- W. Sidney Allen
- and Jonathan G. F. Powell
Our knowledge of the pronunciation of classical Latin is derived from a variety of sources. Most direct are the specific statements of Latin grammarians and other authors (though allowance must be made for the fact that the former tend to be of later date). Other sources are: puns, word-play, contemporary etymologies, and onomatopoeia; the representation of Latin words in other ancient languages; later developments in the Romance languages; the spelling conventions of Latin, and especially any deviations from these; the internal structure of Latin itself and of its metrical patterns (see grammar, latin; etymology).
It is impossible to reconstruct the vocal totality of a language spoken before the invention of sound-recording; but we can make a reasonably good approximation to the sounds of standard urban Latin as spoken around the turn of the era. It should be remembered that the pronunciation of Latin must have varied chronologically, socially, and geographically. In particular, the relatively static nature of the written medium in later antiquity may well have concealed significant changes in pronunciation.
Vowels and Diphthongs
Latin had five basic vowels—a, i, u, e, o—each of which could be either long or short. These may be described phonetically in terms of the degree of raising of the tongue and the part of the tongue involved. Thus a is low (or “open”) central, i is high (or “close”) front, u is high back, e is mid-front, and o is mid-back. The back vowels are also accompanied by lip-rounding. Whilst the short ă and the long ā were of similar quality, the other long vowels were distinctly higher than their short counterparts. In fact, in the Romance languages, the mid-long ē and ō developed in the same way as the high short ĭ and ŭ. The approximate values of these vowels in terms of modern languages are as follows: ă/ā as first/second vowels of Italian amare; ĭ/ī as in Eng. bit/beat; ŭ/ū as in Eng. put/fool; ĕ as in Eng. pet; ē as in Fr. gai or Ger. Beet; ŏ as in Eng. pot; ō as in Fr. beau or Ger. Boot.
In addition to the main short-vowel system, there is also evidence for an unstressed vowel of intermediate quality between ĭ and ŭ. In certain words, in middle syllables followed by a single labial consonant (u/v, p, b, f, and m), early Latin writes a vowel u, which later generally changes to i; for example, optumus → optimus. The earliest, inscriptional example of this change dates from 117 bce, and its official recognition is said to have been due to Caesar. It is likely that, at some time, both the u and the i written in such words were different from the normal vowels so written, most probably being more centralized. There is no good evidence that these spellings indicated, as sometimes stated, a fully front lip-rounded vowel like the German ü or ancient Greek υ. In any case, perhaps already in the classical period, it soon merged with the ĭ of the main system. The sound of the Greek υ was, however, adopted in loan-words from Greek and written with y, as nympha, etc.
The diphthongs ae (earlier written ai) and au had much the same values as those in Eng. high and how. Each of these consists of a glide from a low to a high vowel-quality, and in rural districts the two elements merged into (long) simple mid-vowels; in late Latin, this change became more general. Though represented in spelling as e and o, it is likely that they were of a more open quality then the normal ē/ō, as in Fr. tête and Eng. saw, since they developed in the Romance languages in a similar way to short ĕ and ŏ. The rarer diphthong oe was similar to that in Eng. boy; eu was pronounced as in Greek; the infrequent ui and ei were pronounced roughly as in “gooey” (not as in “we’) and “pay.”
The pronunciation of the Latin consonants calls for no general comment. They were mostly pronounced with the common values of the letters used, with a few exceptions. In particular, c and g were, in classical times, pronounced as voiceless/voiced velar plosives, ias in Eng. cap/gap or kit/get, even before the front vowels i and e, where later, by the process of palatalization, they developed first to “affricates,” as in Italian cento, or by a further process to fricatives, as in Fr. cent or Spanish ciento (the original value survives in Sardinian dialect kentu). The aspirated plosives ph, th, ch occur mostly, though not exclusively, in Greek words and are pronounced as in ancient Greek. In the group gn, as in dignus, it is likely that the g was nasalized, so that the group will have sounded like the ngn in Eng. hangnail.
In its consonantal value, the letter u (V) stood for a semivowel, like the English w; the change to a fricative (= Eng. v) took place around the end of the 1st century ce. The other semivowel, written as i (I), had the value of the English y.
In colloquial Latin, even in classical times, h tended to be dropped and has completely disappeared in the Romance languages. In initial position, it may be considered simply as a voiceless onset of the following vowel, and as such does not prevent elision of a preceding final vowel. Similar considerations apply to the labial nasal m in final position before a following initial vowel; the explanation here is presumably that the consonant developed simply into a nasalization of the preceding vowel (as in Fr. rien from L. rem).
Consonants written double were pronounced long: thus nn in annus as in Eng. unnamed (not as in manner).
For accentual and metrical purposes (see metre, latin), it is important to distinguish between length of vowels (long/short) and quantity of syllables conventionally also long/short, but here for clarity the terms heavy/light (adopted from Sanskrit grammar) are used. A heavy syllable is one containing either a long vowel (or diphthong) or a short vowel followed by more than one consonant; all other syllables are light. An exception to this rule is provided by groups of consonants comprising a plosive (p, t, c, b, d, g) followed by a liquid (r, l); in Plautus and Terence, and in spoken Latin, these groups were treated as single consonants for quantitative (and hence accentual) purposes. But in dactylic verse, they may also be treated as a normal group, in imitation of Greek models: thus in Aen. 2. 663, patris has a light first syllable but patrem a heavy.
Note that in the pronouns hic (nom. masc.) and hoc (nom./acc. neut.), the vowel is short; but in the latter, and after Plautus and Terence the former, c stands for cc, giving heavy quantity, before a following vowel (except by occasional archaism in the case of hic). But in ablative hoc and adverb hic, the vowel itself is long.
The placing of the classical accent was governed by quantity. By the so-called “Penultimate Law,” in words of more than two syllables, the accent falls on the penultimate syllable if this is heavy, but on the antepenultimate, regardless of quantity, if the penultimate is light: thus, for example, conféctus but confícere, cónficit. In disyllables the accent falls on the first.
In prehistoric times, there were clear indications that the accent fell on the first syllable throughout. Even in Plautus, one finds the scansion fácilius, which indicates that the process was not yet complete. In post-classical times, up to and including the modern era in some pronunciations, the accent did not change as such, but was replaced by a stress accent.
Grammarians, following Greek models, tend to describe it in terms of musical pitch, as in ancient Greek, but the detailed plagiarism of their descriptions and the very different placement rules in the two languages, as well as other internal and typological evidence in Latin, make this improbable. Admittedly, stress is often accompanied by variation in pitch, but this does not make pitch the primary feature.
In England the pronunciation of Latin has been particularly subject to change over the centuries, owing to influence first from Old English, then from Norman French, and lastly above all by the “Great Vowel Shift” of Early Modern English. The result was that from the 16th century, the English pronunciation of Latin was so far removed from the classical original and from other “national” pronunciations that it was virtually unintelligible elsewhere. A valuable source for this period is provided by the phonetic transcription of a Latin passage by the phonetician Robert Robinson in his Art of Pronuntiation of 1617. Despite the reforming efforts of Erasmus (De recta Latini Graecique sermonis pronuntiatione, 1528) and early support for these, particularly in Cambridge and in France, vested interests prevented their successful adoption; in England, it was not until the early 20th century that official approval, from Oxford and Cambridge Universities and the Classical Association, was secured for a new and generally acceptable reconstruction. The older English pronunciation survives mainly in legal Latin and in loan-words (e.g., genius). Ecclesiastical pronunciation tends to an Italianate style derived from the Roman Catholic Church. Other European countries have their own traditions of pronouncing Latin but, as in England, academic Latinists have moved increasingly toward a broadly agreed international norm.
- Sturtevant, Edgar H. The Pronunciation of Greek and Latin (2nd ed.). Westport, CT: Greenwood, 1977.
- Brittain, Frederick. Latin in Church: The History of Its Pronunciation (2nd ed.). London, U.K.: A. R. Mowbray, 1955.
- Bonioli, Maria. La pronuncia del latino nelle scuole dall'antichità al rinascimento. Turin, Italy: Giappichelli, 1962.
- Traina, Alfonso. L’alfabeto e la pronunzia del latino (2nd ed.). Bologna, Italy: Pàtron, 1963.
- Gordon, Arthur E. The Letter Names of the Latin Alphabet. Berkeley: University of California Press, 1973.
- Allen, W. Sidney. Accent and Rhythm: Prosodic Features of Latin and Greek. Cambridge, U.K.: Cambridge University Press, 1973.
- Allen, W. Sidney. Vox Latina (2nd ed.). Cambridge, UK: Cambridge University Press, 1978.