Fest. 266. 20 Th. ‘ per-
facul’ antiqui, et per se ‘facul’ dicebant, quod nune facile dicimus; Non.
111. 21 M. ‘facul’ pro faciliter, huic contrarium est ‘difficul’), a word used
by Lucilius in his description of the Roman patricians (vi. 2 M.) :

peceare inpune rati sunt
posse, et nobilitate facul propellere iniquos,

is better regarded as Neut. Sg. for facile, with syncope of -é, as in volup for
rolupe, Neut. of *volupis (ch. iii. § 36), than as Nom. Sg. Mase. of an Adj.-stem
faculo- (ef. sacri-ficulus), since the reduction of -lés to -1 seems to be dialectal
only (e. g. Oscan famel) (ch. vi. § 4), and not, like the reduction of -rés to -r
(e. g. Lat., Umbr. ager, Gk. dypds), shared by Latin. We have dificil Nom. Sg.
Neut. in Varro Men. 46 B. .

quod utrim sit magnum an parvum, facile an difficul.

Simiil (older semol, C. I. L. i. 1175, in MSS. of Plautus semul, e.g. Rud.
760, Men. 405) (Umbr. sumel?) may be Acc. Sg. Neut. of similis, unless it
rather shows the L-suffix of 0. Sl. ko-li, ko-1é, ‘ when,’ to-li, to-lé, ‘ then,’ &e. ;
simul and simulter (a byform of simitliter, used by Plaut. Pseud. 382) corre-
spond to facul and faculter. Deinceps was declined like princeps in O. Lat.
(Paul. Fest. 53. 1 Th. ‘deincipem’ antiqui dicebant proximo quemque
captum, ut principem primum captum; cf. ib. 50.- 5), and may have
been associated with princeps in a fragmentary line on the Lex Repetun-
darum (6.1. L. i. 198. 79): iudex deinceps faciat pr[incipe cessante]. Démus
(ef. Gk. rijpos) was used by Livius Andronicus (Paul. Fest. 49. 27 Th.) ; it is
the reading of the Palatine family of MSS. in Plaut. Truc. 245: qui dé then-
sauris integris demas danunt (demum oggerunt A), and is required by the
metre in Trin. 781. In the Adverbial compounds of versus the terminations
-us and -wm compete in the early literature, e. g. rursum and rursus (also russum,
rusum, &¢., ch.ii.§ 104; rurswm appears to be used in Plaut. after Compounds with
re-, e.g. redeo rursum, and at the end of the line), prorswm (with local sense in
Plaut. Pers. 677 simulato quasi eas prorsum in navem ; Mil. 1193, &c.) and
prorsus (cf. prose oratio), sursum and sursus, but in the classical period one of the
rival forms often has the monopoly, e.g. rursus, prorsus, sursum (for statistics, see
Ritschl, Opuse. ii. 259 ; Neue, Forment. ii®. 743). These compounds show other
554 THE LATIN LANGUAGE. (Chap. IX.

adverbial suffixes in dniversim (Ose. diniveresim), wndse for *wnd-vorse (Pacuy.
Tray. 213 R. éccidisti, ut multa paucis uérba unose obnuntiem). On the
Preposition versus, see § 58. Adverbs in -ter are in classical Latin almost
confined to Adjectives of the third Decl. (Cicero in his earlier writings uses
hionaniter, but finally discarded it for himdnz), but in the older literature are
often formed from O-stem Adjectives [Priscian, ii. 70. 20 H. gives a list of these
from the older writers, e. g. dmiciter, maestiter, Plaut., and the eleventh book of
Nonius is devoted to obsolete Adverb forms, especially (1) Adverbs in -ter from
O-stem Adjs., (2) in -2 from I-stem Adjs, e. g. célére, fidéle, (3) in -tus, e. g. médul-
litus, largitus, comminitus]. The NT-stems have -nter, e. g. vehementer, impiidenter,
instead of -nt-ter (cf. 3 Pl. Imper. férunté for *feront-tod, ch. viii. § 57), by dissimi-
lation, some say, for *-nt-i-ter (ch. iii. § 13, p. 176). From auddax we have audacter
(less commonly audaciter ; see Georges, Lex. Wortf. s. v.), but from fallax, fallae-
-iter, from liquitx loguaciter, from prictx procaciter, &c. Difficulter is more usual than
difficiliter (see Georges) ; ficulter is mentioned by Paul. Fest. (61. 32 Th. ‘ facul’
antiqui dicebant et ‘faculter’), but ftciliter (see Georges) was the form that
competed with facilé (cf. Mart. Cap. iii. 325 cum difficulter dicamus, cur ‘ facul-
ter’ dici non potest?) ; simulter is quoted from Plaut. Pseud. 382 by Nonius
i7o.19 M. Quintilian condemns hoth audaciter and faciliter [i. 6. 17 inhaerent
ei (sc. analogiae) quidam molestissima diligentiae perversitate, ut ‘audaciter’
potius dicant quam audacter, licet omnes oratores aliud sequantur, et ‘emi-
cavit’ non emicuit, et ‘conire’ non coire; his permittamus et ‘ audivisse’ et
‘scivisse’ et ‘tribunale’ et ‘ faciliter’ dicere], regarding audacter and facile as
the true Latin forms.

Other Adverbs that might be called Nominative forms are: éminus and
comminus (the spelling co-minus is due to the analogy of é-:minus ; see Georges
8. v.), Which may be Nom. Sg. of Compound Adjectives (cf. Gk. avrd-xerp, e. g.
Soph. Ant. 1175 Aipwy éd\wdev, airéxerp 8 aipdocera); récens (used with a Perfect
Participle, e.g. Lucr. vi. 791 nocturnumque recens extinctum lumen), which
is equated with Mbens by Charisius (114. 21 K.): utimur sic ‘recens venit,’
quod est pro adverbio nomen, ut ‘libens dixit.’ (Similarly répens comes very
near répenté in phrases like Liv. xxii. 8.1: repens alia nuntiatur clades) ;
pro-tinus (also prd-ténus), ‘stretching forward,’ may be the Nom. Sg. of an
O-stem, as the (somewhat doubtful) form pr%-tenis (Afran. Com. 107 R. cémis-
satum protenis recté domum Digredimur) may be the Nom. Sg. of an I-stem
(ch, v. § 34); sécus in intrin-secus, &c., has been sometimes explained as Nom.
Sg. of an Adj. *seco-, lit. ‘following from within.’ But these, and indeed all
the Adverbs cited as Nominatives, are capable of other explanations ; recens
(like deinceps) as Ace. Sg. Neut.; cominus, eminus [with the other Adverbs in
-8, rursus and other compounds of versus (vorsus)] as augmented with the same
particle -s (p. 573) as appears in Greek péypi(s), duis (and dupi), eb6U(s), &e.
(with demus cf. Gk. jyos, Thos, and for -tenus, -secus, see above) ; praeter, propter,
subier are best explained as suffixless Locatives (ch. vi. § 37) like O. Ind. pra-tar,
‘early,’ sanu-tir, ‘away,’ Lat. super and Gk. trép (cf. O. Ind. upar-i, with the
Locative suffix), Lat. inter and O. Ind. antir (but antdri-k3a-), though they are
also capable of being referred to the O. Ind. ending of local Adverbs, -tra
(I.-Eur. -tré?), e.g. titra, ‘there,’ ydtra, ‘where,’ anydtra (Lat. dliter ?), ‘else-
where,’ *prai-wt, &e. becoming by syncope *prai-ter, &e. ; procul has been simi-
larly explained as *pro-tle (ch. iv. § 105), by Dissimilation from *pro-tre (ch. iv.
§ 84), or as Ace, Neut. Sg. of a compound Adjective, whose second element
§§ 8, 4.] ADVERBS. 555

is from the same root, q3el-, as Gk. rHAe, méAa, but it is more naturally
referred to some extension of the Preposition prd by a co-suffix (ef. réci-procus
from re-co- and pro-co-, O. Sl. pro-kit) ; napér (Superl. nuperrime) may have as its
second component the Preposition per of antio-per (§ 7), sem-per (2), Ose. pert
in petiro-pert ‘ quater,’ and as its first the adverbial particle *nt (O. Ind. ni,
‘now,’ Gk, vi-v, vv, Lat. nu-dius tertius, &e., ch. ». § 10). The existence of
Adverbs in Latin derived trom Nom. Sing. case-forms is thus at once natural
to imagine and difficult to prove.

§ 3. Genitive Adverb-forms. Examples of nox, used for nocti, in O. Lat.
are XII Tab. : si nox furtim faxit ; Enn. 4nn. 439 M. ;
si luci, si nox, si mox, si jam data sit frux;
(cf. Plaut. Asin. 598, and see Gell. viii. lemm. 1). In Lucil. iii. 22 M. it is
qualified by media,.as if nocte :

hine media remis Palinurum peruenio nox.

The parallel Adverbs in other I.-Eur. languages, e.g. Gk. vuerds, O. Ind.
aktés, ‘by night,’ vistos, “by day’ (cf. Germ. nachts), suggest that it is a Genitive
form (see however ch. iii.§ 16). Dius,in O. Lat. ‘ by day,’ may then be likewise
a Genitive form (I.-Eur. *diw-os was the Gen. of the word for ‘ day,’ 0.-Ind.
divas Gen., Gk. Acfés Gen.) and inter-dius. They have also been explained as
suffixless Locatives (ch. iv. § 37), like O. Ind. sa-divas, ‘at once,’ pairvé-dyus,
‘early in the morning’ (so pénés may be a suffixless Loc. of pénus Neut. § 37).
Examples of dius are, Plaut. Merc. 862 noctu neque dius; Titin. Com. 13 R.
noctu diusque. Interdius is more common, e.g. Plaut. Asin. 599: nune enim ésse
/ negétiosum intérdius uidélicet Solénem ;
(other instances in Georges, Lex. Wortf. s. v.), Just as interdii became the usual
word for ‘by day,’ while diz [formed on the analogy of noctit, a U-stem Loe.
(ch. vi. § 37), according to one theory] is seldom found in this sense. (It is
found only in conjunction with noctu, at least in the older writers). Nonius
(98. 20 M.) gives as examples Plaut. Cas. 823 (apparently anapaestic) :

noctique et diu ut uiro subdola sis;
Aul. fr. 4 nee noctu nec diu; Titin. Com. 27 R. (in his play about the life of
the fullones) :

nec néctu nec dit licet fullénibus quiéscant.
(On diu, for a long time, see ch. x. § 12; we have quandius for quamdiu in the
epitaph of a litter-bearer (C. I. L. vi. 6308), quamdius (vi. 13101)]. Mox has been,
like the Adverb nox, explained as an old Genitive-form, or a form with the
Adverbial suffix -s (a variety of -sit, -si, the suffix of the Loc. Plur.?, cf. Gk.
népié and peragv) ; its cognates are O. Ind. makii, ‘soon, quickly,’ an Adverb
from the Adjective makst-, ‘quick,’ O. Ir. mos- (e. g. mos-ricuh, ‘I will soon
come’), moch, ‘early.’ Vix (connected with vicem?) must be of similar formation.
Per-nox is an Adj. (e. g. luna pernocte, Ov.), so in Late Lat. pernox et perdius,

-a, -um.

§ 4. Accusative Adverb-forms. The Adverbial use of the Accusative Case
was a feature of I.-Eur. syntax. In addition to the Latin examples already
given (vicem, id ginus, stcus in virile secus, &c.), may be mentioned partem (e. g.
magnam partem, maximam partem), principium (e.g. Cato, R. R. 157. 1 de brassica
pythagorea, quid in ea boni sit. salubritatisque, prineipium te cognoscere
556 THE LATIN LANGUAGE. [Chap. IX.

oportet). When used alone as an Adverb, the Acc. Sg. partem retains its old
form partim (see Gell. x. 13 on partim hominum venerunt-and Cato’s cum partim
Wlorum erat). If saltem is Acc. Sg. of *salti-, a leap, lit. ‘with a leap, ‘swiftly,
easily, assuredly,’ it has taken -tem by analogy of au-tem, i-tem, &c. [saltim, e.g.
C.G.L. v. 146. 13, is, according to Georges, Lex. Wortf. s.v., a late spelling,
like decim for decem (ch. ii. § 8; cf. B. P. W. xiii. 3101, for the other Adverbial
Accusatives of Verbal Nouns retain, in additicn to the old Stem-suffix -tz-
(elass. -tién-, ch. v. § 42), the old form of the Case-suffix, -im. Other examples of
these Verbal Adverbs are détdtim from the Frequentative ddtare, whence
datatim ludere, to play at ball (Plaut. Cwrc. 296), a phrase used in the famous
description of the coquette (‘ Naev.’ Com. 75 R.): quasi in choro

ludéns datatim dat se et communém facit.

aliam tenet, alii ddnictat, alibi manus

est dccupata, alii peruellit pedem,

alii dat anulum daspectandum, 4 labris

alium fnuocat, cum dlio cantat, attamen

aliis dat digito litteras ;
praesertim from sero, lit. ‘in the front row’ (cf. disertim O. Lat. for diserte) ;
tilatim, at a trot (see,Nonius, 4. 1 M.), connected with tollo; strictim (e. g. strictim
tondere, as opposed to per pectinem tondere, Plaut. Capt. 268) from stringo, to graze,
touch the surface ; pédttemptim, cautiously, like sensim, lit. ‘feeling,’ (Nonius
quotes a byform pedepressim, 29. 1 M.); passim from pando ; cursim, hastily, for
which Virgil and others use cursu [Abl. Sg. of the Verbal Noun-stem cursu-,
like the Comedians’ curriciilé (fugere, abire, percurrere, &c.), Abl. Sg. of
curriculum (cf. Plaut. Trin. 1103, Stich. 337)]; stdétim, at once, lit. ‘standing,’
‘on the spot’ (like iicd, § 7), is the Acc. Sg. of the Verbal Noun stéti- (class.
stdtion-), while O. Lat. statim (for the @, see Nonius, 393. 5M. ; Donat. in Phorm.
v. 3. 7), comes from a bystem sfdti- (ef. stdétu- and sidtu-); in Plautus the word
has the sense of ‘standing to one’s ground,’ e.g. Plaut. Amph. 239 néc recedit
loco quin statim rém gerat ; ib. 276 ita statim stant signa; it does not mean
‘at once’ till Afranius’ time (Langen, Beitr. pp. 16 and 337). Parallel with the
Adjectives in -dto- derived from Nouns, e. g. togatus from tiga, and, like them,
not postulating the existence of a verb in -are (e.g. *togare) (ch. v. § 28), are
Adverbs in -datim like assulatim, from assila, a splinter (e.g. Plaut. Capt. 832
aperite hasce ambas foris, Prius quam pultando dssulatim foribus exitium
idfero), guttatim from gutta, ostiatim from ostium, vicatim from vicus, gradatim
from grddus, &c., also paulatim, pauxillatim, nostratim from noster ; ef. nostrates),
and from proper names, tongiliatim (tongiliatim loqui, ...a'Tongilio parasito, qui
... salutatus convicio responderet ; see Lowe, Prodr.), zopyriatim, Lucil. ix. 74M.
Vir forms its Adverb with -tim, ciritim (so propritim Lucr. ii. 975). Confestim
comes from a stem *festi-. a byform of which, *festién-, is indicated by the Verb
Jestino, from which Virgil (A. ix. 488) coined the Adj. festinus. These Adverbs
in -tim (-sim) were more frequent in early than in classical Latin (Gellius
xii. 15 remarks on their frequency in the historian Sisenna), but returned
into fashion at a later period (see A. L. L. viii. 98). Vicissim (which some
derive from *vic-essi, a Loc. Plur. of *vix, Acc, vicem) is an abnormal form. In
O. Lat. we have also vicissatim, with that substitution of -atim for -im, -m, which
we see in two forms mentioned by Paul. Fest. 79. 12 Th., interatim for intérim
and interduatim (cf. Plaut. Truc. 882?) for interdum. Pré-tinus (also spelt
pro-tenus; see Georges, Lex. Wortf. s.v.) meant (1) forward, onward (of
§ 4.) ADVERBS. 557

space or time), e.g. en ipse capellas Protenus aeger ago, Virg.; sic vives
protinus, Hor. ; (2) without interval of space, e.g. Virg. 4. iii, 416 cum
protinus utraque tellus Una foret, of the traditional connexion of Italy with
Sicily ; (3) without interval of time, forthwith (its usual sense); qud-tenus
(also spelt qua-tinus ; see Georges), whose earlier form quatenos is quoted by
Festus (346. 34 Th.) from a speech of Scipio Africanus, meant originally
(2) ‘as far as,’ of space, then (2) ‘as long as,’ of time, then acquired a causal
sense like our ‘in so far as,’ a usage found first in Lucr. ii. 927, &¢., but
avoided by the classical prose writers, as well as by Virgil ; in later Latin it
took also (4) the sense of ‘ quomodo,’ ‘qua ratione,’ (5) the final sense of ‘ut,’
(6) the consecutive sense of ‘ut,’ (7) the force of an Acc. before an Inf., and
was very widely used (see A. L. L. v. 399). (On the Prep. tenus, see § 54.)
Adverbs in -am are usually called Accusatives Sing. Fem., though a new
theory makes them relics of the old A-stem Instr. Sg. in-am, of which -A was
a byform, and explains Gk. Adverbs in -a like rdya, da, xpipa as Instru-
mentals in -m; see J. F. i. 17). Other examples quoted are: perptram,
falsely, wrongly, said to be Acc. Sg. Fem. of perperus (cf. C. G. L. iv. 141. 19
‘ perperum,’ perversum), sc. viam, an Adjective used by the Dramatist Accius
in his poem on stage technique (Pragmaticon, fr. inc. i. M.), a precursor of
Horace’s Ars Poetica :

. discribere in theatro pérperos

popularis,

along with the derivative noun perpéritiido (fr. ine. ii. M.):

ét eo plectuntir poetae quim suo uitio saépius
duetabilitate dnimi nimia uéstra aut perperitidine ;

protinam, forth, forward, an O. Lat. Adverb, used always of motion forward
from a place of rest (Langen, Beitr. p. 163), e. g. Plaut. fr. 16 G. dare pédibus
protinam sése ab his regiénibus, and sometimes wrongly written protinus in
MSS. of Plautus and Terence (e. g. Ter. Phorm. 190), seems to be an Acc. Sg.
Fem. of an Adjective-stem *pritino-, as continud is Abl. Sg. Neut. of the
Adjective-stem continuus ; pro-miscam (e. g. Plaut. Pseud. 1062 ut med, laetitia
laétus promiscém siet; cf. Paul. Fest. 281. 1 Th.) is similarly related to the
usual pro-miscue (also promisce).

Of Accusative Plural, dlids (sc. vices), at other times, altéras quoted (appar-
ently from Cato) by Paul. Fest. 20. 8 Th., wtrasque, on both occasions, used by
the historian Cassius Hemina (ap. Non. 183. 25 M.: in Hispania pugnatum
bis. utrasque nostri loco moti), and by the comedian Caecilius Statius

(Com. 225 R.:
atque hércle,

utrdsque te, cum ad nés uenis, subfircinatam uidi),

have been called Locative Plural forms like Gk. 6¥pdor, and the occasional use
of fords in the sense of féris (better explained as a vulgarism, e. g. Petron. go,
p. 21.10 B.. III et pridie Kalendas Januarias C. noster foras cenat, the entry on
the engagement-tablets of the wealthy parvenu, Trimalchio ; cf. ib. 47, p. 32.
4B.) has been referred to the same source ; contra is an Adverb, not a Prepo-
sition (but Pers. 13 contra me astat; cf, Psewd. 156), in Plautus and Terence, e. g.
tueri contra (ef. Liv. i. 16. 6 and ix. 6, 8 contra intueri), awro contra vendere, &e.,
and has the final vowel short like frusérd (for the quantity -¢, see Brix, Trin.’
THE LATIN LANGUAGE. (Chap. IX.

558
introd. p. 20) [ne frustré sis, ‘don’t mistake,’ is frequent in Plaut.. but frustra
sum (ef. ita swm, bene sum, &e.) is avoided by classical writers (A. L. L. ii. 3),
though /frustra in Plautus usually occurs with this verb (with other verbs
nequiquam)]. Contré can hardly be anything else than Acc. Plur. Neut. (cf. Gk.
dvria Pl., dvriov Sg.) of *con-t(e)ro- (ch. v. § 16) ; and frusiraé may be the same
case of a stem *frustro-, whence frustrari (ef. Liv. ii. 31. 9 neque frustrabor cives
meos neque ipse frustra dictator ero; Ennius in one of his Saturae (ine. 84 M.)
plays on the words frustra and frustrari for four lines consecutively: ndm qui
lepide péstulat dlterum frustréri, Quém frustratur frastra eum dicit esse
frastra, and so on, somewhat in the style of the Schoolmaster in the Merry
Wives of Windsor]; torvé tweri and similar adverbial Accusatives Pl. Neut.
belong, like dulce ridentem, to the language of poetry. (Servius ad Aen. iii. 594
cetera Graius, &c., quotes from Sallust sanctus alia ; ef. Tac. Ann. xii. 3 juvenem
et alia clarum, and see A. L. L. ii. go).

The Comparative Degree of the Adverb is in Latin (as in Greek) expressed
by the Ace. Sg. Neut. of the Adjective (the Superlative in Greek, but not in
Latin, by the Acc. Pl. Neut.), e.g. pejus (Superl. pessime), longius (Superl. lon-
gissime), The Adverb corresponding to the Compar. Adj. major (pronounced
*maj-jor,’ ch. ii. § 55) is not *majus but mégis ; this magis, properly *ma(h)is (Ose.
mais), takes its g from magnus, as figira for *fihura (ch. iv. § 116) takes its g from
Jingo, and represents either I.-Eur. *mais (*mais?), (Goth. mais), from a stem>
ma-, seen in O. Ir. mar (mir), ‘ great,’ or more probably *maghis from a stem
magh-, seen in O. Ind. méh-, ‘great’ (ef. the Oscan name Mahio-). In either
ease the suffix is -is, the weak grade of the Comparative suffix -ios, seen in
Superlatives like Gk. mAc-io-ros, Goth. ma-is-ts, &e. (ch. vi. § 52°, just as the -w
of Att, mAciy, a byform of mAefov and mAéov, is the weak grade of -ion (see on
these weak grades ch. iv. § 51, and on the Compar. suffix, ch. vi. § 53, and ef.
magis-tro-, minis-tro-, Osc. mins-tro-). This -is seems to have been confused by
the Romans with -is, the Nom. Sg. ending of I-stem Adjectives, like facilis, potis,
so that from mégis, regarded as a Nom. Sg. Mase. or Fem., was formed méigé, a
Nom. Sg. Neut., as poté from pitis (cf. Serv. ad Aen. x. 481 nunc mage sit, &e.:
‘mage’... propter metrum dictum est pro magis, sicut etiam ‘pote’ pro
potis,.. . quod adeo in usum venit ut etiam in prosa inveniatur ; Cicero in
Frumentaria: mage condemnatum hominem in judicium adducere non
posse), whence mavilo (pronounced *mavvolo?, ch. viii. § 97) for mag(e)-volo ; nimis
did not produce a parallel Neuter in -é, *nime, perhaps because there existed
already a cognate Neuter in -ium, nimium (Neut. of the Adj. nimius) with
Coinparative sense, ‘too much’ ; but sétis, a Noun meaning ‘ sufficiency’ (cf.
Jéis, § 7), and properly used in sentences like satis est mihi divitiarum, then
extended as an Adverh to sentences like satis divitiarum habeo, satis dives sum,
developed a Neuter *sute, curtailed to sat‘); plus, if plous on the S.C. Bacch.
(¢. I. L. i. 196, of 186 B. c.) be merely an expression of the sound plas (as plou-
ruma, Clout] on the epitaph of the actor, i. 1297, seem to show ou for a,

 

' Another explanation of sat makes (ef. pénéria). Satin (with the In-

it Neut. Sg. of a stem satu- (cf. Lith.
sotus, ‘satisfying’), whence, with ad-
dition of the suffix -ro-, was formed
the Latin Adj. sdtir (stem satu-10-).
Paene may similarly represent *paenti

terrog. Particle -né) is used in ques-
tions, e.g. satin abiit? ‘has he gone?’
Terence puns on satin and satis in
Phorm. 683: Satin ést id? Nescio
hérele: tantum itissus sum.
§ 5.] ADVERBS. ‘ 559

ch. iv. § 37), will have, like magis and nimis, the comparative suffix -is, and will
represent *plo-is (ef. ploera, ‘ plura,’ an archaism used by Cicero in his laws (Legg.
iii. 3. 6), plo-ir-ume, ‘plurimi,’ on one of the oldest Scipio epitaphs (i. 32, end
of the third cent. B. c, ?)], with root plo-, a variety of the root ple- of Greek
mAciwv, mAeioTos (ch. vi. § 55). That forms like mage, &c. are not mere
expressions of a tendency to drop final s in pronunciation (ch. ii. § 137), we
see from a line like Plaut. Poen. 461 : conténtiores mage erunt atque auidi
minus, where the final of the word is elided (ef. ch. viii. § 78 on -ré and -ris
in 2 8g. Pass.).

For plérum-que, plerum is used in a passage of the historian Sempr. Asellio
(ap. Prise. i. p. 182. 13 H. ut fieri solet plerum, ut in victoria mitior man-
suetiorque fiat), an Acc. Sg. Neut. of the O. Lat. Adjective plerus, used for
example by Pacuvius, Trag. 320 R. :

periére Danai, pléra pars pessim datast.

Other examples of the Adverbial Acc. Sg. Neut. of I-stem Adjectives are:
voliip for volupé (ch. iii. § 36) ; impiiné from impunis, a compound of in and poena ;
vilé, an Adverb in common use in the time of Charisius (116. 7; 187. 7; 183.
14 and 18K.) ; fidel¢, quoted from Plaut. Cupt. 439 (fac fidele sis fidelis) by
Nonius, 512. 59 M. ; sublimé, aloft. For this last we often find sublimen, as in
the Plautine expressions sublimen rapere, ferre, auferre, which is nothing but an
Adverbial word-group sub limen, ‘ under or up to the lintel’ (on the confusion
in MSS. between sublime, sublimem and sublimen, see Ritschl, Opuse. ii. 462) ;
saepé, Neut. of an old Adj. *saepis (whence saepio) of which the Superl. is
possibly to be read in Plaut. Pers. 633 :

Ubi rerum omnitim bonarum cépiast saepissuma,

‘the most closely packed store,’ ‘the densest store’ (fréquens seems to be con-
nected with farcio by a similar transference of meaning).

Gellius (x. 1) says that Pompey consulted various authorities, and finally
Cicero, on the question whether he should write tertium or tertio consul in
the dedicatory inscr. on the temple of Victory, and by his advice wrote merely
tert. The distinction between the two words is a slight one, ‘during the
third year’ and ‘in the third year.’

§ 5. Ablative (Instr.) and Locative Adverb-forms. The normal forma-
tion of Adverbs from O-stem Adjectives was in -2, older -2d (e. g. fucilumed on
the 8. C. Bacch., ¢. ZL. L. i. 196; cf. Falise. rected, Osc. amprufid ‘ improbe,’
Umbr. rehte, totce ‘ publice’), a suffix which in classical Latin was distinctive
of O-stems as -(i)ter of I- and Consonant-stems. The ending -d (the Abl.
suffix, originally -dd ; possibly in some words the Instrumental suffix, origi-

nally -d, but see below) competes, as we have seen, with -¢ in some Adverbs
formed from O-stem Adjectives, and is exclusively used by good writers in
stibito (cf. O. Lat. desubito), omnino, mérito (meritod, C. I. L. i. 190), the Ab].
of meritum, desert (cf. Plaut. Asin. 737 meritissimo ejus), &. Charisius
seems to say that the use of -o for -e in Adverbs was a feature of some
dialects of Latin (193. 16K. non quia negem ultra Safinum interque Vestinos
Teatinis et Marrucinis esse moris e litteram relegare, o videlicet pro eadem
littera claudentibus dictionem). Adverbs formed from O-stem Nouns have -4,
e. g. modo, with Adverb or Preposition prefixed in postmodo, pripémodo [these
560 THE LATIN LANGUAGE. [Chap. IX.

are not early forms ; postmodo is indeed read in Ter. Hec. 208, but propemodum
is the only form used in the old Drama. Asinius Pollio (Cie. Fam. x. 33. 1)
seems to have introduced the form postmodo into prose, and Livy prefers it to
postmodum (see Neue, ii*. p. 600)] ; O. Lat. antigerio (antiqui pro valde dixerunt,
Paul. Fest. 6. 18 Th. ; vel admodum vel imprimis, Gl. Placid. ; ef. Gl. Philox.),
a word described by Quintilian (viii. 3. 25; ef. i. 6.40) as quite obsolete in his
day, apparently Abl. Sg. of a Verbal Noun, an I0-stem (ch. v. § 4), anti-gerium
from anté and géro, a carrying in front ; 0. Lat. niiméro, quickly, or too quickly,
e.g. Plaut. Men. 287 numero hue aduenisad prandium, whence Varro (ap. Non.
352. 32 M.) derives the name Numerius, ‘prematurely born,’ a Beneventan name
first introduced into the Roman aristocracy, according to the tradition, by a
Fabius who married the daughter of a rich citizen of Beneventum (Fest. 178. 32
Th.) ; nwmero, probably used originally with Verbs of motion, may mean ‘ with
musical note or rhythm,’ like Germ. nach Noten ; vulgo from vulgus ; principio
(rarely Ace. Sg. principium) ; impendio (Abl. of impendium, outlay), used as an
Adverb (Gell. xix. 7. ro translates it by impense) by the Republican Dramatists,
especially with magis, minus, e.g. Ter. Eun. 587 impéndio magis animus
gaudebét mihi. Instead of forté, the Nom. fors is sometimes used (e. g. Virg.
A. ii. 139) with an ellipse of sit an [ef. forsitan (first in Terence’) and forsan (first
in Luer.), often written forsitam and forsam (see Ritschl, Opuse. ii. 570) by Anal.
of Advbs. in -am? ef. p. 69], while fortasse and fortassis (in O. Lat. followed by
Acc. and Inf, e.g. Plaut. Asin. 36 ubi fit polenta, té fortasse dicere, but also,
e. g. Rud. 140 fortésse tu huc uocdtus es ad prandium) seem to be parts of
a verb *fortare, to assert. affirm, derived from O. Lat. forctus, strong, as af-fir-
mare from firmus. Opéré, ‘Abl.’ of opus, appears in a great many Adverbial
expressions, e.g. magnigere, tantipere for magn(d) opere, tant(d) opere (cf. Plaut.
Mil. 75 me opere orauit maxumo). Rité is a similar formation from *ris, a by-
form of ritus (cf. O. Ind. rtt-. ‘the fitting time,’ esp. for a sacrifice, rta-, ‘fitting,
suitable,’ P. P. P.. rténa, ‘fitly, duly,’ Adv. (Instr.)].

Téméré (on the quantity of the finale, see A. L. L. iv. 51) is either the Loc. Sg.
of a lost Noun *temus (ef. O. Ind. taimas-, ‘darkness’), lit. ‘in the dark,’ or Ace. .
Sg. Neut. of a lost Adj. *temeris, lit. ‘darkly, blindly’ (cf. O. Lat. témériter). On
the adverbial Locative cases of Nouns in common use, htimi, démi, militiae, &e.,
see ch. vi. §§ 37-38. Diz whether with original -@ (cf. eod die, C. I. L. xi. 4766, an
inser. where -d is not invariably written after a long vowel ; die noine or dze noine,
‘on the ninth day,’ on the Dvenos inser., if this reading be the correct one)
or with original -2d [cf. Falise foied, apparently ‘hodie ’ (Not. Scav. 1887, pp.
262 and 307): foied vino pipafo kra karefo ‘ hodie vinum biham, cras carebo’ |
occurs in a great many Adverbial word-groups, e. g. postr?-die (ef. die crastint),
pri-die, cotti-die and coti-die (spelt quotidie only by precisians, Quint. i, 7.6), péren-
die, hi-die, méri-die (see ch. iv. § 112), from which was formed the Noun meridies.
Cotti-die can hardly represent anything but a compound of dies and the word
quot in some form or other (ef. quot Kalendis?, Plaut, Stich. 60 ; quotannis, and
in Late Latin quot diebus and quot dies) with co- written for quo-, its equivalent
in sound’ (ch. iv. § 137), but what that form was is not easy to say. Some make it
quo-tus, a correlative of totus (ef. totos dies, Plaut. Au’. 73; totis horis, Mil. 212);

 

‘ Whether forsitun really occurs in 2 cotidie, the reading of the MSS. in
Terence is doubtful. (Fleck. Jahrb.  Plaut. Stich 165, ischanged by editors
1894, p. 284.) to quot dies.
§ 6.] ADVERBS. 561

others postulate a *quot-tus (beside qué-tus from quo-) formed from quot- with
that TO-suffix which is used in Ordinal Numerals, quar-tus, quin-tus, &e., just
as quot-umus (Plaut. Pseud. 962. 1173) is formed on the type of septumus ; others
again suppose *quit(?)tus (O. Ind. katithd-) to have been a byform of quétus ;
peren-die, the day after to-morrow, is connected with Ose. perum, without
(originally ‘beyond’; cf. Gk. wépa), of the phrase perum dolom mallom ‘sine
dolo malo,’ and means literally ‘on the beyond day’ ; hé-die seems to join to
die the bare stem h65- (but see ch. iii. § 51 on si-quidem). (For other Abl. and
Loe. Adverb-forms, see ch. vi. § 38.)
adem, at the same time, always with the Fut. or Fut. Pft. in Plautus, e. g.

Trin. 577:

i hae, Lésbonice, mécum, ut coram niptiis

dies cénstituatur ; eddem haec confirmabimus,

is occasionally found without ellipse of dpéra, e.g. Capt. 449:

séquere me, uidticum ut dem & tarpezita tibi :
eddem opera a praetére sumam syngraphum ;

mad may, in some uses at least, have sprung from a similar phrase; cf.
Pseud. 318: be
quia pol qua opera credim tibi,
una opera alligém canem fugitiuam agninis Jdctibus.

Of Adverbs in -2 from O-stems may be noticed : valde [the full form vilide is
found in Plautus, Pseud. 145 (AP), &c.] from validus; Cicero was the first to
use it with an Adj. ; fe? and fermé are related as Positive and Superl. (ef.
p. 185).

§ 6. Adverbs in -tus. The best established in classical usage were anti-
quitus, dwinitus, providentially, funditus, lit. ‘from the ground,’ ‘from the
bottom’ (derived from fundo-, O. Ir. bonn, Gael. bonn, ‘the sole, the ground
or base’; cf. O. Ind. budhnda-, Gk. rv@ynv and rivdag, the bottom of a vessel,
0. Engl. botm), pénitus, lit. ‘from within’ (see § 37 on pénés), rddicitus, intus.
Intus is used not only of motion from within, ev. g. Plaut. Men. 218 euocate
intus Culindrum, but also like Greek évrds (cf. éx7dés), of rest within, e. g.
Capt. 192 ibo intro atque intus subducam ratiinculam ; cf. Lucil. ix. 59 M., who
explains the distinction between ad and apud as the same as that between
intro and intus:

intro nos uocat ad sese, tenet intus apud se.

Quintilian declares the use of intus in the sense of intro (motion to within) to
be a solecism (i. 5. 51). (The suffix -tus, when added to another preposition,
sub, has the same variety of meaning; for subtus, like intus, denotes not merely
motion from, but also rest in, e.g. uti subtus homo ambulare possit, Cato,
R. R. xlviii. 2). But in the older and the later literature many other of
these Adverbs occur. Nonius in his eleventh book, which deals with the
Adverb forms of the older writers, cites commitnitus, publicitus, pugnitus, with
the fist, himdnitus, immortdlitus, largitus, ‘pro large’ ; and elsewhere mentions
dinimitus, germanitus, médullitus [from the Saturae of Ennius (1. 7 M.):

Enni poeta, sdlue, qui mortdlibus
uersts propinas flammeos medullitus],

00
562 THE LATIN LANGUAGE. (Chap. IX.

dctilitus, primitus, sollemnitus, &e. From the Adverb penitus Plautus coined the
Superlative Adj. penitissumus (see Varro, Men. 522 B. ut ait Plautus ‘peni-
tissumae’), which is common in Late Latin. He uses in one passage an
Adj. penitus (Asin. 40 usque ex penitis faucibus), which also found currency
in the late literature (ef. penit@ Adv., Catull. lxi. 178).

§ 7. Adverbial word-groups and compounds. Antioper mpo TovTou, Gloss.
Philox., seems, if the reading is right, to represent *antia [Acc. Pl. Neut. of
*antio-, Gk. dvrios : cf. antiae (se. comae), front curls], with the Preposition per
(on o for u after i in the unaccented syllable, we. g. filiolus, see ch. ili. § 18), as
pérum-per represents pdrum (Ace. Sg. Neut. of *par-o-, little, « byform of
par-uo-, parvus) with the same Preposition (cf. the Umbro-Oscan Numeral
Adverbs, Umbr. triiu-fer, ‘three times,’ Osc. petiro-pert, four times; on the
form pert, see § 38). Topper, a word quite obsolete in Quintilian’s day
(i. 6. 4o ab ultimis et jam oblitteratis repetita temporibus, qualia sunt
‘topper’ et ‘antigerio’ et ‘exanclare’ et ‘ prosapia,’ et Saliorum carmina vix
sacerdotibus suis satis intellecta) seems to be a similar word-group with an
Ace. Sg. Neut. *tod (from the Pronominal-stem to-, ch. vii. § 13; cf. is-tud)
governed by the Preposition per ;, according to the Roman grammarians (see
Festus, p. 532 Th.) the word had two meanings, (1) ‘ cito, celeriter, temere,’
e.g. Liv. Andronicus (a translation of Homer, Od. viii. 138 sq.) :

ndimque nullum peius macerat humd4num
quamde mare saéuom, uis et cui sunt magnae ;
tépper cérpus confringent inportinae tndae,

and in his translation of Bk. xii. 17:
topper citi ad aédis uénimus Circdi,
and of Bk. x. 395:
topper ficit hémines ut prius fuérunt ;
(2) ‘ fortasse,’ a later sense, e. g. Pacuvius (Trag. 424 R.):
topper tecum sit potestas fixit si mecim uelit,

and the historian Caelius Antipater (cotemp. of C. Gracchus): eadem re gesta,
topper nihilo minore negotio acto, gratia minor esset. Sem-per may likewise
have as its first element an Acc. Sg. Neut. *sem (Gk. év), and as its second
the Preposition per (sempiternus! may then be a formation on the analogy of
aeviternus ; but see Suppl. Arch. Glott. Ital. i. 58). These examples suggest that
in paulis-per?, tantis-per, aliquantis-per the first element may be a Neuter-stem in
-is, possibly a weak grade of Comparative -ius (cf. mégis, p. 558), and not, as
is usually thought, an Abl. (Loe. Instr.) Plural form, paulis, ‘by littles,’
‘little by little.’ The -per of these words however may be the Greek -aep of
Gonep, xabdnep, &c, (ch. x. § r), (On nuper, see § 2.) Impraesentiarum is more
naturally analyzed into in praesentia rerum (a phrase which actually occurs in

 

' Explained as *sempe(r)ternus with 2 Paulisper, with the tall form of I
loss of 1 by Dissimilation (ch. ii. to indicate the long quantity, occurs
§ 1c3.. in C. I, L. vi. 27788.
§ 7.) ADVERBS. 563

its full form ; see A. L. L. iv. 11), with suppression of one of two similarly
sounding syllables (see p. 176), than into in praesentia harum, with ellipse of
rerum. Another adverbial word-group, consisting of a Preposition with its
Noun, is affitim, for ad-futim from a lost noun *futis, weariness (cf. fatigo), v. g.
Liv. Andr. Com. 5 R. affatim edi, bibi, lusi; Plaut. Poen. 534 bibas... usque
affatim, where affatim (perhaps better written ad fatim, for an early affatim
would have become *affetim) means ‘abundantly’; it is also used with a Gen.
in the sense of ‘abundance,’ e.g. Plaut. Men. 457 affatim hominumst. Paul.
Fest. 8.34 Th. says: Terentius ‘ affatim’ dixit pro eo quod est ad lassitudinem,
which shows that the grammarians of the Empire still realized that affatim
was a combination of the Preposition ad with a Noun (like ad saturitatem,
Plaut. Rud. 758; ad rivim, Aul. 336, Cist. 304; praeconis ad fastidium, Hor.
Epod. 4. 12). Indeed from the words of Gellius (vi. 7) we gather that in the
second century a. p. the Adverb was divided in spelling and pronunciation
into two words ad fatim, for he speaks of the pronunciation dffatim (like
ddmodum) as unusual, only to be defended on the supposition that the phrase
was one word and not two (quod ‘affatim’ non essent duae partes orationis,
sed utraque pars in unam vocem coaluisset). It was possibly this pronuncia-
tion ad fatim that gave rise to the curtailed Adverb fatim, which was in use in
the time of Servius (4th cent. a. p.) (Serv. ad den. i. 12g ‘ fatim’ enim abun-
danter dicimus). Amussim (Paul. Fest. 5. 3 Th.), Accus. of amussis, a carpen-
ter’s rule, may be a similar late curtailment of ad amussim, though we find
examussim as early as Plautus. Another word-group of the kind is sédiilé,
from sé, a Preposition used in O. Lat. in the sense of its cognate sine (§ 51),
and délus. From the Adverb se-dulo (with w for 6 in the unaccented syllable,
ch. iii. § 26: was formed the Adjective sedulus. That this is the true account
of sedulo is clear from se dulo malo of the Lex Agraria (C. I. L. i. 200. 40), and
from the comparison of phrases like Plaut. Trin. 90 hawd dicam dolo, 480 non
tibi dicam dolo, Men. 228 non dicam dolo, with Capt. 886 quod ego dico sedulo, Ter.
Phorm. 453 ego sedulo hunc diaisse credo (but sedulo is commoner with facio than
with dico) ; though the confusion of sedwo with an Adverb from an Adjective-
stem sedulo- must have been very early, for the byform sedulum is found in a
line of ‘Plautus’ (fr. 41 G.) sedulum est, and in a plebiscitum about weights
and measures quoted by Festus (322. 3. Th.) : ex ponderibus publicis, quibus
hac tempestate populus oetier qui solet, uti coaequetur sedulum, uti quadran-
tal uini octoginta pondo siet, congius uini decem pondo siet, &e. (cf. later
sedule). The Roman grammarians usually explained the word in this way ;
Acron however, in his commentary on Terence, derived the word from an Ad-
jective-stem sedulo-, as falso from the Adj. falso- (ap. Charis. 192. 30; 219. 5 K.).
Another combination of Preposition and Noun is ob-viam, which has its literal
sense in Plautus (usually with esse, ire, venire, &c.); thus in Capt. 791 the
parasite Ergasilus, who is hurrying to give Hegio news of the arrival of his
son, cries out :
éminor intérminorque néquis obstiterit dbuiam,

‘that no one stand in my way’; it has not the sense of praesto in Plautus nor
jn Terence. From obviam was formed the Adjective obvius (but cf. pervium,
pervius), as from sedulo, sedulus. Like obviam (and inter-vias with vias Acc. PL.) is
obiter, a word regarded with suspicion by purists, though Augustus gave it his
sanction, and reproved Tiberius for using per viam instead. The Emperor
Hadrian seems to have reversed his predecessor’s decision (Charis. 209. 12 K.

002
564 THE LATIN LANGUAGE. [Chap. IX.

‘obiter’ divus Hadrianus Sermonum I quaerit an Latinum sit... quan-
quam divus Augustus reprehendens Ti. Claudium ita loquitur: seribis enim
‘per viam’ dyzi 70d obiter). Ilicd is explained by Charisius (201. 17 K.) as in-
lico (the length of the initial 7 could be accounted for by the old form sloco,
ch. iv. § 150), and has often this sense in Old Latin authors (for examples, see
Charisius, 1. v. and Nonius, 325. 6 M.) (in Plautus the local sense is not nearly
so common as the temporal, and is found only with verbs of rest, e.g. Rud. 878
ibidem ilico manete ; 836 illic astate ilico) : Nonius (1. c.) makes it the equivalent
of ‘in eo loco,’ but the sense of ibi is foreign to ‘ico, a fact which tells against
its derivation from *2, an old Abl. Sg. of is, and locus. Jlico has not the sense
of ‘ on that spot,’ but of ‘on the spot’ (e. g. ilico hic ante ostium, Plaut. Lrin.
608), and its change from wu local to a temporal meaning finds an exact
counterpart in our phrase ‘on the spot,’ which may be used in the sense
of ‘immediately,’ ‘ without delay.’ Curiously enough the O. Lat. adverbial
word-group iicet, which in the Comedians has the sense of ire licet (e. g. Plaut.
Capt. 469 :
ilicet parasiticae arti méxumam malim crucem,

‘the profession of diner-out may go hang itself on the highest possible
gallows’), from which by an easy transition it acquired that of actum est [e. g.
Plaut. Cist. 685 perii, opinor, actumst, ilicet, was brought again into fashion
by Virgil in the sense of ilico, e.g, Aen. xi. 468 :

ilicet in muros tota discurritur urbe.

(Charisius quotes the note of a commentator on this line : (ilicet) nunc pro
ilico, id est statim. antiqui pro eas licet,’ and mentions-a somewhat similar
use of the word in a line of Afranius (Com. 215 R.): an tu eloquens ilicet?,
‘have you become an orator all at once?’]. By another freak of language
the confusion of ilicet with a word of similar sound was repeated some cen-
tures later, when it was used (e. g. by Sidonius Apollinaris, fifth cent. a. D.)
in the sense of scilicet (ef. Paul. Fest. 74. 22 ‘ilicet,’ sine dubio). A verbal
group like ilicet is vidélice, which in the earlier period occurs with the con-
struction of vidére licet, e. g. Plaut. Asin. 599:

nune enim ésse
negétiosum intérdius uidélicet Solénem,
Lucret. i. 210:
esse videlicet in terris primordia rerum,

also scilicet, construed like sctre licet in such a line as Plaut. Pseud. 1179: scilicet
solitum esse, ‘of course he used to’ (on these verbal Noun stems /-, vide-, sci-,
see ch. viii. § 34). Sts (for st vis, ch. ii. § 53), sultis Plur., similarly takes an
Inf. in Plaut. Asin. 309 sis amanti stibuenire fimiliari filio.

The word médus enters into several adverbial word-groups, Besides admo-
dum, propemodum (and later propemodo), postmodum (and postmodo), which have
been already mentioned, we have quemadmodum, quimodé [cf. cujusmodi, hujus-
modi, ejusmodi, &c., which in Plautus are scanned as cretics or the equivalents
of cretics (ch. vii. § 22), cuicuimodi (perhaps not in Plautus)], quidammodo, tan-
tummodo and (in Late Latin) solummodo, &c. (ef. the compounds omni-modis, multi-
modis, on which see Neue, ii*. pp. 609 sq.). We find rés in qua-r2, quam-ob-rem,
&e., dies in in-dies, prope-diem.

A Preposition with a Noun (or Adj.) appears algo in dé-mtid, for de nivo (with
§ 8.] ADVERBS. 565

u for unaccented ov, ch. iii. § 24), like de integro, ex-templo (in Plaut. also ex-
tempulo), from templum in its O. Lat. sense of locus (e. g. Acherusia templa, Enn.),
in-cassum, lit. ‘into the empty’ (ef. cassa nua, Plaut.), like its synonyms in Late
Lat. in vanum and in vacuum ; im-primis, cum-primis (ef. apprimé, used with an
Adj. in the sense of a Superlative in ante-classical and post-classical Latin).
A verbal phrase is seen in dum-taxat [taxat is Pres. Subj. of *taxo, a byform of
tango, as viso of video, quaeso of quaero (ch. viii. § 33. 4) for *tag-so], a legal phrase
whose original sense appears in O. Lat. laws liko C. I. L. i. 197. 12: sei quis
magistratus multam inrogare uolet, (quei uolet, dum minoris] partus familias
taxsat, liceto, ‘so long as he assigns a fine of the smaller portion of his
property,’ ‘of less than half of his property’ (for dumtaxat the Oscan expression
is ampert from the Negative particle an- and the Preposition pert, ‘beyond’ ;
thus on the Tabula Bantina: in[im] svae pis ione fortis meddis moltaum herest,
ampert minstreis aeteis eituas moltas moltaum licitud ‘ et siquis eum fortius (? forte)
meddix multare volet, dumtaxat minoris partis pecuniae multas multare
liceto’) (Zvet. I. I. I. 231. 12’. Acta-tum is merely acti, lit. ‘on the act,’
followed by tum, then.

§ 8. Other Adverbs. One of the puzzles of Latin etymology is the O. Lat.
Adverb simiti, the equivalent of simul, found as an archaism on some inscrip-
tions of the Empire in the forms simitu (C. I. L. vi. 7578), and certainly once
(possibly twice) simitur (vi. g290, a slave-girl’s epitaph of 13 B.c., and read by
Ritschl in x. 174, an epitaph in illiterate verse). Nonius mentions simitu
(175. 16 M.) with three examples from Lucilius and Plautus ; so that this
form is well established. Plautus elides the last syllable in Stich. 249 mecim
simitu ut ires ad sesé domum. The later (plebeian) form simitur can hardly
be due to a change of -d to -r like that seen in O. Lat. apor, apur for apud (§ 19) ;
it is more likely to be a corruption caused by confusion with the Impersonal
itur, 3 Sg. Pass. of eo, to go. Simitu has been explained as a Compound of the
root sem- (ch. vi. § 57) in some form or other with iti, the Abl. Sg. of the
Verbal Noun ttus, a going.

Another puzzle is igitur (the quantity of the final syllable cannot be deter-
mined in Plautus), the oldest sense of which is ‘then,’ ‘ thereupon’ (cf. Non.
128. 14 ‘ igitur’ positum pro postea), e. g. the first clause of the XII Tables: si in
ius uocat, niit, antestamino : igitur em capito, a clause well known to Roman
schoolboys of the first cent. B.c. as the beginning of one of their most formidable
lesson-books (ef. Cic. Legg. ii. 4.9 a parvis... Quinte, didicimus ‘ si in ius vocat*
atque alia ejus modi ‘leges’ nominare); Plaut. Cas.215 m6x magis quom dtium ét
mihi et tibi erit, [gitur tecam loquar: nunc uale; Mil. 772 quéndo habebo, igitar
rationem medrum fabricarim dabo; igitur tum (e.g. Most. 689), igitur deind?, igitur
demum are all frequent in Plautus. This makes unlikely the theory that igitur
is a curtailment of quid igitur ?, a form of quid dgitur ?, with a weakened to ¢in
the unaccented syllable of the word-group quid-agitur ? Another theory con-
nects it with Lith. -ktu of téktu, ‘so,’ kdktu, ‘as,’ or with O. Sl. -gda of togda,
‘then,’ igda, ‘hitherto.’ It has also been resolved into the three Pronominal-
stems i- (ch. vii. § 13), ko- (4b. ; digitus is quoted as a parallel case of the change of
-icit- to -igit-; others make the particle *gé (Gk. ye) the second element of igitu],
and to- (ib.) ; the last with the ending -r, seen in O. Ind. tér-hi, ‘then,’ kar-hi,
‘when,’ &.). Its use varies in different authors ; thus it is placed normally
as first word of the sentence in Sallust and Tacitus, but almost never in
Cicero’s speeches; and it is avoided by purists like Terence (except in his
566 : THE LATIN LANGUAGE. (Chap. IX.
earliest plays), Lucilius, Caesar, and the older and the younger Seneca (for
statistics, see A. L. L. iii. 560). Its formation evidently seemed to the Romans
to have something irregular about it (cf. Quint. i. 5.39 ex quo genere an sit
‘igitur’ initio sermonis positum dubitari potest, quia maximos auctores in
diversa fuisse opinione video, cum apud alios sit etiam frequens, apud alios
numquam reperiatur). Another Adverb (?) of uncertain etymology is
siremps in the formula found on old laws, siremps lex esto, ‘the same law shall
hold’ (¢.1. L. i. 197. 12 3 198. 73-3 200. 27 ; 202. (1). 38, (2). 1, &e. ; see Georges,
Lex. Wortf. s. v.), and in the form sirempse (which Charisius calls the Ablative
of siremps, 93. 24; 146. z K.) in the (un-Plautine ?) prologue of the Amphitruo
of Plautus, 1. 73 :

sirémpse legem iussit esse Iuppiter.

One theory declares it to be composed of sis (si vis} and empse, the old Acc. Sg.
Mase. of ipse (ch. vii. § 20), and to have originally meant ‘the aforesaid’
(person), but to have been in course of time wrongly associated with lex esto
(Wien. Stud. 1891, p. 296).

Setius, less (esp. non setius, nihilo setius, also quo setius like quo minus), is the
correct spelling, not sécius, which is not found till the period when t began to
be confused with ci (see ch. ii. § go, and Fleckeisen, Fiinfzig Artikel, p. 28), so that
the word cannot be equated with Gk. focwr (for *fxywy ; ef. fxioros). Though
treated by the Roman grammarians as the Comparative of the Adverb (Pre-
position) sécus and as a byform of the normal comparative séquius (e.g. Afra-
nius 293 R. sin, id quod non spéro, ratio télis sequius céciderit) (from the
root seq"-, ‘to follow,’ Gk. éroya:; lit. ‘following,’ hence ‘inferior ’), it seems
impossible to connect sétius with secus (see § 1) or with sequius. Gellius (xviii.
9. 4), describing a controversy he had heard over the spelling of the O. Lat.
verb instco (insequo), to narrate (Gk. éomov), mentions as an argument adduced
by one of the disputants, that the form sectius was found in Plaut. Men. 1047
(our MSS. in this line of Plautus, a line unfortunately undecipherable in
the Ambrosian Palimpsest, vary between setius quam, sed usquam, and secus
quam) :

haéc nihilo esse mihi uidentur sétius quam sémnia,

and that this sectiws was connected with the O. Lat. insectiones ‘narrationes,’
so that the line really meant ‘nihilo magis narranda esse quam si ea essent
somnia.’ On the strength of this very questionable authority, it has been
attempted to connect sétius with the root seq’-, ‘to follow,’ by the theory that
séctius was the oldest form (from secto-, a participial-stem from sequor ; ef. secta,
sector), which became *stitius, then sétius. But even granting the existence of
this form sectius, a form most probably due to an error in the copy of Plautus
used, the change of ct to tt is a late one (ch. ii. § 95), and the substitution of
at for éttoo doubtful (ch. ii. § 130) to make this theory at allconvincing. It is
much more likely that setius and secus were entirely different words (like mélius
and béné, pejus and mdlé), associated as Comparative and Positive owing to
their similarity of meaning.

Oppidd is another of the problems of Latin etymology. The most likely
theory is that oppidum, a town, meant originally the part on the plain (ob-
*nedum ; ef. Gk. meSiov) as opposed to the arxz. This distinction seems to be
expressly made in the Titulus Aletrinas (¢. I. L. i. 1166) in a passage refer-
ring to the water supply of atown: aquam in opidum adqu[e]arduom. The
§§ 9, 10.] ADVERBS. 567

Adverbial Abl. oppido will then be exactly similar to pléné. In Plautus it
goes usually with a Verb, but in Terence only with an Adj.

§ 9. Numeral Adverbs in -ies (see ch. vi. § 56).

§ 10. Pronominal Adverbs. (On the Relative Adverbs which show
initial v- in the simple word, but as the second element of a compound -cu-,
e. g. Ubi, si-cubt, unde, si-cunde, see ch. vii. § 26).

(t) With suffix -bi (Umbr. -fe, Ose. -f, e.g. Umbr. pufe, Ose. puf, ‘ubi,’
Umbr. ife). The Umbro-Osean forms point to -fi as their original suffix,
with f representing I.-Eur. bh (Gk. -gz), as in the Dat. Sg. of the Personal
Pronouns (e. g. Ose. sifei ‘ sibi,’ I.-Eur. *sebhei, ch. vii. § 5), or more probably
dh (Gk. -6; cf. 0. Sl. kide, O. Ind. ktha). Latin -bi is in O. Lat. -be¢ [e. g. ubei
on the S. C. Bacch. (C. J. L. i. 196) ; other examples in Georges, Lex. Wortf. s. v.-;
on the spelling ube, e. g. Varro R. R. i. 4. 4, see ch. vii. § 6], and is in poetry
scanned as a long or as a short syllable (similarly ibidem), a variation of
quantity which is most naturally explained by supposing that the last syllable
was originally long (representing I.-Eur. -ei), and was in time shortened
through the influence of the short first syllable (avé from avé, &., ch. iii. § 42).
The Osean form however suggests that the short final vowel may not be a late
development (cf. wti- and utei, ch. x.§ 11). Latin examples of this forma-
tion are u-bi, si-cubt, num-cubi, &e., ubi-que, wi, ibidem (rarely ibidem in Plaut.»,
in-ibi [inibi esse like in eo esse (ut), ‘to be on the point of’], intér-ibi (often used
by Plautus for intered, interim), post-ibi (used sometimes by Plautus for postea),
alibi and the less usual aliubi (as early as the Lex Agraria of 111 B.¢., C. I. L.
i. 200. 86 : aliubeiue aliterue), utriibi, utrubique and utrobique (on these spellings,
see Georges s.v.), neutrubi.

(2) With suffix -7 (older -ei). In the language of Plautus ili or illic (older
ollic, Paul. Fest. 231. 2 Th.), éstz or istic may be Dat. Sg., ‘to him,’ or Adverb
(Loe. Sg.), ‘there.’ The scribes of our MSS. however have gone on the prin-
ciple of correcting ili, isti, ‘there,’ to dlic, istic, usually (e. g. Capt. 278, where
the metre requires ili), and almost always illic, istic, ‘to him,’ to ili, isti, so
that the relative frequency of the two forms of the Adverb in Plautus, or
other writers, cannot well be determined. The Adverb from the pronoun
hic seems to have had the enclitic -c(e) at all periods of the literature, hic,
older heic (C. I. L. i. 551. 590. 1007. 1009) ; heicet on the epitaph of an actor,
ie Protogenes Cloul[i] suauei heicei situst mimus,

plouruma que fecit populo soueis gaudia nuges,

may be a graver’s error, for the orthography of the inscription is erratic ;
heice (i. 1049 me heice situm inmature). But on late inscriptions we
occasionally find i (e. g. ii. 3244 hi jacet). [Faliscan he, hei in the formula
he (hei) cupat ‘hic cubat, -ant’ may owe the suppression of its -c to the
initial c- of the following word ; cf. Zvet. I. 1. I. 66.] From the Pronominal-
stem *so- (ch. vii. § 13) we have si, and with the enclitic -c(e), sic! (ef. Umbr. i-sek

 

1 For an example of the older ut rosa amoena homini est quom
spelling seic we may take this touch- primo tempore floret,
ing epitaph of a girl called Flavia quei me viderunt, seic ego Amoena

Amoena [Mitth, (rém.) viii. 150]: fui
568 THE LATIN LANGUAGE. [Chap. IX.

‘item’); from the Relative (Interrog., Indef.) the Adverb qui (which is
perhaps rightly regarded as an Instr. (ef. O. Engl. hwi, 0. Sl. éimi) from a stem:
qui-, ch. vii. § 25), and with the suffix -n(e), guin (ch. x. § 16), as well as the
compounds alio-qui and alio-quin, cetero-qui and cetero-quin, né-qui-quam (ch. vii.
§ 28), and perhaps quippe (if for qui-pe) and quippiam (quipiam) (see ch. x. § 7;
ch. vii. § 28).

(3) With suffix -d, e. g. eo, co-dem (id-eo has only the later sense of purpose,
not the earlier of motion towards, ‘that for that purpose,’ ‘and that indeed
with that object’), quo, quo-cunque, dliquo, dilio, utro, utroque, neutro. These have
the same sense as Adverbs formed with -verswm, -s (older -vorsum, -s), and often
have this participle added, e. g. alidvorsum and aliorsum, Plaut. (for the suppres-
sion of v between the two vowels, see ch. ii. § 53), quorsus, -m. The Preposition
ad is appended in quo-ad, and prefixed in dd-eo (used in a literal sense in the
older literature, e.g. Cato, R.R. xl. 3 sureulum artito usque adeo quo
praeacueris ; cf. adeo res rediit, ‘things have reached such a pass,’ in the
Comedians), and ad-quo, a variety of quoad, ‘so far as,’ for which Nonius
(76. 6 M.) cites two lines of Afranius, Com. 278 R. :

ut scire possis ddquo te expediat loqui,
and 249 R.:
irdtus essem adqu6 liceret.

We have the same suffix -d in Adverbs indicating motion towards, formed
from Prepositions by means of the tro- (70-) suffix, ré-tro, ci-tro, ul-tro, por-ro ;
and that this -d represents earlier -dd, the Ablative case-ending of O-stems,
we see from the spelling porod on an old Praenestine cista (Mél. Arch. 1890,
p. 303). Oscan adpud in a Capua inser. (Rhein. Mus. 1888, pp. 9 and 557.
adptud fiiet) seems to be the equivalent of Latin quoad (cf. Afranius’ adquo) and
to have the sense of ‘so long as,’ Lat. quoad fient. In Umbrian, where, as
in Latin, final -d is dropped after a long vowel, we find a-ni-po with another
sense of guoad, ‘until,’ followed by » Future Perfect, sersitu arnipo . .. pesivis
Sust ‘sedeto quoad precatus erit’ (the -ni- of arnipo seems to be like the -ni- of
Latin dé-ni-cum, ch. x. § 12, and -yo may represent *pdm rather than *péd) ; the
Adverb corresponding to Latin quo is pu-e with that suffix -i (?) (Gk, obto0-1)
written -i, -ei, -e, which is added not merely to the Nom. Sg. Mase. of the
Umbrian Relative, po-i, or po-ci, or po-e (Lat. gut for quo- with 7, ch. vii. § 25), but
to other parts of its declension, e.g. Acc. Pl. Fem. paf-e ; the Adverb from
O. Lat. séd-iiter-que (Plaut.), ‘each separately,’ is in Umbrian sei-podruh-pei,
with uh expressive of the long o-sound (a close o, nearly or altogether u; see
ch. ii. § 20). The fact that the O-stem Abl. shows -uin Umbr. has been used
as an argument that Umbr. wlo ‘illuc,’ postro ‘ retro,’ &c. are Instrumentals.
Similar formations from ile, iste are found at all periods of Latin, #lé (ef.
Umbr. ulu, ulo), ists (for examples, see Georges), but the Adverbs generally
used are illuc, istuc, with final syllable scanned long in poetry, and similarly
from hic we have huc, hither, ad-huc, hither-to (is *hd indicated by the form
horsum for hovorsum?). In the earlier literature the forms with the enclitic
-c(e) end in -oc, with final syllable again scanned long, e. g. hoc (as in Plaut.
Capt. 480, where the parasite is touting for a dinner :

quis ait. ‘hoc’? aut quis profitetur ?),

an archaic form employed by Virgil, A. viii. 423 (see the note of Servius on
the passage) :
hee tune Ignipotens caelo descendit ab alto
§ 10.] ADVERBS. 569

{for other examples, see Georges, Lex. Wortf. s. vv. ; statistics of the spelling
hoc and huc in Plaut., &. are given in Fleck. Jahrb. (Suppl.) 1891, p. 293 n.].
This -wc has been referred to an earlier -oi-ce with the Locative suffix -oi,
which is used in Adverbs of motion towards in Greek, rot, dro, &c., but it
Seems unnatural to regard it as anything but a phonetic development of the
earlier -oc. If the vowel in these endings be naturally long, it must have
been originally ou (ch. iv. § 41), and the parallel Adverbs in Greek will be
not moi, dro but od, Svov, which mean in Greek ‘where,’ not ‘whither.’
But it may have been naturally short ; for the metrical value of the syllable
can be explained by the fact that c really represents cc, from an earlier dc,
*hod-c(e), *is-tod-c(e) (ch. iii. § 51), so that the forms would be really Accusatives
Sing. Neut. (cf. however O. Lat. ile, istuc, and class. hoc, Acc. Sg. Neut.), and
not Ablatives.

Quo, eo, &c. have also other senses in Latin, as with Comparatives ¢o major,
eo minor (cf. Lith. ji with Comparatives ; Welsh po, ‘by how much the’), quo
major, quo minor, &e., from which comes the use of quo for ut in final sentences
with a Comparative, e. g. quo facilius hace fieri possint, and the word-group
quominus, lest, e. g. quominus haec fieri possint ; also quo, whereby, &c. These
too are Ablative forms, originally *quad, *eéd, as we may see from Ose. pod
+ +. mins ‘quominus’ in the Tabula Bantina (Zvet. I. 1. I. 231. 10): nep
Sefacid pod pis dat eizac egmad min{s] deivaid dolud malud ‘neve fecerit quo quis
de ea re minus juret dolo malo’; Osc. svaepod . . . svaepod ‘sive... sive,’
Umbr. svepo ; Umbr. eso (és0) from the stem *ek-so-, also esoc (issoc) for *ek-sok
‘sic’ ; with the last, cf. the Latin gloss : soc, ita (Lowe, Prodr. p. 350, a doubt-
ful form). In a leaden execration tablet (Zvet. I. I. I. 129), written in
Osean (presumably not the best Oscan), we huve svai puh (h indicates the
length of a vowel in Umbrian), a spelling which, if found on a magisterial
proclamation, or any carefully written inscription, might establish the
existence of an Oscan pd (pi), an Instrumental form, beside pod (pid),
the Ablative form. But the character of this inscription diminishes the
value of the evidence of this, as well as of the other ‘ Instrumental’ Adverb,
which occurs on the same tablet, suluh ‘ omnino’ (from the stem sollo-, ‘all,
whole’) ; besides it is doubtful whether h indicates vowel-length in Osean. In
another Oscan execration scroll (I. F. ii. 435), we have the form sullud (the
last letter somewhat doubtful), but the fragmentary state of the inscription
makes it impossible to determine satisfactorily that sullud is an Adverb.

(4) In -4, e. g. qua, ‘in which direction’ (cf. qua... qua, ‘both... and’),
ne-qua-quam, haud-qua-quam, qua-propter (cf. Plaut. Amph. 815 qua istaec propter
dicta dicantur mihi), ea (often appended to Adverbs, propter-ed, praeter-ed, post-
ed, ant-ed, and on theS. C. Bacch. arvorsum ead in the sentence: sei ques esent,
quei aruorsum ead fecisent, quam suprad scriptum est), ea-dem, alia, aliqua,
hac (appended like ea to Adverbs, post-hac, ante-hac, praeter-hac, and in a plebi-
scitum ap. Fest. 322. 8 Th. adversus hac; cf. Osc. post exac ‘posthac’), ilac,
istac. With ténus appended these Adverbs indicate distance : quatenus, ‘how
far’ (on the subsequent development of meanings, see § 4), hactenus, ‘thus
far’ (cf. Hor. est quadam prodire tenus). The formation is evidently the
same as that of Adverbs derived from Prepositions with the suffix -tro- (-ro-),
e.g. extra (exstrad 8. C. Bacch.), supra (suprad 8. C. Bacch.) (cf. Oscan pullad
‘qua,’ an Adverbial Ablative Sg. Fem. of the Relative, formed by adding the
Abl. Sg. Fem. of the Oscan Demonstr. stem ollo- (Lat. illo-, older ollo-, ch. vii.
570 THE LATIN LANGUAGE. [Chap. IX.

§ 18) to the Relative stem po- (ch. vii. § 23) ; [p]allad via uruvu ist ‘qua via
flexa (?) est,’ Zvet. I. I. I. 136. 56). Qua-ad is found for quo-ad on late inserip-
tions, and is a not unéommon variant in good MSS. (see Georges).

(5) In -im, ditrim-que, illim (found in Cicero as well as in the older writers,
whereas istim is doubtful ; see Georges, Lex. Wortf. s. vv.), but usually iline, as
istinc, hinc. The same suffix is found in the forms to which the Adverb sécus
is appended, intrin-secus, altrin-secus, extrin-secus, and apparently in exim, intérim
and délim, though in two at least of these three last words, it has not its usual
sense of motion from. Intérim is equivalent to intér-ca, inter-ibi (Plaut.), while
olim, from meaning ‘at that time’ (often answering to quum in Plautus, e. g.
olim quom caletur maxime, Truc. 65), came to mean ‘at any former time,’
‘formerly,’ ‘once upon a time,’ then ‘at any time,’ ‘ occasionally’ (e. g. Lucil.
iii. 4 M. uiamque Degrumatus uti castris mensor facit olim), and to be used
even of future time (e. g. Hor.: non si male nunc, et olim Sic erit). Olim can
hardly come directly from the Pronoun-stem ollo- of O. Lat. ollus, olle (class. ile),
best analyzed into *ol-so- (ch. vii. § 13), for du- would not become dl- (ch. ii.
§ 130); it is rather to be derived from the stem 6I-, a grade of the OL-(AL-) stem,
without the suffix so- (cf. Umbr. wlo ‘illuc’ with u the equivalent of Lat. 6).

The origin of the suffix -im has not yet been satisfactorily explained.

(6) In-ndé. This suffix seems to be the suffix -m (hardly the Ace. Sg. suffix)
augmented by the particle -de. As the Adverbs meaning ‘thence,’ ille, iste
added to -im the particle -c(e), the corresponding Adverbs from is, qui add to
-m the particle -de (cf. Gk. év@év-3e), which, like -ce, was liable to be curtailed
of its final short vowel in every-day pronunciation (ch. iii. § 36). Thus we
have inde (which should not be derived from the Preposition in, but must go
with unde), indi-dem, and with prefixed Adverb de-inde (curtailed to dein), pro-
inde (and proin), ew-inde (and exin, a different word from exim ; see Georges, Lex.
Wortf. s.v.), &e., unde, si-cunde, né-cunde, undi-que, aliunde, dlicunde. For térinde,
quoted from a speech of Cato by Charisius (224. 14 K. utrinde factiones tibi
pares) we should expect utrunde.

(7) In -wm (-om), the Acc. Sg. Neuter, e.g. tum and with the enclitic -c(e),
tune (cf. nune and etiam-num), quum, oldet quom, with appended jam in quin-
tam (ch. x. § 13), and appended -dam in quon-dam (cf. qui-dam, ch. vii. § 28), dum
(ch. x. § 12), dii-dum (ib.), non-dum. From the Pronominal-stem i- (Lat. ts) an
Adverb in -m was in use in O, Lat. in the sense of ‘then,’ variously written
as im and em (ch. vii. § 19) ; and from the Relative-stem seen in u-bi, &. we
have the um- of un-quam (cf. O. Lat. *wmguis of necumquem ‘nec umquam quem-
quam.’ Fest. 162. 22 Th.), and with negative né- prefixed, nun-quam (ne umquam
is the reading of the MSS. of Plaut. Most. 307); though some make this
originally to have ended in , not in m, and find the form with initial c (like
si-cubi, si-cunde) in the -cun- of qui-cun-que (eh. x. § 2), ne-cun-quem (so interpreting
the O. Lat. word mentioned by Festus), which they compare with Goth. -hun-
of ni hvas-hun, ‘none so ever.’

(8) In -am, apparently the Acc. Sg. Fem., though some make it an Instru-
mental case (suffix -m or -mi, ch. vi. § 36). If Festus is right in quoting tame
as an O. Lat. form of tam from the Carmen Saliare (Fest. 546. 1 Th. ‘tame’ in
Carmine positum est pro tam), all these Adverbs in -am may have originally
ended in a short vowel. From the Relative Pronoun we have quam, the cor-
relative of which is taken from the Pronoun to- (ch. vii. § 13), not from the
Pronoun i-, tam. The two are united in tan-quam. Whether jam, now (ef.
.

§ 10.] 3 ADVERBS. 571
Lith. jati, ‘ already’), is the corresponding formation from the stem i- is not
certain. The indeclinable Adj. n2-quam may be a colloquial compound of quam,
as the Adverb né-quaquam is of quaquam, so that nequam would literally mean
‘a no-how’ (ef. O. Lat. nequalia ‘ detrimenta’ Fest. 162. 23; Paul. Fest. 163.
13 Th.). Another compound is perquam, exceedingly (cf. admodum quam, Plaut.).

The addition of the Preposition dé (§ 27) to quam gives the Adverb quando
(ch. x. § 12), originally temporal, then causal (ef. quandiquidem, dli-quando,
quandd-que (Umbr. panu-pei) (ef. O. Sl. kadu, ‘qua, unde’ ?), while O. Lat.
quam-déis a byform of quam, than, as in a passage of Livius Andronicus (quoted
in § 7): peius. .. quamde mare saeuom, and Lucr. i. 640 quamde gravis inter

, Graios qui vera requirunt. Other derivatives are dliquam in aliquam-diu, &e.,
quam-vis (rarely with Subj. in Plautus, and always with an Adj. or Adv. ;
not in Terence), and quam-quam (ef. tam-quam), n(e)-titi-quam, né-qua-quam, &e.

From the pronominal-stem *no- (0. Ind. na-na, ‘in various places or ways,’
lit. ‘there and there,’ ‘thus and thus’), connected with the stem *eno- (*ono-)
(O. Ind. ana-, ‘this,’ Lith. anas, ‘that,’ 0. Sl. ont) we have nam used in empha-
sizing a question, ce. g. quid ceruss& opus nam ? Plaut. (hence quis-nam, who ?,
O. Lat. quid-nam, why ?), and in the sense of ‘for’; from the stem *do-, -dam
of quon-dam (cf. qui-dam).

(9) With other suffixes: of t-suffixes we have (q) -ta in ¢-td, so (Umbr. itek),
i-td-que, therefore (ch. x. § 8); *uta, as, may be inferred from O. Lat. ali-uta (ef. ali-
ubi, p. 564), otherwise, quoted from the Laws of Numa by Paul Fest. [4. 27 Th.
si quisquam aliuta faxit, ipsos Ioui sacer esto ; ef. the gloss altutea (leg. aliuta),
aliud, amplius, Léwe, Prodr. 432]. The final @ was no doubt originally long, if
every final short vowel became -éin Latin (ch. iii. §37), but there are no traces

ofthis quantity in dtague in the older poetry (on ttague in the Saturnian epitaph
of Naevius, see ch. x. § 8), and probably none in ita either. So that the
shortening of the final vowel under the influence of the preceding short
syllable must have established itself in this word of common use at a very
early date.

(b) -tem in i-tem, au-tem (cf. O. Ind. -tham of ka-tham, ‘how,’ &e. ?).

(ce) -t (originally with a short final vowel, probably i; ef. 0. Ind. i-ti, ‘ thus ’)
in wt, viti-nam, uti-que. The Umbro-Oscan equivalents of Latin ut show an s
after the 4, Umbr. puze, puse, ending originally in -tsi or -tsé, Ose. puz (pous
on the Bantine tablet must surely have ou for %), which in Latin appears in
the local Adverbs us-quam, nus-quam.

Us-que is a different word, derived from the I.-Eur. Preposition ud-, as ab-
s-que from ab (see § 57 for tisque).

The suffix ¢ appears also in aut, which probably ended originally in a short
-i (Gk. ad-re has re for I.-Eur. *-q3é, ch. x. § 2), as we see from Umbr. ole,
Osc. avti. The long -i of uti is probably Loe. -ei (§ 11).

Of d-suffixes, besides (a) -dam of quon-dam (temporal Adv. of qui-dam) and (0)
dum, with idea of time in diu-dum, non-dum, vic-dum, inter-dum already men-
tioned, we have (c) -dem in qui-dem (ch. x. § 6), and with the sense of ‘exactly,’
‘just’ (cf. is demum, ibi demum) in a large number of Adverbs, such as tanti-dem,
ibi-dem, indi-dem, iti-dem, titi-dem (from *toti, the older form of tot, ch. vii. § 29),
as well asin the Pronoun of Identity, i-dem (ch. vii. § 21), with its curious
derivative identidem, repeatedly (explained as ‘idem ante idem or ‘idem tum
idem’), also with the idea of time (cf. dum in inter-dum, non-dum) in tan-dem

(cf, demum), pri-dem.
572 THE LATIN LANGUAGE. (Chap. IX.

§ 11. PREPOSITIONS. Prepositions are Adverbs, which
came to be specially used in connexion with certain cases of the
Noun, or in composition with a Verb. In the early stage of a lan-
guage the cases alone were sufficient to indicate the sense, but as
the force of the Case-suffixes became weakened, or as the necessity
for clearer definition was more recognized, the Case-suffix was
strengthened by the addition of an Adverb. Thus ire monte
might mean ‘to go out of the mountain’ or ‘to go down
from the mountain.’ To indicate the first sense, the Adverb ex
was used, zre monte ex; to indicate the second, the Adverb de, ire
monte de; or ex-ire monte, *de-ire monte. These Adverbs which,
owing to their meaning, are most frequently associated with
particular cases of Nouns, or are used in composition with Verbs,
are called Prepositions ; and the process, by which Latin Adverbs
became Prepositions, may be seen in operation at various periods
of the language. Thus contra, which has hardly passed the
Adverb stage with Plautus and Terence, is a Preposition in
classical Latin and governs an Accusative Case; coram is not
a Preposition till Cicero’s time; simuéd in Augustan poetry and
Silver Age prose; retro not till Late Latin (e.g. vade retro me,
S. Mare. viii. 33, "udgata). It is customary now in writing Latin
to write the Preposition and the Verb in one word, eg. ewire, .
but not the Preposition and the Noun, e.g. ex monte; and this
practice is justified by the fact that a Verb compounded with
a Preposition had, so to speak, a separate life of its own in its
compound form. wiyo, for example, was a different word from
dgo, and so suffered weakening of the vowel a in the unaccented
syllable; pdno ceased to be recognized as a compound of po-
(see § 12) and simo, and changed its Perfect pd-stvi to posui
(ch. viii. § 39. 4); summitto shows that assimilation to which the
internal consonants of a word were liable. On the other hand
a Noun with a Preposition is as a rule not so treated (although
there are not wanting examples like sedulo for se dolo, § 7), unless
a Compound Adjective is formed of the Preposition and the
Noun, e.g. pérégre, Loc. of *peregris compounded of per- and ager
(ch. vi. § 38). Still it must be remembered that in the Roman
pronunciation the Preposition and the Noun formed a word-group
(e.g. cereum-littora, ch. iii. § 124. 6), and in the Roman ortho-
§ 11.) PREPOSITIONS. 573

graphy they were usually written together (e. g. ingalliam, initaliam,
Mar. Victorin. 23. 12 K.), sometimes with consonant-assimilation,
(e.g. summdnus for sub manus, Plaut. Pers. 450). This close
union of the Preposition with its Verb and Noun must have led
at a very early time to the syncope of a final short syllable of
Prepositions ; and it is possible that byforms like Gk. évé and év
may be doublets of very ancient date, representing the forms
assumed by the word when used independently and in compo-
sition (cf. Engl. ‘by’ and‘ be ’-witch, Germ. ‘bei’ and ‘ be *-leben,
similarly ‘ off’ and ‘of,’ ‘too’ and ‘to’ are doublets, one of
which is used as Adv., the other as Prep.). Tmesis, or the
separation of the prepositional part of a Compound, from the
other part, is a feature of the older stage of every language ;
and is common in O. Lat. (e.g. sub wos placo was the archaic
phrase for suppéico retained in Latin prayers; transque dato and
endoque plorato are legal archaisms for ¢rdditoque and imploratoque,
Fest. 444. 30 Th.). An arrangement like sub wos placo, ob uos
sacro (for obsecro vos) (cf. Vedic vi né dhéhi, ‘lend us’) became
the rule in the Celtic languages, thus in O. Iv. at-om-aig
‘adigit me’ the Pronoun is ‘infixed’ between the Preposition
and the Verb, as if we had in Latin ‘ad me agit.’

In the later stages of a language the use of Prepositions
increases more and more. In Latin this culminated in the loss
of Case-suffixes, and the use of Prepositions in their place, as we
see in the Romance languages. As early as the first cent. a.p.
a grammarian points out that iz manus aqua is the phrase in
vogue instead of the older agua manibus (Caper 92. 8 K.). New
distinctions of prepositional meaning were expressed by com-
pounding Prepositions with one another, e.g. de-ew, de-sub, &e.
[ef. abante, C. I. L. xi. 147, Fr. avant], a process which may have
begun at a very early stage ; for I.-Eur. Prepositions often show
an appended particle (Pronoun and Adverb), such as (1) -s(é),
Gk, -ce, e.g. dap, eg, Lat. abs, ex, sus-; (2) -d(é), Gk. ddpov-de,
e.g, Lat. postid, antid, prod- (cf. O. 81. -dti of pré-dit-, ‘before ’);
(3) -ti, e.g. O. Ind. prati, Zend patiy, Gk. sport, mori, Ose. pert- ;
(4) -n(é), e.g. Lat. pone for *pos-ne (cf. Germ. von, O. H. G.
fona and fon). These particles, whose original form is not
always recognizable (thus a Latin -d from -dé might come from
574 THE LATIN LANGUAGE. (Chap. IX.

an original *dé, *di, *dd, *dt, &c., ch. i. § 37), cannot be
separated from the pronominal stems mentioned in § 10, ch. x.
§ 1 (e.g. the ‘Adverbial’ -d« of Gk. ddyovde from the * pro-
nominal’ -3¢ of Gk. 8-3); and it is doubtful how far there
was originally any real distinction between them. In Latin
their original form is especially obscured by the Latin tendency
to syncopate a short second syllable (ch. ii. § 13), a process
which may have led to the confusion of the Preposition endo.,
indti- with the different Preposition en-, im-, in such words as
indit-grédi, in-gredi, indi-pérator, im-perator, and ultimately to
the disuse of endo, indu, in favour of en, in. (In Terence ixaudio
alone is used for earlier 2xd-audio and in-audio, A similar con-
fusion of I.-Eur. *endo and *en may have taken place in Celtic).

And the tendency of a Latin Preposition, because unaccented,
to be obscured brought about that confusion of od- and ab-, de-
and di- (dis-) in Compounds which we see in Late Latin, and
which even in the earlier centuries of the Empire attracted the
notice of the grammarians (Vel. Long. 64. 19 K., &c., on de- and
di- ; in Romance *abdurare, *abaudire, *abtenere have supplanted
obdirare, obaudire, obtinére). A much earlier opportunity of
confusion was afforded by Prepositions which represented different
developments, case-forms, &c. of the same root, e.g. Lat. per,
through, and Umbro-Ose. per, ‘on behalf of, before,’ the equivalent
of Lat. pro (both I.-Eur. *per and *pro being derived from the
same root per-, on which see § 38); and this confusion is very
hard to trace. The readiness too with which a Preposition
changes its meaning is an obstacle in the way of identifying its
cognates in other languages. Oscan dp, op governs the Abl. with
the sense of Lat. apud, while Latin of (governing the Acc.) has
passed from that sense (§ 35) to its classical sense of ‘ on account
of’; O. Ind. 4 with Abl. following has the sense of ‘to,’ but
with Abl. preceding might be translated ‘from’; examples
which show that a difference of meaning between a Preposition
in one language and in another is not a valid proof that the two
words were not originally identical. Much less is the difference
of case governed to be taken into account. In the earlier stage
of every language the Prepositions must have been used with
great elasticity, sometimes with one case, sometimes with another
§ 12,] PREPOSITIONS. 575

(cf. O. Lat. in potestatem esse, &c.), the fixing down of Prepo-
sitions to a particular case being always a feature of an advanced
stage of language. [Servius may thus be right in saying (ad
Ecl. 1. 29 longo post tempore) that post, ante, circum were used
also with the Abl. in earlier times: antiqui enim ‘ post’ ‘ante’
‘circum’ etiam ablativo jungebant, quod hodie facere minime
possumus ; Pompeius (278. 21 K.) attributes ante templo and
propter homine to Pacuvius]. It should be noticed that in
Umbro-Oscan local Prepositions, indicating rest in a place, &c.,
go with the Locative case, not the Abl. as in Latin. Their
position too varied in course of time. In classical Latin a Prepo-
sition, especially a monosyllabic Preposition, precedes the noun
(hence ‘ Pre-position ’), except in particular circumstances (e.g.
metu in magno, &c.; see Neue, ii. pp. 942 sqq. for statistics),
but in the older literature often follows it; and in Umbro-
Oscan postposition is common, e.g. Umbr. asam-ad, ‘ad aram,.
termnom-e ‘in terminum.’ (So our ‘in here’ was earlier ‘ here
in’), In J.-Eur. the Preposition seems to have preceded
the Verb, but to have followed the Noun, while between
the Prep. and the Verb a Particle or Enclitic Pronoun (ch. ii. § 12)
might be inserted (cf. O. Lat. anti-d-eo, § 18 ?, sub vos placo,
Pp. 569).

On the Vulgar Latin treatment of Prepositions in composition
with Verbs, as it is reflected in the Romance languages, see
Meyer-Liibke Rom. Gram. ii. pp. 617 sqq. To the ordinary
Prepositions were added foris, e.g. Vulg. Lat. foris-facere (Fr.
forfaire, Ital. fuorfare), and other words.

§ 12. Ab, ap-, po-, abs, a-, au-, af, absque. Ad, from, is
I.-Eur. *ap (Goth. af, Engl. of, off), a curtailed form of *pd
(O. Ind. dpa, Gk. azo, e.g. O. Ind, apa-i-, ‘to go away,’ Gk. dn-eus,
Lat. ab-eo; cf. Lat. dpid for *apo-d, see below), of which another
curtailment was *po (O. SI. po-, Lith. pa-), found in Lat. pd-sttus,
pono for *po-s(i)no (with Pft. pd-stve changed to pds-ui owing to
a false apprehension of po-situs as if it were posi-tus like méni-tus).
(Po-lubrum, a wash-basin, pd-lire, and Germ. vo-n, O. H. G. fo-na
and fo-n, have also been referred to this I.-Eur. form, § 39). The
form dp- appears in dp-erio, and was no doubt the shape assumed by
the word in such collections as ab templo; ab is due to the same
576 THE LATIN LANGUAGE. [Chap. IX.

Latin preference for -d rather than -p, as substituted 0b for op
(Osc. op ; cf. Lat. op-erio), sub for *sup (ch. ii. § 73). The form
abs (pronounced and often written aps, see ch. ii, § 80), in which the
Preposition is augmented by the particle *-s(é) (Gk. dy), is used
in Composition before Tenues,e. g. before ¢,¢ in abs-traho,abs-condo,
while before p it is, by a law of Latin phonetics (ch. iv. § 157),
reduced to as-, e.g. as-porto for *aps-porto, as-pello for *aps-pello ;
it appears also in the O. Lat. phrase absque me (te, &c.) esset
( foret), equivalent to ‘ si sine me esset,’ where que, like its O. Ind,
equivalent ca in the Rig-Veda, seems almost to have the sense of
‘if? (cf. O. Engl. an for and); at a later period absque me, &c.
was used without the verb, and adsque came to take the sense of
sine, without (4.L.L. vi. 197). That d@ (Osc, aa-manaffed
‘amandavit, Umbr. aha-, aa-, a-, e.g. aha-vendu beside pre-
vendu) is another form of ad, as é of ea (see below), is generally
believed, though it is difficult to see why of and sud did not
develope corresponding forms*6,*si; it may be an entirely different
word, associated with ad because of its resemblance in meaning,
form, and usage!. dw- of aufligio, aufero, &c. is an example of an
association of this kind. It has not been produced from ad by any
phonetic process, but represents a different I.-Eur. preposition,
*aw/(é) (O. Ind. diva, Pruss. au-, e.g. O. Ind. ava-bhr- ‘ au-fero’),
which was brought into requisition in these Compounds before
an initial f to avoid confusion with the compounds of ad, e.g.
affero. (On the confusion between dfluo and affluo, see Nettleship,
Contrib. Lat, Lex.s.v.). A curious Preposition af, used in Cicero’s
time occasionally in account-books, with the name of the person
from whom money had been received, occurs on a few inscrip-
tions, and in O, Lat. afvolant for dvdlant. Whether it is
a dialectal form (cf. Pelign. af-ded ‘abiit’?) with / represent-
ing some IJ.-Eur. aspirate (cf. O. Ind. 4dhi, ‘ on, used with Abl. in
the sense of ‘from’), or a Latin variety of ad (or aw?) with
J produced originally under the influence of some following con-

. sonant (most probably v), it is impossible to say. It may be
a mere (Greek ?) trick of writing, with the symbol F employed
to denote the w- or w- sound, like the Greek digamma (cf. Prise.
1.35. 17 H.).

' Lat. dand W., Teut, 6 are referred to I.-Eur. *4 by Buck, Osk. Spr. p. 25.

 
§§ 13-16.] PREPOSITIONS. 577

§ 13. Ab, abs, a. In Plautus ab is used before vowels and J, s, *; & before
b, p, m, f, v, ¢, g, g (Labial and Guttural sounds); abs (and a) before tu, tuus,
&c.; ab and a before ¢, d, 1, x; in class. Lat. ab is used before vowels and 1, n,
7, 8,J3 & before b, p, f, v ; abs before c, g, t (Cicero began with abs te, but discarded
this expression for a te) ; in Late Latin ab is used before vowels, & before con-
sonants (see Langen, Beitr. 331 ; Georges, Lew. Wortf.s. v.; A. L. L. iii. 148). The
usage of @ in the older period allows of its being a mere phonetic development
of ab, for @ bellomay be simply an expression of the sound abbello (ch. ii. § 1g0),
and so the shortening of a by the Law of Breves Breviantes in Plautus, e. g.
quid & bélio portat?, will not be a case of the shortening of a naturally long
vowel by this law (see ch. iii. § 34).

§ 14, Af. Cicero’s words are (Orat. xlvii. 158) : una praepositio est ‘af,’ eaque
nune tantum in accepti tabulis manet, ne his quidem omnium, in reliquo
sermone mutata est ; nam ‘a-movit’ dicimus et ‘ab-egit’ et ‘ abs-tulit,’ ut
jam nescias ‘a’ ne verum sit an ‘ab,’ ‘abs.’ Quid si etiam ‘ au-fugit,’ quod
‘ab-fugit ’ turpe visum est et ‘ a-fer’ noluerunt, ‘aufugit’ et ‘aufer’ malue-
runt. Quae praepositio praeter haec duo verba nullo alio in verbo reperietur.
Velius Longus (60. 13 K.), who refers to this passage of Cicero, gives as an
illustration of the now obsolete use of af in receipts, af Longo (his own name);
Paul. Festus (19. 31 Th.) mentions afvolant as an actual form used by an
ancient writer. On an inscription of Amiternum (Not. Scav. Oct. 1891) we
have af vinieis, af villa (beside ab castello, ab segete); on the Epistula ad Tiburtes
(C. I. L. i. 201, of ¢. 100 B. c.) af uobeis ; on the milestone of Popillius (i. 551,
of 132 B.c., from Lucania) af Capua (besides ab Regio) ; on a bilingual (Greek and
Latin) inscription ascribed to v. 81 B. c. G. 587) af Lyco; on an inscription
of Praeneste (i. 1143) af muro, and so on.

§ 15. Ad, at, to, I.-Eur. *ad (O. Ir. ad, e.g. at-om-aig ‘ adigit
me,’ lit. ‘ad me agit,’ Goth. at, Engl. at; cf. Goth. at-tiuha with
Lat. ad-dico, Goth. at-baira with Lat. ad-féro) is a different word
from the Conjunction af, I.-Eur. *at (Goth. ap- in ap-pan, ‘ but’),
though often confused with it in Roman spelling (ch. ii. § 76).
On the old form av, e.g. arfuerunt, arvorsum, due to the phonetic
change of d to an r-sound before f, v, see ch. iv.§ 112. This
Preposition, which governs the Acc. in Umbro-Oscan as in Latin,
is found augmented with the particle *s(e) in Oscan, e.g. az
harttim‘ad hortum’ ; but also ad, e. g. adpad ‘adquo’ ‘quoad,’ idad
‘ad id’ In Umbrian we have ad, e.g. ad-fertur ‘ adfertor,
adputrati ‘arbitratu’ (ch. iv. § 112), postfixed to Nouns, e.g.
asam-ad ‘ad aram,’ written ar- in arnipo ‘ quoad’ (§ 10. 3).

§ 16. Ambi-, around, on each side, I.-Eur,*ambhi (Gk. dul; ef.
O. Ind. abhi, Gaul. ambi-, O. Ir. imme, imb-, W.am-, O. Eng]. ymb,
O. SI. obi-), a Locative of the same stem as I.-Eur, *ambho, ‘both’

Pp
578 THE LATIN LANGUAGE. [Chap. IX.

(Gk. dupe, Lat. ambo; ef. O. Ind. ubhé-, Goth. bai, baj-dps, Lith.
abi, O. 8]. oba), appears in Latin compounds in the forms, (1)
amb- before a vowel, e.g. amb-arvale (sacrificiam) ‘ quod arva
ambiat victima’ (Serv. ad Eel. iti. 77), amb-urbiales (hostiae) ‘ quae
circum terminos urbis Romae ducebantur’ (Paul. Fest. 4. 15 Th.),
amb-ustus ‘circumustus’ (2. 4. 17), whence by false analogy
comb-ustus, instead of *com-ustus (unless this rather represents
co-amb-ustus), amb-itus ‘circuitus’ (4. 4. 18); (2) am- before
a consonant, e.g. am-ter-mini (oratores), a phrase of Cato’s ‘ qui
circa terminos provinciae manent’ (Paul. Fest. 13. 9 Th.; Macr.
i. 14. 5; Gl. Philox.), am-plector, am-pendices ‘ quod circum-
pendebant ’ (Paul. Fest. 16. 3 Th.), am-segetes ‘quorum ager viam
tangit’ (2. 16. 1; Charis. 231. 11 K. seems to quote a similar
amfines), ém-icio for am-jicio. The form ambi- in compounds bears
the sense of ‘both’ in ambi-dens (ovis) ‘quae superioribus et
inferioribus est dentibus’ (Paul. Fest. 4. 9 Th.), amdi-lustrum
* quod non licebat nisi ambos censores post quinquennium lustrare
civitatem ’ (‘ Serv.’ ad Aen. 1. 283), ambi-viwm. In Umbro-Oscan
the word appears with an 7-suffix (cf. inter), Osc amfr-et
‘ambiunt, Umbr. ambr-etuto ‘ambiunto,’ but also e.g. Ose.
am-nid ‘cireuitu,’ am-vianud ‘vico, Umbr. an-ferener ‘cir-
cumferendi.’

§17. An-, a curtailment of I-Eur. *ini, ‘on’ (Zend ana, Gk.
dvd, Goth. ana, Engl. on, O.S]. vii for *on) (cf. I-Eur. *init),
(O. Ind. anu, Zend anu) may appear in dén-hélus (also derived from
the root an-, ‘to breathe,’ whence dinimus, &c.), an-quiro [by some
explained as *amb(2)-quiro|, an-tennae, an-testari (or for *ante-
tennae, *ante-testari, ch. iii. § 13, p. 176). Its presence is more
certain in Umbro-Oscan, e.g. Ose. ava-Faxer ‘consecravit,’ an-
getuzet ‘proposuerunt, Umbr. an-tentu ‘intendito,’ am-pentu
‘impendito,’ unless indeed it is here some variety of Lat. in, as
an- the Umbro-Oscan negative prefix (ch. iv. § 81) is of Lat. in-.

$18. Anté, before, I.-Eur. *anti (O. Ind. Anti, ‘ opposite, near,’
Gk. dvri, opposite, instead of, Goth. and, ‘towards,’ Engl. an-swer,
Lith. afit, ‘ on”), a Locative Sing. of some stem connected with
Lat. antes, rows, O. Ind. anta-, ‘ vicinity, end,’ Goth. and-eis, ‘end,’
of which Gk. avra, opposite (cf. dvrnv), is another case. In Oscan
§§ 17-20.] PREPOSITIONS. 579

the Preposition (governing the Acc. as in Latin) appears without
the final short vowel (this loss of a final { is common in Oscan),
e.g. ant pinttram ‘ante pontem;’ but in Latin, though poste was
reduced to post (see below), ant is not written for ante [in Plaut.
Rud. 509, if the reading of the MSS. is right, we must pronounce
ant(e)positast, a quadrisyllable :
quam quaé Thyestae quéndam antepdsitast Téreo.

On antenna and antestor, see above]. With the particle *dé
appended, as in postid, is the form antid- in O. Lat. antid-eo (e. g.
Plaut. Zrin. 545 sed Campans genus Multo Surorum iam antidit
patientia), antid-lac (used by Plautus when three syllables are
required by the metre, antehac being a dissyllable ; cf. antidit, &c.
and anteit, &c.), antid -ed (Liv. xxii, 10. 6 in the Vow of the Ver
Sacrum ; aztea is not found in Plautus, and only once in Terence,
viz. Andr. 52). In antid- the -i of I.-Kur, *anti, not being final,
does not sink to é (cf. anti-stes, &e., ch. iii. § 39).

§ 19. Apud, which is also spelt aput, seems to be the I.-Eur.
Preposition *ipd (of which Lat. ad is a curtailment; see above),
augmented by the particle *d(e), or *t(i), and must have been
originally *apo-d, or *apo-t (cf. Dor. wort). An old form apor,
with that change of -d to an r-sound (before f, v) seen in arfue-runt,
ar-vorsum, &c. (ch, iv. § 112), is quoted by Paul. Fest. 19. 34 Th.
(cf. apur finem on a Marsicinser., Zvet. 1. I. I. 45; apur is quoted
by Mar. Vict. 9.17 K.). On the spelling aput, like at, set for ad,
sed, see ch. ii. § 76. In Oscan dp, op (Lat. 0d) is used with the
Abl. in the sense of Lat. apud, e.g. op tovtad ‘apud populum,’
Gp eistid sakaraklad ‘apud id sacellum.’

§ 20. Cireum, circa, circiter. Cireum, around, is the Ad-
verbial Acc. Sg. of circus (Gk. xpixos, a ring; cf. O. Engl. hring,
with nasalization), which had in O. Lat. the sense of class. cireulus
(Dub. Nom. 573. 4 K ; cf. above, ch, v. § 24), and is used, for
example, by Accius of the moon’s orbit (Zrag. 100 R.) :

quot luna circos dnnuo in cursu institit.
In the early literature cirewm is the only form, whether Adverb
or Preposition, but in class. Lat. a byform cired appears, first
found in Cicero (who uses it in three passages of the Verrine
orations, but afterwards seems to have discarded it), possibly never
Pp2
580 THE LATIN LANGUAGE. [Chap. IX.

in Caesar, but much affected by Livy. Cired is a formation on
the type of supra, extra, &c., perhaps originally employed with
verbs like esse (Cicero’s three examples of the words are: Verr, II,
i. 51.133 canes esse circa se multos ; 1. 48. 126 canibus, quos circa
se haberet; iv. 48. 107 Henna, quam circa lacus sunt plurimi),
owing to a feeling that circum was suitable only for verbs of
motion, e.g. legatos circum civitates mittere, ‘to send ambassadors
a tour of the states,’ ire circum urbem, ‘to goa circuit of the city’
(A. L. L. v. 295). Circiter, an adverbial formation like drevdter,
O. Lat. amiciter (see § 1), came to be restricted to the logical
sense of ‘ about,’ ‘almost,’ e.g. Plaut. Cist. 677 loca haec circiter.
The form circo appears in the Adverb id-circo, as circa in quo-
circa, with the same logical sense (cf. Osc. amnud, ‘because of,
in egm[as tovti|cas amnud ‘rei publicae causa,’ an adverbial Abl.
Sg. Neut. of amno-, a formation with the suffix -no- from the
Preposition am- [ Lat. am-, ambi-| , as comno- ‘ comitium ’ from the
Prep. com-).

Cis, citra, on this side (cf. Umb. gimu, smo, ‘ retro’?), are formed
from the J.-Eur. pronominal root ki-,‘ this’ (Gk. -Ke of odk(, wodAdt,
Goth. hi-na, ‘ this,’ Engl. he, Lith. szis, O. SI]. si), exactly as their
opposites ws, ultra, on that side, from the I.-Eur. pronominal root
ol-, ‘that’ (ch. vii.§ 13), the first by the addition of the particle *s(e)
(p. 573 ; on wés for *oll, see § 56), the second (an Abl. Sg. Fem.) by
the suffix -tero- (ch. v.§ 16). The Adverb citrd (Abl. Sg. Neut. or
Masc.) corresponds to citraé as ultré (e. g. ultro citroque) to ultra,

§ 21. Clam, clancilum. Clam, an Adverbial Acc. Sg. Fem. (?)
from the root fel-, ‘to hide’ (Lat. céd, occitlo, &c.), had in
O. Lat. a byform clam-de, clande (written clade in the MSS. of
Placidus 15.32 G.; but cf. guamde from quam, ch. x. § 11), whence
was formed the Adj. clandestinus. Another O. Lat. form written
callim in the MSS. of Paul. Fest. 33. 6 (‘callim’ antiqui dicebant
pro clam, ut ‘nis’ pro nobis, ‘sam’ pro suam, ‘im’ pro eum) is
more difficult to explain. (Should we read calam, and refer the
form to the Analogy of palam? It may be merely the coinage
of some grammarian to support his etymology of clam). Clam,
which governs the Acc. always in Plaut. and Ter., and perhaps
never the Abl. at any period of Latin (Langen, Beitr. p. 230),
has in the Comedians another, apparently a Diminutive form,
§§ 21-25.) PREPOSITIONS. 581

clanctilum (but cf. procul, § 2), used as a Preposition by Terence,
Adelph. 52 clanculum patres. Cf. the glosses : clanculae ‘absconsae’
(C. G. L. v. 277. 58); clanculum ‘occultum’ (ib. 278. 1).

§ 22. Cém-, (ctim), with, and co- (e.g. céyo for cd-dgo),
L.-Eur. *4ém and *2% (?) (with palatal or with guttural 4?) (O. Ir.
com-, co, W. cyf-, ey, Ose. com, con, co-, Umbr. com, -co, co-) is in
early inscriptions written guom (Bersu, Gudturale, p. 42), like the
Relative Adverb guom, when, because qguo- had the same sound as
co- (ch. iv. § 137). Theo of com became u in the unaccented use of
the word (ch. iv. § 20), and before certain initial consonants (ch. ii.
§ 22), and cum became the recognized spelling of the simple Preposi-
tion, though in compounds, e. g. com-es, the o-form was retained.
On the form co-, e. g. co-eo (Quint. i. 6. 17), O. Lat. co-ventionid,
and the like, see ch. ii. §$ 61, 65. Its original difference from
com- is not certain. Ose. com, with, governs the Abl., and is
prefixed in compreivatud, conprewatud ‘cum privato’ on the
Tabula Bantina (ef. Umbr. com prinvatir ‘cum legatis’), but
Umbr. kum, com is postfixed in the sense of ‘apud,’ ‘ juxta, e.g.
asa-ku, ‘juxta aram,’ veris-co (opposed to pre verir and post
verir), at the gates.

§ 23. Contra (see §§ 1, 4). Ose. contrud in the phrase on the
Bantine Law: svae pis contrud eueie fefacust ‘si quis contra
hoe fecerit,’ is followed by the Adverb (Locative) emeic, as Lat.
arvorsum in the S. C. Bacch. by the Adverb (Abl. Fem.) ead:
sei ques esent, quei aruorsum ead fecisent. It is Abl. of an
O-stem (cf. Lat. contrd-versia), as contra of an A-stem.

§ 24. Coram, in presence of (not a Preposition till Cicero’s
time), seems to be connected with ds, Gen. dris, the face, perhaps
being an Adverbial Acc. Sg. Fem. of a stem *cdso- (*céro-), com-
pounded of the preposition com- (eum) and this noun (cf. O. Ind.
saksid). Incoram with a Gen., e. g. incoram omnium, is found

in Apuleius.

§ 25. Dé, down from, concerning (Fal. de in the phrase: de
zenatuo sententiad, Zvet. [.J.[. 70); O.Ir. di, O.W. di; ef.
O. Ir. di-mér, ‘very great,’ with Lat. de-magis, &c. corresponds
to Ose. dat (e.g. dat senateis tanginud ‘de senatus sententia’),
582 THE LATIN LANGUAGE. [Chap. IX.

which seems to have the particle -t(i) affixed, as per-t, Lat. per

_(§ 38), unless the -t stand for -d, in which case *dad may be an
Ablative (Lat. dé for *déd?). The Umbrian Preposition is da
(with final -d or -t dropped in Umbrian fashion), if da-etom
on the Eugubine Tables (vi. A 28) stands for Lat. demptum
(cf. Ose. da-did ‘ dedat,’ da-dikatted ‘ dedicavit ’).

§ 26. Dis-, apart, is most naturally referred to some byform
of the root dwo-, dwi-, ‘two’ (Goth. tvis-, e. g. tvis-standan, ‘ to
separate’), wanting the w (see ch. iv.§ 71). With the w the same
formation expressed the Numeral Adverb *dwis [O. Ind. dvis,
Gk. 8(F)is, M. H.G. zwis], and is in Latin dis (ch. iv. § 68). Before
a vowel dis- becomes, by the phonetic law of Latin, dir- (ch. iv.
§ 148), e. g. dir-tmo, and before voiced consonants (see ch. iv. § 151)
di-, e.g. di-miveo (dis-mota on the 8. C. Bacch., C. I. L. 1. 196).

§ 27. Endd. (Cf. O. Ir. ind-., e.g. ind-riuth, ‘I attack,’
Gaul. ande-?), also under the form indu, the 7 and uw being ap-
parently weakening of e and o due to the unaccented use of
the Preposition. It corresponds in meaning to 2# (both with
Abl, and Acc.), and was in classical Latin replaced by i, e.g.
class. im-pérdtor, O. Lat. imdu-perator, class. in-grédi, O. Lat.
indu-gredi. It seems to represent an I.-Eur. *en-dé (Gk. evd0-61,
évdo-Gev, évdov), compounded of the Prep. *en (Lat. 7) and the
Prep. *do (cf. Lat. dd-nec; O. Ir. do, Engl. to, Lith. do, O. Sl.
do), the last element being connected with the Adverbial particle
*d(e) (Gk. dduov-de). The final -3 has been preserved from
becoming -é in Latin (ch. iii. § 37) by the frequent use of the
word as the first element in a compound. (So *prd remains pro
in Latin and does not in unaccented use become *pré, owing to
compounds like pré-ficiscor, &c.). Traces of the same confusion of
I.-Eur. *endo- and I.-Eur, *en- are seen in Celtic.

Endo and indu. The form endo occurs, e.g. in the epitaph of Ennius,
quoted by Cicero, in his De Republica (ap. Sen. Epp. 108. 34):

si fas endo plagas caelestum ascendere cuiquam est,
mi soli caeli maxima porta patet,
in a clause of the XII Tables (ap. Fest. 452. 6 Th.): si caluitur, pedemue
struit, manum endo iacito, ‘if he deceives, or attempts to run away, the
prosecutor may arrest him,’ a clause alluded to by Lucilius, xvii. 10 M. ;
si non it, capito, inquit, eum, et si caluitur, endo
ferts manum,
§§ 26-29.] PREPOSITIONS. 583

and in other laws, and is one of the archaisms used by Cicero in drawing up
his code of laws (Legg. ii. 8. 19) ; it is employed too by Lucretiug (vi. 890) endé
mari [cf. the glosses: endoclusa éyxexreopévn ; endo festabat; endo rivum kara
feiOpov ; endodicarit pnvice (C. G. L. ii. 61. 35); endogenia (-ua?) ‘naturaliter
amoena’ ; endoriguum ‘irriguum’ (C. G. L.v. 193. 25)]. The form indo- appears
in a line of Ennius, referring to Romulus and Remus (Ann. 59 M.):

indotuetur ibi lupus femina, conspicit omnis ;

the form indu, e.g. in Ennius, Ann. 298 M. indu foro lato sanctoque senatu
(ef. Lucil. inc. 17 indi foro); in Luer. v. 102 nec jacere indu manus (ef. ii.
r096 indu manu), as well as in the compounds induperator Enn., Lucer., indu-
pedio Lucr., indugredior Lucr., &. By the time of Plautus the word seems to
have dropped out of ordinary usage, for it occurs in his plays only in com-
pounds like ind-audio (Terence knows only in-audio), ind-ipiscor (cf. class. ind-
igeo, indi-géna, ind-éles, &c.) ; and though it occurs at the end of Varro’s Res
Rusticae (iii. 17. 10): ille inde endo suam domum, nos nostram, the phrase is
a quotation from Ennius’ curious experiment in language, mentioned by
Ausonius (Techn. 18) and others: endo suam do, with do, an apocopated form
of dimum, after the type of Homer’s 60. :

§ 28. Erga, ergo. Lrgd, originally local (e.g. Plaut. Zrue.
405 tonstricém Suram Nouistin nostram quae érga aedem
seséd habet ?, if the MSS. reading be right), must be connected
with ergé, on account of, in O. Lat. a preposition or rather
postposition, governing the Genitive, e. g. funeris ergo, XII Tab.
Whether the two words have been differentiated on the type of
ultra and ultré, intra and intro, or whether they came originally
from two different stems, it is impossible to say. ZHrgo has been
explained as a compound of the Preposition ¢ and the Abl. Sg.
of a stem *rego- (from the root reg-, ‘to stretch’), meaning
‘ direction,’ so that its change of meaning would resemble that
of German wegen (originally von wegen). rgd might similarly
represent ¢ *régd, like e régidne, opposite. It is not restricted to
the expression of friendly feeling in Plautus, e.g. Pseud. 1020
ne mélus item erga mé sit, ut erga illam fuit; Cas. 618 aut

quéd ego umquam erga Uénerem inique fécerim.

§ 29. Ex, ec-, 6, out of. I-Eur. *eks (Gk. é, Gaul. ex-,
O. Ir. ess-, W. es-; cf. Lith. isz?) appears to be a compound of
a Preposition *ek and the particle *s(e), as Gk. ay, Lat. ads,
append the same particle to *ap, a curtailment of *apo (§ 12).
In Latin compounds the Preposition often appears before the
letter f in the form ec- in MSS. (cf. Ter. Scaur, 26. 14 K.
effatus, non ‘ exfatus’ nec ‘ecfatus, ut quidam putaverunt ; Ter.
584 THE LATIN LANGUAGE. [Chap. IX.

Mawr. |. 949 K. muto vel partem prioris, si fit hirtum, syllabae,
‘ecfer’ ut dicam, vel illud, ‘hoc tibi effectum dabo’), eg.
ecfodio, ecféro, ecfari (for examples, see Neue, ii*. p. 870), often
corrupted to haec and to et (see Class. Rev. v. 295; Fleck. Jahrb,
1890, p. 771). (Et is often a corruption also of ex, owing to the
fact that the symbols for these words in minuscule writing were
very similar.) This ec- may be merely a phonetic development
of ew before f, as é of ex before d (e. g. é-duco ; cf. sé-décim), m (e. g.
é-mitto ; cf. sé-mestris), &e. (ch. iv. § 151). Corresponding to
Lat. é we find in Ose. ee-stint (apparently with a different sound
from I.-Eur. é, which is in Ose. i, i, e. g. ligud ‘lege ”), eehiianasim
‘ e(ve)hiandarum,’ Umbr. ehiato- ‘ evehiato-,’ easa ‘ ex ara,’ &e.
(see ch. ii. § 6).

Extra. (O. Lat. ewtrad ; cf. the S.C. Bacch., C. LL. 1. 196
exstrad urbem) is an Abl. Sg. Fem. of an extension of ex by
the suffix -t(e)ro- (ch. v. § 16), like in-tra, ct-tra, ul-tra, &c. Oscan
ehtrad (with 2¢ for c¢,as in Uhtavis, the Oscan form of Octavius),
O. Ir. echtar may represent an original stem *ek(s)-tero- or
*ek-tero-.

§ 80. In, in, the unaccented form of O. Lat. en [cf. éngue,
but indltod on the (restored) Col. Rostrata (C.J. L. i. 195)] is
L-Eur. *én (Gk. év, O. Ir. in, W. yn, Goth. in, Lith. 7). The
same form is used in Latin and other languages with the two
senses, (1) in, (2) into [whereas in Greek the second is distinguished
by the addition of the particle *s(e), évs, Att. e/s], and appears
to be a Locative case, formed without the case-suffix -i (ch. vi.§ 37).
(The Greek byform évi shows this case-suffix ; but cf. above, § 11).
Before labial consonants iz became im by the Latin phonetic law
(ch. iv. § 78), e.g. im-pleo, im-mitto, imbello (in war), C. I. L. iii.
4835, &c. On the derivative Prepositions endo, inter, see §§ 27, 32.

Ose. en (in), Umbr. en, have with Acc. and Loe. (not Abl.) the
two senses of Lat. iz, but are postfixed, e. g. Ose. exaisc-en higis
‘hisce in legibus,’ Pel. eite uus pritrom-e, ‘do ye go past or
forward,’ Umbr. arvam -en ‘in arvum,’ arven ‘in arvo,’ fesner-e
‘in fanis.’ Osc. -en with the Abl. imad-en ‘ab ima (parte),’
ewsuc-en ziculud ‘ab eo die (*dieculo),’ which has the sense of
Lat. ad, has been referred to Lat. inde (but see § 10. 6).
§§ 80—-35.] PREPOSITIONS. 585

§ 81. Infra (infera, C.I.L. i, 1166), an Abl. Sg. Fem. like
supra, citra, intra, connected with the Adj. inférus (on which see

ch. v. § 16).

§ 32. Intér, between (O. Ind. antdr, O. Ir. étar; e. g.O. Ind.
antar-chid- ‘inter-scindo’), is formed from 2 by the addition of the
suffix -tero- (see ch. v.§16), like intérior (cf. intrd,intra), as ex-tero-,
&e. from ew. The Oscan form is anter (with Acc., but once with
Abl.-Loe. Plur.), the Umbrian form is anter, ander (governing
the Ace.), both with an- corresponding to Lat. iv-, the Preposition,
as to Lat. ez, the Negative, e.g. Umbr. an-takro- ‘ in-tegro-.’

§ 33. Intra, intis. Jntrd is an Abl. Sg. Fem. like extra,
class. contra (while iztro is an Abl. Sg. Neut. like Ose. contrad,
Lat. contro-versia) (cf. Osc. Entra-, the name of a goddess).
Intus (Gk. év-rés with the I.-Eur, affix -ts, implying usually
motion from, § 1) wavers between an Adverb and a Preposition
in such a phrase as Virgil’s ¢alé intus templo, ‘in such temple,
within ’ or ‘ within such temple’ (cf. Lucr. vi. 798).

§ 34. Juxta, which is first used as a Preposition by Caesar, is
Abl. Sg. Fem. of astem *juxto-, whether this be P. P. P. of a verb
*juao formed from jungo as viso from video, quaeso from quaero
(ch. viii. § 33. 4), or a Superlative with the I.-Eur. Superlative
suffix -isto- (Gk. mAe-ioros, &.). The Adv. juatim is found as
early as Livius Andronicus (Zrag. 11 R.).

§ 85. Ob, I.-Eur. *op(i), apparently a variety of *epi (Gk. éni,
on, to, ém-cbev, behind, O. Ind. dpi, ‘by,’ Lith. api-, ‘around’ ;
ef. Lith. ap-szvésti, ‘to make light,’ with Lat. od-caeco, to make
dark), is in Oscan op (with the sense of Lat. dpud, governing the
ADL. e.g. op tovtad ‘apud populum, tp efsid sakaraklad ‘apud
id sacellum’), and often retains its -p in Latin spelling in
compounds like op-tineo (e.g. optenui on a Scipio Epitaph,
C.I.L. i. 38; cf. Quint. i. 7. 7), op-ério, though in the simple
word the Latin usage substituted the Media for the Tenuis as
the final consonant (cf. ad for ap, sub for sup, and see ch. il. § 76).
In classical Latin it has the sense of ‘before,’ e.g. 0b oculos
ponere, to describe, or ‘on account of’; but in the earlier litera-
ture it had other shades of meaning; cf. Paul. Fest. 193. 7 Th.
586 THE LATIN LANGUAGE. (Chap. IX.

ob praepositio alias ponitur pro circum, ut cum dicimus urbem
‘ obsideri,’. . . ‘obvallari,’ .. . alias pro ad, ut Ennius :

ob Romam noctu legiones ducere coepit ;

Servius tells us that many interpreted 0b Italiam in Virgil, den.
i, 233 as ‘juxta Italiam,’ with the old sense of od: ob enim
veteres pro juxta ponebant. (This variety of meanings has been
explained by the theory that Lat. ob represents, not only I.-Eur.
*epi, *opi, but also an *ebhi, seen in O. Ind. abhi.) By the addi-
tion of the particle *s(e), as ad became ads (e.g. abs te, abstineo),
so 0b became ods, a form occasionally found in compounds before
t-, e.g. obstinet (Fest. 228. 6 Th. o. dicebant antiqui, quod nunc est
ostendit), obstrudant (Paul. Fest. 221. 3 Th. ‘ avide trudant’; Fest.
220.14) (so Umbr. os-tendu ‘ostendito’). (On dmitto, see ch. ili. § 34.)

§ 86. Palam, like its opposite, clam, an Ace. Sg. Fem. (but
see § 1) of some stem, perhaps connected with the Plautine
verb dispalesco (Bacch. 1046) :

periisse suduiust
quam illud flagitium udlgo dispaléscere
(from the root of pdlari, to wander, be dispersed abroad). Others
connect it with palma, the hand, and make it mean literally ‘in
the hand.’ Besides the Adv. palam we have pré-palam, as early
as Plautus, but palam is not a Prep. till the Augustan Age.

§ 87. Pénés (governing the Ace., usually of a person), repre-
sents some case of penus, -oris N., or a kindred stem, from the
root pen- of peni-tus (§ 1), pené-tro, &c., a suffixless Locative
according to some (cf. Dor. aiés), a Loc. Pl. according to others,
who offer a similar explanation of vicissi-m (ch. ix. § 4), sémissi-,
via, mou (cf. § 3). The final syllable may have been prevented
from being weakened to -is by the fact that the stress of the
voice fell on it in the common phrases penés-me, penés-te, penés-
nos, penés-vos, &e. (ch. iii, § 12 a4. 3). Penes is used only with
Pronouns in Plautus.

§ 38. Pér, through (Goth. fair-, Lith. pe), connected with
I.-Eur. *perd, ‘I transport, bring or pass through’ (O. Ind. pr-,
O. Sl. pera; cf. Gk. efpw, meipa, Lat. ex-périor, &c.),
corresponds to Osc. per- of peremust, Fut. Pft. of a verb used
§§ 36-88.] PREPOSITIONS. 587

apparently in the old sense of Lat. perémo (Fest. 266. 31 Th.
‘peremere’ Cincius in libro de Verbis Priscis ait significare idem,
quod prohibere : at Cato in libro qui est de Re Militari pro vitiare
usus est), though the commoner form of the Oscan Preposition is
pert (with the suffix -ti of Gk. zpori, § 11), e.g. comono pertemest
‘comitia peremet,’ am-pert, ‘ not beyond ’ (used like Lat. duntawat,
§ 7), petiro-pert, ‘four times’ (cf. Lat. sem-per, § 7), and to
Umbr. per, pert, e.g. per-etom ‘ peremptum,’ ¢rio-per, ‘ three
times,’ which with the Abl. has the sense of Lat. pro, &. @. nomme-
per ‘pro nomine.’ The intensive sense of per- in per-magnus,
per-quam, &c. (often separated from the qualified word, e.g. per
pol quam paucos reperias, Ter. Hec. 58; hence per-taesus did not
become per-tisus, ch. iii, § 23), is seen in Lith. per-saldis, ‘ very
sweet’; cf. Gk. wepi-pijxns, very long, &c. (cf. Engl. ‘through’
and ‘ thorough’), Again Lat. yer- approaches the usual sense of
Gk. zepi, about, around, in pertéyo, perungo, pervolito. The sense
of * past,’ ‘beyond’ (cf. Ose. am-pert, pert viam ‘trans viam,’
Umbr. pert spinia ‘trans spinam’(?)) appears in per-go, &c.;
and with the implication of wrong or injury (cf. Gk. zapa-Baiva,
map-dpvvpt), in per-jurus [from which pejéro, perjero (see Georges,
Lex. Wortf. s.v.), can hardly be separated, though the ¢ is hard
to explain], per-do, per-fidus, and of difference in perégré
It thus appears that per represents a considerable variety of
meanings, and this variety is increased if we take into account
Umbro-Oscan per, pert. For besides the sense of Lat. pro, on
behalf of, seen in Umbr. xomne-per, &c. in the Eugubine Liturgy:
(tio ... ocre-per Fisiu, tota-per Iovina, erer nomne-per, erar
nomne-per . . . subocau ‘te pro arce Fisia, pro populo Iguvino,
pro ejus (M.) nomine, pro ejus (F.) nomine, subvoco,’ estu esunu
fetu fratrus-per Atiiedie ‘ista sacra facito pro fratribus Atiediis’),
once written -pert in the phrase: Petruniapert natine ‘ pro
Petronia natione,’ it has the local sense of Latin pro-, forward,
in front, in the words, Umbr. yer-ne, per-naio-, opposed to post-ne
(Lat. péne, behind), post-naio-, Osc. Perna-, the name of a god-
dess [cf. I.-Eur. *per- in *per-tit(i), from *wetos-, ‘a year,’ O. Ind.
parut, ‘in the former year,’ Gk. wépvo1, Dor. réputs). The reason
of this is that the I.-Eur. root per- produced a large number of
Prepositions, representing different case-forms, &c., *péri Loe.
588 THE LATIN LANGUAGE. (Chap. IX.

(O. Ind. pari, ‘around’ Gk. epi), *p*rds Gen. (O. Ind. puras,
‘before,’ Gk. mdpos), *pérm (O. Ind. pari,‘ beyond’ ; ef. para, Gk.
népa ; Lat. perem- of peren-die, Ose. perum dolom mallom ‘sine
dolo malo’), and so on. The weak grades of the root, pr-, pr-,
seen in Gk. apd, Goth. faura, faur, Engl. be-fore, appear in Lat.
por- of por-tentum, &c., Umbr. pur-titu, pur-ditom (unless this be
merely a metathesis of prd), and in Lat. pro, Umbr. pro, pru,
Ose. pro, pru, as well as in Lat. prae from *prai (Pel. prai-,
Ose. prae, Umbr. pre), perhaps a Dative form (cf. Gk. mapat,
O. Ind. paré).

§ 39. Pé-, retained only in a few Compounds, po-sttus, po-lu-
brum, porceo, &e., as Teut. *mip, ‘with’ (Goth. mip, Germ. mit,
like Gk. era) is in English retained only in the compound ‘ mid-
wife ’ (see under ad, § 12).

§ 40. Post, poné. Post, behind, from *posti, O. Lat. poste,
posti-d (with the particle *d(e)), adds the suffix -ti (§ 11) to
I.-Eur, *pos (Lith. pas, &c.), which seems to be derived from
*pd- (Lith. pa-,‘ under,’ O. Sl. po, ‘ about’), a curtailment of *4pé
(see § 12). In certain collocations the -¢ was dropped by the
Latin phonetic law (ch. iv. $157), e.g. C.L. L.1. 1454 postempus ;
of Virg. Aen, iii. 1 Marius Victorinus says (22. 11 K.): posquam
res Asiae, non ‘ postquam’; and this os might be further reduced
(before m, &c., ch. iv. § 151) to pd-, e.g. pd-merium (so spelt, not
pomoerium), quod erat post murum ‘post-moerium’ dictum,
Varro L. L. v. 143.

Pone (Plaut., &.) adds the suffix -né (§ 11) to post (Umbr. postne,
opposed to perne ; ef. pustnaio-, pusnao- Adj., opposed to pernaio-),

Umbrian post is joined with the same case as pre (Lat. prae),
eg. post verir Treblanir and pre verir Treblanir, in O, Umbr.
pusveres Treplanes and preveres Treplanes, and similarly Ose.
pust feihtis ‘ post fines,’ while Osc. post ewac corresponds to Lat.
posthac.

In Umbro-Oscan we find a Preposition postin governing the
Ace. case with the sense of Lat. secundum, e.g. Umbr. pusti
kastruvuf, ‘ according to their lands,’ Osc. pastin slagim, ‘ accord-
ing to the locality (?),’
§§ 39-43.] PREPOSITIONS. 589

f § 41. Poste, posti-d, pos, pd-. Poste, which shows the regular change of
i when final to é (i not final is retained, e. g. posti-d, see ch. iii. § 39), is found
in a fine line of Ennius, Ann. 244 M., an exhortation to rowers :

poste recumbite, uestraque pectora pellite tonsis,

frequently in Plautus, e. g. Asin. 915 (see Ritsch], Opuse. ii. 541 sqq.), and pro-
bably in Terence, Eun. 493 (see 4. L. L. ii. 140). Its reduction to post is like
that of animalé to animal, nequé to nec, &e. (ch. iii. § 36). The Adverb postid is not
unfrequent in Plautus (e. g. postid locorwm, Poen. 144, &¢.), as also postid-ed (cf.
antid-ed, antid-hac), compounded with the Adverbial Abl. Sg. Fem. of is (§ 10. 4)
(e. g. postidea loci, Stich. 758, &c.), and has on account of its exclusively adverbial
use been regarded as a compound of post(e) with id, the (adverbial) Ace. Sg.
Neut. of is (cf. post-ed, ad id locorum, Sall., Liv.), though this explanation requires
us to see in postidea, antidea,&c. a pleonastic repetition of the pronouns, post-id-ed,
ant-id-ed. Cicero (Orat. xlvii. 157; ef. Vel. Long. 79. 3 K.) says that he preferred
posmeridianas (quadrigas) to postmeridianas, while Quintilian (ix. 4. 39) seems
to mention the form pomeridiem. (On the spelling pos for post in MSS. of
Plautus, see Ritschl, Opusc. ii. 549; of Virgil, see Ribbeck, Prolegg. p. 442;
of other authors, see Georges, Lex. Wortf. s.v.). The evidence points to Lat.
pos- being not I.-Eur. *pos, but asyncopated form of I.-Eur. *pos-ti (see Stolz,
Beitr. p. 21).

§ 42. Prae, before, I.-Eur. *prai (Lith. pré; cf. O. Ir. ré or
ria, with a dropped final nasal). In O. Lat. also rz, according to
Paul. Fest. 282. 27 Th. (cf. pris-cus, pris-tinus, primus for *pris-
mo-, Pelign, Prisma-, pri-stafalacirix ‘ prae-stabulatrices’), prob-
ably I.-Eur. *pri (Lith. pri, O. SI. pri, Goth. fri-), connected
with I.-Eur. *pro, ‘ before,’ and with Gk. mdépos, O. Ind. puras,
pura, Goth. faura, Engl. be-fore (B. B. xvii. 17), possibly a Dative
formation from the root per- (see above, § 38). The Preposition
is found with the same use as in Latin, but with prominence of
the idea ‘before’ (often for Lat. ante), in the Umbro-Oscan
languages ; Pel. prai-cim, Ose. prai, prae-sentid ‘ praesente’ (with
the usual sense of Lat. praesens; in the Columna Rostrata we
have praesens in its older sense: praesente[d] . . . dictatored
ol[orJom, ‘being in command,’ C. I. L. i. 19 5), prae-fucus * prae-
fectus,’ Umbr. pre verir Treblanir * ante portas Treblanas,’ pre-pa,
‘ priusquam,’ lit. ‘ prae-quam,’ pre-habia ‘ praebeat.’

§ 43. Praetér, past, except, is formed from the preceding by
means of the suffix -tero- like the Adverbs drévi-ter, &c. (§ 2), as
from I.-Eur. *pri is formed Pelignian pritro- (in an epitaph,
Zvet. I. 1. I. 13 eite uus pritrome ‘ite vos praeter’ quasi ‘ prae-
terum in’). (Cf, Umbr. pretro- ‘ prior.’
590 THE LATIN LANGUAGE. (Chap. IX.

§ 44. Pro, por-. Prd, before, forth, is I.-Eur. *pré (O. Ind.
pra, Gk. mpd, O. Ir. ro-, used like the Greek augment with pre-
terite tenses, Goth. fra-, Lith. pra-, O. Sl. pro). The long vowel
seems to be the vowel of the Oscan preposition (or I.-Eur. *pri,
Gk. mptravis, &e.), e.g. pru-hipid ‘ prohibuerit,’ pruter-pan * prius-
quam? and may have been I.-Eur. (Gk. mpo-i, early, O. H. G.
fruo, Lett. pré-jam, ‘ forth’). These Oscan forms suggest that
Lat. pro was not originally *prod, an Abl. form (which would be
in Oscan *prud-, not pru-), so that the prod- of préd-est, prod-ire,
may be a form augmented by the particle *-d(e) like anti-d.,
posti-d, r-ed-, &e. In Late Latin the form prode (cf. Charis. 236.
29 K.) is common, especially with esse (cf. Charis. 237. 8 K., and
see Neue, ii®, p. 662) (prodius: ulterius, longius, a prodeundo,
quoted by Nonius 47. 10 M. from Varro, is generally corrected
by editors to propius). The por- of por-rigo or porgo, porricio for
*nor-jicio, por-r6 (O. Lat. porod), &e. is either a metathesis of
pro-, or represents an I.-Eur. by-form *pr (cf. Gk. wapd) (see
ch. iv.§ 92). In Umbrian we have pro- for Lat. pro, before, e.g.
procanurent ‘procinuerint, affixed in ie-pru (cf. promo- ‘ primus’),
and pur- in a verb corresponding in sense to Lat. porricio and in
form apparently to Lat. prd-do, with P. P. P. pur-ditom, Imper.
pur-dovitu,

§ 45. Pré- and pré-. The variety rd- and prd- in Compounds (the simple
preposition has always the long vowel) is seen more in the early literature
than in the stereotyped usage of the classical age : provehat atque prépellat, Luer.
iv. 194 and vi. 1027 ; Lucr. propagare ; O. Lat. pro-tinam. Pro- almost ousts pré-
in class. Latin, but pré- is normal before /-, e. g. pri-ficiscor, pré-fundo, except in
pro-ficio (for préde-facio, as in Late Latin ?) ; but Catullus (Ixiv. 202) has profudit;
Plautus (Men. 643) and Ennius (Zrag. 293 R.) prafiteri; Plautus (Trin. 149)
profecruus. The Greek mpddoyos is prologus in the Comedians (ef. prdpola, Lucil.
v. 28 M.), mpomivw is propinare. Even in classical poetry we have procuro.
O. Lat. pré-tervus might be similarly explained, were it not for the fact that
there are indications of an old form proptervo- (so in the MSS. of Plaut. Bacch.
612, and in the Ambrosian Palimpsest in Truc. 256: see Liwe, Gloss. Nom.
pp. 142, 184, who connects the form with Gk. apomerns). Probus (ef. O. Ind.
prabhi- ‘ preeminent’) apparently adds to pré the same formation (from the
root bheu- ‘to be’ ?) as super-bus (cf. Gk. dmeppuns) to super.

§ 46. Prdctil is formed from *pré by the suffix -ko- [a suffix
often attached to adverbs, e.g. Lat. posti-cus, anti-cus (ch. v. § 31),
réciprocus from *reco- and *proco-, § 49], and some L-suffix (see § 2).
It is used as a Prep. as early as Ennius (rag. 220 R.)
§§ 44-50.] PREPOSITIONS. 591

§ 47. Prdpa (e.g. Plaut. Cure. 97 prope me est) adds to I.-Eur.
*pré the particle -pe (ch. x. § 1. 4). The sense ‘nearly’ is perhaps
later than Plautus (4.L.D. ix. 165). For Superl. proxime we
should expect *prop-(i)s-ime, ch. vi. § 54).

§ 48. Proptér, near, on account of (in Plautus this latter sense
is always expressed by propter; not by ob, when a person is spoken
of), is formed from the Adverb prope by means of the suffix
-tero-, as praeter from prae (§ 43), cireiter from cirewm (§ 20).

§ 49. Ré-, back, has in O. Lat. a byform réd-, with the addi-
tion of the particle *d(e) (§ 11; cf. anti-d, posti-d), which in
class. Lat. remains in red-eo, red-do, &c. (redi-vivus is peculiar),
but is before a consonant usually discarded for ré-, e.g. ré-ditco
(O. Lat. red-duco, but perhaps only 7e-dux), (before a vowel, not
till Late Latin, e.g. reaedifico; see A. L. L. viii. 278). From re-
was formed the Adverb ré-tré (like in-tré, ci-tré, ud-trd), which in
Late Lat. becamea Preposition, e. g. vade retro me, S. Mare. viii.
33, Vulgata, An Adjective-stem *réco- from this Preposition (ch. v.
§ 31) is seen compounded with a stem *prdco- from the Prep. pré-
in the word reci-procus. In Umbrian this Prep. appears in two
Compound Verbs, re-vestu ‘ revisito’ and re-statu ‘ restituito.’

§ 50. Sécundum, sécis. Secundum, according to, close behind,
&e. is the Adverbial Acc. Sg. Neut. of secwndus, following (§ 4).
In plebeian Latin secus was used for secundum (Charis. 80. 18 K. id
quod vulgus usurpat, * secus illum sedi,’ hoc est secundum illum,
et novum et sordidum est; cf. Caper. 103.12 K.; so on plebeian
inscriptions, secus merita ejus, secus viam, &ec., but also in O. Lat.
authors, for Charis. (220. 14 K.) quotes hoc secus, ‘soon after this,’
from the historian Sempr. Asellio), which may be Nom. Sg.
Mase. of an Adj.-stem (cf. heres secus, ‘h, secundus,’ C. 1. L, iii.
387), or Acc. Sg. of a Neuter S-stem *secus, like tenus (§ 54).
Apparently connected with the Prep. secus are O. Ir. sech, ‘ past,’
W. heb, ‘ without,’ from a stem *seq*o-; and the Latin Adv. secus
appended to Adverbs in -im, e.g. eatrin-secus (§ 10. 5), as wellas the
secus of phrases like secus accidit, non secus atque (Comp. sequius) has
been also referred to our Preposition on the theory that this Adverb
meant originally ‘ following but coming short of,’ ‘ less,’ as O. Ir.
sech meant ‘following and going past,’ ‘ more than ’ (see § 8).
592 THE LATIN LANGUAGE. (Chap. IX.

§ 50a, Simi, used as a Prep. in Augustan and later poetry, and
in Silver Age prose, is perhaps Acc. Sg. Neut. of simzdzs (see § 2).

§ 51. Sind, sd. Sime, if connected with O. Ind. sanu-tar (cf.
sani-tir), M. H. G. sun-der (which have the suffix -tero-), must
represent I.-Eur, *sni (*s*nit), and must have been in O. Lat.
*sene, the change from e to 7 being due to the unaccented use of
the word, as in mihi for *mehi (ch. vii. § 1). The spelling seine
on the Lex Repetundarum (C.J. L. i. 198. 54) is best explained,
like Zeiteras on the same inscription, as a case of ed wrongly used
for % (ch, ii. § 130).

In O. Lat. there is another Preposition of the same meaning,
sé(d) (as an Adverb meaning ‘ apart’), which became obsolete
except in Compounds like séd-i#io, lit. ‘a going apart,’ se-orsum (-8)
(a dissyll. as early as Plautus), sé-diilé from sé dolo (whence the
Adj. sédulus, § 7), sé-curus, sé-cédo, sé-cerno; it is evidently
connected with the Conjunction séd (ch. x. § 5); the d of séd-itio,
&c. need not be the Abl. Case-suffix, but may be the particle *d(e)
(cf. re-d-, § 49). This Preposition occurs in the legal phrase se
fraude, ‘ without hurt,’ free from penalty, written sed fraude on
the Lex Repetundarum of 123-122 B.c. (C.J. Z. i. 198), and on
the Lex Agraria of 111 B.C.(i. 200, ll. 29 and 42, but se dulo malo
1. 40) (cf. Paul. Fest. 500. 6 Th. ‘sed’ pro sine inveniuntur posuisse
antiqui). The compound sed-diterque (cf. Umbr. sei-podruhpet)
occurs in the Nom..Sg. Fem. in Plaut. Stich. 106 sedutraque ut
dicat mihi, This sé(d) has been plausibly connected with the
I,.-Eur. Reflexive Pronoun-stem *swe- (Lat. sé Acc.), and explained
as originally meaning ‘ by oneself.’ The Old Slavonic Preposi-
tions své-né, své-ni, své-nje, ‘except, without,’ in which this root
swe- appears with an N-suffix, suggests that Lat. *se-ne may also
be connected with the Reflexive Pronoun. ‘The first part of the
Compound Verb so/-vo, and the Adjs. s6-brius (cf. é-brius, according
to Charis. 83. 16 K. from dria ‘ vas vinarium’), sd-cors (cf. secordis
‘stultus, fatuus, C. G. L. iv. 282. 52), is of kindred origin. Festus
quotes zes? ‘ pro sine positum’ from an inscription on the temple of
Diana on the Aventine, but the fragmentary condition of the MS.
of Festus for this passage (nesi pro sine positum ... Dianae Aven-
tinen..., 166, 26 Th.) makes it doubtful whether the word is
§§ 50a—54.] PREPOSITIONS. 593

not really the conjunction zisi, used in a context which gave it
the force of sine.

§ 52. Sib, subtér, subtis. Swi, under, is I.-Eur. *tipd
(O. Ind. tipa, ‘ to,’ Gk. 57d, under, for izé, O. Ir. fo, W. gwo-, go-,
Goth. uf), The initial s-, which is found also in the Umbrian
Preposition su(b), e.g. subocau, ‘ subvoco,’ su-tentu ‘ subtendito,’
as well as in super (§ 53), is generally explained as a curtailment
of prefixed ex (I-Eur. *els), so that sud would represent an
I.-Eur. compound Preposition *eks-upo, but is as likely to be
the particle -s(e) (§11). This particle -s(e) is postfixed in the
form sus- for sub-s, e. g. sus-tineo, sus-que de-que (in the O. Lat.
phrase susque deque fero, habeo (Plaut. Amph. 886), explained by
Gell. xvi. 9 as meaning ‘aequo animo sum’), apparently used by
wrong analogy in old forms of swmpsit and sumpserit quoted by
Paul. Fest. 425. 3 Th., suremit and surempsit. Sub-ter is a form-
ation like prae-ter (§ 43), and sub-tus like in-tus (§ 33). The
diminutive sense of Lat. sud in sub-absurdus, &c. is shared by Gk.
bz (e. g. dadAevKos) and O. Ir. fo (e. g. fo-dord, ‘a murmur’).

§ 53. Stipér, stpra,insuper, superné. Super is I.-Eur. *tipér,
*tipért (O. Ind. upari, Gk. iaép for ixép, O. Ir. for, O. W. guor-,
Goth. ufar) with a prefixed s- as in Lat. sub, just mentioned.
Umbr. super governs the Loc., e.g. super kumne ‘super comitio.’
Stipra is an Abl. Sg. Fem. (suprad on the S. C. Bacch.) like
in-tra, ci-trd, &c. The form supera is quoted by Priscian (ii. 30. 3
and 55. 23 H.) from Cicero’s poems, and is found in Lucretius as
well as on C. J. Z. i. 1011. (2). 11 (seep. 181), The Umbrian equi-
valent is sobra, governing the Acc, e.g. sobra tudero * supra
fines.” Super-né (also superné, Adv. of supernus) is formed from
super as poné from post (§ 40). It is not used as a Preposition in
Latin, at least in classical Latin, but is so used (governing an
Acc.) in Umbrian, superne adro ‘super atra (vascula).’ Insuper
is a Prep. as early as Cato (&. 2. xviii. 5).

Z
§ 54, Téniis, apparently the Adverbial Acc. Sg. of a Neuter
S-stem *tenes-,from the root ten-, ‘to stretch’ (cf. tenws N.,‘a cord,’
in Plaut. Bacch. 793 pendébit, hodie pilcre ; ita intendi tenus) is
used as a Preposition as early as Ennius. It takes the Abl. Sg.,

Q4q
594 THE LATIN LANGUAGE. [Chap. Ix.

e.g. Tauro tenus, Cic. (originally ‘from T. in a line,’ then used
for ‘usque ad’), and the Gen. Pl., e.g. crurum tenus, Virg., and
in Late Latin the Acc., e.g. Tanain tenus, Val. Flace. It is not
found in prose till the Silver Age (see 4. D. L. i. 415). Tenus is
appended to Adverbs in -d (older -dd Abl. Sg. Fem.) in the sense
of ‘as far as,’ lit. ‘stretching from,’ e.g. aliqua-tenus, hac-tenus
(on which see § 10. 4; cf. prd-tinus, § 4). 7

§55. Trans, across (Umbr. traf, trahaf), is either the Pres. Part.
of *trdre, in-trare, péné-trare (probably the Nom. Sg. Mase.
crystallized in Adverbial usage; cf. § 2), or is an extension of
an obsolete Preposition *¢ram (an Acc. Sg. Fem. from the same
root) by the same particle *s(e) as appears in the Prepositions of
kindred meaning, wé-s, ci-s. It is clearly connected with O. Iv.
tré or tria, tri, trem-, tar, W. trwy, trach, O. Ind. tirds, all from
the root ter-, ‘to go through, drive through’ (O. Ind. tr-, ‘ to bore,’
Gk. reipw, Lat. téro, &e.). The -ans, pronounced -as (ch. ii. § 66), of
trans became before voiced consonants @ by the Latin phonetic law
(ch. iv. § 151), e. g. trd-do (but transdo C.I.L.1. 198, ll. 54, 58, &c.),
tra-mitto (and trans-mitto). Umbr. traf, trahaf (i.e. traf), tra
governs the Acc. with a verb of motion, e.g. traf sahatam etu
‘trans Sanctam ito,’ the Loc. with the idea of rest, e. g. trahaf
sahate vitla trif feetu ‘ trans Sanctam vitulas tris facito.’

§ 56. Uls, ultra, beyond, on the other side, came from the same
root, I.-Eur. ol- (whence Lat. i/e, O. Lat. olle for *ol-sé, ch. vii. § 13),
the first being augmented by the particle *s(e) (§ 11),’ the
second (an Abl. Sg. Fem.) by the suffix -tero-(ch.v.§16). Thed
has become # before the combination 7 with a consonant by the
Latin phonetic law (ch. iv. § 20), but the original vowel appears in
oltemus (Ose. tltiumo-) in Ennius’ description of Servius Tullius

(4. 337 M.):
mortalem summum fortuna repente
reddidit ut summo regno famul oltimus (MSS. optimus) esset.

(The ovis of the MS. of Varro L. Z. v. 50 isa scribe’s emendation
of vis, the same mis-writing of ws as occurs later in v. 83, or is
due to the correction of o/s to ués, and should not be printed oués,
which would imply és. The shortness of the vowel is proved by
the Romance forms of wltra; cf. Gell. xii. 13. 8 on the extension
§§ 55-58.) PREPOSITIONS. 595

of in, cis, uls to intra, citra, ultra: quoniam parvo exiguoque
sonitu obscurius promebantur, addita est tribus omnibus eadem
syllaba. It is a mistake to suppose that there is an accent to
indicate length over the first vowel of w/tra in Claudius’ tablet
at Lyons.) An original *ol-s(e) must have become *o//in Latin ;
the form w/s is due to a later re-addition of s on the analogy of
et-s, &e. (cf. ch. viii. § 68 on fer-s).

§ 57. Usqué, with long w, to judge from Romance forms like
O. Fr. usque, Fr. jusque for de wsque (for Lat. % would be
represented by o, see ch. ii. § 26), is formed from the I.-Eur. Pre-
position *ud, ‘out, up out’ (O. Ind. ad, Goth. ait, Engl. out) in the
same way as absgue from I.-Eur. *ap(o) (§ 12), so is not
connected with wsguam (§ 10. 9).

The Prepositional use of this Adverb (see § 11), e.g. usque
quintum diem, Cels. (in Cic. only ‘ usque Romam,’ &c.; so Ter.
Ad, 655 Miletum usque, but Cato &. A. xlix. 2 usque radices
persequito), is due to a curtailment of the proper phrase wsque
ad, much as in Attic Greek os (for os eis) came to be used as
a Preposition, e. g. @s Tov Bactréa lévar. The Latin grammarians
point out that wsgue, unlike other Prepositions (cf. p. 573), can
take a Preposition as prefix, e.g. abusque, adusque (Expl. in
Donat. 517. 22 K. nemo enim dicit ‘de post forum,’ nemo enim
‘ab ante’; at vero dicimus ‘ ab usque’ et ‘ad usque’); ab usque
was a poetic inversion of Virgil’s which found its way into Silver
Age prose (4.L.L. vi. 80) ; ad usque (first in Catull. iv. 24) is
likewise a phrase of Augustan poetry and Silver Age prose
(A. L.L, vii. 107).

58. Versus, versum, adversus, adversum, exadversus,
exadversum. Versus apparently a Nom. Sg. Masc., as verswm
is an Acc. Sg. Neut., of the P. P. P. versus, corresponds to the
Celtic Preposition meaning ‘towards,’ ‘against’ (O. Ir. frith, fri,
W. wrth, O. W. gurt). On its Adverbial use, see § 2.

Qq 2
CHAPTER X.

CONJUNCTIONS AND INTERJECTIONS.

§1. CONJUNCTIONS. As Prepositions are hardly separable
from Adverbs of Locality, so Conjunctions are closely connected
with pronominal Adverbs. These pronominal Adverbs, as
we have seen (ch. ix. § 10), are not always capable of being
referred to their proper case-form (e.g. %2, #7), owing to our
imperfect knowledge of the declension of the I.-Eur. pronoun.
Nor is it easy to find their cognates in the various I.-Eur.
languages; so rapidly does the meaning of a Conjunction
alter. Thus Latin éim, which in the older literature is
a particle of asseveration, ‘indeed,’ had by the classical period
appropriated the sense of ‘for’; and in French, pas (Lat.
passus) and point (Lat. punctum) have acquired a negative sense
from their use in the phrases ne... pas, ne... point. A feature
of J.-Eur. Conjunctions is their tendency to append other
Conjunctions or conjunctive Particles (e.g. és in Greek may
append 57, ep, &e., as 37, ds wep); and this habit puts another
obstacle in the way of identifying cognate Conjunctions in
different languages, for in one language they may appear
extended by one particle, in another language by another.
The exact form of these conjunctive Particles is also a difficult
thing to ascertain ; we often see parallel stems in -o, -i, -u, &e.
(e g. Fqto-, *qFe-, *q4i-, *q¥u- are all various forms of the
Relative and Interrogative Pronoun-stem, ch. vii. § 23; -t8 and
-ti appear in O, Ind. u-té, Gk. at-re, O. Ind. i-ti, Gk. &71), and
parallel forms with long and with short vowel (e.g. Negative
*né and *né appear in O, Ind. né and na, Lat. né- and xé-; I.-Eur.
*we, ‘or, O. Ind. va, Lat. -vé); and the tendency was always
§ 1.] CONJUNCTIONS. 597

present to adapt the ending of one Conjunction to the ending of
another Conjunction of similar meaning (e.g. Lat. sadtem for
saltim, adapted to au-tem, t-tem ?, ch. ix. § 4). It will therefore
be best to designate these conjunctive Particles according to their
consonants, as, for example, (1) the T-particle of Lat. tam, %-tem,
u-t(t), Gk. at-re, O. Ind. u-ta(-a), {-ti, ‘so, O. SI. te, ‘and’; (2) the
D-particle of O. Ind. i-da, ‘now ’ (Lat. %dd-neus ?); dum, tbi-dem,
Gk. 57, d¢, 8-5 ; (3) the DH- particle of Gk. év-0a, O. Ind. ki-ha,
‘where,’ O, SI. kti-de, ‘where;’ (4) the P-particle of Lat.
quip-pe, nem-pe, Lith, kai-p, ‘how, as,’ szeip . . . tetp, ‘so... so’;
(5) the N-particle of Lat. nam, num, nem-pe, quis-nam, O. Ind.
hi-né, ‘for, O. SL tu-nti, ‘then.’ These particles are not easily
distinguished on the one hand from the particles affixed to Pre-
positions (e. g. *-ti of O. Ind. pra-ti, Gk. mpo-ri, Osc, per-t ; *-n&é
of Lat. po-ne, Umbr. post-ne, Germ. vo-n), as has been already
mentioned (ch. ix. § 11), nor on the other are they always to be
distinguished from Case-suffixes. Indeed the usage of the oldest
Indian literature, where, for example, the particle k4m is often
added to a Dativus Commodi or to a Dative of Purpose (see
Delbriick, Altind, Syntax, p. 150), and other particles are more or
less allotted te special cases, suggests that the Case-suffixes may
have at the first originated in this way, just as Gk. dy came in
time to be a sign of a Mood of the Verb. Thus not only has
the -s of the Nom. Sg. Masc. been with great probability
referred to the pronominal-stem *so- (*se-) (ch. vii. § 13), but also
the Abl. -d to the suffix *dé expressive of motion, joined with
an Accusative, in the sense of motion towards, in Gk. dduov-de,
Gen. -s (which in Greek and other languages has the function
of an Abl.) to the similar *sé of Gk. GAdo-ce. The person-
suffixes of Verbs may often have had a similar origin. The
*dhi of the 2 Sg. Imper. in O. Ind., Gk. &e., e.g. t-61, is the
asseverative particle *dhi, joined to Imperatives, as Lat. dwm in
dgé dum ; the *-tod of the 2 Sg. Fut. Imper., e. g. quando uidebis,
dato, Plaut., is the Adverbial Abl. Sg. Neut. of the Pronoun *to-,
‘from this,’ ‘ thereupon’ (ch, viii. § 57). And in the declension of
the Pronouns themselves we have clear instances of the progress of
appended particles to case-suffixes in *gé (Gk. ye) used as the sign
of the Acc. Sg. in Goth. mi-k, Germ. mi-ch, *ghi (O. Ind. hi,
598 THE LATIN LANGUAGE. (Chap. X.

Gk. vat-x.) as the sign of the Dat. Sg. in O. Ind. mé-hy-am, ch.vii.
§ 1). The -d of the Acc. Sg. of the Personal Pronouns in Latin
has been similarly explained as the particle *id, so common in the
oldest Indian literature, where it is used to emphasize a preceding
word, so that Lat. ¢éd was originally */é d (cf. tuam id in the
Rig-Veda) (see ch. vii. § 1).

§ 2. (1) Conjunctive.—Que, et, atque, ac, quoque, etiam.
- Qué, I.-Eur. *-q?é (O. Ind. ca, Gk. ré, Goth. -h, e.g. ni-h * ne-que’),
apparently the bare stem of the Relative *q*o-(*q"e-) (ch. vii. § 23),
is in Latin, as it was in I.-Eur., an enclitic appended to the first
word of the sentence. Through Syncope, to which final -é was
always liable in Latin (ch. iii. § 36), it has become -c in nec (neque),
ac for *at-c (at-que), &c., and probably often had this sound before
an initial consonant in the rapid utterance of every-day life. In
some lines of Plautus (Stich. 696, Capt. 246, Poen. 419, &e.) we
must, if the reading of the MSS. be right, scan: démq(ue)
se exdrnat ; pérq(ue), conséruitiim commine, &c. (Skutsch, For-
schungem, 1. p.151). .-Eur. *-q*é gave a relative and indefinite
sense to pronouns, and so in O. Latin, though in the classical
period the fuller ending -cungue (O. Lat. -quomque, e. 2. queiquom-
que, C. I. L. i. 197. 53 198, &e.; see Georges, Lew. Wortf: s.v.) is
preferred, e.g. guem-que Plaut. for quem-cunque (O. Ind. ka8-ca,
Hom. Gk. és re, Goth. hvd-h F.) ; so guis-que, each (cf. O. Ir. ca-ch,
W. pawb, O. W. paup, apparently from I.-Eur. *q*d-q"é or
*q"a-qré). This -cungue seems to be nothing but cwm-que, ‘ when-
ever’ (Hor. C. i. 32. 15), though some connect it with O. Ind.
cand (with ka-, & , ‘whoever,’ &c.), and others make the -cwm-
(-cun-) a byform of wm-(wn-) of wm-quam, &e. (ch. ix. § 10. 7).
The corresponding particle in Oscan is -pid (O. Ind. -cit), e.g.
pokka-pid ‘ quandoque.’ Lat. gud-gué is composed of some part
of the Pronoun-stem guo- (que-) and the enclitic -que (perhaps
the bare Pronoun-stem ; cf. O. Ind. kva-ca, ‘anywhere, in any case,’
from kva, ‘where,’ and ca, Lat. -que). Similarly, at-qué, of the
Preposition (Adverb) ad and the enclitic, lit. ‘and to,’ ‘and
further’; in O. Lat. it often signifies ‘forthwith, e.g. Plaut.
Most. 1050:

quéniam conuocdui, atque illi me éx senatu ségregant.
§§ 2-4.] CONJUNCTIONS. 599

Umbrian aye, when, also spelt api, appei, may be the same
formation as Lat. atgue. The Umbro-Oscan equivalents of Lat.
néqué, Osc. nep, neip, nip, Umbr. neip, wep, have -p for I.-Eur.
*q*é. Et is the I-Eur. Adverb *éti (O. Ind. &ti, ‘ over, Gk. ér.,
further), used in Latin, as in Gothic (ip, ‘and’), for the copula.
It may be that it gradually encroached on the sphere of the
older -gué, for it is noticeable that only -gue, not é, is found in
the (restored) inscription on the Columna Rostrata (C.J. L. i.
195). The Umbrian copula is also et, but in Oscan inim,
a word related to Lat. enim (see below).

In é-iam, et is associated with the Adverb jam, now, the
7 (y) becoming vocalic by the Latin phonetic law in the middle
of a word, as in médius (I.-Eur. *médhyos, O. Ind. médhyas,
Gk. péo(c)os, &e.) (ch. iv. § 67).

§ 8. Atque, ac. On Republican Inscriptions the rule is that atgue be
used before an initial vowel, ac before an initial consonant, and so in the
MSS. of Terence. But in the MSS. of Plautus atque is sometimes used before
a consonant, where the metre requires the pronunciation ac (@. g. Epid. 522),
and in the MSS. of Cato atque is the prevailing spelling (whatever Cato’s
pronunciation may have been) before initial consonants and vowels alike.
The classical authors, as well as Plautus, seem to avoid ac not only before
vowels, but also before c-, g-, g- (see Georges, Lex. Wortf. s. v. ; Skutsch, Forsch.
i. 52; B. P. W. xiii. 312). Atque is the spelling in Republican inscriptions ;
adque occurs in the Res Gestae of Augustus (once), and is in later inscriptions
very frequent, as well as in good MSS. (see Neue, ii*. 953). Aique atque seems
to mean ‘nearer and nearer’ in Ennius, Ann. 519 M. :

atque atque accedit muros Romana iuuentus.

§ 4. (2) Disjunctive.— Ve, aut, vel, sive, seu. -Vé is I.-Eur.
*_wé, (O. Ind. va, e.g. néktam va diva va, ‘by night or by day,’
Hom. Gk. 7-(F)é), probably a curtailment of an I-Eur. Adverb
*iwé (O. Ind. ava, ‘away’). The I.-Eur. particle had also the
sense of ‘as,’ ‘ like,’ seen in Lat. ce-w, O. Ind. i-va, ‘as,’ e-va,
‘thus, later e-vam.

Aut is compounded of I.-Eur. *au [Gk. ad, again, Goth. au-k,
‘also’ (quasi *ad-ye), Engl. eke], another curtailment of the same
Adverb (cf. au-fiigio, auf-tro, ch. ix, § 12), and the particle -ti (§ 1).
Similar are Umbr. o¢e, Osc. avéi and avt, though the latter Oscan
form has generally the sense of Latin autem.

Véi is the old 2 Sg. Pres. Imperative of vdlo (ch. viii. § 58), lit.
600 THE LATIN LANGUAGE. (Chap. X.

‘ choose, as Germ. wohl (e. g. Homer, wohl der grésste Dichter,
‘Homerus vel summus poeta’) was originally Imper. of wollen.

Vel can hardly represent vell for *vels, an old 2 Sg. ‘ Injunctive,’ for it is so
thoroughly a short syllable in Plautus as to be capable of acting as a Brevis
Brevians (ch. iii. § 42), e.g. Poen. 827 uél in lautumiis, uél in pistrino, although
Umbr. heris .. . heris, e. g. heris vinu heri puni ‘ vel vino vel posea,’ is 2 Sg.
Ind. of heri-, ‘to wish’ (whence Herentas, the Oscan Venus).

Other instances of Imperatives used as Particles are piité, for example, Hor.
and dgé; em, the Interjection, probably represents émé, 2 Sg. Imper. of emo,
I take (§ 19).

Swé is compounded of si, older se, and -ve. Before -w, the
curtailed or syncopated form of -ve (as -c of -gué, -u of Interro-
gative -né, &c.), the ci-diphthong was by the Latin phonetic law
(ch. iv. § 66) reduced to @ (as in deus from detu(w)s, ch. tv. § 33).
Lat. si was in Umbrian sve (Ose. svat), and the Umbrian
equivalent of Lat. sive is sve-po ‘ siqué,’

§ 5. (3) Adversative.—At, ast, sed, autem, atqui, tamen,
ceterum, verum, vero. Af is the I.-Eur. Adverb *At(1) (O. Iv.
aith-, ‘back,’ Lith, at-, O. Sl. otti, ‘from’), used in Latin, as in
Gothic (ap-pan, ‘ but’), as a Conjunction. On its confusion in
spelling with the Preposition ad, see ch. ii. § 76.

Ast is a Conjunction found in old laws in various senses (Charis.
2.29. 30 K, ‘ast’ apud antiquos variam vim contulit vocibus,
pro atque, pro ac, pro ergo, pro sed, pro tamen, pro tum, pro
cum, ut in glossis antiquitatum legimus scriptum), especially (1)
‘if further, ‘ and if moreover,’ e.g. Lex Serv. Tull. : si parentem
puer uerberet, ast olle plorassit, puer diuis parentum sacer esto ;
(2) ‘if, eg. XII Tab. ro. 8 (in the curious law referring to the
use of gold in dentistry): ... neue aurum addito, at cui auro
dentes iuncti escunt, ast im cum illo sepeliet uretue, se fraude
esto), and occasionally in the early writers, e.g. with the sense of
“if further, Plaut. Capt. 683 :

si ego hic peribo, ast ille ut dixit nén redit,
with the sense of ‘further’ or ‘but,’ Accius, 77ag. 260 R. :
idem splendet saépe, ast idem nimbis interdim nigret.

It may be a formation from ad, and stand for *ad-s-ti as post for
§ 5.] CONJUNCTIONS. 601

*po-s-ti (ch. ix. § 40), so that its original signification would be
‘further,’ ‘ moreover,’ though, owing to the custom of using it in
the added clause of the protasis in conditional sentences, it came
to acquire the notion of ‘if further, and even of ‘if.’ It is one
of the archaisms used by Cicero in drawing up his code of laws
(Legg. ii. 8. 19, &c.), who gives it the senses of (1) ‘if further,
(2) ‘if’ (so on the law relating to the Ludi Saeculares of Augustus’
reign, ast quid est ‘ siquid est’), (3) ‘further ’ (so in Cicero's trans-
lation of Aratus’s Prognostica, 1. 160). The Augustan poets
revived the use of the word, as a substitute for a¢, where the
metre required a long syllable, and in the second cent. a.p. it
passed into prose. If asted on the Dvenos inscription (usually
explained as adstet) be really ast, it is a byform with the particle
-d(e) like postzd (ch. ix. § 40).

Séd,if we may believe the statement of some grammarians
(Charis. 112. 5 K.; Mar. Victorin. 10. 13 K.; Ter. Scaur, 12.
8 K.; Isid. Orig. i. 26. 24), who argue against the spelling se¢
(ch. ii. § 76), was at some early period sedwm. The word can
hardly be separated from the Preposition (Adverb) sé (séd),
‘apart’ (ch. ix. § 51), and may be a compound of *sé, a byform of
sé, with the Conjunctive particle dwm (see below).

dAutem adds the particle -tem (cf. %-tem) to the I.-Eur. Adverb
*au [Gk. af, again, Goth. au-k, ‘also’ (quasi *ai-ye), Engl. eke],
which is probably identical with the Preposition au- of Lat. au-
féro, au-figio (ch. ix. § 12), and cognate with the I.-Eur. Con-
junction *wé (see under Lat. -ve). The Oscan equivalent of autem
is avt, apparently a ‘ doublet’ of avti, the equivalent of aut (§ 4).
The older usage of autem is seen in passages like Plaut. Merc.
_ et cirrendum et pugndndum et autem itrigandumst in uia,

(cf. sed autem, Rud. 472; et autem, Poen. 841).

Atgui adds to the Conjunction at the particle gui, which is
much used by the early Dramatists as a mere particle of emphasis
[e.g. Plaut. Hercle qui (Pseud. 473), utinam qui, ut qui (Lrim,
637)], and which is either the Abl., Loc. or Instr. Sg. of the
Relative (ch. vii. § 25). Atguin (on this spelling, see Georges,
Lex. Wortf, s.-v.) has the particle -n(e) appended (§ 1, above).

Teimén however, ‘none the less,’ is clearly related to ¢am (ch. ix.
602 THE LATIN LANGUAGE. [Chap. X.

§ 10. 8), so,‘ equally much,’ which was often used in the sense of
tamen in O. Lat. (Fest. 548. 3 Th. antiqui ‘tam’ etiam pro tamen
usi sunt, with examples from Naevius, Ennius, and Titinius ; for
examples in Plautus, see Seyffert, Stud. Plaut. p. 14) (cf. class.
tam-etst and tamen-etsi). Some see in tam-en a relic of the
earlier usage, retained in Umbro-Oscan, of putting the Preposi-
tion im (older en) after the word it governs, e.g. Umbr. arvam-
en ‘in arvum, Pel. pritrom-e ‘ praeter’ quasi ‘ praeterum-in’
(ch. ix. § 30); others suppose that the particle -xé (of guando-ne,
&e., § 1), appended to ¢am, produced tam-i-ne or tamen(e), and
quote Plaut. Mz/. 628, where the MSS. reading points to ¢amine,
as a proof that interrogative -ze appended to ¢am produced this
same form :

tim capularis? tamne tibi diu uideor uitam uiuere.

On the other hand éanne is mentioned by Festus 542. 26 Th.
as the O. Lat. form of tam with interrogative -e, and exempli-
fied by Afranius, Com. 410 R.: tanne drcula Tia plena est
arénearum? Festus also quotes tame, as an old form of tam, on
which see ch. ix. § ro. 8.

Cétérum is the adverbial Acc. Sg. Neut. of the stem cétero-
(Nom. Pl. céteri), as cetera in such a line as Virg. Ad, ix. 656:
cetera parce puer bello, ‘for the rest—you are a boy—deal
sparingly with war,’ is an adverbial Acc. Pl. Neut. Ceterwm of
Plautus’ Truc. 847, &c. is exactly parallel to wewm of Plautus,
Mul. Glor. 24 nisi Gnum epityra ei éstur insanim bene, ‘ but—
one thing,—&c.’ The root of the word is the I.-Eur. pronominal
*ke, which shows the short vowel in the Latin enclitie -ce of
hujus-ce, &e. (ch. vii. § 15), the long vowel in Lat. cé-deri, &e.
(see § 1 on the variation of quantity in pronominal *wé, *né,
&e., and cf. ch. iv. § 33).

Vérum is similarly an adverbial Acc. Sg. Neut. of the Adj.-
stem vero-, true, and verd an adverbial Abl. (Instr.?) Sg. Neut.
of the same stem.

§ 6. (4) Limitative and Corrective—Quidem, immo. The
formation of guidem has not yet been satisfactorily explained.
The qui- may be the bare stem of the Pronoun (see § 2 on qud-
§§ 6, 7.] CONJUNCTIONS. 603

qué); or if Ydem represents *d-dem, quidem may be the Neuter
Pronoun with the suffix -dem (§ 2) *guid-dem (but see ch. vii.
§ 21 on tdem),

Equidem might be similarly explained as e¢-guidem [ef. Plaut.
Pers. 187 et quidem (A), eq. (P)], but is better referred to the
pronominal prefix &- of e-nos, Umbro-Osce, e-tanto-, &c., aug-
mented by the particle -ce in ec-quis, and in ecce (see § 19). Its
association with the first Personal Pronoun in Cicero and other
good writers (see Neue, ii®. p. 96 3) shows that to a Roman the
first syllable suggested a reference rather to ego than to et (cf.
Prise. ii, 103. 5 H.; ecce in Plautus very often refers to the per-
son speaking, 4. L. L. vy. 18). The exact truth regarding its use
in Plautus is not easy to ascertain, for the MSS. frequently write
it for e¢ quidem (e. g. Pers, 187), and editors often substitute it for
quidem after tu, me, &c. to avoid the scansion ti quidem, mé qui-
dem (like siquidem, ch. iii. § 51), or the division of a dactyl between
two words in iambic and trochaic metres, e.g. atque quidem.

Immé (not imo}, according to Brambach, Hiilfshiichlein, s. v.)
has the scansion of a pyrrhic (Uv) according to the MSS. in
passages like Terence, Phorm. 936 :

immo uéro uxorem ti cedo. In ius dmbula,
a scansion which has not yet been accounted for. Nor is the
derivation of the word at all clear. One theory makes it Adver-
bial Abl. Sg. Neut. of us, another analyzes it into in-md, ‘in
magis,’ supposing *md to be an I.-Eur. Comparative ‘ more,’
whence comes Gaul. -ma-rus of Virdo-marus, O, Ir. mar, mér,
‘ great,’ &c.

§ 7. (5) Explanatory.—Enim, nam, namque, quippe, nempe,
nemut. Lnim, in O. Lat. an asseverative particle merely (ef.
class. enim-vero), a usage imitated by Virgil, e.g. A. viil. 84:

quam pius Aeneas tibi enim, tibi, maxima Juno,

mactat sacra ferens,
is most naturally referred to I.-Eur. *eno- (*ene-) (cf. O. Ind. ana,
‘indeed,’ ‘for’), another form of I.-Eur. *no- (*ne-) (§ 1), (cf. dim
from z/e, istim from iste, ch. ix. § 10. 5). The weak point of this
etymology is that it prevents us from connecting the word

 

1 Imo occurs in the Aes Italicense of 176-180 a.D. (C. I. L. ii. 6278, 1. 20).
604 THE LATIN LANGUAGE. (Chap. X.

directly with einom ‘igitur’ of the Dvenos inscription, Pel.
inom (?), ‘et,’ Osc. inim, ‘ et,” Umbr. enom (enum-ek, inum-ek,
with the particle -ce of Lat. hujus-ce, &c.), also enem ‘tum,
which all show the Pronominal root i- of Lat. is (ch. ix. § 19)
prefixed to a form (-nim, -nom) of the root no- (ne-). Lat. enam
(from *e-no-) will stand to O. Lat. eivom of the Dvenos inser.
(from *ei-no- ; cf. O. Ind. éna-, ‘he, éna, ‘so, here’), as O. Ind.
asa-t, ‘ this’ (from *e-so-) to O. Ind. é84-, ‘ this’ (from *ei-so-),

Nam, often used in O. Lat. in questions, e.g. ‘quid cerussa
opus nam?’ ‘why, what is the use of paint?’ Plaut. (cf. guis-
nam), without that definite sense of ‘for,’ ‘because’ to which the
word is restricted in classical literature (but cf. wé¢-nam), is the
same case-form of the Pronominal-stem *no- (§ 1) as guam of
*q¥o-, tam of *to- (Acc. Sg. Fem. ?, ch. ix. § to. 8).

Namque adds to nam the enclitic -gué (§ 2). It is used only
before an initial vowel in Plautus and Terence.

Quippé appends the particle -pé (§ 1) to some case of the
Relative or Interrogative or Indefinite Pronoun, either the Ace.
Sg. Neut. guippe for *quid-pe (cf. quippini equivalent to quidnt),
or (if -ipp- can represent -ip- in Latin; cf. ipsippe ‘ipsi neque
alii’ Paul. Fest. 74. 37 Th., and see p. 116 2.), the Loe. Instr.
Sg., quippe for *qui-pe, or else the Abl. Se. guippe for *guid-pe.

Nempé appends the same particle to a form *xem (the same
case-form of the Pronominal-stem *no-, as -tem of autem is of
the stem *to-), This form *xem appears also in O. Lat. nemut
‘nisi etiam vel nempe’ (Fest. 160. 28; Paul. Fest, 161. 13 Th.).
On the pronunciation xemp(e) before initial consonants, see

ch. iii, § 35,

§ 8. (6) Conclusive.—Ergo, itaque, igitur, rgd has already
been explained, in connexion with its use as a Preposition, e.g.
funeris ergo (ch. ix. § 28), as possibly standing for é *rdgo (ef.
e régione), ‘from the direction,’ and has been compared with
German wegen, M. H. G. vonwégen.

Itiqué, compounded of %té (ch. ix, § 10. 9), and -gué (§ 2), seems,
like zéa, never to have @ even in the earliest poetry. We should
scan the Saturnian line of the epitaph of Naevius (ap. Gell. i. 24.2):

itique péstquam est Orcho trdditus thesatiro (see p. 128 nN).
§§ 8-10.] CONJUNCTIONS. 605

The grammarians of the Empire distinguish itdéque, therefore, from ddéque,
and so, ‘et ita,’ (e. g. Serv. in Don. 427. 13 K. tune corripitur media cum una
pars fuerit orationis, tune vero producitur cum duae), but short a is invari-
ably shown in the Dramatists in both senses of the word. The grammarians’
rule about itaque resembles their rule for the penultimate accentuation of
plerdque, utréque, Nom. Sg. Fem. (cf. ch. ii § 93 on Late Lat dqua, acqua).

Igitur had in O. Lat. the sense of tum (Paul. Fest. 74. 29 Th.
‘igitur’ nune quidem pro conpletionis significatione valet, quae
est ergo. Sed apud antiquos ponebatur pro inde et postea et
tum), as in Plaut. Id. 772:

quindo habebo, igittr rationem medrum fabricarim dabo,

or in the first law of the XII Tab.: si in ius uocat, ni it, ante-
stamino. igitur em capito. The etymology of the word, one of
the most puzzling in Latin, is discussed in ch. ix. § 8.

§ 9. (7) Optative.—Ut, utinam, U/¢, in wishes, e.g. Juppiter
ut Danaum omne genus pereat, is the Conjunction #, that (older
uti, ch. ii. § 36), with suppression of the idea ‘I wish’ or ‘do
thou grant.’

In t#tinam the final 7 of w#(Z) is retained, and nam has its older
sense of a strengthening particle, ‘indeed’ (§ 7).

§ 10. (8) Interrogative.—-Ne, nonne, num, utrum, an,
anne, cur, quare, quianam. In class. Latin -é is the general
interrogative particle, while xonné is limited to questions which
expect an affirmative, zum to those which expect a negative,
answer. This distinction is unknown to Plautus, who uses zonne
hardly at all (e. g. Trin. 789), (-ne being used instead, e.g. Trin.
178, Men. 284, or non, e.g, Stich. 606), and num, numguis without
a negative sense occasionally, e.g. Most. 999. (A list of examples
of the Interrogative Particles in Plautus and Terence is given in
Amer. Journ, Phil. vol. xi. 1890.) It is easy to see how these
meanings came to be attached to non-ne, ‘is... not,’ and num,
‘now’ [Gk. vv; cf. nune for num-c(e), ch. ix. § 10], e.g. nonne
haec ita sunt ?, ‘is not this the case?’ ; num haec ita sunt ?, ‘now
is this the case?’ (with emphasis on the word ‘is’).

-Ne is probably I.-Eur. *né (Zend -na, appended to Interro-
gatives, e.g. kas-ni, ‘who then?’; cf. O. H. G. na weist tu na,
606 THE LATIN LANGUAGE. [Chap. X.

‘nescisne ?’), though it might also represent I.-Eur. *nii (ch, iii.
§ 37) (O. Ind. ni, in the phrase: katha na, ‘how then?’; cf.
Hom. ri vd pot pjxiora yévytat;). On its reduction by Syncope
to -n, e.g. vidén, audin, see ch, iii. § 36.

Utrum is the adverbial Ace, Sg. Neut. of wer, like Gk. wérepov.

An (Goth. an; cf. Gk. av, in that case) belongs to the I.-Eur.
pronominal root seen in Lith. afis, ‘that,’ O. Sl. ont, &e.

Car (O. Lat. quér) is I.-Eur. *q*or [Lith. ku¥, ‘where,’ for
*kar (I. F ii. 420); ef. O. H. G. hwar, ‘where ?’, from I.-Eur,
*q%ér, and O. Ind. kér-hi, ‘when ?’, from I.-Eur, *q*ér], with
a change of 6 to @ in a monosyllable before final -r (ch. iv. § 16)
that has a parallel in fa from *for (Gk. ¢odp). It may also
represent I.-Eur. *qou- (Gk. rod, where ?) with the suffix -r. (On
O. Lat. 6, class. @ for the L-Eur. diphthong ou, see ch. iv. § 41.)

Quaré, which must not be connected with cur, since the
length of the final vowel makes the idea of Syncope impossible
(calcaré becomes calcar, but avaré could not become *avar), is the
Ablative of Cause, just as cui rei Plaut. Truc. 394 (quoi rei te
adsimulare retulit?) is the Dative of Purpose, of the word-group
quae res? (cf, quamobrem ?). Plautus uses cus or quamobrem, quos
vet, but perhaps not guare. (Epid. 597 quare filiam Credidisti
nostram? is bracketed by Goetz.)

Quitinam, in O. Lat. poetry (Fest. 340. 25 Th.; ‘Servius’ ad
Virg. A. x. 6) (not in Comedy, so not colloquial ; Langen, Beitr.
p- 326), and adopted as an archaism occasionally by Virgil
(A. v. 13 and x. 6), is the adverbial (I-stem) Acc. Pl. Neut. of
quis-nam (ch. vii. § 28), as guid-nam is the adverbial Acc. Sg. in
such a phrase of Plautus as: quid tu, malum, nam me retrahis?,
‘plague on you, why do you pull me back ?’”

§ 11. (9) Comparative —Ut, uti, quasi, ceu, quam. Uz, of
which the final short vowel is preserved in #t-nam, ne-utt-quam,
[pronouneed n(e)wtiguam with first, as well as second, syllable
short], #t-gue, is one of those Relative Particles that appear in
Latin with initial ~, but in the Umbro-Oscan dialects with an
initial p-, which is their usual equivalent for an I.-Eur. labiovelar
guttural (see ch. iv. § 135). The Umbro-Oscan forms are dis-
cussed in ch. ix. § 10. 9.
§ 11.] CONJUNCTIONS. 607

Uti is in O. Lat. utei (e.g. on the S. C. Bacch. of 186 3.¢.,
C. I. L. i. 196).

Quast. It is difficult to derive quasi from quam si, though
the two expressions were undoubtedly equivalents: e. g. in the
Republican Laws guasei is the usual form, as in the Bantine
tablet of 133-113 B.c., C. I. L. i. 197. 1. 12, in the Lex Repetun-
darum of 123-122 B.¢., i. 198. 1. 41 (ef. 1.73), but quansei occurs
in the Lex Agraria of 111 B.¢., i. 200. 1. 27; and in Plautus we
have (1) quam st in the sense of the usual gudést in Poen. 241
item ... quam si, (2) gust in the sense of ‘ than if, (guam si) in
Mil. 482, &c. (see Brix’s note on Trin. 265). For the first
syllable of guasi was so thoroughly short that it acted as a
‘ Brevis Brevians’ and made the normal quantity of the final
vowel of -si short as early as the time of Plautus, whereas the
combination zs (ms) properly lengthens a preceding vowel in Latin
(ch. ii. § 144). We are thus driven to suppose that quam (adverbial
Ace. Sg. Fem.) and gud (adverbial Acc. Pl. Neut.) were two
equivalent Conjunctions which were joined with s¢ to denote
(1) as, (2) than if, and that the classical usage made a differentia-
tion of them, assigning the sense of ‘as’ to the combination
qua-si and the sense of ‘ than if’ to the combination guam-si.

Ceu, which is restricted to the Epic and Lyric Poets and a few
Silver Age prose writers (the elder Pliny, &c.), is compounded of
the Pronominal-stem *ko- (ke-), ‘this’ (ch. vii. § 15) and the
particle *wé, ‘as, like’ (§ 4). It has been supposed that as sew
is a syncopated form of sive [*sei-w(8), § 4], cew must come from
a fuller form *kei-w(é), the Locative Case of the Pronoun with
the particle *wé. But of this fuller form there is no trace, so
it is better explained as *cé-ve (cf. cé-tert) (ch. iv. § 33).

Quam is the Acc. Sg. Fem. of the Relative, as tam of the
Demonstrative (ch. ix. § 10). The two words are combined in ¢an-
quam. Its Oscan equivalent is pan (Zvet. LI.1.231.6 mais egm[as
tovti|cas amnud pan pieisum brateis ‘ magis rei publicae causa quam
cujuspiam gratiae’) or pam (ib. 231. 16 pruter pam medicatinom.
didest * priusquam judicationem dabit’). In O. Lat. we havea form
augmented by -d2, guamde, e.g. Liv. Andr. ap. Fest. 532. 8 Th. :

naémque nillum péius mécerat huménum
quémde mare saéuom uis et cui sunt magnae,
608 THE LATIN LANGUAGE. (Chap. X.

(a translation of Hom. Od. viii. 139). This guamde, quande
(Umbr. pane) would probably become *guan (Osc. pan ?), as deinde
became dein (ch, ix. § 10. 6), and would be merged in quam.

§ 12. (10) Temporal.—Quum, quando, dum, donee, ut, ubi.
Quum, O. Lat. guom, is most simply explained as an Adverbial
Ace. Sg. Neut. of the Relative, an I.-Eur. *q*om. Terentius
Scaurus (28. 9 K.) mentions an old form of the word (MSS.
cuine, for which editors read guomne or cume),and quotes a couplet,
from the Carmen Saliare; but unfortunately the passage in the
MSS. is so corrupt that little certain has hitherto been made of
the lines (see ch. viii. § 73). The Umbrian word poune, pone and
the Oscan ptin, pon seem to be compounded of I.-Eur. *q*om
and a particle -dé [cf. O. Lat. guamde for quam (Umbr. pane,
Osc. pan) § 11; and see below on guan-do].

Quandé seems to be the Acc. Sg. Fem. of the Relative with
the I.-Eur. Preposition *do, ‘to,’ or with some form of the
particle -de of O. Lat. guamde, than (§ 11). The Faliscan form
cuando (cu- or ev-) (Zvet. J. [. J. 70 cuando datu) affords presump-
tion that the word did not end originally in -d, so that -do would
not be an Abl. The ‘ Preposition’ -do, ‘ to’ (ch. ix. § 27) would
givea suitable sense, ‘ to what’ (sc. time), for in Plautus the Con-
junction is mainly temporal, though in Terence it is mainly
causal, as guandoquidem is at all periods of the literature (see
Studemund’s Studien, i. pp. 85 sqq.). Varro mentions its use
for guum, ‘ when,’ as a feature of the dialects of Formiae and
Fundi (ap. Charis. 111. 23 K.). In quandone (C. I. L. vi. 25048
nequa ei loci controversia quandone fieret ; 25905 con qua reli-
quias meas quandone poni volo) we have the suffix further
augmented by the particle -ne (cf. dd-nt-cum and see ch, ix.
§ 10; O. SI. ki-da-no ‘ quando,’ beside kti-da ‘ quando’ shows
another form of the N-suffix), and in guandd-qué, whenever, by the
particle -gué, ‘ever’ (§ 2). Quandoc, quoted from the XII Tab.
by Festus 346. 3 Th. (cf. Paul. Fest. 345. 4, 7 Th.) seems to be
a syncopated form of guandogue as née of néqué. The scansion
quanddquidem, found as early as Plaut. (Trix. 991 sdluos quando-
quidem dduenis; some would scan guanddc’dem) seems to be like
stquidem (see ch. ili. § 51).
§ 12.] CONJUNCTIONS. 609

Dum, which is often a mere asseverative particle, e.g. dgé dum
(Gk. dye 87), guidum, how so? primumdum, first of all, is an Acc.
Sg. Neut. from the Pronominal-stem *do-, as ¢wm from *to-, guam
from *q%o- (Gk. 8: is another case-form of the same stem), The
phonetic laws of Latin hardly allow us to connect it with dia,
diés, which come from the root dyew-, diw- ; but the first part of
du-dum, often referred to diw and dies, may come from a stem
*du-, a byform of *do-, whence Gk. djv for *3Fav, O. Sl. davé,
‘olim’ (I. F. ii. 250). Similar parallel stems were *no- and *nu- ;
and as Latin wm may stand for *no-m from the first or *nu-m
from the second, so Lat. dum may stand for *do-m or *du-m.
The temporal sense is clearly seen in the particle -dum in non-dum,
etiam-dum, inter-dum, &c. On late plebeian inscriptions we find
a form dunce, while, e.g. C.L. L. vi. 25063:

ad tu ne propera simili qui sorte teneris,

dune annos titulo nomina ut ipse legas ;
apparently an extension of dum by the particle -c(e), on the
type of nune, tunc (ch. ix. § 10. 7). The connexion of the two
meanings ‘while’ and ‘until’ is seen in archaic Engl., e.g.
Macbeth, iii. 1. 143 while then, God be with you.

Donéc must be considered in connexion with the byforms
donicum and doniqué. Dénicum is mentioned as an O. Lat. form
by Charisius (197.15 K.), who quotes Livius Andronicus :

ibi mdnens sedéto dénicum uidébis

mé carpénto uehénte medém domum uenisse,
as well as Plautus and Cato. Itis naturally resolved into *do-ne
(the Preposition *do, ch. ix. § 27, and the affix -ne, § 1), and cum
the temporal Adverb, ‘to when,’ ‘ till when’ (cf. Umbr. ar-ni-po
and its Latin equivalent guo-a7. Some explain donec as a form
of donicum with the last syllable dropped, but the loss of final
-um in Latin is confined within strict limits [on wihil(um),
no(e)n-(wm), see ch. iii. § 52], and would hardly be allowed in done-
cum. Donec is more naturally explained as the syncopated form
of doni-que (cf. nec and neque, § 18), and doni-que as *done aug-
mented by the particle -gue, ‘ ever’ (cf. quando, when, quandoque,
whenever; also dé-nigue); but the weak point in this account
is that donique is not found till Lucretius, and so is later than
donec. Perhaps the true explanation is that donzcwm was appre-

Rr

:
610 THE LATIN LANGUAGE. [Chap. X.

hended as donee cum, and so with omission of cum became donec,
while Lucretius coined a donigue on the analogy of nec and neque.
Whether Fr. done and the cognate Romance words come from
donec is doubtful [see Kérting, Lat.-rom. Worterb. s.v.; donee in
Petronius 40 (see Friedl. ad Zoc.) and 55 need not be translated
‘ then.’]

Ut (see § 9). Ubi (see ch. ix. § 10. 1).

§ 13. (11) Causal.—Quum, quoniam, quod, quia, quippe.

Quum (see ch. ix. § 10. 7).

Quoniam is a compound of guwm (quom) and jam, the 9 (y)
becoming vocalic by the law of Latin phonetics in the middle of
a word (so I.-Eur. *médhyos, O. Ind. madhyas, Gk. péo(c)os, &e.,
became medius in Latin, ch.iv.§ 63). Its oldest sense is temporal
‘when now’ (with Pres Ind., the Pres. tense being required by
the jam), e.g. Plaut. Trin. 112:

quoniam hine iturust ipsus in Seletciam,

mihi cénmendauit uirginem ;
and it is possible to trace its gradual development from a tem-
poral to a causal sense in the course of Latin Literature (see
Luebbert, Gramm. Stud. ii.).

Quod. T.-Eur, *q%6d (Lith. kad, ‘that,’ after verba declarandi,
&c., also used in the sense of ‘if’) is the Ace. Sg. Neut. of the
Relative O-stem, used like Homeric & in such a line as Od.
1. 382:

Tnréuaxov Oatvpatoy 5 Oapaadréws dydpeve.
In Plautus it is always, or almost always, subject or object of
a relative sentence, e.g. Capt. 586 :

filium tuum quéd redimere se ait, id ne utiquam mihi placet;

from the second cent. A.D, it is used with verba declarandi, &c.
e.g. Apul. Met. x. 7 asserere incipit quod se vocasset.

Quié is an Acc. Plur. Neut. of the Relative I-stem (Slov.
a, ‘if, Bulg. ti, ‘that, because’), and has the same double mean-
ing as guod, (1) that, (2) because. With appended -nam it had
interrogative meaning ‘ why ?’ (see § 10) like gued ? quidnam ?

Quippé (see § 7). (On guatenus see ch. ix. § 4.)

§ 14. (12) Conditional —Si, nisi, ni, sin, sive, seu, modo,
dummodo. 8, O, Lat. se, isa Loc. Se. of the Pronoun *so-, seen
§§ 18, 14.] CONJUNCTIONS. 611

in Lat. ip-se, ip-sa (ch. vii. § 20), as Lith. jéi, ‘if, of the Pronoun
seen in Lat. is, ea (ch.vii, § 19). Greek ci has been connected by some
with the Latin, by others with the Lith. conjunction. Sic, so, is
the same word with the enclitic -e(e) appended (ch. vii. § 15). Ose.
svai, Umbr. sve come from a stem *swo- (whence Goth sva,‘so’),
of which O. Lat. suad ‘sic,’ quoted from an augural prayer by
Festus (526.15 Th. suad ted ‘sic te’) is an Abl. Sg. Fem., as
the Umbro-Oscan forms are Loc. Sg. Fem.; but Volscian se
shows the same stem as Latin. The stems *so-, *swo- were no
doubt originally connected like the two stems of the second Sg.
Personal Pronoun *twe- and *te- (ch. vii. § 3). (On sw- see
ch. iv. § 68,) 3

Nist is a compound of the negative né (§ 18) with si, ‘ not if’
It is probably this word which is quoted in the form esi (ch. ix.
§ 51) in a defective passage of Festus, who explains it as sine
(166. 26 Th.) ; it is spelt nsec on the S. C. Bacch. and the Lex
Repetundarum, and w7se (with ¢ for the ei-diphthong, ch. iv. § 34)
on the Lex Rubria. The change of é to 7 in the first syllable is
due to the unaccented use of the word (as sine for sé-ne, mihi for
mehei) (ch, iti. § 18).

The Umbro-Oscan forms have *swai instead of *sei (Lat. s7),
as their second component, the Negative being represented in
Oscan by the Loc. Sg. form *nei (Ose. zee svae), in Umbrian by
a form no (Umbr. zosve), In that very old Latin inscription,
known as the Dvenos inscription, we find the Negative in another
Loc. form noi (is this a mere graphic variety of nez ?), if noisi
is rightly interpreted ‘unless.’ With the first part of mst we
may compare Ose. ne pon ‘nisi quum’ (Zvet. 1. I. 1. 231. 14 tzie
comono ni hipid ne pon, &c. ‘ is comitia ne habuerit nisi quum,’ &c.).

Ni, I -Eur. *nei, perhaps *né with the deictic particle -i of Gk.
otroo-t, &e. (ch. vii. § 23), had originally the sense of nén or né, as
in guid-ni, quippi-ni (Lith. nei, ‘not at all’; Ose. svwe pis cen-
stomen nei cebnust ‘si quis in censum non venerit, nei-p mais
pomtis com preivatud actud ‘neve magis (quam) quinquies cum
privato agito,’ ez svae ‘ nisi,’ svai nei-p ‘si non’ ‘si nec’), and
still retains this sense in some passages of O. Lat., e.g. Cato:
caueto ni quam materiem doles, and in Virgil’s line: ni teneant
It came however to acquire the sense of msi from its

RYr2

cursus,
612 THE LATIN LANGUAGE. [Chap. X.

use in such phrases as si in ius uocat, ni it, XII Tab., ‘if he
summons him (and) he does not go,’ id ni fit, pignus dato Plaut.,
‘lay me a wager in the event of that not happening,’ lit. ‘ that
does not happen, lay mea wager.’ (See O. Brugmann, Gebrauch
des Cond. ‘ Ni,’ 1887.)

Sim is usually said to represent si-né, ‘if not,’ with the same
syncope of the negative particle as in quiz, lest (§ 16), or as of the
interrogative particle in audin for audis-ne, &c. (§ 10); and this
explanation exactly suits its use in sentences like Cic. Lpp. Kamm.
Xil. 6. 2 qui si consecutus erit, vicimus ; sin —, quod di omen aver-
tant, omnis omnium cursus est ad vos. But this negative sense
of sin, ‘if not,’ is hardly attached to the word in the time of
Plautus. The Plautine use of six has been explained by the
O. Lat. habit of attaching interrogative -ze to the first word of
the relative clause, instead of to the first word of the main clause
[e.g. Ter. Phorm. 923 quodne ego discripsi? instead of: quod
ego discripsi, illudne rescribam ?], so that, e.g. Plaut. Zrin. 309
si animus hominem pepulit, actumst ... sin ipse animum pepulit,
uiuit, might more properly be written: sin ipse animum pepulit?
uiuit. It is however unnecessary to regard the x of siz as either
the Interrogative or the Negative Particle -ze; for it may be
merely that Demonstrative suffix -ne seen in alioguin, &e. (§ 16)
[Quin (see § 16) represents (1) gué with Negative -ne, (2) qué
with Interrogative -ne, (3) gué with Demonstrative -ne]. The
older spelling seé~ occurs, for example, on an epitaph, much
affected by the Romans! (Not. Scav. 1887, p. 180) :

mortua heic ego sum, et sum cinis, is cinis terrast ;
sein est terra dea, ego sum dea, mortua non sum.

Sive and sew (see ch, iv. § 33).

Modo is the adverbial Abl. (Instr.?) Sg. of médus, measure,
limit (cf. Hor. quis desiderio sit pudor aut modus ?), ‘ only’;
dum-modo, ‘ while only. A common sense of the word is the
temporal sense, ‘only a little while ago’ (cf. Caper 96. 15 K.
‘modo’ praeteriti est temporis, et ideo dicendum ‘ modo scripsi,’

 

1 This is a translation of Epichar- xémpos éoriy, :
mus’ epigram (ap. Schol. Hom. Ji. x. el 5€ re yf vexpds éor’, od vexpés,
144): ddA Oeds,

eiul vexpds, vexpds 5 kémpos, vq 8 h
§§ 15, 16.] CONJUNCTIONS. 613

‘modo feci’ non ‘ modo seribo,’ ‘ modo facio’; quamvis quidam
veteres et praesentis putaverint), emphasized in the Praenestine
dialect by the addition of éam (tam modo, inquit Praenestinus,
Plaut. Trin, 609). The shortening of the final -o is due to the
influence of the preceding short syllable (see ch. iii. § 42 on the
Law of Breves Breviantes), like did (originally *datod) Plaut.,
hive, béné, mitlé, &e.

§ 15. (13) Concessive.— Etsi, quamquam, quamvis, licet.
The formation of all these words is evident: e/-st, ‘ even if,
quam-quam (reduplicated), guam-vis, “how you wish’ (like quan-
tum-vis, ‘however much you wish’ or guam-libet, ‘how you please’),
licet, ‘it is allowed,’ ‘ granted.’

§16. (14) Final.—Ut, quo, quominus, quin, ne, neve,
neu, nedum. U# (see § 11).

Quo is the Abl. (Instr. ?) Sg. Neut. of the Relative, used with
Comparatives, quo facilius like co facilius. Qudminiis adds to quo
the Comparative minis, ‘less,’ used in a negative sense (cf.
minimé, ‘by no means’ ; pérum sciens, ‘ignorant’).

Quin is composed of gui, how (Abl.? Loc.? Instr.?), and the
negative particle né (§ 18), and is found with -e (or perhaps xé,
‘lest’) in unsyncopated form in Ter. Andr. 334: efficite qui
detir tibi; Ego id agam mihi qui ne detur, and in a fragment
from some comedy (Com. inc. 47 R.):

haud facile est defénsu qui ne cémburantur préxumae..

In some instances it represents gui (Nom. Sg. Masc.) with -ze,
e.g. nemo fuit quin sciret (qui nesciret); and a construction like
nulla mulier fuit quin sciret, nil tam difficile est quin exquiri
possit, is best explained as a universalizing of gui Masc., as of
potis Mase. in potis est (ch. ix.§ 2), though some regard the gui of
this usage as the Adv., and compare it to that Mod. Gk. use of
nod rév for 6v mentioned in ch. vii. § 23 (J. F. iv. 226).

It is used also in other ways than as a Final Conjunction
(the manifold uses of guim were a favourite theme of Latin
grammarians; see Gell. xvii. 13). Qué in affirmations, e.g.
hercle quin recte dicis, Plaut., may be merely the Adverb gui of
614 THE LATIN LANGUAGE. [Chap. X.

hercle qui, &c. with the Demonstrative suffix -ne (so atguin and
atqui, alioquin and aliogui, ceteroquin and ceterogut ; see on these
forms Georges, Lex. Wortf. s. vv.) ; quin in commands, originally
with Ind. (and so usually in Plautus), e. g. quin dicis?, then by
‘constructio ad sensum’ with Imper. (so usually in Ter.), e.g. quin
dic, is the Adverb gui with the Interrogative particle -ne, ‘how
not?’ ‘why not?’; gwen in a sentence like Plaut. Zvi. 360: quin
comedit quéd fuit, quod nén fuit? (i.e. eumne dicis qui, &c.), is
gui Nom. Sg. with the same particle (cf. Mi. 13 quemne ego
seruaui? ‘you mean the man whose life I saved?’ (see above,
§ 14 on sin).

Né is 1.-Eur. *né, ‘not’ (O. Ind. na, O. Iv. ni), a variety of
I -Eur. *né, ‘not’ (Lat. 2é-, § 18). In O. Lat. 22 (I.-Eur. *nei,
Osc. vei) was used in the sense of xé ($14). In Umbro-Oscan
i corresponds to I.-Eur. é (Lat. ¢), so that Osc. m7 in uz hipid ‘ne
habuerit,’ 22 fuid ‘ne fuerit, Marruc. vi in nita[g]a ‘ne tangat’
exactly correspond to Lat. né.

Névé adds to né the enclitic -vé, or, which in new is reduced by
syncope (cf. sive and sew, § 14).

For nedum (especially used by Livy, also by Cicero, but rarely
by the other authors) ze alone is occasionally found (Journ. Phat.
xx. 177). An early instance of the word, which is not employed
by Plautus, is Ter. Heaut. 454:

satrapa si siet
amidtor, numquam stfferre eius sumptis queat ;
nedum tu possis,
lit. ‘ne(dum) tu te posse eredas dico satrapam non posse,’ ‘ satrapa
non potest, nondum tu potes,’ with which we may compare Plaut.
Amph. 330:

uix incedo indnis, ne ire pésse cum onere existumes.

Nedum is related to ne, as viadum to vix, nondum to non; cf. Liv.
xxiv. 4. I puerum vixdum libertatem, nedum dominationem modice
laturum. On the construction and use of the word, see Harv.
Stud. ii. pp. 103 sqq.

$17 (15), Asseverative Particles.—Ne (nae), -ne. Né is the
spelling indicated by the references to the word in the Roman
grammarians, for they speak of it as the same in form with pro-
* §§ 17, 18.) CONJUNCTIONS. 615

hibitive we (Charis. 189. 2 K.; Diom. 394. 21 K.), and is also
the spelling of the best MSS. (Georges, Lew. Wortf. s.v.), though
there is no reason why there should not have been in Latin two
separate affirmative particles é and nae, as there were in Greek
vn and vai (cf. 8 and dai), representing an original *né and *nai,
a Loc. Sg. Form (ef. Ose. svaf, ‘if, $14). The form *né cor-
responds to *n& of the affirmative suffix -né found in the
Dramatists with Personal arid Demonstrative Pronouns; cf.
O. Scand. pér-na ‘ tibimet’ (e.g. Plaut. Mii. 565:
égone si post hunc diem

muttiuero, etiam quéd egomet certé sciam,

dato éxcruciandum mé;
for other examples, see Amer. Journ. Phil. ii. 51), as *wé to *wé, ‘or,’
*dé to *dé, &e. (§ 1).

§ 18. (16) Negatives.—In-, ne-, nec, non, haud, ve-. Né-
(I.-Eur. *né, O. Ind. né, &c.) is prefixed, not only to Verbs, e.g.
ne-scio, O. Lat. ne-vis, ne-parcunt, &e. (cf. O. Engl. nille, nolde ;
O. SI. nésmi, &c.), but also to other parts of speech, e.g. xé-fus,
u(e)-utequam, neiiter (né-, I.-Eur. *né, O. Ind. na, &c., appears in
néquiguam, &e.) ; in- (I.-Eur. *n, O. Ind. an-, a-, Gk. dv-, a-, &c.)
and vé- (I.-Eur. *we, O. Ind. va-; cf. O. SI. u-bogii, ‘ poor’),
a curtailment of I.-Eur. *aiwé-, O. Ind. ava-, ch. ix. § 12) only to
Adjectives, &c. (but see Langen, Beitr. p. 181 on improbare,
inf iteri, ignoscere, &c.). I.-Eur, *n- (Lat. iv-, older en-, ch. iv. § 81),
the weak or unaccented grade of I.-Eur. *né (ch. iv. § 51), is re-
presented in Umbro-Oscan by an-, e. g. Umbr. antakres ‘integris,’
Ose. amprufid ‘improbe.’

Non is generally supposed to represent *zoen(um), the ‘ doublet ’
of noenum (usually explained as *né-onom, ‘not one’) before an

‘initial vowel, as zzAil, nid was the similar doublet of nzhilum
(ch. 11. § 52); though the absence of a satisfactory parallel for the
change of oe to 6 (instead of the usual #) has led many to see in
the first part of the word some other form of the Negative stem
(cf. Umbr. no-sve ‘nisi,’ § 14), leaving the final x to be
explained as the particle (negative or demonstrative, § 1) *-né
(or *-nii ; ef. O. Ind. na-ni, Hom. ot vv, and see ch. iii. § 37).

Noenii (e.g. Luer, iii. 199 noenu potest) should be written
616 THE LATIN LANGUAGE. [Chap. X.

*oenus, and represents *ne-unusas noenum,*ne-unum (cf. démusand
démum, ch. ix. § 2) with suppression of -s in pronunciation (ch. 1.
§ 126). The other theory makes it differ from moenwm in
appending *ntt instead of num (§ 10).

Haud, which is confined within narrower limits than non in
O. Lat., being used especially with Adjectives and Adverbs,
usually immediately before the negated word, and never in
questions, has been referred (along with Gk. ov) to J.-Eur. *iwé-,
‘away’ (see above on vé-), so that it would properly be spelt *aad.
The Roman grammarians preferred the spelling laud to haut,
e.g. Charis. 112. 8 K. haud...d littera terminatur. od enim
Graeca vox d littera terminari apud antiquos coepit), and were
probably right in doing so; for the byform aw seems to be
the ‘doublet’ (ch. ii. § 136) before an initial consonant [Mav.
Vict. 15. 25 K. cum (sequens) verbum a consonanti incipit,
d perdit, ut ‘hau dudum’ et ‘hau multum’ et ‘hau placitura
refer ’], eg. C.I.L. i. 1007 heie ést sepuleram hau pilerum
pulerai féminae. (On the spellings laud, haut, hau, see Georges,
Lex, Wortf.s.v.) The initial 2- must have been used as a dis-
tinguishing mark to differentiate the word from aut (cf. Prob.
Inst. Art. 145. 9 K. ‘aut’ si sine aspiratione scribatur et in
t litteram exeat, erit conjunctio; si vero ‘haud’ cum aspiratione
scribatur et in d litteram exeat, erit adverbium).

Néc in O. Lat. has the sense of non! [Festus 162. 14 Th. quotes
XII Tab.: ast ei custos nec escit, and Plautus ‘in Phasmate’
(Most. 240) nec recte si illi dixeris], like Osc. neip (svai neip
dadit ‘si nec dedat’), Umbr. neip (sve neip portust issoc pusei
subra screhto est ‘sinec portarit ita uti supra scriptum est’). The
g of nég-dtium, neg-légo (often spelt eclegoin MSS.; see Georges,
Lea. Wortf. s.v.), &e. is variously explained as a phonetic change
of the -c (*q38) of nec (see ch. ii.§ 73), or as the particle *gé (Gk.
ye) of *né-g(é) (cf. Lith, né-gi, ne-gu), a different formation
from *né-q3(8).

§ 19. INTERJECTIONS. Interjections, being for the most
part onomatopoetic words, do not come under the phonetic laws

 

1 Catullus (Ixiv. 83) uses the phrase funera nec funera to express the Greek
rapo: drapor.
§ 19.) INTERJECTIONS. 617

of a language ; their analysis and etymology offer little difficulty.
The Latin interjections need not therefore detain us long. Oh!
ah! st! axe more or less the same sounds that we ourselves use
to express astonishment and surprise, and to enforce silence ; and
, they require no discussion. Many are borrowed from the Greek,
especially the exclamations used at musical or other entertain-
ments, e. g. ewge [in the Dramatists eugé (eugae) with a lengthen-
ing of the final syllable?! like our ‘ bravo,’ ‘ hallo |, sdphos, pilin, as
ours come from the Italian or French, e. g. bravo, da capo, encore.
But some are peculiarly Latin and offer points of interest. Ln
(not to be confused with em, an Interjection of terror, grief,
&e.), which is used by the Republican Comedians, where én (Gk.
jv) is used by the classical writers (in the Comedians éx is used
only in rhetorical questions, e.g. enunguam ?, Plaut. Men. 142,
925), seems to be the Imperative of &mo, lit. ‘take, a sense
which suits well in phrases like em tibi, ‘take that!’ ‘there’s
for you!’ (in giving a blow), e.g. Plaut. dsin. 431 em ergo
hoe tibi. Others make it Adverbial Acc. of is, O. Lat. em,
‘tum’ (Paul. Fest. 53. 37 Th.), which is also, perhaps properly,
spelt am (ch. vil. § 19). Joined with ie (in the Acc. Case) it
produces e//wm [ello (with open e) is still heard in the Abruzzi],
ellos, &c. So ecce (O. SI. ese and se) from the Pronominal stem
*eke- (*eko-) (ch. vii.§ 15) either with appended -ce ; [cf. Ose. eko-,
‘this,’ usually with appended -k (Lat. -ce), ekak ‘hac,’ ekkum
‘item’], or else with doubling of consonant (as in af¢-a/) produces
eccillum, eccillos, &c., whence the Romance forms, Fr. celui, Ital.
quello, &c. (see ch. vii. §.15). Hecwm has been explained as ecce
*hum (the enclitic -ce not being appended to the Pronoun because
it exists already in the Interjection, just as *ecceillwac, &c. is
never found); and this analysis is preferred to ecce ewm, because
the word is used by the Dramatists only when the person referred
to is present on the stage, whereas zs is the Pronoun used of
persons who have been recently mentioned. Still *ecce eum might
be explained as a parenthesis, e.g. Amphitruo eccum exit foras,
‘A. —see him— has come out.’ Accum is the original of the Italian

 

1 Heid, the usual scansion, as in heia, viri, nostrum reboans echo
this refrain of a Late Lat. boating- sonet heia '
song (Poet. Lat. Min. iii. p. 167 B.): appears as heid in Plaut. Merc. 998.
618 THE LATIN LANGUAGE.

interjection ecco, and in Plautus often comes very near ecce, e.g.
Mil. 25 ubi tu es? Eccum, Poen, 279 assum apud te eccum.

Pré (not proh, see Neue, 11%. p. 985) seems to be merely the
Preposition (Adverb) pré, forth, lit. ‘away with it!’

Vae, \.-Eur. *wai (Goth. vai, Lett. wai), borrowed in late
Greek, odai, is the same word as our Noun ‘ woe.’

Agé is, like em, an Imperative used interjectionally, in Plautus
and Terence often with the enclitic dum appended, agedum (like
Gk. dye 87, § 1). The interjectional use of .Imperatives is
a feature of all languages; our ‘lo’ is the Imperative of ‘to
look,’ and we have in modern Italian vie (for venz), tie or te (for
tene), guar (for guarda).

The names of deities occur in herele, me-hercules (sc. Juvet, Paul.
Fest. 90. 11 Th.), we-herele, me-castor, me-dius fidius (‘the god of
good faith, with «dius for deus because the first syllable is
unaccented ?), pol (a curtailment of Pollux), ecastor (better
eccastor, for the first syllable is long, but not long by nature,
since it is shortened by the Law of Breves Breviantes in the
Dramatists; see ch. iii. § 34). The last might represent e¢ Castor,
but éépol can hardly represent e¢ dews Pollux, although the
irregular forms assumed by many of the English interjections,
‘zounds,’ ‘sblood,’ ‘ marry ’ (for ‘ Mary ’), show us the difficulty
of tracing curtailed phrases of the kind back to their origin by
the ordinary methods. Zecéré is either an invocation of Ceres
or ecce re, ‘ lo indeed.’
INDEX?

—++—_

(The numbers refer to the pages; i and J, wand v are treated as identical.)

A, pronunciation, 13 sqq. ; phonetic
changes, 219 sqq. ; in weak grade
of O-root, 258 sq. ; of E-root, 258 sq.;
of E-root, 261; varying with 6, 259;
with 4, 259 sq. ; Lat. @ for I.-Eur.
@ (8), 221 sqq. ; foré, 222; forauin
Agustus, &c., 38, 41 sq. ; in ld, rd,
md, nd, 222; for 6, 234 sq.; Lat. a
for @ lengthened, 220; for 6, 220;
in ld, ra, nd, 219 sq.; not weak-
ened to @ 199; aa written for d,
Io.

-d, shortening of, 210 sq. ; in Nom.
Sg., 210 sq., 373; of gua, Adv.,
&e., 551.

A-Subjunctive, 512 sqq.

ut, ab, abs, Prep., 575 sqq.}; ab- con-
fused with ob-, 574.

abante, 573, 595:

abést, the scansion, 214 sq.

dbicio, 45.

abiegnus, pronunce. of, 138.

Ablative (see Declension), Adv. use
of (see Cases).

Ablaut (see Gradation).

abnuo and abnueo, 476.

abolevi and abolui, 500.

avscisio and abscissio, 112.

abscondi and abscondidi, 502.

absida, 79.

absinthium, 79.

absque, 576.

abstinei (?), 501.

Abstufung, 367.

abstulas, 464. ~

abusque, 595.

abjssus, the scansion, 156.

ac (see atque).

Acca Larentia, 118 1,

accédo, 194.

Accentuation, 148 sqq.; studied at

 

Rome, 151 sq.; Early Law, 157
sqq. ; I.-Eur., 157 Sqq., 165 sqq. ;
of facilis, &e., 158; of Sentence,
165 sqq.; Secondary, 158 sqq.,
161; Paenultima Law, 160 sqq.;
of Word-Groups, 161 sqq., 169 sq. ;
of -ds, 163 ; of addtic, &e., audit, &e.,
163 ; of illéc, &c., 163 ; of tantin, &c.,
163 ; with -gu(e), &c., 163; of Gen.,
Voc. Sg. of 1O-stems, 163 sq.; of
Interj., 164; Vulg. Lat., 164 sq.;
of -itrem, -tlum, 164; with Mute
and Liquid, 164; of Comp. Vb.,
164; of Numeral, 165; and Ictus
in Plaut., 166 sqq.; of sum, 167;
of Pron., 167 sq.; of Prep., 167
sqq-; of Adv., 168 sq.; of Conj.,
169 ; of Auxiliary, 169.

accentus, meaning of, 152, 154.
accepsti (?), 508.

acceptor for accipiter, 115.
accerso, 487.

Accheruns, Plaut., 58.
Acchilles, Plaut., 58.

accipiter, 259.

1 If a word is not found in this Index, the ending or suffix of the word
should be looked for. Thus the reference for furibundus, fremebundus, &e., will
be found under -bundus, the reference for commenticius, &c., under -ticius.
620

Accius, doubling of vow., 8 sqq.; 9g |

for ng, 10 sq. ; ei for 7, 9 sq.

acclinis, 275.

accubuo, 323.

accuratus, 541 n.

Accusative (see Declension),
use of (see Cases).

cer, 260, F. 371; -ris, M., 371.

dcerbus, 180.

acertas for acritas, 365.

acetum, 335.

-dceus (see Suffix -KO-).

Achivi, 196.

acies, acisculus, 347.

acredula, 353.

acrufolius, spelling, 364.

actito, pronune. of, 134; of actum, 139.

Active, endings (see Verb).

actus, Noun, 344.

actutum, 565.

acuo, 260

acupedium, 259.

-dcus (see Suffix -KO-).

Acute Accent, 153 sqq.

ad, Prep., 576; spelling of, 76 sq. ;
ay, 288, 99.

adagio, 291.

addues, 515; -it, 515. (See arduuitur.)

adeo, Adv., 568; accent. of, 166.

adeps, alipes, 287.

adessint (?), 466.

adgredimur, the scansion, 475.

adgretus, 285,

Adjectives, dist. of Gender (see
Gender) ; Decl. of (see Declension) ;
Compar. of (see Comparison) ; form
of (see Suffixes); Numeral (see
Numerals); Pronom. (see Pro-
nouns); used as Part., 540, 543;
from Part., 540 sqq.

adjuro (~jue-), Fut. Pft., 507.

adnitit, Perf., 508.

adolesco, 481.

adoritur, the seansion, 475.

adpetissis, 462.

adquo, 568,

adsum, pronounced ass-, 313.

advenat, 464.

adventicius, 337.

Adverbs, 548 sqq.; Nom. forms,
553 Sqq.; in -ter, 553 sqq.; Gen.
forms, 555; Acc. forms, 555 sqq.;

Adv.

 

Compar. of. 557 sq.; Abl., Instr.,

THE LATIN LANGUAGE.

Loe. forms, 559 sqq.; in -tus, 561 ;
in -tim, 548; in -e, 548; word-
groups, 562 sqq.; -mente, 552; of
doubtful origin, 565 sqq ; Num.
(see Numerals) ; Pronom., 567 sqq. ;
Compound, 360 sq.

adversus (-m), Prep., 595.

adulescens, spelling of, 197.

aduncus, 259.

adusque, 595.

ae, pronune. of, 37 sqq.; and e, 42,
242; for au, 42; for Gk. n, 42 sq.;
for a, 242 (see also AT).

Aecetiat, 188.

aedes, 241, 346.

aedilis, 340; aidiles, Nom. Sg. 376.

aeditumus (-twus), 405.

aegrotus, 484.

-aei- for -ai-, 242.

aemidus, 258.

Aenea, Nom., 373.

aequanimus (-itas), 123, 364.

aequipero, 192.

aequus, pronune. of, 42.

-aes in Gen. Sg., 381 sq.

aes, 157. :

Aesculapius, 242; spelling, 198.

actus, 173.

aevum, 241, 251, 348.

af, Prep., 576 sq.

affatim, 563.

afluo and affluo, 576.

age, 600, 618.

ager, 221,

agilis, 332.

Agma, Io sq., 60, 65.

agmen, 292.

agnomen, 294.

agnus, 235; F., 370.

ago, 221; egi, 497; actum, pronunce.,
139; age, 600, 618 ; axim, 465 sq.

agoed (?), the scansion, 373 ”.

agricola, 317.

Agrigentum, 197.

agulum, 334.

Agustus for Aug-, 38, 41 sq.

Ahala and Ala, 54.

Ahenobarbus, 364.

ahenus, 265; spelling of, 55,57. _

AI, phon. changes of, 241 sq. ; AI,
25I sq. ; ai on inscrr., 242.

-ai of Gen. Sg., 381 sq. ; Dat., 386.

aio, 546, 265; pronunce., 53 ; spelling
INDEX,

Wo, 8; ai, aie, Imper., 546 5 aibam,
«iebam, 49x ; pronune. of ai-, 43.

-al- from I.-Eur, 1, 279.

ala, 293.

alacer, pronune. of, 18 ; -ris, M., 371 ;
~ecer, 18, 198.

albeus for alv-, 51.

albico. 488.

albogalerus, 361, 364.

Albsi for Albensi, 177.

albus, 223.

alebris, 334.

Aleria, 197.

ales, 351.

Alexander, -ter, Alixentrom, Acc., 73.

Aljius, dial. for Alb-, 80.

alia, Adv., 569.

alias, Adv., 557.

alibi and aliubi, 567.

alica (hal-), 56.

alicunde, 570.

alienus, 449.

alio, Adv., 568.

alioqui (-n), 568, 614.

aliorsum, 549.

aliquamdiu, 571.

aliquando, 571.

aliquantisper, 562.

aliquis, 447.

aliquo, Adv., 568.

-Glis (see Suffix -LI-).

dliter, 554; cf. 553.

dlitus and altus, 335.

alium, all-, 115 ; -eum, 22. (See alum).

dliunde, 570.

alius, 449 ; -t8, 375.

aliuta, 571.

allexi, 505.

alnus, 309.

alo, 223.

Alphabet, 13qq.,5; Gk.letters, 4,11 sq.;
Claudius’ reforms (see Claudius).

alter, 449, 452; -tus, Gen., 450.

alteras, Adv., 557.

*alternas (?), Nom. Pl., 398.

alternis, Adv., 551.

alteruter, 450 ; -rtra, 450.

alucinor, 488; spelling of, 117.

alum, 333. (See alium.)

alumnus, 327.

-am of quam, tam, &e., 549.

am-, Prep., 578.

amarus, 259.

 

621

amasius, 305.

ambages, 345 sq., 221.

ambegna (-igna), 229.

ambi-, Prep., 577.

ambio, 505 ; -issit, 466,

ambo, 451.

ambulo, 547.

amendo (-ando), 200.

amicio, 505, 578.

amicus, 337.

amitto (amm-), 109, 114.

ammentum, spelling of, 114.

amnego for abn-, 80.

amnis, 282,

amnuo for abn-, 80.

amo, 274; -ento, 3 Pl, 519.

amoenus, 246.

ampendices, 578.

amplant, 483.

amplector 578; -oct-, 467.

amplio, 485.

amsegetes, 578.

amtermini, 578.

GMUTCHy 33, 75+

amussim, 563.

amygdala, -iddula, 198.

an, Conj., 606.

an-, Prep., 578.

Anaptyxis (see Parasitic Vowel).

anas, 274.

ancaesus (-7sus), 198.

anceps, older, -cipes, 178.

ancilia, 287.

anclabris, 334

ancora, 155, 190; -ch-, 59.

anculus, a servant, 178 ; -cilla, 333.

ancunulentae, 196,

ancus, 259.

angina, 326.

ango, 271; -ustus, 223, 356.

anguila (-illa), 115.

anguis, 338.

anhelo, 199, 578; alen-, 98; -ll-, 112,

animadverto, 362.

animus, -Ma, 223.

annus, 117, 314.

anquind, 247.

anquiro, 578.

anser, 272.

antae, 274.

ante, 578; ant(e)positus, 579; antec,
579, 569; antehac, 569 ; antid-, 579 ;
antidhac, 579.
622

antennae, 578.

antes, 578.

antestamino, 519, 578.

antiae, 562.

anticus, 337.

antideo, 575, 579:

antigerio, 560.

antioper, 562

antiquus, 337.

antistes, 350.

antruo, -dr, 289.

anuis, Gen., 384.

-dnus (see Suffix -NO-).

anus, 333-

Aorist (see Tense-Stems).

aper, 222.

aperio, 475.

Apex, over long vow., 4,129,134, 161 sq.

apinae, 58.

apio, Vb., 480.

apiscor, 480.

aplustrum, 96.

Apocope, 203 sqq. ; accent in, 153, 161.

appello, 4723 -amino, 519

Appenninus (Ape-), 117.

Appius (see Claudius’.

cpprime, 565.

apricus, ‘178.

Aprilis, 178.

aprugnus, -unus, 294.

apud, 579 3 -or, -ur, 288.

aqua, 223; trisyll. (?’, 87; 4G, acg-,
87; -ai, Gen., 382.

aquild, the scansion, 210.

Aquilonia, 286,

-ar- from I.-Eur. r, 279.

ar for ad, Prep., 288, 99.

ard, Usd, 305.

*aramen for aeramen, 201.

aranea, 292.

aratrum, 330.

arbiter, 288.

arbor, 290 ; -08, 356; -osem, Acc., 305.

arboretum (-bustum), 306, 235.

arbutum (-itum), 197.

arceo, 223.

arcesso, 487 ; -ivi, 506; «accerso, 487.

arcubii, 176.

arcus, 300; decl. of, 344; F., 344.

ardea, 279.

ardeo, 486.

ardus, 184.

arduuitur (2), xii Tab., 288, 515.

 

THE LATIN LANGUAGE.

are-(facto), 490 ; arf-, 184.
arena (see harena).
argentum, 296.

arger for agg-, 288.

Argiletum, accent of, 161 sq

argumentum, 336.

argutus, 484.

aries, 261 ; =jete, 144.

-Gris (see Suffix -RI-).

arispex, 29.

-Grius (see Suffix -IO-); ousted by
-dris, 321.

armus, 279.

aro, 223.

arquites, 300.

ars, 341.

artena, 172.

Article, Def., 452; Indef., 410

Articulation, Basis of, 30.

artio, 485.

artus, 343.

Aruncus for Aur-, 42, 40.

arvum, 323 3 -uUus, 323.

-ds (see Suffix -TI-) ; accent of, 163.

-as, Nom. Pl1., 398; Gen. Sg., 381.

aser, blood, 26r.

asinus, 305.

asom fero, 539.

asp- for absp-, 310.

asper, aspr-, 185.

aspergo (-argo), 200.

aspernor, 470, 486.

Aspirates, phon. changes of,279 sqq. ;
Gk. in Lat. orth., 4, 11 8q., 54, 57
sqq. 72, 99 sq.; Tenues (see
Tenues Asp.).

asporto, 210.

assentior (-0), 521.

Asseverative Particles, 614.

Assibilation (see Palatalization).

assidue (-0), 550.

Assimilation, of Cons., 311 sqq.; of
Prep., 312 sq.; of unacc. vow.,
201; of final cons. (see Sandhi).

-asso, Vb.-forms in, 462 sqq.

assulatim, 556.

ast, 600 sq.

asted, Dven. Inser., 514.

-aster, -ast(ryinus (see Suffix -TERO-,
Suffix -D-),

at, Conj., 600 ; spelling, 76 sq.

Atella, 312.

ater, 81.
INDEX.

Athematic, Conjug. (see Verb) ; Pres.
Part. Act, 54r.

-atim of Adv., 556.

atque (ac), 598 sq., 122; spelling,
599 ; atgue atque, 599.

atqui (-n), 601, 614.

atritus, 485

atrox, 259, 354.

attat, 617 ; accent of, 164.

attigas, 464.

attulas, 464.

-atus, e.g. dentatus, P.P.P., 483.

AU, phon. changes of, 242 sq.;
pronune., 37 sqq.; weak grade of
OU-root, 261; -au- for dvé, 243 ;
AU, 252.

au-, Prep., 576.

au, Interj., 38.

aububulcus, 235.

auceps, 180.

audacter (-iter), 554.

audeo, 486; -si, 522; -sus sum, 522;
-sim, 465 ; aussus, IT2.

audio, 307.

ave (see have).

Avernus, 197.

averruncassis, 462.

averta, 197.

Aufidus, 250.

augeo, 482, 243.

augur, -ger, 198 ; -ra, Accius, 48.

avillus, 235-

aula (olla), 41 ; aulla, 112.

Aulius, 267.

aureae, 261.

aurichalcum, 41.

auriga, 261.

auris, 243.

Aurora, 243, 356.

aurugo (-igo), 37-

aurum, 243.

aus- (os-) in osculor, &e., 41, 262.

aus for arus, 52.

ausculto, 243 5 «sc-, 41.

auspex, 180.

aussus, T12.

austerus, -ris, 338.

austium, 262.

cut, 599.

autem, 601, 571.

autor for -ct-, 89, TT9.

autumo, 180, 235-

avunculus, aunc-, 49, 172 5 amen, 172.

 

623
Auxiliary Vbs., 511; accent., 169.
auxilla, 333.

Avyayibhiva, Compd., 360 sq.

-dx (see Suffix in Gutt., -KO-.

-X, &. g. MUredx, 355.

axin, 465 sq.

axis. 305, 338.

axites (-tiosi), 352.

B, pronune., 78 sqq.; for Gk. p, 11;
for v, 47, 49 sqq.; phon. changes
of, 282; for dw-, 265, 268 ; for bh,
282 sq.; for dh, 289 sq. 3 br for sr,
303, 308.

baca, spelling of, 116 sq.

Bahuvrihi, Compd., 360 sq.

balbus, 282, 358.

balbutio, 488.

*baliolus (?), 287.

baliaena, 48, 58; spelling, 117.

balneum, balin-, 173.

-bam of Impft., 489 sqq.

barba, 283.

barbactum for verv-, 52.

barbar(us), 374.

barca, 184.

basilica, bass-, 115 ; basis, bass-, 115

Basis of Articulation, 30.

batillum (see vat-).

battuo, spelling of, 113.

beatitudo (-tas), 341.

Belena (?), 48.

bellum, O. Lat. duellum, 268.

bellus, 326.

bene, -%, 551 ; ben(e), 184 ; -merens, Adj .
540; -volens, 352; and -lus, 540;
-ficus for vené-, 51.

benignus, pronune. of, 138.

Benuentod, 184.

*berbix for rervex, 52.

bessi-, 409.

BH, phon. changes of, 282 sq.

bhé-, ‘ to speak,’ 457.

bher-, ‘to carry,’ 457.

bheu-, ‘ to be, 458.

-bi of ibi, &e., 551, 567-

bi- of bidens, &e., 411.

bibo, Vb., 468; -i, Perf., 503; -er(e),
Inf., 537.

bicorpor, 376.

bigae, 196.

-bilis (see Suffix -DHLO-).

bimus, 144, 294.
624 THE LATIN

bint, 411.

bipinnis for -pen-, 23.

bis, 411.

blasfémus, the scansion, 156.

blatta, 314.

-bo of Fut., 491 sqq.

bonus, 326 ; Compar., 406.

bos, 253; bobus, bu-, 250.

-br- for mr, 269 sqq.

-bra (see Suffix -DHRO-).

braca, spelling of, 116.

bracchium, 58 ; spelling of, 117.

Breathing, Gk. in Panhormus, &c., 57.

Breath-Stops (see Tenues).

Breves Breviantes, 210, 126, 129 sq.,
201 sq.

brevis, 227, 292

breviter, 553. ‘

Britanni (Britt-), 115.

Broken Reduplication, 358.

Bruges for Phryg-, 36, 58.

-brum (see Suffix -DHRO-). $

bruma, 407.

Bruttii (-ri-), 29.

-bs-, -bt-, pronune. of, 79.

bucetum, 335.

ducina, spelling of, 117.

bulba for vulva, 50.

-bulum (see Suffix -DHLO-).

-bundus, e. g. errabundus, 545.

Burrus for Pyrrh-, 36, 75.

burrus, 75.

-bus, Dat. Pl., seansion of, 404.

bustar (bo-), 205, 250.

bustar (cf bustum), 250.

butirum, the scansion, 156.

*butis (-tt), 116.

Butrio, 33.

buxus, 75.

C, the letter, 2, 6 sq.. 76; pronune.,
84 sqq.; palatalization of, 87 sq. ;
for qu, 299 sqq., 315; eo for tl,
283 sq.

cadaver, 541.

caduceus, 288,

caducus, 337.

cadui for cecidi, 509.

Caecilius, Cec-, 42; Caeic-, 242.

caecus, 242,

cuedes, 346.

caedo, 242 3 cecidi, 496 ; cedre, 184.

caelebs, 48,

 

LANGUAGE.

caeles, 352.

caelum, the spelling coe-, 44.

caementum, 285; -ta, F., 400.

caeruleus, 275.

calamitas, 286 ; kad-, 286,

calamitosus, 353.

calandae for cale-, 23.

calcar, 203.

caldus for -lid-, 173.

cale-(facio), 490 ; calf-, 173, 184.

calicare, 95.

caligo, 355.

callescerunt, Cato, 481.

callim (?), 580.

*calmus for -lam-, 198.

calor, N., 356.

calumnia, calvor, 327.

calvus, 323.

calx, heel, 355.

calz, lime, 95 ; -ls, 107.

Calypsinem, 155.

camellus for -élus, 115.

Camena, 308.

camera (-mar-), 197.

Camerina (-mar-), 197.

Camillus, 308.

cammarus (ga-), 74.

Campans for -nus, 182.

campester, 330.

cancer, 96.

canes, Nom. Sg., 346 ; -nés, Pl., 399.

canicula, 347.

cano, 223 ; -nte, Carm. Sal., 459; Perf.,
501, 509.

Canopus, 75.

cunus, 307.

capac, 355.

caper, 276.

caperro, spelling of, 117.

capesso, 462.

capiclum for -tulum, 83.

capio, 298; cepi, Perf., 502.

capis, bowl, 83.

capistrum, 331.

Capito, 349.

Caralis, -lar-, 93 3 -Tar-, 93.

Carda (-dea), 317.

Cardinal f (see Numerals),

cardus for -duus\ 174.

carictum (?) for -rec-, 23.

cdrint (?), 515.

carmen, 271 sqq.; C. Saliare, 5, 245,
459 ”.
INDEX.

Carna, 317.

cairo, 278, 273, 349.

carpatinae, 75.

canpo, 279.

cartilago, 279.

Carvilius Ruga, letter G, 7.

- CHUS, 491.

cuscus, 307.

Cases, 366 sqq. ; Strong and Weak,
367; suffix ousted by Prep., 573;
Adv. use of, 548 sqq.; of Nom.,
553 sqq.; of Gen., 555; of Acc.,
555 sqq.; of Abl., Instr., Loc., 559
sqq. (See Declension.)

cassis, -ida, 354 ; casila, 286.

cassus, empty, 565.

Castérem, 155.

cusus, -SS-, TIO sq.

Cato (Elder), -ae for -am, 61, 493 n.

catulio, 484.

catus, 258, 541.

cavaedium, 362,

cauculus for -calc-, 96.

cauda, 41.

caved, 235 ; cave, accent, 169 ; pronunc.,
49 ; cari, 499.

Cauneas (cave ne eas), 169.

Caurus, 258.

Causa, -Ss-, ILO Sq.

Causative Vbs., 477, 481 sq.

causis (cave sis), 49.

cavus (cov-), 234 sq.

-ce, Particle, 432 sq.

cédo, 432, 518; cette, 284.

cédo, Perf. cessi, pronunce., III.

cedre for caedere, 184.

cedrus, 289. _

celer, 351; -rissimus, 407.

-cello, 486.

celo, 488 n., 227.

celox, 354.

celsus, 229.

cena, 277; -atus, P.P.P., 520, 542;
-0@-) 44, 277.

-cendi, Perf., 501.

censeo, 273; -ento, Pass., 519.

centum, 418 sq. ; plex, 418 sq. ; -tussi-,
409; -centum for -ti, 418 sq. ; -tesi-
mus, -tensu-, 418 sq.

cerebrum, 296 ; -ber, 370.

ceresium, cherry, 18.

cerno, 472; crevi, 500.

certo (-€), 550

Ge

 

s

625

cervix (Sg., Plur.), 355.

Cerus, 329.

cesaries for cues-, 43.

cetero-, 2443; -rum, Oonj., 602; -ra,
Adv., 602 ; ~rogui (-n), 568, 614.

cette, 284.

-cetum, 335.

cen, 607.

-ch- (Gk. x), -cch-, Plaut., 58.

Change of unaccented vowel (see
Weakening).

Chersonensus, the spelling, 136 7.

Chi (see Aspirates).

Chius (Adj.), scansion of, 132.

-ci- and -ti-, 82 sqq.

eibus, decl. of, 344.

ctcindela, 333.

ciconia, Praen. conea, 22.

cicur, quant. of i, 485.

*ctcus for -icc-, 116,

cieo (cio), 481 ; citus, 335, 541.

cincinnus, 315.

cinctutus, 335.

cingo, -nxi, -nctus, pronune., 140.

cinis, 357.

-cinor (-cinium), 488.

circes, 352.

circo-,-um, Prep., 579 3 -4, 579 ; ~tter, 580.

Circumflex Accent, 153 sqq., 161.

cis (citra), 580, 432; -ter, 432.

cistella, 333.

citera for -thar-, 190.

citrus, 289.

citus, 541 ; ci-, 3353 -t0, Adv., 551.

civicus, 337.

-c- for -tl-, 81.

clades, 219.

clam (clanculum), 580; clamde, 580.

clamo, 279.

clandestinus, 580.

clango, 471.

clarare (-ere), 484.

Claudius (App.), reforms alph., 6,
105; (Emp.), reforms alph., 3 sqq.,
36, 47 $q., 79 (see also Cho-).

claudo, 180, 252; clu-, 40, 196.

clavis, 347.

clavus, 298.

clepo, 298 ; -psi, 505.

cliens (clu-), 29.

clipeus (clu-), 29.

clivus, 275, 323.

cloaca (clu-), 37:
626

Clodius, 41; Cla-, 42.

Cloelius, 250; Clowl[t}, 246.

Close Syll., quant. of vowel, 133 sqq.

clueo (-u0), 473, 295-

-clum (-culum, q. v.) (see Suffix -LO-).

clunis, 250.

Clutémestra, the scansion, 202.

co- for guo-, 300.

-co of albico, &e., 479.

cow (from coeo), 318.

codlesco, 481.

coculum, 300.

coelum for cael-, 44.

coemptionalis senex, 143.

coena for ce-, 44, 277.

coepio, 545 ; -pi, 502; scansion of, 143 ;
-plus sum, 522.

coero, cot-, for curo, 248.

coetus, 142, 39.

cogito, 143.

cognatus, spelling of, 114.

cognomen, 294.

cognosco, con-, 294; Perf., 509.

cogo, 143.

cohibeo, scansion of, 143.

cohors, 183.

co(h)um, 235.

cojicio, pronune. of, 53.

coinquo, 311.

colina (?), the spelling, 236.

Collective, Compd., 360 sq., 365;
Noun, 399.

collega, 318 ; -gius for -m, 370.

collis, 271.

collum, gender of, 369.

collum for -l-, 112.

cilo, 227; spelling of, 300.

colober for -lub-, 37.

coloephia for -ly-, 36.

colonia, 321.

columen (culm-), 185.

columna, pronune., 69 ; -lom-, 37, 69;
C. Rostrata, 7; -mella, 69.

colurnus, 97.

colus, 300.

com- (cum) (co-?), Prep., 580; bef. v-,
JF, 50 8q., 66, 99 sqq.; bef. n-, gn-,
114 ; bef. s-, f-, 136 sqq. ; quom, 581 ;

__cum bef. n-, 121; ¢. €0 c. quiqui, 448.

~ combretum, 227.

comburo, 578, 144.

comes, 350.

comis, 307.

 

THE LATIN LANGUAGE.

commendo (-man-), 200.

commentus, 335.

comminus, 5543 -m-, II5.

commircium, the spelling, 220.

communis, 247.

como, ~mpst, 505.

compages, 346.

Comparison, of Adj., 404 sqq.;
Compar., 404, 406; Superl., 405,
407; irreg., 407 sq.; of Adv., 550.

compellare, 472.

Compensation, length by, 314.

comperendinare, 486.

compesco, 192.

compitum, 194.

complere, in Romance, 489.

Compounds (Noun and Adj.), 358
sqq.; accent of, 161 sqq.; A-stems,
363 sq.; O-stems, 364; I-stems,
364; U-stems, 364; N-stems, 364 ;
R-stems, 365; Dent., Gutt. Stems,
365; S-stems, 365; in Pacuv., 360 ;
in Luer., 360; in Plaut., 362;
(Verbs), 362 sq.; accent of, 164;
influence on Simple Vb., 468;
(Adv.) (see Adverbs) ; (Prep.) (see
Prepositions).

concapit, xii Tab., 379.

concino, -ut, 509.

conditio, 341 ; -cio, 88.

coned, Praen., 22. —

conesto for cohon-, 143.

confestim, 556.

confeta (sus), 318.

confuto, 309.

conger (go-), 743 gu-, 33-

congius, 280,

congruens (-wus), 540.

conitor, bpelling of, 114.

coniveo, 302 ; -nixi, 499.

Conjugations (see Verb).

Conjurtctions, 596 sqq. ; accent, 169 ;
vari¢ty of meaning, 596; of stem,
596j of vow.-quant., 596; Con-
junctive, 598 sq. ; Disjunctive, 599
sq.; Adversative, 600 sqq. ; Limi-
tative, 602 sq.; Explanatory, 603
sq.; Conclusive, 604 sq ; Optative,
605; Interrog., 605 sq.; Compar.,
606 sqq. ; Temp., 608 sqq.; Causal,
610 ; Condit., 610 sqq.; Concessive,
613; Final, 613 sq. ; Assev., 614 sq. ;
Neg., 615 sq.
INDEX.

conjus (-nz), 69, 358.

conquaeisivei, 240,

conquiniscor, 470.

consacro for -sec-, 200.

conscribillo, 479, 487.

considero, 488.

consilium, 286 ; -sid-, 286.

consiptus, 195.

Consiva, 199.

Consonant, lost in group, 309 sqq. ;
Stems and I-stems, 338, 341 (see
Suffixes),

consternare, 470, I92.

consul, pronune. of, 136 sq.

contages, 346.

contamino, 292, 294.

conterere, in Romance, 489.

conticinium, 194.

continuo, 557. .

contio, 67 ; pronune., 141; covent-, 250;
-net-, 310.

contra, 581 ; -@, 557.

Contraction, e.g. cra for cera, 1977;
e. g. mg for magnus, 125; Contr. Vb.-
forms (see Verb).

controversia, 581.

contubernium, spelling of, 193.

contudit, 496.

contumelia factum itur, 538.

conubium, spelling of, 114.

conucella, 273.

convicium, 225.

conviva, 318.

convollo (-rell-), 228.

copia, 144 ; copis, Adj., 144.

copula, 143.

coquino, 470.

coquo, 467 ; -int(?), 515.

coquus, 291 ; spelling of, 299.

cor, 279; scansion, 122, 215.

corallium (curali-), 34.

coram, 581.

corbus for -vus, 51.

cordatus, 483.

corigia for corr-, 114.

Cornelis, 372.

cornicen, 192.

Corniscas, 404.

cormx, 347.

cornu, 279.

cornus, eornel, 279.

corolla, 333+

corona, 59; spelling of, 59.

 

627

corruptus, cor-, 114; -mpt-, 471.

Corus, 258.

0s, 259.

cosentiont, Scipio Epit., 520.

cosmis, Dvenos inser., 307.

cothurnus, 33.

cotonea, 75.

cottidie, 560; spelling, 227 sq.

coventio, S. C. Bacch., 250.

courauerunt, 246,

coxa, 298.

crabro, 220.

cracli for clatri, 97.

crapula, 197.

Crasis, 142 sqq.

Crassus, story of, 169.

crastinus, 325. ,

crates, 219, 279.

crebesco for -br-, 95.

credo, 479; -duam, -im, 514.

creo, 329.

crepa for capra, 98.

crepo, -ut, -avt, 499, 506; -itus, P.P. P.,
542.

crepus, 98.

crepusculum, 273.

cresco, 479 ; -évi, 500.

ceretariae for cet-, 96.

creterra, 118.

cribrum, 330.

crimen, 336.

crinis, 339.

crista, 339.

crocio, 476.

crudelis, 340.

crudus, 298.

cruentus, 352.

cruor, 298.

crustum, pronune. of, 141; -trum, 96;
-tlum, clustr-, 97.

Crustuminus (Cl-), 93:

-ct- for Gutt. with ¢, 291, 293.

-ctum, e.g. virectum, 335.

-cu- for quu, 86 sq., 300.

-cubi, 446.

cubiculum, scansion of, 175.

cubo, -ut (-avt), 506, 499.

cuculus, 290; -ullus, 115.

cudo, 486 ; -di, 502.

cui (see qui, quis) ; cui ret?, 606.

cuicuimodi, 445, 564.

cujus, Poss., 443, 447; cujds, 447.

cujuscemodi, 444.

288
628

culfus (#dATos), 59.

-culi, Perf., 50x.

culmen, 2353; (colum-), 185.

culmus, 328.

-culo-, Dim. (see Suffix -LO-).

culpa, col-, 236.

-culum, scansion of, 146, 175 sqq. (see
-clum).

cum (see com-).

cumbo, 471.

cumprimis, 565.

cunae, 258.

cunchin for co-, 33.

cunctus, 541 1.

-cundus, 544 sq.

cunila (2), 155.

-cunque, 598.

cuntellum for cult-, 97.

cupa (-pp-), 116.

cupio, 476; -is, 475.

cur, 606,

curd, 2473 coi-, coe-, 248 ; ‘cou-, 246,

curbus for -rv-, 51.

curia, 180.

curiosus, 353.

curriculo, 556.

curro, 239; cecurri,
Romance, 509.

cursim, 556.

curtina for co-, 34.

curtus, 239.

custos, 308.

-cutio (quatio), 196.

cutis, 260.

cygnus, 292; ci-, 36.

503; Perf. in

D, pronunce., 80 sqq.; and -t in at, ad,
&c., 76 sq. ; phon, changes of, 285
sqq.

D-particle (see Particles),

-d, Abl., 391 sq.

Daimatia (De-), 17.

-dam of quidam, &c., 552.

Dama, spelling of, 117.

damma, spelling of, 117.

danas esto, 183.

damnum, 328.

Danubius for -uv-, 51.

danunt, 531.

dapsilis, 340.

dasi (?), O. Lat Inf., 537.

datatim, 556.

Dative (see Declension).

 

THE LATIN LANGUAGE.

dautia, 286,

-de of inde, &e., 570.

de, Prep., 581 ; confused with di-. 574.

deabus, 403.

debeo, scansion of, 143.

debit for -lis, 376.

debilito, 176.

decem, 416; -cim, 19, 21; -cimus, 416,
Decius, 416.

decimanus, 326.

Declension (Noun, Adj.), 366 sqq. ;
Nom. 8g., 371 sqq. ; Gen., 379 sqq. ;
Dat., 385 sqq. ; Acc., 387 sq.; Voe.,
388 sqq.; Abl., 390 sqq.; Instr.,
392sqq.; Loc., 395 sqq. ; Nom. Pl.,
3978qq. ; Gen., gorsq.; Dat., Abl.,
Loc., Instr., 4o2 sqq.; Ace., 404;
(Pron.) Pers., 421 sqq.; Demonstr.,
431 sqq.; Rel., 443 sqq.; Pron.
Adj., 450 sqq.; (Verb) (see Verb).

declino, 470.

decor, M., -us, N., 356; decorus Adj., 356.

decreiuit, 22,

decussi-, 409.

dedro, 531; -ot, 531.

defendo, 486.

defrudo, the spelling, 196, 40.

defriitum, 261 ; -fri-, 197.

defuctus, 471.

degener, 356.

Degrees of Compar. (see Comparison),

degunere, 4.72.

Dehnstufe (see Gradation),

deinceps, 553.

deinde, 570; dein, 122.

delenio (-lin-), 199, 225. 4

delicatus, 287.

delico, 286.

delirus (-ler-), 199, 22.

delubrum, 331.

-dem of idem, &c., 441 ; of tandem, &e.,
552.

demo, -psi, 505.

Demonstratives (see Pronouns).

demum, 549 3 -S, 553- ;

Denominative (see Derivative).

dens, 540.

Dentals, phon. changes of, 283*sqq.

dentio, 485.

denuo, 564.

Deponent, 519 sqq.; Past Part. of
Act. Vb., 520, 5423; Perf. of Neut.
Vb., 522; bef. Pass. Inf., 522; Act.
 

INDEX.

Perf., 520; Pres. Part., 520; Act.
by-form, 527 sq.; Pass. use of, 522,
542; -art like Gk, -evew, 521.

deprensa, 336

derbiosus, 268.

Derivative Verbs, 478, 483 sqq.

descendidi, 502.

: desciso, pronune. of, 479,

deses, 358.

Desideratives, 478, 482, 484.

desidero, 488.

desilui (-ivi), 499.

desitus sum, with Pass. Inf., 522.

destino, 470, 472.

Determinative Comp., 360 sq.

detestatus, Pass., 542.

detondi (-tot-), 504.

devas Corniscas, 404.

deunx, 409.

devoro, Fut. Perf., 507.

deus (cf. div-),244; dei, di(i), Nom. Pl.,
399, 21; deum, -orum, Gen., 402;
deis, diis, Dat., 21; diibus, 404;
-dius (?), 618.

dextans, 409.

dexter, 285 ; -timus, 405 ; -tera, Adv., 550.

DH, phon. changes of, 289 sq.

DH-particle (see Particles).

dhé-, ‘to put,’ 457.

di- (see dis-) for de-, 574; for bi-,
412; difor z, 105.

dicae for -am, 492.

dicax, 355.

dicis, Gen., 358.

dico, 243 ; -ces(?), 2 Sg., 526; -c (-ce),
Imper., 518; -«i, 495, 497, 5043
-"0, 463; -xim, 465; -xerd, 212;
dicebo, 492, 494; dicturum, O. Lat.,
537-

-didi, Perf., 496, 502.

dienoine, Dvenos inscr., 560.

diequinti, 397; pronunc., 212.

dies, 252 ; pronunc., 24, 30,133; gend.,
369 n.; Nom. Sg., 377; Gen., 382 ;
dit, 382; in word-group, 169; in
Comp. Adv. 560; dienoine, Dvenos
inser., 416.

Diespiter, decl. of, 364.

difficul, N., 205; -lter (-liter), 553.

Digamma, Lat. F., 2,5; for v, 8.

Digentia, 287.

digitus, 76; -ct-, 185.

dignus, 293; pronune., 138 sq.

 

629

dilexi, Perf., 505.

diloris, 412.

dimico, 194 ; -avi <-ui), 499.

dimidius, 409, 159; de-, 30.

Diminutives, 333, 336 sq.

dimminuo, the spelling, 314.

dingua, O. Lat. for lingua, 286.

dinummium, 412. :

dinus for divin-, 52.

Diovem, 263 sq.

Diphthongs, phon. changes of, 239
sqq.; pronunc., 37 sqq. ; shortened
bef. cons., 251 sq. ; when final, 213 ;
Gk, in Lat. orth., 43 sq. ; Gk. e, 244.

dirimo, 582.

dirrumpo, the spelling, 314.

dis-, Prep., 582.

discerniculum, 333-

disciplina, 176; -plic-, 97.

disco, 477; didict, 50r.

disertim, 556.

dispalesco, 586.

dispennite for -nd-, 64.

Dissimilation of 1, r, 275; syll. lost
by, 176.

dissipo, 304; spelling of, 193.

distennite for -nd-, 64.

distinguo, 471; spelling of, 301 sq.

Distributives (see Numerals).

Dite, Voc., 389.

diu, 555; -tinus, 325.

dives, dis, 408; Compar., 408.

divisi, Perf., 498.

Division of Syll., 124 sqq.

divissio, the spelling, 110 sq.

dius, Adv., 555.

divus (cf. deus), 244; devas, 404.

dixeram illis, pronune. of, 123.

-dd of cupido, &c. (see Suffix -N-).

-do, Vbs. in, 486; Perf., 502; of condo,
&e., 457.

do, I give, decl. of, 457; da, Imper.,
518; dedi, 495 sq.; dedro (-t), 531,
124; datus, 222 (see duo).

doceo, 259, 482; -eunto, 3 Pl., 519..

dodrans, 409.

Dolabella, 331.

dolitus, 485.

dolus, 318.

domnus for -min-, 185.

domo, Vb., 474, 481; -ut, 506 n.

domus, 258; decl., 344; -mos, Gen.,
380, 384.
630

donec, 609; -icum, 609; -ique, 609.

donum, 232.

dorsualis, 340.

dos, 341.

dossum for -rs-, 96.

Double Cons., pronunce. of, 108 sqq. ;
for Single, 113 sqq.; written, 3, 8;
TI, 7 sq.; VV, 7; Vow.,3, 9 sq.

Doublets, 120 sqq , 204.

drachuma, 145.

Dropping (see Loss).

drua for tr-, 289.

Drusus, 289.

Dual, traces of, 366, 400.

Dvandva, Comp., 360 sq., 365.

dubito, 482.

dubius, 411.

dicenti, 419 ; -tum, O. Lat., 418 sq.

duco, 466; duc, Imper., 518; -xi,
pronunce., 498 ; ductus, pronune., 542-

duellum, 268.

Dvenos Inscr., 2.

duicensus, 411.

duidens, 411.

Dvigu, Comp., 360 sqq.

duis, O. Lat. for bis, 411.

dum, 609, 570.

*dumpa, 286.

dumtaxat, 565.

dumus, 237.

dune, 609.

duo for do, 515 5 -im, 515.

duo, Num., 410 sq.; scansion, 411 ;
-um, Gen. PL, 412; -a, Neut. Pl,
412; -decim, 416; -centi, 419 ; -vice-
simus, 417 ; -deviginti, 416.

Duodecim Tabb. (see Twelve Tables),

duonus, 268.

duplex (-us), 411.

dupundius (-on-), 197.

-dus, e.g. pallidus, 353 sq.

Dusmius, 237.

dw-, phon. change of, 265 sqq.

B, pronune., 18 sqq.; for oe, 44; for
ae, 42 sq.; for 7 in hiatus, 19,
22; in atonicsyll., 25, 30; é for 7,
25, 29 sq.; phon. changes of, 223
sqq. ; for short vow. in atonic syll.,
Ig sq., 194; for a after j, 17; for?
after 7, 230, 232 ; for.-i, &¢., 205 sq. 3,
@ for @ lengthened, 224; for -cyé-,
-ché-, 224; in grade of é-root, 260 ;

 

THE LATIN LANGUAGE.

not weakened to i, 199; shortening
of -2, 211 sq.; O. Lat. @ for ez, 244
sq. ; Gk. (see Eta).

-é, loss of, 204 sq.; ‘Dat.’ in, 387;
with -i in Abl. Sg., 390 sqq.

-@ of Adv., 548.

2, ex, Prep., 583; ec-, 583.

E-grade of Root (sce Gradation).

#-Subjunctive, 512 sqq.

edidem, Adv., 561.

eapse (-@?), 441.

-ebam of legebam, &c., 490.

-ebo, Fut., 3 Conj., 493.

-ebris of funebris, &e., 196.

ebrius, 592.

ecastor (ecc-), 618.

ecce, 617; -cere, 618; -wm, 617, 435;
-illum, -istum, 432.

eclesia, the spelling, 115.

ecquis, 4477.

Ecthlipsis, 309 sqq.

ecus for equus, 86.

ecepol, 618.

edice, Imper., 518.

édo, Vb., decl. of, 456 ; adi, 497 5 -am,
-im, 512; Imper., 518; essus, the
spelling, 112 ; -tus, 309.

‘ -ado, -tdula (see Suffix -D-).

edulis, 340.

edus for haedus, 42.

egestas, 326.

ego, decl., 421 sqq.; scansion, 422;
mihi, scansion, 422.

egregius, Voc. of, 389.

egretus, O. Lat., 285,

EI, on inserr., 244 sq.; for 7, 9; for
i, 245, 22; for @,22; phon. changes:
of, 243 sqq.; for atonic ai, oi, 243
sq. ; BI, 252; O.Lat. é for ei, 244 sq.

ei-, ‘ togo,’ 456.

ejero, 199.

einom, Dvenos inscr., 604.

-cis, Nom. Pl., O-stem, 398.

ejulo, pronune. of, 53.

-cius, Prop. Names in, 320.

ejuscemodt, 437.

-el-, phon. change of, 228 sq.

-éla (see Suffix -LO-).

elicui, Perf., 505.

-élis (see Suffix -LI-).

Elision, 144 sq.; of -m, 61 sq., 144;
-8, 123; -t of -d, -%, 381, 383 (see
Hiatus).
INDEX.

elixus, 293.

-ella (-us), -l-, 11a sqq., 115.

ellum, 617.

-ellus (see Suffix -LO-).

-em- for I.-Eur. m, 273 sq.

-em, -im, Acc. Sg., 388; for -am, Fut.,
493 7.

em, Interj., 617.

em, ‘tum,’ 438 ; ‘eum,’ 438.

emem, ‘eundem,’ 438.

eminus, 554.

emitor for im-, 30.

emo, 505; emi, 502; -psi, 505; -ptus,
spelling of, 70; emeru, 531; -psim,
466. ~

-endus, -undus, Ger., 544.

emungo, 471.

-emus of supremus, &e., 407.

-en- for I.-Eur. n, 273 sq.

en, Prep. (see i).

en, Interj., 617.

Enclitics, 165 sqq.

endo, indu, 582 sq.; and in-, 583; en-
doque plorato, 573.

eneco, spelling of, 194.

enim, 603 ; -vero, 603.

Ennius, introduced double cons.,
3, 8.

enocilis, 197.

enos, Carm. Arv., 425.

-ens (see Suffix -NT-).

-ens- for Gk. no (?), 1362.

ensis, 274.

-entia, -ium (see Suffix -NT-).

-entior, Compar., 407.

enubro- (-nib-), IgI.

enunquam, 617.

-énus (see Suffix -NO-).

eo, Adv., 568 sq.

e0, Vb., decl. of, 456 ; eunt-, Part., 541 ;
-tit, Perf., 214.

Epirus, accent of, 155.

epulonus, 348.

eques, 330.

equidem, 603.

equifer, 361.

equio, 484.

equus, 226; spelling, 300; pronune.
of e-, 42.

-er- for ri, 231 sq. ; for -dr- in faeneris,
&c., 34 (see also E).

:er, from -ros, 374; M., -ris F., 371.

ercisco, pronune. of, 479.

 

631

erémus, the scansion, 156.

erga, 583.

ergo, Prep., 583, ; Conj., 604; accent,
166; -d, 212.

erro, 308.

erugo, -cto, 298.

erumna for aer-, 43.

-runt, -ére, 3 Pl., 531 sq.

ervum, 196.

és, ést (see edo, decl. of).

és, est (see sum, decl. of).

-es (see Suffix -T-) ; -es, -is, Acc. PL,
404; Nom. Pl. 399; O-stems, 398.

esc, 310.

esc0, 479.

-ésimus, -ensumus, 418.

essere foy esse, 536.

essis (?), 466.

-esso, Vb.-forms in, 462 sqq.

(e)st, (e)s, Procope of, 121.

-ester (see Suffix -TERO-’.

et, Conj., 599.

Eta (Gk.), Lat. ae, 42 (see -ens-).

etiam, 599; -dum, 609; -num, 570;
-nune (-nn-), 62, 69, 121,

etst, 613.

-etum, e.g. arboretum, 335.

EU, pronune. of, 39 sqq.; phon.
changes, 245 sq.; EU, 252.

evenat, 464.

euge (-ae), 617; accent, 164.

-eus, -eum (see Suffix -10-).

-éx of remex, &c., 358.

exadversus (-m), Prep., 595-

examen (-agm-), 292.

examussim, 563.

exaurio, the spelling, 475.

exemplum, 271. ;

exfuti, 309.

exiet (?), Fut., 493.

eximius, 319.

exinde, 570.

exolesco, 481.

expergitus, 542.

expers, 192.

explenunt, 531.

explodo, 196.

explorato (-e), 559.

exsugebo, 494.

exta, 3II.

extemplo, 565.

extinguo, spelling of, 301 sq.

extispica, 358.
632

extra, 584.
exo, 475+

F, the letter, 2, 5, 291; pronune., 98
sqq.; for Gk. ¢, 11; from I.-Eur.
bh-, 282 sq. ; -bh-, 283; dh-, 289 sq. ;
-dh-, 289; ghw, 297; gh¥, 302;
for h, 294 8q., 56; b, 78, 80.

Fabaris, 95.

Jabula, 334.

Sacesso, 462.

facetus, 335.

faciae for -am, 492.

Facies, 3455 -18, 347.

facilis, -ciil, N. (Adv.), 553; -cile, Adv.,
554 3 -ciliter, -culter, 553; -cilumed,
8. C. Bacch., 559.

facto, 457; decl. of, 458; cal.e)f, &e.,
183 sq., 488 ; -is, -%4, 475 5 Seci, 497,
502; facte), 518; faxo (-im), 465 ;
-xet (?), 508; fefaked, Praen. fib.,
504; feked, Dvenos inser., 528.

Jacundus, 545.

Sfaenisicia (fen-), 42.

Jaenus (fen-), 42.

fagus, 221.

Falla, 355.

Fal, 486.

fama, 328.

James, 345.

Jamex, 355-

Somilia, 193; -& (?), Nom., 211; -as,

- Gen., 381.

Jamul for -lus, 374.

fanum, 307.

Jar, 357-

farcio, 476; decl., 458; -rsus, 542.

-fariam, 552. |

Sarina, 357-

Jarnus, 279, 294.

Sarreus, 277.

fastidium, 176.

fastigium, 277.

JSastus, decl. of, 344.

Satigo, 563.

Satim, 563.

Satum, 542; -us, 370.

Fatuus, 324.

fatuus, 324.

Savilla, 235.

Savor, 357-

Saux, 355-

Sebricula, 333.

 

THE LATIN LANGUAGE.

Jecundus, 545.

Jefaked, Praen. fib., 504.

fel, 295.

Seles, 346.

Jélix, 354.

fello, 225.

JSemina, 225, 327.

Feminine (see Gender).

femur, decl. of, 349 sq.

-fendo, Vb., 486 ; -di, Perf., 501.

Sere, ferme, 561, 185.

Seriae, 307.

Serio, percussi, 545 ; ferinunt, 531.

Jero, decl. of, 457; fer, 5175 tult, 545,
494, 503 ; tetuli, 494, 497.

Serox, 354.

JSertum, a cake, 310.

Serveo (-v0), 476; -bui, 51.

Serus, 297.

JSescemnoe (?), 398.

Sestino, Vb., 472; -us, Adj., 556; -ato
(-im), Adv., 550.

Sestus, 307.

Jetigo tor fat-, 18.

Setus, 344.

Siber (feb-), 229.

Jibula, 467.

Sicedula (-cella?), 353.

Jidele, Ady., 559.

Sidelia, 290.

Fidenae, the scansion, 127 7.

Jides, 345 5 -@, Gen., 383.

Jidicina, 358.

Jido, 243; in Romance, 488.

Jiducia, 337-

Jidus for foed-, 356 ; fidustus, 356.

Fifth Decl. (see Suffix -YE-, -E-). *

Jigel for -ulus, 375.

Sigtina, 184.

Figo, 467 ; -ai, 499 ; -xus, 542; figarus, 534.

Jigura, 291.

Jilius, 225, 22; -te, -i, Voc., 389 sq. ;
Jiliabus, 403.

JSilix (fé-), 229.

Final, Cons., pronune. of, 119 sqq.  -d,
122; -m, 67 8q., 123; -t, 124 5-1, 97;
-s, 108, 123; -t, 123; double cons,
IIg, 122; Vowel, short, 203 sqq. ;
long, 207 sqq. ; syll. in -m, 216 sq,

findo, 469 ; fidi, 495, 501.

Jingo, 297 ; -nxi, pronunc., 140 ; -nctus,
471.

Jini (-e), Adv., 552.
INDEX.

Jinitimus, 405.

fio, 545, 522; fi, 522; fieri, 132;
O. Lat. -re, 522 ; fitur, 522.

firmus, pronunce. of, 141.

First, Pers. Sg., ending, 524 sq.; Pl.,
529 ; Syll., accent. of, 157 sqq.

Jwo, O. Lat. for figo, 467.

Jixulae, 467.

flagro, 222 ; confused with fragro, 92.

flamen, a blast, -mm-, 118.

flavus, 279.

Slecto, 486.

flemina, 258.

Jfleo, 476.

Jlecuntes (?), 352.

flo, 476.

Flora, 356.

Slorere, in Romance, 489.

flos, 258.

Sluentum, 352.

Siu, 484 n. 5 -xi, 499.

Soculum, 289.

Sodico, 488.

Sodio, -odi, Perf., 502.

Soedifragus (-erif-), 365.

foedus, N., 356 ; fid-, 356.

Jons, pronune. of, 136.

for, decl. of, 457 ; fari, 221.

foras, 550, 557-

forceps, 178.

Sorcilla, the spelling, 239.

Jorctus, O. Lat., 182 sq., 541.

forem, 545 n.

Joris, 551; in Vb. Comp., 575.

forma, pronune. of, 141.

Sormidolosus, spelling of, 197.

Sormonsus for -os-, 69.

Sormus, 302.

Sornax, 239.

Sors, 278; Adv., 560 ; forsun, -m, 560;
Sorsitan, -m, 560; f. fuat an, 516;
forte, 560.

Sortasse (-is), 560.

Fortes and Lenes, 71 sqq.

Jortis, 541; ret, 342.

Sortuito (-u), 550.

JSorum, 289.

Foslius, 307.

Fovea, 295.

Soveo, 289, 302,

Fourth Decl., blends with Second,
343 Sq.

fr- for mr-, 269 sqq.

 

633

Sraces, 270.

Fractions (see Numerals).

Sragro and flagr-, 92 ; fragl-, 92.

Sragum, 306.

Srango, 222; fregi, 502; fractus, pro-
nune. of, 139.

Srater, 221.

Sraxinus, 279.

Frendo, 486 ; (-deo), 486 n. ; fressus, the
spelling, 115.

Srequens, 559.

Frequentatives, 478, 482 sq.

Srigidus, frid(d)-, 30, 119; -gd-, 185 ;
-aria, 172.

Srigus, 306.

Sviguttio, 488.

Srugi, 407; Compar., 408.

Srumentum, -mint-, the spelling, 23.

Srundes for -ond-, 31, 33.

Sruniscor, 470, 237.

Sruor, 484 n.; -imino, 519.

Srustra, (-4), 557 5 -au-, 40; -or, Vb., 558.

Srustrum for -tum, 96.

Suga, 239.

Sugio, 476 ; fugi, 502.

Jui (see fuo). .

Sulgeo (-9o), 476; -Isi, 505

Sulgorio, 485.

Fulica, (-lc-), 236.

Sulmentum, 310.

Sulvus, 235.

Suma, ‘terra,’ 295.

Sumus, 237.

JSunambulus, 364.

Fundanius, Gk. mispronunce. of, 58, 99.

fundatid, Luceria inser., 519.

JSunditus, 561.

JSundo, fudi, 502 ; exfutus, 309.

Sunera nec funeru, 616 n.

JSunerus (?), 356.

Sungor, 471.

Suntes for -ont-, 33.

Suo, -am, 515 ; fu, 518 ; fui, Perf., 545 ;
scansion, 132, 508 sq.

fur, 233; scansion, 215.

Surca, 239.

Surnus, 239 ; for-, 239.

Furo, 297-

Survus, 306.

Suscus, 306.

Fusio-, 305.

fussus, the spelling, 113.

futtilis, 309 ; spelling of, 117.
634 THE LATIN
Future (see Tense-stems) ; Fut. Perf.
(do.) ; Fut. Imper., 516 sqq.

G, letter, 2 sq., 6sq.; pronunc., 84
sqq.; from I.-Eur. Gutt. Asp., 291,
296 sq., 298, 302; Gutt. Ten., 292
sqq.; g3, 301 sq.; for gu, gor sq. ;
¢-, 72, 74 sqq.; G, phon. changes of,
296 ; G, 298 ; G, Zor sq.

gaesum, 305; spelling of, 112.

Gaius, 252 ; pronune., 53.

gallicinium, 194.

gallina, 370.

garrio, 277.
gaudeo, 479; gavisi, 522; -us sum,
522.

Gaulish, mispronune. 27.

gelu, 296, 261.

gemma, 273.

Gender, 368 sqq.; of N-stems, 349;
in Adj., 370 sq.; Fem. O-stems,
369; Mase. and Neut. O-stems
mixed, 369 sq.; Fem. A-stems
and Neut. O-stems, 400; disuse
of Neut., 369 sq.

gener, 271.

genetrix, IQI ; -nit-, 200.

genista, 195.

Genitive (see Declension) ; 10-stem,
accent, 163 sq.; Gerund. of Pur-
pose, 383 7.

geno for gigno, 459, 465.

gens, 341.

genu, 296 ; -nva, the scansion, 144.

genuinus, 326,

genus, 225,

germen, 271 sqq.

Gerund, &c., 543 sqq.

gesticulor, 488.

-gg- written for ng, 10 sq.

GH, phon. changes of, 296 sq. ; GH,
298; GH4, 302.

ghe- (gho-), Dem. Pron., 430.

gigno, 468 ; geno, 459, 465.

gingrina, 483 ; -rio, Vb., 483.

-gintd, the scansion, 418.

glaber, 290.

glacies, 261.

glans, 302,

glarea, 288.

*glerem for -lir-, 30.

globus, -mus, 80.

glos, 296.

 

LANGUAGE.

glosa, spelling of, 112.

gluma (glubo), 282.

-gm-, pronune. of, 89 ; -gn-, 64, 70.

Gnaeus, 294.

gnar- (Adj.), -us, 220; -uris, 541 ; (Vb.)
-itur, 485; -ivisse, 485 ; -igavit, 488.

gnatus (nat-), 541.

gnoritur, 485.

Gnosus, spelling of, 117.

-gnus (see Suffix -NO-),

-go (see Suffix -N-) ; for -guo, 301 sq.

gobius, 74.

 goerus for gy-, 36.

gorytus (co-), 74.

grabattus, 118.

Gracchus, Varro’s deriv., 93.

Gradation of Vowels, 253 sqq. ;
Weak grade, 255 sqq. ; of yé, wé,
256; E-grade, 255 sq.; O-grade,
255 8qg.; 8-8, 258; 6-0, 258; d-d,
258 sq. ; 6-4, 2588q.; A-d, 259; 4-3,
259 sq.; &-6, 260; 1-1, 260; 6-0,
260 ; ti-U, 260 sq. ; é-4, 261 ; Su-au,
261 sq.

gradior, 476.

gradus, 222.

grallae, 285.

gramae, -mmosus, 118.

grando, 297.

granum, 219 sq.

grates, 341.

gratis (-iis), 403, 551.

gratus, 279.

gravastellus, 330.

Grave Accent, 153 sqq.

gravedo (-%do), 23, 353.

gravis, 301 ; -id, 210; *grevis, 18. *

Greek, Aspirates (see Asp.) ; Diph-
thongs (see Diph.); Letters (see
Alphabet); Loanwords, parasitic
vow., 70 sq.; accent, 155 sq.; for
nuances of feeling, 182 ; for excla-
mations, 617 ; Mispronunce. of Lat.,
27, 45,58, 114. sq.; Orthography, in-
fluence on Lat., 12, 576 ; Phonetics,
infl. on Lat., 28, 32, 152 sq. ; Tran-
scription of Lat., ch. ii. passim, 135;
Lat. of Gk. (see under Gk. name
of letter, e.g. Eta, also Aspirates,
Diphthongs, Breathing, Tenues).

Grimm’s Law, 31.

groma, 96.

grus, 298.
INDEX. 635

-gu- for gu, 86 sqq., 301 sq. ; pro-
nune., 84 sqq.

guberno, 74.

gulfus (see cul-).

gummi, 74.

-guo (-go) in Vbs., gor sq.

gurges, 301, 358.

gurgulio, 275.

gusto, Vb., 482; -us, Noun, 296.

Gutturals, letters, 2 sq., 6 sq., 10 sq.;
phon. changes of, 290 sqq. ; three
series, 290 sqq. ; Proper (or Velar),
297 Sq.

gutus (-tt-), 116.

gyla, the spelling, 29, 36.

gyrus (goe-), 36 sq.

H, pronune. of, 53 sqq.; to denote
vowel-length, 54; hiatus, 265; for /,
56, 294 sq. ; dropped bet. vow., 54,
294 ; from gh, 296 sq. ; from gh, 2908.

habeo, 280; scansion of Comp. 143;
‘to dwell,’ 483.

habito, 482.

hacetenus, 433.

haedus, 242; faed-, 56; ed-, 42.

haereo, 242; -ssi, the spelling, 112;
“SUTUS, 542.

Half-long Vowel, 127.

halica (al-), 56.

halo, 220.

hanser (see anser).

harena, spelling of, 56.

hariolus, spelling of, 56; far- (?), 56.

haruspex, arisp-, 29.

hasta, 308. oe

haud,616 ; haw, 120, 122; -quaquam, 569.

havé, pronune. of, 49, 56, 127”.

haurio, 475; -ssi, the spelling, 112;
-surus, 542.

hebes, 351.

hec, O. Lat., 433.

Hecoba, O. Lat., 197.

hedera, haed-, 43.

Hedonei, Gen., 381.

heia, 617 2.

helvus, 276, 229.

hem, Interj., 617 ; pronune. of, 61.

Hercules, spelling of, 197.

Herentas, 482.

heri, 264, 396; (-e), 25 sq.; -slernus,
pronune. of, 135.

heries (heriem Junonis), 345+

 

Heteroclite, Nouns, 367 ; locus, -ca,
400; Vbs., 545 (see Comparison,
irreg.).

heu, Interj., 39.

Hiatus, 144 sq. (see Prosodical H.)

hibernus, 269 sq.

hic, Pron., decl. of, 430 sqq.; pro-
nune., 433 3 huic, pronunce., 44.

hic, Adv., 567, 433-

Hidden Quantity (see Close Syllable)

hiems, 358, 297 ; -mps, 7o.

hilaris, 338 sq. ; -rus, 182.

-hilaritudo (-tas), 341.

Hiluria for Ilyr-, 36, 115.

hinnuleus, pronune. of, 118.

hio, 476.

hiquidem, 433.

hircus, spelling of, 56.

hirrio, go.

hirsutus, hirtus, 229.

hiuleus, 337.

hoc (Adv.), O. Lat., 568.

hocedie, 433.

hodie, 561; pronune., 84.

holus (hel-), 228 sq.; -atrum, 362.

homicida, 364.

homo, 349; hem-, 367 ; hum-, 33, 236;
-ullus, 333 3 -wncio, 337; -~unculus, 337.

hordeum, 298.

horitur, Enn., 482.

horreo, 277.

horsum, 568.

hortor, 482.

hortus, 296.

hospes, 178, 298.

hosticapas, 187, 371, 373-

hosticus, 337.

hostis, 298, 341.

huc, Adv., 568.

*nucare, 486,

hui, Interj., 39-

hujuscemodi, accent. of, 162.

humane (-iter), 554.

humerus (see um-).

humilis, 338.

humus, 236; decl., 344.

Hydruntum, 289.

I, the letter, 3, 7 sq.; written for IT,
7 sq. ; tall form, 4, 8 sqq., 47, 133
sq., 137 .: doubled in aiio, &e., 8,
47, 53; symbols of long, 9; pronune.,
23 sqq.; phon. changes, 230 sqq. ;
636 THE LATIN
Lat.? for atonic vow., 193 sqq. ; for é
bef. ng, gn, &e., 225 sq., 229 sq.; in
hiatus, 19, 21 sqq.; in tuncine,
&e., 206 sq. ; for % in optimus, &e.,
189, 23 sqq. ; é- prefixed to st-, &.,
102, 105 sqq. ; -i dropped, 204 sq. ;
Lat. 7 for @ in filius, &e., 224 sq.;
for atonic ai, oi, 243 sq.; for ei,
243 sqq.; for-% lengthened, 230;
varying with i, 260; -2 shortened,
213.

J, the letter, 7; pronunc., 44 sqq.;
I-Eur. Y, phon. changes of, 262
sqq.; Lat. j for Shy, 263 sqq. ; for
dy-, 263 sq. ; dropped bef. accented
vow., 144.

I-stems, mixed with Cons.-stems,
338, 401 (see Suffix -I-).

I-Subjunct. (Opt.), 513 sqq.

-t- in Fut. Perf, 510; Perf. Subj,
500 ; 3 Conj. Vbs. in -io, 475.

i-t, %-i, in stupila, &e., 37.

ja- pronounced je-, 15, 17.

Jacio, jaceo, 473 3 jeci, 502.

Jaculum, 332.

Jajentaculum, 17.

Jajunus, 17.

iam, ‘eam,’ Acc. Sg., 437.

Jam, 579.

Jandudum for -md-, 66, 121.

janitrices, 274.

janto, 17.

Janua, 264 ; jen-, 17.

Januarius, pronune. of, 15 3 Jen-, 17.

Janus, decl. of, 344; -is, Carm. Sal.,
339.

-ibam, -iebam, Impft., 491.

ibi, 567 5 -dem, 571, 567.

-ibo, Fut., 493.

-ic of illic, &e., 551.

-icanus, e.g. Afr-, 327.

tcl, Perf., 502.

-icius (see Suffix -KO-).

Ictus, and accent. in Plaut., 165 sqq.

-icus (see Suffix -OK-).

idcirco, 580; icc-, 314.

idem, 431 ; decl. of, 441 sq.

identidem, 571.

ideo, 568.

idolatria, 176.

idélum, the scansion, 150.

-idus, e.g. pallidus, 353 sq.

je- for ja-, e.g. jecto for jacto, 15, 17.

 

LANGUAGE.

Jjecur, decl. of, 349 ; joc-, 41.

-ieis, old spelling of -eis, Dat. Abl. Pl.,
Ig, 21.

jejunus (jaj-), I7 ; pronune., 53.

-ie(n)s of Numeral Adv., 408.

iens, euntis, Gen., 541.

Jjento (ja-), 17 ; jejent-, 17.

-ier, Inf. Pass., 536 sq.

-térem, accent of, 164.

igitur, 565, 605 ; accent., 169.

Ignatius for Egn-, 229.

ignis, 229.

ignoro, 485.

ignosco, 363, 615.

-igo of navigo, &e., 479.

-igo (see Suffix in Gutt.).

-tit of abiit, &e., 528; of audiit, &e.,
132.

tlicet, 564.

ilico, 564.

ilignus, 293, 229.

-ilis (see Suffix -LI-); -iis, of Pass.
Adj. (do.).

illac, Adv., 569.

ile, 430, 436 sq.; pronunc., 122;
accent., 167; -ui, Dat. Sg., 452.

illée and illéx, 135 n.

dex, the spelling, 112.

illic (-t), Adv., 567, 432.

ilicio, -exi, pronune. of, 498, 139.

ulim, Adv., 570.

illimodi, 431.

-illo of Dim. Vbs., 479, 487 sq.

illuc, Adv., 568.

-illus (see Suffix -LO-) ; (-ius), 115.

ilustris, 293.

-im of sensim, &e., 548; of illim, &e.,
551; (-em) Acc. Sg., 388.

im, ‘eum,’ 438.

imago, 521.

imbilicus for umb-, 29.

imeum (?), ‘eundem,’ 438.

imitor, 521.

immanis, 339.

immo, 603.

impendio, 560.

Imperative (see Moods); as Particle,
600 ; accent of addiic, &e., 163.

Imperfect (see Tense-stems).

impero, 192.

Impersonal Pass., 520 sq.

impetrio, 485.

impleo, 473.
INDEX.

impliciscor, 480.

implicitus, P. P.P., 542.

Imporeitor, 279.

impraesentiarum, 562.

imprimis, 565.

improbo, 615.

impudenter, 554.

impune, 559.

-imus of Super]. (see Comparison).

imus, infimus, 407.

in (en), Prep., 584; im, _im-, 50 8q., 66,
69,99 sqq., 121; bef. s-, f,136sqq.

in-, Neg., 615; with Vb., 615, 363.

-ina (see Suffix -N-).

incassum, 564.

Inceptives, 476 sq.,4.79 8qq.} -ésco, 134.

incipisso, 462.

incitega, 197.

inclino, 470.

inclutus (-lit-), 239; spelling of, 197.

incogitabilis, 334.

incoho, spelling of, 57.

incolomis, the spelling, 192.

incoram, 581.

incubus (-bo), 348.

incurvicervicus, Accius, 360.

indaudio, 583.

inde, 570; pronunc.,
570 Sq.

Indefinite Pron. (see Pronouns).

Indeterminate Vow., 257.

indigena, 583.

indigeo, Vb., 583 ; -us (-ens), Adj., 540.

Indo-European, languages, 218 ; pro-
totype of word, 218; alph., 218 sqq.

indoles, 345, 583.

indu (see endo).

indugredior, 583.

induo, 475.

induperator, 583.

industrius, 189.

inebrae aves, 191.

infans for infandus, 182.

inferebis, 494.

inferus, -fer, 3743; fra, Prep. 585;
sfera, 181 ; infimus, imus, 407.

Infinitive, 535 sqq.; as Imper., 517;
Pres. Act., 535, 537; Pass., 536 sq. ;
Fut. Act., 536 sq.; Pass., 536, 538;
Perf. Act., 536; Pass., 536; Hist.,524.

, infit, 546.
infiteor, 615.
ingens, 274, 541.

122; -didem,

 

637

inger, Imper., 526.

ingratis (-tis), 551.

inibi, 567.

inipite (?) 198.

Initial Syll. (see First Syll.).

inlicite, the scansion, 475.

inpetrator, 22.

inquam, 524; Conj. of, 545 sq.

inquies, -etus, 182.

inquilinus, 227.

insane, -wm, 550.

insciens, -us, 540.

insequo, -co, 566; decl., 545.

inserinuntur, Liv. Andr., 531.

insons, pronune. of, 136,

insperatas, Nom. Pl., Pompon., 398.

instar, 205.

instigo, 284, 471.

institii, Perf., 508.

Instrumental, 548 (see Declension) ;
Adv. use of, 559 sqq.

insuper, 593.

int (2), 3 Pl., 456.

intellexi, Perf., 505.

intemperies, Sg., -ae, Pl., 347.

inter, 585.

interatim, 556.

interdiu (-s), 555.

interduatim, 556.

interdum, 609.

interduo, 515; -im, 515.

interealoci, accent. of, 162.

interibi, 567.

Interjections, 616 sqq.; from Gk.,
617; Imper., 618; accent. of,
164.

interim, 570.

Interrogative, Particles, 605 sq. ;
Pron. (see Pronouns).

intervius, 563. Fi

intolerans for -andus, Laev., 518

intra, Prep., 585; -tro, Adv. 561. *

intrare, 474.

intus, Adv., 561; Prep., 585.

inventio, 274.

invito, Vb., -us, Adj., 299.

-inum for -tnum, 23.

inunt, 3 Pl., 531.

involucrum, 329.

-inus (see Suffix -NO-),

jocus, 264.

-iélum, accent. of, 164.

-tor (see Comparison).
638 THE LATIN

Iotacismus, 27.

Ioues (Dvenos Inscr.), 264.

ipse (-us), 430, 4408q. ; isse, 79; ipsima,
-issumus, tpsippe (-pse, -pte), 441;
eapse (-c), 441.

ira, spelling of, 245.

iri in Fut. Inf. Pass., 538.

Irregular Verbs, 545 sqq.

-is of Cornelis, &e., 375.

is, Pron., 430; decl. of, 437 sqq.;
ejus, pronune. of, 53.

-isco for -esco, Incept., 480.

Issa, 79.

-issimus, Superl. (see Compar.).

-isso, Vb., 488.

istac, Adv., 569.

iste, 430; decl., 435; pronunc,, 122;
accent., 167.

istic (-1), Adv., 567.

istuc, Adv., 568.

-it, 3 Sg. Perf., 527 sq.

ita, 5713 -idem, 571.

Italia, the scansion, 1277.

-ilanus, e.g. Abder-, 327.

itaque, 604 8q., 571; scansion, 604.

itare, 482.

item, 571.

-iter, Adv., 549, 553.

iter, decl. of, 349.

Iterativés, 478, 482 sq.

iterum, 330; 550.

-itia (-wm) (see Suffix -I0-).

itur, Impers., 520.

ju- pronounced ji-, 15.

jubeo, 481; jussi, pronune. of, 110 sq. ;
O. Lat. joussei, 498.

Jucundus, 545.

judex, 182,

Jugatinus, 325.

Jugis, 338.

jugmentum, -gum-, 336, 292.

juger-, 245; tugra for-era, Lex Agr., 184.

jugum, 237, 264.

Julius, 250.

ium, ‘eum,’ Luceria inscr., 437.

jumentum, 336.

Jungus (2), 471.

junior, 408.

juniperus (-pir-), 374, 1925 Ji-, 35.

Junix, 345.

Jupiter (-pp-), 246, 389; spelling, 116;
decl., 377; Jov-, 263 sq.

Jjurigo, O. Lat., 173.

 

LANGUAGE.

jus, ‘broth,’ 237; ‘law,’ 264.

-ius, Compar., scansion of, 406 n.

jusjurandum, 358.

Justus, 356.

juvenalis, -ilis, 340.

juvencus, 264.

juvenis, 239; Compar., 408.

juventa, 334.2. ; -tas (-tus), 341.

juvo, 476 ; juerint, 508.

juata, 585.

-ix of felix, &c. (see Suffix -KO-, Gutt.);
of cornix, &e. (see Suffix -I-)

K, the letter, 2,6 sq ; pronune., 84
sqq.; phon. changes of I.-Eur. K,
295 sq.; of K, 297 sq.

kadamitas, 286.

Kappa in Lat., 72.

Karmadhfraya, Comp., 360 sq.

ke- (ko-), Dem. Pron., 429 sqq.

L, pronune. of, 89 sqq.; bef cons.,
96 sq. ; I-Eur. L, phon. changes
of, 275 sq.; L, 278 sqg.; Lat.7 for d,
80, 82, 285 sqq.; for 7, 92 sq.; for
n,.96; for ll, tog sqq.

1- for tl-, 283 sq.

-l, vow. shortened bef., 213; decl. of
Nouns in, 376.

Labials, phon. changes of, 281 sqq.

Labiovelar Gutt., phon. changes of,
299 sqq. (see Q¥, g% Q@He.

labes, 345.

labium, 261.

labo, 303.

labrum, 180.

lac, spelling of, 122; lact (-te), 378. ~

lace, 307.

lacer for -ratus, 540.

lacesso, 462.

lacio, Ig.

lacrima, 223; spelling, 57 sq.

lacus, 301.

lacusta for loc-, 201.

laevus, 242.

lambero, 479.

lambo, 471; -bi (?), Perf., 501.

lambrusca for lab-, 65.

lammina, -mn- (-nn-), 184.

lana, 279.

lancino, 470.

langueo, 306.

lanius (-i0), 348.
INDEX,

lanterna (-mpt-), ‘Jo.

lapis, 353-

larignus, 293.

Larinum, 288.

larix, 286.

larva (-rua), 46.

lassus, 258.

latex, 355.

laticlavus, 361.

Latona, 349.

ldtus, 219 sq., 541.

lavo, 235; -vi, 499; lautus, lo-, 250 ; -luo
in Compounds, 196.

laurus, 286.

lautia, 286.

lector, pronune. of, 139.

lectum (-s), 542; pronunce., 139.

Legato Pronunciation, 131.

legatus, 541.

legirupa (-ger-), 373, 192.

legitimus, 405.

lego, 260; légi, 502; lectum, pronunce.,
139.

Lenes and Fortes, 71 sqq.

Length (see Long Cons., Vow.)

Lengthening, by Position (see Pos.) ;
by Compensation, 314; bef. nf, ns,
136 sqq.; gn, gm, 138 sq.; ct, x,
139 sq.; net, nx, 140; r with cons.,
140 Sq.; s with cons., 141.

lenibat, 491; -ibunt, 493.

lenocinor, 488.

-lens (-lentus) (see Suffix -NT-).

lentus, 252.

lepesta, 286.

leptis for nep-, 96

Letters (see Alphabet).

Levana, 326.

Leucesie, Carm. Sal., 245.

levir, 200, 242.

lévis, 292.

lévis, 244.

lex, 260. ‘

-lexi (-legi), Perf., 505.

-li-, syncopated after cons., 171.

libertabus, 403.

libet (tub-), 29.

libra, 289.

licet, 613.

lien, 349.

ligula, 272.

‘Ugurrio, 291, 482.

limitrophus, 176.

 

639

limus, mud, 328.

lingo, 471.

lingua, 229, 286.

lingula, 272.

linio, 483.

lino, 470.

linquo, 469 ; liqui, 502.

linter (lun-), 29.

liquare, -ere, 484.

liquor, 268.

lira, 199.

littera, spelling of, 117; 1. canina, go.

litus, spelling of, 117.

lixa, 293; -ivus, 323.

-ll- for 1d, In, 1s, 275 ; dl, 285; nl, 271;
rl, 277.

Locative (see Declension); Adv. use
of, 559 saq.

locus, -cc, Pl, 400; O. Lat. stlocus, 303,
307, 564; in wordgroup, 170.

Loebasius, 248.

loedus, O. Lat., 248.

Long, Cons., orth. of, 3, 109; Vow.,
3 8q., 9 Sq. (see Quantity).

longus, spelling of, 236.

loquella, the spelling, 112 sqq.

loquor, 284.

Losna, 292.

Loss, of Cons. in Group, 309 sqq.;
of final syll. in -m, 216 sq.

lotus, 250.

lubricus, 306.

lubs, Marso-Lat., 12, 177.

luceo, 481.

lucerna, 237.

luci claro, 396.

Lucilius, i and et, 9, 27; g for gg, 11;
a for d, 10, 14; e and ae, 42; 7, 90;
pellicio, 97; numeri, Gen., 383 7.;
mille, 420; illi, 437.

Lucipor, 183.

Lucretius, use of Compounds, 360.

luctus, pronune. of, 135.

lucus, 250.

ludus, 287 ; loed-, 248.

lues, 345.

lumbus, 290.

luna, 292.

lupus, 291; F., 370.

lurco, 179.

*lir(d)dus for liri-, 37.

listra, listra, 141.

lustro, 293.
640 THE LATIN
lutra, 289.

lux, 276 (see Luci).

lympha, 286 ; spelling, 11, 36.

M, the letter, 7; pronunc, 60 sqq.;
final dropped, 68 sq., 123; assimi-
lated, 121; I-Eur. M, phon.
changes of, 268 sqq.; M, 273 sq.;
Lat. m for n, 269; bef. v-, f-, 50 sq.,
66, 99 sqq. ; for p, 281; for b, 282sq.

-m, vow. shortened bef., 213; final
syll. dropped, e.g. nihil(um), 216
sq.; of 1 Sg., 524.

macer, 223.

macero, 488.

madeo, 223, 473.

maereo, maestus, 306, 542.

magis (-e), 558.

magister, 232.

magistratus, 343.

magnanimus, 364.

magnopere, 362,

magnus, 261.

Maia, -ti-, 8; -di-, 105.

major, 292, 408; pronune., 53.

malé, 551 3 mal(e)-, 184; -ficus, 51.

malignus, pronune. of, 138.

malo, 547; -Ul-, III sqq. ; malim, 515.

malogranatum, 364.

mitlus, 307.

Mamers, 95.

mamilla, 113.

mamma, 118 n., 363.

mamor for marm-, 95.

mamphur (?), 197.

mandare, 485.

mandere, 472; Perf. of, 501.

mane, 396.

maneo, 4763 -nsi, 505.

mani- (-no-), 183; Manes, 339.

manico-, a handle, 337.

manifestus (-nuf-), 193.

maniplus, pronune. of, 94.

mansues (-tus), 182.

mantele (-llum), 117.

manudlis, 340.

marcerat for -cidat, 288.

mare, 338.

maredus for -didus, 288,

maritimus, 405.

Marius, 320.

marmor, 18 ; pronune., 95.

Marpesius (-pessos), 117.

 

LANGUAGE.

Marpor, 185.

marsuppium (-p-), 117.

Marsus, 84.

mascel, 375.

Masculine (see Gender).

Maspiter, 278.

massa, 104.

mateola, 19.

mater, 219.

matertera, 405.

matruelis, 340.

*mattinus for -tut-, 184.

mattiobarbulus, 197.

mattus, 185, 309.

maturrime, 407.

mavolo, 547.

maximus, 407; pronunce., 139.

Media Prosodia, 161.

Mediae, pronune. of, 71 sqq.; M.,
Tenues and Asp., phon. changes of,
279 8qq-

mediastinus (-tri-), 330.

medioximus, 407.

medipontus (-1-), 287.

meditor, 521, 287.

Meditrina, 347.

medius, 226.

medius fidius, 618,

medullitus, 561.

nY>

| mehe, O. Lat., 422,

mehercle (-cules’, 618.

mejo, 466.

mel, scansion of, 122.

meles, 346.

Melica for Med-, 287.

melior, 406,

meltom (2), 406,

melum for mail-, 18.

membrum, 270.

memini, 270; -mento, 517.

memor, 541.

mendicus, spelling of, 245.

mendum, N. (-da, F.), 400.

menetris, for meretrix, 96.

mens, 274; -mente, Adv., 549, 552.

mensa, pronune. of, 67.

mentio, the scansion, 212,

-mentum (see Suffix -TO-).

mercedimerae, 540.

mercennarius (--), pronune, of, 118.

merda, 306,

mereo, -eunt 3 Pl, 519;
the scansion, 202.

merebatur,
INDEX.

meretrix, 347 ; menetris, 96.

merga, 351.

merges, 351.

mergo, 285.

meridies, 288; med-, 288,

merito, 559.

mers for -rx, 107.

merto for -so, 482.

merula, 288.

messis, 340.

-met of -~omet, &e., 421, 423 sq., 429.

Metathesis of 7, 1, 91, 97 sq.

Metellus, 486.

metior, mensus, 471.

méto, 486; -ssui, 499.

metuculosus, 333.

metus, F. (O. Lat.), 343, sq.

meus, 426 sqq.; mi, Voc., 427.

Mezentius, Med-, 104.

mica, 306,

migro, 302,

miles, 287; scansion, IIg, 215; -ex,
108.

mille, 419 sq.; milia, spelling of, 112
sq.

milvus (-uos), 46.

mina (pra), 64; in word-group, 169
sq.

minerrimus, 407.

Minerva, 306, Igo.

mingo, 455

-mini, 2 Pl., Ind., 534; Imper., 517.

minimus, 407 ; ~me (-mum), Adv., 550.

miniscitur, 200.

minister, 232.

-mino, Pass., 517, 519.

minor, 407.

mi(n)sterium for -nist-, 173, 202.

minuo, 471.

Mircurios, Praen. inser., 229.

mis, Gen., 421 sq.

misceo, 479; mixtus, -stus, 107.

misellus, -serulus, 333-

miser, 306; my-, 29.

mitto, pronune. of, 117; misi, O. Lat.
meissel, 499, I12.

mitulus, myti-, 37.

ml, phon. change of, 270 sq.

-mn-, pronune. of, 64, 69 sq.

mo- for meo-, Poss., 426.

moderor, 356.

modestus, 356.

modo, 612; -b, 551; Comps. of, 559.

 

641

modus in word-groups, 169, 564.

moe-, O. Lat. for mu- in moerus, &e., 248.

moles, 345 sq.

mollibit (?), 493.

monedula (-er-), 288.

moneo, 477; moneris, -int for monu-, 506.

-monium (-ia) (see Suffix -10-).

Monosyllables, shortening of, 215 sq.

monstrum, 331.

Months, Gender of, 369.

monumentum, -nom-, 201 ; -tus, 370.

Moods, 511 sqq.; Subj. 511 sqq. ;
Opt., 511 sqq.; Imper., 516 sqq.

mordeo, 303, 482; Perf. of, 505; mo-
mordi, 497; mem-, 503.

morigeror, 485.

morior, 473; -imur, -iri, 475; -tuus,
324; -tus, 174.

moror, 483.

mors, 278,

morcus for -bus, 52.

mostelum, 331.

moveo, Perf, in Romance, 509.

mox, 555.

mr, phon. change of, 270 sq.

-ms-, phon. change of, 270.

mucus (-cc-), 116.

mulceo, -Isi, 505.

mulgeo, 296 ; -lsi, 505; -lctus, 279.

mulier, 287 ; -iérem, accent. of, 164.

muiltimodis, 362.

multus, Compar., 408.

Mummius, dedicatory inserr., 11.

-mungo, Vb., 471.

munus, 247.

-mur, x Pl, Pass., 534.

*murca for am-, 107.

muriola, spelling of, 117.

murmur, 315.

-mus, I Pl. Act., 529; scansion, 529.

mus, 237.

mMusct, 239.

muscipula, 365.

musimo, 7I.

Mutation of Vowels (see Gradation).

Mute and Liquid, vow. lengthened
bef., 94; accent. of penult with, 164.

muto, 247; for -tor, 522.

muttus, 58.

mutuus, 324.

myser for mis-, 29.

N, pronune. of, 60 sqq. ; L.-Eur. N.,

Tt
642 THE LATIN
phon. changes of, 271 sqq.; N, 273
sq.; Lat. » for 1, 96; for m, 270;
for gn-, 292, 294.

N-particle (see Particles).

Naepor, 183.

Naevius, 294.

Nahartis, Nart-, 54.

nam, 604, 571.

Names, of birds, &c., changed by
Anal., 201; Proper, form of Italic,
319 sq. ; in -o, 348 sq. ; in -ds, 375.

namque, 604.

nanciscor, 261, 480;
(-netus, 471.

naris, 272.

narro, 483; pronune., 118; -7, 118 sq.

Nasal, pronune. of, 60 sqq.; Gutt.,
Io sq., 60, 65; Vb., 469sqq.; 3 Pl,
e. g. danunt, 530 sq.

nassiterna, the spelling, r12.

nasus, 259; -ssum, ITZ.

nitrix, 355-

natus (gn-\, 541, 219 8q., 294.

naufragus, 180; -ium, 252.

nevis, 221, 252.

-net-, pronune. of, 64, 70.

-nd-, pronune. of, 64, 70.

-nde of unde, &e., 551.

-ndo-, of Gerund, 543 sq.

-ne, Interrog., 605 sq.

-ne, Affirm., 615.

né-, Neg., 615.

né, nae, Affirm., 614 sq.

né, Neg., 614; accent. of, 166.

nebula, 226.

nec, ‘non,’ 616.

nec (see neque).

necerim, ‘nec eum,’ 440.

neco, 481; -avi (-ut), 499.

necto, -xul, 509.

necubi, 446.

necumquem, 570.

nedum, 614.

nefas, 615.

Negatives, 615 sq.

neglego, 616; -xt, 505.

nego, 486.

negotium, 616.

nemo, 449.

nempe, 604; scansion, 63; pronunc.,122.

neo, 476, 225; neunt, 519.

nepos, 351,272; -Us, 32.

neptis, 351 ; lept-, 96.

-cio, 480; -ctus

 

LANGUAGE.

nequalia, 571.

nequam, 571.

nequaquam, 569.

neque (nec), 122.

nequeo, 547; -quinont, 531 ; -quitur, 522.

Nero, 271.

nescto, 547.

nesciocube, 446.

nesi, 592.

neve (neu), 614, 122.

Neuter (see Gender); in -r, decl. of,
349; in -es, 355 sq.

neuter, 450; pronunc., 143.

neutiquam, pronune. of, 143.

-nf-, lengthening of vow. bef., 136 sqq.

nt, OT.

nicto, 293.

Nigidius, on h, 55 ; Agma, 65 ; accent
of Valeri, 163, 390.

nihil (nil), 216 sq., 144, 57-

nimis (-ium), 558.

nisi, 611; -se, 25.

nitedula (-ella), 333, 353-

nitor, Vb., 521 ; -tito, 519; -wus, 294.

nittio for -ct-, 89.

nivit, Vb., Pacuv., 455.

mix, 272,

-nm.-, phon. change of, 271 sqq.

no, 476.

nobilis, 334.

noceo, 481; -ivus, Adj., 323.

noctu, 555.

noenum (-t), 615.

nola (from nolo), 318.

nolo, deel. of, 547; -l-, 111 sqq.; -lim,
515.

nomen, 294 ; -clator, 364.

Nominative (see Declension); Adv.
use of, 553 sqq.

non, 615, 216 sq. il

nonaginta, 417.

nondum, 5°70, 609.

nongenti (noning-), 419.

nonne, 605.

nonus, 416.

*noptia for nuptiac, 37.

nos, decl. of, 424 sq. ; -ss, I12.

NOSCO, 4793 gN-, 294; -vi, 500.

nostras, &e., accent. of, 163.

*noto for na-, 15, 17.

notus, (gn-), 233.

novem, 415 8q.3 -decim, 416; -venus.
416.
INDEX.

Novensiles (-d-), 286.

novicius, 338.

novitas, 341.

novus, 226,

nox, 234 sq.; Adv., 555.

-ns-, pronune. of, 63 sq., 69; for s
after long vow., 69 ; from -nss-, 112;
vow. long bef., 136 sqq.

-ns of triens, sextans, &c., 409.

-nt, 3 Pl., 529; -nto, 531; -ntor, 5353
~ntur, 534.

nubes, 346 nubs, 182.

nudipes, 361.

nudius tertius, 260, 377.

nudus, 260, 235, 179.

nullus, 449; pronunce., 113.

num, 605.

Numasioi, Praen. fib., 305.

Number, 366; Dual (see Dual).

Numerals, 408 sqq.; Fractions, 409;
accent. of, 165.

Numerius, 560 ; -ri, Gen., 383 n.

numerus, 270 ; numero, Adv., 560.

nummum, Gen. Pl., 402,

nune, 570.

nuncubi, 446.

nundinae (nond-, nound-), 251, 180.

nunquam, 570.

-nunt, 3 Pl. e. g. danunt, 530 sq.

nuntius, 180; novent-, nont-, 250 sq.;
pronune., 141,

nuper, Adv., -rus, Adj., 180, 553.

NUYUS, 239 5 -T, 343; NO-, 37-

“nus, e.g. facinus, 356.

nusciosus (-citiosus), 96.

nutrio, 249.

O, pronunec. of, 30 sqq.; phon.
changes of, 232 sqq.; 6 in e-roots,
258; 0 in &-roots, 258; 4-0, 260;
6-4, 259; Lat. 6 for lengthened 6, 233;
Lat. 6 for u, 17 sq.; for e with
w, I, 225 sqq.; for atonic vow. bef.
Lab., 192 sqq. ; after qu-, 300; Lat.
6 for au, 40 sq.; for ou, 248 sqq.;
not weakened to #, 199.

O-stems, blend with U-stems, 343
sq. (see Suffix -O-, -YO-, &c.)

-6, 1 Sg., 524; of Rufo, &c., 348 sq.;
of quo, Adv., &¢., 551; of vero, Adv.,
&e., 550; for-2 in Adv., 559; short-
ening of, 212 sq.

ob (obs-), Prep., 585; ousted by ab-, 574.

 

643

0b vos sacro, 573.

obinunt, 531.

obiter, 563.

oblitterus for -atus, 543.

obliviscor, -lise-, 52; -litus, Pass. 543.
oboedio, 196, 246.

obrussa, 198.

Obscure Vowel, 185 sqq., 257.
obsequium, 321.

obsolesco, 481.

obsono, 488.

obstetrix, 191.

obstino, 472 ; obstinatus, 310.
obstringillo, 487.

obtineo, spelling of, 78 sq.
obviam (-us, Adj.), 563.

‘occanui, Perf., 509, 198.

occidamus (?) for -céd-, 199.

occillo, 487.

occulo, 227 ; oquoltod, S, C. Bacch., 227.

occupo, 470, 486.

octor, 259, 406.

ocris, 259.

octo, 415; -decim, 416; -ginta, 417 8q.;
-tag-, 418 ; -tuag-, 418 ; -tingenti, 419 ;
-tavus, 220.

oculus, 234.

odium, 259.

odor, 235.

OF, pronune. of, 39 sqq.; spelling,
246. (See OL.)

oenus, 248, 410.

ofella, 113.

offendices, 272.

offendo, 486; -sa, Noun, 542.

officina, 174, 349. 7

OI, phon. changes of, 246 sqq. ; OI,
252 sq.; OI, OE on inserr., 247 sq.

oiet, Interj., 39.

oino$, 410.

Oinumama, 193.

oinuorset, 178.

-ol- from el, 228 sq. ; 1, 278 sq.

ol-, Dem. Pron., 430.

olentica, 337.

oleo, 287.

olim, 5703; vu
436.

oliva, 228 ; -wm (oleum), 196.

olla (aula), 41.

olle (-us), 436; oloes, olorom, 436.

Olympus, accent. of, 155.

Omega, Lat. transer. of o, 44.

oliorum, 436; olitanus,

Tta2
644

Omicron, Lat. transer. of, 33.

omitto, T13 sq., 202,

ommento, 80.

omnimodis, 362.

Omnis, 450; omnino, 325.

Onomatopoetic Verbs, 483, 476.

-onssus (-dsus), 353.

-ont, 3 Pl., O. Lat., 529.

-onus, O. Lat. for -d, 348 sq.

onustus, 306,

opera, 400,

-opere, Adv., 560.

operio, 475; for ap-, 18.

operon, 485.

opificina, 174.

opilio (up-), 34.

opinor, 472.

opiparus, 192.

Opiter, 180.

opitulor, 485.

opituma, 174.

oppido, 566,

opportunus, opor-, 114.

Optative (see Moods).

optimus, 406,

optio, 369.

opulens (-lentus), 352.

opus, 485.

oquoltod, S. C. Bacch., 227.

-or- from I.-Eur. r, 278 sq.

-or, 1 Sg. Pass., 533; Nom. 8g., e.g.
calor (see Suffix -S-),

Orata, as a nickname, 4o.

orbis, 239; orbs, 182,

orbus, 258.

orca, 239.

ordia prima for primord-, 362.

Ordinal Numerals (see Num.),

ordo, pronune. of, 141.

orede, 261,

orichalcum, 202,

oricla for auricula, 40 sq.

orno, 33, 310; pronune., 141.

-orum (-um), Gen. P]., 402.

-os, Nom. 8g., e.g. O. Lat. colos, 356.

os, mouth, 358.

os, bone, 405.

os- (aus-), e.g. osculor, 41.

Oscus, 310. nae

ostendo, 310.

ostium, 262 ; udt-, 34.

-dsus (-0ssus, -onssus), 112, 353.

OU, phon. changes of, 248 sqq.; OU,

 

THE LATIN LANGUAGE.

253; Lat. ow for eu, 245 sq.; on
inserr,, 246 ; written for %, 246; for
tt, 246,

-ov- for -uv-, 33.

ovifer, 361.

ovis, 235.

*ovum for dv-, 34.

-ou (see Suffix in Gutt.),

oxime, 407.

P, pronune. of, 78 sqq.; phon.
changes of, 281 sq.; dial. from q},

299 sq. ; not final, 77; bet. m-n, &c.,

0.

po ads (see Particles).

pacunt or pag-, xii Tabb., 465 ; paciscor,
465 ; ~peciscor, 200,

Pacuvius, use of Compds., 360.

paedora for -res, 356.

Paelignus, pronune. of, 138.

paene, 558 n.

paeninsula, 360.

paenula, 197.

Paenultima Accent.-Law, 160 sqq.

pagina, 326.

palam, 586.

Palatal Gutt., phon. changes of, 295
sqq. (see K, G, GH).

Palatalization of Cons. bef. y, 263 ;
of t, 81 sqq.; of c, 87 sq.; 2, 91, 98;
¥, OT, 98.

palea, 279.

palma, 279, 328.

palor, 586.

palpebra (-tra), 331.

palumbes, 346.

pdlus, -tis, the scansion, 214. *

pdlus, 293.

Panda, 318 ; -dna porta, 318.

pandiculor, 488.

pando, 472; -di, Perf., 502.

pango, 259; Perf. of, 504 ; panmi, 504 ;
pegi,.497; pepigi, 495.

panis, 339.

pannucia, 337.

panus, pannus, 117.

papae, accent. of, 164.

papaver, 541.

Papirius Crassus, his use of r for z, 6.

Parasitic Vowel, 145 sqq.; with J,
r, 93 sqq.; in Gk. loanwords, 70
sq.; u bef. J, 193 sq., 197 Sq.

parcarpus (?) for panc-, 273.
INDEX.

parcepromus, 360.

parco, Perf. of, 505; -rsi, 505.

parcus, 318.

parentatid, Luceria inscr., 519.

parentes, 465.

paret (-rr-), 117.

paricidas, 371, 373, 117 sq.

pario, 279; -tre, 475 ; -ibis, 492 8q.

paro, to equalize, 485.

pars, 278; partem, Adv., 555.

Participles, 539 sqq.; Pres. Act.,
540; Fut., 540; Perf., 541; Pres.
Pass., 540; Perf., 541 sq. ; -tus (-sus),
542; truncated, 543; used as Adj.,
5409 sq.

Particles (see Conjunctions); T-,
597; D-, 597; DH-, 597; P+, 597;
N-, 597; -ce, 432 sq.; relation to
Case-suffix, 597 sq. ; to Pers.-suff.,
597; Imper. used as, 600.

partim, 556.

parum, 562; parumper, 562.

parvus, 562; Compar., 406 sq.

pasco, pastum, 310. .

passar for -er, 201.

passim, 556.

Passive, 519 sqq.; Impers., use, 520
sq.) 523; With Acc., 521; Person-
endings (see Verb) ; Inf. (see Inf.) ;
Part. (see Participles).

pastillus, 339.

pateo, 476, 222.

pater, 222; p. familias, 381.

Patricoles, 197.

patrisso, 488.

patrocinor, 488.

patronus, 349.

patruelis, 340.

paucus, 243.

paveo, pavio, 473-

paulatim, 556;  -lisper,
spelling of, 111 sq.

pax, 259.

-pe, Particle, 597.

pecten, 349.

pectino, 488.

pecto, 479 5 ~«t, 505; Cut, 509.

pecu, 281 ; pecus, 354.

pedepressim, -temptim, 556.

peditaster, 330.

pedo, 307 5 pepedi, 501.

pejero, 48, 199 3 perj-, 313, 587.

pejor, pronune. of, 53-

562; -lum,

 

645

pelegrinus for per-, 93.

pellex for pael-, 115.

pello, 472 ; pepuli, 504.

pelluviae, 285.

pendeo, pendo, 473.

penes, Prep., 586.

penetro, 586, 594.

penitus, Adv., Adj., 561 sq.; -te, Adv.,
562,

penna, 313. 4

penuria, 558 1.

penus, 586.

per, e.g. paulisper, 562.

per, Prep., 586 sqq.

peragro, pronune. of, 94.

percello, 486; -culsus, 542.

percussi, Perf., 545.

peregre, 396.

peremne, I9l.

peremo, 587.

perendie, 560, 588, 192.

perendino (see comperendino).

eres for ped-, 8x.

perfacul, 198.

Perfect (see Tense-stems) ; quant.
of penult of S-perf., 134 sq. ; accent.
of -tt, &e., 163 5 scansion -izt, 214.

perfidus, 587.

pergo, 587.

periculum, scansion of, 175 sq.

perjero (see pej-).

Periphrastic Tenses, 510 sq.

pernd, 251.

pernix, 3543 pern-, 141.

pernox, 361, 555.

perperam, 557-

perplovere, 466.

perquam, 571.

Personal Pron. (see Pronouns).

Person-endings (see Verb).

perstroma, 172.

pertineo, 476.

pertisus, 195.

pervenat, 464.

pervicus, 317 8q.

pes, 286.

pesna (?), 313-

pessum, 539-

petesso, -isso, 462,

spetiolus (2), 76.

peto, 468; -ivi, 506.

petorritum, 300.

Petreius, 300.
646 THE LATIN

petulcus, 337-

phalerae, 190.

phasellus, -ll- for -l-, 115.

Phi (see Aspirates).

Philippus, accent. of, 155.

Pi, Lat. transer. of, 75.

piaculum, 333; O. Lat. -colom, 199.

pidato (pedatu), 19, 21.

pigmentum, pronune. of, 139.

*pilla for pila, 115.

pilleus (-l-), 117.

pilumnoe poploe, Carm. Sal., 368.

pinaria for pen-, 200,

pinguis, 292.

pinna, 229.

pinso, 471; ~t0, 470.

pirus, 374.

piscis, 232.

piscosus, 353.

Pitch-Accent, 148 sqq.

pituita, pronune. of, 52.

pius, 265; scansion, 131.

placenta, Igo.

plaga, 318.

Plancus, 179:

plango, 471.

plaudo, plodo, 41.

plaustrum, plostrum, 41.

Plautus, use of Compds., 362; dial.
Plotus, 242.

plebes, 376.

plecto, 486.

plenus, 324.

-pleo, to fill, 458, 473, 223.

pleoris (?), Carm. Arv., 408.

plerumque, plerus, Adj., 559.

plico, 200, 468.

Plinius, 225.

plisima (?), Carm. Sal., 408.

plodo, 41.

*plopus for populus, 98.

plostrum, 41.

pluo (plovo), 466,

Pluperfect (see Tense-stems).

plurimus, 408.

plus, 408, 558; -ra (-ria), 401.

pi-, Prep., 575, 588.

podex, 307.

poella for pu-, 37.

poend, 246 sq.

Poenus, 246.

poeta, 373.

pol, 618.

 

LANGUAGE.

polio, 575.

Polla, 41.

pollen, polenta, 367.

Pollio, spelling of, 112; -6, 212.

Pollux, 179, 182 ; -luces, O. Lat., 245.

polubrum, 575, 331-

pomerium, 588.

Pompeius, 300.

pono, 258.

pondus, 356.

pone, Prep., 588.

pono, 178 ; posut (posivt), 499 ; -s8-, I15.

Pontius, 300.

popina, 300.

populus, scansion of, 146; -loi Romanoi,
O. Lat., 387.

populus, in Romance, 98.

-por for puer, 183, 185.

por-, Prep., 590.

porca, 279.

porceo, 588,

porcus, 277.

porricio, 485.

porrigo, porgo, 545, 178.

porro (QO. Lat. porod), 568.

porrum, 279.

Porsenna, spelling of, 23.

posco, 477; Poposci, 501 ; pep-, $03.

Position, length by, 129 sqq.; bef.
Mute and Liq., 94, r29 sqq.

Possessive, Compd., 360 sq.; Pron.
(see Pron.).

possum, decl. of, 546; Inf. in Romance,
536 ; potestur, 522.

post, pos, po-, poste, postid, 588 sq. ;
postea, 569; posteac, 437; posthac,
569; postibi, 567; postmodum (-0),”
559-

posterior, N., 378.

posticus, 337.

postmeridianus, spelling of, 589.

postridie, 560.

Post-tonic, Syncope (see Syncope) ;
Vow.-change (see Weakening).

postulo, 179.

postumus (-remus), 407.

*poteca for apoth-, 107.

Potina, 349.

potior (-i0), 484, 546.

potis, 233.

potis (pote) sum, 546.

poto, 232.

potus, 520, 54a.
INDEX.

-pp- for p, 116,

prae, 589 ; prae let tremonti, Carm. Sal.,
530.

praebeo, scansion of, 143.

praecipes (-ceps), 182.

praeco, 180, 187.

praecos \-coguus), 358.

praeda, 143.

praedopiont, Carm. Sal., 189, 472.

praefamino, 517.

praefiscini, 1923 -ne, 396.

praemium, 143.

Praenestine, conea, 22, 106, 177 ; tam
modo, 613; Syne., 177; fibula, 188.

praes, 180, 187.

praesagio, 259, 486.

praesens, 589.

praesertim, 556.

praestigiae, 95.

praestino, 472.

praesto (-tu), 178.

praestolor (-tul-), 34.

praeter, 589 ; -ed, 569; -hac, 569.

praeterpropter, 521.

praetor, 350.

prandeo, -di, Perf., 502; -sus, 520, 542.

pratus for -m, 370.

precor, 296.

precula for perg-, 76, 97+

prehendo, 471, 42, 132; -di, 501 ; prendo,
57, 143; Perf. in Romance, 509.

prelum, 307.

premo, 307 ; -8si, 499.

Prepositions, 572 sqq.; written with
Noun, 168, 572; oust Case-suff.,
573 ; Compound, 573; with many
Cases, 574; assim. in Comp, Vb..
3l2 sq.; accent., 167 sqq.

Present (see Tense-stems).

Pretonic, Syncope (see Syncope) ;
Vowel-change (see Weakening) ;
6, 6 changed to @(?), 159, 222.

pri, Prep., 589.

pridie, 560.

primilegium for priv-, 52.
primordia, 362.

primus, 410; primumdum, 609.
princeps, 178; pronune., 141.
principio (-tum), Adv., 560.
prior, N., 378.

priscus, 337-

pristinus, 325.

pristris for pristis, 96.

 

647

privicloes, Carm. Sal., 403.

privignus, 181, 325; pronune. of, 138.

pro, Prep., 590; -3, 590.

pro, Interj., 618.

probunto, 519.

probus, 590 ; -bowm, 246,

procapis, 182.

procestria, Igl.

Procope, 107; (¢)st, 121,

procul, 590.

procus, 258.

prod-, Prep., 590; prodius, 590.

brodigium, 291.

prodinunt, 531.

produit, 515.

profestus, 199.

proficio, 590.

proficiscor, 480.

profitemino, 519.

profligare, 470, 486.

progenies, 345 sqq. ; progenie (Scip. Ep.),
48, 506.

progredimino, 519.

prohibeo, scansion of, 143; -bessis, 462.

proiecitad, Luceria inscr., 519.

proinde, 570; proin, 122.

proles, 345.

prolixus, 293.

prologus, 590.

promenervat, Carm. Sal., 194.

promiscam, 557.

prompsi, Perf., 505.

promunturium, spelling of, 197.

pronis for -nus, 339.

Pronominal, Adverbs, 567 sqq.; in
~bi, 567; -7, 567; -0, 568 sq. ; -d,
569; -im, 570; -nde, 570; -um,
570; -am, 570; T-suffix, 571; D-
suff., 571; Adjectives, 449 sqq.

Pronouns, 421 sqq.; Pers., Refl., 421
sqq.; Poss., 426 sqq.; Dem., 429
sqq.; Rel., Indef., Interrog., 443
sqq.; Decl. of (see Declension) ;
accent. of, 167 sq.; illéc, &e.,
163.

Pronunciation, 13 sqq.

pronus, 326; -nis, 339.

propages, 346.

propagmen, 292.

prope, 591 3 -modum (-0), 559.

Proper Names, Italic, 319 sq.

properus, 374.

propino, 590, 488.
648 THE LATIN

propinguus, 358.

propitius, 194.

propius for -pri-, 95.

propritim, 556.

propter, 591, 179 ; -ea, 569.

proptervus, 590.

prorsus (-m), 5533 cf. 549.

prosa, 553-

Proserpina, 98; Prosepnai,
382, 184.

Prosodical Hiatus, 132, 144 sq., 209
sq.

prosperus, 257.

prosternere, in Romance, 489.

Prosthetic i, 102, 105 sqq.

prosum, Vb., 590.

protervus, 590.

protinus (-tenus), 554, 556, 200; -am,
5573 ~t8, 554.

protulum for prothyr-, 190.

proximus, 591.

prurio, 487.

*psalli, Perf. (?), 50x.

-pse, -pte, Particles, 440.

puber, 356.

pubes, 346.

‘Publicola (Popli-), 76.

publicus, 287.

Publius, 287.

puer, -re, Voc., Fem. in O. Lat., 374;
-por in Compd., 183, 185.

-puertia (-rit-), 174.

pugil, 376.

pulcer, spelling of, 12, 59 sq.

puleium (-egi-), 48, 292.

pulenta for pol-, 33.

pulex, 355.

-puli of impuli, &e., 501 ; pulsus, 278.

pulto (-so), 482.

pulvis, 235.

piimex for piim-, 37.

pumilio (pom-), 34.

pungo, 471; pepugi,
pronune., 140.

pupa (-pp-), 116.

purus, 542; purime, 407.

pusillus, 305.

puter, 260; puteo, 260.

puto, 482 ; -td, 211, 600.

putrefacio, scansion of, 212,

plitus, 542.

(not -ais)

503; punctus,

Q, the letter, 3, 7; qu, pronune.,

 

LANGUAGE.

84 sqq.; length by Position, 87;
I.-Eur. Q3, phon. changes of, 299
sqq.

qato-, qiti-, qtu-, Pron. stems, 443
sqq.

qoi (Dvenos inser.), 445.

qua, Adv. (quaad), 569 sq.

quadra, 413.

quadraginta, 417 sq.3
quarr-, 418.

quadrans, 409.

quadrigae, 196.

quadrigenti (-ing-), 419.

quadruplex (-plus), 414.

quaequalis, 448 n.

quaero, 487 ; -sivi, 506.

quaeso, 462, 487; -wmus (-imus), 487
-8s-, II2.

qualis, 451.

quallus, the spelling, 112.

quam, 570, 607; of unquam, &e., 552;
quamde, 570, 607; -libet, 613; -vis,
613; -obrem, 606 ; quamquam, 613.

quando, 608, 571; -0, 213; -que, 571;
“quidem, 571, 608; quandoc, 608;
quandone, 608.

quansei, 607.

Quantity, 126 sqq.; overmastered by
accent, 129 (see Shortening,
Lengthening, Scansion); changed
bef. cons.-group, 133 sqq.; of vow
in close syll., 133 sqq.

quantus, 451 5 -mvis, 613.

quapropter, 569.

quare, 606.

quartus, 413.

quasi, 607 ; -se, 25,

quasillus, 305.

quatenus, 557.

quater, 413 5 -nus, 414.

quatio, u in Comp., 196.

quattuor, 413 sq.; spelling, 414, 113;
-ttor, 414; -decim, 416.

que, 598 sq. ; enclitic, 166 sq.; qu(e),
598.

queistores, 242.

quercus, 29I ; -nus, 294.

querella, the spelling, 112 sqq.

queror, 227.

querquera, 315.

quetus for quie-, 142 sq.

-qui- for Gk. «v, 36.

qui, Pron., 443 sqq.; pronune., 39,

accent., 165;
INDEX.

443 cud, pronune., 39, 44, 4463;
spelling, 87.

qui, Adv., 446, 568; -qui (-n), 613;
-dum, 609.

quia, 610; -nam, 606,

quicumque, 448 ; -dam, -libet, -vis, 447.

quidem, 6o2; siq-, ttig-, &e., 216.

quies, 182.

quiesco, pronunce. of, 134; quetus for
quie-, 142 sq.

quin, 613.

quinque, 414, 229; pronune., 414;
-ndecim, 416 ; -n(c)tus, 70 ; pronunce.,
140, 4143; -ngenti, 419; -ncentum,
O. Lat. 419; -nquaginta, 417;
-NCUNL, 300; -NUS, 414.

quippe, 604; pronune., 122.

quirquir, 288.

quis, 443 Sqqd.; pronunc., 85; -que,
~piam, -quam, -quis, 448.

quisquiliae, 315.

quo, Adv., 568; Conj., 613; -ad,
568 ; -circa, 580 ; -modo, -6, 212.

quod, Conj., 610.

quoiatis (cujas), 447.

quoiei, Dat., 445.

quoiguoimodi, 445.

quom for cum, Prep., 581.

quominus, 569, 613.

-quomque (-cungue), 598.

quondam, 571.

quoniam, 610.

quoque, 598.

quorsus (-m), 568.

Quorta, 413.

quot, 451; -annis, 560 ; -idie (see colt-).

quotumus, 561.

quotus, 451.

quum, 608, 570.

R, pronune. of, 89 sqq. ; I.-Eur. R,
phon. changes of, 276 sqq.; R, 278
sq.; Lat. r for s, 303 sqq., ror,
105; for d, 285, 288, 80 sqq.; for J,
92 sq.; for n, 96; dropped in prae-
st(r)igiae, &e., 91,95; bef. cons., 97.

-r, long vow. shortened bef., 213 sq. ;
of Passive, 523, 533-

rabies, 347; -es, Gen., 383.

rabo for arrabo, 17'].

rabula, 177.

racemus, 306.

radix, 220.

 

649

Raius for Rav-, 252.

ramentum, 312.

rapio, 476.

rapo, a robber, 475.

ratio, 340.

ratiocinor, 488.

ratis, 307.

ratus, 259.

ravastellus, 330.

raucus, 180.

Raudus for -vid-, 185.

raudus, 248 sq.

-re, 2Sg. Pass., 533; Inf. Act., 535
sqq.

re-, red-, Prep., 591.

reccidi, Perf., 504.

recens, Adv., 554.

recidivus, 322.

reciprocus, 337.

Recomposition, 199 sq. (see Re-
formation).

recordor, 483.

recta, Adv., 550.

recupero, 488.

reddo, 468, 114; reddibo, 493.

red(d)uco, 114; redux, 591.

rederguo for -arg-, 198.

redinunt, 531.

redivia, 286.

redivivus, 591.

Reduction (see Weakening, Shorten-
ing).

reduncus, 259.

Reduplicated, Present, 468 ; Perfect,
496 sq., 502 sqq-; form of red.,
502 sq.; assim. of red. vow. to
stem-vow., 503; loss of, 503 8q-;
Noun, 358, 363.

Reflexive Pron. (see Pronouns).

refriva faba, 178.

regind, 370.

regnum, pronune. of, 138.

rego, 296; rexi, 505; pronune. of, 139,
498.

regula, 318.

Relative Pron. (see Pronouns).

reliquus, pronunc. of, 323;
O. Lat., 46.

reluvium, 286.

remex, 358.

remulcum, pronune. of, 142.

ren, 264.

reor, ratus, 259.

-cuos,
650 THE LATIN

repandirostro-, Pacuy., 360.

répens, 268; Adv., 5543; -nte, 551.

repo, 307.

repperi, Perf., 504.

reppuli, Perf., 504.

repulsa, 542.

requies, decl. of, 346.

res, 252, 225; in word-group, 169.

reses, 358.

respondeo, Perf. in Romance, 509.

Res(ti)tutus, 176.

retro(r)sum, 549.

rettuli, Perf., 503 sq.

retus for retitus, 543.

reverti, Perf. of, -tor, 520.

1eUs, 244.

rex, 260, 276.

Rho, Lat. transer. of, 12, 59.

-ri-, syncop. after cons., 171, 179 8q.

-ri, Inf. Pass., 535 sqq.; -rier, 536 sq.

rien, 264.

rigor, 306.

ringor, 471.

-ris, 2 Sg. Pass., 533 ; and -er in Adj.
M., F., 371.

rite, 560.

ritus, 560.

rivalis, 244.

Rivers, gender of, 368.

rius for rivus, 52.

-rl-, pronune., of, 97.

-rm- for nm, 271 sqq.

-ro, e. g. lambero, 479.

robigo, 348 ; rub-, 34.

robur (-0r), 356, 379; -US; 379.

robus, Adj., 248.

rodus, 248 sq.

Roma, 307.

ropio, 76.

Rostrata Columna (see Col. Rostv.).

rota, 258.

rotundus, 544 sq.

Rough Breathing (Gk.) (see Breath-
ing).

-rr- for rs, 277 Sq.

-rs-, pronunce. of, gt, 96.

rubeo, 4.76.

ruber, 239.

rubicundus, 545,

rubigo for rob-, 34.

rubus, 307.

rubustus for rob-, 34.

ructo (see eructo).

 

LANGUAGE,

rudentes, 467.

rudis, 249, 338.

rudo, 307.

rudus, 248 sq.

rues, 345.

rufus, 248,

Ruga (see Carvilius).

rumen, 307.

rumentum, ‘ abruptio,’ 314.

Rumina ficus, Rumon, 307.

rumpia, 33.

vumpo, 4713 rupi, 502; -mptus (see
corruptus).

rupes, 346; -pp-, 118.

rurt (-€), 396.

rursus (-m), (7rus-), 549, 553-

-rus, 2 Sg. Pass., 534.

Rustic Latin, & for %, 19, 25, 30;
veha, 22; @ for 7, 24 sq., 29 8q.;
frundes, 31, 33; tundo, 33; 0 for au,
40 sq.; e for ae, 42 sq.

rusticus, 337-

8, voiced written z, then r, 6; pro-
nunc. of, tor sqq.; phon. changes
of, 303 sqq.; Lat. s for ss, 305 sq.,
109 sqq.; for ms, 1368q.; for th in
Not. Tir., 58; O. Lat. s, class. 7,
305; dial. s, 305.

8-Perf., Vb.-forms (see Tense-stems).

s- from ps-, ks-, 303.

-s, O. Lat., e.g. colos, 356; 2 Sg.,
525.

sabulum, 304.

sacena (scena), 261, 184.

sacerdos, 179 8q.

sticri-, sticro-, 183.

sactus for sanct-, Jo.

saepe, 559-

saepes (se-), 42.

Saeturni, 242.

Saguntum, 104.

sagus, 259 sq.

sal, 223.

Saliare Carmen (see Carmen Sal.).

salignus, 293.

salio, 223.

salix, 278, 223.

sallo, 285, 479.

saltem, 559.

*salvaticus for silv-, 201.

sumbucus (sab-), 65.

Sanates, 183.
INDEX.

sancio, 470; sanctus, pronune., 140;
sact-, sant-, 70.

Sandhi, 120 sqq.

sanguis (-en), 377.

saplutus, 104.

sarcing, 320.

sarmentum, 310.

satis (sat), 558; satin, 558n.

satur, 5587.

satura (-ira), 197.

Saturnian Metre, 128 7n., 132 ”., 159.

satus, 222.

saucius for -atus, 543.

savium, 268.

¥sauma for sagma, 89.

saxum, 261.

sc- lengthens final vow., 131; isc-
(see Prosthetic 2).

scabellum (-mill-), 283.

scabo, 223, 281, 259; -bi, Perf., 502.

scaena (sce-), 42.

scaeptrum (sce-), 42.

SCHEvUS, 242.

Scaliger’s Law, 361, 363, 365.

scalpo, 279.

sealprum, 333.

scamnum, 283.

scando, -di, Perf., 502.

Scansion, traditional, 1277. ; errors
in late literature, 128 sq. (see
Shortening).

Scaptensula, the spelling, 136 n.

scateo (-to), 476.

scaturio, 482,

scauria for sco-, 41.

scelerus (?), 356.

scelus, 229.

scena, a priest’s knife, 184, 261.

scend, (kNVH), scue , 42.

scheda, sc(h)i-, 23.

scilicet, 564.

scindo, 280, 471; scicidi, 495, 501;
scisc- (?), 503; scidi, 495, 501.

Scipio Afr. (Min.), ve- for vo-, 228.

sclis (stl-), 307, 83.

Scloppus, 307.

-sco, Incept., 477.

scobis, 259, 338.

scopulus, 197.

scoriscus for corusc-, 29.

scriba, 318.

seribo, 282.

scrobis, 306.

 

651

scrofa, 80.

scerupulus (scrip-), 29.

sculna, 184.

sculpo, 279.

86-, ‘to sow,’ 224.

se, Pron., 424.

se (sed), Prep., 592; se fraude, 592.

secespita, 261.

secius (see set-).

seco, 298; si-, 23.

Second Pers., Sg. ending, 525 sq.;
Pl., 529.

Secondary Accent, 158 sqq., 161.

secordis, 592.

secratum for sacr-, 18.

secta, 566.

sectius (?), 566.

secundum, Prep., 591.

secundus, 411.

securus, 592.

secus, Noun, 552; Adj., 591; Adv.,
552; Prep., 591.

*SCCUS, 552, 554-

sed, O. Lat., ‘ himself,’ 424.

sed, Conj., 6or.

sedda for -ll-, 287.

sedecim (sexd-), 416.

sedeo, 285 sq., 473; Sedt, 497, 502.

sedes, 345 8q.

seditio, 592.

sedulo, 563; -us, Adj., 563.

sedum, 601.

sedutraque, 450.

seges, 351.

segmentum, 293.

selinum, the scansion, 156.

seliquastra, 287.

sella, 287.

semel, 410, 229.

semermis, 364.

semi-, 409, 225.

se(mi)modius, 176.

semissi-, 409, 586.

semper, -iternus, 562.

semptem for sept-, 66.

senatus, decl. of, 343; -ti, Gen., 380,
384 ; -tuos, Gen., 384.

seneca, 337; senecio, 337.

senecta, 334 n.

senex, 271, 354; decl. of, 367.

Sentence-Accent, 148 n., 165 sqq.

sententia, 352.

senus, 415.
652

seorsum, 592.

septem, 4153; -decim, 416; -tuaginta,
417 sq. ; -tingenti, 419; -timus, 415;
-tenus, 415; -tuennis, 415; sempl-, 66.

septentriones, 269.

Septidoniwm for Septiz-, 104.

septimus decimus, accent of, 163.

sepulerum, 334; spelling of, 57, 59 sq.

sequius, 566.

sequor, 520.

Serena for Sir-, 30.

serius, Adj., 267.

serpillum, 197.

serpo, 277.

serra for sera, 115.

sero, 468 ; sevi, 500; sctus, 222.

sesamum, spelling of, 198 sq.

sescenti, 419.

sescentoplagus, 364.

Sescuncia, 300 ; -onc-, 236.

sesqui-, 409.

sesse for sese, 112.

sestertius, 409 ; -ium, Gen. Pl., 402; 418.

setius, 566. .

Selus (ZijO0s), 104.

seu, sive, 122,

severus, 226 sq.

SEX, 4155 sexaginta, 417.

sextans, 409.

Seatius (-st-), 415.

Shva, 257.

Short Vowel (see Quantity).

Shortening, of vow. bef. vow., 131
sqq.; bef. n with cons., 141 sq. ;
bef. 7 with cons., 142; bef. -1, -m,
-7, -t, 213 sq.; of final vow., 207
sqq.; of final syll. long by position,
215 sq.; of monosyll., 215 sq. ; after
short syll. (see Breves Breviantes) ;
of long diphth. bef. cons., 25%sq.

si, 610. ee

sibi, 4243 -€, 25.

Sibilants, phon. changesiof, 302 sqq.

sibilo, 78; -f-, 80 ; su-, 30. *,

sic, 567; pronune., 121 sq. ; seine, 433.

siccus, 447.

Sicilicus, mark of double cons., 4, 8.

sicubi, 446.

sido, 468 ; sidi (sedi), 503.

sidus, 267.

stem, 514 sq.

*sifilo for sib-, 30, 78, 80.

signum, pronune. of, 138 sg.

 

THE LATIN LANGUAGE.

silenta for -tia, Neut. Pl., Laev.,4o1, 352.
silicernium, 287.

-sim, Subj., 465.

simila, 286.

similis, 338.

simitu (-tur), 565.

simplex (-plus), 410.
simpludiarea funera, 410.
simul, Adv., 553; Prep., 592.
simulter, 553.

simus for sumus, 29.

sin, 612,

sinatus for sen-, 200.

sincinia, 410.

sinciput, I41.

sine, 5923 sei-, 592.

Single Cons. for Double, 113 sqq.
singnifer for sign-, 66.

singuli, 410.

sino, 471.

sins, Carm. Arv., 518 7.
siquidem, the scansion, 216.
sis for si vis, 52; sultis, Plur., 181.
siremps(e), 566.

sisto, 468; Perf. of, 503.

sive (seu), 122, 600.

-so, Vb.-forms in, 462 sqq.
so- (se-), ‘this,’ 430.

so- for suo-, Poss., 426 sqq.
so-, O. Lat. Demonstr., 430, 432.
sobrinus, 303.

sobrius, 592.

soc (?), ‘ita,’ 432.

socer, 192, 227; -erus, 374.
sociennus, 545.

sociofraudus (-uf-), 364.

socius, 262.

socors, 592.

SOCrus, 3445 -, 343.

sodes, 265, 486.

Sofia, scansion of, 150.

sol, the scansion, 215.

soldus for -lid-, 185.

solea, 287.

solemnis for -ll-, 111.

solerare, ‘ solidare,’ 288.
solinunt, 531.

solitaurilia, 8.

solitus sum, Perf. Dep., 522.
solium, 287.

sollemnis (-nn-),70 ; -mypn-, 70 ; sole- 111.
sollicitus, 361.

sollistumus, 407.
INDEX.

sollo-, 8, 16,

sollox, 354.

solum, 287.

solvo, 592; -lui for -lvi, 48; -lutus, 260.

solus, 449.

somnium, 319.

somnolentus, spelling of, 192.

somnus, 227.

Sonant L, M, N, R (see L, M, N, R).

sonivius, 323.

sono, 488 n.; sonui, 506.

sopor, 227).

sorbeo, 283; -psi for -bui, 509; -billo,
487.

sordére, -ére, 476.

sordes, 345.

soror, 2277.

sortus, P. P. P., 542.

sp- lengthens final vow., 131; isp-
(see Prosthetic 2).

sparsi, Perf., 505.

spatiarus, 534.

species, decl. of, 346.

specio, 472; Spi-, 23.

sperno, 472.

Spes, 257, 345 2.

spica, spe-, 25-

Spiritus (see Breathing).

spondeo, 482;  spoponii, 496; spe-

(sposp- ?), 503.

spongia (-ea), 22.

spuo, 264.

-ss- for tt, 304, 309; for x, 102; re-
duced to s, 110 sqq.

-ssere, Inf., 465.

-(s)sim, Subj., 465.

-(s)so, Vb.-forms in,. 462 sqq.; Perf.
of, 506.

sta-, ‘to stand,’ 457.

Staccato pronunciation, 131.

stagnum, pronune. of, 138.

sttitim, 556; ste-, 15.

Statina, 349.

statod (Dven. inser.), 519.

status, 221; as P. P. P. of sum, 542.

ste for iste, 435, 167.

stelio (-ll-), 117.

stella, pronune. of, 112.

Stem, Noun and Adj., 316 sqq.;
suffix (see Suffixes) ; interchange of
U- and O- stems, 343 sq.

sterilus, 338.

sterno, 470, 219; stratus, 219, 306.

 

653

sternuo, 471.

sterquilinium, the spelling, 301.

stetim for stat-, 15.

-sti, 2 Sg. Perf., 525.

stilicidium, spelling of, 112.

stilla, 487.

-stinguo, 471.

stipendium, 116,

stircus for -erc-, 20, 229.

stl-,O. Lat., Stlaborius,Stlaccius, stlembus,
stlis, stlocus, stloppus, 307.

stlattarius, 219.

stlis (scl-), 83, 307.

sto, decl. of, 457; steti, 501; Perf. in
Romance, 509; status, 221, 542.

stolidus, 235.

storax, 37.

siramen, 279; stratus, 219.

strenna for -n-, 116.

strenuus, 323 ; -nn-, 113.

Stress-Accent, 148 sqq.

strictim, 556.

stridi, Perf., 502.

stringo, 229; strictus, 542.

stritavus, O. Lat. for irit-, 196.

Strong, Cases, 367; Root-grade (see
Gradation).

strues, 345.

strufertarii, 361.

struo, in Romanee, 489.

struppus (-opp-), 58.

studeo, 476.

stultus, 235.

stupa (-pp-), 116.

stupila for stipula, 37.

suad, ‘sic,’ 611.

suadela, spelling of, 115.

suadeo, 482, 259; pronunce., 53.

Svarabhaktic Vowel (see Parasitic).

suavis, 221; pronunc., 53.

suavisaviatio, 361.

sub, sus-, Prep., 593; sub vos placo,
572.

subaediani, spelling of, 195.

subaxet (?), 508.

*subilo for stb-, 30.

Subjunctive (see Moods).

sublimen, 5593; -Us, 338.

suboles, 345.

Subordinate Words, 165 sqq.

subrigo, 545.

subrimii haedi, 193.

subsicivus, 323.
654 THE LATIN

subtel, 199, 213.

subtemen (-egm-), 292, 70.

subter, 593 ; -tus, 561; Prep., 593.

subverbustus, 306.

subula, 334.

succidaneae porcae, 195.

sucerdae, 260.

sucus, 76; spelling of, 116.

sudus, 307.

suesco, 481.

suffio, 267.

Suffixes (Noun, Adj.), 316 sqq. ; -O-,
-A-, 316 sqq.; -I0-, 318 sqq. ; -UO-,
322 sqq.; -NO-, 324 sqq.; -MENO-,
327 sq.; -MO-, 328; -TEMO-, 328
-RO-, 328 sqq.; -TERO-, 329 sq. ;
-cro-, 329; -TRO-, 329 sqq. ; -DHRO-,
329 sqq.; -LO-, 331 sqq.; -TLO-,
332 sqq.; -DHLO-, 332 sqq. ; -TO-,
334 sqq.; -KO-, 336 sqq.; -I-, 338
sqq.; -NI-, 339; -MI-, 339; -RI-,
-LI-, 339 sq.; -TI-, 340 sqq.;
-TION-, 340 sqq.; -TAT(I)-, 341
sq.; -TUT(I)-, 341 ; -U-, 342 saq. 5
-TU-, 343 sq.; -YE-, -I-, -E-, 344
sqq-; -EN-, 348 sq.; -YEN-, 348
sq.; -WEN-, 348 sq.; -MEN-, 348
sq.; -R-, 349 sq.; -ER-, -TER-,
350; -T-, 350 sqq.; -NT-, 352;
-WENT., 352 sq.; -D-, 353 sq;
Gutt., 354 sq. ; -S-, 355 saq.3 -ES-,
355 sqq.; -YES-, 357; -issa, 365 ;
-ia, 365; -itto-, 365.

Suffixless Stems (Noun. Adj.), 357 sq.

*sufilo for sib-, 30.

sugo, 76.

Suila, Sy-, 29, 36.

sultis, for si vultis, 181.

sum, 237; decl. of, 455 sqq.; enclitic,
167; Perf., 545; es, Imper., 518;
sim, 5143; simus for stimus, 29; ero,
492; eram, 490; forem, 227, 545%.
(e)st, Procope, 121.

summosses, Hor., 507.

summus, 4073; -opere, 362.

sumo, -psi, 505; suremit,
surenrpsit, 593-

suo, 264, 484 n.

suovitaurilia, 361.

supellex, decl. of, 367 ; -erl-,97 ;-pp-, 118.

super, Prep., 593; Adj. (-rus), 374.

superbus, 590.

supercilium, 195.

595; 593;

 

LANGUAGE.

Superlative Degree (see Compari-
son); ¢ for «win, 189.

superne, 593-

superus (-per), 374.

Supines, 538 sq.

supparum (sip-) 29.

Suppression of Syl, e. g. ar(ci)cubti,
176 sq.

supra, 593; -perd, 181, 593.

suremit, 505, 593; -Psit, 593.

surgo, 178; surrext, 545, 505; suregtt,
505; sortus, 542. ‘

surput, Perf., 178; surptus, 178.

sursum (-8), 549, 553-

-sus (-tus), P. P. P., 542.

sus-, Prep. (see sub); susgue deque, 593.

sus, SOW, 260,

suspicto, 225.

sustuli, Perf., 545. :

suus, 426 sqq.; monosyll., 426 sqq.;
O. Lat. so-, 426 sqq.

swé-, phon. change of, 227.

Sylla, 29.

Syllable, Close (see Close Syll.);
Division, 124 sqq.; First (see First
Syll.); Suppression of (see Sup-
pression ; Syncope); Syllabie Writ-
ing, 12, 177.

sylva, the spelling, 11, 29.

Syncope, 170 sqq., 150 sqq.; Prae-
nestine, 177; by old Accent-Law,
178 sqq.; final syll., 181 sq.; pre-
tonic, 183 sq.; post-tonic, 184 sq.;
final vow. (see Apocope) ; vow. in
final syll., 203 sqq.

Synizesis, 142 sqq.

T, pronune. of, 80 sqq. ; and d, final,
76 sq.; phon. changes of, 283 sq. ;
for d bef. r, 285, 289.

T-particle (see Particles).

-t, 3 Sg. 526 sqq.; vow. shortened
bef., 214.

-ta of tta, &e., 552.

tabes, 346.

Tables, Twelve (see Twelve).

taceo, 476.

taeter, 289.

talis, 451%.

talus, 293,

tum, 570; -me, Carm. Sal., 570; -elsi,
602 ; tanne, 602, 69.

tamen, 601 3; -etsi, 602.
INDEX.

tango, 471; tago, 464; tetigi, 495.

tanquam, 570°.

tantus, 451; -tidem, 571; -tisper, 562;
-tummodo, 564.

Tarentum, accent of, 155, 197.

tarpessita, 104.

tata, 118 n., 363.

Tatpurusha, Comp., 360 sq.

*axitare, 482.

-te, 2 Pl. Imper., 529.

tector, pronune. of, 139; -tum, 542.

Tecumessa, 64, 71, 1458q.

ted, 423.

teges, 351.

tego, 303; text, 505; pronune. of, 139.

tequila, 318.

tela, 293.

Telesia, 287.

Telis for Thetis, 75.

telum, 293; -Ul-, I12.

-tem of item, &e.4 552.

temere, 560,

temperi, spelling of, 192, 356.

temperies, 344.

temperint (?), 515.

tempestas, 342; -tus, O. Lat., 342.

templum, 565.

Tempsa, Temese, 181.

tendo, 486 ; -sus, 542.

tenebrae, 270.

leneo, 476 ; -tus, 542; tetini, 501.

tenor, tenus, 355.

Tense-Stems, 459 sqq.; Aor. and S-,
459 sqq.; Pres., (1) Them. E-grade,
466 sqq.; (2) Redupl., 468; (3)
Nasal, 469 sqq.; (4) YO-, 472 sqq.
(5) Inceptives, 476 sq., 479 sqq-:
(6) Causatives, 477, 482 sq.; (7)
Desideratives, 478, 482, 484; (8)
Iteratives, 478, 482 sq.; (9) Deri-
vatives, 478, 483 sqq.; Stem-suf-
fixes, 478 sq., 486 sqq.; Impft. (Ind.
Subj.), 489 sqq.; Fut., 491 sqq.;
Perf. (Ind. Subj.) 494 sqq.; (1)
Redupl.; 496 sq., 501 sqq.; (2) Un-
redupl., 497, 501 sqq.; (3) S-, 497
sqq., 505, 508; (4) V-, 499 sq.,
505 sqq.; (5) irreg., 509; Plupft.
(Ind. Subj.}, 509 sq.; Tenses with
Auxil., 510 sq.

tento for -mpt-, 70.

Tenues, (Lat.) pronunce. of, 71 sqq. ;
(Gk.) Lat. transer. of, 74 sq.;

 

655

(I.-Eur.) phon. changes of, 279
sqq.; Ten. Asp., 280, 308.

tenuis, 274; pronune., 46, 174; -vid, 144

tenus, Prep., 593; hactenus, &c., 569.

tenus, Noun, 355.

-ter, AdV., 549, 553.

ter, 412; scansion, 119; ternus, 412.

terebra, 331.

teres, 351.

tergus, 302.

-terior (see Comparison).

terminus, 269 ; termo, 327.

tero, terui for trivi, 509.

terreo, 481.

terrimotium, 362.

terruncius, 412.

tertius, 412; -o (-wm), Adv. 559.

tesca (-qua), 337°

tesera for -ss-, 115.

testamentum, 277.

tetini, Perf., 501.

tetuli, Perf., 494, 497.

-th- for s, 58.

Thelis, for Thetis, 286,

Thematic, Conj. (see Verb); Vowel,
453.

thensaurus for thes-, 69, 1367.

Theta (see Aspirates).

Third Pers., Sg. ending, 526 sqq.;
PL, 529 sqq.

-ti- and -ci- confused, 82 sqq.

tibi, spelling of, 423 (see tu).

tibicen, 364.

-ticus (see Suffix -KO-).

tilia, 225.

-tim, Adv., 548.

tingo, 225; tinctus, pronunc., 140.

tinnio, pronune. of, 118; tintinnio (-no),
483.

-tinus, Adj. (see Suffix -NO-); Adv.
(see tenus).

-tis, 2 Pl., 529.

tis, Gen., 423.

-tiwus (see Suffix -UO-).

Tmesis, 187, 573.

-t. Fut, Imper., 516.

to-, ‘this,’ 430.

toga, 255.

Toitesiai, Dvenos inscr., 305.

tolero, 488.

tollo, sustuli, 545.

tolutim, 556.

tondeo, 486; totondi, 496.
656

Tone-Accent, 148 sqq.

tongere, 259.

tongiliatim, 556.

lopper, 562.

-tor, Imper. Pass., 533 sq.

torculus, 300.

-toriwm (see Suffix -I0-).

tormentum, 310.

torpedo (-ido), 23

torpeo, 476.

torqueo, 482; torsi, 310; tortus, 310.

torreo, 4777.

torris, 339.

tot, 451 ; totidem, 451, 571-

-tote, Fut. Imper., 517.

totus, 450; -tt-, 116.

totus, 451-

-tr- for dr, 81, 289.

-tra, Adv., 569.

tra-, ‘to go through,’ 458.

trabes (-bs), 376.

trado, -nsd-, 594.

trahea for -ha, 318.

traho, in Romance, 489.

tramitto, -nsm-, 594.

trans, tra-, Prep., 594.

transmarinus, 362.

transtineo, 4.76.

Trasumennus for Tars-, 97.

Trees, gender of, 368.

tremo, 499 ; tremebit, 494 ; tremonti, 530.

tres, 412 sq.; tré-, tri-, 412; tredecim,
416; trecenti (-wm), 418 sq. ; triginta,
417 8q.; accent., 165; trienta for trigi-,
418 ; tricenus (-g-), 418 ; tricies (-g-),
418; trinus, 412; triplex (-us), 412.

tribunal(e), 205.

tricae, 58, 116.

tric(h)ilinium for tricl-, 94.

triens, 409.

triginta dies, accent. of, 169.

tripodare, 256.

*trippa, T19.

tristus for -is, 368.

tritavus (strit-), 196.

triumphus, spelling of, 59.

-tro, Adv., 568.

Troja, pronune. of, 53.

*trono for tono, 95.

trudis, 338.

trudo, 486.

-trum, (see Suffix -TRO-) ; with é, 191.

-tt- for ct, 86, 89; by Syne, 284.

 

THE LATIN LANGUAGE.

tu, decl. of, 423 sq.

-tu, 2 Sup., 538 sq.

tuber, 270.

tuburcinor, 488.

tudes, 351.

-tudo (see Suffix -TUT(I)-, -D-).

tueor (-2t07r), 476.

-tuiri, Fut. Inf. Pass., 538.

-tulas, attulas, 464.

tuli (see fer).

-tum, I Sup., 538 sq.

tum, 570; tunc, 570.

tumba, 36.

tundo, tutudi, 496; tunsi (?), 471 3 tun-
sus, tusus, 471; tundo for tondeo, 33.

-tur, 3 Sg. Pass., 534.

turba, 239.

turbinés, the scansion, 399.

turdus, 308.

-turio, Desider., 478.

turnus for to-, 31, 33.

-turo-, Fut. Inf., 537; Part., 540.

turtur, 363.

-tus, -sus, P. P. P., 542; of funditus,
&e., 548.

tus, 58.

Tuscus, 278.

tuté, Pron., 423 sq.

tutela, spelling of, 115.

tuus, 426 sqq.; monosyll., 426 sqq.

Twelve Tables, 7, 565.

tympanum (typ-), 272.

Tyrannio, taught Accent., 151, 154.

U, V, the letter, 3,7 sq.; Vand VV,
267 sq., 52; wu for %, 10; Gk. (see
Upsilon) ; U, pronune., 34 sqq.; V,
44 sqq.; ii-sound, 25 sqq.; I.-Eur.
U, phon. changes of, 237 sqq., 260 .
sq.; V \W), 265 sqq.; Lat. % for o,
235 sqq., 31 sqq.; for atonic vé-,
196; in weak-grade of root with
labiovel., 239 ; for atonic vow. bef.
lab.,, 192 sqq.; for 6, 33 sq.; Lat. a
for 6, 233; for % lengthened, 237;
for eu, 245 sq.; for ou, 248 sqq. ;
Lat. v from g¥, 301 sq.; for b, 47,
49 sqq.; dropped bet. vowels, 52;
dropped after cons., 52 sq., 144.

-d, shortening of, 213; of cornu, &e.,
scansion, 377 sq.

a-t and i-t, e.g. stupila, 37.

V-perfect (see Tense-stems).

©
INDEX.

U-stems, blend with O-stems, 343
sq. (see Suffix -U-).

vddo, 467.

cadum, 467.

vae, 618.

valde, 561 ; -lid-, 174.

Valeri (Voe., Gen.), accent. of, 164.

vapor, 299.

vapulo, 522.

Variation of Vowels (see Gradation).

varir, 279, 355+

Varro, i and ei, 9; deriv. of Gracchus,
1477. ; On divus, 244.

vas, a surety, 2g0.

vas, a vessel; vassa, Pl., 112.

vates, 346 sq., 221; O. Lat. vatius, 375.

vatillum, spelling of, 51.

ubba for obba, 33.

uber, 290, 250.

ubi, 567 5 -e, 25.

-ubris, e.g. lugubris, 196.

-uc, Adv., 551.

-ticus (see Suffix -KO-).

udus, 180.

-ve, Conj , 599; enclitic, 166 sq.

ve- for vo-, 228,

ve-, Prefix, 615.

veclus for -tul-, 83

vectis, 341.

vegeo, 482, 206.

vegetus, 296, 335.

veha for via, 22.

vehemens, ve-, 54, 57; -ter, Adv., 554.

veho, 226; -xt, 499.

rel, 599 84.

Velar Gutt. (see Gutt. Proper, Labio-
velar).

velatura, 308.

velim (see volo).

vella for vi-, 29.

vellico, 488.

vello, older vo- (?), 228; -li, 502; vulsi, 509.

vellus, 266

velox, 354.

veltrahus for -rtrag-, 93.

velum, 333 ; -ll-, 112.

vendo, 472, 488; -ditus,in Romance, 542.

veneficus, bene-, 51.

venenum, 326.

veneo, 488, 5393 -iri, 522.

venio, 473; vent, 500; Perf.in Romance,
509 ; -venat, 464; -ventio, 274.

ventus, 251.

 

657

venum, 539; -cdo, 472.

Venus, 356.

vepres, 346. :

Verb, 453 sqq. (see Tenses, Moods,
Pass., Dep., Inf., Part., Sup., Ger.) ;
Conjugations, 454; Athem. Conj.,
453 sqq.; Them. Conj., 453 sqq.;
Contracted forms, 463 sq.; Person-
endings, 522 sqq.; 3 Conj. Vbs. in
-to, 475; atonic form of Simple Vb.,
468 ; in -o, -e0, 476; -urrio, 482; -uo,
484 7.; Onomat:, 483; Compound,
485; in -uttio, 488; -cinor, 488 ; -isso,
488: 1 Conj. predominant, 488;
Noun-stem in -@, 490; Irreg.,
545 sqq.; Contracted Perf.-forms,
V-perf., 506 sqq. ; S-perf., 508.

verberit (2), 515.

verbum, 290

verécundus for -2-, 202.

vereor, 4733 verébamini (?), 202.

Vergilius, spelling of, 23.

vermina, 310.

vermis, 339.

Verner’s Law, 157.

vernus, 324.

_ vero (-€). 550.

verres, 277].

verro, 468 ; vo-, 228; -ri, Perf., 502.

VErTULA, 277], 337-

Verrugo, 337+

-versus (-m), Adv., 549.

versus (-m), Prep., 595.

Verticordia, 361.

verto, 266; vo-, 228; -ti, Perf., 502.

Vertumnus, 327.

veru, 3OI.

verus. 266; -m, -0, Conj., 602; -o {-e),
Adv., 550.

vespa, 266.

vesperi, 396.

vesperna, 324

vestis, 305, 341.

veto, 288 ; vo-, 228.

vetranus, for -ter-, 184.

Veltd, the scansion, 212.

vetus, 356; Compar., 407; veter, 356;
-lustus, 407.

Ufentina, 250.

-ago (see Suffix in Gutt.).

-ui (-dvi), Perf., 506.

UI, pronune. of, 39, 44; vi-, pronune.,
29.

Uu
658 THE LATIN

-vi-, Syncope of, 171 sq., 180 sq.

ribix, 355.

vicem, Adv., 550.

vicenus(-9-’, 418; -cesimus,417; -cies,418.

viciniae, Loc., 397.

vicissim, 556, 586; -satim, 556.

viclus for -tul-, 83.

victria, N., 371.

vicus, 295.

videlicet, 564.

video, 473, 266, 232; -di, 502, 497;
viden, pronune., 163; vissus, spell-
ing, 112.

viduus, 268,

vieo, 266.

vigeo, 229.

vigil, 376.

vigilandé, the scansion, 213.

viginti, 417 sq. ; accent., 165 ; vinti, 418.

vile, Adv., 559.

vilicus, the spelling, 112 sq.

villa, pronune. of. 112 sq. ; ve-, 29.

vilum, 333, 179.

villus, 229.

vinarius, 321.

vincio, 470.

vinco, 471, 298; vici, 502.

vindemia, 178.

vindex, 362.

vinea (-ia), 22.

Vinnius, pronune. of, 118.

vinolentus, spelling of, 192.

violens (-tus’, 352.

vir, 260; pronunce., 29; -wm, Gen. Pl.,
402.

virdis, &e., for -rid-, 171, 185.

virecttm, 335.

vireo, 260.

viritim, 556.

virus, 267.

vis, Noun, 230; vis, Plur., 399 345, 1.

vis, 2 Sg. (see volo, decl. of).

viso, 462; -si, 502; visso, 112.

vissit for vixit, 107.

vita, 179; vitam vivitur, 521.

vitex, 230.

vitis, 266, 341.

vitus, 344.

viveradix, 192, 361.

vivd, 301; vivebo, Nov., 492; rixi, 499 5
-88-, 107; vivitur vitam, 521.

 

virus, 230,

LANGUAGE.

vix, 5553 -dum, 614.

-ul- for li, 232; from 1, 278 sq.

uleiscor, ultus, 310.

uligo, 287.

-tilis ‘see Suffix -LI-).

-ullus (-lus), T15.

ullus, 449; pronune., T13.

ulmus, 279.

ulna, 260, 179.

uls, ultra, 594 Sq. ; pronune., 142; ol-
timus, 236.

ulula, 363.

ululo, 474.

Ulysses (Ulixes), 286, 200.

-um (-orum), Gen. Pl] , 402.

umbilicus, 283; imb-, 29.

umbo, 283.

umerus, 236: spelling, 56

und, Adv., 561.

Unaccented Vow., weakened (see
Weakening) ; Syncope of (see Syn-
cope).

unco, 476.

uncus, 259.

unde, 570; pronune., 122,

undecim, 416; pronune., 141.

undeviginti, 416.

undique, 570, 206,

-undus (-endus), Ger., 543.

unguis, 293.

unguo, 301 ; -netus, pronune., 140.

unicus, 337.

universus, 178; -sim, 554.3 oinuorsei,178.

Unomammia, Plaut., 364; Oinwmama
193.

unose, 554.

unquam, 570.

UNUS, 409 Sq. 3 -Quisque, 449.

vo-, O. Lat. for ve-, 228,

-vo- atonic changed to vu, 267.

-uo, Vbs. in, 484.

Vocative (see Declension) ; of IO-
stems, accent., 163 sq.

*rocitus, empty, 18.

vocivus, 18; voco for va-, 15, 18.

voco, 228, :

Voices (see Passive, Deponent).

Voice-Stops (see Mediae).

Volaterrae, 228.

-vollo (?) (ve-), 228.

volo, I wish, decl. of, 456 sq.; accent.,
169; velim, 515; volam, 516; volimus,
4567. ; Inf. in Romance, 536.
INDEX.

volturus, 374.
Volumnus, 327; -nius, 228.
voluntas, 541.

volup, 553-

voluplas, 342.

vomo, 267, 506 n, 228.
vopte, 426.

toro, 301, 228.

vorro (ve-), 468, 228.
vorto (ve-), 467, 228.
vos, decl. of, 425 sq.

Vowel, Grades (see Gradation);
Quantity (see Quant., Shortening,
Lengthening).

vox, 358.

upilio (op-), 34.

Upsilon, Lat. transcr. of, 4, 11, 36,
248,

upupa, 315, 363.

urbanus, 325.

urbs, 239; spelling, 78 sq.

urceus, 239.

urged, 239, 482.

UINA, 3IO.

-urnus (see Suffix -NO-).

uro, 466; ussi, 497; pronune., III;
ustus, pronune., 255.

-urrio, Vbs. in, 482.

ursus, 239.

-us, Gen., e.g. Venerus, 384; decl. of
Neuts. in, 355 sq. (see -rus, 2 Sg.
Pass.).

usque, 571, 595.

ustium for ost-, 34.

usurpo, 173.

ut, 606, 605 ; accent., 166; uti, 607.

uter, Noun, 289.

uter, Pron., 450; -que, 450.

uterus (-um). 369.

Utica, 37.
utinam, 605. -
utor, 247; utarus, 534; wssus, I12;

utunto, 519.
utrasque, Adv., 557:
utrimque, Adv., 570.
utrinde, Adv., 570.
utro, Adv., 568.
utrubi, Adv., 567.
utrum, Conj., 606.
-ut(t)io, Vbs. in, 488.

 

659

-utus, P. P. P., 542.
-uu- written for #, ro.
-H(v)i. Perf., 508 sq.
vulgo, 560.

vulnus, spelling of, 236.
vulpes, -346.

vulpinor, 472.

vultur, volturus, 374.
-uus (see Suffix -UO-).
uxor, written voxor in MSS., 5.
vy- for vi-, e. g. vyr, 29.

W, Lat. expression of, 7 sq.; L.-Eur.
W, phon. changes of, 265 sqq.

Weak, Cases, 367; Grade of Root
(see Gradation).

Weakening, of Atonic Vowel, 185
sqq.,148 sqq.; syll. long by Position,
I9i sq.; short syll., 194 sq.; bef.
r, 192; bef. Lab., 192 sqq.; of
Diph , 195 sq.; of Diph. in Hiatus,
196; of jé, vé, 196; 0, u, t, 196 sq. ;
Gk. loanwords, 197 sq. ; long vow.
unweakened, 199; short, 198 sq. ;
re-formation, 199 sq.; weakening
in pretonic syll., 200; by Assimi-
lation, 201; in final syll., 203 sqq.

wel-, ‘to wish,’ 456 sq.

Wharton's Law, 159, 222.

Winds, gender of names, 369.

Word-Groups, 361 sq., 365; accent.,
T6I sqq., 169 sq.

X, the letter, 2, 5; written 2s, cx,
&c., 5; pronune. of, ror sqq.; -ss-
for, 102 107; for Gutt. with s, 291,

-x (see Suffixes in Gutt.).

Y, the letter, 4, 11: Gk. (see Upsi-
lon); Lat. expression of y-sound,
7 sq.; pronune of Lat. y, 34 sqq. ;
I-Eur. Y, phon. changes of, 262
sqq.; Lat. y for i, u, 29 (see U).

Z, the letter, 4,5 sq ; Gk. (see Zeta);
pronune. of, tor sqq ; I.-Eur, Z,
phon. changes of, 303 sqq.; Lat. 2
for j, 49.

Zabulus, for diabolus, 105.

Zeta, Lat. transer. of, 4, 11, rot, ro4.

THE END.
p-
p.

- 415, 1.

- 420, 1.

sp eu tp UP POP OP OP

ADDENDA ET CORRIGENDA

<i pes

. vii. Prof. Stolz has now published a Lautlehre der lateinischen Sprache (Leipz.

1894), giving the phonetic laws of Latin, (ef. chap. iv of this book)
along with a general introduction to the study of the language.

» 344, § 95. The new number of the American Journal of Philology (vol. xv.

p-. 194) has a paper by Mr. L. Horton-Smith, in which this view of
the origin of the Gerundive -ndo- is supported by a strong array of

arguments.

« 159; 1.12 for
. 198, 1. 3 5
.235,1.7abim. ,,
. 238, l.gabim. .,
» Bye, da tg ys

276, 1. 8
291, 1, 11 ab im. ,

356, 1.

pp Ww an

1.
528, 1. 22 ”
551, 1. 19

v

illic read illic.

Ann. Epigr. ,, Ann. Epigr.

on- + «OM.

orbs ..  orbis, late Lat. orbs.
déveri «  déveri,

citlare we ealare,

*qénqité a *pemqye.

*modes- +  *modes- (whence moderor).
SenUS 1 GENUS.

*egd s Fego.

*egdm »  *egdm,

fefaced »  fefaked.

U-stems +  Vow.-stems.