Latin pronunciation

Latin pronunciation
by: Gn. Dionysius Draco

Introduction

Latin pronunciation is an uncertain field. No one has been able to revive an ancient Roman to see how he pronounced his language, and even so, there probably were a lot of regional varieties in Italia, and especially in the conquered provinces. Nevertheless, from the way the Romans wrote, and from comments of authors who were occupied with language and oratory or who commented on the pronounciation of their contemporaries, we can assemble a set of fairly commonly accepted rules. However, no manner is completely orthodox, and always has its flaws.

Take note of the following symbols: <..> for a grapheme (character) and [..] for a phonemic realisation (actual sound). Also remark that this document has been written mainly for the use of anglophones (as it's written in English), but that in order to understand and pronounce Latin, knowledge of other languages comes in very handily.

Alphabet

Since most Indo-European languages use an alphabet largely derived from the Roman or Greek one (although there are exceptions, such as Yiddish) the Roman alphabet is very easy to grasp by intuition. Their alphabet consisted of 24 characters:

A B C D E F G H I (K) L M N O P Q R S T V (X Y Z)

It has to be pointed out, that the <K, X, Y, Z> were marginal characters. The <K> was gradually replaced by the <C>, and the last three letters were loaned from Greek, and only appear in Greek (or other oriental) loanwords. Also note that the original Roman alphabet makes no distinction between and <V> and that the <J> was only added in medieval times; interesting to remark is also that the Romans did not employ small and capitalised characters as the Greeks did.

Consonants

The most important thing to note is, as in the majority of languages, one character does not always equal one sound, and one sound is not always represented by one character.

 as the voiced plosive [b] in all germanic languages, French and a host of other language.
 as the voiceless plosive [p] in English, French, Dutch, German and a number of other languages.
<F> as the fricative [f] in most languages.
<PH> was pronounced as an aspirated [p], id est, a normal [p] accompanied by a puff of air. This sound also exists in Etruscan and Greek, which both had a significant influence on Latin, former in earlier stages, latter in later eras. In Greek loanwords such as <physica>, the <PH> can also be pronounced [pf] or even [f]. In Greek dialects, there was a great variety in this.
<C> is pronounced [k] in all instances, and is a voiceless guttural stop.
<K> originally represented [k], but when <C> came to take this function over, the <K> only remained a marginal grapheme, and only occurs in words such as <Kalendae> or in writings of hellenised authors using Greek words (Greek only has the <k> and doesn't have a <c>)
<G> is pronounced as [g] as in the English word <goat>, and is a voiced guttural stop.
<Q> originally came from Etruscan and probably indicated a [k] pronounced with rounded lips, but eventually came to represent the [kw] as we know it today. Note that the grapheme <Q> is always followed by <V> (or ).
<CH> occurred in Greek and Etruscan loanwords, but also in genuine Latin words. In latter case, its pronounciation is probably an aspirated [k], as described under the <PH>. In case of loanwords such as <Christus>, the Latin pronounciation can be used, but better would be to use the [kx] or even [x]; the so-called "ach-Laut", a guttural fricative which does not exist in English, but does in a great many other languages.
<H> is pronounced as the [h] in most other languages, but was usually dropped by the lower classes, and in later times (by the end of the empire) it was gradually abolished altogether with the fall of the western Empire.
<M> as the labial nasal [m] in English and most other languages poses no problem.
<N> as the dental nasal [n] in English and most other languages poses no problem.
<NG> is pronounced as in the English word "thing". Beware not to voice the <G> of this consonant when another syllable follows.
<S> is pronounced as the English, French, Dutch or German [s].
<Z> was a marginal phoneme, and usually occurred in Greek loanwords such as <zephyrus>, and was pronounced [dz] or [ts], comparable to the German <z> in for instance <Zeitung>.
<T> poses no problem, and sounds like the English [t].
<TH> occurred in Greek loanwords such as <thorax> and is pronounced like the English "thorn".
<R> is a simple alveolar tap [r] as in modern day Italian and Spanish. It was never pronounced as a uvular [R] as in French, Danish or German.
<RR> is an alveolar trill [rr], which is actually an elongation of the consonant above. Most languages do not have this distinction, but Latin had.
<X> was also a marginal phoneme, and like the <Q> actually represents two sounds: [ks]. English speakers should be mindful that the <X> is also pronounced [ks] in the beginning of a word. It occurs in Greek loanwords and names such as <Xenophon>.
 was pronounced [j] as in English <yes> or Dutch <ja> when it preceded another vowel. In medieval Latin this was shown by replacing the character by <J>.
<V> was pronounced as the English <w> with strong lip-rounding. It was only pronounced this way if it preceded another vowel that was not the <V> itself. Most textbooks however make this distinction clear by using the to show where the actual <V> needs to be pronounced [u].

Vowels

The free vowels of Latin are relatively easy to deal with. Basic Latin had five vowel pairs, with the addition of another (marginal) pair in later times. All free vowels have short-long-opposition which is indicated by the [:] which means long and the [ ] which means short.

<A> is pronounced [a:] in stressed syllables and [a] in unstressed syllables, very much as in modern Italian, Dutch or German. Anglophones should try to isolate the first sound of the diphthong in "might" in order to produce this sound.
<E> has different forms of pronounciation, its most common forms being [e:] and [e] in most cases (same rules as with the <A> to discern between long and short), respectively as the first part of the diphthong in "face" and the short vowel in "dress" in English.
 was, depending on the stress, pronounced [i:] as in "see" or [i] as in "me" in English--supposing, of course, that it doesn't precede a vowel in the beginning of a word.
<O> was, depending on the stress, pronounced [o:] or [o] as in the first part of the diphthong in English "so" or as can be found in many North-English and Scottish dialects.
<V> was, depending on the stress, pronounced as [u:] in "goose" or [u] in "you" in English, but probably with more lip-rounding and backing, resembling an [o:] at times. Even in the upper class, it was a frequent spelling mistake of nobles to write <O> when they should have written <V>. Important to note is that a <V> followed by another vowel is usually pronounced [w]. In most Latin transcriptions, this will be shown by actually using a <v> and making the difference clear by using the for the real vowel.
<Y> represents a problem for people who speak English only. Phonetically, it is represented by [y:] or [y]. If you don't know how to form this sound, try to form a long [i:] with rounded lips instead of spread lips. With some practise, it should work. You can also ask advice from people who know German, French or another language which has this sound. The <Y> is a marginal vowel, however, and only occurred in Greek loanwords such as <cyclops>. It's very probable that the Romans themselves were often unable to correctly produce this sound.

Last to deal with are the diphthongs. These require special attention.

<AE> is not recognised as a diphthong by church Latin, and pronounced [e:], but is recognised that way by modern Latin pronounciation, effectively being the glide [ae]. The written sequence <AI> essentially has the same sound. It's important to get this correct, as it is a frequent sound in Latin.
<OE> occurs mostly in Greek loanwords such as <phoenix>, and was probably pronounced [oi] (where the "o" is as in "lot" and the "i" as in "kit").
<EI> is pronounced as [ei].
<UU> is not just a longer stress or duration of the [u] but effectively two separate vowels, seperated by a small stop (a glottal stop). It is not a real diphthong, but it's still a sequence of two sounds, its pronounciation being [uu].
<II> is the same, but with the [i] sound, so the result is [ii]. In both cases, the stress is put on the last vowel.