F^Jjt^Ti l^ipilf 
 
 'Mil 
 
 m? 
 
 ; .1 .' «i 
 
 1 . .■■ . ^..^■■^^. .</ 
 
 ,^<t!4?;'^^ 
 
 ' 1 ' ', ' I' 
 
 ■'f-l' ,^'ibJr- ^ 
 
 
THE N1IWLSERI[S OF SIHIIDIieD SCHOOL-BOOKS 
 
 COMPRISES STANDARD "WORKS 
 
 lu every department of imtniction and of every grade. The teacher io want of a book for anyimr- 
 
 titular pnrpose or class, yrill always find the best of ile kind in onr eatalogne. No other soplcs even 
 
 claims to be as complete as this. Jfoue is so extensive or so judiciously selected. Among so many 
 
 volumes a hiB:h Stan lard of merit is maintained, as it is our aim never to permit our imprint upon a 
 
 poor or nmrorthy bo, k. It Is also our plan to make books not for a class or sect, but for the vhole 
 
 country— nnontectiona >la to parlies and creeds, while inoulcaUng the great principles of politiool 
 
 freedom and CbiisUanity. upon which all right-minded persons are agreed. Hence, and from their 
 
 almost nniversal circulation, the name—" National Series." Among the principal volumes are 
 
 Farlcer & Watson's Readers— in two distinct series, each complete in Itself. The 
 
 National B.eaaers, of full grade, in large, elegant volumes, adequate for every want of 
 
 the moEt thorough and highly graded schools. Tlie Independent Meadere, in smaller 
 
 vohimes, for Common Schools. Low in j)rioe, but in no other respect inferior to the companion 
 
 series. Spellers complete to accompany either series.. 
 
 Davies' Blathematicsr-^rit/tmetic. ^Zrre&ra. fieomet-,-^. 
 
 <XI PRESENTED l>o 
 
 TO THE LIBRARY OF 
 
 l^^ifT'fiu 4iU- ^ M^3iJ!^/^i^^^^^**^. 
 
 
 .uistorj^.— MOKTKiTB's Child's V. S.-Mes. Wilukd's Series; F. S. and tTnlversal-BKEiED'S- 
 England— BICOED'S Bome-Snmmary of Hist., in lOJ pages— Bible Hist.— Ecdesiaalical Hiat. 
 
 Fen and JPencil — Beers' Bound-hand Pomnanship-Copy-book Cover^National Steel Pena 
 — Smith i Maeiins Bookkeeping— Chapiun's Drawing Book— Drawing Cards— AliEK's Map 
 Drawing. 
 
 Ifatural Science.— 'Soman t Poethe's First Book— Peck's aanofs Philosoplff— 'Pobtee's 
 Chemistry— McIiJTTEE's Astronomy— Paoe's Gcology-njAEVls' Physiology— Wooij's' Botany— 
 Chahbees' Zoology— Peck's Mechanics— Baetlkt's College Philosophy. ~" 
 
 Important Works also are PtiJOI.'s Trench Class Book-DwiOHT's Mythology-HmraiHO- 
 TOK's Fine Arts— Chakplik's Political Economy— MAKsriEM's Oovemment Manual— Aidik's 
 Ethios- Brooks' Mannal of Dovotion— Tracy's School Record, Ac. 
 
 Tlie Teacher's Hftrorj^TOaslsts of over so volumes of strictly professional llteralnre, asPAGE'l 
 Theory and Praclicc-HoLEEOOK's Normal Methods— Noeihehd's Teacher's Assistant, 4o. 
 
 A DESOKIPTirE CATALOQTTE of all these and many more may be obtatoed by enclosing a 
 stamp to the Publishers, 
 
 A. 2. BAME2 & COMPAIY, 
 
 Irrational Educational Publishers, 
 111 & 113 WILLIAM STREET, NEW YORK. 
 
_ Cornell University Library 
 
 PA 2117.B63 
 
 Latin Dronunclation.An Inquiry into the 
 
 3 1924 021 614 007 
 
The original of tliis book is in 
 tine Cornell University Library. 
 
 There are no known copyright restrictions in 
 the United States on the use of the text. 
 
 http://www.archive.org/details/cu31924021614007 
 
LATIN PRONUNCIATION 
 
 PROPER SOUNDS OF THE LATIN LANGUAGE 
 
 THE CLASSICAL PEEIOD. 
 
 BY 
 
 -WALTEE BLAIR, A.M., 
 
 PBOFEBSOB 07 LATIN IN HAMPDEN SIDNEY COU.EGE, VIBalNIA. 
 
 A. S. BARNES & COMPANY, 
 
 NEW YORK AND CHICAGO. 
 
 1874. 
 
Entered according to Act of Congress, in tfte year 1870, l)y 
 
 WALTEE BLAIE, 
 
 In tlie Office of the Librarian of Congress, at Wasliington. 
 
PREFACE. 
 
 The following treatise was originally prepared by 
 appointment of the Educational Association of the State 
 of Virginia. Haying heen faTorahly received by that 
 body, it is now respectfully ofEered to a larger circle of 
 teachers and scholars, in the hope that thereby some 
 contribution may be made towards a just settlement of 
 the still Tery uncertain practice of our schools in the 
 matter of Latin Pronunciation. 
 
INTRODUCTION. 
 
 Quid enim tarn neeessarium qttam recta locuiio^ — Quint. 
 
 MANY languages haye been richer than the Latin 
 in variety of sounds; but no people were ever 
 more attentive at once to discrimination and refinement 
 in their utterance than the Romans. Their poets and 
 orators were held to a strict account in this respect ; and 
 the sedulous endeavor of these to satisfy the i claims of a 
 sensitive ear is apparent in their written performances. 
 
 In the way of direct testimony, moreover, the writers 
 on the fine art of elocution furnish abundant evidence 
 of the exacting judgment of a Eoman auditory. They 
 make it certain that during the most cultivated period 
 of the Latin literature, all who pretended to any refine- 
 ment of letters, and even the general crowd, which was 
 subject to the polishing influences of the metropohs, pos- 
 sessed and cherished the most delicate sense of propriety 
 in the rendering of the sounds of their language.* 
 
 When, therefore, it is recollected further, that the 
 compositions which have come down to us from that 
 distant period were never designed to be absorbed through 
 the eye, in silence, from a written page, but were dis- 
 
 * Cic. Orat, c. 44. . . . . "aures; quarum est iudicium superbis- 
 Bimum." Cf. c. 48. 
 
6 IN TEODUCTIOlf. 
 
 tinctly and consciously planned for tlie interpretation of 
 the tongue, we must feel how hopeless is the efEort to 
 penetrate to any true and intimate knowledge of the 
 Latin language, not only in its beauties but eyen in its 
 essential nature and powers, until we shall render the 
 signs before us back again into the sounds which they 
 were designed to represent.* 
 
 This necessity is patent with regard to poetry, the 
 whole texture of which is persistently controlled by the 
 exigencies of recitation: but it is also true (though a 
 somewhat closer observation is necessary to reyeal the 
 fact), that many an appearance, both regular and occa- 
 sional, in the structure of the plainest prose is directly 
 traceable to the influence of sound. 
 
 But notwithstanding the astonishing neglect of our 
 ancestors, for generations, of the proper sounds of the 
 Latin language ; perhaps no thoughtful student of the 
 present day needs any longer to be convinced of the im- 
 portance of their revival. And should there be any to 
 whom it is still necessary to justify the demand made 
 for an inquiry into the Latin Pronunciation ; if we can 
 satisfy them at all, we cannot hope to do so in a better 
 way than by inviting them to enter with us upon the 
 inquiry itself. Therefore, to this we address ourselves 
 at once, assuming for our present duty the endeavor to 
 satisfy as far as we can, rather than to justify the ques- 
 tion which has been assigned us. That question is, to 
 ascertain the " correct Accentuation, Quantity, and Pro- 
 nunciation of the Latin Language." These are three 
 
 * Hie enim est usus litterarum, ut custodiant voces et velut deposi- 
 tum reddant legentibus, itaque id exprimere detent quod dictwH 
 sumus. Quint. I, 7, § 31. 
 
IlfTBODUCTION. 7 
 
 distinct modifications of the yoice in speech, closely asso- 
 ciated it is true in practice, so as to exert an important 
 influence upon each other, yet quite separate in their 
 nature and causes. The pitch, the stress, the time, or 
 Accent and Quantity, are accidents of the vocal tone, 
 while Pronunciation, as commonly understood, has to 
 do with the essential quality and stamp of the uttered 
 sound. In a particular examination of these three, then, 
 the natural order to be observed will be the reverse of 
 that just stated, and it is the Pronunciation which must 
 engage us first.* 
 
 And here we conceive our task to be purely a practical 
 one, that is, to indicate, as far as we may be able, what 
 were the true and actual sounds of the Latin, furnishing 
 if possible, reasonable evidence of their authority. Of 
 the evidence now available to us, it is to be remarked 
 in the outset, that all of it is probable in its nature, so 
 that its principal force must arise from concurrence and 
 accumulation. Such arguments, therefore, as shall be 
 sufficient to establish the convictions which we seek 
 must of necessity be various, while a certain extent of 
 detail also cannot be avoided. 
 
 Some seem to have supposed it possible to arrive 
 directly at the desired results by a simple application 
 of the general principles of phonetics. 
 
 These have found it easy to construct a neatly arranged 
 dogmatic system /or Latin sounds ; but in offering us a 
 plan already suspicious from its very perfections, they 
 omit to demonstrate the actual connection between their 
 theory and the facts. The Latin pronunciation was cer- 
 
 * The present inquiiy will extend no further. 
 
8 IKTBODTJCTION. 
 
 tainly simpler and more regular than that of oiir ovn 
 language ; but a perfect simplicity and regularity will be 
 far from appearing to the candid student : nor will he 
 be able to content himself, fpr example, with the dogma, 
 that the " pronunciation of the diphthongs in' common 
 use is to be determined, at once and infallibly, by the 
 well established sounds of the component elements."* 
 Such consistency and symmetry is, no doubt, to be pre- 
 ferred to the hopeless variety and accident to which our 
 practice has abandoned the Latin which we read : but 
 reformers must be careful, or reaction may oyerreach 
 itself. 
 
 The only fair and reasonable way of obtaining a satis- 
 factory knowledge of the sounds which, in the classical 
 age, were actually in use among the Romans, is by com- 
 paring together such statements of the ancients, and such 
 other particular facts, bearing upon the matter, as are 
 within our reach. As has been already intimated, we 
 shall find that the best of this eyidence is only probable. 
 
 The ancients had but an imperfect acquaintance with the 
 physiology of the vocal organs and of the natural causes 
 which determine the quality of sound ; and hence it was 
 out of their power to describe the elements of their 
 speech in any more certainly unmistakeable way than by 
 appealing to the familiar sensation of the ear in hearing 
 it. Therefore in all that the old Romans have to say 
 concerning the sounds of their tongue, there is nothing 
 which furnishes in itself an adequate account for foreign 
 ears or remote posterity, f 
 
 * J. F. Ricliardson, Boman Ortho^y (N. T., 1859), p. 29. 
 
 f Cic. Orat., c. 49 "rerum, verborumque judicium, pruden- 
 
INTRODUCTION. 9 
 
 Cicero discourses profusely on the proprieties of enun- 
 ciation, and some of Ms remarks will furnish a useful 
 preliminary to the more particular inquiry which we have 
 to undertake. Especially in De Orat. Ill, cc. xi-xiii, and 
 Orat., cc. 45-48, as well as in many passages of " Brutus," 
 we have a great number of general precepts for the man- 
 agement of the tongue, the breath, and the tone. — The 
 voice must be sustained and uniform (permanens) ; the 
 tone must be natural and unaffected {recta, simplex); 
 syllables must be distinctly heard, but not inflated nor 
 expanded ; letters must not be slurred, nor offensively 
 thrust into prominence ; the drawl {latitudo verborum, 
 dilatare syllabas), and harsh and strident, coarse, thick, 
 and deep sounds are condemned as " rustic " and inelegant ; 
 while a full-rounded, sonorous tone, sweet without weak- 
 ness, gentle without effeminacy \vox {plena), canora j and 
 suavts, lenis opposed to mollis^, concise but clear, imparts 
 the indescribable but indispensable " ring " and " color " 
 of "urbanity." {Brut. 46, §§ 171, 172.) 
 
 These precepts already put some limit to an indefinite 
 range of possibility for the Latin sounds. But from 
 Cicero and other philosophical writers on the use of 
 language, we cannot obtain the special and minute in- 
 struction which we need. The nature and power of tto 
 several " elements " to which the single letters correspond 
 was not discussed by those writers, as being a subject 
 which, in a scientific aspect, was too particular and too 
 obscure {"subtilior cognitio ac ratio literarum"),* but in 
 
 tise est : wcum autem et numerorum, a/ures sunt jndices : .... in 
 Ulia ratio invenit, in his sensus artem." 
 * Cic. Compare also Quint. I, 4, § 6. 
 
IG INTEODUCTIOK. 
 
 its popular aspect, too familiar for the more eleyated and 
 more general aims of philosophical disquisition. (See 
 Cic. d. Or. Ill, c. 13.) 
 
 Yet there were some, and they too the most learned of 
 their age, who eveu as early as the time of Cicero, made 
 long and careful studies of the elements of pronunciation. 
 Marcus Ter. Varro (b.c. 64), " doctissimus Romanorum," 
 is, for us, the first, and Priscianus CEesariensis (instructor 
 at the court of Constantinople, a.d. 510), among the last 
 of a long series of grammarians, to whose instructions we 
 must now turn, in order to gather, by inference, and not 
 without great care and pains, the information which 
 might have been easily and more certainly had by spend- 
 ing an hour with the Roman boys in their elementary 
 school. It is impossible to undertake here an estimate 
 of the responsibility of the different Latin Grammarians 
 severally. An effort shall be made, as far as • may be,- to 
 keep in mind a reference to this in the use which is to 
 be made of their remarks in the following pages. Those 
 of a later time often copy their predecessors freely, while, 
 on the other hand, occasionally discrepancies occur. 
 
 But it must be remembered in the general, that even 
 the later writers of this class represent an uninterrupted 
 literary tradition, which remained to a great extent, if 
 not entirely, unimpaired among scholars long after the 
 popular speech of the Eomans had entered upon its rapid 
 course of corruption.* 
 
 The statements of grammarians, however, are not 
 
 * The ;^opular speech was full of corruptions at all times. Taci- 
 tus' allusion (Dial. d. Orat. 33) to the quotidiani scrmorm foeda ae 
 pudenda titia, was scarcely made irrespective of the sounds, while 
 applying also to the choice and other management of words. 
 
INTRODUCTION. 11 
 
 enough, to satisfy us : since they presuppose a practical 
 knowledge of the very sounds which they describe. 
 When, for example, Priscian says : " Vocales .... per se 
 prolatcB nomen suum ostenduni j" that is a valuable 
 statement, and carries us a great way towards a, perfect 
 ^knowledge of the sounds of the vowels: but while it 
 serves to remove many possible doubts, it nevertheless 
 stops short of the last point ; and we learn from it, after 
 all, only that the name of the vowel I (e. g.) contains the 
 sound of I. The accounts of grammarians, therefore, 
 must be brought into comparison with separate lines of 
 probable argument. Among these — 
 
 1. The tradition of scholars is entitled to a place ; and 
 the inherited, if even perverted, sounds of the modern 
 Eomance languages may be usefully compared, with a 
 cautious regard to their history. 
 
 2. The Greek rendering of Latin sounds is instructive 
 so soon as it is made to appear (as may be and has been 
 done), that the Greeks endeavored to indicate the Latin 
 sounds, which they heard in Eoman mouths, as faithfully 
 as their alphabet would permit.* 
 
 3. The third source of concomitant evidence is the 
 face of the language itself, as seen in its records which 
 have been preserved to us. Since the value of this evi- 
 dence depends upon the assumption that it was the 
 practice of Latin orthography to accommodate itself 
 faithfully to the sound, and to change with its changes, 
 it is of importance to show that such an assumption is 
 
 * Certain deviations from the uniformity of this practice have 
 been happily elucidated by StreMIce in Kuhn's Zeitseh. f. Vergl. 
 Sprachf., Vol. I, p. 211 and after. 
 
12 IFTEODUCTIOK. 
 
 well founded. Aside then from internal probability for 
 this, and besides the fact that the ancient grammarians 
 generally proceed upon that assumption, we are not at a 
 loss for passages in the Latin authors, which both directly 
 and implicitly declare, that correspondence of form with 
 sound was aimed at, and understood in practice also to 
 be (certainly in the main) the fact.* 
 
 Quintilian says directly (I, 7, 11) : " Verum orthograjMa 
 quoque consuetudini servit, ideoque scBpe mutata est." 
 Cicero (Orat., 48), says that the ancient hooks of Ennius 
 show that their author said Bums instead of Pyrrhus. 
 Again {lb., c. 46) : "Jam [ut censorice tabulce loquuntub) 
 'fdbruin' .... audeo dicere,non 'fabrorum.'" Gellius 
 (N. A., IX, 14, §§ 31-33), in order to prove that Lucilius, 
 among others, said {dixerunf) facie, etc., instead of faciei, 
 etc., quoted a passage from that author's writings, con- 
 taining the form in question. And so, eTcrywhere among 
 the grammarians. 
 
 But this very principle, being continued into later 
 times, has served to corrupt the orthography of the MSS. 
 which we possess ; few of which transcripts are of earlier 
 date than the ninth century, and none more ancient than 
 the fourth. For a period earlier than that, therefore, the 
 only unquestionable authority for Latin orthography is 
 to be found in the monumental records. The late valu- 
 able collections which have been made of those inscrip- 
 tions have added much stability to this department of 
 evidence, for the Latin sounds. To the study of them is 
 
 * So that the existence of two forms of the same word is evi- 
 dence also of two sounds; e. g., " Sibt, plenum est: sit, imminit- 
 tum : licet utroque uta/re." Cio. Orat., 47. 
 
INTEODUCI'IOK. 13 
 
 due the chief merit of Prof. "W. Oorssen's masterly work 
 on Latin Pronunciation. Perhaps we should not here 
 regret the want of access to original documents, since the 
 results of Oorssen's extended labors He before us ; and 
 these we shall not hesitate in using, so far as may suit 
 the phrposes of this more humble and restricted inquiry. 
 
 It may be allowed us to mention as other modern 
 works, from which we haye derived instruction land taken ' 
 facts : Lachmann's Corfiinentary on Lucretius j various 
 essays in Kuhn's Zeitschrift fiir Vergleichende Sprach- 
 forschung; and Seyffert's Latin Grammar (Branden- 
 burg, 1798), from which particularly have heen drawn 
 many extracts from the Latin Grammarians. Those who 
 are acquainted with the more or less well known names 
 of scholars who have written specially on the subject 
 assigned us, will miss the most of them from this short 
 list. Of the works of some, particularly of Schneider, we 
 have ourselves deeply felt the need 
 
 Having defined the purpose before us to be a limited 
 and practical one, we shall endeavor to direct our inves- 
 tigation solely towards results which are to be of ser- 
 vice in the school, and will inquire, therefore, only into 
 the sounds actually in use among the Romans in the 
 period commonly known as the classical age. No allusion 
 shall be made to the popular or provincial variations, 
 except when these can be made use of to illustrate the 
 main object of our search. For the same reason, and to 
 the same extent, it will be proper to exclude all parti- 
 cular account of the early and late history of the Latin 
 sounds. 
 
 And yet, with all restrictions, our undertaking remains 
 a considerable one. Without a particle of positive evi- 
 
14 INTEODUOTIOK, 
 
 dence, our only resource is in the multiplication of prob- 
 able proofs ; and so also we cannot escape the risk of 
 prolixity, since there is no choice between dogmatism 
 with brevity on the one hand, and a spreading out of 
 details on the other, if eyen a feeble hope be indulged of 
 producing, in the minds of any, a reasonable convifction. 
 This conviction we believe to be the prominent demand 
 •of the present time with respect to the important ques- 
 tion in our hands : for while few are found disposed to 
 undertake a hopeless defence of that utter abandonment 
 into which the pronunciation of the Latin has fallen 
 among us, yet before adopting another plan all think 
 (and justly think) themselves entitled to exact a show of 
 its authority. To give up even a corrupt and barbarous 
 manner of pronunciation, requires a sacrifice of habit, 
 which will not be generally conceded without a sufficient 
 reason. 
 
 Firmly believing that the true Latin sounds may in 
 the main be known, and may be sufficiently defended, 
 but not without a due sense of the difficulty of our 
 performing the task in a satisfactory and suitable man- 
 ner, we proceed to the details of our inquiry. 
 
PRONUNCIATION OF LATIN. 
 
 The fundamental units of speech (elementa) were 
 represented in the classical period of the Latin language 
 by the following signs {litterce) : 
 
 A, B, C, D, E, F, G, H, I, (K), L, M, ^, 0, P, 
 Q, E, S, T, V, X, [Y, Z]. 
 
 The whole group is called by Tacitus Uteratura, but 
 by later authors often compendium, and commonly by the 
 Greek name Alphdbetum. Severally their names were 
 (Vid. Prise. P. p. 540) : 
 
 A, Be, Ce, De, E, eF, Ge, Ha, I, Ka, eL, eM, 
 elSr, 0, Pe, Qu, eK, eS, Te, V (and Vau), iX, 
 (Tpsilon, Zeta). 
 
 It will be convenient, in a particular examination, to 
 consider the vowels first. 
 
 A stream of air vibrating from the vocal chords, and 
 forcing its own way through the buccal tube, which is 
 left unmodified by any activity of the organs situated 
 along its course, will produce inarticulate sound only. 
 
16 PEONUNCIATIOlSr OF LATIN. 
 
 But the same stream, accompanied by the eksiest, and 
 (if we begin with the root of the tongue) the first modi- 
 fication of the size and shape of the buccal tube, gives 
 rise to a simple vocal sound, which for many reasons, 
 historical as well as natural, deserves to hald the place 
 always assigned it at the head of the Latin alphabet. 
 This is the vowel A, m&de by a conformation of the 
 buccal tube through the action of the root of the tongue 
 and the soft palate.* The whole extent of the tongue, 
 in advance of the root, remains in repose {suspensa lin- 
 gua), as do all the muscles of the jaws and lips, which 
 are separated by allowing the lower jaw to fall to the 
 extent to which its weight will carry it {ridu patulo). 
 
 The resulting sound is that of a in English bar, or 
 French l)arbare.\ That this was the sound, and the only 
 sound of the Latin A, is to be argued from the combined 
 force of a variety of considerations, as : 
 
 1. Every form of tradition maintains this sound, and 
 no authority supports another. 
 
 2. The history of internal changes of Latin words, 
 the modifications and combinations to which this vowel 
 has been subject, furnish evidence in the same sense. 
 
 3. A comparison with other languages — whether •his- 
 torically or actually instituted — favors the same. (Com- 
 pare, for example, Latin a with Greek a in equivalent 
 forms.) 
 
 * For an exact account of the physiology of the vocal organs, see 
 Max Miiller's Lectures on Science of Language, Second Series. 
 
 \ It is known that the English and French a have a difference of 
 tint. We believe that the Latin was more strictly the same with 
 the French a : but it may be for practical purposes also sufficiently 
 —"11 represented by a in English " bar." 
 
PBOWUSrCIATIOK OF LATIN. 17 
 
 4. The testimony of Latin writers, before their spoken 
 language had lost its authority, points in the same 
 direction. 
 
 But it will not be necessary to deTclop all this eyi- 
 dence. The sound of Latin A is so little disputed, that 
 we may safely assume it and pass on at once to a par- 
 ticular remark — the only one necessary to be added. 
 This is, that the short sound of A differs not at all in 
 kind, but only in quantity, from the long sound. "While, 
 therefore, long A has the sound of a in English tarter, 
 the short sound of Latin A is not heard in English hatter. 
 The English rarely uses the short Latin a in its purity. 
 It is sometimes accidentally given in unaccentuated places 
 in English words, in which it is common, howeyer, to 
 hear either an uncertain sound or another sound of Eng- 
 lish a, as in the first syllables of papd,, ahdJi, marauder. 
 In German, French, and Italian the short sound of Latin 
 a is regularly heard. We have it in French battre. 
 
 It is obvious that the English short a of hatter differs 
 not only in quantity but also in quality from the long a of 
 barter. But of such a difference in quality between the 
 long and the short A of the Latin, we can infer nothing 
 indirectly, and we hear nothing directly. All accounts 
 speak only of one sound.* 
 
 * For the ten 'oaHeties, spoken of by Priscian (ap. P. p. 539), do 
 not belong to the nature of the sound, but are due to the various 
 combinations of accent and quantity, and even a,spwation, since he 
 enumerates ha among them. Upon consulting the passage, rather 
 long for insertion here, it will be plain, that if we remove by 
 elimination all the elements of the author's mischievous confusion, 
 we shaU have left only the one sound of A. Much of the obscurity 
 and apparent contradiction found in the ancient Grammarians (and 
 there is no little of it), is simply due to a want of accurate and close 
 
18 PKONUKCIATIOK 03? LATIiT. 
 
 Marius Viciorinus (P. p. 3453) : " A litera, rictu patulo, 
 suspensa neque impressa dentibus lingua, enuntiatur." 
 
 Priscian (P. p. 540) : " Vocales .... per se prolatse 
 nomen suura ostendunt." 
 
 Nay, the identity of sound is certainly contemplated by 
 Lucilius as quoted by Ter. Scawrus (P. p. 3355, ap. Cors- 
 sen I, p. 140) : 
 
 "'A' primum longa, breyis syllaba, nos tamen unum 
 Hoc faciemus, et uno eodemque ut dicimus pacto, Scri- 
 bemus : ' pAcem, -plAcide, jAnum, Aridum, Acetum,' 
 ' "Ap£f, "Apef ' Grseci ut faciunt." 
 
 Hence, 
 
 long A is equal to a in Eng. barter, 
 short A « « " a " Fr. lattre. 
 
 3E. 
 
 Latin E is in form the G-reek epsilon. A correspond- 
 ence iu sound, also, to the sound of that letter in the 
 Greek, is to be inferred from 
 
 Latin, Secundus, Greek, "Se/covvdof, 
 " Geta, " Tt-Taf, 
 
 Tubero, " Tovfiipuv, 
 
 " Servilius, " lepovtkiog, 
 
 " Porsenna (and Por- " Uopaewag (and IIop- 
 sena), O'qvag), 
 
 discrimination between permanent qualitiea and accidental condi- 
 tions of sound — distinctions which have become familiar and neces- 
 sary to the more rigorous method of modern times. 
 
PEONUNCIATIOK OP LATIK. 19 
 
 and, assuming rj^e, 
 
 Latin, Suetonius, Greek, SevrjToviog, , 
 " Aurunculeius, " AvpowKovk-qlog, 
 
 " Zeno, " Z^vuv, 
 
 " zelotypia (Pliny), " ^rjXoTvnia, etc., etc. 
 
 The practice of scholars generally makes a true dif- 
 ference in the timbre * or essential tone of e and g ; so 
 that the e of egi is as clearly separated from the e of ego 
 as are the vowel sounds heard in English bake and beck — 
 and that quite irrespective of the quantity. Such a 
 practice must be regarded as inconsistent with the words 
 of Priscian quoted above (p. 18), the import of which is, 
 that each vowel had but one proper sound. The dif- 
 ference, however, between a short rendering of a in bake 
 and & in beck, although real, is small, and to insist upon 
 a difficult reform in so minute a particular may well be 
 thought by some an instance of misdirected diligence 
 (molestissima diligentiae perviersitas. Quint. I, 6, 17). 
 For us who have no longer an appeal to the criterion of 
 the ear, it is impossible to decide positively for the 
 authority of either of these two sounds as compared with 
 the other, yet probability seems to favor the belief that 
 it was the sound now heard in egi, which contained the 
 true stamp of the vowel e. 
 
 This sound seems best adapted to the deviations in the 
 rendering of E among the Komans, which must now be 
 considered. The first of these is indicated by the inter- 
 
 * WHiat by the French is called the timbre of a vocal tone, and 
 by Prof. Tyndal (in imitation of the German) the dang-Hnt, is vari- 
 able only with the shape of the vocal tube. A change of timbre, 
 therefore, is both physiologically and actually to the ear a change 
 ofwwd. See Tyndal on Sound : Lecture V, 
 
30 
 
 PEONUNOIATIOK OF LATIN. 
 
 change of e and m, well attested in yery niimerona 
 instances from early times on. The following all belong 
 to the classical period, or were older : 
 
 Hedus, 
 
 cena, 
 
 merere, 
 
 cespes, 
 
 penuria, 
 
 Murena, 
 
 levis, 
 
 ve! 
 
 ne ! 
 
 Hsedus, (Varro, ap. Lach. Luc. p. 35). 
 caena. 
 
 mserere, 
 
 csespes, 
 
 psenuria, 
 
 Miirsena, 
 
 laevis, 
 
 vsb! 
 
 On inscriptions, in- 
 ) stanced by grammarians, 
 and in MSS. 
 
 nse 
 
 Varro, L. L., VII, 96. 
 
 ponunt, alii non . 
 no7i Mmsium. 
 
 In pluribus verbis A ante E alii 
 . , ac rustici Pappuni Mesium, 
 
 Now assuming the proper sound of e to be that of a 
 in English lake, experiment shows it to be diflBcult to 
 dwell upon this sound in certain situations (namely, in 
 an open final syllable or before a dental or labial letter), 
 without mixing with it an after ring of i. This done, 
 the result will nearly coincide with what will hereafter 
 appear to have been the sound of ce in Latin : so that the 
 6 which was often confounded with a must have sounded 
 like ei and ai in the English rein, rain ; veil, vail, etc. 
 [And such was, perhaps, the " E plenissimum" of Cicero, 
 d. Orat. Ill, 13, 46.] 
 
 Another iinpurity of e is mentioned by Quintilian, I, 4, 
 8, and 5, 33 (and Gell. X, 34), in the case of the word here, 
 which he says was written heri both by the ancients and by 
 the Emperor Augustus, while in the pronunciation of his 
 
PKONUKCIATION OF LATIiT. )81 
 
 own time it was hard to decide for the e or the i. What 
 was the exact nature of this obscure sound must now be 
 left to conjecture. It seems to be indicated by Velius 
 Longus (p. 2235, P.) as that of the second syllable of tibei 
 for tibi, as spoken by some whom he blames ; and must 
 have been heard in those places where the orthography 
 fluctuated between e and i, as in caeretes, casrites, etc. 
 (See Liv. XXII, 1, and Drachenborck : " Utrumque Cceri- 
 tes et Omretes apud ipsum Livium invenitur") 
 
 Quint. Inst. Orat. I, 7, 24 : " Sibe et quash scriptum in 
 rnuUorum libris est, sed an hoc voluerint auctores, 
 nescio : T. Livium ita Ms usum ex Pediano comperi, 
 qui et ipse eum sequdbatur ; hcec nos I littera fini- 
 mus." 
 
 Further, the short e must have had an uncertain, indif- 
 ferent, or neutral sound, not only in final er as in acer,* 
 linter, pater, inter, etc. (Oorssen), but also in iniellego, 
 protenus, donee, saltern, and many others, f 
 
 The normal and approved sound of B, therefore, was, 
 e. g., in ^ 
 
 e of sgi, = a in English bake, 
 while e in ego = the same sound shortened. 
 
 Deviations in practice were, on the one hand, to a slight 
 after augment of i, in the case of long e, making eplenis- 
 simumj thus e in 
 
 * Cf. acris, patris, etc. 
 
 f We shall see, later, that the neutral sound of e was often inter- 
 changed with that of i, in early Latin, aa in timedus, naxeboa (Eib- 
 beck, Com. Ant., p. 10). 
 
33 PEONUNCIATION OF LATIK. 
 
 frenum \ 
 
 1. heres r = ai in English refrain, etc. 
 levis * 
 
 2. e, in last syllable of here, and others, had an uncertain 
 sound between e and i. 
 
 3. e, as in ?(for, etc., was neutral, as in English other. 
 
 I. 
 
 The terms exilis and angusta, applied by the ancients 
 to the vowel I, are well suited to the sound which tradi- 
 tion assigns to that vowel. This is the sound of ee in 
 English reed, which is made with a greater contraction of 
 the resounding space of the vocal tube, than is the case 
 in sounding any other vowel.* [I spiritus, prope dentibus 
 pressis." Martin. Capella.] 
 
 Latin, Ariminum ; Greek, 'kpiiiivov ; Italian, Eimini. 
 
 That the Latin short * generally partook of that pecu- 
 liar dullness which belongs to the English i in rid, cannot 
 be shown. 
 
 Corruptions of the I sound were at all times very com- 
 mon. Prom the nature, the direction, and the history of 
 these corruptions, as well as from the remarks made upon 
 them by the ancients, much probiable evidence may be 
 elicited, pointing to the above-named sound as the proper 
 one of I. 
 
 * For particulars consult Tyndal's Lectttres on 8ownd, p. 200, and 
 Max Miiller's Lee. on Science of Lang., 3d Ser., p. 
 
PRONUNCIATION OF LATIN. 23 
 
 Sucli inferences must be left to suggest themselTCS. 
 The facts which appear are hriefly these : 
 
 In earlier times, and in the sermo rusticus of Cicero's 
 time, I in some cases was pronounced with a certain ful- 
 ness which approximated to the sound of e. Men spoke 
 diequinti and diequinte, pristini and pristine, proclivi and 
 proclive (A. Gell. X, 24), heri and here (Quint. I, 6), tibi 
 and tibei (Velius Longus, P. 2235). For this fluctuation 
 ill orthography there was a middle sound neither quite i 
 nor quite e. [Quint, neque e plane auditur neque i.] 
 Lucilius proposed to regulate the practice of his time hy 
 indicating this obscure sound by EI, and confining its 
 use to the plural number in the case of vowel stems,* 
 thus genitive singular pueri, nominative plural puerei. 
 We will follow this no farther as a point of grammar, 
 and introduce it merely in evidence of the existence of 
 the sound in question. That it is evidence of such impu- 
 rity of sound in the vowel i in LuciUus' time, and not a 
 mere form of grammatical distinction addressed to the. 
 eye, may easily be shown jf but for brevity will be taken 
 as true on the opinion of Quintilian (I, 7, 13, 15-18). 
 The common people of Cicero's time went further and 
 pronounced full broad e instead of i. For this is the 
 most likely interpretation of De Orat. Ill, 12, 46. Quare 
 Gotta noster, cuius tu ilia lata, Sulpici, nonnumquam imi- 
 taris, ut Iota literam tollas, et E plenissimam dicas, non 
 mihi oratores antiquos, sed messores videtur imitari. But 
 neither of these sounds received in the classical times the 
 sanction of scholars, and down to a late period to which 
 
 * See Scaurus, p. 2255 ap. Seyffert, p. 142. 
 t Tet see Aul. GeU. XIX, 14, in fin. 
 
84 PfiOHUNCIATION or LATIN. 
 
 they continued to be more or less heard they incurred 
 the disapproyal of grammarians, who declared for the 
 thin, fine, proper sound of i {tenuitatem i litercB. See 
 Vehus Longus, P. p. 2335). 
 
 There was another impurity of the i sound heard in 
 the approximation and confusion of short i and short u, 
 as in simus* for sumus. 
 
 The words 
 
 maxumu$ 
 
 contumaz 
 
 lacrumcB 
 
 existumat 
 
 optumus 
 
 monumentum 
 
 alumenta 
 
 nolumus 
 
 lubido 
 
 aurufex 
 
 artubus 
 
 AVI yvyvi j%i\ f /v\n n i An/t 
 
 
 TflCCTlCUpiUTn 
 
 aucupare 
 
 and others \ 
 
 are mentioned by grammarians, who say of them that 
 they contained the uncertain sound, which they describe 
 as " something half way between i and u," " thicker than 
 i, thinner than u," — a sound, says one (Priscian, Putsch, 
 p. 539), which " seems to be that of the Greek y." Cors- 
 sen observes upon these forms, as well as upon a large 
 number of others, which he has drawn from inscriptions 
 of all periods, that the uncertain sound of i or u appears 
 to have been used almost exclusively before labials m, b, 
 p, f. In this observation he is somewhat more discerning 
 than Priscian, who, however, adds to his list of words in 
 
 * So spoke the Emperor Augustus, accordmg to Suet. Vit. Octav. 
 87. 
 
 \ Tibi is for tvhi (Sansc. tubyam). 
 
PROlf UNCIATIOK OF LATIN. 25 
 
 whicli this sound was heard (on the testimony of Donatus) 
 the following : video, vim, virtus, vitium, vix.* 
 
 Finally, for our guidance in practice, we have to 
 observe : 
 
 1. That the u sound in such words was the ancient 
 one, which may be gathered from the fragments of the 
 ancient comedians and tragedians (Eibbeck), froin the 
 MSS. of Plautus and Terence, as well as from the state- 
 ments of grammarians. 
 
 2. That the decision of Cicero and jOsesar was given in 
 favor of i; and 
 
 3. The sanction of grammarians to the i sound. 
 Agnaeus Cornutus (Putsch, p. 3384). Lacrumm an 
 
 lacrimcB ; maxumus an maximus, et si quae similia sunt 
 scribi debeant, quBBsitum est. Ter. Varro tradidit, Caesa- 
 rem per i eiusmodi verba solitum esse enuntiare et scri- 
 bere ; inde propter auctoritatem tanti viri consuetudinem 
 fadam. Sed ego in antiquiorum multo libris, quam 
 Caius Caesar est, per u pleraque scripta invenio : optumus, 
 intumus, pulcherrumus, dicundum . . . .j melius tamen 
 et ad enuntiandum et ad scribendum, i literam pro u 
 ponere, in quod iam consuetude inclinavit. 
 
 The normal sound of I, therefore, when long, was as e 
 in English re' tail; when short, as e in retail. 
 
 THE CONSONANT I (SEMI-VOWEL). 
 
 Besides the sound which we have seen represented 
 by I as a vowel, this letter played also another part — 
 
 * And for some other rather surprismg instances of this sound, 
 see Vel. Longus, P. p. 3316 (in prodire, etc.). 
 2 
 
36 PEONUNCIATION OF LATIN. 
 
 that, namely, of a consonant, or (as, in Tiew of all the 
 facts, it is better called), a semi-vowel 
 
 The organs being all set for the pronunciation of the 
 vowel I, the intonation was more or less completely with- 
 held ; thereby reducing the function performed more or 
 less completely to that of a consonantal check. This 
 happened only when I occurred immediately before a 
 vowel sound in the same syllable.* There are two cases. 
 
 1. In the beginning of a word, as in Janus, Juppiter, 
 iungo, and in the beginning of the second member of a 
 compound word, as in abiectus, iniuria, eieetus, diiudico. 
 In such a situation, the consonant I was sounded like the 
 English y consonant: thus, Yanus, Yuppiter, abyedus, 
 diyudico. 
 
 - In defence of this sound, which is that of tradition, it 
 need only be said, that there is no allegation in antiquity, 
 nor any evidence of another sound — of such, for instance, 
 as that of / in French or of g before i in Italian and 
 English, f Further, any such distinct consonantal power 
 could not have belonged to a letter which was continually 
 allowed to relapse into the original vowel sound of i, 
 and was always spoken of as essentially the same letter. 
 
 * If I was ever a consonant in the end of a word, as ai (aj), hei 
 (TuS), is not certain. 
 
 f Evidence of this must not be supposed to exist in the fact, that 
 Jams, dimis, dies, etc., are made from the same root (Sansk. die.). 
 Jovis (Diovis) dropped its d not because the I was relied upon to 
 represent alone the dj (dg) sound, but because the i, becoming hard- 
 ened into the consonant sound, rejected the preceding d as incom- 
 patible with its pronunciation. A parallel appearance is that of 
 VeUum from duellum. So that Italian giorno from Latin diurnua is 
 not in any way to be compared here. 
 
PEONUNCIATIOK OF LATIN. 37 
 
 called by tlio same name,* and marked by the same 
 sign. 
 
 The resolution of the consonant into the vowel sound 
 may be abundantly shown, and will appear in the next case. 
 But no better proof can be wanted than this : that ancient 
 grammarians thought it worth- while to point out par- 
 ticularly ' ttie cases in which I was not a vowel but a 
 consonant. Thus P. Mgidius (ap. Aul. Gell. N. A., XIX, 
 14, 6), the learned contemporary of Cicero, says, " i et u 
 vocales semper subditse .... Si quis putat praeire u 
 in his: 'Valerius,' 'Vennonius,' ' Volusius,' aut t in his : 
 ' iampridem,' ' iecur,' ' iocurri,' ' incunduni,' errabit, quod 
 has literse, cum prseeunt, ne vocales quidem sunt." And 
 it is manifest that no remark could be more utterly use- 
 less and uncalled for to those who were sensible of any 
 such difference between the vowel i and the consonant i, 
 as is imphed in sounding the latter like several modern 
 languages sound their /.f 
 
 The fact, that some proposed to derive the name 
 Janus from eo, ire, and to write it Eanus, is very signi- 
 ficant in this connection. (See Forcellini sub. v.) 
 
 2. The second case is the occurrence of i after a 
 vowel in the preceding syllable and before a vowel 
 in its own syllable, within a simple word: as in aio, 
 maior, maiestas, eius, Jiuius, Veii, Pompeii. In this situ- 
 ation i discharged the function of a double consonant, 
 according to the very explicit testimony of the ancients 
 {"pro duplici accipitur consonante"). Now let any one 
 
 * As Quint. I, 4, 11. Atqui littera IbM insidit, coniicit enim est 
 ab iUo iacet. 
 
 f The passage alluded to will be better appreciated if consulted 
 more at length. 
 
28 PRONUNCIATION OF LATIN. 
 
 consider the statement of Priscian, that the i in TrXiia 
 must be treated as a double consonant in this way (like 
 English y in destroyer) ! Can it possibly be supposed 
 that a double consonant, like, say, English gg, could then 
 be contemplated (Trogga or Trojja) ? If this were the 
 case, nothing could be accomplished in the sound by 
 doubling; for Trojja could not be distinguished from 
 Troja, more than caussa from causa (with s sharp). But 
 if this were not the case, and the consonant i was nothing 
 but the vowel i rendered without tone {i. e. English y, 
 consonant), then it is evident not only that the effect of 
 it might be doubled in the supposed situation, but that 
 it would necessarily be doubled. And Cicero's manner 
 of writing atio, Maiia, and the manner of others in writ- 
 ing Pompeiii, was only a faithful rendering of the sound 
 inevitably heard, provided ai, etc., be not taken as a 
 diphthong (against which Priscian warns), and provided 
 the consonant i was sounded like English y. [Compare 
 French ennui with ennuyer ; envoi with envoyer, etc.] 
 
 Quint. I, 4, § 11 : " Sciat etiam Ciceroni placuisse aiio, 
 Maiiamque geminata i scribere, quod si. est, etiam 
 iungetur ut consonans. [Thus, ay-yo, or in fact also 
 a'i-yo.] For another use of II, see Quint. I, 7, § 14 
 
 Priscian, I, 4, 18;* ap. P. p. 545: Et i quidem mode 
 pro simplici, modo duplici aecipitur consonante : pro 
 simplici, quando ab ea incipit syllaba in principio 
 dictionis posita subsequente vocali in eadem syllaba, 
 ut Juno, Jupiter, pro duplici autem, quando in medio 
 dictionis ab ea incipit syllaba post vocalem ante se 
 
 * Ed. Krehl, as quoted by Auf recht in ZeiUch. f. Vergl. Sprachf., 
 1, p. 225. 
 
PEOSriTNOIATION OF L^TIN. 29 
 
 positam subsequente jjuoque Toeali Id, eadem syllaba, 
 lit maius, 2}eius, eius, in quo loco antiqui solebant 
 geminare eandem i literam et maiius, peiius, eikis, 
 K.B.! scribere, quod non aliter pronunciari posset, quam si 
 cum superiore syllaba prior i, cum sequente altera 
 proferretur, ut pei-us, ci-ius, mai-ius. Nam quam- 
 vis sit consonans, in eadem syUaba geminata jungi 
 non potest. Ergo, non aliter, quam tellus, mannus 
 proferri debuit. Unde Pompeiii quoque, genitivum, 
 per tria in scribebant, quorum duo superiora loco 
 consonantium accipiebant, ut si dicas Pompeiii \i. e. 
 Pompeyyi^. Nam tribus Hi junctis. qualis possit 
 syllaba pronunciari ? Nam postremum. i pro vocali 
 est accipiendum, quod Csesari doctissimo artis gram- 
 maticEe placitum a Victore quoque in arte gramma- 
 tica de syllabis comprobatur. Pro simpliei quoque 
 in media dictione invenitur, sed in compositis, ut 
 iniuria, adiungo, eiecius, reiice. Virgilius in Bucq- 
 lico proceleusmaticum posuit pro dactylo : 
 
 Tityre pascentes aflumine reiice capellas. See also 
 the same, I, 9, 50. 
 
 I consonant is rendered in Greek by iota, as in : 
 nojuTrjytof, Brjioi (Veii), rd'ipg. 
 
 Finally, then, Latin I consonant = English y in yam, 
 young,_ etc. 
 
 o. 
 
 The Towel is to be identified with the Greek omi- 
 cron, as well in its nature as in the shape of its written 
 sign. The sound approved as its proper one in the 
 
30 • PEONUKOIATIOK OF LATIN. 
 
 classical period seems to have been biit^ one and simple ; 
 and there is nothing to disturb the accepted opinion, that 
 this sound was the same which is heard, first short and 
 then long, in the English word postpone. 
 
 In the practical rendering of Latin short o, those who 
 speak the English language need to be guarded against 
 using the short o of English not, etc., between which and 
 the simple shortened sound of o (as in English post) 
 there is a qualitative difference — a difference due, as 
 .every qualitative difference must be, to a different atti- 
 tude of the parts engaged in modifying the extent and 
 shape of the room in which the voice resounds. The 
 English is in the habit of deadening the sound of all its 
 short vowels. This is done with the o of not, and a pro- 
 longation of that sound will not give the o of post, but 
 rather that of frost. Of such deviation in the simple o 
 sound of the Latin we have no proof. Probably no one will 
 question that the o of sermo was sounded like the same 
 vowel in sermonis j and yet there is no reason for supposing 
 the character of the vowel sound to have been otherwise 
 heard in arbor or arioris, excejpt a supposed necessity for 
 such a change, arising from the closure of the syllable. 
 Now this supposition is natural, to be sure, to the speaker 
 of the English language, but it cannot on that account 
 only be applied to the Latin. And if any will -maintain a 
 second sound for o, they must sustain their opinion by 
 some sort of evidence drawn from the Latin itself, or from 
 some natural law of speech of universal application.* 
 
 * The in unaccented, so to speak, neglected places, as in the 
 ancient nominatives ./ifes, servos, and perhaps in many other forms, 
 was no doubt often blurred ; but in thus losing its own character- 
 istic tone it merely became neutral at most. This neutrality befel 
 
PRONUNCIATION OF LATIN. 31 
 
 How the Greeks heard. Latin o may be inferred from 
 
 'Podavog, for Ehodanus (Polyb.) 
 
 Kevoiidvoi, " Cenomannij " 
 
 UoaTov/iiog, " Postnmius, 
 
 VbinT], " Eoma, 
 
 Kamruktov, " Capitolium, etc.* 
 
 In the forms Plato, Zeno, etc., it is reasonable to sup- 
 pose that the o had the same sound as the w in TlXdruv, 
 etc., and yet it is not likely that the final o of Latin 
 names was commonly sounded long. 
 
 So much for the recognized regular sound of o. There 
 were irregularities in practice. 
 
 1. Grammarians state, and MSS. show (as the inscrip- 
 tions are said to do also), that in the ancient Latin, o 
 appeared written in many places where u was afterwards 
 the accepted orthography of the classical age. 
 
 Priscian (P. p. 554) says : Multa praeterea yetustissimi 
 etiam in principalibus mutabant syllabis, ut cungrum pro 
 congrum, cunchin pro concMn, huminem pro hominem 
 proferentes, funtes pro fontes, frundes pro frondes. 
 
 most of the Latin vowels when situated in unimportant places, so 
 that frequently they went practically for nothing : witness vinclum, 
 soedum, cuxipter (for accipiter) ; spiclum (for ^nculum) ; the Greek 
 renderings AevrAof, Ka-Aof, for Lentulus, Catulus ; supra (for 
 supera) ; and when defatigatus was written defetlgatus (Ter. And. 
 667^ that e should probably not be taken with the true sound of e, 
 and representing thus a new sound for a, but only as an accidental 
 rendering of a neutral a. 
 
 * This comparison will have to go for what it is worth. Some, the 
 most perhaps, among us suppose the Greek o = o in English riot. 
 This is notour opinion. The modem Greeks know no distinction 
 of sound between u and o. 
 
33 PRONUNCIATIOIJf OF LATIN. 
 
 TJnde Lucretius .... in Lib. Ill ; Atqui animarum etiam, 
 quwcunque Aclierunte profunda : pro Acheronte .... 
 Quse tamen a junioribus repudiata sunt, quasi rustico 
 more dicta. TJ quoque multis Italise populis in usu non 
 erat, sed e contrario utebantur o: unde Romanorum 
 quoque vetustissimi in multis dictionibus loco ejus o 
 posuisse inveniuntur, poplicum pro publicum, quod testa- 
 tur Papyrianus de Orthog., polchrum pro pulchrum, col- 
 pam pro culpam dicentes, et Hercolem pro Serculem, et - 
 maxima Digamma [v] antecedente, hoc faciebant, ut : 
 servos pro servus, volgos pro vulgus, Davos pro Davus. 
 
 There must have been a transition period when o thus 
 situated (chiefly after v, or before s or m, in final sylla- 
 bles) had a sound approximating u, and that sound must 
 have been heard even in the time of Cicero, in such words 
 as cervos, equoni, servom, cBvom, rivom, nativom, etc. This 
 was retained so late for the sake of a grammatical the- 
 ory, and was subjected thus for a time, till it was relieved 
 of this duty, to the anomalous representation of the u 
 sound, which had become the prevalent one for the situa- 
 tion above indicated. 
 
 Quint. I, 7, 26. Nostri preeceptores servum cervumque 
 U et litteris scripserunt, quia subjecta sibi vocalis 
 in unum sonum coalescere et confundi nequire't : nunc 
 TJ gemina scribuntur ea ratione, quam reddidi. But the 
 uncertainty here indicated was early removed, and the o 
 and the u clearly separated. 
 
 There was another fluctuation of orthography in which 
 the was concerned, namely, the putting a for au. This 
 occurred in contemporary forms of the same stem, and 
 to a limited extent in all periods of the language. 
 
PRONUNCIATION OF LATIN. 33 
 
 Cato wrote, dehorito, for dehaurito j 
 
 The ancient comedians, lotus, " lautus ; 
 
 (SeeEibbeck&Priscian.) phstrum, " plaustrtm J 
 Cicero, plodo, " plaudo ; 
 
 exploditur, " ezplauditur; 
 5 clodicat and claudicafj 
 
 while in the time of the empire also we find : 
 
 (Suet. Vit. Tib. Cses. II) Clodius and Claudius; 
 ( " Vesp. 22) plostra " plaustra ; 
 
 Polla " Paula, and others. 
 
 It may he supposed with Corssen (I, p. 148), that this o 
 had. the fullest possible sound. 
 
 To repeat, then ; the stamp of the Latin o sound was 
 that of in English note. Hence 
 
 sounded like o in English expose; 
 " " " exposition, 
 
 (a;nd not like o in English expdsitor). 
 
 Deviations. — 1. A confusion of o and u, which once 
 obtained, was disallowed by the classic Latin. 
 
 2. Already in the time of Cicero (and by him), o was 
 sometimes written for au — an inaccuracy which had its 
 seat probably in the popular speech, and which showed 
 itself in literary forms to a small extent in all times. It 
 is safe to presume that this o contained only the radical 
 sound above given, but ^as given as full as might be. 
 (Compare English clothes, earlier cloathes.) 
 
34 PKONUKCIATION OF LATIN. 
 
 Marius Victorinus (P. p. 3454) says: "V literam, quo- 
 tiens enuHtiamus, productis et ooeuntibus labris effere- 
 mus.'' 
 
 Professor Tyndal {on Sound, p. 200) : " For the pro- 
 duction of the sound u (English oo in Jioop), I must push 
 my lips forward, so as to make the cavity of the mouth as 
 deep as possible, at the same time making the orifice of 
 the mouth small."* These two statements, so nearly 
 coincident, point to an identity of sound in Latin «< and 
 English 00 in pool, hoop. This is the traditional sound of 
 the Latin Towel, and the probable evidence within our 
 reach contributes to support the same. 
 
 The Greeks rendered the Latin m by their ov, thus: 
 
 Sallustius, SaAovariog, 
 
 Tertulliahus, TeprovXiavog, 
 
 Lugduntim, AovySovvov^ 
 
 Superbus, lovTrspPog, 
 
 Brutus, Bpovrog. 
 
 Also, 
 
 Ifumitof, "Nov/iriTbip, 
 
 Oorbulo, KopPovXuv, 
 
 popiilus, TTWiTovAovf . 
 
 These forms are from MSS. (of Polybius, Plutarch, and 
 Dion Cassius). It is said that inscriptions show : 
 
 * Compare Max MuUer'a Science of Language, 2d Series, p. 130. 
 
PBONUNCIATIOir OF LATIN. 35 
 
 for tuum, 
 
 rovofi, 
 
 " laborum, 
 
 Xapopovjj; 
 
 " suo, 
 
 aovu, 
 
 " singularii, 
 
 aiyyovXdpioi, 
 
 and many such ; but we can dispense with further exam- 
 ples when we have the statement of Marius Victorinus 
 (p. 2454) : " u litteram, quam nisi per ov conjunctam, 
 Grseci scribere ac pronuntiare non possunt." It is veiy 
 clear, therefore, that such a pronunciation as corbewlo for 
 coriulo is quite out of the question. It was entirely in 
 the power of the Greek to express this sound, but it can 
 never be supposed to have been heard in KoppovXwv. 
 
 But while the statement of Victorinus, amply supported 
 by examples, is sufficient to remove all doubt that the 
 proper sound of u was that of Greek ov ; the fact is that 
 there were various other ways of writing in Greek the 
 Latin syllables containing this vowel. Some of these 
 must be taken to. indicate an entire surrender of the u 
 sound in obedience to certain influences operative in the 
 Greek language. Thus Strehlke..(ZetfccA/.- /. Vergl. 
 Sprachf., I, p. 333) shows admirably how Lucius, Lucul- 
 lus, etc., were rendered,* Aeviaog, AevicovXXog, etc., in 
 obedience to a principle of popular etymology, pointing 
 to kevKog ; just as AoXofisXXag (besides AoAa/Je'AAaf) for 
 the sake of the supposed element doXog. 
 
 But other varieties of Greek rendering cannot be 
 accounted for in this way, and are found, moreover, to 
 correspond with analogous fluctuations in the Latin 
 orthography itself, both together pointing to certain 
 varieties of sound in the Latin u. 
 
 * Besides more rarely the ordinary Aoi/ctof, etc. 
 
36 
 
 PKONTJNCIATIOK OP LATIN. 
 
 1. These are : o for u, as seen in 
 
 Mofifiiog for Mummius, 
 
 DfflTopVtvof " Saturninus, 
 
 KaXiyoXag " Caligula, 
 
 KopoyKaviog " Coruncamtis, 
 
 'I6(iag " Juba, 
 
 Tlonkiog (also Uovnhog) " Publius, etc. 
 
 Besides tlie widespread ancient use of o for later u in 
 Latin words, as in 
 
 polsi 
 
 for 
 
 pulsi, (Rib. Com., p. 136.) 
 
 exfociont 
 
 tt 
 
 exfociunt. 
 
 vivom 
 
 i{ 
 
 TiTum, (Ter.) 
 
 Polvius 
 
 u 
 
 Fulvius, 
 
 Hecoba 
 
 £i 
 
 Hecuba, ■\ 
 
 probaveront 
 
 a 
 
 probayerunt, S- (Quint. I, 4, 16.) 
 
 dederont 
 
 a 
 
 dederunt, ) 
 
 ssquom 
 
 a 
 
 sequum, \ 
 
 quom 
 
 a 
 
 quum, ■ WTer.) 
 
 quoi 
 
 ei 
 
 cui. ) 
 
 A similar orthography was practiced to a limited 
 extent also in classical times.' It is said * that the best 
 MSS. not only of Plaut. and' Lucrei, but also of. Virgil 
 and Cicero hare such forms as : 
 
 volt. 
 
 Tolnus, 
 
 Toltis, 
 
 avolsa, 
 
 TolgUS, 
 
 revolsum, ^ 
 
 Tolpes, 
 
 Voltumus, etc.. 
 
 * Corssen's Aussprache, 1, 260. 
 
PRONUKCIATIOM- OF LATIN. 37' 
 
 while MSS. of Liyy have promuntorium, and those of 
 Ovid and Pliny show both coralium and curalium. (Met. 
 IV, 750.) 
 
 Thus in some cases even in the classical age (chiefly 
 before liquids, as appears in the above examples), there 
 must have been a sound given to m approximating to the 
 0, perhaps analogous to that which is heard for o in Eng- 
 lish front and London. But the tendency, in the best 
 period of the language, was to distinguish and purify 
 the u J and when the word avolsa was written avulsa, 
 we must suppose that the u was sounded as in English 
 full {i. e., its own sound). 
 
 2. A second impurity of the Latin u was a tone Kke 
 that of Greek v [alluded to apparently by Velius Long., 
 p. 2215, from Verr. Plac], heard by the Greeks, at least 
 sometimes, in : 
 
 Eomulus, which appears as 'PujuvAo?, ' 
 
 Marullus, " " " MdpvXkog, 
 
 TuUius « « « Tvk^iog, 
 
 Capua " " " Kanvrj, 
 
 and Cures " " " Kvpeig. 
 
 This must have been the half-way sound spoken of by 
 Quintilian (I, 4, 8), when he says : medius est quidam U 
 et I literaa sonus; non enim sic optimum dicimus ut 
 opimum." In optimum, therefore, even after the change 
 from the older orthography of optumum, was heard this 
 middle sound, not quite u, not quite i, i. e., without 
 doubt the French u, or German il. This must have been 
 the sound of u in such words as : 
 
38 PRONUNCIATION OF LATIN. 
 
 sacrifuco, / \ sacriflco, 
 
 magnuflcTis, I jmaguificus, 
 
 simufex (signuficem),\ . ,. /signifex, 
 
 ° „ *■ ° ' Ifor some timef .„ 
 
 carnufex, J, „ ^, Icarnifex, 
 
 , , , (before they were),., , 
 
 lubet, \ .^^ ■' flibet, 
 
 Jwritten, I , 
 
 lacruma, / Vlacrima, 
 
 manufestus, ( Imanifestus, 
 
 maxume, \ /maxime. 
 
 So also probably in 
 
 Tolumus, sumus, 
 
 nolumus, emolumentum, 
 
 and in many other words, which like 
 
 inclutus, and inclytus, 
 
 obStupui, " obstipni, (Ter.) 
 
 Brundusium " Brundisium, 
 
 remained of uncertain orthography during the classical 
 period, (when i had become estaiblished as the regular 
 form in carnifex, libet, mstimo, victima, etc., etc.). 
 
 In these words the tai\. sound* of u (besides that form 
 of writing) was antique, and its remains in Cicero's time 
 were characterized by that author as rustic. The more 
 refined and approved sound was, however, not quite that 
 of i, as stated by Velius Longus (P. p. 3216) ; while 
 grammarians directed it to be nevertheless so written j 
 and cite C. J. Csesar as authority. 
 
 Mar. Victor. (P. p. 2465) : "voces istas (i. e. proximum, 
 etc.) per i scribite." lb. — " C.'Ceesar per i scripsit ut 
 
 * Atque lllis (antiquis) fere placuisse per u talia acribere et enun- 
 tiare. Vel. Long. 
 
PBONUNCIATION OF LATIN. 39 
 
 apparet ex titulis ipsius." (Cf. Cassiodorus ap. P. p. 
 2284.) 
 
 Agneeus Cornutus (P. p. 2284) : Lacrumm an lacrimce ; 
 maxumus an maximus, et si quae similia sunt scribi 
 debeant, quaesitum est. Terentius Varro tradidit, 
 Csesarem per i ejus modi verba solitum esse enuntiare 
 ft scribere. 
 
 [To represent this half-way sound was the object of 
 one of the three characters {\-) which the Emperor 
 Clatidius attempted to add to the Latin alphabet. Vid. 
 Tn<?. Ann. XI, 6. 13, and Suet, in Claud, c. 41, and the 
 Grammarians.] 
 
 RECAPITULATION. 
 1. The proper sound of u, 
 
 as in Brutus, Bructeri, bule, 
 Gr. 
 
 Jirutus, Jiructeri, bule, ) -p, . , 
 
 BpOVTOf, BpOVKTEpOl, POV^^T], ) -f' ' 
 
 The proper sound of t^ = English u in pull. 
 
 II. In classical times there were a few instances, ap- 
 parently accidental, in which syllables written with u had 
 something of the o sound. To pronounce o for u after 
 the ancient style, as dederont for dederunt, wquom for 
 mquum, was thought inelegant in the time of Cicero. 
 Yet the short u, especially in a final syllable before s and 
 
 .m, often had a neutral sound easily exchangeable with the 
 neutral sound of o. 
 
 III. TT sometimes (but more frequently i) is written 
 to answer to a sound like the French u. Thus, m in Latin 
 incluta, like u in French lutter. 
 
40 PRONUNCIATION OF LATIN. 
 
 CONSONANT "V^ (SEMI-VOWEL), VAU. 
 
 Withhold the intonation from the Towel u, and its 
 function is reduced to that of a consonantal check, equal 
 to the English iv, which, by the slightest alteration, may 
 be hardened to the adjacent consonant v. This change 
 from Yowel to consonant passed upon the Latin u when 
 it was followed in the same syllable by a Towel sound. 
 And the nature or power of the Latin v consonant was 
 just that of the English v, in which case it was fully a 
 consonant, or it only approached this, and retaining 
 something, but not all, of its original vowel character, 
 was equivalent to the English w. Accordingly, the evi- 
 dence which we have to adduce points to a normal sound 
 for this consonant, which is a' medium between English 
 V and w. It is soft like w when, after s, g, and q, in the 
 same syllable, it is followed by a vowel ; hard like v in 
 other situations. Thus, suetus, sanguis, quisquam, 
 sounded swetus, sangwis, qwisqwam, and seruus, uulgus, 
 ■servus, vulgus.* 
 
 In the way of evidence we remark : 
 
 1st. That this u is pointedly and most expressly dis- 
 tinguished by the grammarians from the u vowel ; as by 
 Nigidius (ap. Gell. XIX, 14, 6): « Fin Valerius, Volu- 
 sius, etc., is not a vowel at all." Quintilian (I, 7, ^6) 
 says that the writing ceruum with the sign which belongs 
 to the vowel u, does not represent the sound that is 
 heard. Again (I, 4), in his " seruus et uulgus -^oHcum 
 digammon desideratur." That is, " a suitable representar 
 
 * See Appendix, Note A. 
 
PBONUSrCIATIONOF LATIN'. 41 
 
 tive of the sound of the first u in seruus, etc., would be 
 made by a character like that of the (lost) ^olic F" 
 Accordingly the Latin grammarians very generally speak 
 of the consonant u by the name Digamma, as Agnseus 
 Cornutus (P. p. 2282) : " Ntfs hodie V literam in duarum 
 literarum potestatem coegimus ; nam modo pro Digamma 
 scribitur, modo pro Vocali." Velius Longus (P. p. 2323) : 
 " V literam Digamma esse — debemus advertere — in eo, 
 quod est QVIS." See also Mar. Victorin. (P. p. 2461).* 
 If we knew precisely and independently the sound here 
 contemplated as that of the Digamma, our inquiry touch- 
 ing the sound of u consonant, would now be satisfied. 
 But our knowledge of the Digamma's sound is conditioned 
 by the very question upon which we are now engaged. 
 It is necessary, therefore, to proceed, and to our first ob- 
 servation — ^that the u was truly a consonant — we add a 
 second, viz. : 
 
 2d. That even when a consonant the u was always felt 
 to have such intimate relations to the vowel u that they 
 each respectively easily passed into the place of the other, 
 as when for , 
 
 tenuior, we find tenvior,f 
 
 quattu-or, " " quat-tuor (quattvor), 
 extenuantur, " " extenvantur, 
 genua, " " genva, 
 
 and conversely for 
 
 silva, " " silua,J etc. 
 
 * Also p. 2463. 
 
 f Lachman on Lucret. p. 192-3, and Corssen IT, p. 167. 
 % In ancient Latin it would seem that a difference between the 
 vowel and the consonant u was scarcely appreciated, if we will 
 
43 PEOSrCNCIATION OF LATIN. 
 
 Destroy tlie connection of v with following vowel in 
 the same syllable, and v immediately reverts to u, thus : 
 
 From ancient plovo comes finally plno ; 
 
 « floyo' " « fluo; 
 
 From the root gavid, when i falls out, we have gaudeo ; 
 
 " " avid, " " audeo; 
 
 " navita, " " nauta; 
 
 With faveo, instead of favitum, " fautum ; 
 
 " caveo, " cavitum, " cautum; 
 
 " lavo, with falling out of i, " lautum. 
 
 And if any think it likely that the U of lautum, etc., 
 retains the consonantal power of the same u when fol- 
 lowed by a vowel (like modern Greek ai'TOf, pronounced 
 aftos), then to be considered is the further form latum.* 
 In the same way we get Opiter,'lor Aupiter, from avi- 
 pater.f 
 
 Seeing then, 1st, that u was certainly a consonant, and 
 as such, therefore, essentially distinct from u vowel ; and 
 2d, that this consonant was at the smallest remove from 
 the vowel u, the natural probability is in favor of the , 
 English w as its sound. 
 
 Compare now the Greek equivalents : 
 
 OvaKKatovg, for Vaccaeos, 
 liepovtXiog, " Servilius, 
 
 judge by the words of Scaums (De Orthog., P. p. S3S1) : antiqui — 
 — ignorantea earn {u) . praBpositum Vocali, consonantis vice fungi 
 et poni pro ea litera quse sit/. 
 
 * Cicero de Divinations II, 40, supposes cauneas intelligible for 
 cme ne eas. 
 
 f And vid. Prise. P. p. 560 : Non potest Vau, id est Digamma, in 
 fine syllabse invenirl> 
 
PEONUNCIATION OF LATIN. 43 
 
 OvTJ'ioi, 
 
 for 
 
 Veii, 
 
 'EkovTjTia, 
 
 te 
 
 Helvetia, 
 
 Kovadoi, 
 
 te 
 
 Quadi. 
 
 This orthography of the Greek favors the opinion that 
 the Greeks heard just our w sound in the words above 
 given.* But the passage from to to v is short and easy, 
 and there is reason to believe that the latter came to be, 
 and in the classic age was, nearer the actual sound of u 
 consonant in the majority of cases, i. e., prqbably every- 
 where except in the combination qu, gu, su (of which 
 more particularly below). For this opinion there is sup- 
 port to be found in the general tendencies of the Latin 
 speech ; but even to glance at this would carry us too far, 
 and we will adduce here only : 
 
 1st. The circumstance that, besides the spelling of ov, 
 the Greeks employed also /? to represent Latin u conso- 
 nant, thus : 
 
 Bijj'ot,- for Veii; 
 
 <i>Xd(iio^, as well as ^XaovCog, " Flavins, 
 'EA/3jyT(a, « Helvetia, etc., 
 
 and these forms were in use at the same time with the 
 others in ov. 
 
 2d. The statement of the grammarian Velius Longus 
 (P. p. 2233): "VHteram Digamma esse interdum, non 
 tantum in his (vocibus) debemus animadvertere, in qui- 
 bus sonat cum aliqua aspiratione, ut in Valente et Vitulo, 
 etprimitivo, et Genitivo, sed etiam in his quibus confusa 
 haec litera est, in eo quod est Quis. 
 
 * Ow is also the modem Greek equivalent for English to, as in 
 QMXliyyTuv for Wellington, etc. 
 
44 PEONUNCIATION OF LATIN. 
 
 Now in view of all that we have heretofore seen of the 
 nature of this letter u, the natural conclusion (it seems 
 not too much to say), the irresistible conclusion from 
 such language is, that Velius Longus understood by the 
 ■ u, "sounded with an aspiration," the sound of English v, 
 as Valente ; while for that sound of it in which he says it 
 was "combined, or confused (confusa)" the English w 
 answers for it exactly, as Qwis. 
 
 That the w sound was that heard in u after g, s, and q, 
 seems to have been certainly the opinion of Priscian, who 
 says (Lib. I) : Est, quando amittit V vim literse tam 
 vocalis quam consonahtis {i. e., becomes a sort of half 
 way thing between the two), ut cum inter Q et aliam 
 Yocalem ponitur . . . . uti : QVISQVAM. Hoc idem ple- 
 rumque patitur etiam inter ff et aliquam yocalem, ut : 
 sanguis, lingua. S quoque antecedente u, et sequente a 
 vel e hoc idem fit, ut : suadeo, suavis, suesco, suetus. 
 
 It will be seen that Priscian here makes no distinction 
 whatever as to the sound of u in gu-, su-, qu-. 
 
 It has been frequently affirmed * in modern times that 
 the u in qu was silent in the classical Latin, and that 
 these two letters together represented no other sound than 
 that of the guttural tenuis h; thus ki, kae, kod, kis, etc., 
 instead of kwi, kwse, etc., for qui, quas, etc. There is 
 certainly some show of ground for this opinion, and since 
 that is true, we may not perhaps (even upon the proof 
 we have brought), assume the matter to be entirely set- 
 tled, until an investigation is undertaken more extended 
 and particular than is now possible to us. Such an 
 
 * See the bold statement made by J. P. Richardson, in Roman 
 Orthoepy, p. 43 and after. 
 
PE0NU2!fCIATI0N OF LATIST 45 
 
 investigation will certainly make it evident that in a 
 number of particular cases the u after q was not heard,* 
 in some regularly, as guum (cum), quotidie (cotidie), and 
 others : in some generally, as equus (ecus ?), coquus f 
 (cocus), and others: in some occasionally, as quoquef 
 (sounded koke), and others. Other reasons, moreover, 
 and reasons too of apparent force, have been given in 
 favor of the opinion that the dropping of the u element 
 was the rule in qu. But our belief is that the superior 
 weight of evidence lies decidedly on the other side. We 
 shall have to content ourselves here with a few of the 
 many very clear statements of grammarians. 
 
 Vel. Long. (P. p. 2323) : " V literam Digamma esse — 
 debemus advertere — in eo, quod est QVIS." 
 
 Quintilian XII, 10, § 30 : " Duras et ilia (sc. Q) sylla- 
 bas facit, quae ad conjungendas demum subjectas 
 sibi vocales est utilis, alias supervacua, ut equos hac 
 et equum scribimus; cum etiam ipsse hse vocales 
 {i. e., uo and uu) duse efficiant sonum, qualis apud 
 Grcecos nullus est, ideoque scribi illorum litteris non 
 potest." But what easier than e«tof, Inovfi ? 
 
 Scaurus (P. p. 2253) : " Q litera aeque retenta est, — — 
 quia cum ilia V litera conspirat, quoties consonantis 
 loco ponitur, id est, pro vau litera, ut: QVIS, 
 QVALIS." 
 
 * It ought to be considered that u was not unfrequently dropped 
 in sound after other consonants also, as after «, g, and t. Cf. 
 smium for sxiaimim, urgeo for urgueo, and the old Latin sarn for 
 Buam, Us for tuU, etc. Vid. Schleicher, Zeitseh. f. Vergl. Sprachf., 
 II, 377. , 
 
 f See the pun of Cicero reported by Quintilian I, 6, 3, 47. 
 
46 PEONUNCIATIOK OF LATIN. 
 
 Marius Victorinus (P. p. 2461) : "Fos vero, qui nori 
 liabemus hujus vocis {i. e., Digamma) — notam, in 
 ejus locum, quoties una Tocalis, pluresve junctae 
 
 unam syllabam faciunt, substituimus V literam 
 
 At cum (V) prseponitur vocali, tunc accipitur pro 
 consonante, ut est ANGVIS, BXTINGVIT, IIN- 
 GrYA, PELVIS." 
 
 We are ready to admit that in qu there was a some- 
 what closer union (" fusion ") of the two elements than 
 was the case in gu, or su. But further, there is a state- 
 ment of Priscian, quoted by Corssen (Aussp. I, p. 37), 
 which indicates his belief that the u of gu becomes more 
 or less influenced (assimilated) by the following vowel. 
 It may be taken for what it is worth : u autem quamvis 
 contractum, eundem tamen [hoc est y] sonum habet inter 
 q et e Yel i Tel w diphthongum positum, ut " que, gucB," 
 nee non inter g et easdem vocales, cum in una syllaba sic 
 invenitur, ut 'pingue, sangi^is, linguce.' This places the 
 u of qui, qum, etc., in precisely the same category with 
 the u of pingue, Ungues, etc., and further attributes . to 
 .the u, when followed by the i and e sounds, a character 
 somewhat accommodated to the same : that is, the u of qui 
 and of lingucB are here contemplated as having a sound 
 like u in the French Mui, while before a and o (quo, 
 linguam) the sound would be that of French ou in 
 Edouard. 
 
 In the following extract from Q. Terent. Scaurus (ap. 
 P. p. 3261) there is more than we have space to develop. 
 We quote it for its bearing upon the points already 
 made: 
 
 Quis quidam per cuis scribunt, quoniam supervacuam 
 
PEON UNCIATIOK OF LATIN. 
 
 47 
 
 osse Q literam putent ; sed nos cum ilia («. c. Q) V. literam 
 et S [«. e. su, as in suetus], quando "tertia ab ea vocalis 
 ponitur, consentire jam demonstravimus ; . autem in 
 dativo ponemus ut sit differentia, cui et Qui, id est dativi 
 et vocativi singularis ctt. 
 
 Beda, de Orthogr., ap. P. p. 3333 : Goquus, coqui, prima 
 syllaba per C, secunda per q scribenda est [as in the 
 above quotation, q is supposed to involve hy implica- 
 tion the after sound of -a]; non enim dicimus Quo- 
 QVEBE, sed GoquEEE. Cf. Aul. Gell. X, 11, § 9. 
 
 It remains to compare the, 
 among "which we find : 
 
 Greek equivalents for Qu ; 
 
 Kovadoi 
 
 for 
 
 Quadi, 
 
 TopKovarog 
 
 ei 
 
 Torquatus, 
 
 XTjKovavoi 
 
 li 
 
 Sequani, 
 
 Kovipivog 
 
 ti 
 
 Quirinus, 
 
 inpEKovevTeg 
 
 t£ 
 
 obsequentes ; 
 
 also, 
 
 
 
 K.otvTog 
 
 for 
 
 Quintus, 
 
 KoaSoc 
 
 a 
 
 Quadi, 
 
 ^TiKoavoi 
 
 ct 
 
 Sequani ; 
 
 and, before i, 
 
 
 
 Kvpivog 
 
 for 
 
 Quirinus, 
 
 TapKvvia 
 
 ee 
 
 Tarquinia, 
 
 'AKvXag 
 
 le 
 
 Aquila, 
 
 Kvplreg 
 
 ce 
 
 Quirites. 
 
 All these point directly to the sound of English w, for 
 that of u in qu, except the last case, viz : kv for qui : 
 
48 PBONUNCIATION OF LATIN. 
 
 and this seems to give eyidence of an extreme attenuation 
 of the u of qu when followed by i. If Kvplvo^ alone 
 were found, this would indicate Kyrinus as the sound of 
 Quirinus ; but when we find in the same author (Strabo), 
 both KvpivLog and Kovpivoq (Oorssen's Aussp., I, p. 37), we 
 are forced even here to strike a mean, and to assume the 
 normal sound of u in this situation to have been like 
 that of French u in cuir. The fairest inference, then, 
 which can be drawn from this fluctuation of Greek ren- 
 derings, is that practically the force of u in qu was often 
 lost before i, leaving its trace only in the attraction of 
 the following i towards the u sound ; but that often also 
 the u was retained in its natural w sound, somewhat 
 refined by the influence and in the direction of the sub- 
 sequent vowel. And here, therefore, so far as this evi- 
 dence is concerned, we shall have to suppose an irregu- 
 larity of pronunciation within very narrow limits which 
 can no longer be precisely ascertained. For our own 
 practice there is no other resort than the employment of 
 the normal sound. 
 
 In view of all that has been said, therefore, we infer, 
 that in Latin the consonant F sounded like English V 
 when it began a word or a syllable (unless in combination 
 with preceding s, g, or q). When, alter s, g, and q, it 
 began a syllable, the intonation of the vowel u was par- 
 tially supplied to it, yielding a sound like English w. 
 After s and g, the u was sometimes lost ; and after q the 
 same was more frequently the case, although, with the 
 exception of a few words, the particular instances in 
 which the u thus vanished, cannot now be sufficiently 
 identified. Therefore : 
 
PRONUNCIATION OF LATIN. 49 
 
 Validus, seruus, uulgus, sounded Validus, servus, vulgus 
 
 (comp. Eng. Valid). 
 Suavis, suesco, " Swavis, sivesco 
 
 (comp. Eng. Sweet). 
 Unguentum, lingua, " Ungiventum, lingtva 
 
 (comp. Eng. Unguent). 
 . Quesiio, relinquo, quamquam, " Qwestio, relinqwo, qwam- 
 
 [qwam (comp. Eng. Question). 
 
 DuEiNG the classical age the vowel Y was not theo- 
 retically (in strictness not even actually) a member of 
 the Latin alphabet. Cicero (Nat. Deor., II, c. 37) admits 
 twenty-one letters only in the Eoman list. 
 
 The poet Attius made no use at all of Y in his writ- 
 ings, according to the testimony of Marius Victorinus 
 (p. 2456) : Accius — nee z literam nee y in libro suo 
 retulit. Shortly after the time of Attius, however, the 
 foreign sign was commonly adopted for the more perfect 
 writing of foreign words. Cicero alludes to this practice 
 as modem in his day. ■ Ofat., c. 48 : Burrum, semper 
 
 Ennius, nunquam Pyrrhum : Bruges non Phryges 
 
 Nee enim (antiqui) Greecam literam adhibebant; nunc 
 autem etiam duas. Yet the Y was not admitted to a 
 place in Latin words, and Quintilian (XII, 10, 27) still 
 treats it expressly as a foreign element : . . . . incundis- 
 simas ex Graecis litteras iion habemus, vo,calem alteram 
 alteram consonantem .... ut in Ephyris et Zephyris. 
 
 Accordingly, in a systematic treatment of Latin letters 
 3 
 
60 PRONUNCIATION OF LATIN. 
 
 the Y can claim no place. Still, for practical purposes, 
 we require to know how the Komans rendered that ele- 
 ment of Greek words, employed by them, which they 
 represented by this sign. 
 
 In order to this, it will- be enough to present the fol- 
 lowing brief considerations. 
 
 Latin Y, it must be repeated, was not .merely some- 
 thing like the Greek T, but was understood, was designed, 
 to represent that very sound itself. (See Quint. 1. 1.) 
 
 Now Greek T was sometimes rendered into Latin by 
 IT: {Bruges, Burrus, Eurudica). It was also at the 
 same time (later far oftener)' represented by I : {Stigio 
 for liTvyiu, Sisipus for Xiav<j>og, etc.). It would there- 
 fore seem to have had a sound between that of Latin U 
 and Latin I. ISTow we know that a middle, uncertain 
 sound of this sort was very familiar to the Latin speech ; 
 and the final proof that the Y represented precisely or 
 nearly precisely that sound is found in direct statements 
 of gramtoarians to that effect, as thus : 
 
 Mar. Victor. (P. p. 3465) : Sunt qui inter u quoque et i 
 literas supputant deesse nobis voces (?), sed pinguius 
 quam i, exilius quam u; sed (pace eorum dixerim) non 
 vident y literam desiderarij .... Priscian, speaking of 
 the same sound, upon the testimony of Donatus, says 
 (P. p. 539), sonum y grsecae videtnr habere.. An interest- 
 ing confirmation of the same is found in certain forms 
 reported from inscriptions of the time of the Emperor 
 Claudius, which show T represented by the sign invented 
 by that Emperor for the sound in question : as, Mg V pti, 
 Che nus for ^gypti, cycnus, and others. 
 
 There can be no doubt, then, that the sound given to 
 
PRONUNCIATION OF LATIN. 51 
 
 Y by the Eomans -was between u and i, that is, the sound 
 of French u or German u — the German ii more particu- 
 larly perhaps, since that is more subject to be sounded 
 like pure I, which was the most frequent corruption of Y 
 in Latin. 
 
 DIPHTHONGS. 
 
 A diphthong is the sound resulting from an effort 
 to pronounce two vowels, the one after the other, at one 
 breath. The sound, therefore, of a diphthong is a unit, 
 but begins while the organs of the vocal tube are set for 
 the sound of one vowel, and does not end until after 
 those organs have assumed the attitude necessary for 
 rendering another. And further, as meanwhile the breath 
 must not be suspended or interrupted, the resulting 
 sound will also necessarily be modified by the intermediate 
 transitory conformation of the vocal tube during the 
 change. 
 
 This is the sense in which it is necessary to receive the 
 old definition of diphthong, as given e. g. by Terent, 
 Maurus (P. p. 2392) : 
 
 " Diphthongos eas 
 
 Graeci dicunt magistri, quod duse junctse simul 
 
 Syllabam sonant in unam." 
 
 (Cf. Priscian, P. p. 561.) 
 
 Upon the failure of any of the conditions named, the 
 di|ili^hong iHo ' &%s4iili^i[j 'fai!s';'"ahd' w^"" Me^ get in its 
 stead, the sound bf a vOwel and a consoaant, as aj (pro- 
 
52 PBOsruNoiATiojir of hxiis-. 
 
 nounced ay) for ai, or two vowels, as ai (pronounced a-ee) 
 for -ai (aula'i), or one long vowel, as Pretor for Prmtor, 
 usus for oisus. 
 
 The Latin language was once very rich in diphthongs, 
 but gradually surrendered them by yielding to the im- 
 perfections which have just been pointed out ; until the 
 only diphthongs recognized by ancient grammarians as 
 heard in the classical age -were four, viz. : m, au, eu, and m. 
 Some admit ei, but allow it to be a diphthong only in a 
 very few cases ["ista rarior Diphthongus EI;" Ter. 
 Maur., p. 3393], while others distinctly reject it, as 
 Diomedes : " Ex his Diphthongus EI ; cum apud veteres 
 frequentaretur, usu posteritatis explosa est." And fur- 
 thermore, we shall find good reason to distrust the 
 integrity of some of the admitted four. The most of 
 them, indeed, nearly succumbed to a corruption, which 
 none of them entirely escaped. There are fluctuations 
 here, which fall within the classic period, and which do 
 not square with that theory of beautiful uniformity and 
 consistency which some have been satisfied to assume for 
 the Latin pronunciation. This must now be made to 
 appear by a particular examination. But first, a few of 
 the statements of grammarians, as to the number and 
 kind of diphthongs in the classical LatLa. 
 
 Terent. Maurus (P. p. 2393) : 
 
 " Porro Diphthongos Latini quatuor fixas habent. 
 Quatuor ideo separavi, quinta quod sit rarior." 
 
 Priscianus (P. p. 561) : Sunt igitur vocales prsepositivse 
 aKis vocalibus aubsequentibus in eisdem syllabis : a, e, 
 o; subjunctivse : e, u, ut (s, au, eu, m. Jquoque apud 
 
PEONUNCIATIOK OF LATIN. 53 
 
 antiques post e ponebattir, et ei Dipthongum facie- 
 bat, quam pro omni i longa scribebant more antiquo 
 
 GrsBCorum Sunt igitur Diphthongi, quibus 
 
 nunc utimur, quatuor. 
 
 Diomed. (P. p. 422) : Cum .... duse vocales jungun- 
 tur, ut CB, au, eu, os, ei, yi. Ex his Diphthongus ei j 
 cum apud veteres frequentaretur, usu posteritatis 
 explosa est. Item yi grseca potius, quam latina est. 
 
 The use of ei for I (as mentioned by Priscian), was 
 condemned ah-eady in Cicero's time by Nigidius, 
 ap. Aul. Gell. N. A., XIX, 14, 8 : Grascos non tantee 
 inscitiae arcesso, qui ov ex o et n scripserunt, jquantte 
 nostri fuerunt, qui {e)i ex e et i: illud enim inopia 
 fecerunt, hoc nuUa re subacti. 
 
 Upon consultation of the organs of speech, it seems 
 likely that the sound of a would lie between that of the 
 diphthong ai and that of one of the simple elements a 
 or e. Accordingly, in that well ascertained progression 
 which the Latin made in the diminution of its vowel 
 sounds, we find this regular descent : Oaisar, Ccssar, 
 Cesar ; praifectus, prwfectus, prefectus j quaistor, qumstor, 
 questor. On the testimony of those who have access to 
 the careful reports which have been made of Latin 
 inscriptions, we know that forms like these belong 
 respectively to the ante-classic, the classic, and the post- 
 
54 PBONUNCIATIOK OF LATIN. 
 
 classic periods.* The first transition, that of ai to cb, is 
 in entire conformity with ithe statements of the ancient . 
 grammarians, who represent the change as so far com- 
 plete^n the cultivated age of Latin literature that every 
 instanfe of ai employed in that period was to be regarded 
 merely as a vestige of antiquity. 
 
 Quint. I, 7, § 18 : 2E syllabam, cuius secundam nunc 
 E litteram ponimus, varie per ^ et / efEerebant ; qui- 
 dam semper ut Grseci, quidam singulariter tantum, 
 cum in dativum vel genitivum casum incidissent, 
 unde pidai vestis et aquai Vergilius amantissimus 
 vetustatis carminibus inseruit. In iisdem plurali f 
 numero E utebantur, hi Syllm, Galhm. 
 
 But the further transition, that namely of a into o 
 is equally manifest from inscriptions on which are found : 
 from the first century after Christ, Lelio for LcbUo, prime 
 for primm (genitive singular) ; from the second century, 
 condite (dative singular) ; from the third century, aque, 
 jMfrie, for aguw, etc. (genitive singular) ; from the fourth 
 century, que for quce (nominative singular), ahd many 
 others. Of course the change of sound which preceded 
 these changes in orthography must have been more or 
 less gradual, and for a time it may have been hard for 
 the ear to decide whether the sound heard was one which 
 would be more truly represented by cb or by e. Such a 
 complete uncertainty is shown by inscriptions of the 
 third century after Christ, on which are found such 
 remarkable forms written together as : 
 
 * See Corssen's Aussp,, I, p. 178. 
 
 f Corssen shows from inscriptions that cb was frequent in the 
 plural before it was common in the singular, for ai. 
 
proktikciatioh' of latin. ,55 
 
 Dative. Geititite. 
 
 bonae femine, mire sapientiaB, 
 
 Impise Juste, vitas nostre, 
 
 Fructuose filise, Coccejae Severe, 
 
 As ■well also as the e alone, as : 
 
 Dative. Genitive. 
 
 Julie felicissime, sancte memorie, 
 
 Tulliane Marcelle, vite sue, 
 
 mee pudicissime, benignissime femine. 
 
 HoTv^ thorough was, in later times, the prevalence of 
 the e sound in place of the Latin a, scarcely need be 
 mentioned ; and no more convincing proof of it can be 
 given than the frequ,ent occurrence in the copied MSS. 
 of ancient authors which have reached us of a mistake 
 by which cb is written where e properly belongs ; as, e. g., 
 sjjrmtorum for spretorum.* 
 
 The principal question for us, therefore, is : when did 
 that sound which in classical Latin is represented by cb cease 
 to be spoken like ai, and when did it begin to be undis- 
 tinguishable from the sound of e ? The most satisfactory 
 of the approximate answers possible to this question, will 
 be found by consulting the varying orthography of the 
 monumental records of the language according to its 
 different periods. Those records are beyond our reach, 
 but all who will consult Corssen's elaborate comparison 
 of the forms which he has drawn from that instructive 
 source, will be ready to accept his conclusion, that the 
 
 * So in the Cd. Puteanus (eiglith century) of Livy, XXI, 63. 
 Vid. Fabri ad loc, cf. sesculus, esculus, csella, cella, cena, casna, 
 caespes, cespes, Bsecnlum, seculum. 
 
56 PEOKUNCIATIOIfr OF LATIN. 
 
 testimony of inscriptions is about tlie following: "As 
 early as the Syrian war (say B.C. 190), the sound ai was 
 giving way to that of m, and this sound prevailed uni- 
 versally after about the time of the Gracchi (say B.C. 130). 
 But already in ancient times cb (at least by the vulgar) 
 was frequently sounded like e, and from the earliest 
 times of the Empire this e for m was heard in the mouths 
 of the cultivated; and, gaining in use and favor, iy 
 about the third century a.d. this was the exclusive sound 
 of that sign." 
 
 The first part of this testimony, i. e., the abandonment 
 of the sound as well as the form ai before the classical 
 age, is affirmed also, directly and indirectly, by QuintiUan 
 and the grammarians. Thus QuintiUan (as we have seen 
 above) .says : "The syllable which we now make m they 
 {i. e., the ancients) nsed to ^pronounce (N.B., not only 
 write !) ai — some invariably, others only in the genitive 
 and dative singular." * And Virgil's use of gictai, aulai, 
 he ascribes to the Poet's love for the antique. Compare 
 also Priscian, P. p. 738.t 
 
 The sound of cb, therefore, was not that of ai. But 
 before the bloom of the Latin literature was passed, it 
 
 * This, in order to distinguish the singular forms in m {ai) from 
 the nominative plural, as appears also in a similar but slightly dif- 
 ferent rule laid down by Nigidius (ap. Gell. XIII, 86). The 
 inscriptions, however, show that these rules were not observed in 
 practice. 
 
 f In the face of these statements, and in spite of all the evidence 
 of a changed orthography. Prof, Richardson (Roman Orthoepy, 
 N. Y., 1859) lays down the law, " M (= AI) sounds like ay, the 
 English adverb of affirmation." Will he render Virgil's ^oem 
 after this fashion? (Vid. .iBneid, III, 886.) And with his qu = k, 
 ■will he read Inf ernike lacus AyayayM insula 'K.iileay f 
 
PRONUNCIATiaN OF LATIN. §7 
 
 ■was to a Tery large extent confounded with the e sound, 
 •nath which it was soon after entirely identified. That in 
 the early classic period, however, there was a separate 
 sound for m, at least in the city of Rome, may be seen 
 from Varro de L. L., VII, 96: "In pluribus verbis A 
 ante E alii ponunt, alii non . . . . ac rustici Pappum 
 Mesium non Mcssium" Again, Lib. 4, c. 9 : " In latino 
 rure Hedxjs, quod in Ukbe, ut in multis, A addito, 
 H^Dtrs. It is important to notice the terms of these 
 statements. When taken together with what we know 
 of the history of this diphthong, they seem to point us to 
 the characteristic sound of which we are in search. This 
 was the sound which very soon was e in everybody's 
 mouth ; but now (in Varro's time), at least with the city 
 people, it was e " with an a sound before it" * The two 
 elements are not spoken of as though they were of equal 
 importance. The e evidently has the most weight in the 
 mind of the writer. The e being already there, as it 
 were, "the a is put in before it by some." Hsedus is 
 Sedus " with an a added." It can hardly be doubted 
 that the sound so described was that of ae' and not d'e: 
 d'e could hardly be distinguished in practice from the 
 old ai, while ae' is on the high road to e — just what the 
 history of the sound demands. For the earlier part of 
 the classical period, therefore, we infer that the sound of 
 CB was that pf ai in French raison, while before the end 
 of the classical period, it is certain that the fine distinc- 
 tion between that sound and the sound of the simple 
 vowel e was to a great extent ignored, and soon after 
 became extinct, f 
 
 * See Appendix, Note B. f See Appendix, Note C. 
 
58 PEONUKCIATION 3? LATIN. 
 
 AV, the diphthong fullest of tone, was the only one 
 of the Latin diphthongs which maintained its integrity 
 throughout the period when good Latin was spoken. 
 The natural sound of a combination of a and u, if the 
 first element be given with stress, is just that of English 
 ow in now, or ou in flour. And the records of the lan- 
 guage, the comparison of the Greek, the tradition of 
 scholars, and the absence of important adverse evidence 
 all conspire to confirm the probability of this as the true 
 sound of au. 
 
 1. The written sign AV, is said to be very abundantly 
 found in Latin inscriptions, from the earliest to the latest 
 period. 
 
 2. The Greek rendering of au maybe seen in KXavdwg, 
 ^avarvXo^ (Strab.), KvXipKioi (Latin Aulerci) ; while the 
 Latin renders the Greek av by au in Aulis, aulcedu^, 
 Auletes, etc., etc. 
 
 3. The diphthong au was not entirely exempt from the 
 influence of the general devocalizing tendency of the 
 Latin, and from the time of the earliest records onward 
 (through the classic period and afterward), it was not 
 unfrequently reduced very nearly, if not quite, to- the 
 simple sound of its second element, or to that of the 
 cognate vowel o.* The first (i. e., the change to u) could 
 be effected by removing the stress of utterance to the 
 second element (au instead of du), which would leave the 
 first an easy prey to neglect or assimilation. The second 
 
 * See Appendix, Note D, 
 
PEONUKCIATION OF LATIK. 
 
 59 
 
 change involved a further departure, the substitution, 
 namely, of kindred o for u, and the assimilation of the a, 
 as in the first case, thus ; du, do, ao, 5.* But whatever 
 the mode, the facts are : 
 
 
 (m for au) 
 
 rudus 
 
 for raudus, 
 
 adrudus 
 
 " (adraudus), 
 
 defrudo 
 
 with fraudo, 
 
 frudavi 
 
 " fraudavi, 
 
 cludo 
 
 " claudo ; 
 
 
 (o for au) 
 
 Pola for 
 Plotus . " 
 
 Paula, ?, , , 
 Plautus, r^'^'^-''^^ 
 
 dehorito " 
 
 dehaurito, {Cato.) 
 
 plodo " 
 
 plaudo, ( Cicero.) 
 
 clodicat as well as claudicat, " 
 
 plostra for plaustra; {Suet. Vesp. 22.) 
 
 And in the fifth century, a.d., when Diomedes wrote, 
 
 clostra was heard as well as claustra, 
 coda " « " " " Cauda. 
 
 For the interpretation of these facts, it must be re- 
 marked first, that the change into u was very rare. And 
 those few cases in which this sound was heard as a 
 representative of earlier au, so far as they have any bear- 
 ing upon the general sound of the diphthong in question, 
 imply the probabihty of English oio as that sound. Thus 
 
 * Gruter (quoted by Corssen, I, p. 168) reports Aordius from a 
 very ancient tablet found at Spoletum. 
 
60 PRONUNCIATION OF LATIN. 
 
 cldudo {du = English oio) need only be sounded claudo 
 with, no change in the quality of the Yowel sounds, and 
 the result is already almost identical with cludo. 
 
 The number of cases in which o appears for au is Tery 
 much more considerable; but still, in mere point of 
 numbers, these instances are not sufficient to warrant the 
 belief of anything like a generally preyailing confusion 
 of the sounds of o and au — a confusion which some 
 modern scholars have been too prone to assume. The 
 conclusion thus stated may be defended by argument ; 
 but we happen to possess an anecdote told by Suetonius 
 (Vit. Vesp. c. 32), which settles beyond dispute that in 
 the time of Vespasian, at least, there was a clear" difEerence 
 between plausira and plostfa, Florus and Flaurus. The 
 emperor, who was rebuked by his courtier for saying 
 2}lostra, is characterized by the historian, just before and 
 in the same immediate connection, as being fond of low 
 expressions. There are other indications also,* that o 
 for au was in most cases a yulgarism, and without the 
 sanction of the learned. We may therefore safely belieye, 
 that not only properly but actually as a general rule 
 among the educated, au had its own sound, i. e. 
 
 AY sounded like ou in English cloud. 
 
 In many instances among the people, and in some also 
 in the practice of the cultivated classes, au was allowed 
 through carelessness, or for some other reason, to sink 
 into the sound of o (may-be a little thicker than the 
 common o ?). In these cases the orthography accom- 
 modated itself to the change ; and we have now no right 
 
 * See CoTssen's Aussp., I, p. 167. 
 
PBONUNCIATIOK OF LATIN. 
 
 61 
 
 to suppose that any word which we find written with au 
 was sounded with o. 
 
 CE. 
 
 The sign CE is representative of the most unstahle, 
 and therefore the most uncertain of all the Latin vowel 
 sounds. If it only marked one stage of the descent oi, 
 ce, e, we might surely infer at least a close approximation 
 to its sound, as we do for that of w in the scale ai, ce, e. 
 But the old Latin oi descended through ce in three direc- 
 tions, namely to u, to i, and to e (or m). Thus : 
 
 oisus. 
 
 cesus, 
 
 usus. 
 
 coiravit. 
 
 coeravit, 
 
 curavit. 
 
 ploira, 
 
 ploera. 
 
 plura. 
 
 noiv^, 
 
 poena, 
 
 punire, etc. ; 
 
 foidus, 
 
 foedus, 
 
 fidus. 
 
 ohog, 
 
 
 vicus. 
 
 quoi, 
 
 
 cui. 
 
 also, 
 
 (while the cases ending in -i of stems in o, have in this 
 i the remains of former -oi, as populi for pojmloi, Romani 
 for Bomanoi. The intervening form cb is said to appear 
 in a few acfcidentally preserved ancient instances, as 
 Pilumnce, poplce, nominative plural) ; 
 
 and. 
 
 moiros, 
 noivrj, 
 
 pomoenum, 
 poena, 
 
 pomenum, 
 
 paenitet (and poenitet), 
 
63 PRONUNCIATION OF LATIN. 
 
 coena, csena (and cena, Cato and Ter.), 
 
 obscoenus, obscsenus (and obscenus), 
 
 Coelius, CaBlius, 
 
 coeteri, cseteri (and ceteri), etc. 
 
 From tbese obseryations we should be obliged to 
 infer that the sound of ce diyerged from that of oi in at 
 least three different ways. And when now we turn to 
 inquire in what order of time those yarious forms of 
 orthography prevailed, and to what extent respectiyely, 
 we find in Corssen's exhibition drawn from inscriptions, 
 
 1. That oi (haying begun from the earliest times to 
 give way to (b) had ceased entirely to be written before 
 the classical period. 
 
 % That from an early period i was, collaterally with cs, 
 a substitute for earlier oi in the terminations of cases of 
 nouns. (In this place m gaye way entirely to i in the 
 classical period.) 
 
 3. That from the time of Plautus on, m in the stems 
 of words gaye way to u, which became the prevailing ortho- 
 graphy after the time of the Gracchi and through the 
 classical age. 
 
 4. The few words which during this time retained the 
 written form ce were pronounced by the later Latins with 
 an c* 
 
 The most positive inference, which we seem entitled 
 to draw from such evidence, is, that along with the sign 
 of oi, the sound also of that diphthong had disappeared 
 from the Latin before the classical age. And the pro- 
 posal, therefore, of some modern scholars to return to 
 
 * See Corssen, I, p. 307. 
 
PKONUKCIATION OF LATIN. 63 
 
 that antiquated sound for ob is not only in coniiict with 
 the broadly marked general tendency of the Latin towards 
 a redaction of its diphthongal sounds, but singularly at 
 variance with the particular history of the diphthong ce. 
 The zeal of reform has impelled some to deny the exist- 
 ence of any irregularity or obscurity in the Latin pro- 
 nunciation : but these cannot help us h.ere ; nor can we 
 adopt any such adventurous dogma as prescribes for our 
 practice a sound which was obsolete for Cicero. 
 
 There is certainly a consolation for our ignorance in 
 the belief that the Eomans themselves did not know for 
 any length of time together a certain sound for this 
 diphthong. And if now it be thought necessary for us 
 to fix some uniform sound, by which in our practice to 
 render those words which, in printed texts furnished us, 
 appear written with the sign ce, we must choose between 
 the sound of e, as in Latin fetus, French pritre, or that 
 of German oe (6) in Goethe. The first of these sounds is 
 recommended : 
 
 1. By considerations of convenience, chiefly in view of 
 the fact that in almost every instance (except, namely, in 
 the word ccetus and possibly a few others) the words 
 
 ■ which are spelled in our texts sometimes with cs, are also 
 frequently (some as frequently) found with m, as : olscoi- 
 nus, olsccenus, and even obscenus (French obscene) ; 
 pcenitet, pmnitet ; mceror, mceror, etc., etc. 
 
 2. Where better reasons fail, the authority of the post- 
 classic Latin deserves to be heard ; and this speaks for e. 
 On the other hand, the general turn taken by m into ii, 
 in the practice of the cultivated Eomans of the classical 
 period, establishes a good degree of probability for the ■ 
 prevalence at that time among the educated of a sound. 
 
64 PRONUNCIATION OF LATIN. 
 
 ■whicli conjecture can represent no better than by that 
 which the German scholars have adopted; the sound, 
 namely, of their own w — [nearly the French eu in/ewr].* 
 
 Thus the oe, in prmlium would sound nearly like 
 
 in English ivorld or 
 i " " whirl. 
 
 "The diphthong c::, if we except Greek words, occurs 
 only in Jieus, heu, and eJieu, in ceii, seu, and neu, and in 
 neuter and' neutiquam." (Zumpt, Lat. Gr. p. 1.) The 
 orthography of these words was uniform, and there is no 
 reason to suppose that the integrity of this diphthong 
 was ever disturbed except by the license of poetry — it 
 being rendered sometimes in yerse by diaeresis, according 
 to the testimony of Servius ad. ^n. II, 69. " Heu modo 
 est una syllaba, sed interdum propter metruni duae sunt, 
 ut est : He u ! quam pingui, etc." Supposing, then, 
 both elements to be heard, as is necessary for a true 
 diphthong (nam singulse vocales suas voces habent; 
 Priscian, P. p. 561), it becomes only further necessary to 
 know which element has the stress. There is no other 
 Latin diphthong in which the stress does not associate 
 itself by preference either with one or the other of the 
 two combined constituents, and upon the selection thus 
 made depends the principal coloring of the sound- E'ow 
 
 * See Appendix, Note B. 
 
PEOKUKCIATIOK OF LATIK. 65 
 
 we know of no case in whicli Latin eu has passed into e, 
 but some instances are reported from the early Latin 
 containing eu which was afterwards heard as u. Such 
 forms are Leucesio (Carm. Sal.), Teurano (Sc. d. Baccan.), 
 and others. This appearance argues in favor of the pre- 
 dominance of the latter element, which is also after the 
 analogy of w and m. We infer, therefore, that Latin eu 
 had the sound in perfect conformity with its simple ele- 
 ments, with the stress on the last. This is the sound of 
 English ew in few, pew, etc. 
 
 CONSONANTS. 
 
 Having found it convenient to consider the vowel 
 sounds separately, we pass now to the remaining charac- 
 ters of the Latin alphabet. In a systematic view of these 
 elements it would be necessary to divide them into sev- 
 eral classes according to the organs (or parts of organs) 
 chiefly engaged in their utterance : but for our present 
 purpose, which is special and particular, it will be best, 
 not indeed to forget the nature and power of letters 
 which is shown by the science of Phonetics, but to 
 keep a limited inquiry of this nature free from the 
 embarrassing admixture of general considerations. The 
 letters which remain to be treated of, therefore, we shall 
 consider under the one category of Consonants (among 
 which the breathing h also will be admitted to a place) ; 
 and take them up in the common order of the alphabet. 
 
66 PKONUNCIATION OF LATIN. 
 
 There is eyery reason to believe that the Latin B was 
 the same with our own. Tlie English 5 — the "soft 
 check" of the lips — described by M. Miiller (Lee. 3d 
 Ser., p. 156), will be found to suit, as no other consonant 
 ivill, the description of the Latin h, as given by Marius 
 
 Victorinus (P. p. 3454) : B et P literse dispari inter 
 
 se officio exprimuntur ; nam prima [5], exploso c mediis 
 labris sono ; sequens [ P], compresso ore, velut introrsum 
 attracto vocis ictu, explicantur. 
 
 Owr b (a compromise between the labial hard check 
 and the labial breathing) stands half-way between p, 
 on the one hand, and v, on the other. Precisely the 
 same was true of the Latin b j as is plainly shown by the 
 interchange now of h and jj, and now of h and v. If 
 these fluctuations were contemporaneous they would in- 
 dicate a very uncertain sound for b in practice, still 
 -implying, however, a normal sound equal to English 5. 
 But when it becomes evident that the interchange of b 
 with p belonged mainly to the early (ante-classic) Latin, 
 and that it was only in the later history of the language, 
 that b began to be confounded with v, then there can be 
 no reasonable doubt but that during the intervening 
 (classical) period the prevailing actual pronunciation of 
 5 was just that of our labial media. 
 
 In proof of the ancient interchange of b and ;j we have 
 a statement made on the authority of Papyrianus, that 
 Bnnius said, nupo, scripo, and, on the other hand, rebo, 
 
PRONUNCIATION OF LATIN. 67 
 
 for the later niibo, scribo, and repo. Cicero (Orat. 48, 160) 
 says of the same author, " Burrum semper Ennius, num- 
 quam Pyrrlium" And Q. Ter. Scaurus (P. p. 2352) says: 
 " Grseci nvpptav, nostri Byrriam. ; et quern Purrum, 
 antiqui Btcrrumj .... item Publicolam BobUcolam; 
 alii scapillum, alii scabillum dicunt." In most cases p had 
 become fixed instead of b before s and t in the classic age, 
 as in scripsi, nupta. Forms like opsonium, sup persona 
 (Eib. com. p. 273) were discontinued after early times, 
 but the b in obtineo and similar cases had a sound very 
 like p in Quintilian's time (aures magis audiunt p, Lib. 
 I. c. 7, § 7). B was therefore at this time different from 
 p, but nearer to it than to any other sound. After the 
 classic period it became softened to v j and inscriptions 
 of the late empire are said to show, 
 
 incomparavile for incomparabile, 
 
 venemerenti " benemerenti, 
 
 liventer " libenter ; 
 
 also, 
 
 fobere for fovere (a.d. 344), 
 
 fabente " fayente ( " 367), 
 
 bixit " vixit ( " 409), 
 
 and great numbers of similar forms. This v sound for 
 5, was perpetuated into the new languages * made from 
 the Latin [cf. English vervain, Latin verbena, English 
 tavern, Latin tdberna (modern Greek -ajSepva pronounced 
 taverna)], but did not prevail in the classical Latia 
 ifself.t 
 
 » The existing MSS. of ancient authors are said to be infested 
 ■witli errors from this source. Danuvius for Danubius, etc., etc. 
 f See Appendix, Note P. 
 
68 PKONUNCIATION OF LATIIJ-. 
 
 o. 
 
 Quintilian (I, c. 7, § 10) says : ' ISTam E quidem in nul- 
 lis verbis utendum puto, nisi qu» significat, etiam ut sola 
 ponatur. Hoc eo non omisi, quod quidam earn, quotiens 
 A sequatur, necessariam credunt, cum sit C littera, quse. 
 ad omnes vocales vim suam perferat. This language 
 merits careful attention, as it brings the weight of 
 Quintilian's authority directly to bear upon some of the 
 most important points raised in the dispute which has 
 been many times renewed over the proper sound' of 
 Latin c. The points made by Quintilian are these: 
 " K liaving become obsolete ought not now to be em- 
 ployed in the spelling of any words except those few for 
 which it is also used by itself {K.) as an abbreviation (as 
 K. for Kmso, or Kalendw) : that letter is not only anti- 
 quated but also superfluous as an element in words ; and 
 those persons are in error who hold that the guttural 
 tenuis must be given by Ic when an a is to follow it (as, 
 e. g., in Kaput, Kalumnid). For c is perfectly adequate to 
 take the piiace of k before a as well as hefore all the other 
 vowels, since it maintains its own sound in connection with 
 them all." 
 
 In view of the simple and unmistakable character of 
 such an announcement, when we consider farther, that 
 it is confirmed both directly and indirectly by later 
 grammarians, with never a word to the contrary as re- 
 gards the classical period, it would seem that we were 
 already at the end of our inquiry. For the sound of h is 
 entirely undisputed, and that, said Quintilian, was the 
 
PKOUUNCIATION OF LATIN. 69 
 
 sound of c. But the practice of all modern Europe, except 
 among the Greeks, has agreed to reject the dogma of 
 Quintilian except so far as it applies to c before a, o, 
 and u. In that situation all sound it like k ; but before 
 e and i, m, eu, m, and y, a hiss is substituted, varying in 
 its character according to the people who employ it. 
 
 Thus the Latin words cicercula, cssdes 
 are sounded 
 
 by Germans and Slavonians, tsitscrkula, tsaedes, 
 " Italians, chicherkula, chaedes, (?) 
 
 " French and English, siserkula, ssedes. 
 
 Of all these sounds the most defensible is that of ts, 
 and those who prefer for reasons of convenience to accept 
 the authority of the late Latin can satisfy themselves of 
 a change during the late Empire at least of ci, followed 
 by a vowel, into a hissing sound, a sound which as early 
 as the sixth century became established for the situation 
 indicated. But even these persons can find no reason 
 beyond the analogies of their own tongue (which it is 
 frivolous to adduce without ancient support) for a hissing 
 sound of c before single i, or before e, m, ce, eu, y. Our 
 design, however, is now to return to the higher ground 
 occupied by Quintilian, and to offer some further reasons 
 in defence of his declaration that c in the classical age 
 was h everyiohere. It will be convenient to proceed 
 according to the objections raised by those who defend 
 the hissing sounds. 
 
 Those objections are : 
 
 1. " There is an inherent absurdity in supposing the 
 Eomans to have pronounced Kilcero, Jcesso, pahis for 
 
70 PRONUNCIATION OF LATIN. 
 
 Cicero, cesso, pads." No-w every student of language 
 knows how hazardous it is to assume impossibilities of 
 sound ; and it is further quite plain to every one that the 
 only reason (!) which underlies such a statement as the 
 above is (to omit other languages) that the English says 
 Sisero, sesso, pasis. 
 
 Weak as is such reasoning, it has stood, and still stands, 
 more in the way of fair argument on this subject than 
 any other obstacle. It may be worth while to confront 
 it with a counter objection of the same form but more 
 valid. Shall it be thought that the Komans could not 
 have said Icilcero, etc., and yet that they could have said 
 esca, essm ; * bucca, bucsce ; Perdiccas, Perdicsw j Marcus, 
 Marsej jloccus, flocsij pax (paks), pasis; docui, doserej 
 cadere, sesidi ; dico, disis j dos{i)tum and docfum ; au- 
 das(i)ter and audacier ? jK7(;ero has against it the anal- 
 ogies of modern European languages ; while essce, disis, 
 if so sounded, would obliterate the most important 
 element of the stem, in violation of general probability, 
 and of a distinctly felt desire with the Latins to keep the 
 same stem throughout the inflected forms of words. 
 The ofEence to Latin ears of such fluctuations in the 
 stem of declined words may be gathered from the words 
 of Cicero (Orat. 48, § 160), where he speaks of the absurdity 
 of such a change, and is to be inferred also from the face 
 of the language : so that we may be quite sure of a strong 
 tendency to say doheo, dokes, etc., with doJcui and doktum, 
 or if doseo, etc., then also dosui, etc. But doJcui is undis- 
 puted, and hence a better reason than the shallow 
 objection above stated must be brought to defend a 
 sound like ihseo. Accordingly, great pains, have been 
 
 * The same will apply to estaa, etc. 
 
PEONUKCIATION- OF LATIN. 71 
 
 taken to furnish a reason from the Latin orthography ; 
 with what success remains to be inquired. But be- 
 fore we leave the consideration of general probabilities 
 we would ask : if c in excisus is equal to 5, why do we not 
 sometimes find exisus written by neglect, as we often do 
 exatiare for exsatiare, exul for exsid, etc., etc. ? * 
 Once more ; on reading the fragment of Titinius : 
 
 Eeliqui acus aciasque ero atque erse nostrse . . . . , 
 
 and this of Atilius : 
 
 Cape caede, Lyde, come conde . . . . , f 
 
 is not the design of "atmominatio" sufficiently evident 
 to make it incredible that akus and asias, Kape and 
 Smde, or {tsade, etc.) could have been the sound of those 
 adjacent words ? We have seen already that qu involved 
 the h sound. When, therefore, Nonius (6, 17) says, 
 "ilUcere est proprie illaqueare," we can see how he 
 should say this of ilUkere, but can scarcely believe it 
 would ever have been said of illisere {illitsere, etc.). 
 
 2. There is an allegation of evidence in the orthography 
 of the Latin for a hissing sound of c ; but the argument 
 drawn from this source will apply at best only to c when 
 followed by i succeeded by another rowel. The fluctua- 
 tion, it is said, b'etween the forms -cio and -tio, -cius and 
 -iius, -ciicm and -tium in the end of majjiy Latin words is 
 a proof of a similarity of sound between ci and ti fol- 
 lowed by a vowel, which can only be understood by 
 supposing both the c and the t when so placed to have 
 
 * This particular objection applies only to the English and 
 French pronunciation of e. 
 t Eibbeck, Com. Lat. Frag., pp. 27 and 115. 
 
ra PEOKUNCIATION OF LATIN. 
 
 degenerated into that hissing sound like ts or Italian zz, 
 which may be easily shown to have been used for each 
 of them in the post-classical Latin. Now before such a 
 sweeping conclusion can be allowed, it is necessary to 
 establish the fact of the alleged fluctuation not only, but 
 also that it prevailed during the classical period in such a 
 sense that it was indiffierent for the sound in the case of 
 a given word whether it was written with ti or ci. But 
 recent investigations of inscriptions made with this view, 
 show, according to the testimony of Hiibner and Corssen, 
 that the fluctuation in question was extremely rare until 
 after the early Empire. Almost the only authentic in- 
 stances are said to be the proper names : 
 
 Mucins, 
 
 Mutius, 
 
 Lacia, 
 
 Latia, 
 
 Marcius, 
 
 Martins, 
 
 Aecius, 
 
 Attius, 
 
 Volcacius, 
 
 Volcatius, and others. 
 
 There are, now, two ways of explaining this variety of 
 form: one by assuming a common sound for Marcius and 
 Martius — say Martsius — the other by supposing a dia- 
 lectic variety in the sounding of those names, so that 
 some said MarTcius while at the same time others spoke 
 Martius. The first explanation seems hardly credible in 
 view of the entire silence of Grammarians concerning 
 any such sound for c ; while the second does not demand 
 remark from them, since Markius and Martius were in 
 fact two*" names for the same individual, just as Atus 
 (or Attus) Clausus, in Sabine, and Appius Claudius in 
 
 * See Appendix, Note G. 
 
PKONUNOIATIOlir OF LATIK. ■'5'3 
 
 Roman style. (See Liv. II, 16.) And so when Plutarch 
 wrote Mapnof, he preferred to adopt the form with t, and 
 when he wrote Uopicia he chose the k sound heard in 
 Forcia. 
 
 This c, then, as well as c everywhere in Latin, is ren- 
 dered uniformly by & by Greek authors contemporary 
 and nearly contemporary with the classical period : thus 
 not only Karwv, but Ktsepuv and Kataap, KeXaog, Aiiuog, 
 JlXanevTia, etc. Now it is plain from such forms as 
 OvTj'ioi, Ko'ivTog and Kvivrog, 'EXPrj-ia and 'EAovrjrla, 
 AevrXog, etc., etc., that the Greeks tried to render as 
 nearly as possible the sounds of proper names which they 
 heard in the mouths of the Eomans. Their imitation 
 was not hampered by technicalities or traditions of or- 
 thography ; and was limited only by the powers of their 
 alphabet. 
 
 And so if Plutarch, in the early part of the second 
 century after Christ^ had heard the Eomans say Martsius, 
 Portsia, there is every reason to believe that he would 
 have written them JAdpr^iog, or Mapraiog, Jlopraia : just 
 as the modern Greeks commenced (at least as early as the 
 sixth century) now and then to write Bever^ia for Venetia, 
 T^iPiTd, Nd/Ja for Civita Nova.* 
 
 Further affirmative evidence for the hard sound of c 
 may be dispensed with, if, as is hoped, sufficient offset 
 has been found for the alleged objections to Quintilian's 
 general statement. Scholars, it may be added, are now 
 commonly agreed that that statement was correct, and 
 that in the classical Latin, C was equal to K. 
 
 * Sophocles' Modem Greek Grammar, p. 18. 
 
 4 
 
74 PKONUSrCIATIOif OF LATIlf. 
 
 Do 
 
 A remark of Quintilian, that " D was akin to T,'' is 
 in conformity with the accepted opinion, that the' Latin 
 D represented one of the members of the lingual scale, 
 t, d, til (as in English the). From a more particular 
 description by Martianus Capella, which distinguishes 
 between the t and the d, we have the further information 
 which we need ; showing that d was the media of the 
 above series, and equivalent, therefore, to the English d. 
 " D, appulsu linguae circa superiores dentes, innascitur. 
 
 T, appulsu linguae dentibus impulsis, excutitur." 
 
 Similar statements of other Grammarians are not want- 
 ing ; but in the absence of objection it does not seem 
 necessary to produce either those statements or other 
 corroborative evidence of the sound just indicated. 
 
 It remains only to be said with regard to the practice 
 of the classical age, that d in the end" of words frequently, 
 perhaps generally, was sounded hard like t. For not 
 only in the MSS. of older authors * do we find a2mt, set, 
 Tiaut, but these forms are said to appear also on inscrip- 
 tions of the late Kepublic, as well as afterward, along with 
 the other spelling apud, sed, liaud. "Well authenticated 
 instances of the same kind may be found in aliut, quit 
 (for quid), illut, quitquit, it (for id), and others. 
 
 Before the classical age, the confusion of d and t seems 
 to have been more general. Before Cato's time " {mare) 
 
 • * See Ribbeck's collection of Fragments, and the editions of 
 Flautus. 
 
PEOKUKCIATION OF LATIK; 75 
 
 Hadriaticum " had been called Atriaticum. And Quin- 
 tilian, I, c. 4, § 16 : " Quid D litterae cum T qusedam 
 cognatio ? Quare minus mirum, si in vetustis operibus 
 urbis nostrae et celebribus templis legantur Alexanter et 
 Cassantra." 
 
 This language clearly implies that in the classical age 
 a distinct separation had been efiected between the two 
 letters, at least generally. The tendency to the hard 
 sound was now confined to the end of words; but that it 
 still existed in practice for that situation (like in Ger- 
 man Kind, Brod, etc.) is shown by the above examples and 
 by such remarks as those of Quintilian (I, 7, 5) and of 
 Charisius (P. p. 87): ' Haud' similiter d litera termiua- 
 tnr ; dvde enim, grseca vox, d litera terminari ccepit, .... 
 sed et per t scribi, sonus Tocis admittit. 
 
 • 
 
 D, therefore, not in the end of a word, as in dodrans, 
 was sounded like English d j in the end of words, more 
 like English f. 
 
 in. 
 
 The nature and actual power as well as the history of 
 Latin P have been plentifully announced and discussed 
 by the ancient Grammarians. They, not omitting to 
 quote certain remarks of Varro, Cicero (Orat. c. 48), and 
 Quintilian (I, c. 4), have expressed so decidedly the dis- 
 tinction to be observed between the sound of F and that 
 of the consonants akin to it, and have even pointed out 
 so plainly the oflGice of the organs engaged in rendering 
 
76 PRONUNCIATION OF LATIN 
 
 the proper sound in question, that we cannot attain our 
 practical end in a better way than by simply making a 
 few quotations. In these it will be seen that for its 
 origin F is usually traced to a sound either like that of 
 the vanished ^olic Digamma (the character for which is 
 supposed to be preseryed in that of P *), or like that of 
 Greek ; which ((/>), it is important to observe, is every- 
 where in these discussions contemplated as- having, not 
 the sound of English F, but oip followed (in union) by 7*, 
 as would be heard in a close union of p and h in the 
 English words TJpham, soii,pJiouse \ (leaving away U- and 
 sou-). 
 
 Further, from the separation made between the sound 
 of /, and that of the ^olic Digamma, as well as that of 
 (j>, of I, and of h, we have left only the sound of our own 
 F, which is, moreover, indicated by the descriptions of 
 the manner of utterance. 
 
 1. F was not equal to Digamma, or Latin V {vau). 
 
 Mar. Victorinus, P. p. 3468 : Scripseram autem vobis 
 . . . . F X — ^olis duntaxat, idem valere, quod apud 
 nos V, cum pro consonanti scribitur, vocarique fiav 
 et Digamma. 
 
 [That Latin F once had this sound of later V (uni- 
 versally taken by Grammarians as the classical repre- 
 sentative of Digamma) is affirmed by Cornutus 
 
 * Agnaeus Cornutus, P. p. 3383 : Est quEedam litera in P literse 
 speciem figurata, quae Digamma nominatur, quae duos apices ex 
 Qamma litera habere videtur. 
 
 f Except among the ^olians, where (j> is supposed equal to/. 
 
 :{ This is, of course, not meant for the sign of Latin F, but fop 
 Digamnuv. 
 
EBOIf DlfCIATION OF LATIN. 77 
 
 (Pi -3282)" on the eTidence of Fotum, Firgo; and 
 others, found by him in certain ancient books for 
 votum, virgo, etc.] 
 
 2. F was not equal to <p, or its Latin equivalent (in 
 Cicero's time) ph. 
 
 When Cicero (Orat. c. 48) says of Ennius' substitute 
 for Greek (/>: "Vi patefecerunt Bbugbs* non Phrt- 
 Giis ; ipsius antiqui declarant libri. Nee enim 
 Grmcarn liter am adhitebant: nunc autem etiam duas" 
 \i. e., ph] : he means evidently that Ennius, if he 
 had not inaccurately spoken with B, would have 
 been forced to use / as the nearest then available 
 equivalent for <p, since ^ itself could not be employed, 
 and the more exact mode of rendering by means of 
 ph had not in Ennius' day come into use. 
 
 Priscian speaks very satisfactorily to the same points 
 (P. p. 542) : F, u^olicum Digamma, quod apud anti- 
 quissimos Latinorum eandem vim, quam apud ^oles 
 hdbuit <p ; eum autem frope sonum, quern nunc 
 hcibet f significahat p cum aspiratione, sicut etiam 
 apud veteres Grcecos pro (p, p et it. Unde nunc 
 quoquc in Grcecis noridnibus antiquum scripturam 
 servamus, pro 4>, p et h ponentes, ut : Orpsetts, 
 Phaeton. Postea vero in latinis [with a correspond- 
 ent in Greek] placuit verbis pro p et jt, Fscribi, ut : 
 FA3IA . . . . ; loco autem, Digamma, V pro conso- 
 nante, quod cognatione soni videbatur affinis esse 
 Digamma ea litera. Thus, indisputably, was the 
 
 * The editions previous to Emesti's give the reading Fruges. 
 
78 PKONUNCIATION OF LATIN". 
 
 sound of /, as in fanum, a medium between v in 
 vanum and- </> (with the rr audible) in (pauov, and 
 equal, therefore, to /in English Fan. 
 
 With this agree now also the prescriptions for its 
 enunciation. 
 
 Priscian (P. p. 543) : Hog tamen scire debemus, quod 
 non tamfixis Idbris est pronuntianda f, quomodo pli r 
 atque hoc solum interest inter f et ph. 
 
 Terent. Maurus (P. p. 2388) : 
 
 Imum superis dentiius adprimens labellum, 
 
 Sp'iramine lent 
 
 Hanc ore sonabis 
 
 To the expression " spiramine leni," if it is to be under- 
 stood strictly of the " spiritus leuis," we must oppose the 
 language of Quintilian (XII, 10, 38 and 39), where it is 
 plain that a very rough breathing is contemplated. 
 Admitting the rough breathing, this description coin- 
 cides with that given by Max Miiller (3d Ser. Lee. p. 146) 
 for the English/. 
 
 " A sixth barrier is formed by bringing the. lower lip 
 against the upper teeth. This modifies the spiritus asper 
 to /." 
 
 a. 
 
 E'o one doubts, and there is no room for doubting, that 
 this letter was in full play as a familiar element of Latin 
 words during the classical period. It is therefore uh- 
 
PKONUNCIATION' OF LATIN. 79 
 
 necessary for us here to set forth, or to espouse either side 
 of the historical question concerning the time when its 
 sign was introduced into the Latin alphabet. The state- 
 ment made by Plutarch, and often repeated after him, 
 that Sp. Carvilius Ruga (A. U. 0. circ. 520) first gave the 
 sign G a place among the Latin letters (assigning it the 
 place of the disused z), has been much questioned as to 
 its accuracy, in later times — and with reason. But quite 
 independently of the merits of that statement as a piece 
 of history, the mere fact that it was made involves two 
 inferences of importance to us. 
 
 It shows, first, that scholars knew of a time when a 
 separate sign for G had not been thought necessary ; and, 
 secondly, that they knew that that time had long ceased 
 to exist, and that, before the classical period, a distinct 
 mark for the G sound had been recognized as a neces- 
 sity. 
 
 Such inferences are supported by more certain evi- 
 dence : which will show, first, that the sounds of C and 
 G were never far asunder ; second, that in the classical 
 age they tvere distinct ; and, third, that the difference lay 
 in this : that in the scale of k mutes {k, g, %) C was the 
 tenuis and G the media. 
 
 1. Lachmann (Lucr. p. 148) reports from the Cd. Lei- 
 densis of Lucretius (called by him oMongtis), the readings 
 centis for gentis, vacari for vagari, conspercunt for con- 
 spergunt, grecis for gregis, lucebat for lugebat, etc., etc. 
 
 Such spelling may, indeed be not a faithful copy of the 
 archetype, but only a mark of inaccuracy in the German 
 copyist of the ninth century. The same doubt, however, 
 cannot attach to the statement of Marius Victorinus 
 (P. p. 2459), that the ancients put C for G in " CaMno 
 
80 PKOlirUNCIATIOlf OF LATIN. 
 
 for Gabino, lece for lege, acna for agna." Allowing that a 
 separate sign and a separate sound of G existed in those 
 early times, this confusion in orthography shows all that 
 we want now, namely that the sound of G and C must 
 have lain yery near together. And the same is further 
 to be seen in the perpetuation of for G (or G for C) in 
 forms which continued always, to retain their old spell- 
 ing, even after the establishment of clear distinction 
 between the two closely related consonants. The reason- 
 ing here need not be long. Maximus Victorinus (P. p. 
 1945) says, .... d.uplex est x : constat enim aut ex g et 
 s Uteris, aut ex c et s, ut puta rex, regis; pix, picis. 
 Quippe ante x literam, quse postea in compendium inventa 
 est, rex per gs \regs\, item pix per cs \^pics\ veteres 
 scribebant. The g sound, therefore, when combined with 
 subsequent s was so nearly identical with the c sound 
 under the same circumstances that it was found con- 
 venient to represent both groups by the same sign x. 
 
 The interchange of G and C, both regularly and acci- 
 dentally, may be pointed out in a great many other 
 combinations also : thus agere, actum, actor ; lugere,'luc- 
 tus, etc., etc.; negotium for necotium, gurgulio and cur- 
 culio, co7igordia and concordia, Progne for UpoKvrj, cygnus 
 as well as cycnus for iivievog. These examples, which 
 might be multiplied, are not arranged with respect to 
 the time of their origin, since there is s^o design of show- 
 ing here the steps in the history of the alliance between 
 g and c. Having shown that the two sounds were always 
 very nearly related, it is now to be shdwn : 
 
 2. That in the classical pronunciation there was, never- 
 theless, a clear distinction between them. It is plain 
 that this may be fairly inferred from some of the very 
 
PKONUNGIATION GF LATIN. 81 
 
 instances whicH we haye; cited for showing the approxi- 
 mation of the sound. For if it is true that lece once 
 serTcd sufiBciently to represent what was afterwards heard 
 as lege, it must be equally true that for the once fully 
 established lege, lece would no lori,ger serve. 
 
 Cicero (de Nat. Deor. II, 36, § 67) says : Mater autem 
 est a gerendis fmgibus Ceres tanquam Geres : casuque 
 prima litera itidem immutata, ut a Greecis : nam ab illis 
 quoque At/jujJtt^p quasi Tij iii]rrip nominata est. And 
 Quintilian writes precisely to our point in the words 
 (1,5,12): Nam duos in uno nomine faciebat barbarisr 
 mos Tinga Placentinus (si reprehendenti Hortensio credi- 
 mus) preculam pro pergula dicens, et immutatione, cum 
 c pro g uteretur, et transmutatione, cum r praeponeret 
 antecedenti. 
 
 Gr, therefore, was clearly different from the kindred C 
 — ^but in what ? The answer to this question will com- 
 plete our principal inquiry, and can be sufBciently found 
 in the words of Priscian and Mar. Victorinus. 
 
 ■ Priscian (P. p. 549) : Inter C sine aspiratione, et cum 
 aspiratione est G. 
 
 Mar. Victorinus (p. 3454) : C etiam et G, ut supra, 
 sono proximae, oris molimine nisuque dissentiunt; 
 nam C reducta introrsum lingua hinc atque hinc 
 inter molares surgens, hsrentem intra- os sonum 
 Tocis excludit;. G vim prioris pari linguae lapsu 
 palate suggerens, lenius reddit. 
 
 Thus is the Latin G seen to be the media of the 
 i-mutes, or our own hard G. 
 
 The soft sound of G so familiar to our practice when 
 
82 PKONUNCIATIOK OF LATIN. 
 
 this consonant occurs before e and i sounds (Italian 
 Giorgio; English George), could not have been known 
 to the classical Latin. 
 
 Evidences of so important a variation of sound could 
 not fail, if the variation had existed ; but no such evi- 
 dence is visible until a very late period — so late as to be 
 beyond the scope of our inquiry. As to any such double 
 power of the letter G, ancient Grammarians in all that 
 they have to say about that letter are entirely silent ; and 
 in maintenance of the uniform hard sound we might 
 repeat a great part of the considerations already presented 
 in connection with the discussed softening of the tenuis 0. 
 Such repetition will be needless, and we will call atten- 
 tion only to one passage of Aulus Gellius (IV, c. 9) : 
 " Masurius Sabinus . . . . ' Beligiosum,' inquit, est ... . 
 verhum a ' relinquendo ' dictum, tamquam ' cmrimonim' 
 a * carendo.' " 
 
 Such supposed derivations, founded, as they are, on a 
 similarity of sound, do not allow us to suppose that we 
 should here contemplate harere and seritnonia, or relin- 
 kuere and relidjiosum. Therefore, 
 
 Latin G = English G in gag. 
 
 The sign H in the Latin alphabet represented that 
 efEort of speech known both to us and to the Latin 
 Grammarians as the rough breathing. 
 
PEONUNCIAXION OF LATIN. 83' 
 
 This conclusion is indicated unmistakably by the 
 history of the character as employed in Latin (its origin 
 being identical with that of the G-reek Upoacftdia dabela, 
 or spiritus asper), and by what has been said about H by 
 the Latins, Grammarians, and others, as compared with 
 what we know of the nature of our own H. 
 
 As to the first point, Priscian (P. p. 1345) says: H, 
 eiia-6v. Vetustissimi enim quique Grseci pro aspiratione, 
 H scribebant, quam habebant HEOATON in principio. 
 And ]y[ar. Victorinus (P. p. 2459) : Grseci, ....■&, (p, %> 
 priusquam a Simonide'invenirentur, exprimebant iuxta 
 T, et iuxta 77 et k, aspirationis notam H ponendo.* 
 
 As to the second point ; the nature of our H, especially 
 in view of the physiology of its iitterance, gives rise to 
 the question whether in strictness it deserves to be called 
 a letter. But precisely the same question was debated by 
 the Latin Grammarians with respect to their H. We 
 know of ours that it is made by a rush of air driven 
 through the open glottis f by the contraction of the 
 muscles of the abdomen. Of theirs it was said by Mar. 
 Victorinus (P. p. 2455): H — profundo spiritu, auhelis 
 faucibus, exploso ore fundetur. 
 
 Through a deficiency in Victorinus' observation these 
 two descriptions are not quite coextensive, but they lie 
 together, and point in the same direction. 
 
 When Donatus says (P. p. 1737) : H interdum con- 
 sonans, interdum adspirationis creditur nota. He refers 
 not to more than one distinct function of H, but to a 
 difference of opinion among Grammarians, as to whether 
 
 * See Appendix, Note H. 
 
 f Vid. M. Miiller, Science of Language, 3d Series, p. 139, £ 
 
84 PEONUNCIATIOB^ or LATIN. 
 
 H ought to be regarded and called a consonant, or not. 
 And among the Grammarians who maintain respectively 
 the two sides of this dispute, all admit, on the one hand, 
 that H represents an aspiration ; and none claim more, 
 on the other hand, than that this aspiration effects so dis- 
 tinct and indiyidual a modification of the adjacent vowel 
 sound, that it deserves a regular place among the con- 
 sonants. Thus Velius Longus (P. p. 2218) : Accedit 
 huic, q.uod, si accedens literse esset, cum ipsa litera enun- 
 tiaretur ; nunc vero et ante Vocalem, et post Vjcalem 
 Bonat sic, quomodo est alia syllaba ca, et alia ac, et alia 
 la, alia ab, sic alia ha, alia ah. Nam quod ex H quoqiie 
 exiStimant quidam coUigi posse consonantem et adsigni- 
 flcantem, apparet ex eo, quod aut accedens, aut recedens 
 immutat significationem ; siquidem aliud est hira, aliud 
 dira canam, exemplum. 
 The difEerence, therefore, between 
 
 a and ha, 
 
 dbeo " habeo, 
 
 iste 
 
 (( 
 
 histrio, 
 
 ortus 
 
 a 
 
 hortus. 
 
 aut 
 
 ce 
 
 haud, 
 
 ordo 
 
 es 
 
 hordeum. 
 
 omen 
 
 u 
 
 homo. 
 
 oro 
 
 te 
 
 horror, etc., etc., 
 
 was just this, that the initial vowels of the latter forms- 
 were preceded (or accompanied) by what was on all 
 hands acknowledged to be an aspiration — this aspiration 
 being regarded by some as merely an accident of the 
 vowel, that is to say, an extraordinary impulse imparted 
 to the column of air, whose intonation makes the vowel 
 
PEONtrNGIATION OF LATIN". 85; 
 
 sound; while others considered it an element of so pre-: 
 cise and separable a nature as to be justly entitled to the 
 name of consonant. In all this we have a description 
 well nigh perfect and complete of our own H. 
 
 The Latin H was therefore the same with the Eng- 
 lish H. 
 
 It now remams to inquire how far the integrity of the 
 H was respected by the practice of the Latin tongue. 
 We know what has been the fluctuating history of the 
 aspirate consonant in modern languages; how, through 
 eVery stage of attenuation, it has finally vanished from 
 words in which it once constituted a tegular (often an 
 organic) part ; and how accident and caprice has often 
 secured it a more or less temporary, or a conceded and 
 permanent place, to which there was not the faintest 
 historical title. 
 
 An this happened also to the Latin H. 
 
 And it would be possible, if necessary, to collect many 
 particulars in support of Quintilian's statement, which 
 follows (Lib. I, c. 5, § 19, sq.) : H . . . . cuius quidem 
 ratio mutata cum temporibus est ssepius. Parcissime ea 
 Teteres usi etiam in vocalibus, cum cedos ircosque dice- 
 bant, diu deinde serratum, ne consonantibus aspirarent, 
 ut in Graccis et in . triumpis ; erupit brevi tempore 
 nimius usus, ut choronm, chenturioiies, prcechones adhuc 
 quibusdam inscriptionibus maneant, qua de re CatuUi 
 nobile epi^amma est.* Inde durat ad nos usque 
 
 * We must make room for this epigram (LXXXIV) here : 
 
 Chommo&a dicebat, si quando eommoda vellet 
 Dicere, et hirmdias Arrius insidias. 
 
86 PEOKUNOIATION OF LATIN. 
 
 menter et compreliendere et mihi, nam melie quoqne pro 
 me apud antiques tragcediarum prsecipue scriptores in 
 yeteribus libris invenimus. 
 
 If, now, for our practical purpose, an attempt be made 
 to ascertain the particular instances in which, during the 
 classical period, the H was actually heard in its proper 
 force (and therefore also properly written), we shall find 
 so much conflict in the evidence which remains to us, as 
 to leave us no hope of satisfaction. Corssen's* interroga- 
 tion (Vol. I, p. 50, S5'.) of inscriptions, and the most 
 ancient MSS., hring to light the contemporaneous exist- 
 ence of (among others) the following forms : 
 
 harundo, 
 
 arundo. 
 
 haruspex. 
 
 aruspex. 
 
 harena, 
 
 arena, 
 
 heres. 
 
 eres. 
 
 hordeum. 
 
 ordeum. 
 
 Hammon, 
 
 Ammon, 
 
 Hiberns, 
 
 Iberus, 
 
 vehemens, 
 
 vemens. 
 
 prehendo. 
 
 prendo, etc., etc. 
 
 Et turn mirifice sperabat se esse locutum, 
 Quum, quantum poterat, dixerat hinsidiois. 
 Credo sic mater, sic liber avunculus eius, 
 Sic maternus avus dixerit, atque avia. 
 Hoc misso in Syriam, requierant omnibus aures, 
 Audibant eadem haec leniter et leviter. 
 Nee sibi postilla metuebant talia verba ; 
 Quum Bubito adfertur nuntius horribilis : 
 lonios fluctus, postquam illuc Arrius isset, 
 lam non lonios esse, sed Bionios. 
 
PBON-UNCIATION OF LATIN. 87 
 
 Grammai'ians report for the early classical Latin : 
 
 mihi 
 
 and 
 
 mi, 
 
 nihil 
 
 ti 
 
 nil, 
 
 cohors 
 
 u 
 
 coors, cors, and others.* 
 
 In MSS. of later date the increasing tendency to drop 
 the 7i must be ascribed, no doubt, in large degree, to the 
 operation of that softening tendency which resulted in 
 the final banishment of the sinritus^ asper, as well from 
 the Italian as from the Greek languages, and to an 
 ignorance among copyists of the proper Latin sound. 
 
 The Cd. " quadratus " of Lucretius, which, according 
 to Lachmann, seems to be the work of a German copyist 
 in the sixteenth century, has the forms : 
 
 aborret, orror, orrifico, orrida, proibere, distraitur, is 
 for Ms, etc. 
 
 The improper addition of H seems to have been a mark 
 of antiquity or rusticity, if we judge by the remarks of 
 Quintilian and Catullus given above, and the language 
 of Gellius (II, 3), which follows : 
 
 jETliteram sive illam spiritum magis quam literam dici 
 oportet, inserebant earn veteres nostri plerisque vocibus 
 verborum firmandis rohorandisque, ut sonus earum esset 
 
 viridior vegetiorque Sic ' lachrimas,' sic 'sepul- 
 
 chrtim,' sic 'ahenum,' sic 'vehemens,' sic ' incohare,' sic 
 'helluari,' sic ' halucinari' sic 'honera,' sic ' honustum,' 
 dixerunt. In GelHus' time, therefore, it must have been 
 considered proper to say not only onera, onustum, lacri- 
 
 * See Appendix, Note I. 
 
88 PRONUNCIATION OP LATIN. 
 
 mas, inchoare (pronounced incoare), etc., but also vemens, 
 alucinari, and in the same way probably a number of 
 other words which are more or less frequently still 
 written with the h in the M^S. which have been handed 
 down to us. 
 
 Since, then, in the matter of adding or withholding 
 the H, the practice of the Romans themselyes was never 
 very determined and uniform, our practice can only con- 
 form itself to some authority, or when this fails, to the 
 probability furnished in the best attested forms of or- 
 thography. For authority (such as it is), we have some 
 lists given by the Grammarian Phocas (De Aspiratione, ap. 
 P. p. 1732, sq. Lind., I, p. 353, sq.), too long to be given 
 here. For the rest, wherever H is properly written in a 
 Latin word, there is no choice left us but to render it, 
 and that with the force of English H.* 
 
 K. 
 
 The sign K had fallen into general disuse in the clas- 
 sical period ; and, but for the very few scanty remains of 
 its employment, which will be mentioned, we should have 
 nothing to say of that letter here. 
 
 Quint. Inst. Orat., I, vii, 10 : Nam K quidem in nullis 
 verbis utendum puto, nisi quae significat, etiam ut sola 
 ponatur. It was the opinion of some Grammarians that 
 the K should have a somewhat more extended use — that- 
 
 * Of the combinations ph, th, eh, rh, we shall come to speak 
 hereafter. 
 
PEONUlfCIATIOBr OF LATIK. 89 
 
 it should be put, namely (instead of C), in all places 
 where its sound was followed by the vowel a. This state- 
 ment Quintilian adds to the above : Hoc eo non omisi, 
 quod quidam earn, quotiens A sequatnr, neeessariam 
 credunt, cum sit littera, quse ad omnes vocales vim 
 suam perferat. 
 
 This dogma of orthography was often repeated by 
 later Grammarians, who founded it on the reason that K 
 already contained A in the sound of its name — Ka, while 
 the C was sounded Cfe, that is, Ke (see Scaurus de 
 -Orthog., ap. P, p. 2353). It was sometimes further con- 
 tended (as by Scaurus, 1. 1.) that it was the C, and not the 
 K, which should be universally surrendered. And all 
 this implies, of course, that which is also directly stated, 
 that the consonant K was equal to the consonant C* 
 
 With this our inquiry is at an 'end. It may be added, 
 however, upon the testimony of the best MSS., and upon 
 that of inscriptions too, according to Corssen's report, 
 that in practice the K was banished more completely 
 even than was sanctioned by Quintilian : who himself, 
 moreover, seems in I, 4, 9 to admit the fact that in his 
 time this sign was only used, as we see it now — in certain 
 few abbreviations — K, quae et ipsa quondam nominum 
 nota est. [As K. for Kcbso, Karthago, EalendcB-l 
 
 * Prisdan L. I. (ap. P. p. 543) : K Bupervacua est, ut supra 
 diximus, quse quamvis scribatur, nullam aliam vim habet, quam C. 
 
90 PRONTTN CIATIOK- OF LATIK. 
 
 The nature of L is thus described by Mar. Victoriuus 
 (P. p. 3455) : L, quae validum nescio quid, per partem 
 palati, qua primordium dentibus superis est, lingua 
 trudente, diducto ore, personabit. Add the further cir- 
 cumstance, which is recognized by all the Latin Gram- 
 marians, that L is a liquid or sem.i-vowel, or, in other 
 words, intonable, and we have here a full account of 
 the English L; which is given by the author of the 
 Principles of Pronunciation, in Webster's Dictionary, in 
 these words: "This letter has only one sound, which 
 consists of an efflux of vocalized breath, or voice, over the 
 sides of the tongue, while its tip is pressed against the 
 gums of the upper front teeth." Latin L is therefore the 
 same with English L. 
 
 Corroboration for this may be found, if needed, in a 
 comparison of Latin L with Greek A ; and further, in 
 observation of the affinities and modifications to which 
 this letter was subject in the history and growth of Latin 
 words, and the practice of Latin speech, from which it 
 vnll appear, as shown by Bopp (in § 30 of his Compara- 
 tive Gr., V. 1, and elsewhere), that L appears in Latin 
 imder the same conditions and in the same relations as 
 those in which the same letter is also employed in Greek, 
 Gothic, Sclavonic, and the other languages of the same 
 stock. Here should be consulted also Corssen, Max 
 Miiller, and Pott (Etymol. Forsch. II, 97), whom Corssen 
 and Miiller both quote. But we need not expand our view 
 so far. Since there is no objection to the point we have 
 
PRONUNCIATION OF LATIN. 91 
 
 to make, we are fully warranted in accepting the undis- 
 puted tradition of the power of Latin L, supported by 
 the statement of Victoriaus reported above. 
 
 In respect to variation or modification in the rendering 
 of L we have been able to find no intimation of any- 
 thing more than a slight difference in the force or vigor 
 of its utterance in different situations. On this subject 
 there is a statement of Pliny, quoted by Oorssen from 
 Priscian, as follows : L triplicem, ut Plinio videtur, sonum 
 habet : exilem, quando geminatur secundo loco posita, 
 ut ' ille, Metellus,' plenum, quando finit nomina vel syl- 
 labas, et quando aliquam habet ante se eadem syllaba 
 consonantem, ut 'sol, silva, flavus, clarus,' medium in 
 aliis ut 'lectum, lectus.' 
 
 Thus, the second of two L's put between vowels was 
 sliglit ; as must, indeed, have been necessarily the case 
 if both were sounded alike. For the same must have 
 been true of other consonants likewise, as, e. g., of the 
 second s in caussa, aussus, fussus, odiossus, mentioned 
 by Victorinus (p. 2456) as the earlier, more pronounced 
 mode of rendering (and writing) causa, etc. Greatest in 
 strength or weight was the L at the end of a word (no 
 doubt also at the end of a close syllable), and after an- 
 other consonant at the beginning of a syllable; and 
 neither particularly strong nor particularly weak, the L 
 in all other situations — ^in all but strength, however, 
 always simple L.* 
 
 * See Appendix, Note K. 
 
9a PEON UN CI All ON OF LATIK. 
 
 ■ M.. 
 
 It is the tmiyersally accepted tradition that the M of 
 the Latin was the same with that of the cognate lan- 
 guages — the same, therefore, with our own. The tradition 
 is suflBciently supported by these words of Mar. Victo- 
 rinus (ap. P. p. 2455) : At M, impressis invicem labiis, 
 mugitum quendam intra oris specum, attractis naribns, 
 dabit. Compare also Qnintilian, XII, 10, § 31 ; and VIII, 
 3,45. 
 
 There were some situations, however, in which the 
 practice of iie Eoman speech declined to render M with 
 its normal force and expression ; and other situations in 
 which the sound of M was not only obliterated but 
 supplanted by another sound under the assimilating 
 influence of a following letter. Without going minutely 
 into the physiology of M, it will be seen at once that 
 when the labial organs are in the proper position for the 
 production of this letter, the degree of distinctness with 
 which it may be rendered, will still depend upon another 
 condition, namely, the degree to which the necessary 
 column of breath, without obstruction by the tongue, is 
 allowed access to the hollow of the mouth (s^ecws), where 
 it is to reverberate (mugire). The lowest degree of dis- 
 tinctness (next to absolute inarticulation) is that which 
 is heard when the breath is not only excluded from the 
 cavity of the mouth but is debarred a free passage through 
 the nose, but, being restricted behind that organ, only 
 reverberates there. This is the French nasal M. One 
 degree higher, perhaps, may be set the sound supposed 
 
PEONUNOIATION OF LATIK. 93 
 
 tob'e represented with us by the m in the written form 
 Hm ! , for here the breath, so long as let on at all, is 
 given free passage throiTgh the nose. Between this 
 minimum approach to the full sound of m, and the com- 
 plete rendering of the same, many degrees of approxima- 
 tion are possible, degrees, however, which we have no 
 means of recording. It remaiiTs now to be said that 
 somewhere within the scope of these the final M of the 
 Eomans fluctuated. While some distinctions are to be 
 known, we shall never be able to determine the distinct- 
 ness which properly belonged to final M in- all cases. 
 Discrimination is here difficult in itself, and so the 
 Eomans found it also evidently in their own practice.* 
 
 We now introduce some statements of the ancients 
 touching the value of M, and will then proceed briefly to 
 arrange and apply the same. 
 
 Quint. IX, 4, § 40. Atc[ui eadem ilia littera [M], quo- 
 tiens ultima est et vocalem verbi sequentis ita con- 
 tingit, ut in eam transire possit, etiam si scribitur, 
 tamen.parum exprimitur, ut Multum ille et Quan- 
 tum erat J adeo ut psene cuiusdam novae litterse 
 sonum reddat. Neque enim eximitur sed obscuratur 
 et tantum aliqua inter duas vocales velut nota est, 
 ne ipsEB coeant. 
 
 Priscian I, 555, (Vid. Forcel.) : M obscurum in extremi- 
 tate dietionum sonat ut 'templum,' apertum in 
 principio ut ' magnus,' mediocre in mediis ut ' um- 
 bra.' 
 
 Donat. ad Ter. Adel. II, 1, 53 : Mussitare .... dictum 
 • Bee Appendix, Note L. 
 
94 PKOKUNCIATIOK OF LATIN. 
 
 vel a muto, yel ab M: quae littera est nimium pressae 
 vocis ac pene nullius : adeo ut sola omnium quum 
 inter vocales incident, atteratur atque subsidat. 
 
 Vel. Longus (P. p. 2338) : NonnuUi synalceplias quo- 
 que observandas circa talem scriptionem existima- 
 Yerunt, sicut Verri]js Flaccus, ut ubicunque prima 
 vox m littera finiretur, sequens a vocali inciperet, M 
 non tota, sed pars illius prior (M) tantum scribere- 
 tur, ut appareat exprimi non debere. 
 
 Quint. VIII, 3, § 45 : — si cum hominibus notis loqui nos 
 dicimus, nisi hoc ipsum homitnbus medium sit, in 
 prsefanda {Ji. e., in obscoenum intellectum) videmur 
 incidere; quia ultima prioris syllabae littera, quse 
 exprimi nisi labris coeuntibus non potest, aut inter- 
 sistere nos indecentissime cogit aut continuata cum 
 insequente in naturam eius corrumpitur [thus, cun 
 noHs]. Of. Cic. Orat. 45, § 154. 
 
 These statements contain explicitly and implicitly the 
 following points : 
 
 First (generally), m in the end of words is obscure, 
 weak, uncertain, and often almost nothing ; and this is 
 shown, 
 
 .Second (particularly), by the almost complete evanish- 
 ment or subsidence of M (with the vowel preceding) at 
 the end of a word in verse,* when a vowel sound fol- 
 lows in the opening of the next word ; and. 
 
 Third (particularly) ; everywhere, in prose as well as 
 in verse, final M, in the succession of words, is assimilated 
 
 * And alao in prose, when the connection was close. 
 
PKONUXCIATIOK OF LATIN. 95 
 
 by the influence of an immediately following consonant 
 (of a certain class), provided the connection be close, as 
 it is, e. g., between the preposition and its case. (As in 
 cun notis, cun nobis ; for which, besides Quintilian, as 
 quoted above, see also Cicero, Orat. 45, 154.) 
 
 Wo must add to this third case, that the assimilation 
 here spoken of as occurring at the end of a word was 
 demanded by stronger reason, and, in fact, generally was 
 effected at the end of a syllable in the midst of a word. 
 In this situation, therefore, M was assimilated to N 
 before 0, Q, and G [concipere, congueri, congerere) ; also 
 before D and T {condere, eundem, contingere, etc.) ; and 
 before P, S, the consonants I and V, and before N. Fur- 
 ther, before vowels, and S, I, V, N", M, and GN", the M 
 was often lost. See Lachmann on Lucr., p. 136, touching 
 the forms coopertus, cocoleretur, coicere, coventionid, conu- 
 iium, comqvisse, cognomen, etc. 
 
 These analogies should lead us to conclude that whei'e 
 we find MM written as in comminus (well attested in 
 most ancient MSS. for more common cominus), the first 
 M should be very weakly rendered, i. e., no attempt made 
 to give a distinct utterance to two M's ; and so for all 
 those eases as well, where the MM is the iinchanging 
 orthography. And further, similarly, for such forms as 
 umquam, cumque, numquid, tamquam, it is necessary to 
 suppose that the M was so uncertain that many heard 
 and many spoke, as many also wrote, unquam, cunque, 
 nunquid, tanquam, etc. 
 
 Before the labials B and P within a word, the M was 
 moderately distinct, according to Priscian, e. g., " umbra" 
 of the two, less distinct probably before P, as in compo- 
 trix (written also conpotrix, etc.). [Compare here com 
 
9.G PUOlfUNCIATION OF LATIN. 
 
 in English. " composmg," rather tlian in French "com- 
 poser."] 
 
 It is plain that in all the instances just named the 
 utterance of complete, clear (apertum) M could only be 
 effected by a stoppage, highly inconyenient and offensive 
 in the midst of a word. 
 
 But further we are -told by Quintilian, that not only 
 thus in the midst of a word, but also in the rapid 
 sequence of words, this stoppaige was " unseemly " in the 
 end of a word when closely followed by a sound requiring 
 a rearrangement of the organs. The question, therefore, 
 now reverts to the sound of final M as it was heard in 
 practice. This diflScult question is by no means satisfac- 
 torily disposed of by Priscian's " M obscurum in extremi- 
 tate dictionum sonat," for there are several degrees of 
 obscurity through which it is possible for the M to range, 
 not to say q.nything about its transformation by assimila- 
 tion. There is reason for supposing a good deal of 
 iiuctuation in the Eoman practice here, and certainty in 
 particular cases is clearly no longer attainable for us. 
 The final M appears to have been less respected in 
 ancient times than it was afterwards in the classical 
 period.* The forms diee for diem, " Recipie apud Cato- 
 nem pro recipiam, ut alia eiusmodi complura " f show 
 this ; and the same is probably the explanation of the 
 old forms tamejX cume (and quome, S. C. do Bach.), in 
 which the irrational e was added for the preservation of 
 the otherwise final M. The writing of the m in 
 
 * After the classical period the M became weaker again. 
 
 f Festua, in Jordan's Cato, p. 90. 
 
 X Vid. Scaurus de Orthog., ap. P. p. 2261. 
 
PEOSTTTlf CIATION OF LATIK. 97 
 
 etc., and' the dropping of the e in tam{e), etc., are both 
 signs of an increase of force in the final M. . 
 
 If we weigh the language of Quintilian in discussing 
 the order of the words cum Jiominibus notis, and Cicero's 
 similar remarks on the forms nobiscum, mecum, tecum 
 (Orat. § 154), it will be manifest thai the fanciful reason 
 which they assign was certainly not the principal reason 
 for the order nobiscum rather than cun nobis j and it 
 cannot be reasonably doubted that in the close connection 
 alluded to, a place was sought for cum in which its final 
 letter would net be lost in another sound or in nothing- 
 ness, or cause an offensive stoppage by being given in its 
 own sound.* It is fair, therefore, to infer that in mecum, 
 etc., and also in the sequence cum Jiominibus the final 
 M was weak, to be sure, but nevertheless a distinctly 
 ai^dible M. But for the pronunciation of cum Jiominibus 
 we must call attention very particularly to what Quin- 
 tilian says of the pronunciation of M final before a vowel 
 opening the following word. It will be remembered that 
 his language is very explicit, and, 
 
 1. His statement is inconsistent with our commonly 
 received dogma of Ecthlipsis by which the M and its 
 preceding vowel go for naugJit. Donatus too (Lib. Ill, 
 seq. 4 &,p. L.) does not countenance the dropping of the 
 M in the situation alluded to, when he says : Ecthlipsis 
 est consonantium cum vocalibus aspere concurrentium 
 quaBdam difflcUis ac dura coUisio, ut 
 
 Multum ille et terris iactatus et alto. 
 
 The M remains then, for, it must be to it that the 
 
 * Vid. Servius in II Donati Editionem, ap. P. p. 1797 : Nemo 
 enim dicit cum me, cum te, propter cacophaton. 
 
98 PEONUNCIATION OF LATIK. 
 
 " dura coUisio " is due : it is very much weakened though, 
 for, 
 
 2. Quintilian says that it is faintly rendered and hardly 
 recognizable particularly as M ; but, 
 
 3. While he speaks of the M here as a sort of middle, 
 severed thing between the two vowels, yet he evidently 
 regards it as passing over to the opening vowel of the 
 subsequent word with which it is brought in close con- 
 tact (ita contingit, ut in eam transire possit). The 
 conflict between the statements, that the M passes over 
 to the second vowel, and remains as a sort of mumbled 
 connection between the two is not great, and will be 
 accounted for if we think of the M as actually by the 
 speaker attached to the second vowel while he remembers 
 that it belongs really to the first. 
 
 In this way we shall get Multu' mille, Quantu' merat, 
 cu' mhominihus, etc., etc., in which the m must be 
 obscurely rendered, i. e., as nearly suppressed as possible 
 consistently with its still remaining M, " Neque enim 
 eximitur, sed obscuratur." The previous vowel, more- 
 over, will be given irrationally, i. e., with but faint trace 
 of its own characteristic tone, and so short as not to count 
 for quantity. 
 
 In accordance with what has been said ; if we take the 
 sentence : " numquam ego pecuniam neque meam neque 
 sociorum per ambitionem dilargitus sum," we shall find 
 a clear complete M only in the first syllable of meam 
 {M apertum) ; in the first syllable of amUtionem it is 
 dimmer, weaker, but still M { if mediocre) ; in the sound 
 of the first m of numquam, we have the same weakness, 
 out of which grows an uncertainty of sound in practice 
 
PEOKUNCIATIOK OF LATIN". 99 
 
 more like n than m, as nunquam * {M incertum). Lastly, 
 the final M is, in each of its occurrences, more obscure 
 still than the m mediocre (M oiscurum), but with some 
 differences; thus in numquam the final m is almost 
 nothing {pcene nulKus vocis) as far as numquam is 
 concerned, but passes over in a weak utterance to join 
 the e of ego closely following, thus : numqua' mego (with 
 which we may compare English d in the connection 
 "baffled investigation," if pronounced haffl^ dinvestiga- 
 tion). In all the other instances in which final M is 
 found in the above sentence, nothing more can be said 
 than that it is obscure, for we are not entitled to read 
 pecunian neque, or mean neque, since we have not here 
 the close connection of cun nobis. In pecuniam meam, 
 if we had it, this connection would obtain, and there is 
 every probability that we should sajpecunia' meam. 
 
 The faintness of m in sum is said by Corssen to be sup- 
 ported by the form su, found on an ancient inscription. 
 
 The Latin N" was naturally and normally, for the most 
 part also actually, the same with our own N. This 
 appears from tradition ; from the ancient description of 
 the formation of this letter; by implication also from 
 ancient allusions to the same of a less direct kind ; and 
 
 * It is necessary to remark, however, not in conflict with this but 
 in addition to it, that in this place we shall have the sound of N 
 advlterinum, as descrihed in the next section. 
 
100 PEONUNCIATION OF LATIN. 
 
 from an obserTation of the part played by N in the con- 
 stitution of vords and their changes. 
 
 For the normal force of N we present the following - 
 description of Mar. Victorinus • (P. p. 2455) : JV vera sub 
 cotivexo palati lingua inhmrente,.gemi?io naris et oris spi- 
 ritu explicaUtur : in which, when due allowance is made 
 for the usual want of positiye and experimental accuracy, 
 we see indicated with a good degree of certainty our own 
 N — the nasal letter described by Helmholtz and Miiller 
 in MUUer's Science of Lang., 3d. Ser., pp. 158-9. 
 
 The acddents and affinities of Latin N" in contact with 
 other sounds are conformable to the nature of our "N as 
 indicated, and the identity is further seen in the circum- 
 stance that the Latin Grammarians uniformly class N" 
 with the semi-YOwels. N", therefore, in Latin, like M, 
 was a weak consonant. The firmest sound which it had 
 was generally given it in the beginning of a word. This 
 is fairly inferred from the fact that N in this situation is 
 scarcely eyer lost or changed.* The changes to which N" 
 was subject when variously placed within a word are too 
 numerous to be named and suitably estimated here. 
 
 The most important of them may be seen briefly stated 
 in Forcellini's Lexicon, and fully expanded and discussed 
 by Corssen (Ausspr. I, p. 93, ff.). As we haye not space 
 for so wide a view, we shall give this author's conclusions 
 in his own words, embracing his highly probable opinion 
 of the value of 'N final ; paiising first only to name one 
 peculiar sound of N", pointed out and described by Nigi- 
 dius Figulus, and others, and called by them A/" aduUeri- 
 num. This is a sound like ng, acquired by N when it 
 
 * We cannot speak absolutely here, while we find, e. g., Lympha 
 for Nymjiha. See Festus in Forcellini's Lex. under Lym/phatus. ' 
 
PEONUNCIATION OF LATIN. 101 
 
 occurs immediately before the sound of c, q, and g (there- 
 fore also before x). 
 
 Aul. Gell. N. A. XIX, 14, 7 : Item ex eodem libro {sc. 
 covwientariis Nigidii) verba Ikbc sunt : Inter literam 
 N et G est alia vis, ut in nomine ' anguis ' et ' angari ' 
 et 'ancorm' et 'increpat' et 'incurrit' et 'ingenuus.' 
 In omnibus Ms nan verum n, sed adulterinum poni- 
 tur. Nam iV non esse, lingua indicio est ; nam si ea 
 litera esset, lingua palatum tangeret. 
 
 This language ^s sufficiently unmistakable, but any 
 who wish more proof may, by consulting Mar. Victorin., 
 pp. 2462-3, 2465, and PrisSian, p. 556, see it placed 
 beyond a doubt that the sound indicated as that of N 
 adulterinum was that of English w-in "rancor," "fin- 
 ger," " hungry," etc. 
 
 This is embraced in Corssen's conclusions, which are as 
 follows : 
 
 1. " jN" had the well defined, steady, lingual sound in 
 the beginning of words, and within a word between two 
 Towels,* and (in the classical language) before lingual 
 mutes." 
 
 2. " N" has a weak, dull sound, like Sanscrit Annsioara, 
 and like the n in such German words as Oans, Zins, 
 Sense, when it occurs within a word immediately before 
 s, and (in compound words) also before the semi-vowels 
 / and V, and before the strong labial breathing /; it has 
 
 * We have seen one instance (viz.: Ny7ivpha)rD.'w\x\ch,'va. the 
 beginning, n was not so steady : in the middle of a word between 
 two vowels, it seems also not to have been always quite firm, at 
 least in the popular speech ; for we have Messana turned by the 
 people (" vulgus ") into Measala. See Senec. de Brev. Vit., 13, 5. 
 
103 PEOKUKCIATIOK OF LATIST. 
 
 a weak, dull sound also within a word after m, likewise 
 also in the end of a word." [So that in nomen, etc., the 
 last n is fainter than the first.] 
 
 3. " The sound of N is guttural in its character when 
 it occurs immediately before c, q, g, ch, x. This sound 
 of n the Latins sought at various times to represent in 
 various ways ; by g [as in aggulus for angulus ; see Pris- 
 cian], by nc [nuncquam for nunquam], by c {nucquain], 
 and (occasionally in yery late times) by dropping all sign 
 [as in prici23i for principi]. The sound itself corres- 
 ponds to French nasal n [?] and to the n in German 
 ' Danlc' ' sinken,' ' Klarig,' ' singen.' " [English ' thanh,' 
 etc.] " 
 
 P, Idbris spiritus erumpit. Marrian. Capella. Mar. 
 Victorinus (P. p. 2454) : B et P literm — dispari inter 
 se oris officio exprimuntur ; nam prima exploso e mediis 
 Idbiis sono ; sequens (P) compresso ore, velut introrsum 
 attracto vocis ictu, explicantur. Webster's Dictionary: 
 " P is formed by closely compressing the lips,, and sepa- 
 rating them suddenly with an explosive emission of 
 breath." Add the uniform testimony of tradition to 
 these declarations, and nothing more is necessary to show 
 that, Latin P = English P. 
 
 The only noticeable confusion of the sound of P with 
 that of any other letter is to be seen in the interchange 
 of P and B, the sonant of the same organ. The history 
 and import of that interchange has already been dis- 
 
EKONtTNOIATIOIf OF LATIN. 103~ 
 
 cussed in connection with the letter B. The circum- 
 stances and the time (mostly ante-classical) of the fluctua- 
 tion, point rather to a stiffening of B than to a softening 
 of P. And it has been further made to appear that the 
 practice of the classical time made a clear distinction 
 between the two. 
 
 Nothing more remains to be said, except that in one 
 situation we may take it for granted that the P had a 
 weak sound — that situation, namely, where not unfre- 
 quently it seems to have been employed by insertion to 
 protect an M against the corrupting influences of follow- 
 ing N, T, or S, as in emptus, sutnjpsi, contempnere, 
 (BTumpna (A. Gell., VII, 16, 9, H.). 
 
 Q. 
 
 Q was one of three signs employed by the Latin for 
 representing the guttural tenuis (English K). Through- 
 out the classical time it was prescriptiyely assigned and 
 confined to that situation in which the guttural tenuis 
 was to be followed by V, that is, to the combination QV. 
 In itself considered, Q was held to be entirely equiyalent 
 to G or K. This is abundantly stated on authority, and 
 appears from the fact that many linguists from the early 
 times of the cultivated Latin sought to banish the Q 
 entirely (equally with the K). 
 
 Mar. Victorinus (P. p. 2456) : Licinius Calvus Q litera 
 non est usus. 
 
104 PKONUNCIATIOlf OF LATIK. 
 
 Quint. I, 4, 9 : Q, cuius similis efEectu specieque, nisi 
 quod paulum a nostris obliquatur [compare 9 and 
 9], Koppa apud Graecos .... 
 
 Priscian, L., 1 : Quamyis in varia figura et vario 
 nomine fuit K, et Q et C, tamen quia nnam Tim 
 habent tarn in metro quam in sono, pro una litera 
 accipi debent. 
 
 Accordingly Qura would be equal to cura, 
 Degus " " " " decus, 
 Pequnia " " " " pecunia; 
 
 and these and similar forms are often found on inscrip- 
 tions of ancient date (see Schweitzer's de titulo Mum- 
 miano, in Kuhn's Zeitsch., II, p. S??). The established 
 practice of the later classical period, however, added one 
 more restriction to the use of Q, which was-ithen con- 
 fined to use before V when that element was followed by 
 another vowel. (See Quint. 12, 10.) [The effect of V 
 in this connection has been discussed already, in another 
 place, p. 44.] ' 
 
 By a simple comparison of Latin E with the same let- 
 ter in cognate languages, we soon arrive with certainty 
 at the nature and power of that letter up to certain limits. 
 That is to say, it is easy to ascertain its identity with 
 {e. g.) the English E. But this is still to speak very 
 loosely ; for while our Grammarians have not thought fit 
 
PEONUIirCIATION OF LATIN. 105 
 
 to employ more than one name and sign for E, yet we 
 are familiar with several quite different renderings of 
 thib element. 
 
 There are, namely, with us at least three clearly dis- 
 tinguishable modes of uttering E. First we have the fiat 
 , or dull E, as in " far," " future ;" second, the unfolded, 
 rolled or round E, as in "wrap," "reap," "farrier," 
 "futurity;" and, third, the "trilled" E — not at home, 
 indeed, in English, but heard, e. g., in the French "rap- 
 port." The question must then recur : With which of 
 these E's is the Latin E to be identified ? For an answer 
 to this question we find the material but scant and slen- 
 der. But such considerations as we have to offer seem to 
 favor the belief * that the Eomans sounded their E in 
 each of the three ways above indicated, and that the 
 difiference was with them, as with us, determined to some 
 considerable extent by the situation of the E. 
 
 The separate character of the three E's will be best 
 discerned by reference to the organs of speech ; and it 
 must be observed that in the utterance of each there is 
 the vibration of some soft part in the vocal tube ; and 
 that further, in the case of each, there is a tendency to 
 contact between the point of the tongue and the hard 
 palate, but a short distance behind the teeth. This ten- 
 dency never becomes perfect effect (in which case the E 
 would cease to be the semi-vowel, which it always re- 
 mains), but the degree and mode in which it approxi- 
 mates attainment constitutes a large part of the difference 
 between the E's ; while the other elements of distinction 
 are referable to the pai-ts which vibrate (whether the soft 
 
 * In one case rather the conjecture. 
 
106 PEONTTN^CIATION^ OF LATIN. 
 
 palate or the tip of the tongue), and to the space which 
 confines the vibrating air (whether this be the mouth 
 only, or the mouth together with the pharynx and the 
 cavity of the nose). Some other points of difference 
 there are, but these will be enough for our purpose. 
 
 The dull K, then, is given when the point of the 
 tongue lies nearly passive in the front of the slightly 
 opened mouth. Unless there is smne tendency, however, 
 of the tongue towards the contact of its tip with the 
 palate, the E cannot be rendered at all, and may fade 
 entirely away, as sometimes in English farther, etc. In 
 this E, the soft palate vibrates and the nose is open. 
 
 In the rolled or round E the tongue becomes active, 
 makes a vigorous effort as though towards contact with 
 the palate, but its tip curls backward into the mouth 
 instead, just short of the contact. The air is now let on, 
 the soft palate vibrating and the nose being open as 
 before. Under these conditions, when the tongue un- 
 folds, the rouhd E is delivered. Let it be noted, however, 
 that until the tongue is unfolded, the sound of this 
 E may, instead of being discharged, be prolonged at 
 pleasure. 
 
 In the trilled E the tip of the tongue partially suc- 
 ceeds in making the contact alluded to, but in its per- 
 sistent efforts to maintain it, is repeatedly defeated by 
 the passing column of air. The tip of the tongue thus 
 vibrates against the hard palate behind the teeth. The 
 soft palate does not vibrate, and the nose is closed. 
 
 Whether the trilled or rattling E was known to the 
 Eomans it is not easy to say with certainty, and suppos- 
 ing it known, still more difficult is it to know the limits 
 of its use. It may be that most persons will refer to this 
 
PBONUJSrCIATION OF LATIN. 107 
 
 E, the words of Mar. Victorinus (P. p. 2455) : E, quae 
 Tibratione Tocis in palato, linguffi fastigio fragorem 
 tremulis ictibus reddit. 
 
 Further, it seems not unlikely (more cannot safely be 
 said) that this was the sound of that E which in Latin 
 was readily interchangeable with L ; since L was also a 
 " trilled " letter, according to Helmholtz and Miiller, in 
 Science of Lang., 2d Ser., p. 149, f. 
 
 The interchange of E and L is pointed out by Bopp 
 (§30, Comp. Gr., Vol. I), by Pott (Etymol. Forsch., 
 II, 97), and fully by Corssen (I, p. 80). And the readi- 
 ness with which the transition was made is well illustrated 
 by the jocose use of Hillus for Hirrus by Cicero (Epp. 
 ad Fam. II, X, 1).* This furnishes, however, only a 
 slight ground for conjecturing a trilled sound of E, and 
 on the other hand if Bentley (Hor. Serm. I, 3, 47) is right 
 in supposing Hirrus called " db hirriendo," it is probable 
 that we must recognize in this word the litter a canina, 
 and in that case we have no argument left, that we 
 have discovered, for the trilling of E in Latin, except 
 what may be found in the words of Victorinus, quoted 
 above. 
 
 Passing now to the rolled or round E (English roiv), 
 a regard to the organs of speech, and a general view of 
 the practice of many languages — the Latin included — 
 leads us to determine this as the normal E. For the 
 Latin particularly, we argue that Persius must have felt 
 this, and must have had this round E in view in the use 
 of his phrase littera canina (Sat. I, 109) : sonat hie de 
 nare canina Littera, That E was the letter here referred 
 
 *.See Appendix, Note M. 
 
108 PKONUNCIATIOK OF LATIK. 
 
 to, is shown bywords of Lucilius, reported by Donatus in 
 Ter. Adelp. II, 4, 18, and further by Donatus in Ter. 
 And. Ill, 4, 18, where he says of the word irritatus : 
 Ducitur autem verium a canihus, qui restrictis dentibus 
 hanc literam B imitantur. Too much importance must 
 not be attached to obserTations like these, which have 
 often been abused ; but as far as they are worth inter- 
 preting, they point plainly to the round R, as above 
 described. An easy experiment will show that by causing 
 the voice to dwell on the E in {e. g.) rasp, we effect the 
 nearest possible imitation of the snarling of a dog. 
 
 The nose is open now, as Persius knew, and resounds 
 with the air vibrating from the soft palate ; but this is 
 not the case with the trilled E. 
 
 The same will be thought the most likely sound of the 
 E, which in the history of Latin words appears so very 
 often in the place of former S ; as in 
 
 Furius for Fusius (Liv. Ill, 4), 
 
 Papirius -" Papisius (Cic. Bp. ad Fam. XXI), 
 
 qumrundum {qumrendum) and qumsundum, in a fragment 
 of Bnnius (Eibbeck, Trag. p. 33) ; in connection with 
 which is to be noted, the remark of Festus, p. 358,, M. : * 
 qucBsere ponitur ah antiquis pro qucerere. 
 
 Corssen makes a full exhibition of this very wide- 
 spread phenomenon, and points oiit the similar activity 
 in the point of the tongue in the utterance of S and the 
 utterance of that E of which we now speak, and which 
 we have called the round E. It is true that we have here 
 only another basis for conjecture, but we add it to what 
 
 * See Ribljeck, 1. 1. 
 
PBONUNCIATIOK OF LATIK, 109 
 
 has beeii already said, in support of what seems to be 
 certainly the most probable belief, namely that the round 
 E (our own normal R) was the normal, and most com- 
 mon, R of the Latin. 
 
 We haye now to inquire whether the sound aboye 
 described as that of flat or dull R was ever heard, and 
 if ever, where it was heard, in Latin words. 
 
 Whether we think of the efEect of the sound, or have 
 regard to the relaxation of the point of the tongue in 
 uttering this letter, it is plain that, on comparison with 
 the normal E, the chief character of the dull E is 
 weakness. 
 
 We argue now then briefly that final E in Latin words 
 of more than one syllable is weak, since the vowel before 
 it is weak, and therefore incapable of receiving a strong 
 check after it. We hold, accordingly, that probability is 
 unquestionably in favor of dull, weak E (as in English 
 farther, river) in the end of Latin words, as in jular, 
 nectar, pater, acer, duumvir, soror, mmror, satur, robur, 
 etc. The vowel before E final is short and without stress, 
 while in many cases it must be supposed so weak and 
 indifferent as to be almost irrational, as in satur, robur, 
 acer, duumvir, and many others. With such a sound 
 before, and no sound to follow, the R must be faint. 
 
 s. 
 
 In. such allusions to the Latin S, as have heretofore 
 been occasionally made, we have assumed for it that 
 
110 PRONUNCIATIOK or LATIIT. 
 
 sound commonly known as the sharp s. In proof of this 
 as the normal character of Latin S, we need only refer to 
 the Toice of tradition and the sustaining evidence to be 
 had by even a cursory series of observations of the part 
 played by S within the Latin and the languages allied to 
 it. We need not introduce these observations here, 
 partly because a sufficient number of them are familiar 
 to all, partly because some of them will appear below, 
 and partly because the tradition, undisputed, is enough 
 to establish this one point, viz. : that the normal force of 
 Latin S was that of the English S in "see" "bless." 
 
 Prom regard to the organs engaged, this S may be 
 called the t hiss, While the soft S of our own and other 
 languages may be marked as the d hiss (z). The ques- 
 tion remains, whether the Latin did not also use the soft 
 rendering of S. We have for answer to this, that in what 
 we have found said of 8 by the ancient Grammarians, no 
 allusion, at least no distinct allusion, can, be found to 
 the soft sound. 
 
 The importance of this negative argument is great, and 
 other reasons must be strong if they shall be sufficient to 
 establish the soft sound. Such reasons are to be sought 
 in the conditions of the use of S ; and some, accordingly 
 (as Corssen), have wished to interpret certain appearances 
 among those conditions, as indicating a soft sound for 
 this letter in the midst of a word between two vowels. 
 
 But what shall we say, then, to the doubling of S in 
 Cicero's time in such words as caussa, aussus, divissio, 
 odiossus, etc. ? Quint. (I, 7, 20) : Quid ? quod Ciceronis 
 temporibus paulumque infra, fere quotiens S littera 
 media vocalium longarum vel subiecta longis esset, gemi- 
 nabatur, ut caussce, cassus, divissiones ? 
 
PEON-tJNCIATION OF LATIN. Ill 
 
 We scarcely need the assurance of Vel. Longus (P. p. 
 2233) : " S vero geminata Tocis somirn exasperat," or 
 that doubled S is intended to mark the sharpest sound 
 of S. 
 
 And from this it appears that the single s in causa, 
 etc., when so written, must also have had the sharp 
 sound ; if, as Corssen admits, it is plain from that fluc- 
 tuation in orthography between ss and s (which he 
 further illustrates from inscriptions), that there " could 
 not have been an essential difference " in the pronuncia- 
 tion, in the situation alluded to, between the doubled 
 and the single sign. For if Corssen, admitting this, will 
 still speak cauza for causa, he must also say that cauzm 
 was the sound that Cicero, Virgil, and others, thought 
 proper to indicate by caussa; but in this position he will 
 find himself pointedly opposed by the words quoted 
 above from Vel. Longus, and will further lay himself 
 under the necessity of saying iuzzi or iuzi for iussi ^or, 
 according to the tesimony of Quintilian, in continuation 
 of the passage quoted above, this word was, shortly before 
 Cicero's time, also written iusi). 
 
 The readiest and most reasonable conclusion which can 
 be drawn fi'om that fluctuation in orthography seems to 
 be simply this. There may have been a tendency at 
 times to weakness in the utterance of s occuiTing between 
 two vowel sounds, the first of which was long. Possibly 
 to guard against this, certainly for some reason, there 
 were those who, about Cicero's time, thought it best to 
 mark s in this place by a double sign (not to change a 
 d liiss into a t Mss, but) simply to mark unmistakably 
 the normal strong t hiss. When this doubling of the s 
 
112 PEONUNCIATIOSr OF LATIIT. 
 
 became again unusual,* it was because the single s could 
 be relied on for the purpose ; while the same opinion had 
 influenced the writers of an early period in using only 
 one s in iussi, and probably other perfect forms. 
 
 It is also easy to suppose that some slight variety in 
 syllabication might prevail in such forms in the practice 
 of a limited time, and that on such variety might be 
 based a particular idea with some as to the proper form 
 of orthography. Again, the first of two s's may have been 
 now used, and again supposed unnecessary, for metrical 
 purposes, to mark as long the foregoing vowel sound. 
 But in any of these suppositions the identity in character 
 between ss and simple s must be taken for granted. 
 Hence, the significance of thiS' phenomenon of doubling 
 the S is altogether in favor of the sharp sound of single 
 S in the midst of a word between two vowels, f 
 
 Final S, in words of more than one syllable, deserves 
 particular notice, because of a weakness to which this 
 consonant, like others, was subject in this situation. It 
 does not follow, however, that weak 8 is soft S. [It has 
 not perhaps been sufficiently remembered that a t hiss 
 maybe weak or faint, without being turned into a d hiss : 
 thus in English business (biznes) we have a strong d hiss 
 
 * It wag disapproved by Quintiliau, and condemned by later 
 Grammarians, as by Q. Ter. Scaurus, P. 3357, and others. 
 
 f Corssen has several independent reasons to urge for the soft 
 sound of S between two vowels. But they are not unanswerable. 
 We are a little surprised to see to what an extent that learned 
 author is led in this particular by the analogy of modern languages. 
 Moreover, in citing Italian cosa as containing soft s, we find him in 
 conflict with Dr. Mahn (Lehrbuch der Ital. Sp.), who says, " S is 
 almost always sharp ; as in ' sentire,' ' cosa,' ' caaa.' There are 
 only a few exceptions — perhaps in a dozen words." 
 
PEOKUNOIATIOK OF LATIM-. 113 
 
 (z) followed by a weak t hiss (s)]. To say nothing of 
 other considerations, the weakness of final S during the 
 best time-of the Latin, is probable in view of the manifest 
 exceeding faintness of that letter in the end of words 
 both before and after the classical age. For the ancient 
 weakness of final S we need only point to the extensive 
 dropping of S from the end of grammatical forms. Like- 
 wise also in the post-classic time, on inscriptions of 
 the late Empire are said to occur the forms JiKo (for 
 fiUos, nominative case), Mariu, mtati, and many others 
 with loss of final S (Corssen, 1, 120). It is hence plain, 
 that during the classical age itself the final S could not 
 have been strong ; yet we have evidence that it was then 
 thought proper to sound it stronger than in the earlier 
 or the later times, when it often faded into nothing. 
 
 Cic. (Orat., c. 48, § 161) says : Quin etiam, quod iam 
 subrusticum videtur, olini autem politius, eorum ver- 
 borum, quorum eaadem erant postremae duse literse, quse 
 sunt in ' optumus,' postremam literam detrahebant, nisi 
 vocalis insequebatur.* Ita non erat ofEensio in versibus, 
 quam nunc fugiunt poetse novi. Ita enim loquebamur 
 [al. loquebantur] : Qui est omnibu' princeps, non omnibus 
 princeps, et, Vita ilia dignu' locoque, non dignus. 
 
 S final, accordingly, although weak, should still be 
 heard distinctly, and that too with its normal sound. 
 
 But, lastly, there is reason for believing S to have had 
 not only the weak but also the soft sound (2) when it 
 occurred at the end of a word, and was preceded immedi- 
 ately by the letters B or N, as in plebs, scobs, mans, demens, 
 
 * A vowel sound immediately following, if in close connection, 
 would practically destroy the final situation of S. 
 
114 PEOliTUNCIATIOH' OF LATIN. 
 
 dodrans, etc. The Latin S (sharp) is, as to the effect 
 exerted by it on a preyious contiguous consonant, per- 
 fectly analogous to Grreek 2. Thus, while in Greek Kg or 
 yq = I, so in Latin cs or gs = x. In Greek r or tJ before 
 ? are dropped ; so in Latin, as in " miles," " palus." But 
 in Greek (3 before g becomes rr {n. s., (pXeip, stem (pXefi-) : 
 thus in Latin we ought to haVe aps, apsque, apscondere, 
 opsonium, opsecro, etc., besides, nupsi, scripsi, lapsus, etc. 
 The B might be rejected, also, or assimilated as in iusi 
 or iussi, but could not stand as B before the normal S. 
 Accordingly, when such forms as aps, opsonium, etc., 
 found in the old Latin (as seen, e. g., in the critical edi- 
 tions of Plautus), came to be.written ahs, obsonium, etc., 
 this must in all probability have been in consequence of 
 a softening of the S. 
 
 Again, the v falls out in Greek before o in At/isffi (for 
 Xiiievai), etc. So also often in Latin n falls out before S, 
 as in 
 
 toties 
 
 for 
 
 totiens, 
 
 quinquies 
 cosol 
 
 
 quinquiens, 
 consol. 
 
 prsQgnas 
 
 iC 
 
 praegnans, etc, etc. 
 
 Now N, being the nasal of D, is much more easily con- 
 sonant with the d hiss, or soft S, than it could <be with 
 the hard or normal S. We infer, then, that it was the 
 hard S before which the !N" fell away, as in toties, while 
 that sound of the S must have been the softened one, 
 which admitted N before it, in the same syllable, as in 
 totiens. 
 
 So that we ought to give the S soft in mens, amxms. 
 
PBOHrUNCIATIOM- OF LATI]<r. 115 
 
 etc., -srhile, however, it remains hard as usual in men-sura, 
 man-sit, etc. 
 
 S, then, is normally and generally in Latin the hard 
 or t hiss, the strength of which is moderated in the end 
 of a word. The hiss is transferred to the d contact when 
 S occurs in the end of a syllable, before a syllabic pause, 
 and immediately preceded by the consonants n or i. 
 
 T. 
 
 It will not be thought necessary to adduce here proofs 
 of the normal power of Latin T. The conditions of the 
 use of this letter in Latin words, its appearances and 
 disappearances, furnish evidence coincident with that 
 which more obviously arises from comparing the use of 
 the same consonant in Latin forms and in corresponding 
 forms of cognate languages ; and all this evidence, most 
 of which lies very near, justifies the unquestioned prac- 
 tice of tradition in giving to Latin T the same power as 
 that of our own (lingual-dental tenuis) T. 
 
 We must infer that T in the end of words had a weak 
 sound ; and this in the case of polysyllables, from a gen- 
 eral regard to the barytone prosody of Latin words. But, 
 secondly, evidence of the same is also seen in the dis- 
 appearance of t from the end of words (monosyllables 
 included) during the period antecedent to the time when 
 Latin words assumed their settled forms of orthography. 
 Thus, to speak now only of nouns, a t sound originally 
 
116 PEONUNOIATIOX OF LATIN. 
 
 belonged to the ablatiye case, as in dolot iov dolo, malot 
 tor malo, capitet for capiie, concionet for condone, iuret 
 for iure, suot for suo, etc. These words, found in the 
 XII tables, were written, it is true, dolod, hapited, cond- 
 oned (cf. coveiitionid), toured, souod, etc. But this d 
 (representing Sanscrit ^, sign of the ablative case) was 
 sounded, as we have before seen, not differently from t. 
 There was here, therefore, a dropping of the final t sound. 
 Such dropping of T becomes more 'striking proof of 
 the general weakness of that letter in a final situation, 
 when instances occur in which it sometimes falls away 
 and sometimes not. And this is the case, for instance, in 
 the Haut and Hau (both for Haud ) in MSS. of Plautus. 
 "We have therefore in tangit, cantat, mittit, essentially the 
 same T twice, but weaker in the last place.* 
 
 There is only one, but that in effect a very decided 
 corruption of T, which now demands our attention. It 
 is maintained and very commonly admitted that already 
 in the classical age, an assihilation was superinduced 
 upon this letter by following I, when the I was followed 
 again by another vowel sound. 
 
 It is well known that the representative of this Latin 
 T, which appears in the derived forms of modem lan- 
 guages, is in them given by some form of a hiss. (Latin 
 iustitia, Italian giustizia, Spanish justida, French justice, 
 
 * It is probable that final d was practically not distinguished in 
 sound from final t. This was argued before at pp. 69, 70. To the 
 evidence there given we add a remark from Phocas' de Aspiratione 
 (P. p. 1724), incidentally confirming the same point in the case of 
 Tiaud ■= haut. Adverbium (haud) autem discretionis causa dasian 
 [H] habet prsepositam, ut ne cuniunctio fuerit, veluti legimus: 
 'Aut portum tenet.' 
 
PKONUNCIATION OF LATIN. 117 
 
 etc., etc.) This fact used as an argument requires the 
 assumption of a hissing sound for the Latin T (more 
 strictly TI) in the situation indicated, at a time early 
 enough for the Latin letter to have served, as it un- 
 doubtedly did, as a common source for the Spanish c, 
 Italian z, French p, etc. 
 
 But as this falls far short of proving- a hissing sound 
 for T in the classical time of the Latin, other evidence 
 must be had for the use of such a sound in that early 
 time. 
 
 Here, then, is to be considered again that fluctuation 
 in orthography between ci and ii found by those who 
 have had access to ancient inscriptions, upon stones of 
 date as ancient as the early Empire. The value of this 
 phenomenon as bearing upon the question of a similar 
 hissing sound for both ci and ti we have already con- 
 sidered in connection with the letter 0. 
 
 We ventured then to express the opinion that the few 
 earliest observed interchanges of C and T do not require us 
 to suppose a third sound difEerent from the proper sound 
 of each of these ; that is, that nuntius and nunhius might 
 have existed together as dialectic varieties (as moihie for 
 moitie in Canada, and 'ay/cXid for 'av-Xia in modern 
 Greek, etc.). If this is so, the> argument for a hissing 
 sound of either ci or ti which is founded on the inter- 
 change of these letters, must remain uncertain until a 
 hissing sound is established for either the ci or ti upon 
 other grounds. 
 
 Now this is easy to do, in the case of both, in the late 
 and corrupt days of the Latin, but attended with ever- 
 increasing difficulty (according to the exhibition of those 
 who have had access to the most ancient records) on pene- 
 
118 PEOKUKCIATIOKT OP LATIN. 
 
 trating backward towards the classical era. In default of 
 data sufiBcient for a settled opinion, we cannot do more 
 than record the conclusion which will generally, no 
 doubt, be regarded as justly reached by Corssen. That 
 is : that the T before I followed by another vowel hegan 
 to be heard with a hiss in the very best times of the 
 classical Latin ; that the assibilation became more uni- 
 versal as well as more raarked with the advance of time. 
 As to the sound itself ; it was in the late Latin marked 
 by z, as in iustizia, as well as also by Grreek f and t^, by 
 ts, and by tc. Comparing Sova^iovefi (for donationem), 
 'aKT^io (for actio), Constantsa (for Gonstantio), milizia 
 (for militia), Bincentce (for Vincentim), all late forms ; we 
 see that the hiss sometimes was felt to represent the T 
 only, sometimes the TI together, and that the sound 
 probably was about that of ts or tscJi — such a ts, e. g., as 
 might be put in place of ce in the end of English ' ounce,' 
 thus, ountsj or such a tsch as might represent the pro- 
 nunciation of te in English "righteous" thus, "rit- 
 tschus." 
 
 lustitia, according to this, would sound iustitsia or 
 iustitschia. 
 
 The limits to the hissing sound of T in ti must be 
 briefly stated. 
 
 1. The I must be consonantal I, and consequently short. 
 Hence, when the I is long, as, e. g., in totius, the T has 
 the pure sound. Further, in Greek words (as u^gyptius), 
 as the I is not consonantal, the T cannbt bfe hissed. 
 Again, in the old infinitive passive forms, as nitier, quo- 
 tier,* etc., the I was not consonantal, as appears, and the 
 T was not hissed. 
 
 * For we never find nmer, nitsier, etc., written for this form. 
 
PRONUN-CIATIOlSr OP LATIN. 119 
 
 2. For manifest natural reasons it is commonly agreed 
 as improbable that the T could have been hissed when it 
 was immediately preceded by another T, or by a hiss, as 
 S or X. (See Zumpt's Lat. Gr., p. 7. 
 
 Notwithstanding a dispute which has caused some 
 doubt concerning the time when X was adopted into 
 the Latin alphabet, and notwithstanding that some 
 writers have wished to fix this as late as the time of 
 Augustus, yet it is certain that X appears in the oldest 
 records of -Latin which have been preserved to us; and 
 many facts might be adduced to prove that the letter 
 was commonly, if not universally, used throughout the 
 classical period. Quintilian (Lib. I, c. 4, § 9) makes allu- 
 sion to X as occupying the last place in the Latin " com- 
 pendium," or alphabet : and as to the power of the letter, 
 the same author remarks, that the sign was not indis- 
 pensable for use in the Latin. That implies, of course, 
 that its office could be fully performed by other means 
 at the disposal of the Latin alphabet. Accordingly, all 
 the Grammarians affirm, and their declaration is abun- 
 dantly confirmed by collateral evidence, that X was a 
 monogram representing the combination OS or GS. So 
 perfect was the equivalence that many, especially in 
 earlier times, thought proper always to write cs or gs 
 instead of resorting to the simpler figure. ^ And Nigidius 
 
130 PRONUNCIATION OF LATIN. 
 
 Figulus, the learned linguist and Latin purist of the time 
 of Varro and Cicero, is said never to have used the X. 
 (See Mar. Victorinus, P. pp. 2456 and 3466.) 
 
 Max. Victorinus (P. p. 1945, Eind. 277) : Ex his (semi- 
 vocalibus) duplex est x : constat enim aut ex g et s 
 Uteris, aut ex c^ et s, ut puta rex, regis j jjix, picis. 
 Quippe ante x literam, quae postea in compendium 
 inventa est, rex per gs \regs], item pix per cs [pics\ 
 veteres scribebant. 
 
 Turning now to the monumental records (Corssen I, 
 p. 124), we find that indiscriminately throughout the 
 classical time it was common to find such forms as saxsum, 
 maxsume, proxsimum, deduxsit, faxsit, etc. In this we see 
 evidence that the S element of the X was heard as 
 sharply as possible, and in that point there is no dis- 
 tinction observed between the X which stood for gs and 
 the X which stood for cs. Further, from what we know 
 of the nature of G we are prepared to believe that when 
 thus closely combined with s, the sound would be vir- 
 tually, in all probability, precisely the same as that of 
 the c element in cs. 
 
 We do not therefore suppose two sounds, one harder 
 and one softer for Latin X, but believe that during the 
 classical age that letter was sounded like our own sharp 
 X in "mix." 
 
PEONUNCIATION OF LATIN. 121 
 
 z. 
 
 In tlie case of Z, as once before also, in the case of Y, 
 ■we haTe to consider the sound of a Greek letter as it was 
 heard among the Eomans ; who employed it sometimes 
 in the rendering of Greek words. X was the twenty-firs b 
 and last letter of the Latin alphabet for the Ciceronian * 
 and later classical age ; and however true it may be that 
 the Z was not unknown to Latin usage at a much earlier 
 date (see Vel. Long., P. p. 3316, and Agn. Cornut., p. 
 3386), yet it is plain that the common consciousness of 
 the classical age was quite unmindful of any importance 
 belonging to such an historical fact, if, indeed, its exist- 
 ence was thought of. Certain it is, that during the 
 period of which we have to speak, Z was recognized as a 
 Greek letter. So the Grammarians : 
 
 Asper Junior (P. p. 1735) : Sunt autem .... literse 
 
 latino sermoni accommodatse una et viginti, quibus 
 
 Graecorum accedunt duas z ei yj nam Mezentium et 
 
 ■ Hylam et alia nobis peregrina nomina scribere et 
 
 enunciare proprio sono non possumus. 
 
 Donat. (P. p. 1737, Lind. L, p. 6) : Y et z remanent, quas 
 litteras propter Graeca nomina admisimus. 
 
 Dioined. (P. p. 421) : Z consonans semivocalis duplex 
 graeca, quae propter graeca vel barbara nomina ad- 
 mittitur, ut: Zenon, Zacynthus, Mezentius, gaza. 
 
 * Cf. CSc. de Nat. Deor., II, c. 37. 
 
123 PEOKtTNCIATION OF LATIX, 
 
 Pro hac veteres duabus ss utebantur, ut Messentius, 
 et pitisso, tablisso, etc. 
 
 Our question then is : IIow did the Romans sound the 
 Greek Zeta ? Without doubt, as the Greeks themselves 
 did. But this is an answer satisfactory to those only 
 who, independently of this our inquiry, know what the , 
 sound of Greek Zeta was. 
 
 The statements of the Latin Grammarians are, upon 
 their face, so perplexing and inconsistent touching the 
 power of this letter, that a reconciliation of their views 
 upon any likely hypothesis seems at first almost hope- 
 
 As Diomedes above, so Agnceus Cornutus (P. p. 2286) 
 says : . . . . pro ilia [Z] {in antiquis libris) ss pone- 
 bantur, ut : crotalizo, crotalisso ; malieuo, malacisso, 
 et his similia. 
 
 Vel Long. (P. p. 2317) : Mihi videtur .... esse aliud 
 z, aliud sigma .... 
 
 Taking ss to represent only strong, clear, decided sj 
 these statements say, the first that z is (or was virtually 
 considered to be) equal to s ; the second that z is not 
 (at least not properly) equal to s. 
 
 Again, in Vel. Long. (P. p. 2216), we have the report 
 of Verrius Flaccus, that it was the opinion of some in 
 his time, that z consisted of a and t5. {sd.) 
 
 So Max. Victorinus (P. p. 1945, Lind. I, 277) : Literae 
 peregrinse sunt Z et Y, quasi peregrinse a nobis prop- 
 ter Graeca qusedam nomina assumptte sunt, ut Hylas, 
 Zephyrus; quae si non essent, Hoelas, et Sdephirus 
 
PRONUNCIATION OF LATIN. 133 
 
 diceremus. — • Mezentium per s et d Mesdentium 
 scriberemus. 
 
 On the other hand, Mar. Victorinus (P. p. 2453) : Sic 
 et z, si modo latino sennoni necessaria esset, per d et 
 s literas faceremus. (ds.) 
 
 Further, Priscian (P. p. 552) : Antiqnissimi quoque 
 Medentius dicebant pro : Mezentius. 
 
 But, Vel. Longus (P. p. 2211') : Scribe enim. per unum 
 Z, ct consule aurem ; non erit al^rixo-? quomodo 
 aSexaq ; — . 
 
 Thus we have here these statements : 
 
 1. Z was anciently represented in Latin by ss : but it 
 was not equivalent to pure s. 
 
 2. Some persons in ancient times put a <? in the place 
 of z in foreign words : but if the ear be consulted, d will 
 not give the proper sound. 
 
 3. A third way which seemed proper to many for 
 exhibiting the phonetic power of z was a combination of 
 
 'the two just named; they wrote it by sd; but others 
 gave it by ds. 
 
 Now it, appears to us that there is only one hypothesis 
 upon which may be reconciled and harmonized so many 
 diverse endeavors to indicate the same sound ; and that 
 hypothesis is the assumption that the sound intended was 
 that of Modem Greek Zeta and English z. Submitting 
 this sound to the test, we find — 
 
 1. N"o surprise in seeing it represented by the cognate 
 
124 PEONUNCIATION OF LATIN. 
 
 S* and are at the same time quite ready to be told that 
 it is not sigma, or pure 8. 
 
 2. If we remember that our Z or soft S is a d-hiss, or a 
 hiss passed over the tongue when just ready to finish the 
 d-contact, it will not be strange that some more mindful 
 of this last condition of the utterance, should have sought 
 to characterize it by the d alone (these not feeling that 
 the z was a double consonant). Others, of course, would 
 be discerning enough to see that d could not be a true 
 and complete mark of the sound. 
 
 3. Others again might naturally think it best and right 
 to represent both the hiss and the d-attitude of the 
 tongue. The interesting fact that both sd and ds were 
 employed to render it, seems to indicate this as the only 
 aim of those who wrote so : that is, that each of these 
 compounds were intended as modes of indicating a 
 d-hiss; the one meaning, as it were, a hissed d, the 
 other a d hissed. 
 
 Finally, when almost all the ancient Grammarians 
 agree in calling Z a double consonant, we feel obliged 
 to regard this as an imperfect and inaccurate observation ; 
 in which view we have the support of Velius Longus, 
 who says (P. p. 2217) : Denique siquis secundum naturam 
 vult excutere hanc Kteram, id est Z, inveniet dupUcem non 
 esse, si modo illam aure sinceriore ezploraverit. ' In our 
 opinion, therefore, 
 
 Latin Z = English Z. 
 * See Appendix, Note N. 
 
PEONUNCIATION OF LATIN. 125 
 
 Ch, Ph, Th, Rh. 
 
 "With respect to these combinations it will be enough 
 to report the statement of Diomedes and Priscian. 
 
 Diomedes (P. p. 417) says: ^quoque interdum conso- 
 nans, interdum aspirationis creditur nota. Hebc si c 
 mutse subjuncta fuerit, % notat grascam : si p prsepo- 
 sita fuerit (j> significat. Item si t prseposita fuerit 
 aspiriationi, pro ponitur grteca ; sicut p et s simul 
 positse ip grsecam afferunt. Cseterum h vocalibus 
 nunquam supponitur nisi in interiectione ah. 
 
 Priscian (P. p. 543) : Sciendum tamen, quod hie quo- 
 que error a quibusdam antiquis Grsecorum gramma- 
 ticis invasit Latinos, qui ^ % 5 semivocales putabant ; 
 nulla alia caiisa, nisi quod spiritus in eis abundet, 
 inducti. Quod si esset verum, debuit c quoque vel t, 
 addita aspiratione, semivocalis esse ; quod omni 
 caret ratione. Spiritus enim potestatem literse non 
 mntat: unde nee vocales addita aspiratione aliaa 
 sunt, et alias ea adempta. Hoc tamen scire debemus, 
 quod non tam Axis labris est pronuncianda /, quo^ 
 modo^A; atque hoc solum interest inter fet ph. 
 
 Ch, ph, th, and rh, therefore, are nothing but c, p, t, 
 and r, with the subaddition of h, which does not change 
 the nature and power of the previous part of the com- 
 bination. Accordingly, ch was sounded like hh in khorn 
 
126 PRONUNCIATION OF LATIN.. 
 
 (remains of English ink-horn), and so for the others, so 
 that 
 
 ch was sounded like ch in Sanscrit chanda, achanda, 
 ph " " " ph " " pjialla, utphulla, 
 
 th " " " th " " atha, thurvami, 
 
 rh " " " p " Greek 'Pu>, 'appijv. 
 
PBONUNCIATIOBT OF LATIIf. 
 
 127 
 
 8 
 S °i 
 
 3 P 
 
 
 
 
 1 
 
 
 
 
 ^ 
 
 
 
 
 g 
 
 
 
 
 SO 
 
 •ri © ^ w 
 
 
 
 o 
 
 3 -s B f< 
 
 1 
 
 
 g 
 
 
 i 
 
 i 
 
 gg i3g g 1 
 
 P 
 
 ^ 
 
 8 
 
 
 ti 
 
 *H 
 
 a 
 
 
 *r4 
 
 o 
 
 
 
 TS 
 
 S>t 
 
 
 5 s ; 
 
 d 
 
 II 
 
 
 
 - s 
 
 ,d 
 
 
 
 r4 
 
 •♦3 
 
 
 2 5 2 5 3 
 
 .3 
 
 
 1 
 
 
 ^ 
 
 
 
 
 ■g 
 
 q S> t- O 
 H _D w p— I 
 
 -a "g » "3. 
 
 
 B W ^ W O ^ 
 
 
 
 «-^ 
 
 
 .s ^ 
 
 §1 -- 
 
 3 -a p. ff 
 
 P3 
 
 bo 
 
 3 « 
 
 ^ 
 
 t3 
 
 
 o 
 
 
 (3 - 
 
 bo 
 
 a 
 
 ma 
 
 o S 
 
 ^ l!3 J9 I'U 3«»- 1^ 'O 'O 3$ 1^ '^ 1^ 
 
 ■ ■ l__l 
 
 ;3 « 
 
 
 •Sj -S 
 
 
 a,.r. 
 
 <13 
 
 3 
 
128 
 
 PKONUIJCIATION OF LATIN. 
 
 M 
 
 a 
 
 I 
 
 ii 
 
 fe o 
 
 1 
 
 
 'S 
 
 d 
 
 
 o g 
 
 § 
 
 ^ 03 O N 
 
 1 
 
 =8 ,a 
 
 t3 
 
 I 
 
 q -3 
 
 ^ ^ 3 
 
 bp 
 
 08 .rj 
 
 
 -a 
 1- I 
 
 r-H eS o °3 
 
 
 .g 
 
 
 
 e OS , 
 
 a -e 
 
 ^ as 
 
 O IS 
 
 9 9-" 
 1-1 h-5 
 
 s>, I, 
 
PRONUNCIATION- OF LATIN. 
 
 129 
 
 
 m 
 
 ^ Si 
 
 9 IS 
 
 
 9 
 
 g-9 
 
 I 
 
 M 
 
 O 
 
 
 5 ^ 
 
 w 
 
 6 s, 
 
APPENDIX. 
 
 Note A, page 40. 
 
 On the sound of u after, and In the same syllable with », g, q. 
 
 That the combination su sounded like English sw, and not like 
 English sv (therefore, swam, etc.), is very strongly conGnned by an 
 examination of such words as, anterior to their life in the Latin 
 speech, contained the sound sv. There are numbers of such. And 
 if we compare, e. g., 
 
 Sanscrit svas with Latin suus, 
 " s-casar " " soror, 
 " svwpnas " " somtms, 
 " s'iedas " " sudor, 
 
 and many other forms, we see that the Latin was averse to the 
 connection sv, when v was a full consonant. To avpid that con- 
 nection, the Latin proceeded to reduce the given forms to its own 
 preferred softer sound, either by rejecting the v altogether, as in 
 the above examples except the iirst, or by investing the v with the 
 character of a vowel, as in suus for svas (although in old Latin the 
 forms SOS, sis, etc., for suus, suis, etc., are said to be found also). 
 Now then, when, corresponding to Sanscrit SBOdus, we have Latin 
 stMvis, it is certainly to be presumed that the u of suavis is not 
 sounded like English v, but that the utmost approach to a conso- 
 nantal sound will be represented by the English w, like the vowel 
 u in quattuor (when that word is pronounced with two syllables). 
 F after t is also inadmissible in Latin, as is well shown in an 
 interesting article by Grassmann, in Kuhn's Zeitschrift fttr Ver- 
 gleichende Sprachforschung, Vol. I, p. 1 and after, where he seeks 
 to extend the same observation to lev (g«) {jgv). 
 
APPEKDIX. 131 
 
 -Thus [g in Sanscrit being represented by the guttural tenuis 
 (k, c, q) in Latin], Sanscrit fcara, dog, would give Latin qvanis, or 
 kvanis {cvanis) ; but the Latin will not have a d so situated, and 
 drops it, leaving canis. Again, when (vas passes into Latin the ■» 
 is changed to r, and we get eras : so also with (u, to grow, from 
 which the Latin has (not gves-co but) eresco. And in the Latin 
 rendering of Sanscrit agvas, another mode of avoiding ev (or g») is 
 adopted, that namely of noealizing the « ; for so we get equus. Now 
 this equus is a perfect parallel to suits in respect of the value and 
 origin of its first u ; and from the point of view furnished by com- 
 parative grammar, no one has more reason for supposing u to be 
 silent in equus, than may be alleged for its silence in svus, svamum, 
 guattvor. And further, if, confining the view to the Latin alone, , 
 it be insisted that there is direct testimony to the fact (in Pris- 
 cian) that equus was sounded equs (or ecus), tliis may be admitted, 
 as well as some other (even numerous other) instances of the same 
 sound as indicated by oculvs and oqulus, secutus and sequutus, etc. 
 On the other hand, we can show also the forms sos, sis, for suus, 
 suis, savium for suavium, quattor for quattuor, urgeo for urgueo ; 
 and are we, therefore, everywhere to pronounce, e. g., sus, sam, for 
 suiis, suam ; savis, sesco, for suavis, sueaco ; angis, Unga, for anguis, 
 Ungua f All as reasonably as torket for tdrquet. 
 
 Note B, page 57. {^.) 
 
 Gellius, XVI, 12, 8 : M. Varro . . . . M. Catonem et ceteros 
 tetatis eius ' feneratorem ' sine a litera pronuntiasse tradit, sicuti 
 'fetui ' ipse et 'feeunditas ' appellata. 
 
 A very instructive passage on the sound of m occurs in Quint. Ic 
 5, 17. It would occupy too much space to develop all that it 
 implies. The principal point is that the m of Asice, etc., ought not 
 to ie spoken with " divisio," i. e., so that both elements be distinctly 
 heard ; while on the other hand the (e of Phaethon {^ac8ov) ought 
 to be so sounded, but that nevertheless P. Varro had rendered this 
 Greek name per " complexionem," i. e., giving the ce in one sound 
 . (therefore, Phmthon — Phethon). Notice the accent of 'taeffuv. 
 
133 APPENBIX. 
 
 Note C, page 57. 
 
 The following words of Varro seem to indicate, first, a difference 
 between the sounds of ffi and of e, and second, that that difference 
 was mry small. 
 
 Varro, L. L., Lib. 6, c. 5 : Obsccenum dictum a scena, ut Graeci 
 (lis enim oktiv^ dicitur). Ea (ut Accius scribit) sccena. In 
 pluribus rectis A ante E alii ponunt, alii non ; ut, quod partim 
 dicunt scceptrum, partim dicunt sceptrum. 
 
 Note D, page 58. 
 
 There are a few instances also of the settlement of au into the 
 sound of its first element a. For such is the history of the first syl- 
 lable of arrugia from aurwm, and Arruntius (also spelled Aruntius) 
 from Aurunca, a town of Campania. [Arruntius is for Arruncius 
 (cf. AUius, Accius), an enlargement of Arruncus, (cf. Atus, Attiu», 
 Appius, Servus, Bermvs, TuUus, TiiUius).'\ Cf. Auruneuleius formed 
 on the same stem. 
 
 Note E, page 64. {(E.) 
 
 Max. Victorinus (p. 1946) (Lind. I, 377) says : " We have taken 
 from the Greek Y and Z for the spelling of Greek words, as e. g., 
 Hylas, Zephyrus. If we had not done this we should have said 
 Hcelas and Sdephirus." If the Greek T was what we hold it to be 
 (nearly the French u), and this ffi of HcelaB was no nearer to it than 
 the 8d of Sdephirxis was to the true sound of Z ; then our assump- 
 tion for the normal sound of m is, to say the least, not weakened by 
 the above declaration. It should have its weight, however, in 
 assisting our estimate of the true tone of the diphthong, which we 
 have not felt entitled to indicate more definitely than by supposing 
 it nea/rly like that of German Oe in Oel. 
 
APPENDIX. 133 
 
 Note P, page 67. 
 
 The affinity of B with V (Digamma) was, however, not obscurely 
 felt in early times also. Priscian, ap. P. p. 709 : " Habebat hsec P 
 litera sonum quern nunc habet V, locio consonantis posita, unde 
 antiqui (before the classic period) AF pro AB scribere solebant, 
 .... Sifilum quoque pro StbUum, teste Noiiio Marcello de doc- 
 torum indagine, dicebajit." 
 
 NoTS G, page 73. 
 
 In favor of this double sound may be adduced on authority of 
 MSS. the forms Atus and (feminine) Aca (and Attus, Acca), natta 
 and nacca, also nacta, stloppus and scloppus (?). Here there can be 
 no idea of a compromise by a hiss for the c and t. And before any 
 one will undertake to say that there is any reason in the general 
 principles of Phonetics, any physiological or other natural difficulty 
 in the contemporaneous employment of two such different sounds 
 as k and t in the same word, he should consider some facts stated 
 by Max Muller (Sci. Lang., 2d Ser., pp. 181, 182) : " No two conso- 
 nants would seem to be more distinct than k and t. Nevertheless, 
 in the language of the Sandwich Islands these two sounds run into 
 one, and it seems impossible for a foreigner to say whether what he 
 hears is a, guttural or a dental. The same word is written by 
 Protestant missionaries with k, by French missionaries with t." 
 Again, " we are told by Careful observers that the lower classes in 
 Canada habitually confound t and k, and say mekier, moikie, for 
 metier, moitie." 
 
 Consider also the interchange of the equally diverse guttural and 
 palatal k and p, as in (Sanscrit paO-ami) Greek jrell-ru, Latin (pop- 
 ina and) coQu-o, 
 
 Note H, page 83. 
 
 The Greeks separated their sign into f- and -/, its two elements, 
 to serve, the one of them for the Saaela, and the other for the V"^"?. 
 
134 APPEKDIX. 
 
 /- became >■ and ^ ; -/, similarly, -> and ^. Priscian (P. p. 1287) : 
 Quid est Dasea ? Flatilis, quae liac notatur figura/-. Quid est Psile ? 
 Siccitas, quse notatur sic, -j. The Latins evidently considered the 
 TJiOJi as practically equivalent to nothing. Cf. Phocas ; de Aspirat. 
 Segm II, ap. Lindeman. Omnia quae post i, a f uerint habentia, 
 cwrent aspiratione, ut Janus. So also : ordo ignorat aspiratiouem ; 
 Jlion negat dasiau. 
 
 Note I, page 87. 
 
 It must be noticed that the organic character or ancient title of 
 the H did not protect it even in early times from dissipation ; thus 
 mi for mihi corresponding to Sanscrit mahyam ; nil for nihil, from 
 nehilum. Hiltim probsibly = Jilum (and not as is commonly said 
 for pilus) lilce liarena corresponding to Sabine fasena ; made also 
 arena in Latin. In mdus (mdua) for hcedus (Quint.) the lost h was 
 organic substitute for g, as seen in Gothic gaits, old high German 
 gdz. (Cf. Corssen's Aussp., p. 47 ; and Lottner in Kuhn's Zeitsch. 
 f. Vergl. Sprachf., Vol. 7, p. 184, sqq.) 
 
 Note K, page 91. 
 
 No evidence of an L mouiUS in Latin has appeared to our investi- 
 gation ; but rather proof could be offered, if there were space for it, 
 that Latin L was in all places essentially the same. This much we' 
 may say, that the appearance of cdius compared with Greek u/l/lof, 
 and milia by the side of millia is no sign of the vdcalization of an L. 
 In the first case, it is the second X of (UAof which needs to be 
 accounted for, and not the i of alius', which is the more immediate' 
 representation of the Sanscrit y in any as; and in the second case, 
 there is no vocalization, but a falling away of I. Among the 
 instances of the doubling of a consonant, i. e,, of course, giving the 
 same consonant twice, Victorinus mentions I, without any mark of 
 a distinction between it and other consonants in this respect. The 
 doubling, namely, was employed simply to give strength to the 
 
APPENDIX. 135 
 
 consonantal element, or to fortify the length of the previous vovrel. 
 Victorinus (p. 2456) : Annius, LvcvijLva,:Memmino. So was relligio 
 sometimes written just as JuppUer, and puUua (from puerulus) 
 just Bspuppua. 
 
 ^ Note L, page 93. 
 
 This blunting, deadening, more or less complete suppression of 
 the closing M, may be fitly compared with the Sanscrit Anuswara, 
 which also had its origin in the M sound. The deadening of the 
 Latin M seems, however, to have gone farther than the Anuswfira, 
 since in its highest instance it involved the almost complete enerva- 
 tion and obliteration of the foregoing vowel ; so that, in place of 
 hoth consonant and vowel, there was often left only a sort of " after- 
 noise," or weak, irrational utterance, which is represented by e in 
 dice, etc., which the ancient Romans (in Cato's time) spoke and 
 wrote for dieam, etc. (Vid. Quintil. IX, 4, and Jordan's Catonis 
 ReUqq., -p. 90.) This e was probably Tiext to nothing, as it was also 
 in the imperative form dice, before this fell to die. (Vid. Nebv. 
 Eeliqq., ap. Ribbeck, p. 69.) But as the Anuswara made position 
 in poetry, so also the Latin final M, if it was followed in a verse by 
 a consonant, was restored to the proper functions of a consonant 
 (not clear M, but a consonant), while the vowel before it also 
 resumed its characteristic color and power. 
 
 Note M, page 107. 
 
 It is said that on a wall in Pompeii was found scribbled : 
 ALMA VILVMQ. CANO TLO 
 
 Note N, page 124 
 
 It must be admitted that the choice of ss instead of simple » for 
 representing s in the middle of a word seems to militate against 
 
136 A P P E K.D I X . 
 
 the soft sound of z which we contemplate. Whilst we do not feel 
 able to remove this difSculty satisfactorily, we can at least elimi- 
 nate it from the present discussion by remarking that the objection 
 found here will lie in all cases with equal strength, and in some 
 cases with much greater strength, against any of the other sup- 
 posed sounds of e, except the very one of sharp s itself. 
 
The J^ational Series of Standard School-jBooks, 
 
 MENTAL PHILOSOPHY. 
 
 Mahan's Intellectual Philosophy H 75 
 
 The subject exhaustively considered. The author has evinced learning, candor, 
 and independent thinking. 
 
 Mahan's Science of Logic 2 oo 
 
 A profound analyBis of the laws of thoqglit. The eystem possesses the merit gf 
 being intelli^hle and self consistent. In addition to the antbor's carefully elabo- 
 rated views, it embraces results attained by the ablest minds of Great Britain, Ger- 
 many, and Krance, in this department. 
 
 Boyd's Elements of Logic 1 35 
 
 A systematic and philosophic condensation of the subject, fortified with additions 
 from Watts, Abercrombie, Wbately, &c. 
 
 Watts on the Mind •' 50 
 
 The Improvement of the Mind, hy Isaac "Watts, is designed as a guide for the 
 attainment of useflil knowledge. As a text-book it is unparalleled; and the disci- 
 pline it affords cannot be too highly esteemed by the educator. 
 
 MORALS. 
 
 ■♦. » ■» 
 
 Peabody's Moral Philosophy 1 25 
 
 A shOit course ; by the Professor of Christian Morals, Harvard University— for 
 the Freshman Class and for High Schools. 
 
 Alden's Text-Book of Ethics 60 
 
 For young pupils.' To aid in systematizing the ethical teachings of the Bible, 
 and point out the coincidences between the instructions of the sacred volume and 
 the sound conclusions of reason. 
 
 Willard's Morals for the Young '?5 
 
 Lessons in conversational style to inculcate the elements of moral philosophy. 
 The study is made attractive by narratives and engravings. 
 
 GOVERNMENT. 
 
 Howe's Young Citizen's Catechism .... % 
 
 Explaining the duties of District, Town, City, County, State, and United States 
 OfilcerSj with rules for parliamentary and commercial ousinesa— that which every 
 fliture " sovereign " ought to know, and so few are taught. 
 
 Young's Lessons in Civil Government . . l 35 
 
 A comprehensive view of Government, and abstract of the laws showing the 
 rights, duties, and responsibilities of citizens. 
 
 Mansfield's Political Manual 1 25 
 
 This is a complete view of the theory and practice of the General and State Gov- 
 ernments of the United States, designed as a text-book. The author is an esteemed 
 and able professor of constitutional law, widely known for his sagacious utterances 
 in matters of statecraft through the public press. Eecent events teach with em- 
 phasis the vital necessity that the rising generation should comprehend the noble 
 polity of the American government, that they may act intelligently when endowed 
 with a voice In it. 
 
 39 
 
2%e' JVaiional Series of Standard SchoolSooks. 
 
 MODERN LANGUAGE. 
 
 ■ o ^ » 
 
 French and English Primer, $ lo 
 
 German and English Primer, lo 
 
 Spanish and English Primer, JO 
 
 The names of common objects properly illustratod and arranged in aas/ 
 
 lessons. 
 
 Ledru's French Fables, 75 
 
 Ledru's French Grammar, l oo 
 
 Ledru's French Reader, . . • «. l oo 
 
 The author*s long experience has enabled liinl to present tha most thor- 
 ongbly pracfcical text-books extant, in this branch. The systaia of pro- 
 nunciation (by phonetic illustration) is original with this author, and will 
 commend itself to all American teachers, as it enables their pupils to se- 
 cure aq absolutely correct pronunciation without the assistance of a natlra 
 mnster. This feature is peculiarly Taluable also to " self-taught" student:!. 
 The directions for ascertaining the gander of French nouns— also a great 
 Etumbling-block — are peculiar to' this work, and will be found remarkably 
 competent to the end proposed. The criticism of teachers and the test of 
 tUe school-room is invited to this excellent Gcrie3, with confideaco. 
 
 Worman's French Echo, i 25 
 
 To teach conversational French by actual practice, on an entirely new 
 plan, which recognizes the importance of the student learning to think in 
 the language which he speaks. It furnishes an extensive vocabulary of 
 words and expressions iii common use, aiid suffices to free the learner 
 from the embarrassments which the peculiarities of his own tongue aro 
 likely to be to him, and to make him thoroughly familiar with the use 
 of proper idioms. 
 
 Worman's German Echo, 12"! 
 
 On the saiue plan. See Worman's G-erman Series, paga 42. 
 
 Pujol's Complete French Class-Book, ■ • . 2 25 
 
 Offers, in one volume, methodically arranged, a complete French course 
 —usually embraced in series of from five to twelve books, including tho 
 tulky and expensive Lexicon. Here are Grammar, Conversation, and 
 choice Literature— selected from the best French authors. Each branch 
 is tlioronghly handled ; and the student, having diligently completed tha 
 course as prescribed, may consider himself, without further application, 
 au/ait in the most polite and elegant language of modern times. 
 
 Maurice-Poitevin's Grammaire Francaise, • 1 00 
 
 American schools are at last supplied with an American edition of thia 
 famous text-book. - Many of our best institutions have for years been pro- 
 curing it from abroad rather than forego tlie advantages it offers. Tha 
 policy of putting students who have acquired some proficiency from tha 
 ordinary text-books, into a Grammar written in the vernacular, can not 
 to too highljr commended. It affords an opportunity for finish and review 
 at once ; while embodying abundant practice of its own rules. 
 
 Joynes' French Pronunciation, 30 
 
 Willard's Historia de los Estados Unidos, • 2 00 
 
 The History of the United States, translated by Professors Tolon and 
 Db Toknas, will be found a valuable, instructive, and tatcrtaiaiJij read- 
 ing-book ioV gj'ttOl.'b Clti-dCS. , ^ 
 
The Ji/'ationat Series of Standard Sctiool-Sooks. 
 
 Pujol's Complete French Olass-Book. 
 
 TESTIMONIALS. 
 From I^op. Eixia Peissneb, Vnian CoUege. 
 I take great plcaeure in recommending Pujol and Van Norman's French ClaBS- 
 Book, as there is no Piencli grammar or class-book whlcU can be compared with 
 It in completeness, system, clearness, and general utility. 
 
 Ipr'om, Edwakd Noeth, PnsUent of HaimUton College. 
 I have caTPfiiily examined Pi\jol and Van Nonnan's French Claes-Book, and am 
 Batislled of Ita superiority, for college purposes, over any other heretofore Ufccd. 
 We shall pjt fail to use it with our nest class in french. 
 
 JVr»»»A. OintTis, Free' t of Cincinnati Literary and Scimtijlc Institute. 
 I am confident thqt it may he made an instrument in conveying to the student, 
 ht flrom els months to a year, the art of speaking and writing me French with 
 nlmoBt native fluency and propriety. 
 
 From HinAM Okoutt, A. M., Prin. Glenwood and TUden Zadiea' Seminaries. 
 
 1 have used Pujol's French Grammar in my two seminaries, esclueively, for 
 more than a year, and have no hesitation in saying that I regard it the beet text- 
 book id this department extant And my opinion is confirmed by the testimony 
 of Prof. F. De Launay and MademoiGcllc Marindin. They assure me that the 
 book is eminently accurate and practical, aa tested in the school-room. 
 
 ifVoin Pnos'. TnEO. F. Bn Fuhat, Scbrew Educational Institute, Memphis, Tenn. 
 M.Puiors French Grammar is one of the best and most practical works. The 
 French language is chosen and elegant in otyle— modern ard easy. It is far su- 
 perior to the other French class-boolio in this country. The selection of the con- 
 versational part is very good, and will interest pnpils ; and being all completed in 
 only one vofnme, it is especially desirable to have it introduced in our schools. 
 
 From Peop. Jahes H. Woeman, Bordentown Female CoUege, JV. J. 
 The work is upon the same plan as the text-books for the study of French and 
 English publishod in Berlin, for the study of thoEO who have not the aid of a 
 teacher, and these books are considered, by the first authorities, the best books. 
 In most of our institutions, Americana teach the modem languages, and hereto- 
 fore the trouble has been to give them a text-book that would dispose of the 
 difficulties of the French pronimcfation. This difficulty is succeBsfuIly removed 
 by P. and Van N., and I have every reason to believe it will soon make its way 
 into most of our best schools. 
 
 From Pbop. CnAELES S. Don, Ann Smith Academy, Lexington, Va, 
 I cannot do better than to recommend " Pujol and Van Norman." For compre- 
 hensive and systematic arrangement, progressive and thorough development of 
 all grammatical principles and idioms, with a due admixture or theoretical knowl- 
 edge and practical exercise, I regard it as superior to any (other) book of the kind. 
 
 From A. A. Fokstee, Prln. Finehurst School, Toronto, C. W. 
 I have great satisfaction in bearing testimony to M. Pujol's System of French 
 Instmction, as given in bla complete class-book. For clearness and comprehen- 
 siveness, adapted for eU classes of pupils, I hove found it superior to any other 
 work of the l^d, and have now used it for some years in my establishment with 
 great snccesa. 
 
 JVwn Peop. Otto Feddee, Maplewood Institute, Pittsfleld, Mass, 
 The conversational exercises will prove an immense saving of the hardest kind 
 of labor to teachers. There is scarcely any thing more trying in the way of 
 teaching language, than to rack your brain for short and easily intelligible bits 
 of conversation, and to repeat them time and again with no better result than 
 extorting at long intervals a doubting " oul," or a hesitating " non, monaieur 
 
 ^"For ftirther testimony of a similar character, Bea specla) ciroalar. and 
 current numbers of the Educational Bulletin. 
 
 41 
 
The JVaiional Series of Standard School-Hooks. 
 
 GERMAN. 
 
 A COMPLETE COUESE IN THE GEEMAH. 
 
 By JAMES H. WORMAN, A. M. 
 
 Worman's Elementary German Grammar -W 50 
 Worman's Complete German Grammar • 2 oo 
 
 These volumes are designed for intermediate and advanced classes respectively. 
 
 Though following the same general method with "Otto" (that of 'Gaspey'). 
 ©ur author differs essentially in its application. He is more practical, more sys- 
 tematic, more accurate, and hesides introduces a number of invaluable feature* 
 which have never before been combined in a German grammar. 
 
 Among other things, it may be claimed for Prof. Worman that he has been 
 the first to introduce in an American text-book for learning German, a system 
 of analogy and comparison with other languages. Our best teachers are also 
 enthusiastic about his methods of inculcating the art of speaking, of understanding 
 the spoken language, of correct pronunciation ; the sensible and convenient origi' 
 nal classiflcation of noims (in four declensions), and of irregular verbs, also de- 
 serves much praise. We also note the use of heavy type to indicate etymological 
 changes in the paradigms, »nd, in the exercises, the parts which specially illustrate 
 preceding rules. 
 
 Worman's Elementary German Reader . . i 25 
 Worman's Collegiate German Reader . . . a oo 
 
 The finest and most judicious compilation of classical and standard German 
 Literature. These works embrace, progressively arranged, selections from the 
 masterpieces of Goethe, Schiller, Komer, Seumfe, Uhland, Freiligrath, Heine, 
 ScUegel, Holty, Lcnau, Wieland, Herder, Lessing, Kant, Fichte, Schelling, Win- 
 kelmann, Humboldt, Bankc, Eaunier, Menzel, Gerrinus, &c., and contains com- 
 plete Goethe's " IpLigenie," Schiller's " Jungfran ;" also, for instruction in mod- 
 em conversational German, Benedix's "Eigeiisinn." 
 
 There are besides. Biographical Sketches of each author contributing, Notes, 
 explanatory and philological (after the text). Grammatical References to all lead- 
 ing griunmars, as well as the editor's own, and an adequate Vocabulary. 
 
 Worman's German Echo l 25 
 
 Consists of exercises in colloquial style entirely in (he German, with an ade- 
 quate vocabulary, not only of words but of idioms. The object of the system de- 
 Teloped in this work (and its companion volume in the French) is to break up the 
 laborious and tedious habit of translating the thougJite^ which is the student's 
 most effectual bar to fluent conversation, and to lead him to thinJc in the language 
 in which he speaks. As the exercises illustrate scenes in actual life, a considera- 
 ble knowledge of the manners and customs of the German people is also acquired 
 from the use of this manual. 
 
 Worman's German Copy-Books, % Numbers, each 15 
 
 On the same plan as the most approvecl syetems for English penmanshlpf with 
 progrcBBive copies, 
 
 43 
 
The JSTailonal Se7'ies of Siauda7'd School'!Books, 
 
 Worman's German Grammars. 
 
 TSSTIMOIf lAIiS. 
 
 From Prof. R. W. Jones, Pete,r&bv,rg Female College, Ya. 
 From -what I have seen of the -work it is almost certain / sJiall introduce it inb« 
 this institution. 
 
 From Prof. G. Campeeu,, Vniversity of Minnesota. 
 A 'Valuable addition to our school-books, and will find many friends, and do great 
 
 good. 
 
 From Prof. O. H P. Corpeew, Mary Military Inst , Md. 
 I am better pleased with them than any I havo ever taught I liave already orderpd 
 through our booksellers. 
 
 From Prof. E. S. Kendall, Vernon Academy, Conn. 
 I at once put the Elementary Grammar into tha hands of a class of beginners, and 
 have used it vyith great satisfaction: 
 
 From Prof. D. E. Holmes, Serlm Academy, Wis. 
 Worman's German works are superior. I shall use them hereafter in my German 
 classes. 
 
 From Prof. Magnus BnciiuoLTZ, I/ircem College, Ohio. 
 I have examined the Complete Grammar, and find It excellent. You may rely that 
 U will bo used here. 
 
 Front Prin. Tnos. "W. Tobey, Paducah Female Seminary, Ky. 
 The Complete German Grammar is worthy of an exteooive circulation. It is a<2- 
 tnirahly adapted to the class-room. I shall itsc it 
 
 From Prof. Alex. Eosbnspitz, Houston Academy, Texas. 
 Bearer will take and pay for 3 dozen copies. Mr. 'Worms.Q deserves the approbation 
 and esteem of the teacher and the thanks of the student. 
 
 From Prof. G, Malmkne, Augusta Seminary, Maine. 
 The Complete Grammar cannot fail to givs great satisfar.iion by tho simplicity 
 of its arrangement, and by its completeness. 
 
 From Prin. Oval Pikket, Christian University, Mo. 
 Just such a series as is positively necessary. I do hope the author will succeed as 
 well in the French, &c., as ho has in the German. 
 
 From Prof.- S. D. IIillman, Dickinson CoUr-ge, Pa. 
 The class have lately commenced, and my examination thus tar warrants me in say- 
 ing that I regard it as the best grammar for instruction in the German. 
 
 From Prin. Silas Liveumore, Sloomfleld Seminary, Mo. 
 I have found a classically and scientifically educated Prussi&,n gentleman whom I 
 
 S repose to make German instructor. I have shown him both your German grammars. 
 [e has expressed his approbation of them generally. 
 
 From Prof. Z. Test, ITowland School for Young Ladles, K. Y. 
 I shall introduce the books. From a cursory examination I have no hesitation in 
 pronotmcing the Complete Grammar a decided improvement on the text-books at 
 present in use in this country. 
 
 From Prof. Lewis Kistler, Northwestern Univa^sity, HI. 
 Having looked through the Complete Grammar with some care I must say that you 
 have produced a good hook; you may be r.warded with this gratification — that your 
 grammar promotes the facility of learning the German language, and of becoming 
 acquainted with its rich literature. 
 
 From Pres. J. P. Rous, Stock-well Collegiate Inat., Ind. 
 I supplied a class with the Elementary Grammar, and it gives complete sattttfao' 
 Hon. Tho conversational and reading exercises are well calculated to illustrate the 
 principles, and lead the student on an easy yet thorttugh coursB. I think tho Com 
 pletd Grammar equally attractive. 
 
 43 
 
JVationlal Series of Standard School- Soo&s. 
 
 THE CLASSICS. 
 
 LATIN". 
 Silber's Latin Course, $i 25 
 
 The book contains an Epitome of Latin Grammar, followed by Kcading ExcrcifleR, 
 vith explanatory Notes and copious References to the leading Latin Grammars, and 
 also to the Epitome wliich precedes the work. Then follow a Latin-English Vocabu- 
 lAry and Exercises in Latin Prose Composition, being thus complete iit itself, and a 
 very suitable work to put in the hands of oiio about to study the language. 
 
 Searing's Virgil's /Eneid, 2 25 
 
 It contains only tho first six books of the ^neld. 2. A' very carefully constructed 
 Dictionary. 3. Sufficiently copious Notes. 4. Grammatical reforehces to four lead- 
 ing Grammars. 5. Numerous Illustrations of tho highest order, G. A superb Map 
 of the Mediterranean and Adjacent countries. 7. Dr. S. II. Taylor's "Questions on 
 the ^neid." 8. A Metrical Index, and an E^ssoy on the Poetical Style. 9. A photo- 
 graphic fac simile of aii early Latin M.S. 13. The text according to Jahn, but para* 
 graphed according to Ladewig. 11, Superior mechanical execution* 
 
 Blair's Latin Pronunciation, ...... l 00 
 
 An inquiry into the proper Eounda of tlie Language during the ClaBsical Period. 
 By Prof. Blair, of Hampden Sidney College, Va. 
 
 Andrews & Stoddard's Latin Grammar, *i so 
 
 Andrews' Questions on the Grammar, • *o 15 
 
 Andrews' Latin Exercises, *i 25 
 
 Andrews' Viri Romae, *i 25 
 
 Andrews' Sallust's Jugurthine War, &c. *i so 
 
 Andrews' Eclogues & Georgics of Virgil, *l 50 
 
 Andrews' Caesar's Commentaries, . . ■ . *i 50 
 
 Andrews' Ovid's Metamorphoses, . . . *i 25 
 
 GREEK. 
 
 Crosby's Greek Grammar, 2 oo 
 
 Crosby's Xenophon's Anabasis, i 25 
 
 Searing's Home r's Iliad, • • • 
 
 MYTHOLOGY. 
 Dwight's Grecian and Roman Mythology. 
 
 School edition, §1 25; University edition, *3 00 
 
 A knowledge of the fables of antlqjity, thus presented in a Bystcmatic form, is as 
 ladispenBable to the student ofgeneral literature as to him vho would peruse intelli- 
 gently the classical authors. The mythological allusions so frequent in literaturo arc 
 readily understood with such a Key as tbia, 
 
 44 
 
The JVatlo»al Series of Standard SchoolSooks. 
 
 SEARING'S VIRGIL. 
 
 SPECIMEN FEAGMENTS OP LETTERS. 
 
 "I adopt it gladly."— Piaar. V. Dabket, Lmtdoun School, Va. 
 
 "I like Searing's Virgil."— Prop. Bkistol, Eipon College, Wis. 
 
 "Meets my desires very thoroughly."— Pkop. Ci^rk, Berea College, Ohio. 
 
 " Superior to any other edition of Virgil."— Pres. Haix, Macoa College, Mo. 
 
 " Shall adopt it at once." — Prik. B. P. Baker, Searcy Female Institute. ArJe. 
 
 "Tour Virgil is a beauty."— Yrot. W. H. Db Mottb, Illinois Female College. 
 
 "After use, I regard It the best."— Peik. G. H. Barton, Home Academy, N. T. 
 
 " We lilio it better every day."— Pri». E. K. Bdbhble, AUentown Academy, Pa. 
 
 " I am delighted vrith your Virgil."— Prin. W. T. Leonard, Pierce Academy, Mass. 
 
 " Stands well the test of class-room."— Prin. F. A. Chase, Lyons Col. Inst., Iowa. 
 
 "I do not see how it can be improved."— PRm. N. F. D. Browne, Chart. Hall, Md. 
 
 "The most complete that I have seen."— Pmn. A. Brown, Columbus High School, 
 Ohio. 
 
 " Our Professor of Language very highly approves."— Supt. J. G. James, Texas 
 Military J^tUute. 
 
 " It responds to a want long felt by teachers. It is beautiful and complete." — 
 Prop. Brooks, University of 3/Rnnesota. 
 
 " The ideal edition. We want a itew more classics of the same sort "— Prts. C. P. 
 P. Bancropt, Lookout Mountain Institute, Tenn. 
 
 *■ I certainly have never seen an edition so complete with important reqniBites for 
 a student, nor with such fine text and general mechanical execution." — Pres. J. K. 
 Park, University of Deseret, Utah. 
 
 "It is charming both in its design and execntion. And, on the whole, I think it 
 »s the best thing of the kind that I have seen."— Prop. J. Db F. Richards, Pres. 
 pro tern, of University of Alabama. 
 
 " In beauty of execntion, in jndicionB notes, and in nn adequate vocabulary, it 
 merits all praise. I shall recommend Its introdnctiou."^PRBS. J. K Patterson, 
 Kentucky Agricultural and Mechanical College, ^ 
 
 " Containing a good vocabulary and judicious notes, it will enable the industrious 
 student to acquire an accurate knowledge of the most interesting part of Virgil's 
 works."- Prop. J. T. Dunklin, Fait Alabama College. 
 
 *'It wants no element- of completeness. It is by far the best classical text-book 
 with which I am acquainted. The notes are just right. They help the etndeut 
 when he most needs help."— Prin. C. A. Bdukbr, Caledonia Grammar School, Vt. 
 
 "I have examined Searing's Virgil with interest, and find that it more nearly 
 meets the wants of students than Ihat of any other edition with which I am ac- 
 quainted. I am able to introduce it to some extent at once."— Prin. J. Easter, 
 Fast Genesee Conference Seminary. 
 
 " I have been wishing to get a sight of it, and it exceeds my expectations. It is 
 a beautiful book in every respect, and bears evidence of careftil and criticsl study. 
 The engravings add instruction as well as interest to the work. I shall recommend 
 it to my classes."- Prin. Chas. H. Chandler, Olenwood Ladies^ Seminary. 
 
 "A. S. Barnes & Co. have published an edition of the first six books of Vlreil's 
 .£neid, which is superior to its predecessors in several respects. The publishers 
 have done a good service to the cause of classical education, and the book deserves 
 a large circulation." — Prop. George W. Collord, Brooklyn Polytechnic, N. T. 
 
 " My attention was called to Searing's Virgil by the fact of its containing a voca- 
 bulary which would obviate the necessity of procuring a lexicon. But use iu tha 
 class-room has impressed me most favorably with the accuracy and just proportion 
 of its notes, and the general excellence of its grammatical suggestions. The gen- 
 eral character of the book in its paper, its typography, and its engravings is highly 
 commendable, and the fac-simile manuscript is a valuable feature. I take great 
 pleasure in commending the book to all who do not wish a complete edition of 
 Virgil. It suits our short school courses admirably."— Hbijby L. Boltwood, Master 
 of Princeton High School, III. 
 
 45 
 
The JVatlonal Series o/' Sta7idard School- Hooks. 
 
 RECORDS. 
 
 Tracy's School Record, '. .*lo 75 
 
 Tracy's Pocket Record, *65 
 
 For keeping a Bimple but exact record of Attendance, Deportment, and Scholar- 
 slip ; containing also a Calendar, an extensive list of Topics for Compositions and 
 CoUociuies, Tliemes for Short Lectures, Suggestions to Young Teachers, etc. 
 
 The pocl^et edition is of smaller size, with blanks on the same plan, for con- 
 venience of handling, etc. 
 
 Brooks' Teacher's Register, *i 00 
 
 Presents at one view a record of Attendance, Recitations, and Deportment for the 
 whole term. 
 
 Carter's Record and Roll-Book, *i 50 
 
 This is the most complete and convenient Eecord offered to the pnhlic. Besides 
 the nsual spaces fov General Scholarship, Deportment, Attendance, etc., for each 
 name and day^ there is a space in red lines enclosing six minor spaces in blue for 
 recording Recitations, 
 
 National School Diary, Per dozen, *i oo 
 
 A little book of blank forms for weekly report of the standing of each scholar, 
 from teacher to parent. A great convenience. 
 
 REWARDS. 
 
 National School Currency Per set,*li 50 
 
 A little box containing certificates in the foi-m of Money. The most entertaining 
 and stimulating system of school rewards. The scholar is paid for his merits and 
 fined for his shortcomings. Of course the most faithM are the most successftil in 
 business. In this way the use and value of money and the method of keeping 
 accounts are also taught. One box of Currency will supply a school of fifty pupils. 
 
 TACTICS, 
 
 The Boy Soldier, 75 
 
 Complete Infantry Tactics for Schools, with illustrations, for the use of those who 
 would introduce this pleasing relaxation from the confining duties of the desk. 
 
 46 
 
Mu Mdhmdiu. 
 
 And Only Thorough and Complete Mathematical Series. 
 
 IKT TI3:B,BDE3 F-A-HTS. 
 
 /. COMMON SCHOOL COURSE. 
 
 Davies' Primary Arithmetic— The Hmtlamental principles dlsplnj-ed In 
 
 Object Iiessons. 
 Davies' Intellectual Arithmotic— Heferring all operations to the nnit 1 03 
 
 the only tangible basis for logical development. 
 Davies' SSlements of Written Arithmetic!— A practical Introduction to 
 
 the whole subject. Theory subordinated to Practice. 
 Davies' Practical Arithmetic^*— The most successM combination of Theory 
 
 and Practice, clear, exact, brief, and comprehensive. 
 
 //. ACADEMIC COURSE. 
 
 Davies' University Arithmetic.*— Treating the subject exhaustively as 
 
 a science^ in a logical series of connected propositions. 
 Davies' Elementary Algrebra.*— A connecting link, conducting the pupil 
 
 easily from arithmetical processes to abstract analysis. 
 Davies' University Algebra.*— I"or institutions desiring a more complete 
 
 but not the fullest course in pure Algebra. 
 Davies' Practical IlXathematics.— The ecienco practically applied to the 
 
 useful arts, as Drawing, Architecture, Surveying, Mechanics, etc. 
 Davies' Elementary Geometry.— The Important principles In simple form, 
 
 but with all the exactness of vigorous reasoning. 
 Davies' Elements of Surveying.— Be-written is 1870. The simplest and 
 
 most practical presentation for youths of 13 to IS. 
 
 ///. COLLEGIATE COURSE. 
 
 ■Davies' Hourdon's Algebra.*— Embracing Sturm's Theorem, and a most 
 
 exhaustive and scholarly course. 
 Davies' University Algebra.*— A shorter course than Bonrdon, for Instltu- 
 (^ tions have less time to give the subject. 
 
 Davies' tegendre's Oeometry'.— Aclmowledged iAeore?^ satisfactory treatise 
 of its grade. 300,000 copies have been sold. 
 
 Davies* Analytical Geometry and Calculus.— The shorter treatises, 
 combined in one volume, are more available for Ameijcan courses of study. 
 
 Davies' Analytical Geometry. I The originid eompendiums, for those ds- 
 
 Davies' Diff. & £nt. Calculus. ' siring tp give iUl thne to each branch. 
 
 Davies' Descriptive Geometry.— With ajiplication to Spherical Trigonome- 
 try, Spherical Projections, and Warped Surlbc^^. 
 
 Davies' Shades, Shadovrs, and Perspective- A succinct exposition of 
 the mathematical principles involved. 
 
 Davies' Science of B(latheinaticSir»-^or teachers, embracing 
 
 I. Gbammab op Abithmktic, hi, Looio and TTthitt ov Mathxkatios, 
 
 n, OUTUHBS OF HATHEDIATIOS, IV. MaTHBMATIOAL DlOTXOHABT. 
 
 KXTB HAT BB OBTAIITXP FHOM THE FVBIiISHBBa 
 
 BY TEACHERS ONLY. 
 
"FOURTEEN WEEKS" III iML SCIENCE. 
 
 BT 
 
 Jf. GOHMAH STEELE, A.M. 
 
 NATURAL PHILOSOPHY, 
 ASTRONOMY. 
 
 CHEMISTRY, 
 
 GEOLOGY. 
 
 Tbesc volumes conBtitate the most availahle, practical, and attractive text-books on 
 the ScieuGea ever pablished. Bach volume may Be completed in a single term of study, 
 
 THE FAMOUS PRACTICAL QUESTIONS 
 devised by this author are alone sufficient to place his books in every Academy and 
 Grammar School of the land. These are questions as to the nature and cause of com- 
 mon phenomena^ and are not directly answered in the text, the design being to test 
 and promote an intelligent use of the student's knowledge of the foregoing prmciples. 
 
 TO MAKE SCIENCE POPULAR 
 la a prime object of these books. To this end each subject is invested with a charm- 
 ing interest by the peculiarly happy use of language and illustration is which this 
 author excels. 
 
 THEIR HEA VY PREDECESSORS 
 demand as much at the student's time for the acquisition of the principles of a single 
 branch as these for the whole course. 
 
 PUBLIC APPRECIA TION. 
 The author's great success in meeting an urgent, popular need, is indicated by the 
 faet (probably unparalleled in the history of scientific text-books), that although the 
 first volume was issued in 1867, the yearly sale is already at the rate of 
 
 I^ O E. T ^^■ T K O "CJ S -A- ISr 13 "V O t. XJ 3S«I El S - 
 
 PHYSIOIOCY AND HEALTH. 
 
 By EDWARD JARVIS, M.D. 
 9 ELEMENTS OF PHYSIOLOGY, 
 PHYSIOLOGY AND LAWS OF HEALTH. 
 
 The only books extant which, approach this subject "with a proper view of the trae 
 object of teaching Physiology in schools, viz., that Bcholars may know how to take 
 care of their own nealth. The child instructed from these works will be always 
 
 "^onsi^eii the Lilies." 
 
 BOTANY. 
 
 WOOD'S AMERICAN BOTANIST AND FLORIST. 
 
 Sliis new and eagerly expected work is the result of the author's experience and 
 life-long labors in 
 
 CLASSIFYING THE SCIENCE OF BOTANY. 
 He has at length attained the realization of his hopes by a wonderfully ingenious pro- 
 cess of condensation and arrangement, and presents to the world in this smgle moder- 
 Bte-Bized volume a COMPXjETB MANUAL. . 
 In 370 duodecimo pages he has actually recorded and defined 
 
 NEARLY ^,acx, SPECIES. 
 The treatises on Descriptive and Structural Botany are models of concise statement, 
 which leave nothing to be. said. Of entirely new features, the most notable are the 
 Synoptical Tables for the blackboard, and the distinction of species and varieties by 
 variation in the type.' 
 
 Fraf. Wood, by tnis work, establishes a just claim to his title of tlie great 
 
 AmERICAIV EXPONENT OP BOTANY. 
 
 Jlliwl 
 
w 
 
 MSi^ 
 
 Mil'