SECOND EDITION AUGMENTED

NEW YORK HENRY HOLT AND COMPANY 1894

Copyright, 1890, BY HENRY HOLT & Co.

CONTENTS.

PAGE

I.

INTRODUCTORY,

........

5

II.

SOURCES OF OUR INFORMATION,

........

7

III.

THE LATIN ALPHABET,

........

12

IV.

SOUNDS OF THE LETTERS,

........

15

V.

SOUNDS OF THE DIPHTHONGS,

........

31

VI.

A SHORT BIBLIOGRAPHY OF THE SUBJECT,

........

37

LATIN PRONUNCIATION.

I.

INTRODUCTORY.

THIS short manual is primarily intended for those who, being interested in the study of Latin, have accepted the Roman method of pronunciation upon the authority of the Grammars, but have either not been able to command the time to make themselves familiar with the arguments upon which this system is based, or have been repelled by the technicalities employed in treating the question from the standpoint of the specialist. It is believed that the following pages will be found to give in simple form the main facts bearing upon this interesting question; and that nothing has been introduced that is either unnecessary or obscure. For those who may wish to pursue their investigations farther after mastering these facts, a bibliography of the subject is given at the end.

The Roman method of pronouncing Latin has now received the approval of all Latinists of authority in Europe and America, as giving substantially the pronunciation employed by educated Romans of the Augustan Age. It has been formally adopted at our leading Universities. The most recent Grammars of the language recognize no other method. Thus, one great reproach to classical scholarship seems likely to be soon removed, and one universal pronunciation of the noblest of the ancient languages to receive general acceptation.[1] This little book will more than accomplish its object if it shall have aided ever so slightly in discrediting the barbarisms of a method which, to use the expression of a distinguished scholar, "ought long since to have followed the Ptolemaic system of astronomy into the limbo of unscientific curiosities."

[1]. It is natural that the Roman system should make its way more rapidly into use in this country than in Europe, not because Americans are more given to experiments, but because here in the United States the inconveniences of having no standard system have been more sharply felt. New England being wholly settled from Old England, long continued the English system of pronouncing Latin. In the Middle States, the Germans and Dutch introduced their own methods; in the South and West, the French pronunciation came in quite frequently; and all over the Union, the Catholic clergy in their schools and colleges have propagated the traditional usage of their Church. Hence a Babel of pronunciations and systems existing and practised side by side, in a picturesque confusion such as no European country ever knew; and hence the general willingness to accept a single method, especially one that is based upon historic truth.

II.

SOURCES OF OUR INFORMATION.

A QUESTION of much interest to the student of Latin, and one that does not always receive a satisfactory answer, relates to the sources of our information.

What knowledge have we of how the Romans pronounced their own language nineteen hundred years ago? How is it possible after so long an interval to reconstruct the laws of a pronunciation which prevailed at a given period of the remote past?

Briefly summarized, the sources of our information are six in number.

(1) Statements of the Roman writers themselves, which modern scholarship has laboriously collected. These are of different degrees of explicitness, and of different degrees of value. It is evident that a statement of Cicero, however brief, is more trustworthy and more convincing, with regard to the usage of his own time, than whole pages of testimony in a writer like Priscian who wrote in the sixth century, by which period the language had become corrupt.

We may, then, broadly divide the ancient authorities on this subject into two groups,—the first consisting of those writers who themselves belonged to the classical age; the second, of those grammarians and commentators who have left us very full statements, though the date at which they wrote somewhat impairs the value of their testimony.

The chief classical authorities to whom appeal can be made are M. Terentius Varro, a contemporary of Cicero, whose treatise on the Latin language has in part come down to us; Cicero himself, from whose rhetorical works one can gather many valuable facts; and M. Fabius Quintilianus, the author of the treatise Institutio Oratorio, in twelve books. It is not merely when these authors speak of definite points of language and pronunciation that they are valuable; sometimes a casual remark, an anecdote, or a pun, may be of very great importance, as will be seen from time to time in the following pages.

Of the other writers on language who treat the subject very minutely, a great number might be cited[1]. The most important are Terentianus Maurus, who wrote, perhaps about the third century, a poem on letters, syllables, feet, and metres, which is twice quoted by St. Augustine; Verrius Flaccus, the tutor to the grandchildren of the Emperor Augustus and author of a work on the meaning of words which has come down to us in a later abridgment; Aulus Gellius, who, toward the end of the second century, compiled a huge scrap-book on a variety of subjects, many of them of great linguistic interest, and, with the exception of a few chapters, still extant; Priscianus Caesariensis, who wrote under Justinian at Constantinople eighteen books of grammatical commentaries which form the most complete grammar of antiquity; and Aelius Donatus (A.D. 333), whose elementary treatise was so highly thought of in the Middle Ages that the name "donat" (Chaucer) was used as a generic term for a grammar.

From these and many other writers one gathers a great mass of instructive facts; and their very silence is sometimes as significant as what they say.

(2) The orthography of the language itself as seen in the inscriptions. Latin orthography was in the main phonetic (Quintilian, I. 7. 11). The language was pronounced as it was spelled. But as is always the case, changes in orthography lagged a little behind changes in the pronunciation. Hence even the blunders made by an ignorant lapidary in cutting an inscription are often a source of information to us.

(3) The representation in Greek letters of Roman sounds. A number of Greek writers treated of Roman history, Roman biography, and Roman geography. In so doing they were obliged to represent many Latin names and words in Greek characters. But many of these writers had no particular knowledge of the Latin language, and hence spelled these Latin names and words phonetically. Their method of doing this is both interesting and instructive. The writers of this sort who are oftenest cited are Polybius (B.C. 175), the friend of the younger Scipio and the author of a General History of Rome from the Second Punic War down to the conquest of Macedonia; Strabo the geographer (24 B.C.); Diodorus Siculus, the contemporary of Julius Caesar and author of an Historical Library in forty books; and Plutarch (A.D. 80), the best known of the Greek writers on Roman subjects[2].

(4) A critical comparison of all the modern languages of Europe that are derived from the Latin (Italian, French, Spanish, Portuguese) with reference to those points wherein they all agree. This source of information is of less importance than one would think, because these languages are not derived directly from the classical Latin, but from Latin that was either provincial or modified by foreign influences. Still, this comparison is useful in corroborating facts that are elsewhere learned, and is of positive value when not contradicted by other evidence.

(5) The traditions of scholars, and especially of the Roman Catholic Church, which in its rites has employed Latin continuously from the first century down to the present time. The rhymes of the early Christian hymns also have a bearing on this subject.

(6) The general principles of the science of phonology, which are now well established and understood, and are of great value in detecting erroneous assumptions which would otherwise pass unchallenged.

From these six sources can be gained a very accurate understanding of how Latin was pronounced in the days of Cicero and Caesar. It is not too much to claim that the system of pronunciation upon which scholars are now agreed, differs less from that of the Romans of the Augustan Age than does our modern pronunciation of English differ from that of Shakespeare and his contemporaries.

[It is not always remembered that only very gradually was the true pronunciation of Latin lost in Europe. Scholars long retained the essential features of it, and by the fact of their constant intercourse long prevented the growth of local and national variations from the established method. Great teachers like Erasmus passed from country to country, lecturing in Latin at the universities of Italy, Germany, Holland, Trance, and England, teaching pupils of all nationalities, and being everywhere understood without any difficulty, for Latin was the lingua franca of the educated, and one general pronunciation of it prevailed. Even in England, it was only after that country's isolation, political and religious, in the sixteenth century, that an "English pronunciation" arose, and this was long protested against, e.g. by Cardinal Wolsey, by Milton, and as late as the last century by Ainsworth (1746) and Philipps (1750). For the Continental traditions, see Justus Lipsius in his Dialogus de Recta Pronunciatione Linguae Latinae; and Erasmus,

De Recta Latini Graecique Sermonis Pronunciatione (Basic, 1528). In Scotland, the Continental sound of the vowels was long retained, on which see the incident imagined by Sir Walter Scott in his novel The Fortunes of Nigel, ch. ix.]

[1]. Schneider in his Elementarlehre der Lateinischen Sprache cites more than fifty ancient authors. Besides those mentioned above, reference is often made to Velius Longus, Servius, Marius Victorinus, Macrobius, and Martianus Capella.

[2]. Others are Josephus, the Jewish historian; Dionysius of Halicarnassus; Appian; and Dio Cassius,—the last a Roman who wrote in Greek.

III.

THE LATIN ALPHABET.

IN its earliest form, the Latin alphabet consisted of 21 characters,—A, B, C, D, E, F, Z, H, I, K, L, M, N, O, P, Q, R, S, T, V, X. These letters were derived from the alphabet used by the Dorian Greeks of Campania. At a very early period the letters K and Z fell into disuse, although K continued to occur in a few ancient abbreviations, such as Kal. for Kalendae, K. S. for carus suis, K. K. for calumniae causa (a legal phrase), KK. for castrorum, KA. for capitalis; and the use of Z was subsequently revived in transliterating Greek words. Originally, the character C had the sound which was afterwards given to G; but when K was abandoned, C took its place and its sound; while a new letter, G, was formed by slightly changing the original C. Plutarch says that the character G was first employed by Spurius Carvilius about the year 230 B.C. In Cicero's time the letter Y was introduced to represent the sound of the Greek Υ; but its presence in a word always marks a foreign origin, so that the character can scarcely be regarded as an

essential part of the Roman alphabet. About the year A.D. 44, the Emperor Claudius tried to introduce three new symbols into the alphabet: (1) the inverted diagamma to mark the consonantal sound of V; (2) the character known as "anti-sigma" to express the sound denoted by the Greek ψ (ps or bs); and (3) the sign , which was to have the sound of the Greek υ, i.e. of modern French u or German ü. It may be mentioned also, that consonants were not doubled in writing Latin until the practice was adopted from the Greek by Ennius (B.C. 239-169), who in various ways conformed Roman usages to those of the Greeks.

The Roman alphabet, like the early alphabet of the Greeks, lacked distinctive characters for the long and short vowels. This defect, which was partly corrected in Greek by the adoption of the letters η and ω (traditionally ascribed to Epicharmus of Syracuse, B.C. 500), was never fully remedied in Latin, though at different times various devices were employed to distinguish between ā and ă, ē and ĕ, ū and ŭ, ō and ŏ. These were:

(1) The doubling of the vowel when long, as in modern Dutch; thus, vootum = votum; aara = āra. This method was persistently used by the poet Attius[1].

(2) By the use of a species of accent (apex) over the long vowel. This became quite general in the Augustan Age.

(3) The length of the vowel ī was denoted sometimes by making it longer than the other letters and sometimes by writing it ei; thus, DICO, PVEREI.

The Roman numerals V, X, L, C, D, M originated in various ways[2].

V represented originally the open palm with the thumb extended, just as our 0 (zero) is thought to represent a closed hand.

X perhaps = , an old form of θ;; according to others, it is merely two V's placed together.

L = = or χ, a Greek letter which the Romans did not need in their alphabet and hence used only as a numeral.

C = , another form of θ, and confounded with C as though standing for centum.

M = φ, becoming first CI and then M, as though standing for mille, D is one half of this figure, or I[3].

[1]. Quintilian, I. 7, 14. When i is doubled it always denotes the consonantal i (j); e.g. maiior.

[2]. Cf. Ball's History of Mathematics, pp. 119, 120.

[3]. See, on the whole subject, Taylor, The Alphabet (London, 1883); Kirchhoff, Geschichte des Griechischen Alphabets (4th ed, Gütersloh, 1887); Berger, Histoire de L'Écriture dans l'Antiquité; (Paris, 1891); Cantor, Vorlesungen über die Geschichte der Mathematik (Leipzig, 1880); Martin, Les Signes Numéraux et l'Arithmétique chez les Peuples de L'Antiquité; (Rome, 1864); and Friedlein, Die Zahlzeichen (Erlangen, 1869).

IV.

SOUNDS OF THE LETTERS.

1. A: ā had the sound of a in English "far"; ă had the sound of a in English "trespass."

There is no disagreement of opinion regarding the proper pronunciation of Latin a. All the modern languages derived from the Latin practically agree in the sounds which they give to this character. Furthermore, its pronunciation is described for us by Terentianus Maurus (p. 328 in the edition by Keil); by Marius Victorinus (p. 32 in the edition of the same editor); and also by Martianus Capella (III. 261).

[NOTE.—It must be remembered in the pronunciation of the Latin vowels that the short vowel does not differ in quality from the corresponding long one, but only in quantity, i.e. it occupied less time in pronouncing. This is an important distinction between Latin and English.]

2. B: had in general the sound of English b; but before s or t, the sound of p.

(a) The ordinary sound of Latin b is described for us by Martianus Capella (III. 261); and by Marius Victorinus (p. 32 Keil).

(b) That it was sounded like p when it stood before s we know because very often in inscriptions it is so written, e.g. pleps for plebs; Araps for Arabs; urps for urbs. In certain verbs this usage has modified the common orthography, e.g. scripsi for scribsi from scribo; and opseguor for obsequor. And so before t, as we learn both by the spelling of certain words (optulit, scriptum); and from the statement of Quintilian (I. 7. 7): "When I pronounce the word obtinuit, our rule of writing requires that the second letter should be b: but the ear catches the sound of p."

3. C: always had the sound of English k.

The facts upon which this statement is founded are as follows:

(a) The pronunciation of this letter is so described for us by Martianus Capella (III. 261) as to prove it a hard palatal.

(b) C took the place of an original k in the early alphabet as previously stated; and in succeeding ages at times c reappears in inscriptions indifferently before the various vowels. Thus we have the form Caelius alternating with Kaelius, Cerus with Kerus, and decembres with dekembres,—showing that c and k were identical in sound. Quintilian (I. 7. 10) says: "As regards k, I think it should not be used in any words...This remark I have not failed to make, for the reason that there are some who think k necessary when a follows; though there is the letter C, which has the same power before all vowels."

(c) In the Greek transliteration of Latin names, Latin c is always represented by k; and in Latin transliteration of Greek names, k is always represented by Latin c. And we know that Greek k was never assibilated before any vowel. Suidas calls the C on the Roman senators' shoes, "the Roman kappa."

(d) Words taken into Gothic and Old High German from the Latin at an early period invariably represent Latin c by k; thus, Latin carcer gives the Gothic karkara and the German Kerker; Latin Caesar gives the German Kaiser; Latin lucerna gives the Gothic lukarn; the Latin cellarium gives the German Keller; the Latin cerasus gives the German Kirsche. Also in late Hebrew, Latin c is regularly represented in transliteration by the hard consonant kôph.

[Advocates of the English system claim that Latin c had the sound of s before e or i because every modern language derived from the Latin has in some way modified c when thus used. It is true that modern languages have so modified it; but, as already noted, the modern languages are the children not of the classical Latin spoken in the days of Cicero, but of the provincial Latin spoken five or six centuries later. There is no doubt that at this late period, Latin c had become modified before e or i so as to be equivalent to s or z. Latin words received into German at this time represent c before e or i by z. But had this modification been a part of the usage of the classical language, it would have been noticed by the grammarians, who discuss each letter with great minuteness. Now no grammarian ever mentions more than one sound for Latin c. Again, if Latin c had ever had the sound of s, surely some of the Greeks, ignorant of Latin and spelling by ear, would at least occasionally have represented Latin c by σ,—a thing which none of them has ever done. It is

probable that the modification of c which is noticed in the modern languages was a characteristic of the Umbrian and Oscan dialects and so prevailed to some extent in the provinces, but there is absolutely not the slightest evidence to show that it formed a part of the pronunciation of cultivated men at Rome.]

4. D: had regularly the sound of English d; but at the end of words nearly that of t.

(a) The position of the vocal organs in uttering this letter is described by Terentianus Maurus (p. 331 Keil); Marius Victorinus (p. 33); and Martianus Capella (III. 261).

(b) That final d was sounded like t is clear from the positive statements of Quintilian and from the fact that in inscriptions, as well as in the best manuscripts of Plautus and Vergil, we find almost indifferently ad and at, apud and aput, haud and haut, quid and quit, as well as adque and atque and many others.

[At about the fourth century A.D., di before a vowel began to be pronounced somewhat like the French j, just as in Aeolic Greek we find ζά for διά. Hence in the modern languages g and j arise out of Latin di. Compare Latin diurnus with the Italian giorno and the French jour.]

5. E: ē had the sound of English e in "they" or of the French ê; ĕ had the sound of English e in "net".

(a) The position of the vocal organs in pronouncing e is described by Terentianus Maurus (p. 329 Keil); Marius Victorinus (p. 32); and Martianus Capella (III. 261). It is regularly represented in Greek transliterations by ε when short, and by η when long.