The Pronunciation of Latin

Faulty pronunciation can have serious consequences. When Hannibal wished to go to Casinum, his guide, misled by his pronunciation, brought him instead to Casilinum. Hannibal was now in danger of being trapped by the Romans, so, after flaying and crucifying the guide, he tied torches to the horns of 2,000 cows and drove them off into the night; having tricked the Romans into thinking that this was his army on the move, he led his men to safety in the opposite direction.

In our ignorance or neglect of ancient practice, Latin pronunciation tends to be influenced to a considerable extent by the phonology of the speaker’s native language, always with the proviso that Church Latin has given a certain degree of uniformity in Catholic countries.

Perhaps most significantly, the Romans’ known tendency to slur word-endings, or not to pronounce them at all, is all but universally ignored. This tendency, now particularly evident in French (e.g. Les châteaux sont magnifiques), contributed greatly to the development of the Romance languages, with their abandonment of so much of the inflection-system.

Nearly every professional Classicist educated in Great Britain has been trained to pronounce Greek as if it were Latin, even though the rules for vowel-length are even more clear-cut than in Latin and almost every word has its accent marked; some such scholars, who themselves pronounce, for example, the word for “market-place”, ἀγορά, incorrectly as ágora, as if the final vowel were short and the accent were on the first syllable, may even concede that the correct pronunciation agorá strikes them as slightly pretentious, not far short of “Paree” for “Paris” in an English sentence.

The Romans did not regard Latin as exactly the same in writing as in speech. Suetonius records that Augustus did not observe the rules for orthography established by grammarians, preferring to follow the opinion of those who thought that words should be spelled exactly as they are written (Life of Augustus 88). We do not know enough about the grammarians’ rules to assess them properly; even so, largely because writing had not been established in Rome long enough to permit radical discrepancies to evolve, classical Latin is simple to read and spell.

By contrast, because orthography has not kept pace with changes in the spoken language over a very long period, English has a rather challenging spelling system and would profit greatly from a systematic overhaul. (It is extremely difficult to implement spelling reforms, as the German attempt in 1996 has demonstrated.) Elbows rhymes with nose and toes, but not with eyebrows, intestine with chin, but not with spine, lung rhymes with tongue, breast with chest, gum with thumb, brain with vein and eye with thigh, while the vowels do not sound the same in blood and tooth, ear and heart (but compare artery); cf. also calf, knee, knuckle, muscle, palm, wrist.

The combination ee is written in many different ways: convene, debris, fiend, he, key, people, protein, quay, sardine, seem, ski, team.There are at least nine different pronunciations of the combination ough: bough, cough, dough, enough, hiccough, lough, ought, through, thorough. (The different words spelled slough, meaning a “mire” and “to shed a layer of skin”, rhyme with bough and enough respectively.) With f as in enough, i as in women, and sh as in station, fish may be spelled ghoti, a suggestion attributed to George Bernard Shaw. (This is perhaps no harder to accept than the fact that fish and the Latin piscis have a shared IndoEuropean ancestry.)

There is a markedly higher incidence of dyslexia among native speakers of English than among those who speak languages with a more regular orthography. Manx, the language of the Isle of Man, is perhaps the European language in which pronunciation and orthography are most at variance, the language itself being a descendant of Old Irish whereas the spelling system is mostly English. The last person to speak Manx as his first language died in 1974.

Finnish may be unusually difficult to acquire as a second language, but the extremely regular correlation between the written and spoken forms of words makes it very easy for Finnish children to learn to read, and Finland has a remarkably low level of dyslexia. It has been established that, not surprisingly, the brain expends more effort in reading words with an irregular spelling.

Since spoken and written Latin always differed, there will always have been discrepancies and difficulties in spelling. The Appendix Probi (wrongly attributed to the scholar who coined the phrase fetutinae grammaticae), a document written at an unfortunately indeterminable date between the 3rd and 7th centuries, attempts to maintain correct Latin orthography by listing some 227 common errors, and they are an invaluable illustration of the early development of the various Romance languages; for example:

In earlier Latin, s was frequently dropped at the end of a word after a short vowel and before a word beginning with a consonant, a sign that it was pronounced lightly if at all. In the late Republic, Lucretius and Cicero continue the practice in their poetry and Catullus has one instance, in the last line of his last poem, but, towards the end of his life, Cicero describes it as subrusticum “rather rustic” and it was not favored by later poets.

The elision of final m, however, when followed by a word beginning a vowel or the aspirate h, was standard at all periods. The early 6th century grammarian Priscian says that m at the end of a word is obscurum, at the beginning apertum (“open”), in the middle mediocre (“average”); the precise significance of these terms is opaque, but the general sense is clear.

Observing that no Greek words end in a μ (mu), Quintilian regrets that many Latin words end with an m, a letter which sounds like the mooing of a cow. He also laments the absence of upsilon and zeta, noting that Latin speech seems to shine more cheerfully when it borrows Greek words like ζέφυρoς (zephyros; “the west wind”) and ζώπυρoν (zopyron; “spark”). (Such opinions are, of course, largely subjective: in an early protest against Greek intellectual dominance [c. 300 BC], Appius Claudius Caecus had complained that pronouncing zeta made one grin like a dead person suffering from rigor mortis.)

The distinguished grammarian Verrius Flaccus, who was tutor to Augustus’ grandsons, suggested that, when m was to be elided, only the first half of the letter should be written. It seems reasonable to wonder how it would then differ from an n.

Other such enterprising, if sometimes rather eccentric, visual aids to comprehension are attested. When i was consonantal (= j) rather than vocalic, it might be written twice; e.g. iiam for iam (pronounced yam).

i and j, like u and v, were not fully distinguished in the writing of English till the 19th century. Whereas the only function of the first i of iiam was to indicate the value of the second i, the doubling in w has a more creative purpose. It originated in the 7th century, when the Latin alphabet had to devise a letter representing the sound w found in Germanic and Celtic languages. w is called “double u”, even though it is written as a double v, because Latin had not at that stage evolved a distinction between u and v. Despite the potential confusion with f, the “long s” (as in “profeƒsor”) evolved in Carolingian miniscule in the 9th century and persisted in English till the 19th century.

A long vowel might be written larger than a short one, or a circumflex or other such mark might be used to distinguish a long vowel; contrast e.g. malus “a bad man” and mālus “apple-tree”.

Although macrons are used throughout the textbook, standard modern texts of Latin authors do not give such guidance.

The practice of adding a superscript dot to i (and j) arose in the medieval period, in reaction presumably to the difficulty in distinguishing in the new cursive scripts between words such as minimum “very little” and nimium “too much”.

Before becoming emperor, Claudius devised three new letters, a backwards and upside-down f to represent consonantal u, a backwards c to replace bs and ps (as x replaces cs and gs), and a broken h, apparently representing the Greek upsilon. These letters appear occasionally on public inscriptions of the Claudian era, but were subsequently abandoned.

By no means all languages written with the Latin alphabet use the same 26 letters as does English; j, k, q, v, w, x, y and z are found in Irish only in words borrowed from other languages.

facile “easy” is spelled the same way in the original Latin as it is in English, French and Italian, but all four languages have different pronunciations:





q is always followed by u in Latin, as almost always in English and many other languages. Since u after q is not simply silent, but also unnecessary, having no effect on the pronunciation, it might without loss be systematically omitted, as in e.g. “qerulous”, “qirky”, “qixotic”, “qizzical”. (Roman grammarians observed that q and also k, which is found very rarely in classical Latin, could both be dispensed with, since their function is covered by c, which is never soft.)

Not only does Latin not have combinations difficult for the English speaker, it does not have combinations as difficult as those found in some English words of Germanic origin, such as sixth, eighth, twelfth. Many words borrowed from Greek, such as isthmus, ophthalmology, phlegm (but not phlegmatic), can also be challenging.

All of the few words beginning with schw- included in the OED are modern borrowings from German. The splendid German word Angstschweiss means “sweating because of fear”.

The combination vr is extremely rare in English. The word “very” came into English from Norman French vrai “true”, but restores the intervening e found in Latin verus. The French surname of Brett Favre (from the Latin faber = [black]Smith) is routinely pronounced “Farve” by football fans in Green Bay and elsewhere.

English allows many variations in vowel-length; e.g. bake, ball, balloon, bar, bat, bay. If a Latin vowel is not long, it is short. The length of the vowels in most Latin words is known, thanks to the strict metrical practices of classical poetry. There are very few positions in a line of verse where a poet had the option of using either a long or a short vowel, and a vowel could rarely be considered either long or short as required by the meter. For example, the a in sacer is short, but the combination of consonants following the a in sacrum etc. allows it to be considered either short or long; hence, for example, Tibullus has sacras with a short first a at 1.3.11 and 30, but sacram and sacra with a long first a in lines 18 and 25 of that same poem. Such flexibility is ignored in the marking of macrons in this course. The challenge of determining vowel-length is not a new one: St. Augustine complains more than once that he had difficulty with it, even though Latin was his mother tongue.

St. Augustine was not the most hard-working of pupils and vividly recollects the floggings which he had received at school. He never regarded faulty pronunciation as a serious flaw:

h in Latin is not so much a letter as an aspirate (a compound of the preposition ad and the verb spiro “breathe upon”) or rough breathing, in Latin spiritus asper. The OED s.v. aspirate observes that “Some writers have altered this word to asperate, after the spiritus asper of the Latin grammarians, an ingenious but unfounded conceit”. Catullus 84 is an amusing, if rather silly, epigram directed against a certain Arrius for his excessive aspiration. The grammarian Nigidius Figulus, a contemporary of Catullus, regarded misuse of aspiration as a sign of rustic speech.

u and v were not differentiated, with both serving to express both vocalic u and consonantal v: a scribe might write uacuum, adhering more closely to the original Greek letter upsilon (the Latin alphabet is adapted from Greek), whereas a stone-mason would find VACVVM easier to incise. Many modern texts continue to print uacuum, but vacuum, though rather a hybrid, avoids the problem of using the same letter both as a vowel and as a consonant, and this is the system which will be used in this course (e.g. vulpes uvas vulturis vult vorare “The fox wishes to devour the vulture’s grapes”); on the other hand, j is not much used nowadays for consonantal i.

In conventional orthography, the letter k is rarely found other than in Kaeso, a rare praenomen, Kalendae, -arum fem. 1 “the first day of the month” (our word “calendar”), Karthago, Karthaginis fem. 3 “Carthage”, and the unclassical rendering of the Greek κύριε ἐλέησον (kurie eleison “Lord, have mercy!”). In one of the oldest surviving copies of a Latin text, however, the fragmentary papyrus of the elegies of Gaius Cornelius Gallus, dating from the 20s BC, the name of a different Cato is spelled with a K. (K is almost never followed by any letter other than a.)

Isidore (Etymologies 1.18) notes that one reason why word-accents were devised was to ensure a clear distinction between words, for example, to prevent viridique in litore conspicitur sus (‘and on the grassy shore a pig is seen’) at Vergil Aeneid 8.83 is not read as viridique in litore conspicit ursus (‘and on the grassy shore a bear sees’).

There are a few known exceptions to the rules for accenting Latin words. Grammarians note that the archaic preposition ergo “for the sake of” and the adverb pone “behind” were distinguished from the conjunction ergo “therefore” and the singular imperative verb form pone “put!” by being accented on the final syllable.

Quintilian regretted the predictability of Latin accentuation in contrast to Greek, which was much more variable and musical.

These differences are especially obvious when Latin borrows Greek words. For example, the accent on φιλοσοφία (philosophia) is on the short penultimate syllable, of θησαυρός (thesaurós) on the final syllable; in adopting these words, writers of Latin seem to have been unable to decide whether to retain the original accent or make the words conform to the Latin system by accenting them as philosóphia and thesaúrus.

At Marriage of Philology and Mercury 3.272, Martianus Capella asserts that Greek words should be pronounced according to Latin rules unless they retain their Greek forms. This does not help with a word such as philosophia, which is identical in both languages except, crucially, for the length of the final vowel. A later grammarian ruled explicitly that philosophia should retain its Greek accent, but the only surviving occurrence of the word in verse, in a satire by Marcus Terentius Varro, shows the last syllable to be short, necessitating the accent philosóphia.

Variation and confusion in pronunciation is familiar in English also. Whether it means “head-band” or “slice of meat”, fillet in British English rhymes with “skillet”; in the latter sense in American English, however, it retains not only the French spelling, filet, but also the French pronunciation. On the other hand, in contrast to the American mústache, the British moustáche retains the French spelling and accentuation. AC Milan was founded in 1899 as the Milan Cricket & Football Club, but its name was changed by the Fascist government in 1938 to A(ssociazione) C(alcio) Milano; the English spelling was subsequently restored, but with the Italianized accentuation Mílan.

Since the Greek system was so flexible, the difference in sense between two quite different words spelled in the same way might be distinguished by the accent alone, and this could cause problems. When Euripides’ Orestes was first performed in 408 BC, Hegelochus, the actor in the title role, provoked much unlooked-for merriment when he mispronounced γαλήν’ (galén’) in line 279 as γαλῆν (galén): instead of “after stormy waves I once again see calm waters”, the audience heard him say “after stormy waves I once again see a weasel”. Modern editions of classical Greek texts have an accent marked on almost every word. This system is attributed to one man, Aristophanes of Byzantium, the head of the Library at Alexandria in the early second century BC; earlier writers, such as Thucydides, Sophocles and Plato knew little of written accents. In 1982, the Greek Ministry of Education drastically simplified the system.

It is perhaps natural to imagine the scholars in the Library and Museum at Alexandria as being intensely intellectual and desiccatedly devoted to their ivory-tower studies. (A contemporary epigram by Timon of Phlius describes the Alexandrian scholars as “overfed parasites continually squabbling in the Muses’ bird-cage” [Supplementum Hellenisticum 786].) It is therefore pleasant to note that Aristophanes himself had as his rival for the love of a girl who sold garlands (a prototype Eliza Doolittle) an elephant which every day brought her fruit and fondled her with its trunk (see e.g. Plutarch, On the Cleverness of Animals 18).