NEGATIVE 
 NO. 91-80083-11 
 
MICRO! ILMED 1992 
 COLUMBIA UNIVERSITY LIBRARIES/NEW YORK 
 
 1^ 
 
 as pari of the 
 Fouiicialions of W'csie;;' Ci\ ilizatioii Preservation Project 
 
 55 
 
 Fuiided bv the 
 NATION.VL ENDOW MENT FOR THE HUMANITIES 
 
 Reprcxluciioiih may noi be made wiiliout permission from 
 
 Columbia University Library 
 
COP^^IGHT STATEMENT 
 
 The copyright law of the United States -- Title 17, United 
 States Code - concerns the making of photocopies or other 
 reproductions of copynghted niatenal... 
 
 Columbia University Librar\ reserves the right to refuse to 
 accept a copy order if, in its judgement, fulfhlment of the order 
 would involve violation of the copyright law. 
 
AUTHOR: 
 
 MUNRO, HUGH ANDREW 
 
 TITLE: 
 
 FEW REMARKS ON THE 
 PRONUNCIATION OF... 
 
 PLA CE : 
 
 CAMBRIDGE 
 
 DATE: 
 
 1871 
 
Restrictions on Use: 
 
 Master Negative # 
 9/ - S0OS3 - // 
 
 CX:)!AJMBiA iiNiXI'KsirV LUiRAKHiS 
 
 I'RHSHRX'A'fKlN DEPARTMEN^r 
 
 BliU lOCRAIMilC MirROf-OKM TARCFT 
 
 Ongiuai Material as i-iimed ^ Lxisling [bibliographic Record 
 
 mf-tr-w-T' 
 
 -fT«p~"""'"T<r 
 
 — I > ii i « J HI * ■ * I* " "■« ' * " y 
 
 
 (119 
 
 877*15 
 
 2 l.lunro, Iluc^i rJicIrow JolmotonG, 1019-1685 
 
 A few ren^arks on tlio pronuiiciation of Latin, ■ '■'% 
 vilUi a poGtccript, by H./u JJ.Iuriro. . . Cambridge, 
 Deichtoii,1071« 
 
 3G p# ' ^^: ^^* 
 
 • • A 
 
 \ 
 
 ^ 
 
 .1. 
 
 L 
 
 4 
 
 "J 
 
 »•• .. 
 
 TECHNICAL MICROFORM DATA 
 
 I'li,M S!ZH:„J:_:,J[^jQA_. 
 
 IMAGE PL.ACnMHNT: ..TACIIA^JB Illi 
 
 DATE FILMED: ^ . /0^_T^ INITIALS 
 
 FILMED BY: Rl SFARC 1 1 PUBLICATIONS, INC WOODBRIDGE, CT 
 
 REDUCTION RATIO 
 
 JJ 
 

 r 
 
 Association for Information and image Management 
 
 1 1 00 Wayne Avenue. Suite 1 1 00 
 Silver Spring, Maryland 20910 
 
 301/587-8202 
 
 Centimeter 
 
 mi 
 
 1 2 3 
 
 iiiiliiii iiiiliiii II ilii 
 
 Jm 
 
 4 5 6 7 8 
 
 iliiii iiiiliiii mil III mil III iiiilim iii 
 
 9 10 n 12 
 
 liii liii liiiiliiiiliiiiliiiili nil 
 
 13 14 15 
 
 nil II iiiilim nil 
 
 Tl 
 
 "fT'frf'VT " "1" 
 
 T] 
 
 "fTrrriTi vVi pTi r 
 
 1 1 "n"f' |"('p"|" "1" 
 
 1 "I'T'fT T 1 
 
 Inches 
 
 1.0 
 
 I.I 
 
 1.25 
 
 1^5 
 
 |7J 
 
 |ao 
 
 Li, 
 
 KUbb. 
 
 2.8 
 
 3.2 
 
 13.6 
 
 nil 4.0 
 
 1.4 
 
 2.5 
 
 2.2 
 
 2.0 
 
 1.8 
 
 1.6 
 
 MflNUFfiCTURED TO fillM STflNDPRDS 
 BY APPLIED IMAGE, INC. 
 
 mm^si^stts^ss^M 
 
giirrs 
 
 M^l 
 
 j 
 
 
 ijilivuvvj 
 
 GIVEN BY 
 
 Hubh Lvai 
 
 ? 
 
 """WtSI 
 
y-imwi. 
 
 ^ 
 
 A FEW EEMABKS 
 
 ON 
 
 THE PUONnNCIATION OF l,\TT\ 
 
 f 
 
 WITH A POSTSCRIPT 
 
 BY ^ 
 
 Hf A^ J? MUNRO 
 
 .-nT 
 
 FELLOW OF TRINITY COLLEGE 
 AND PROFESSOR OF LATIN IN THE UNIVERSITY OF CAMBRIDGE 
 
 ■^ 
 
 I 
 
 '» 
 
 k 
 
 CAMBRIDGE : 
 DEIGHTON, BELL, AND CO. 
 
 LONDON: BELL AND DALDY. 
 
 1871. 
 
Cambriticje : 
 
 PRINTED BY C. J. CLAY, M. A. 
 AT THE UNIVEUSITY PRESS. 
 
 o 
 
 o 
 
 CO 
 
 f 
 
 o 
 6 
 
 The following Remarks, privately distributed some months ago, I now 
 publish with a postscript, in the endeavour to comply so for as is in my 
 power with the request of the headmasters of schools who met at Sherborne 
 last year. To ensure complete uniformity in the pronunciation of Latin 
 is an impossibility: it is a delusion to suppose that there is a foreign as 
 opposed to an English method. In England itself there is no one un- 
 varying system. If the scheme which is proposed in the Oxford circular, 
 as finally revised by the Oxford Philological Society, be taken to represent 
 the minimum of change desirable, and some such scheme as is sketched out 
 in the following jjages be placed at the other extreme, the whole inter- 
 vening ground might I think be covered without exhibiting much greater 
 differences in practice than exist among us at p^^sent. At the same time 
 it seems to me desirable to endeavour in theory after a worthy ideal ; and 
 the most essential part of such an endeavour is in my opinion to take for 
 our standard the Italian vowel system, if even in practice we should many 
 of us fall short of it. 
 
 
 0* 
 
 o 
 
 < 
 
 1—2 
 
 41^717 
 
In discussing the pronunciation of a dead laiirruaoe It is well 
 to remember 'the shrewd Sicilian's' Xu(/>c Ka\ f.^firaa airiaTdv. And 
 I should probably have gone on to the en! <>f mv life in lieincr 
 sober and mistrustful in tln\ matter, if it had not been fonvd on 
 my attention from many ditierent quarters wliirl, [ cidd n-t dis- 
 regard. Nearly two years ago Mr Cornish of Eton, m Li^ uwn 
 name and that of several of his colleagues, urged me to print 
 something on the matter. For many reasons T declined nt 1)1.3 
 time to enter on so slippery a course. Soon after some friend< lieie, 
 to whose judgment I could not but defer, among them I)i LigLtfuut 
 and our Public Orator, pressed me to try a reform. Tims stimnlnted 
 I gave some lectures on the subject more than a year ago, n\v\ . v, r 
 since have continued in lecturing to adhere to the system I ilien 
 traced out. Last term Professor Palmer wrote to me that they 
 were thinking of a reform at Oxford: at his request i sent a 
 pretty full summary of the plan I pursued. This was received with 
 very great courtesy by him and the distinguished Committee ap- 
 pointed to consider the matter. They w^ere not however mriiu-d 
 to go so far as I had gone; and they have iince circulated a private 
 paper stating what course they were disposed to recommend. It is 
 with reference especially to this paper that f jiint these remarks. 
 Personally I should have been disposed to bow at once to such 
 high authority; but I have been almost forced to move for the M~ 
 lowmg reasons. On the one hand the Head-Master of W'inehester 
 wrote to me a month ago to inform me that 'at a conference of 
 Schoolmasters held at Sherborne this Christmas... ii was resolved 
 to ask the Latin Professors of Oxford and Cambridge to issue a 
 joint scheme of Latin pronunciation, to ensure uniformity in any 
 changes contemplated'. On the other hand not only did 1 • iiink 
 myself, but I found it to be the general opinion of those whom I 
 consulted, such as Dr Lightfoot, Mr John E. Mayor, Mr Jebb, ^Ir" 
 

 allowiirjr the 
 
 G 
 
 tfiai \\L might with advantage push reform farther than 
 i pnprr prn poses. ^\r ATayor says 'I confess that I would 
 
 rath" ! k* t p 1 > (lii existing pronunciation than accept any com- 
 promise'. Lastly that (Hstiiiu'ii-lM-l -'ii"!:!!' .-iU'l -r.-thiiii.-ir'.aii .Mi il. 
 J. iluhy Iia> pubiishud a pa])ur, in \siii.'ii liu declares iiimself in favour 
 of a romph-tv' srliomo of icfonn. it i- \vh]\ great diffidence there- 
 fore that 1 issue tliese i-finnrk-, \'<r the -nlc piirp..,-r ,. 
 (piestions iiivstlvfd to be con^idiTcd ;ind diM-u-.-td. 
 
 I wish thfU to (h'claru my lull coiicuricucc iii the changes piu- 
 posed in the Oxford pnpor and mv rcasonq for e""!'!'-!' -^i^ fMi-fln'i-. 
 I hold that rofonii, wlirtlici- partial ^r cnnipiff c. vh.snM !„■ nnij. r- 
 taken f>r !t> nww sake and the sake of the an'i.iii 1 tii-uat'^e, not 
 to mak'j uurstdvcs mor** rnlulligiidc Lu * uLhcr Lauii-ic<idiiig nations*, 
 who aro iiof int<']]i<j-ihh' jo r-nr-]] f>f]i'-r winiout spec'i.'i] <-nltivation. 
 A iMtii 'hni in - Litm is at first as nnintclligible t<» .tn Italian, as 
 o!i!x is, an. I more absnr 1 , a Spaniard cannot be understood by 
 ]' r. n« hijiaii ui liduu] , a Scuichman's bro^jue, while retaininii some- 
 thini; of fh(^ prt^pi^^ voAvel sounds, has most of our own disagreeable 
 p* ' idi iiiin >, i. uiipleasing and l)ut partially intelligible to us, and 
 cannot be unlorstood hv Spaniard, Frenchman or ItaHan\ 
 
 Are \s.j tliuii ^aiid this is a \iUu ipie^iion) to endeavour to observe 
 (juanfifv svstoTnatically to distinguish between long and short, and 
 1""-' ini h Ml r, syllables? If this is to be done, we must break 
 alik' witii .ill existing pronunciations, Italian as well as English. 
 'lis tu miiv . f iho accent over quantity is perhaps more marked 
 i'i ^^1' if ikiin than in our own reading of Latin. We learn from 
 ( icero an i «^>ninMl;;in that rhythm or a due admixture of lono- and 
 }^ if -vJkihk- \\,i- important in prose as well as verse; and for 
 ^!'" 'h '0 < h>< rving quantity, I seem to feel more keenly the beauty 
 ol ( icero's style and Livy's, as well as Virgil's and TTorace's. The 
 same T fin I to be the case with those in whose judL^ient and 
 kn v 1 mI^. 1 confide. Mr Mayor writes to me: *As regards quantity, 
 C. .1 Shrewsbury, a most experienced and intelligent teacher of 
 elementary classics, tells me that since he has made his boys dis- 
 tiTiguish cano, amis and canus, lego, legis and lex, Icgis, and sound 
 
 ' I have a sufficient knowledge of the ordinary Scotch method and care for no 
 contradiction however flat. If in Edinburgh or elsewhere any profess a superfine 
 system^ ac<;eptablo alilve to gods and men, to Spaniard, Italian and ancient Roman, that 
 is |)ot Scptch, bwt s^me ideal which common mortals would fain attain to, but cannot. 
 
 * 
 
 all long syllables long, and short short, in v liatever positions, he finds 
 them perfect in quantity for verse composition'. 
 ; Though we break however with all existinir systems, iLauan 
 
 appears to me to offer many vahiabic auls which it wnnld lio mn^f 
 iinwi^o to neglect. English seems so utterly dilT. ivnt iii all its t.aies, 
 its entire vocalisation, from ..],1 Ikniin, ijiat (.ftcn \\o cannot find in it 
 even single sound- to give as the representative of a Latin sound. 
 Ilic Italian of liTfratni'c has been tixcd for six ecMtiirirs ; tlie more 
 wo pxamino thr twn, tlir moro ^vo fed that tlio liomano-Tuscan of 
 to-day is csm ntially the Latin of the 7th or Sth conturv : that 'Siede 
 ^•■i trrrn «h.vr nata lui ' must represent very nearly tin- 7th century 
 pronunciation oi 'sedet (il)laterra deiibi nata fni ' : that race and 
 elimate and rnncl, else have made the Mingua Toscana m bocca 
 Ivomana' to inherit in a higher degree than any other lanu'iao-e the 
 refinements of old Latin. Let me not be misuiiderstood : i feel 
 most strongly the tintli ■■•( I )r Ridding's judicious words, wlien ho 
 writes: 'the point which would be likely to cause the greatest 
 difficulties, would be very subtle distinctions oi -h ties of vowel 
 sounds. But if any such wx^re proposed, \\._ ::iiuuld iia\t t.. kt 
 boys be rough in it, and tlvv would be rough in it. T f I thrro 
 is so much to be said in favour of doing a thing as thoroughly as 
 possible, that I would say no more than just this, that a subtle 
 foreign pronunciation will not be realized at school I think'. \\ hat 
 I mean is this: our English sounds are so different from vdnt v>r' 
 must suppose the old Latin to have been, that, by looking only to 
 them, we should probably fall into such slipshod ways as t . nn ake 
 our new pronunciation hardly better, perhaps more distasteful liiaii 
 our present. I do not i)ropose that every one should learn Tta]i m 
 in order to learn Latin. What I would suggest i- tli at those who 
 know Italian, should make use of their knowledge and should in 
 many points take Italian sounds for the model to be followed, that 
 those who do not know it, should try to learn fmm others the 
 sounds required, or such an approximation to them as may be pos- 
 sible in each case. 
 
 In seeking to recover in some degree the old pronunciation, we 
 have many great advantages in Latin, compared ^\dth Greek : 1. 
 from the literature developing itself comparatively late, and so not 
 stereotyping the orthography : we see in the first volume of the 
 Coi-pus inscr. Latin, a map as it were of the language spread op- u 
 
s 
 
 before us, and feel sure that cliauge of spelling meant systematical 
 thaiige of pruiiuiiciation : coira, coera, ciwa; aiquos, aequos, aecus ; 
 qiiei(]vo))\fpi(\ qfficKmqite, etc. etc. : 2. from the far less complexity of 
 soiuul -. Iijiiih iius mostly disappearing and the two chief ones left, 
 ae and an, being easy to pronounce: 3. from the invaluable service the 
 
 u- m k.:r|»iii_; liie accent lu most cases uii liie 
 
 1' 
 
 i!!aii>. ha V r rch'. 
 
 'It 
 
 riglit syilubiu, cAt-n winlf chafi'j'iTi'j' it^ Tiatnrc. IM.-uiv of n*^ T fcnr are 
 (|uit<' ii]irr)ns(-'!<.u> of t!i.' i\r\>\ \\r M\\. tlM^in: Ihii, IkmI we been left 
 to our own !ight>, t ho coiii'ti -i'Mi \i\ l.niin ini-hi lia\.' h.'on as disas- 
 trous as Hi dn-ok. In ol)-or\ in,; (|uahni v u c shall ^uil k '■.•!. tho acc^juL 
 in its proper pkicc, but its tyrnnnioni proilnminance v-Hl ho nbated. 
 
 At first th*' Latins seem to have been careless enouuh in matters 
 
 of granmiai' and 
 
 H • ■ \ I I 1 » ' * 
 
 ^'r MM fh« tinieof 1 hi nius onwards 
 - ' inn !i pains and attention to 
 
 j>r>'!!nhoi;tf ,.ai 
 this nai i- ai nf" -'fMniinari;!! ' 
 thu.su UiaLiurs Liiai by tiiu liha uf Cicero an 1 \ ugil tiiu language had 
 nttaino.] n ]-m rf. (--tion as great as that of Attic in its palmiest days. 
 Tho -hnring o\ . i .t jin-J syllables, once its great weakness, had been 
 so inurh oarected bv careful culture that, n \ irgil's antiquarian 
 prujLuiicu.^ had uui ^loud ni i h. way, we may infer fro?n the example 
 of Ovid tliat rlivir,]! ..f long syllables an! many sliort ones would 
 ^>*>'^' <«^'i> '-^^ 'h a}.|M ncd. Every change in pronunciation seems to 
 ha\o ho.'!! o.-iivhiMv ni:iik.-ii 'hv a change in ihe spelling. We may 
 thus 1 ihink approx miiiLu lo liiu true i^roimnciation. This approxi- 
 mntio.n it ni iv he said, will after all be a rude one. Vuiy well: that 
 may be an argument for doing nothing at all ; but not I think, if we 
 try a reform, for J.ing it imperfectly. With this preface I will pro- 
 cuud Lu siiuw where it seems to me we might safely go beyond the 
 Oxford rirrnlnr in correcting our pronunciation of the different 
 
 i 'f' '^: i^h. r th )? 1 will say a few words about quantity, accent and 
 
 elisioi!. 
 
 * a .^iiuuki liavu the sound of a in father : a that of the first a in 
 prtpfy\ As tin first a of jyajya would seem in English mouths to be 
 sometimes a short a, sometimes a short /, sometimes a short u, and as 
 it is well to accustom the English to open the mouth and expand the 
 chest. 1 w ui.i udd; (;r still better, a should have the sound of the 
 accentui oh a of the unaccentuated Italian a: amiita, padre, pa- 
 drone. < ' course a and every short vowel should be pronounced 
 short, when the syllable in which they occur is only lengthened by 
 
 pUSitliJli. 
 
 'e (and ae) should have the sound of a in cake : e of tho first (/ u\ 
 aerial,' The first a of aerial has to my ear a very vague sound : 1 
 would add : or better, let e have the sound of the Italian closed e 
 (e): e, whether the syllable is short or lengthened by position, and ca' 
 that of the Italian open e (e) : arena, ride, but bene, temere : est 
 Coats'), but est ('is'); lectas (partic), but lectus ('bed'): Caesar, 
 musae, Aeaeae. Thus in Italian^ as a ride e' represents tlio h)na h 
 represents the short Latin e ; while Latin ae is invariahlj r* pi o>Lniod 
 by e: Cesare, s^colo, etc. Diez compares the German lehen, ^vegen 
 for the open, legen, heben for the close e. In Knglish perhaps pear 
 will give a notion of open, pain of close e. In Italian hh. \ d > n-t 
 distinguish between naturally long and short vowels, wli-n tlaj 
 syllables are long by position ; but we should do so in Latin T flunk : 
 mens, mentis. In Italian too the open and close sounds are only per- 
 ceived in the accentuated syllables. 
 
 In Lucilius' time the rustics said Cecilius pretor iur Caecilius 
 praetor: in two Samothracian inscriptions older tlian B.C. 100 ^tho 
 sound of ac by that time verging to an open e), we find rmiste piei 
 and muste: in similar inscriptions fjivariu piei, and mystae : Paeligni 
 is reproduced in Strabo by iUXtyvoi: Cicero, Virgil, Festus and 
 Servius all alike give caestus for Kearo^ : by the first centnrv i^orlmp^; 
 sooner, e was very frequently put for ae in woid- like taeter: we 
 often find teter, erumna, mestus, presto and the like : soon inscriptions 
 and Mss. began pertinaciously to offer ae for e : praetium, praeces, 
 quaerella, aegestas and the like, the ae clearly representing a short 
 and very open e: sometimes it stands for a long e, as often in 
 plaenus, the liquid before and after making perhaps 'the e more 
 open {(TKr^vr) is always scaena) : and it is from this fonn plaenus liua 
 in Italian, contrary to the usual law of long Latin e, we have pi^no 
 with open e. With such a pedigree then, and with the genuine 
 Latin ae always represented in Italian by open e, can we hesitate to 
 pronounce the ae with this open e sound ? 
 
 'I should have the sound of e in he, i of e in behalf : I mi lud 
 prefer: l shall have the sound of the accentuated, i of tho nna*- 
 centuated Italian i : timidi. 
 
 'o and d should be sounded as at present': in this I cannot 
 acquiesce : what is the present o ? noii, bos, pons, honos ? or, 7ios, hos, 
 domos ? these o's we English utter with totally different sounds : we 
 have scarcely in English or in English-Latin a genuine o, except j . r^ 
 
 
10 
 
 haps before r : roar, mores : then what is our Anglo-Latin o ? how 
 does (liff( r in dnnvim. and donuml Here too the close and open 
 Italian o represent respectively the long and short Latin o, on the 
 exact analogy of e. Let us then represent o by the close, o by the 
 upLii luilian u : the name of the painter Benozzo Gozzoli gives a 
 specimen of tho two o's. Or 1 care not if we take the long and short 
 CiTinati M : r./me, (j<^>lil : H'r '^wv jn;rp...se. Here too az* has a curious 
 a!la^^^\ wirji dr : ih,' Liiiti nu becomes in Italian 'prn o: bro, hde : I 
 wuiild jjiuiiuuiicc LliU^ in Laliu ; phsfnnn, Clhdius, corns, i'uiliap.^ 
 to<j the fact flint ffJaria, viftorin and iiic roinnion tormi Tint ion -orio 
 have ill Italian thr njM'n -*, mi-ln >!n'\\ ihai thy corrcsj)oinhn- r; '\u 
 Latin was (>|Mn hy coming iMiwt.n \_\\n litpiids, <i t)efore one: 
 cuhipai'c plciiits above. 
 
 ' v7 slionld liavo the sonrid nf o in mlio, v of u in fruition* : or, of 
 accentuated aii'l imarcontuated Italian u respectively: tumulo, tumulto. 
 For thit 1 argre class of wor<ls, comprisinor all superlatives an 1 many 
 otliur kind.^ ul' iiuiui.^ and ui \uiij.-, (^uiiiLiliaii <:^i, i, 8) gives a valuable 
 hint : • th( ro is a middle sound between u and ?' ; for we do not pro- 
 ii'tune. ojifn/ucm [pptumum) either as optimum^ or as optumumJ 
 
 ' au shonl 1 have the sound of oio in owl'; I should prefer the 
 Italian an wiiicli gives more of the u, than our owl, cow. 
 
 ' eu slmiild be sounded as at present': for Greek words, adopted 
 into Latm, let Greek authorities tell us what is right : of Latin words 
 there are In two or three, heu, ceu, sen: I prefer the Italian eu which 
 gives yuLL more of the e, than the English you sound of these words 
 does : ' m' as we in we' : here too in Latin we have but two or three 
 s n 1 d 1 \\ . ! d < , cut, hui, phui, huic. 
 
 'oe shoul 1 h iv. I he sound of a in cake': here too (putting Greek 
 ^^■ui\h out of the question), when hateful barbarisms like coehivi, 
 coena, moestus are eliminated, oe occurs very rarely in Latin: coepi, 
 poena, moenia, coetus, proelia, besides archaisms coera, moerns etc., 
 where oe, coming from oi, passed into u. If we must have a simple 
 soun i, 1 should take the open e sound which I have given to ae: 
 but I should prefer one like the German o. Their rarity however 
 makes the sound of oe, eu, ui of less importance. 
 
 ' ei should have the sound of i in idle': surely this cannot be right. 
 V>\\\ this too" is a diphthong which has practically disappeared from 
 Latin, owing to the people's dislike to complex sounds: we find hei 
 (more correctly ei): ei (dat.) and rei are sometimes monosyllables, 
 
 I' 
 
 I'! 
 
 If, 
 
 h ^ 
 
 / 
 
 and Horace has Pompei, Voltei, Virgil Fenei. But in the older lan- 
 guage there are thousands of ei's, later i or e: surely we are not to 
 pronounce all these with the English i sound, in dofianrp aliko of 
 euphony and consistency. I should infinitely prefer eiih. i- the Latin 
 and Italian long e, or long i; i.e. to pronounce ojiiiieis either as 
 omnes or omnis. But as the diphthong is inip..iLani, i wuuLi nmcl! 
 rather give it ihu iLaiian or Latin e sonnrl qnirklv f. Unwed liv an 
 ItaliaT) or Latin i sound. Then there is an jnipuiant da^s ..t words 
 "^ ^''lii'li ?b<' nvfnrd paper takes n.> note: are we to -iN-.- the Kn-]i>h 
 t sound to such forms as eius, Pompeias, Setanas^ And here I will 
 luke together a large elass of similar wnrd^ in ai, ei, oi, ui, whirli 
 have really two z's, a voAvel and a consonant, and which in njd times 
 were often so uiiiten, as we see in inscriptions and oood .M^s. : 
 Quintilian tells us that Cicero preteired -11110 J/a/e/z/enio o-eminata 
 i scribere'; and we know fiY^m Prisrian that Caesar in jn's d(' anal..^ia 
 spelt PoDipeiii (gen.) with three 2's, and explained ho\v they were 
 .'til u, be pronouncedi \\\> English shew in these words our usual 
 undaunted inconsistency: we say JIluu bui major, Graius but Troja, 
 eju6 buL Pompeiiis; Seius, while we call his -011 Sejanvs. In ^m h 
 words the z has a double force, that of the vowel together with tliat 
 of the consonant i (our y): the Greeks always write llo^7r7,KOs\ 
 not UofiTreio^. In all these cases I conclude we shcnld give the 
 long Latin or Italian a, e, i sound respectively, followed by an En- 
 glish y or Italian j sound: Gra-yus, Ma-ya, ma-yor, Tro-ya (this 
 word has the open sound in Italian), e-yus, Pompe-yus, Be-yanus, 
 cu-yus. So with the compounds of iacio: e-yicit, db-yicit, re-yicit', 
 though we should always write them with a single i: eicit etc. : Gams 
 is a dactyl, Gains a nonentity. The or e of proin, proinde, prout, 
 dein, deiyide, when not forming a distinct syllable, is elided, does not 
 form a diphthong, and must be treated as cases of elision Ijctween 
 two words: in iieUtiquam e is elided as mmli a< in nurnquam, nullus: 
 the Greek eu and yi I refuse to pronounce upon. 
 
 We come now to consonants: the Oxford paper proposes that the 
 consonant i, or j, should have the sound oi y m yard: that consonant 
 u, or V, should be sounded as at present. That we should sound 
 consonant i as our y I am quite agreed : equally persuaded am I that 
 we should give consonant u the nearest sound possible to the vowel 
 u, the sound that is of our English w. This I hold to be called for 
 by the whole inner structure of the language: comp. iuvenis, iunior; 
 
 1—4 
 
* 
 
 I 
 
 12 
 
 noverat, norat\ motus, momen, mintms, nundinae, etc. etc.: by the fact 
 that the Greeks employed their ov to form words whicli must liave 
 })eei! uHtTly l)arbaroiis to their ears, in order to reproduce precisely 
 
 lliu Kuiii;ui MJiuid-.. OL-i.\//ry, dhovevTo^, and many others even more 
 rpinil<iv( : lastly by clear external evidence. Gellius is fond of quot- 
 in ' Cicero's friend Ni^^idius Fimilus, next to Varro the most learned 
 of tlie Iloiiinns, Ki> u the passage about the vowels cited by Gellius 
 
 uL Uiu end oi hi.-, Jiuh Look seems Lu luc Lo shew that the consonant 
 
 ?/ in VnJerius, etc. had the sani-^ rnlnti.-Ti to tlie vowel, as the i of 
 
 » 
 
 iecui'y etc. had to the vowel i; an' I liin in l",i]i cases they were as 
 near to the vowel sound n- iht^y couM v\.!l be. Still more con- 
 vmcinL; is liii' eiaiuiis passage m x i: luiless vos was SQunded woSj 
 tlip vtr.rv would ■<(^n^^. to linvn nn x^^^v^\ or meaning. Now Gellius 
 quoting I'i-nhis covers the whole classical period. Why should we 
 then renounce the advantage we have over others in our lo, surely 
 a nobler sound, to us at least, than v1 
 
 The circular shrinks from giving c and g uniformly the sound of 
 h and hard g\ and leaves ci and ti (and ? si) before another vowel to 
 be sounded as at present. As for special reasons I have spoken of 
 these points so fully in an Appendix, I will only say that, since keUy 
 kin, get, give are such genuine English sounds, I see no reason for 
 not allowing them in Latin, and many reasons for the contrary; and 
 that our rxishios, fashiams and the like are hardly compatible with a 
 reformed system. 
 
 The circular does not touch on other consonants : I wish to make 
 a few remarks on some of them: hs, ht should always be sounded, 
 generally written, ps, pt: lapsus, aps, apsens, apstulit, Araps, urps, 
 opscenus, optulit, supter: and generally assimilation should take place 
 in pronunciation, if not in spelling; ace-, not adc-, imm-, imp-, 
 coll- etc. 
 
 d and t we treat with our usual slovenliness, and force them up to 
 the roof of our mouth: we should make them real dentals, as no 
 doubt the Konians made them, and then we see how readily ad at, 
 apud apiit, illud Hint and the like interchange : f seems from what 
 (^fiuntilian says to have been sounded with a stronger breath than 
 wo ohipioy; L)UL i suggest uo change: m before q had a nasal sound: 
 quaynquayyi, numqiiam: final m was sounded slightly and indistinctly, 
 as I - proved by its elision and the testimony of grammarians : qiiu 
 
 i 
 
 
 13 
 
 I avoid, pronouncing cu or qioo: cum or quom, ecus or eqiios: r v\e 
 should sound more strongly and distinctly than we do at ])]' -ent. 
 
 Of 5 I would say a few words, as it has many interesting analogies 
 in Italian: s between two vowels has in Italian aij 1 I'l in h a soft 
 z sound like our rose: I would thus sound it between two vowols m 
 Latin: 7'osa, viusa, miser. But Avords of tlii^ Ixin.] inl.Mtin are cuni- 
 paratively very few, and in Italim there are most suggestive excep- 
 tions to .9 being soft between two vowels: ni cosa, riso, etc. and 
 in lIio adjective ternnnaiiiai -u.-^u \l is sounded as uui- s in .^ml : tlicse 
 words represent causa [caussa), risus {i^issus), exanijiles of tliat very 
 
 lar^re class of 
 
 o 
 
 .^]b.■l, 
 
 Quintiiiin 
 
 •20 
 
 sjh'aks 
 
 lie tells us tliat 
 
 Cicero and Vir-il wrote cassus, caussae, divissiones. There arc \ast 
 numbers of such words, in which ss was the original spelling, a lost 
 consonant having been assimilated, and the vowel was alwa3^s long. 
 The old Latin pronunciation seems to have been to dwell on the long 
 vowel or diphthong, and sound the ss as a single sharp s, as iii the 
 Italian words quoted: cau-ssa, cd-ssus, ml-sit {mi-ssit), ml-ssus, 
 iu-ssus, ru-sum (ru-ssum) for ryrsum, odJo-sus {-ssus) etc. etc.: the 
 ss and s seem to have been sounded alike. At the beorinnin2[ and end 
 of words too, and at the beginning of syllables, and before consonants, 
 s is always sharp in Italian, and should be so in Latin: sol, stella, 
 de-sero, ni-si, quasi, hos, nos, sonus. 
 
 There are 5 letters or unions of letters wholly alien to the old 
 language and brought into it for the sole purpose of reproducing 
 precisely Greek sounds: ?/, z, ch, pli, th: we have abundant evidence 
 that ?/, or Greek v, had some sound between i and u, probably like 
 either French u or German it; and one of tliese sounds I should wish 
 to give it. Of 2; I do not feel competent to speak. The modern 
 Greeks sound 6, cj) as we do, ;^ like a strong Scotch guttural : in old 
 Greek and Latin it seems to be generally agreed that the tenues 
 c, J), t were distinctly sounded and an h sound appended. I should 
 not venture to suggest such a pronunciation for Latin ph and th; but 
 should prefer it for ch, as this would not be a difficult sound, and 
 the Scotch or German ^uttural is strano^e to the EnHisli tonofue. 
 
 gn was sounded as we sound it, not as the Italians and French 
 pronounce it. Though I do not propose to change the sound of 71 
 before c and g: anceps, aiigo and the like; it seems to have been 
 nasal, nearer a g sound, and many grammarians wished to write 
 agceps, aggo, aggulus, as the Greeks actually did for similar reasons: 
 
I ill 
 
 i 
 
 ) ^ 
 
 ayytX. 
 
 14 
 
 ty^parr}';] though oddly enough both ItaUans and modern 
 Greeks ha\ here a clear ii sound. 
 
 hi modern Latin pronunciation quantity is systematically neg- 
 lected: attention to it seems to me essential in any reformed 
 method, attention too to the iiiiiural length of vowels when long by 
 posit in Jn T^itin there is no tj or ov Tjicilius unluckily for us 
 111 \ i II _: L I u u 1 1 . i 1 out of fashion tlie poet Accius' invention for noting 
 nuLLirLdly lun^ syllables by donMinu ihnn. though \\'- iind inahv 
 traces of this in tlic oldci* iiisrii |.! ioii.>; Jlaui'cu,^, ^^cfao^/u/xo"; ^o ee for 
 e, J tor 7. as vlximus as well a- vivo: ou f ^r " ns pouhlicom. Apices 
 were often usrd afterwai'd- m all ages to innrk ii.itui-.-dlv L-iig sylla- 
 
 e 
 
 bles: MdrtL.s, j'<'rtrU\ Im a h ! iiese usages are noted 1)\ n-nm iliaii. W 
 knnw too tliat tlio vowol of flic snpino and ooi^iuito parts of the verb 
 \va> n\\\:\\^ l^ng by nature, il' t lie vowel of iIh' lovsent indie, tlinugh 
 vhat was i-l lowed by a medial: aactus, leectus {])iiYtic.),h\xt factus, 
 Lctas ^.-Libst.): Cicero (Orator § 159) tells us also that every vowel 
 wla n fnllfuvod hv ns or vf bornnic long by nature: Insanus, Infelix, 
 but indoctus: coonsuevit, coonfecit, but composuit And this is borne 
 i>\\\ by abundant other evidence: we find in Greek KXrjfjL})^ \\\r)/ji€v- 
 T(K\ OvaXijis Oi'uXti'Tu^ and ilaj like. Priscian too (ii G3) tells us 
 th.ii '7" nml. fli.^ preceding vowel long by nature: reegnum, staag- 
 num, ben Lj mis, mallgnus, ahieegnus, pnvlgnus: and this is confirmed 
 by nm iiiidin- in inscriptions more than once the apex of a naturally 
 long vowel attaciiud lo regni, regnOy and also signu, digni, and in 
 Greek the form 'Vny^'^^i- ^^''" must not be misled by the wrong accents 
 MapKo^ !'■( M,"p/vf*<, Mayro<; for ^! 77i'09, there being conclusive testi- 
 mony for th« 1 n.ili a the vowel. The rhythm of prose as well as 
 verse will be improved, if we attend to such p.nuts: amaans ainaan- 
 tis, doceens doceentis, legeens audieens, but legentis audientis; amaan- 
 dus, doceendas, but legendus, audiendus: Moonstrum horreendum 
 Informe ingeens: Insontem Tnfaandoo indicioo, and the like. An 
 extruded consonant too often leaves a naturally short vowel long : 
 eXy ee\ sex, seescenti, seemis; Sextius, Seestius {aijaricoBiaTepov nihil 
 novi) ; ees, eest from edo. By comparing Cicero (de orat. ill § 183) 
 with QuintiUan (l, 5, 18) we learn that in the time of the former the 
 prose pronunciation was illtuSy unius, etc.: in the time of the latter 
 illlus, unlus, ho and subsequent grammarians holding the shorten- 
 ing to be a poetical licence. 
 
 Plautus and Terence, following the usage of common life, never 
 
 h*^ 
 
 15 
 
 lengthen a short vowel before a mute and liquid : compare on this 
 point Aristophanes with Euripides, Euripides with lb m. i : and in 
 prose we should always keep such syllables short. \\ ixcu iu the 
 learned verse such syllables are lengthened, we should still sonnd 
 the vowel short, and lengthen the syllable by separating distinctly 
 the two consonants : Gnatum ante ora patris, pat-rem : Ju Lycum 
 nJg-ris oculis lugroqve: similis volucri, nunc vera voluc-ris, 
 
 'i'hc Italians, as I hnvr nlrondv olisrrvod. linvo dono ns an in- 
 calculable service by keepmg m most cases the accent on thu riglit 
 syll.dilc. though the loss of quantity lias clian^vd its nature. Jt 
 woidd be well tu recai tlie accent to the right ])lace in the cases 
 vl: 
 
 ere Avo now Ti02:lect to f]o 
 
 to (b'aw it forward towards 
 
 enclitics: arniuque, omnidve as w<ll as ciintsque ; tcDt'iuc; to jjro- 
 nounce tanton, posthdc, postSa, jyraeterea, adeo (adw), qniprLiiiLis 
 abdris, inter se, apudmest, etc. 
 
 Tn respect of elision T wmdd only say tliat, by comparina' Plantn.s 
 with «>\id, vv. may see how much the elaborate ciiltivati<»ri .t the 
 language had tended to a more distinct sounding of final syllab.K-. ; 
 and that but for Virgil's powerful iuliuence the elision oi h'W^ 
 vowels would l^nve almost ceased. Clearly we must not altogether 
 pass over the elided vowel or syll. in m, except perhaps in the 
 case of e in common word.s, que, neque and the like. 
 
 In conclusiuii 1 wuuid repeat that, li we are to refurm ^nv jao- 
 nunciation at all, it would be well to do it as tla»! 'Uulilv a> wo 
 can, and get rid of as many of our Sliiljboleths as possible; :ind 
 would suggest that exact uniformity ibn - not exist among us nuw, 
 aa^l liued iiuL be luuked uijuii as indispensable in a reiurmed system. 
 At ;ii! ovents ' libera vi nnimam moarn '. 
 
i\ 
 
 16 
 
 APPENDIX. 
 
 An article wliicli has just appeared in the Academy of Feb. 1/5 by Mr Max 
 Miiller, ' on the pronunciation of c before e, i, ?/, ae, eu, oe\ and is argued 
 out wkh his usual power, will help no doubt to make innovation more 
 difficult here. His chief objection to change would seem to be the same 
 as that urged in the Oxford circular, that it could not ' be attempted with- 
 out intolerable offence to the ears of all the Latin-reading nations'. He 
 speaks of 'fear of ridicule', * a dislike of the harcih and disagreeable sound 
 of s\ich words as Kikero, fakit\ This difficulty has never struck me as of 
 such very great weight; and my ear has already accustomed itself to look 
 on Kikero^ skelus, sk'io and the like as even more euphonious than their 
 former sounds. Of course I assume that Sisero, Sesar, SrphaluSj sinic and 
 the like are still to be English for the new Kikero, Kaesar^ hjnkus, just as 
 much as for KtKepwi/, Kato-ap, Ke^aXo9^ kvvik6%. Our present English pro- 
 nunciation of Latin appears to afford some arguments to the j^oint. Some 
 centuries ago we pronounced with the rest of Europe (I assume now the 
 new and corrected sound of the vowels) cana^ cam and the like, as kana^ 
 kara: when the revolution took place in our vowel sounds, we said kena^ 
 kera^ not sena^ sera. Now that we propose to reform our vowel sounds in 
 cena^ cera, why should we find kenay kera more offensive than sena, sera 1 
 Our English k is common before all vowels alike and such consonants as 
 it can precede in Latin, and is at least as euphonious as s or tch: kettle and 
 kin are not less mellifluous than settle and sin : Kikero I prefer to Tchi- 
 tchero; and I doubt whether Kikero is to an Italian more offensive or 
 strange than Sisero, as they too have abundant k (ch) sounds before e and 
 i. Assuredly the many Greek words like Cilicia^ Cihyra^ scena, cithara, 
 Cithaeron I would rather have with theii- Greek than their Latin sounds. 
 
 Quite the same is my experience with the very numerous cases of -ci, 
 -si, -ti before another vowel: vicies, visio, vitium; species, spatium, ratio, 
 gratia, solacium. Habit here too is all-powerful, whichever direction it 
 takes. The common English pronunciation of Greek words like Avo-ta? 
 is I believe Avshta?, TleAoTrovi'Tyshto/, MiATysluot and the like, though custom 
 seems to permit a more correct sounding of the a. The pronunciation 
 of the oldest Greek scholars within my recollection, such as the late Bishop 
 Butler and Mr George Burgess, proved that some generations ago Greek 
 
 »> 
 
 
 u 
 
 Hi 
 
 ^ 
 
 r 
 
 17 
 
 was in mairy points sounded more like Latin than it is now. Bishop 
 Blom field was fond of telling an anecdote about a Freshman examined by 
 Porson. The Freshman talked of yScXsluoi/ : Porson intimated a preference 
 for /3eX-Ttov. The Freshman politely allow ud the Professor to please him- 
 self; but had all his life been accustomed to helshion and intended ic) stirk 
 to it. I think it not very unlikely that before his degree he became 
 reconciled to piXnov, and that if the will were present, it would take us 
 less time to exchange rayshio for ratio, speeshiees for sjjekies. 
 
 Nay if we keep within the limits of the Oxford pa2:>er, we shall be 
 forced to many awkward inconsistencies. Suppose we are comparing the 
 successive forms of words which we see collected in the first volume of the 
 new Corpus Inscript, such as coii^a, coera, cura ; Cailius and Caelius; 
 Coilius and Coelius, Caicilius and Caeciliiis, we must pronounce Koira, 
 sera, kura ; Kailius and Selius ; Koilius and Selius ; Kaisilius and Sesilius. 
 The more ancient pvlcer and Gracci will be pulser and Graksi, the more 
 lecent pulcher and Gracchi will be pulker and Grakki: co'epi and coepi 
 will be ko'epi and sepi. And so with an indefinite number of terminations : 
 haca and hacae will be haka and hasae, siccus and sicci will be sikkus and 
 siksi. Long-suffering as we are oil such points with our present system, 
 a partially improved method would perhaps render them intolerable. The 
 Italian shuns such inconsistencies by substituting ch ( = k) for c : secco, 
 secchi, and lungo, lunghe. 
 
 It is doubtful whether our improved y sound of j will not by contrast 
 make such inconsistencies appear even more flagrant. Habit lets us 
 acquiesce in our English w^ay of j^i'onouncing such words as ioci, iugi, 
 coniugibus SLiid the like: but will not yosi, yuji; conyujibus be somewhat 
 uncouth ? The Italians practically reverse this process, and give our j 
 sound to the consonantal i and our k and hard g sound to the c and g, by 
 writing giuochi, gioghi. This gi in fact is the almost universal substitute 
 for the Latin j, aiutare (adiutare) being quite exceptional. 
 
 But though to my present feeling to reform the pronunciation of j for 
 instance and leave that of c unchanged w^ould almost be worse than to do 
 nothing, the important point is to know what is right or probably right. 
 However firmly one may have held the common belief that the sound of the 
 Latin c was in all cases the same as k or our k, the fact of such an autho- 
 rity as Mr Max Miiller calling it in question, must make one hesitate. 
 Still a variety of considerations compels me to retain my former belief. 
 
 He points out with much force that it does not follow, because Greeks 
 and others in transferring Latin words into their own language always 
 represented c by k, that therefore the sound of the two letters was always 
 identical. And yet the fact that Greek and barbarian, Goth and German 
 
 1—5 
 
I- 
 
 ] 
 
 Jf 
 
 18 
 
 alike, do reproduce the Latin c by k is such a prima facie argument of 
 identity or near resemblance, that strong counter evidence is needed to 
 rebut it. Halm's Grammar and Dictionary shew that the Albanian has 
 sounds representing most of the modern corruptions of the Latin c, such as 
 various cr and f sounds. The cice7', which must have been imported into 
 those countries in early times, perhaps by Atticus on his farm at Buthrotum, 
 is represented by KytKyepe: this y (or Germany) sound being exceedingly 
 common in Albanian before all vowels, a and o as well as c and i. Now 
 when I think of the Greek KtKepojv and then of his own eponymous cicer 
 re})roduced on one side by the Albanian KyiKytpc and on the other by the 
 German klcher, each of these languages shewing only the first and to them 
 most natural deviation from the pure k sound, the concentrated force of 
 the three impresses me strongly'. 
 
 For the Greeks, though indeed they did represent /by </>, took much 
 pains to reproduce the most peculiar Latin sounds. How trying must it 
 have been to the eyes and ears of a Greek — unless he wished to laugh at 
 the barbarians — to find in his Polybius Iloo-Tov/xios 'FrjyovXo^ (Postiimius 
 Regiilus), in his Cassius Dion OvovXroypvov (Vulturni), in his Dionysius 
 OuoAovo-Ktos (Volscius), in his Ptolemy VvipoveSpovfx, and the like. If the 
 Latin -ce and -ci had anything of an s sound, why could not the Greeks 
 represent them by some combination of $ or f or cr, such as were used in 
 Byzantine times 1 The Greeks would probably have given to these sounds 
 some conventional meaning, as to those odd accumulations of ov : nor do 
 I think they would have cared for the quantity of such barbarous words; 
 or, if they had cared for it, would have hesitated to change it. Indeed any 
 consideration of quantity seems to me to a])ply with tenfold force to the 
 supposition of an s added to the k sound in Latin, so long as quantity was 
 regarded, or to the Italian tch, which surely must have been anterior to the 
 English or French s sound. 
 
 Yet more weighty to my mind is the fact that the Romans in all cases 
 expressed k by c. In old times they could only reproduce Greek words in 
 the rudest way; but for several generations this nation of philologers 
 expended vast energy in overcoming this difficulty. For this purpose they 
 introduced no less than five * diacritical' letters or combinations of letters, 
 
 1 It strikes me as improbable that Ulfilas, after years of intercourse with Eoman 
 dignitaries in Constantinople during its early days, and living with his flock in the 
 midst of Latin-speaking nations, should have his got Latin words through any ♦ Greek 
 transliteration'; and, as to the form aivaggeU, surely although in modern Greek 77 
 and in Italian ng are alike sounded as ng, the very fact that the Greeks put 7 for v and 
 that some of the best Roman Grammarians wished to write in Latin aggulus, aggens, 
 iggerunt and the like, prove that it was different in ancient times. 
 
 41 ill 
 
 I 
 ) 
 
 
 „r 
 
 
 I 
 
 '■I 
 
 19 
 
 y, 5f, cJiy ph, th, in order to reproduce wdth the nicest accuracy every Greek 
 sound; and schooled their tongue to utter words which once were most 
 strange to them. At first content with Teses, they finally brought them- 
 selves to adopt TJieseus, a sound and intonation most alien to a Roman 
 ear. Long satisfied with Saguntum, with sejmncs or sepirm, hicinus or 
 licinus, they came at last to Zacynthus, zephyriis, lychnus, containing each 
 of them three letters or combinations of letters utterly foreign to^them. 
 So that at length they learnt to revel in such sweet sounds as Anthem, and 
 Mnestheus, and Actias Oreithyia. 
 
 Why then, when they had got to Cepheus, Cephalus, Chalcis, cUh>ra 
 and the like, if c was not exactly equivalent to k, did they not adopt here 
 too a ' diacritical ' letter ? One was at hand, more ready for use than any 
 of the five adopted, their own k, now lying idle, with only an antiquarian 
 value before a in a few words or symbols of words. And on this point the 
 dekemhres of no. 844 of the Corpus inscr. vol. I seems to have some bearing. 
 This is one of nearly 200 short, plebeian, often half-barbarous very old 
 inscriptions on a collection of ollae. The k before e or any letter except a 
 is solecistic, just as in no. 831 is the c. instead of k. for calenclas. From 
 this I would infer that, as in the latter the writer saw no difierence between 
 c and k, so to the writer of the former k was the same as c before e. 
 Perhaps keri tells the same tale, if, as Mommsen assumes, it be the geni- 
 tive of cerus (creator). 
 
 The following too appears to me to have no small significance. In 
 Cicero's time from an abuse of Greek fashions the aspirate was permanently 
 attached to a few Latin words. Cicero tells us (Orator § 160) that till late 
 in life he had persisted in saying pulcros, Cetegos, triumpos, Cartagimm-, 
 but after a hard struggle e\dl habit and public opinion forced him to insert 
 the h in these words. It appears now from inscriptions and QuintHian 
 (i, 5, 20) that this h, which in some words was permanent, in others not 
 was attached to c alike before a, o, u and e, ^: in the 1st vol. of the Corpus 
 inscr. we find Volchacia and Achilio (Acilio); often Fulcher, but also 
 Fulcer. We have Gracchus and Graccus, GraccUs and Graccis : Quintilian 
 refers to what he calls Catullus' 'nobile epigramma' Chommoda dicehat, 
 and says that some inscriptions still extant have choronae chenturiones 
 praechones. It is I believe generally allowed that the ancient sound of 
 6, (t>, X was that of the tenuis with a distinct h sound attached to it. Pnt 
 even conceding that ch was like the modern Greek or Scotch or German 
 guttural, in either case I do not well see how the aspirate could have been 
 attached to the c, if c had not a k sound, or how in this case c before e or i 
 could have differed from c before a, 0, u. 
 
 And finally, what is to me most convincing of all, 1 do not vw ]] m .lor- 
 
1^ 
 
 20 
 
 ijtand liow in a people of Grammarians, where for 700 years from Ennius 
 to Priscian the most distinguished writers were also the most minute 
 philologers, not one, so far as we kno^-, should have hinted at any difference, 
 if such existed : neither Ennius, Accius or Lucilius, the three greatest of 
 the early poets j nor Cicero, Varro or Ctesar; nor Pliny or Quintilian, nor 
 Gellius, Charisius, Donatus, Servius or Priscian. Lucilius devoted whole 
 books to such slight matters as the use of fervit or fervet ; i or ei in termi- 
 nations. Cicero in his Orator and elsewhere dwells on what seem to us 
 very trivial minutiae. Varro asserted that lact was rif^ht, lac wronir; 
 Cajsar in his ' de analogia', addressed to Cicero, maintained that Varro and 
 lact were both wrong, lac alone right. He told Cicero that the genitive of 
 their common friend Pompeiius' name ought to have three H^ and explained 
 how they v.- ere to be pronounced ; but seems to have said nothing of the 
 ^'s in Cicero. Quintilian tells us how to pronounce the i of ojnimus, the 
 final e of here, and much else of an equally important nature. And all 
 know that Gellius, Servius, Priscian and the rest are brimful from first 
 to last of the most insignificant details : but of a soft c not one syllable. 
 
 Nay, what is even more to the point, Priscian relates at length how 
 Pliny heard three different sounds of ^: an 'exilis sonus' as in ille: a 
 'plenus' as in sol: a'medius' as in lectus. So Priscian himself finds 
 the n of nomen to be * plenior ', that of annis to be * exilior ' j and not only 
 is there a difference in final m, but the m of marjnus ^apertum sonat', the 
 m of umhra ' mediocre \ Of c ovhl ypv, singular indeed if its sound differed 
 perceptibly before different letters ; for surely the distinctions in the letters 
 just enunlerated cannot have been so very great. 
 
 Quite as little classical authority can I find for our strange confusion 
 of sounds in many classes of words, important from their great number, as 
 they happen to occur in so many common inflexions : I speak of ce, ci, se, 
 si, ti, coming before another vowel, to all of which we give the same Hebraic 
 ^eAshioi/ sound: iaceam, placeo, iacies, faciunt, coiulicio; nausea, caesius 
 divido', ratio, gratia, retia, otivjn, indiUiae, etc. etc. The modern confu- 
 sion of sounds here comes I believe not from classical times, but from the 
 'colluvies gentium' which met together on the breaking up of the old 
 world. Mr Muller sajs Corssen has 'proved (p. 5i) that from about 
 200 A.D. words with ti began to be spelt with ci How was that possible 1 
 if ci was always pronounced ki, then assibilated ti could never have been 
 written ci' The 'never' is surely too much: Ribbeck in his prolegomena 
 to Virgil, p. 241, gives dozens of instances where one or other of his capital 
 Mss. writes c for ^ or ^ for c; such as ac for at, tetera for cetera, tumulos 
 for cunmlos, etquis for ecquis, in none of which can the two letters hav e 
 had the least similarity of sound. But he gives not a single instance of 
 
 iff 
 
 il 
 
 
 I 
 
 
 21 
 
 confusion in a capital Ms. between the ci and ti in question: thes< ^F s. 
 write without fail dicio, solacia, fades, proditio, seditio, ratio, spatium. 
 And yet almost every line of Latin offers opportunities for blundering on 
 this point. When we consider this, the half-dozen instances in Corssen 
 seem quite inadequate to prove confusion between ci and ti. For there are 
 but six which have even a prima facie look of sufficiency : the most pro- 
 mising of these is renunciationem from a Poman inscrij^tion of a.d. '2\ \. 
 But when we examine its pedigree, we find that Orelli copies it from 
 Reinesius' collection 'quibus nihil imperfectius vitiosiusque extet,' says 
 Lie. Gronovius: 'ipse lapides nullos viderat,' says another scholar: 'who 
 exceeds all bounds in saxa violentius grassando,' says a third. When we 
 remember then that in Reinesius' time remmciatio was the recognised 
 spelling, that one instance after another of conditio for example vanishes 
 when it can be put to the test, surely the chances are a hundred to one 
 that the c is due to Reinesius or some previous transcriber, not to the old 
 Roman chiseller. Two more of unknown age are due to old copies taken 
 when ocio at least was a received spelling: two more are published by 
 Renier from a copy taken by a French officer at Medjana in Africa, Africa 
 great mother of barbarisms and heresies. The Gth has an unquestionable 
 voucher: Mommsen's inscr. reg. Neap. 109 has disposicionem. It was copied 
 at Salerno; but it must be late and is very barbarous, containing also 
 rivocaverit, distituta, jwpidusqiiae, an unmeaning suetad, the language 
 being in part unintelligible. Had Corssen applied his vast industry to 
 post- classical times, he might have collected without effort 100,000 clear 
 instances of the confusion in question, the only reason with many ap- 
 parently for writing racio, spaciurn, faties, speties being that the spelling 
 was wrong. We still see some relics of this barbarism of the middle a^^es 
 in coruUtio, solatium, novitius, tribunitius, nuncius, and the like\ 
 
 We have however late classical authority of the 5th century for a cor- 
 ruption of ^i (not ci) : Servius tells us that medius was pronounced medsius, 
 something like the Italian mezzo \ Pompeius, probably of the same age, 
 informs us that it is a fault to say Titius, not Titsius. If therefore we 
 prefer the 5th century to the age of Cicero and Quintilian, we should say, 
 not Tishius, llorashius, but Titzius, Horatzius : but then to be consistent 
 we should also say medzius, commodzius. From the strange emphasis with 
 which Pompeius asserts that Titsius is right, Titius wrong, I should infer 
 
 ^ The supposed concupiscenciae for concupiscentiae in an acrosticli of the African 
 Commodian, a comparatively early, but very barbarous writer, vanishes, when it is 
 Been that the sense and context require Turn (or, Tu) vigila for Cum vigiJas. This 
 correction which I made myself many years ago, has been made by Haupt in a recent 
 number of the Hermes. 
 
22 
 
 that this was a new fashion; and that laiktio represented to Ulfilas the 
 sound of lectio in his day, while kautsjogSLve the sound oi cautio in the year 
 551. In Servius' time the natural feeling for quantity was utterly gone: 
 1! 111.] t., I), learnt as artificially as it is learnt now. But in earlier classical 
 ^''>^*^ " "^i pronunciations were out of the question. Indeed if we are to 
 observe quantity, whicli many of us think a vital part of reform, I hardly 
 kiiusv iiuw with any of the modern fashions of pronouncing we are properly 
 to enunciate ratio and Horatius, fades and solacimu, 
 
 Tkinitt College : February 1871. 
 
 ■I 
 
 POSTSCIM 
 
 The preceding Remarks were privately circulated some months ago 
 among the masters of schools, the members of the Universities and 
 others who I thought had a right to be con^uitcd, m nnirht feel an 
 interest in the points discussed. I now publish thim wiih tin' pre- 
 sent supplement for more reasons than one. 1) i ring tin months of 
 February, March, April, May and June of this year several questions 
 concerning pronunciation were argued at length in the pages c^f the 
 Academy, by myself and others. Again our distinguished scholar 
 and historian, the Dean of Ely, wrote a paper in the Contemporary^ 
 for April on 'the classical pronunciation of Latin'. :\ii Uoby too has* 
 recently published his Latin Grammar in which the sounds of the 
 language have been treated of very fully and very ably. Tii most 
 points I am happy to say we are entirely agreed: on the few in which 
 we differ I will take this opportunity of saying something. Whether 
 the headmasters of schools have acted wisely in raising the question of 
 pronunciation at all, I am unable to decide. As I said in the Aca- 
 demy, it may be better to do nothing: the spectre of Greek is ever 
 before my eyes. It seems unreasonable to reform our Latin piu- 
 nunciation and leave our Greek unchanged. And I for one know 
 not what to do with the latter and its avenging accent. If however 
 we are to change, I am for as complete a change as possible. 1 1 . this 
 I am glad to have the support not only of Mr Roby but of Mr J. 
 Rh;^s in his very valuable letter, printed in the Academy of '\I:f\ 1 
 in which he shews what may be gained for Latin pronunciation from 
 the Welsh. 'In the Oxford circular' he says 'it is proposed to make 
 only a minimum of change in the present pronunciation ui liiu Laiii 
 vowels. This seems to be treating the present fashion with too much 
 tenderness: if indeed a change be desirable, let it be an adequate 
 one '. Few of the headmasters and others to whom I sent my Re- 
 
' 24 
 
 marks, have told me what their views and wishes are. Yet most of 
 those iiuiii whom I have heard appear to be in fawur of a complete 
 rliange. Mr Grignon of Felstead for example tells me that he and 
 his boys alike find a complete more easy and agreeable than a partial 
 change; and Mr Farrar of Marlborough does not think it more diffi- 
 cuiu Let us try our utmost, we shall at best be far enough from 
 perfection. 
 
 One of the questions discussed at greatest length in the Academy 
 was whether consonant u was to be sounded like EnoHsh lu or EnMish 
 V. The sole champion of the v was Prof Kobinson Ellis, who argued 
 the point at great length and with a zeal which far outstripped his 
 discretion. So far as I, or any one else whom I have consulted, can 
 see, his premisses had little or no bearing on the real questions at 
 issue, and his conclusions as little concern with his premisses. Thus 
 w^hen I liad quoted Gellius in p. 9 of my Remarks to shew that 
 Cicero's friend Nigidius Figulus pronounced vos as wods, he retorts 
 upon me that the passage on the contrary is in favour of a v sound, 
 and clinclies his retort by the dictum of a later grammarian ' V ore 
 constricto labrisque prominulis exhibetur'. The grammarian is really 
 telling us by what position of the mouth and lips we are to pro- 
 nounce the voiuelu: if Mr Ellis does not feel the humour of this, 
 then surely his sense of the ludicrous is undeveloped. I regret to 
 speak thus of a learned brother Professor; but clearly this is a case 
 of afZ(l)OLu ovTOLv (piXocv K.T.X. and Mr Ellis ought in charity to re- 
 member the thankless labour he has imposed on me and others. Let 
 me again refer to the elaborate exposition of this question by 
 Mr Roby in p. xxxii— XLII of his grammar ; and quote from the 
 Academy some words of my own : I cannot help inferring from the 
 hideously barbarous forms in which the older Greek writers express 
 Latin words, that ov came near to the sound of consonant u and that 
 Cassius Dion for instance could precisely reproduce aduentus only by 
 dhovevro^ : that /3 on the other hand was a substitute like the </> fjr 
 /, or the Italian and French gu or g for the old Teutonic w. But 
 even in French we find ouest for west, and Italian and Spanish 
 guai, French oiiais, whether they come from Latin vae or an older 
 form of German 7veh, as well as guastare and gater, which they would 
 seem to have caught up from their Teuton masters before the w 
 sound of vastare had passed into the Romance v, all tell the same 
 tale. This is well illustrated by what Mr Rh^^s tells us in his 
 
 m 
 
 25 
 
 most important letter referred to above: Latin words beginning 
 with V begin with gw in Welsh : Giuener = Vener-is, gwenwyn 
 = venemnn, g wain = vagina, giuaiul = vallum. But in words taken 
 from English initial v is represented by h or in: thus we have man- 
 tais, menter, milain, herf, etc. from vantage, venture, villain, verb, etc. 
 Now Welsh is especially valuable here, when we consider the period 
 \\ ithin which such words must have been borrowed from the Romans. 
 Mr Ellis greatly darkened counsel by his uncritical treatment of 
 the confusion which finally took place in many words between h and 
 consonant u. After exposing this at length I thus summed up: 
 neither the old Greeks nor Romans, I believe, had our and the 
 Romance v sound: h and consonant u were kept quite distinct in the 
 best classical period. Perhaps towards the end of the second century, 
 when many other symptoms of decay began to shew themselves in 
 the language, a lazy confusion of the two crept into the utterance 
 of many words. But amavi was pronounced amai before the v got 
 its present Italian sound, which was not fully developed till post- 
 classical times. We sliall thus I think best account for Italian 
 amai = amavi, as well as amavo = amabam; for ho hai ha ebhi = habeo 
 habes habet habui, as well as avevo = habeham, etc. The early cor- 
 ruptions of laborai and the like for laboravi, etc. seem to postulate 
 a quasi iv pronunciation of consonant u, 
 
 I will now refer to the few points in which there is a serious 
 difference of opinion between Mr Roby and myself He has become 
 somewhat more favourable to Italian ; w^e are agreed now about e and 
 o; but differ somewhat as to e, ae and o. At present I will only 
 say that I do not know how the facts I have briefly touched upon 
 in page 9, connecting old Latin with modern Italian, can be looked 
 upon as of slight moment in the discussion : surely on such slippery- 
 ground one fact is worth a hundred theoretical considerations. Every- 
 day I feel more and more certain that no consistent reform can be 
 carried out, unless we look up to the Italian vowel system as an 
 ideal to be aimed at, if it cannot be always attained. As I said in 
 the Academy, if any change is made, be it gre^t or small, I am 
 convinced that the mainstay of an efficient reform is the adoption 
 essentially of the Italian vowel system : it combines beauty, firmness 
 and precision in a degree not equalled by any other system of which 
 I have any knowledge. The little ragged boys in the streets of 
 Rome and Florence enunciate their vowels in a style of which princes 
 
26 
 
 27 
 
 iiiigiit be }.i -ud. :Nuxt to liiti Italian I would acquiesce in the 
 Gennnn vowel system. The Scotch is to me disagreeable, the French 
 
 alinu^r h|,jirr< 
 
 i I I S 
 
 M lloby asserts that 'the Italian has sprung not 
 ir 111 tlio cultivated language, but from one or more rustic provincial 
 dialects', that ' Italian is not the child of classical Latin, but of one 
 or more unsubdued dialects'. I on the other hand hold it to be 
 d. in.iistrable that the Romano-Tuscan is the child of cultivated 
 Latin, falling to pieces, and caught up and subdued by German 
 mouths. With reference to the Latin I and u as represented by 
 Italian close e and o, to wliich Vi: Tloby applies this theory of 
 dialects, I would call attention to an interesting article which has 
 juM a|>peared in the iiurmes, on Roman names in Greek inscriptions 
 and literature. It is there shewn that the oldest inscriptions and 
 writers continually represent Latin I by e, that for instance the name 
 Tiberius is habitually reproduced by TeySe/9to9, until Tiberius' acces- 
 sion to the empire officially decided its spelling in Greek. This will 
 explain the connexion of Tevere,fede and the Hke with Tiberis, fides 
 etc. far more simply than 'one or more rustic dialects'. 
 
 To come to 5: Mv lloby and I differ with regard to the compara- 
 tively few cases in which s, not representing a real ss, comes between 
 two vowels. But here too I would appeal to the interesting analo- 
 gies in Italian to which I refer in p. 13. The Italian too is strongly 
 supported by late Greek: we find Kaaao<; (casus), Kovpioaao^ (curiosus), 
 ^afiwaaa (famosa), i^Kovaararo^; (excusatus), i^KovaaareveLv (excu- 
 sare) and the like. What is the meaning of tliis aa, if there was 
 no difference between the s of casus (cassus) and casa, of rdsus 
 (rossus) and rosa ? i quite agree with Mr Roby that in the case of 
 ns which always made the preceding vowel long, the n was almost or 
 quite mute and the s sharp: comp. 7)^ = ens, toties = totiens : Varro 
 and Pliny tell us through Charisius that frons =fros, /runs =frus : 
 formonsiis too = formdssus = formdsus, all three being pronounced 
 alike. The case is much the same with prorsum prdssum prdsum, 
 rursum russum rilsum, and similar forms. 
 
 Only one other point have I to contest with Mr Roby, one which 
 however I did not expect to have to contest. In p. LXXiii he writes 
 'Mi Munro takes Priscian's statement (n 63) that the vowel before 
 gn was always long, as meaning that the gn makes it long by nature'. 
 But I would ask how I could possibly take his statement as meaning 
 anything else, when after an elaborate explanation of the long vowel 
 
 ■»^ 
 
 
 of adjectives of places in -Inus and -amis, ho tlms continues: gnus 
 quoque vel gna vel gjium terminantia longam habent vocalem paeniil- 
 timam, ut regnum stdgnum bemgnus mallgnus abiegnus privlgnus 
 Paellgmis. Priscian distinguished most carefully vowrl^ Innj bv 
 nature, and short vowels which had position: thus, elsewhere he 
 draws conclusions from the short i oi firmus, virgo and vix. M.\ R. i v 
 and I are I believe pretty well agreed as to the general liiLiu^ of 
 Priscian ; that for his age he was a very learned man ; but that age 
 was the 6th century. Of necessity therefore he was wholly de- 
 pendent on older authorities for all that he tells us worth know- 
 ing, and when he trusts to himself, he falls into imbecihty. It is 
 absolutely certain then that the passage in question comes directly 
 or indirectly from some old and good authority; since we can see 
 from the loose gossip of Gellius that even in his day the most learned 
 men were losing the instinctive feeling for the natural length of 
 vowels which had position ; certain I repeat that it is servilely copied 
 in fact from some one like Varro, or Verrius Flaccus, or Valerius 
 Probus, one who had the same indisputable right to speak on such 
 points, as Cicero has when he tells us that the in of insanus and 
 infimus is long, of inclitus short by nature. Surely without Cicero 
 this would seem at least as improbable, as that the vowel before gn 
 is long by nature in ignarus, cognitus, etc. 
 
 How could Priscian of himself know anything of what he here 
 tells us ? But Mr Roby says he cannot agree with me and thinks 
 Priscian unsupported in his statement : ' Priscian on his principles 
 could come to no other conclusion ; for he held that gn began the 
 final syllable and that gn made a preceding syllable common, i.e. 
 allowed a short vowel to remain short (l 11, Ti 10^ Hence finding 
 all words which ended in gnus had the penultimate long, he con- 
 cluded the voiuel must be long'. And surely in all this Priscian is 
 quite right, and his statement may be supported by a variety of 
 arguments: gn did allow in Latin a short vowel to remain short 
 before it, and yet in every Latin word the vowel is inflexibly long 
 by nature : si-gnum, tl-gnum, i-gnotus, co-gnatus, etc. etc. the g and n 
 being locked in one another, and the vowel rendered long by nature; 
 just as in the very similar instance of ns we find toties = totiens, 
 K\TJfirj<; = Clemens, and again and again in old Latin cosol, cesor, 
 cosentiont, cdstanti and the like ; while before all other consonants 
 except gn and ns the con or com remains short by nature and does 
 
28 
 
 not lose its consonant : thus we have congredior, congliitino (as well 
 as ingredlory ingluvies and the like), never cogr^edior, etc. ; while with 
 gn not only do we find co-gnitus, co-gnatus, as well as i-giiotus etc., 
 but the g too is absorbed and disappears in co-nexus, co-nitor, co- 
 iiuhium^ co-niveOy the double n in all these words being a gross 
 and admitted solecism. The con remains short I repeat before all 
 other consonants ; for not oidy have we c6 before vowels, but Con- 
 sentius tells us (p. 400 Keil) the curious fact that Lucilius, dropping 
 per metaplasniuni one r, says ore corupto. And, though formerly I 
 was biassed by the authority of Lachmann, I now feel no doubt that 
 his MSS. are quite right in giving to Lucretius (vi 1135) natura 
 coriiptum. 
 
 Why I dwelt on this point in my paper was mainly in conse- 
 quence of Corssen's note (ii p. 2G5) where he assails Hermann 
 and Ritschl's irrefragable dictum that the old scenic poets never 
 allow a mute and liquid to give position to a short vowel: they 
 always have lugri, Idhruni and the like, while the hexameter poets 
 much prefer to lengthen these syllables when the mute is a medial. 
 He quotes a great number of these words in which he says a short 
 vowel is lengthened by Plautus and Terence before gn. And indeed, 
 just as the vowel should be short in insanus, consul, sapiens and 
 the like, but for this peculiar power of ns; so should it be short in 
 ignarus, ignotus etc. but for this cognate power of gn: also in sig- 
 num, tignum etc. as shewn by sigillum, tigillum etc. In adopting the 
 Greek KVKvoi;, UpoKvr} Plautus could not say either ciignus or cfignus, 
 Frogne or Prdgne: he wrote cuclnus and Procina. The short vowel 
 followed by the two distinct consonants were an impossibility to his 
 and his hearers' ears: in Latin words he and they knew only 
 si-gnum, tl-gnum, co-gnatus, etc. Yet at the same time, though he 
 knew only lahra, pigri, while Horace and Ovid preferred lahra and 
 'pigri, Horace can write Donatura cygni, Ovid Ad mandata Prognes, 
 proving surely thereby the statement of Priscian; for their MSS. 
 give gn and not en, and even if we assume that they wrote en, gn 
 was always the prevailing spelling and therefore pronunciation. 
 And why does Mr Eoby overlook the fact I stated that in inscrip- 
 tions of high authority, not likely to err on such points, Ave find 
 slg7ia, digna, the long i implying that the vowels are long by na- 
 ture? we Hnd too regni and regno with the apex, and Corssen says the 
 root-vowel is short, though he may be wrong in this. Ptolemy gives 
 
 
 
 29 
 
 the name of 'Vrj^voi. to a people in Britain, "KwyvoL to a people of 
 Germany, in both cases doubtless getting the names from Roman 
 sources, while two cities in India he names Ko^vaBavha and Koymi/- 
 hava, and, but for Latin influence, the short vowel would seem most 
 natural judging by all analogy. But 'the Latin words Egnatia, 
 Egnatius occur not unfrequently in Greek with e'. The words are 
 no more Latin than Diognetus, Poljjgnotus, Progne, cygnus: the town 
 is Pencetian, akin to Greeks, not Latins; and Mommsen tells us that 
 the native name is Gnathia, the genuine Latin form Gnatia; and cer- 
 tainly our two oldest authorities Horace and Mela know no other 
 form. If it be affirmed that Gnatia is Egnatia with the e dropped, that 
 would only strengthen my position. The Egnatii were no Latins : 
 until after the social war they were always bitter enemies of Rome. 
 I still therefore adhere to Priscian and to my opinion: and have no 
 doubt that what has been said of gn applies to gni as well: agmen, 
 t^gmen, etc. I have much difficulty in perceiving what Mi iljby 
 gains for his argument by the fact that the Verona palimpsest of 
 Livy (or any other manuscript) * always divides w^ords with gn 
 occurring at the end of a line between the g and n, so as to give 
 the g and n to separate syllables': the 'always' is self-refuting, 
 and proves, as Mommsen himself to whom he refers justly says, that 
 * veriloquii in talibus nullam omnino rationem haberi'. Mr Roby 
 I feel sure will not maintain that ig | nominia and ig \nota are ri^^htlv 
 divided; or that l-gnominia and i-gnota are not unquestionably true: 
 I should say the same of henig \ nitateni : heni-gnitateyn can alone be 
 correct'. ne\ glegens is equally absurd in the other direction. 
 
 The curious Latin Christmas chant which Constantine Porphyro- 
 genitus has preserved for us, written in Greek characters for Greek 
 tongues, and handed down traditionally for generations, lets us see 
 that many corruptions had long existed in the pronunciation of both 
 languages, while c had continued hard. One of the verses is as 
 follows: KoviJb KpovKT](f)L^ov<s icTT €T aeTTOvXrov; ev rep^ia Bteppeaov- 
 pe^cT, i.e. cum crucifixus est et sepultus, in tertia die resurrexit: 
 here we see that r] = i, as in modern Greek; and ti = fi, as illustrated 
 above in p. 21 at the end of my Remarks; but the c of crucifixus 
 is still hard. As then Professor Palmer in the Academy of March 
 15 informs us that the majority of the Oxford Philological Society 
 have now decided that c, g and t should be sounded hard before all 
 vowels alike, I should have concluded that enouo^h and more than 
 
30 
 
 eiiuu^ii 'had iili'jiidy buuii said on this head, if it had not been for the 
 Pt nil of Ely's article in the Contemporary Review for April, which 
 after :\ few general remarks is wholly devoted to the much vexed 
 L trill c. The opinions of so eminent a scholar cannot be slighted: 
 1 siiciil liiLrciuio repeat here the greater part of a letter of mine 
 wliirii rqiprrirrMl in tl,. AfMdemy of May 1: it will serve as a supple- 
 ment to the Appendix lo my Remarks. 
 
 : Let in. 1h gill by stating that the 'one or more persons of impul- 
 sive temperament' include among others the Philological Society of 
 Oxf rl I rid the headmasters of the forty or fifty greatest schools in 
 J.nuiii 1 Possibly enough, as I have always asserted, it may not be 
 wuith "11! while to attempt to recover the utterance and intonation 
 of a ducid language; but, as I have tried to show in my printed re- 
 marlcR, we have greater facilities in my judgment for approximating 
 tit ill' pronunciation of old Latin than Dr Merivale believes to exist. 
 We know exactly how Cicero or Quiutilian did or couhl spell ; we 
 kiiuw the syllable on which they placed the accent of almost every 
 word: and in almost every case we already follow them in this. I 
 have the conviction that in their best days that philological people 
 t.>.k vast |t liiis to make the writing exactly reproduce the sounding; 
 and liiat if Quui Lilian or Tacitus spelt a word differently from Cicero 
 ftr Livv lip also spoke it so far differently. With the same amount 
 ?'f evidence, direct and indirect, we have for Latin, it would not I 
 thnik be worth anybody's while to try to recover the pronunciation 
 ui i rench ui i^]nglish ; it might I think be worth his while to try to 
 roonv ^r that of German or Italian, in which sound and spelling accord 
 more nearly and accent obeys more determinable laws. 
 
 But on these general questions nobody will care for what his 
 neiglibuiu says or thinks. I proceed therefore to the special topic 
 of T)r ^Trrivale's paper, the pronunciation of the Latin c, of which he 
 wishes ' to say a few words before the hour has struck for fixing it 
 inovocably'. And here I would remark that he appears to me to be 
 too ii ir.i iii Madvig and 'his followers', whose theory as to the sound 
 of c does not depart more widely from our or the Italian practice 
 than does Dr Merivale's own; for he is the only scholar of distinction 
 
 s 
 
 31 
 
 I ever heard of who has maintained that the oil Creeks souii 1* 1 
 avT\o<; as anshlos, or that the old Latins pronounced exanclare as 
 exaiishlare, porta and porca as po7^sha. Dr Merivale puL- puui Cicero 
 once again to the question. He not only says what is true enough, tlint 
 the hardness of the Greek does not prove that of the Latin c; ^ n he 
 is not sure that the k was hard. I showed however in t !• t|>pendix 
 to my Remarks that deer was reproduced in Albauian by KycKXtpe, 
 in Gorman by kicher, which allows I think of a more concln>ivp in- 
 ference. This inference I will try to strengthen here. Yiv III v> in 
 the instructive letter spoken of above tells us that in W elsh c and g 
 are always hard, whatever vowel may or once did immediately fulluw 
 them. He jrives this list of words taken from the Latin, where c 
 precedes e or i: cella = cell, cera^ cwyr, ce7'tare = certhain, cingula — 
 cenr/l, cista = cist and cest, cicuta = cegid, civitas = ciived, civitat-is — 
 ciwdod, cippus = cyff, plur. cyffion. In 'inlaut' Latin c regularly be- 
 comes g in Welsh: cancelli = ex gell, ascendeve = esgyn, descendrre = 
 disgyn, dejicere = diffyg, locellus = llogell, inaceria = inagwyr, medici- 
 na = meddyginiaeth, pascere =pesgi, jjraeeeptum = pregeth. ' Welsh 
 words ofive no evidence of a hesitation between ci and //. ibus nata- 
 licia is postulated by nadolig\ 'It would be useless to multiply in- 
 stances, as there are no exceptions tending to prove the Latin c to 
 have been pronounced soft before i or e, excepting a few of modern 
 date. The same argument maybe extensively supported b) iii>uaices 
 from the other Celtic languages'. 
 
 Mr G. W Alason too of Morton Hall has sent me a list of Wish 
 words derived directly from the Latin ; among them ceirios = cerasus, 
 carchar = career. No one I think will doubt that these words were 
 taken directlv from the livinoj Latin, as the Latin words were from 
 the living Greek. Equally certain it is that Gothic karkara an 1 
 German A'erZ:er came directly from the living Latin, anl lirsche too 
 directly or indirectly. Thus then the old Siceliot and Italiot /cdpKa- 
 po? {KapKapoi'), career^ carehar, karkara, kerker, were all five sounded 
 hard; or else a soft Latin c reproduced the hard Greek k. while all 
 the three languages which took the word directly from the Latin 
 agreed in returning to the hard k sound, though they all might have 
 used sibilants to express it. Which h^^pothesis is the more likely? 
 The argument is not affected if, as "Or "Merivale w< nild deem possible, 
 Kdp/capo<; was pronounced sarsaros: K€paao<f, cerasus, ceirios, kirsclie, 
 tell the same tale: klott} of Homer is in Latin cista : this same 
 
32 
 
 83 
 
 word, borrowed from Latin, reappears with hard c or h in Welsh and 
 other Celtic languages, in old German and new German, in Scotch, 
 Dutch, Icelandic, Swedish and Danish, under one or other of these 
 forms, cist, cest, cisde, ciste, kest, kista, kiste, hist : compare too Caesar^ 
 Kalaap, and Gothic, German and English, keisar, kaiser, kesar, etc. 
 And yet, as Corssen shews, when the Germans borrowed words after 
 the Romance sounds were introduced, they employed a sibilant : zelle, 
 zirkel, zither. 
 
 Dr Merivale casually opens his Propertius and finds that the MSS. 
 give cilices for silices: the MSS. of Propertius are of the late middle 
 ages, and from those times 1(),()()() similar instances can be got. The 
 same may be said of the arscedat of Turpilius, where the true reading 
 IS by no means so certain as Dr Merivale seems to think it: the best 
 MSS. of Nonius, from whom the passage comes, belong to the earlier 
 middle ages. Such blunders being past numbering in post-classical 
 tmies, unless it can be shewn that they were usual in classical times, 
 my inference would be the exact contrary of Dr Merivale's. And so 
 with the ' fluctuation so common in our printed books and MSS. of ci 
 and ti pure', unless it can be shewn that such fluctuation took place 
 in classical times; and I flatter myself I can shew it did not. And 
 here let me express my absolute conviction that with regard to the 
 words he specifies every scholar whose judgment is held to be su- 
 preme on such points, such as Ritschl, Mommsen, Huebner, Haupt, 
 will at once declare that the sole classical forms are condicio, contio, 
 nwitius, novicius : reason, authority and etymology are here at one! 
 Accius and Attius, Mucius and Mutius, may all exist, but the dif- 
 ferent forms point to different origins. Portia I know in Shake- 
 speare, but only Porcius and Porcia have classical authority ; and so 
 with Aehutius, Caedicius, Sidpicius, Itius. *lt is clear that both 
 writers and printers gave a common pronunciation to ci and ti pure\ 
 Yes; but what writer or printer before the 7th century? 
 
 Dr Merivale asks, *Is not porca= porta, the gateway or space 
 between two furrows? Again, is not porca=:7rlpTc,, the pregnant 
 sow or heifer V porcus then will be the pregnant boar. ' Is not Mars 
 identical with J\Iarcus, and accordingly Jfartius = Marcius T So we 
 are to pronounce, it seems. Marsh MarMs Marshiiis, porsha porshis. 
 a surmise, with some confidence, that if one transcriber wrote exa7i- 
 tlare and another exanclare, it was because the word was pronounced 
 exanshlare. And this too represented the Greek cli^rXo^ (anshlos)' 
 
 i 
 
 I have never before known English;iian, German, Frenchman or 
 Italian go such lengths in sibilation as this. For the first porca let 
 me refer Dr Alerivale to furrow, furche, and Grimm's law ; for the 
 second to farrow, farch; [ferkel, porcellus), and the same law. See 
 too Max Muller on the Science of Language, vol. I p. 296 (6th ed.): 
 ' the Latin porca, a ridge between two furrows, is derived from porcus, 
 hog; and the Gqyvluxxi faricha, furrow, is connected y^ii\\ farah, boar'. 
 Mars Martis Martins are all in rule : Marcus, comparing Mamercus 
 and Tiberius, I believe to be a child or favourite of Mars ; and from 
 Marcus comes Marcius, as from Quintus (Quinctus) comes Quinctius, 
 from Sextus Sextius. But what all this has to do with the sound of c 
 or t I cannot conceive. Ought a foreigner to maintain that the Jc has 
 an s sound in Tompkins and Wilkins, because they are etymologically 
 connected with Thompson and Wilson^ Exancidare and exanclare 
 (anculo, ancilla) are the sole classical forms : the word has no more 
 concern with avT\o<; than with the German antlitz, or English antler. 
 The learned Theban who first wrote exantlo wished to connect it 
 preposterously with Greek : perhaps it was he who first discovered 
 coeluni sylva hyndjres and similar flowers of etymology. Surely the 
 time is come to discard such fancies. 
 
 But Dr Merivale concedes that 'the usage of the 12th or 14th 
 century is not conclusive as to that of any preceding one'; and adds, 
 ' we possess probably no manuscript authority which goes back to the 
 classical or nearly to the classical ages '. Of Virgil alone, or of por- 
 tions of him, we possess at least six MSS. which go back to late 
 classical times ; and every one of them utterly repudiates any such 
 confusion of c and t — unless, that is, we are to pronounce ac and at as 
 ash on the analogy of anshlos; sic and sit as sich; for these little 
 words are occasionally confused. 
 
 How can infitiae come from the Homeric dficpaair] any more than 
 fateor and conjiteor from 0?;o-/? Gellius knew no secius : he himself 
 said sctius ; he or Varro seems to have found sectius in a manuscript 
 of Plautus: we are then to pronounce quinctus and quintus as 
 quinshus on the analogy of anshlos; vita and victus as visha and 
 vishus. Is an extruded c before t so foreign to Latin ? what of artus 
 fartus sartus ? Of course Varro, hke everybody else before A.D. 600 
 odd, ^^e\i jwetiuni with a t', but the jingle 'nunc prece, nunc pretio' 
 I prefer with the sounds not entirely coincident, i.e. either as we or 
 the Italians pronounce it, or as I believe Ovid and Horace pronounced 
 
 I 
 
34 
 
 it, 'nunc preke, nunc pretio\ I am compelled therefore to doubt 
 that it * was doubly alliterative'. 'Varro' Dr Mcrivale adds 'has a 
 curious passage on the word feciales : quod fidei publicae praeerant ; 
 nam per hos fiehat {i.e. hi faciehant) ut justum conciperetur bellum, 
 ^i...ni foedere fides pacis constitueretur : as if to his ear the d and c 
 ill Fidiales, Foediales, and Feciales had much the same sound. 
 Which is the most likely ; that it was the sound of k, or of dh, th, or 
 sh V Let Dr Merivale rest assured that to Varro, and his countrymen 
 for six centuries after him, fetiales was the only known ^orn\, feciales 
 a nonentity. We have merely therefore to mediate between t and 
 d, an easy enough feat in the case of Varro's etymologies. 
 
 Dr Merivale comes at last 'to the testimony of monumental 
 inscriptions ; and about these there can be no misapprehension'. My 
 experience, alas ! is widely different. He observes that whoever 
 glances over Orelli's Corpus Inscriptionum, will find conditio and 
 condiciOj statio and stacio, nuntius and nuncius, fetialis and fecialis. 
 It is true he will ; but let me shortly illustrate this point. Of all 
 these words there is one and only one right spelling — condicio, statio ^ 
 nuntius, fetialis: whenever you find in Orelli the wrong spelling, it is 
 taken from some old collector who took it from some one else, or 
 blindly followed the depraved spelling of his time ; and whenever a 
 competent observer can test such spelling, it is found to be a blunder. 
 Thus in no. 4132 Orelli gives conditione: Mommsen has copied from 
 the original and published the same inscription, and reads condicione : 
 'utri creditis?' Another inscription Orelli takes from Gruter, Gruter 
 from an old and worthless authority Apianus. Gruter gives this 
 inscription twice over : once he prints condicione, three times con- 
 ditione: his authority Apianus once prints condictionel All these 
 works I have before me at this moment. Did space and patience 
 allow, all the other false spellings might be routed in the same way : 
 patritius is absolutely without authority. It is met with in Orelli, 
 not ' often ', but in no. 723 : this he takes from Gruter ; Gruter, as 
 Orelli remarks, gives this same inscription twice : in the one place he 
 prints patritios, in the other patricios ! Orelli published his collection 
 in 1828, and we are now in the year of grace 1871. 'On the other 
 hand aedilicius, sodalicium are spelt with c': yes, and c is alone right, 
 t wrong de facto and de iure. One or two of the names mentioned 
 have a double spelling, but also a double etymology ; but SulpiciuSy 
 Caedicius, Aebutius, are the only true forms. ' The inference seems 
 
 irresistible'. What inference? Surely this, that, while an ancient 
 stonecutter could not confound ci and ti pure because to him the 
 sounds were totally different, a modern transcriber, before the quite 
 recent introduction of a real epigraphical science, could not be trusted 
 not to confound them because the sounds were to him the same. 
 Thus wdth regard to the just-mentioned sodalicium the accurate and 
 accomplished Huebner for his new volume of Spanish inscriptions 
 copies from the original and prints sodalicium in no. 3730, while a 
 Spanish scholar in the year 1760 reads sodalitium, 
 
 I marvel how Dr Merivale could have made all the startling 
 deductions he has made, without refiecting on the countless chances 
 of error from the carelessness or ignorance or prejudice of transcribers of 
 transcribers, printers, correctors, etc.: thus in the supplementary volume 
 to Orelli, published by Henzen, one of the greatest of living epigra- 
 phists, I could scarcely believe my eyes when I saw in no. 5438 trihu- 
 nitia. Huebner however informs us that Henzen has written to tell him 
 that this is a mere misprint, the original having of course ti^ibu7iicia. 
 
 And now we are advancing to the final and triumphant climax, 
 ' But of all these monuments there is none that comes to us with 
 more authority both from its age and its authorship, than the 
 marmor Ancyranum, the inscription on which is the undoubted 
 composition... of the Emperor Augustus'. 'Now in this inscription 
 we find the word patritius spelt with a t» We find also the two 
 forms tribwiicia and tribunitia once or twice over. It is plain that to 
 the ear of Augustus either form sounded alike, and he therefore 
 wrote them indifferently. And so, if I may rely on the transcript 
 given by Egger (Ilistoriens dJAuguste), it would seem to be absolutely 
 demonstrated that in the best classical age ci = ti pure... and their 
 common sound, as I have already said, can only have been a sibilant, 
 jjuch as s or sh or ch\ But the transcript may not be relied on : let 
 Egger be a warning not to pin one's faith on the Reinesii, Apiani and 
 Gruters of old. For the last six years I have treasured, as one of the 
 most precious works for the study of Latin orthography, Mommsen's 
 elaborate monograph on the ' Res gestae divi Augusti'. Every con- 
 ceivable pains has been taken to ensure the minutest accuracy, every 
 letter has been studiously pored over. We find then patriciorum 
 once, tribunicius thrice : I need not say the t never occurs. Augustus 
 cashiered an officer who wrote isse for ipse. What would he have 
 done to one who spelt 'patritius or tribunitius ? At the very least he 
 
m 
 
 ■«.*,■ ^^ 
 
 » 
 
 36 
 
 ;;;;;'i;"';;- --' iu,u u. k.op poo,- .,, ; l company. T^^ Dr Mcrfvalo 
 
 '^ :'">'■ "V. n ,nake ,t probal.le to me. that Divus Angustus once 
 
 ■;> i^ : -ote ,„ ^,, l..,,.,,,,.-,, or triUunit.po., i u.li .enonnce for 
 
 . ,. I.:.n. ,.ni,,grapL3 -.a,,, enunciation; T ,nH „V ,,eccavi, not 
 
 >ecc. „ , . I u ui i^romise to say il/a.,/,. Marshus, McrMus, and- But no 
 
 1 < :un..i |,!o,l,.o myself to pors/w and exanshlare. 
 
 Trinity College : September 1871. 
 
 i 
 
 CAM[;iii f)^; (.; 
 
 !''^'^'T£D BV C. J. CLAY, M.A. A'l 
 
 r THK UNIVKRSITV PTJKSS.