Traditional English pronunciation of Latin

The traditional English pronunciation of Latin, and Classical Greek words borrowed through Latin, is the way the Latin language was traditionally pronounced by speakers of English until the early 20th century.

Since the Middle Ages, speakers of English (from Middle English onward) have pronounced Latin not as the Romans did, but according to a traditional scheme borrowed from France. This traditional pronunciation became closely linked to the pronunciation of English, and as the pronunciation of English changed with time, the English pronunciation of Latin changed as well.

At the end of the 19th century, this Anglo-Latin pronunciation began to be superseded in Latin instruction by a reconstructed Classical pronunciation, closer to an earlier Roman pronunciation, and with a more transparent relationship between spelling and pronunciation. By the mid-20th century, classroom use of the traditional pronunciation had all but ceased. The traditional pronunciation, however, survives in academic English vocabulary:

Overview

In most cases, the English pronunciation of Classical words and names is predictable from the orthography, as long as long and short vowels are distinguished. For Latin, Latinized Greek or for long versus short α, ι, υ Greek vowels, this means that macrons and breves must be used if the pronunciation is to be unambiguous. However, the conventions of biological nomenclature forbid the use of these diacritics, and in practice they are not found in astronomical names or in literature. Without this information, it may not be possible to ascertain the placement of stress, and therefore the pronunciation of the vowels in English.

Note that the following rules are generalizations, and that many names have well established idiosyncratic pronunciations.

Stress placement

Latin stress is predictable. It falls on the penultimate syllable when that is "heavy", and on the antepenultimate syllable when the penult is "light".

(In Greek stress is not predictable, but it may be ignored when pronouncing Greek borrowings, as they have been filtered through Latin and have acquired the stress patterns of Latin words.)

A syllable is "light" if it ends in a single short vowel. For example, a, ca, sca, scra are all light syllables for the purposes of Latin stress assignment.

Any other syllable is "heavy":

Latin diphthongs may be written ⟨æ⟩ or ⟨ae⟩, ⟨œ⟩ or ⟨oe⟩. Long vowels are written with a macron: ā ē ī ō ū ȳ, though this is a modern convention. Greek long vowels are ει, η, ου, ω, sometimes ι, υ, and occasionally α. (Long α is uncommon.) For example, Actæon (also written Actaeon) is pronounced /ækˈtiːɒn/ ak-TEE-on or /ækˈtiːən/ ak-TEE-ən. A diaeresis indicates that the vowels do not form a diphthong: Arsinoë /ɑrˈsɪnoʊiː/ ar-SIN-oh-ee (not *AR-si-nee).

The importance of marking long vowels for Greek words can be illustrated with Ixion, from Greek Ιξιων. As it is written, the English pronunciation might be expected to be */ˈɪksi.ɒn/ IK-see-on. However, length marking, Ixīōn, makes it clear that it should be pronounced /ɪkˈsaɪ.ɒn/ ik-SY-on.

When a consonant ends a word, or when more than a single consonant follows a vowel within a word, the syllable is closed and therefore heavy. (A consonant is not the same thing as a letter. The letters x [ks] and z [dz] each count as two consonants, but th [θ], ch [k], and ph [f] count as one, as the pronunciations in brackets indicate.) The English letter j was originally an i, forming a diphthong with the preceding vowel, so it forces the stress just as æ, œ, z, and x do.

Secondary stress

If more than two syllables precede the stressed syllable, the same rules determine which is stressed. For example, in Cassiopeia (also Cassiopēa), syllabified cas-si-o-pei-a, the penult pei/pē contains a long vowel/diphthong and is therefore stressed. The second syllable preceding the stress, si, is light, so the stress must fall one syllable further back, on cas (which coincidentally happens to be a closed syllable and therefore heavy). Therefore the standard English pronunciation is pronounced /ˌkæsi.ɵˈpiːə/ KAS-ee-o-PEE-ə. (Note however that this word has the additional irregular pronunciation of /ˌkæsiˈoʊpiːə/ KAS-ee-OH-pee-ə.)

Long and short vowels in English

Whether a vowel letter is pronounced "long" in English (/eɪ, iː, aɪ, oʊ, juː/ ay, ee, eye, oh, you) or "short" (/æ, ɛ, ɪ, ɒ, ʌ/ a, e, i, o, u) is unrelated to the length of the original Latin or Greek vowel. Instead it depends on position and stress. A vowel followed by a consonant at the end of a word is short in English, except that final -es is always long, as in Pales /ˈpeɪliːz/ PAY-leez. In the middle of a word, a vowel followed by more than one consonant is short, as in Hermippe /hərˈmɪpiː/ hər-MIP-ee, while a vowel with no following consonant is long. However, when a vowel is followed by a single consonant (or by a cluster of p, t, c/k plus l, r) and then another vowel, it gets more complicated.

Regardless of position, stressed u stays long before a single consonant (or a cluster of p, t, c/k plus l, r), as in Jupiter /ˈdʒuːpɨtər/ JEW-pi-tər.

Note that in many dialects a syllable followed by r tends to be closed regardless of position, and while the long/short distinction described above is maintained, the r has its own effect on the vowel, as in Elara /iːˈlɛərə/ ee-LAIR-ə (a long but closed syllable ending in r).

Alphabet

Anglo-Latin (hereafter A-L) includes all of the letters of the English alphabet except w, viz.: a b c d e f g h i j k l m n o p q r s t u v x y z. It differs from Classical Latin in distinguishing i from j and u from v. In addition to these letters the digraphs æ and œ may also be used (as in Cæsar and phœnix). These two digraphs respectively represent mergers of the letters ae and oe (diphthongs, in Greek and Latin) and are often written that way (e.g. Caesar, phoenix). However, since in A-L both ae and oe represent a simple vowel, not a diphthong, the use of the single letters æ and œ better represents the reality of A-L pronunciation. Despite being written with two letters, the sequences ch, ph, rh, th represent single sounds. The letter x, on the other hand, usually behaves like a sequence of two sounds (being equivalent to cs).

Conversion of Greek to Latin

A-L includes a large amount of Greek vocabulary; in principle, any Greek noun or adjective can be converted into an A-L word. There is a conventional set of equivalents between the letters of the Greek and Roman alphabets, which differs in some respects from the current mode of Romanizing Greek. This is laid out in the tables below:

Vowels Diphthongs
Greek letter α ε η ι ο υ ω αι ει οι υι αυ ευ ου
Romanization a e ē i o u ō ai ei oi ui au eu ou
Conversion to Latin a e e i o y o æ i œ yi au eu u
Consonants
Greek letter ʻ β γ γγ γκ γξ γχ δ ζ θ κ λ μ ν ξ π ρ ρρ σ
ς 		τ φ χ ψ
Romanization h b g gg gk gx gch d z th k l m n x p r rr s t ph ch ps
Conversion to Latin h b g ng nc "nx" nch d z th c l m n x p r
rh		 rrh s t ph ch ps

Rh is used for Greek ρ at the beginnings of words, e.g. ῥόμβος (rhombos) > rhombus. Rarely (and mostly in words relatively recently adapted from Greek), k is used to represent Greek κ. In such cases it is always pronounced [k] and never [s] (as it might be if spelled c) : e.g. σκελετός (skeletos) > skeleton not "sceleton".

Greek accent marks and breath marks, other than the "rough breathing" (first in the list of consonants above), are entirely disregarded; the Greek pitch accent is superseded by a Latin stress accent, which is described below.

Frequently, but not universally, certain Greek nominative endings are changed to Latin ones that cannot be predicted from the tables above. Occasionally forms with both endings are found in A-L, for instance Latinized hyperbola next to Greek hyperbole. The most usual equations are found below:

Endings
Greek ending -η -εια -ον -ειον -ος -ρος
after a consonant 		-ειος
Romanization -ē -eia -on -eion -os -ros -eios
Latin ending -a -ea
-ia 		-um -eum
-ium 		-us -er -eus
-ius

Examples:

Consonants

Letters and sounds

Phonemes

The underlying consonantal phonemes of A-L are close in most respects to those of Latin, the primary difference being that /w/ and /j/ are replaced in A-L by /v/ v and /dʒ/ j. The sound /θ/ th was borrowed from Greek.

Phonemes of A-L Labials Interdentals Alveolars Palatals Velars Glottals
Stops voiceless /p/ /t/ /k/
voiced /b/ /d/ /ɡ/
Affricate (voiced) /dʒ/
Fricatives voiceless /f/ /θ/ /s/ /h/
voiced /v/ /z/
Nasals (voiced) /m/ /n/
Approximants (voiced) /r/
/l/

Consonantal allophones

Greek consonant clusters

Several word-initial clusters, almost all derived from Greek, are simplified in A-L by omitting the first consonant:

In the middle of words both consonants in these clusters are pronounced (e.g. Charybdis, Patmos, Procne, prognosis, amnesia, apnœa, synopsis, cactus, captor); medial chth and phth are pronounced /kθ/ and /fθ/ respectively, as in autochthon and naphtha.

Polyphony

The letters c, d, g, h, n, s, t and x have different sounds (phonemes) depending upon their environment: these are listed summarily below.

Letter c d g h n s t x
Underlying sound /k/ /d/ /ɡ/ /h/ /n/ /s/ /t/ /ks/
Primary phonemes /s/ /dʒ/ ∅ /ŋ/ /z/ /s/ /z/, /ɡz/
Secondary phonemes /ʃ/ /dʒ/ /ʃ/
/ʒ/ 		/tʃ/
/ʃ/ 		/kʃ/

The full set of consonantal phonemes for A-L is almost identical to that of English, lacking only /ð/.

Sounds of A-L Labials Interdentals Alveolars Palatals Velars Glottals
Stops voiceless /p/ /t/ /k/
voiced /b/ /d/ /ɡ/
Affricates voiceless /tʃ/
voiced /dʒ/
Fricatives voiceless /f/ /θ/ /s/ /ʃ/ /h/
voiced /v/ /z/ /ʒ/
Nasals /m/ /n/ /ŋ/
Approximants /w/ /ɹ/
/l/ 		/j/
Miscellaneous environments

Environments that condition the appearance of some of these phonemes are listed below:

Sound affected Spelling Environment Resulting sound Examples
/h/ h between a preceding stressed and a following unstressed vowel ∅ cf. "vehement, annihilate"
after x exhibitor
/n/ n before velars /k/ (c, ch, k, q) and /ɡ/ g /ŋ/ incubator, fungus
/s/ s between two vowels /z/ miser, Cæsar, Jesus
between a vowel and a voiced consonant plasma, presbyter
after a voiced consonant at the end of a word lens, Mars
/ks/ x initially /z/ Xanthippe
in the prefix ex- before a vowel or (silent) h in a stressed syllable /ɡz/ exemplar, exhibitor

The change of intervocalic /s/ to /z/ is common but not universal. Voicing is more common in Latin than in Greek words, and never occurs in the common Greek ending -sis, where s is always voiceless: basis, crisis, genesis.

Palatalization

The most common type of phonemic change in A-L is palatalization. A-L reflects the results of no less than four palatalization processes. The first of these occurred in Late Latin, the second in Proto-Gallo-Romance, the third and fourth within the history of English. While the first two palatalizations are universally used in variants of A-L, the third and especially the fourth are incompletely observed in different varieties of A-L, leading to some variant pronunciations.

Some of the occasions on which palatalizations 3 and 4 fail to take effect should be noted:

Summary

Palatalization Sound affected Spelling Environment Resulting sound Examples
1 /t/ t when not initial, following s, t, or x, and before the semivowel i /s/ annunciator (from annuntiator)
/s/ usually changes to /ʃ/ by Palatalization 3
2 /k/ c before front vowels e, æ, œ, i, y /s/ circus, census, Cynthia, foci, proscenium, scintilla, successor
/ɡ/ g /dʒ/ Gemini, regimen, algæ, fungi, gymnasium
3 /s/ c, t
(sc, ss) 		when not initial, before semivowel i and e /ʃ/ acacia, rosacea, species, inertia, ratio
fascia, cassia
/ks/ x /kʃ/ cf. "complexion"
/t/ t /tʃ/ cf. "question, Christian, bestial, Attius"
/z/ s /ʒ/ Asia, ambrosia, nausea, Persia
4 /d/ d when not initial, before (usually unstressed) open u /ju/, /jə/ /dʒ/ educator, cf. also gradual
/s/ s, ss /ʃ/ cf. "censure, fissure"
/ks/ x /kʃ/ cf. "luxury"
/t/ t /tʃ/ spatula
/z/ s /ʒ/ cf. "usual"

See further the section on the "semivowel" below.

Degemination

Following all of the above sound changes except palatalizations 3 and 4, "geminate" sequences of two identical sounds (often but not always double letters) were degeminated, or simplified to a single sound. That is, bb, dd, ff, ll, mm, nn, pp, rr, ss, tt became pronounced /b d f l m n p r s t/. However, for the purposes of determining whether a syllable is open or closed, these single consonants continue to act as consonant clusters.

Other notable instances involving degemination include:

The following combinations, derived from Greek, are also pronounced as single consonants:

Syllables

The simple vowels of A-L (a, æ, e, ei, i, o, œ, u, y) can each have several phonetic values dependent upon their stress, position in the word, and syllable structure. Knowing which value to use requires an explanation of two syllabic characteristics, openness and stress.

Openness

Openness is a quality of syllables, which may be either open, semi-open, semi-closed, or fully closed.

Fully closed syllables

Fully closed syllables are those in which the vowel in the middle of the syllable (the vocalic nucleus) is followed by at least one consonant, which ends or "closes" the syllable. Vowels in fully closed syllables appear:

Semi-closed syllables

Semi-closed syllables are formerly closed, unstressed syllables that became open due to the merger of two following consonants of the same sound. For the purpose of determining vowel reduction in initial unstressed syllables they count as open.

Semi-open syllables

Semi-open syllables are formerly closed, unstressed syllables that are followed by a sequence of consonants that can stand at the beginning of a syllable. Since instances of obstruents+r or l are already considered open, semi-open syllables are practically restricted to instances of s+obstruent, bl, and in some cases perhaps tl. Vowels in initial semi-open syllables may be treated as open for all purposes except for determining the value of u, which is still closed in semi-open syllables.

See further the section on initial unstressed syllables below.

Open syllables

Open syllables are those in which the nucleus is followed:

Stress

Primary stress

Stress is another characteristic of syllables. In A-L, it is marked by greater tension, higher pitch, lengthening of vowel, and (in certain cases) changes in vowel quality. Its exact concomitants in Classical Latin are uncertain. In Classical Latin the main, or primary stress is predictable, with a few exceptions, based on the following criteria:

Primary stress can therefore be determined in cases where the penult is either closed or contains a diphthong. When it contains a vowel that may have been either short or long in Classical Latin, stress is ambiguous. Since A-L does not distinguish short from long vowels, stress becomes a lexical property of certain words and affixes. The fact that decorum is stressed on the penult, and exodus on the antepenult, is a fact about each of these words that must be memorized separately (unless one is already familiar with the Classical quantities, and in the former case, additionally with the fact that decus -ŏris n. with short -o- syllable became in late Latin decus/decor -ōris m. with long -o- syllable: 'Dómine, diléxi decórem domus tuæ').

Secondary stress

Secondary stress is dependent upon the placement of the primary stress. It only appears in words of four or more syllables. There may be more than one secondary stress in a word; however, stressed syllables may not be adjacent to each other, so there is always at least one unstressed syllable between the secondary and primary stress. Syllables containing semivowel e or i are never stressed.

Secondary stress in words with three or more syllables before the primary stress is less predictable. Such words include those of five syllables with penult primary stress, and all words of six syllables in length or longer. The following generalizations about such long words may be made:

Unstress

Unstressed syllables are all others. They are always adjacent to a stressed syllable; that is, there can never be more than two unstressed syllables in a row, and that only when the first one follows a stressed syllable.

Semivowel

Several sound-changes in A-L are due to the presence of the "semivowel", an alteration of certain front vowels. Originally ordinary vowels, they acquired at different points in history the value of the glide /j/ (a y-sound like that in English canyon). Subsequently, their value has fluctuated through history between a consonant and a vowel; the term "semivowel" thus reflects the intermediate historical as well as phonetic position of this sound. The environment in which the semivowel was produced was as follows:

  1. The vowel was e (æ, œ), i (ei), or y.
  2. The vowel came immediately before a vowel or diphthong.
  3. The vowel was not in the initial syllable: e, æ, ei, i and y in rhea, mæander, meiosis, fiat, diaspora, hyæna, did not become semivowels.
  4. The vowel was unstressed: e, æ, œ, ei, i in idea, Piræus, diarrhœa, Cassiopeia, calliope, elephantiasis did not become semivowels.

Examples of words where e, i, y became semivowels include: miscellanea, chamæleon, nausea, geranium, rabies, Aries, acacia, ratio, fascia, inertia, halcyon, polyanthus, semiosis, mediator, Æthiopia, Ecclesiastes.

The effects of the semivowel include the following:

  1. Though always in hiatus with a following vowel, semivowel i and y are never pronounced like long i or y (e.g. /aɪ/); historically semivowel e could also be distinguished from "long e" (formerly [ɛː] or [eː]). In current varieties of A-L, semivowels are pronounced in a variety of ways:
    • Most frequently as /i/: labia, radius, azalea, præmium, cornea, opium, Philadelphia, requiem, area, excelsior, symposium, Cynthia, trivia, trapezium. In British Received Pronunciation, the prescribed pronunciation was once /ɪ/.
    • In some dialects or registers of English as /j/, e.g. junior pronounced [ˈdʒuːnjər].
    • Merged with a following -es or -e ending, as in Aries, scabies.
    • They are usually deleted following the palatals /ʃ/, /ʒ/, and /dʒ/: Patricia, consortium, Persia, nausea, ambrosia, Belgium.
    • Occasionally a semivowel is retained after a palatal sound: ratio, sometimes Elysium. This type of pronunciation is an artificiality, as the sounds /ʃ/ and /ʒ/ resulted from an absorption of the original /j/ in the sequences /sj/, /zj/. The pronunciations with /ʃi/ and /ʒi/ result from a re-introduction of the i sound to conform with the spelling. This pronunciation was, however, recommended by academics, and as such is common in the pronunciation of A-L phrases such as ab initio, in absentia, venire facias.
  2. The consonant t changed to /s/ and then to /ʃ/ before the semivowel arising from i: minutia, inertia, nasturtium.
  3. The sibilants /s/ (including ss, sc, c, and t) and /z/ (usually spelled s) are usually palatalized before the semivowel:
    • /s/ > /ʃ/: cassia, fascia, species, militia
    • /z/ > /ʒ/: amnesia, ambrosia
  4. The vowels a, e, æ, and o in an open antepenult syllable become long if a semivowel appears in the next syllable:
    • radius, Asia, azalea, area
    • anæmia, chamæleon
    • genius, medium, interior
    • odium, cochlea, victoria
This lengthening takes place regularly in antepenultimate syllables. It is less regular in syllables further back. On the one hand, there are words that do seem to lengthen before a semivowel in the next syllable:
  • Æthiopia, Ecclesiastes, mediator, negotiator, variorum.
On the other hand, some words have short vowels:
  • gladiator, apotheosis, Meleagrus, polyanthus (and other words containing poly- followed by a vowel).
In general, those words with lengthened vowels in pre-antepenult syllables before a semivowel in the next syllable are those that are derive from a word with a regularly lengthened vowel in an antepenult syllable, e.g., Æthiopia from Æthiops ("Ethiopian"), Ecclesiastes from ecclesia ("church"), mediator from medium, negotiator from negotium ("business'), variorum from varius ("manifold"). The failure of gladiator (from gladius, "sword") to have a long vowel is anomalous.

Vowels

Mergers

The most notable distinction between A-L and other varieties of Latin is in the treatment of the vowels. In A-L, all original distinctions between long and short vowels have been obliterated; there is no distinction between the treatment of a and ā, etc., for instance. However, the subsequent development of the vowels depended to a large degree upon Latin word stress (which was preserved nearly unchanged in the mediæval period), and as this was in part dependent upon vowel length, in certain cases Latin vowel length contrasts have been preserved as contrasts in both stress and quality. However, the immediate governing factor is not length but stress: short vowels that were stressed for various reasons are treated exactly like stressed long vowels.

In addition to the merger of long and short vowels, other vowel mergers took place:

The merger of æ and œ with e was commonly recognized in writing. Sometimes forms written with æ and œ coexist with forms with e; in other cases the form with e has superseded the diphthong in A-L. Consider the following:

The following words are usually spelled with e, though they originally had æ:

In other cases, particularly names, the forms with the diphthongs are the only correct spelling, e.g. ægis, Cæsar, Crœsus, Œdipus, onomatopœia, pharmacopœia, Phœbe, phœnix, Piræus, sub pœna.

The sequences ei, æi, œi (distinguished in writing and pronunciation from ej, the vowel followed by a consonant, as in Sejanus) are sometimes retained in spelling preceding a vowel. In such cases the sequence is invariably pronounced as a simple vowel, sometimes i (as in meiosis, pronounced as if miosis) but more usually e: Cassiopeia, Deianira, Pleiades, onomatopœia, pronounced as if Cassiopea, Deanira, Pleades, onomatopea.

The result was a system of five vowels, a, e, i, o, u. These would subsequently split, according to their environment, into long, short, and (eventually) unstressed variants; and these variants would eventually also be altered in various dialects of A-L dependent upon neighboring sounds. However, in phonemic terms, A-L still has only five vowels, with multiple allophones.

In addition, there were the diphthongs, ai, oi, ui, au and eu. Of these, ai and au eventually monophthongized, eu merged with the open variant of u, and yi merged with the "long" i. Only oi and ui remained as true diphthongs, but both are extremely rare.

Realizations of a, e, i and o

The vowels a, e, i, o have three primary phonemes: long, short, and reduced; each of these may, in turn, have allophonic variations based on their phonetic environment, including whether they are stressed; in an open or closed syllable; their position in the word; and neighboring consonants. One of the most common environmental alterations of a vowel is due to the presence of a following r. Such vowels are called "r-colored".

Short vowels

This is the default value for vowels, observed:

  1. In closed monosyllables
  2. In stressed closed penult syllables
  3. In all antepenult syllables, open or closed, which receive primary stress, except for those lengthened due to a following semivowel
  4. In all syllables with secondary stress
  5. In fully closed unstressed syllables which immediately precede, but do not follow, a primary or secondary stress (usually in the first syllable of a word), with exceptions for certain prefixes

All short vowels have variants colored by a following r sound when the r is followed by a different consonant (not r) or by the end of the word. In addition, there is a variant of short a that only appears after a /w/ sound – chiefly found in the sound qu /kw/.

Short vowels 17th c. American British Australian Type 1 Type 2 Type 3 Type 4 Type 5
a /æ/ ɐ æ æ æ pax mantis, pallor, malefactor camera, marathon, calculus anæsthesia, saturnalia antenna, magnificat
ar /ɑr/ ɐr ɑɹ ɑː aː par, Mars argus, catharsis arbiter, Barbara arbitrator, pharmacopœia narcissus, sarcophagus
e /ɛ/ ɛ ɛ ɛ e rex sector, error, præceptor, interregnum Gemini, Penelope memorandum, impedimenta pentathlon, September, spectator
æ /ɛ/ quæstor Æschylus, diæresis prædecessor, æquilibrium
œ /ɛ/ Œdipus
er /ɜr/ ɛr ɝ ɜː ɜː per vertex, Nerva terminus, hyperbola perpetrator Mercator, persona
i /ɪ/ ɪ ɪ ɪ ɪ nil isthmus, lictor, cirrus, narcissus simile, tibia, antithesis, Sirius, delirium simulacrum, administrator, hippopotamus scintilla, dictator
y /ɪ/ lynx, Scylla, Charybdis chrysalis, synthesis, Thucydides, Syria symbiosis hysteria
ir /ɜr/ɪr ɚ ɜː ɜː circus, Virgo Virginia
yr /ɜr/ thyrsus myrmidon
w-colored a /ɒ/ ɔ ɑ̹ ɒ wɔ quantum
o /ɒ/ non impostor, horror optimum, conifer, metropolis propaganda, operator October, thrombosis
w-colored ar /ɔr/ ɔr ɔ̹ɹ ɔː ɔː quartus
or /ɔr/ cortex, forceps formula cornucopia torpedo

Exceptionally, monosyllables ending in es are pronounced with the rhyme /iːz/, e.g. pes, res. This pronunciation is borrowed from that of -es used as an ending.

Exceptions to the pronunciation of short y generally involve prefixed elements beginning with hy- in an open syllable, such as hydro- and hypo-; these are always pronounced with a long y, e.g. hydrophobia, hypochondria. This pronunciation is the result of hypercorrection; formerly they were pronounced with a short /ɪ/, as is still the case in the word "hypocrite" and (for some speakers and formerly commonly) hypochondria.

Prefixes may also behave in anomalous ways:

  1. The prefix ob- in unstressed syllables may be reduced to /əb/, even when it closes a syllable: cf. "obsession, oblivion".
  2. The Greek prefix en-, em- in a closed unstressed syllable may be reduced to /ɨn/, /ɨm/: encomium, emporium.
  3. The prefix ex- in an unstressed syllable may be reduced to /əks/, /əɡz/, despite always being in a closed syllable: exterior, exemplar.
  4. The prefix con-, com- is reduced to /kən/, /kəm/ when unstressed: consensus, compendium, regardless of whether the syllable is closed or not.
  5. The preposition and prefix post(-) is anomalously pronounced with "long o": /poʊst/: post-mortem and cf. "postpone"; also thus in words in which post was originally a preposition (postea, postquam) but not in other derivatives, being pronounced with short o in posterus, posterior, postremo, postridie.

Long vowels

Long vowels are those that historically were lengthened. By virtue of subsequent sound changes, most of these are now diphthongs, and none is distinguished by vowel length—however, the term "long" for these vowels is traditional. "Long" vowels appear in three types of environments:

  1. a, e, i and o are long in an open monosyllable
  2. a, e, i and o are long in a stressed open penult syllable
  3. a, e and o are long when in an open syllable followed by semivocalic i and e
  4. a and o are long when they precede another vowel in hiatus; i and e are long in the same environments, but only when they are not semivocalic (i. e., when they are in the initial syllable or receive primary stress). Hiatus may be original, or may arise from the deletion of h between a stressed and unstressed syllable
"Long" vowels 17th c. American British Australian Type 1 Type 2 Type 3 Type 4
a /eɪ/ aː ei eɪ æɪ a, qua crater, lumbago radius, rabies chaos, aorta, phaëthon
ar /ɛər/ e ɛː eː pharos area, caries
e /iː/ eː i i i e, re ethos, lemur, Venus genius idea, creator
æ /iː/ Cæsar anæmia, chamæleon æon, mæander
œ /iː/ amœba, Crœsus diarrhœa
ei /iː/ Deianira, Pleiades
er or ær /ɪər/ i ɪə ɪə serum, Ceres, æra
ɪ bacterium, criterion, materia
i /aɪ/ ɛi ai aɪ ɑe i, pi item, Tigris, saliva, iris, horizon (i remains short, e.g. trivia) miasma, hiatus, calliope
y /aɪ/ hydra, python, papyrus (y remains short, e.g. Polybius) hyæna, myopia
o /oʊ/ oː ɞu əʊ əʉ O, pro bonus, toga odium, encomium, opprobrium boa, Chloe, cooperator
or /ɔər/ ɔ ɔː o chorus, forum, thorax emporium, euphoria

Reduced vowels

Reduced vowels appear in unstressed syllables, except for:

Initial unstressed syllables

A variety of possible realizations are available for open, semi-open, and semi-closed initial unstressed syllables, including (for e and i) long, short, and reduced variants. Fully closed initial unstressed syllables are always short.

Open and semi-open unstressed vowels
in absolute initial position		 17th c. American British Australian Examples
a /ə/ ɐ ə ə ə amœba, anemone, ascesis
e /ɨ/e i, ɪ, ə i, ɪ, ə i, ɪ, ə Elysium, emeritus, epitome, erotica
æ /ɨ/ ænigma
œ /ɨ/ œsophagus
i /aɪ/ɛi ai aɪ ɑe idea
y /aɪ/ hyperbola, hypothesis
o /ɵ/ o ɞu, ɞ, ə əʊ, ə əʉ, ə Olympus
Initial-syllable open/semi-open
unstressed vowels		 17th c. American British Australian Examples
a /ə/ ɐ ə ə ə papyrus, placebo, saliva, basilica
e /ɨ/e ɪ, ə ɪ, ə ɪ, ə December, thesaurus
æ /ɨ/ Mæcenas, pæninsula, phænomenon
i /aɪ, ɨ/ɛi, ɪ ai, ɪ, ə aɪ, ɪ, ə ɑe, ɪ, ə criteria, tribunal, minutiæ, cicada
y /aɪ, ɨ/ lyceum, psychosis, synopsis, chrysanthemum
o /ɵ/ o ɞu, ɞ, ə əʊ, ə əʉ, ə November, rotunda, colossus, proscenium

The variation in the value of the initial open unstressed vowel is old. Two different types of variation can be distinguished; the older use of a "long" vowel for i, y, o (and their variants); and more recent variations in the value of the reduced vowel.

No completely general rule can be laid down for the appearance of an initial unstressed long vowel, although such vowels must have appeared before the shortening of geminate consonants, as they are restricted to fully open syllables. The most general tendency is for long vowels to appear when i and y are either preceded by no consonant or by h, e.g. idea, isosceles, hyperbola, hypothesis. The prefixes in and syn never have long vowels: inertia, synopsis. I and y also tend to be short when the next syllable contains an i or y, short or long: militia, divisor.

O is a little less likely to appear with a long value in this location; or, at any rate, it is harder to distinguish the long value from the reduced vowel.

Unstressed e and i in open syllables had merged by the early 17th century; their reduced reflex is often transcribed [ə], but by many speakers is still pronounced as a high front lax vowel, distinct from the [ə] derived from a. For such speakers, the first syllables in Demeter and Damascus are pronounced differently. The sound is not identical to the short vowel [ɪ], but is more central: [ɪ̈], and here transcribed as /ɨ/

Unstressed o, also often transcribed [ə], is by many speakers pronounced with considerable lip-rounding: [ɞ], here transcribed /ɵ/.

Semi-closed initial
unstressed vowels		 17th c. American British Australian Examples
a /ə/ ɐ ə ə ə addendum, appendix, calliope, farrago
e /ɛ, ɨ/ ɛ ɛ, ə ɛ, ə e, ə ellipsis, Ecclesiastes, erratum
i /ɪ, ɨ/ ɪ ɪ, ə ɪ, ə ɪ, ə Illyria, cf. cirrhosis
y /ɨ/ syllepsis
o /ə/ ɔ ə ə ə collector, oppressor, opprobrium, possessor

The partially closed initial unstressed vowels began as short vowels, but were later reduced.

These are the same sounds as in the preceding chart, but without the option of the "long" vowels and much less rounding of the o.

proscenium does not fall in this group, apparently because felt to be pro+scenium.

Medial unstressed syllables

All vowels in medial unstressed syllables are reduced to /ə/ or /ɨ/, regardless of whether they are in open or closed syllables.

Medial unstressed vowels American British Australian Examples
a /ə/ ə ə ə diabetes, emphasis, syllabus, diagnosis, melancholia
e /ɨ/ ə ɪ, ə ə impetus, phaethon, malefactor, commentator, Alexander
i /ɨ/ ə ɪ, ə ə animal, legislator
o /ə/ ə ə ə hyperbola, demonstrator
y /ə/ ə ɪ, ə ə platypus, analysis, apocrypha
Vr /ər/ ɚ ə ə interceptor, superficies

Open and closed u

The pronunciation of the letter u does not depend upon stress, but rather upon whether the syllable in which it appears is open or closed. There are no "long" and "short" variants of either type of u, but there are reduced and r-colored variants of both types.

Open u

The underlying sound of open u is /juː/; it shares developments with the homophonous diphthong eu, which can however appear in closed syllables.

The sound [j] in /juː/ and its variants is deleted in various environments:

After the following consonants when they precede u in an initial, final, or stressed syllable:

In some dialects, particularly of American English, /j/ is deleted after the following consonants when they precede u in an initial, final or stressed syllable:

/j/ is not deleted in the following environments:

Unstressed open u may retain rounding in some positions for some speakers, varying between [ʊ] and [ə] ([jʊ] and [jə]). It is here transcribed /ʉ/ or /jʉ/.

Open u
Environment Examples with /j/ Examples without /j/
/juː/ in stressed syllables jɵu humor, uterus, tribunal, euthanasia ɵu rumor, verruca, junior, Jupiter
/jʊər/ in stressed syllables, r-colored jɵ furor ɵ juror
/jʉ/ in unstressed initial syllables jɵ musæum, urethra, euphoria, eureka ɵ superior
/jʉ/ in medial unstressed syllables jɵ, jə calculus, nebula ɵ, ə spatula
/juː/ in unstressed final syllables jɵu ɵu impromptu, situ, passu
/jʉ/ in unstressed hiatus jɵw, jəw amanuensis, innuendo

Closed u

Closed u appears only in closed syllables, except for instances of the prefix sub- before a vowel. It has reduced and r-colored variants, as shown below. r-coloration only appears when the r is followed by a different consonant (not r) or the end of the word.

Closed u
Environment American British Australian Examples
/ʌ/ in stressed syllables ʌ ɐ a sulfur, alumnus, ultimatum
/ɜr/ in r-colored stressed syllables ɚ ɜː ɜː laburnum, murmur, præcursor
/ʌ/ in initial fully closed unstressed syllables ʌ ɐ a ulterior
/ə/ in initial open or semi-closed unstressed syllables ə ə ə suburbia, curriculum
/ə/ in medial unstressed syllables ə ə ə illustrator
/ər/ in all r-colored unstressed syllables ɚ ə ə murmur, sequitur, saturnalia

Diphthongs

Diphthongs in A-L are distinguished from simple vowels by having no long or short variants, regardless of position or syllable type. The only diphthongs that are at all common are au and eu. For variations in the pronunciation of the latter, see Open u. Au is, rarely, reduced in an unstressed syllable to [ə]: Augustus pronounced as if "Agustus". Such words may be pronounced with the full value of the diphthong, however.

Diphthongs American British Australian Examples
ai /eɪ/ ei eɪ æɪ Achaia, Maia, Gaius
au /ɔː/ ɒ ɔ o aura, pauper, nausea, autochthon, aurora, glaucoma, mausoleum
eu /juː/ ju, u ju jʉ neuter, euthanasia, zeugma
oi /ɔɪ/ ɔi ɔɪ oɪ coitus, paranoia
ui /(j)uː.j/* (j)ui (j)uɪ (j)uɪ alleluia, cuius
yi /aɪ/ ai aɪ ɑe harpyia

Note that ui is generally disyllabic, as in fruc.tu.i, va.cu.i, tu.i. The monosyllabic words cui and huic were traditionally pronounced /kaɪ/ and /haɪk/.

Endings

The pronunciation of the final syllables of polysyllabic words do not always correspond to what might be expected from the constituent phonemes. Some endings also have more than one pronunciation, depending upon the degree of stress given to the ending.

Three types of endings can be distinguished:

Vowel alone

The first class consists of vowels alone, i.e. -a, -e, -æ, -i, -o, -u, -y. In this class, the vowels are generally long, but -a is always /ə/.[1]

Letter American British Australian Examples
a /ə/ ə ə ə circa, fauna, mania, quota
e /i, iː/ i ɪ i ante, epitome, posse, simile
æ /i/ algæ, larvæ, vertebræ
i, y /aɪ/ ai aɪ ɑe alibi, Gemini, moly
o /oʊ/ ɞu əu əʉ ego, Pluto, torpedo
u /juː/ (j)u ju jʉ (in) situ

In British Received Pronunciation, -e and -æ were once transcribed as /ɪ/, but now agree with most other varieties of English with /i/. Words deriving from Greek long -ê (η) end in /iː/ unless assimilated.

In the words mihi, tibi, sibi, by an old tradition, the final i was pronounced like final e above (i.e., as if spelled mihe, tibe, sibe).

A late and purely academic pronunciation distinguished final -ā from -a by pronouncing the former like "long a", /eɪ/: for instance, Oxford professor A. D. Godley rhymed "day" with Rusticā. That this was not the usual pronunciation can be told from such forms as circa, infra, extra, in absentia, sub pœna, all of which have an originally long final vowel: circā, sub pœnā, etc. This use is distinct from the older tradition (in use in the 17th-18th centuries) had made all final a's "long", regardless of their Latin length.

Vowel plus consonant cluster

The second class consists of vowels consonant clusters such as ns, nt, nx, ps, x. In this class, the vowels are always short, except for u, which may be reduced to [ə].

Letter American British Australian Examples
a /æ/ æ æ æ climax, phalanx
e /ɛ/ ɛ ɛ e biceps, index
i /ɪ/ ɪ ɪ ɪ matrix, phœnix
o /ɒ/ ɑ ɒ ɔ Cyclops
u /ə/ ə ə ə exeunt, Pollux
y /ɪ/ ɪ ɪ ɪ pharynx, oryx

Vowel plus consonant

The third class consists of vowels followed by the consonants l, m, n, r, s, t. The treatment of these endings is inconsistent. Generalizations include:

  1. All vowels are reduced before final r for /ər/: Cæsar, pauper, triumvir, Mentor, sulfur, martyr.
  2. All vowels are reduced to /ə/ before l: tribunal, Babel, pugil, consul.
  3. Except sometimes before t, a is reduced to /ə/ before any of this class of consonant: animal, memoriam, titan, atlas.
  4. All instances of u are reduced to /ə/ before any of this class of consonant: consul, dictum, locus.

The remaining endings are: -at, -em, -en, -es, -et, -im, -is, -it, -on, -os, -ot. Of these, -em, -im, -is, -it, -on, -ot have two possible pronunciations, one with a short vowel and one with /ə/. Final -es and -ies are alike pronounced /iːz/.[2] Final -eus may be pronounced containing the diphthong eu, juːs, or as if it were semivowel e followed by the ending -us, i.əs. However, even when pronounced as two syllables, -eus counts as a single syllable for the purpose of determining vowel length; that is, the syllable preceding the ending is considered the penult.

Ending American British Australian Examples
at
/æt/, /ət/		 æt æt æt magnificat
ət ət ət fiat
em
/ɛm/, /əm/ 		ɛm ɛm em idem, ibidem
əm əm əm item, tandem
en /ən/ ən ən ən lichen, semen
es /iːz/ iz iz iz Achilles, appendices, fæces
ies /iːz/ rabies, species
et /ɛt/ ɛt ɛt et videlicet, scilicet, quodlibet
eus
/juːs/, /i.əs/		 (j)ɵus (j)us (j)ʉs Perseus, Nereus
iəs iəs iəs
im
/ɪm/, /əm/		 ɪm ɪm ɪm passim
əm əm əm interim
is
/ɪs/, /ɨs/ 		ɪs ɪs ɪs ægis, crisis, hypothesis
əs əs əs
it
/ɪt/, /ɨt/		 ɪt ɪt ɪt exit
ət ət ət deficit
on
/ɒn/, /ən/ 		ɑn ɒn ɔn icon, marathon
ən ən ən bison, siphon, horizon
os /ɒs/ ɑs ɒs ɔs chaos, pathos, pharos
ot
/ɒt/, /ət/		 ɑt ɒt ɔt aliquot
ət ət ət

This last pronunciation of -os is the expected one; however, in the masculine accusative plural, where the ending is historically -ōs, the academic prescription was the pronunciation /oʊs/.[3] Such an ending is not found in English loan words or proper names.

Sample text

O Fortúna, velut luna, statu variabilis, semper crescis aut decrescis;

diaph: /ˈoʊ fɔrˈtjuːnə ˈviːlət ˈluːnə ˈsteɪtjuː ˌvɛəriˈæbɨlɪs ˈsɛmpər ˈkrɛsɪs ˈɔːt dɨˈkrɛsɪs/

Amer: [ˈɞu fɔɹˈtɪunə ˈvilət ˈlɵunə ˈsteitɵu ˌveɹiˈæbəlɪs ˈsɛmpɚ ˈkɹɛsɪs ˈɒt dɪˈkɹɛsɪs]

Brit: [ˈəʊ fɔːˈtjunə ˈvilət ˈlunə ˈsteɪtɪu ˌvɛːɹiˈæbəlɪs ˈsɛmpə ˈkɹɛsɪs ˈɔt diˈkɹɛsɪs]

Aust: [ˈəʉ fɔːˈtʃʉnə ˈvilət ˈlʉnə ˈstæɪtʃʉ ˌveɹiˈæbəlɪs ˈsempə ˈkɹesɪs ˈot diˈkɹesɪs]

vita detestabilis nunc obdúrat et tunc curat ludo mentis aciem,

diaph: /ˈvaɪtə ˌdɛtɨˈstæbɨlɪs ˈnʌŋk ɒbˈdjʊəræt ɛt ˈtʌŋk ˈkjʊəræt ˈluːdoʊ ˈmɛntɪs ˈeɪʃiɛm/

Amer: [ˈvaitə ˌdɛtəˈstæbəlɪs ˈnʌŋk əbˈdɪuɹæt ˈɛt ˈtʌŋk ˈkjɵɹæt ˈlɵudɞu ˈmɛntɪs ˈeiʃiəm]

Brit: [ˈvaɪtə ˌdɛtəˈstæbəlɪs ˈnɐŋk ɒbˈdjuɹæt ˈɛt ˈtɐŋk ˈkjʊɹæt ˈludəʊ ˈmɛntɪs ˈeɪʃiɛm]

Aust: [ˈvɑetə ˌdetəˈstæbəlɪs ˈnaŋk əbˈdʒʉɹæt ˈet ˈtʌŋk ˈkjʉɹæt ˈludəʉ ˈmentɪs ˈæɪʃiəm]

egestátem, potestátem, dissolvit ut glaciem.

diaph: /ˌɛdʒɨˈsteɪtɛm ˌpɒtɨˈsteɪtɛm dɨˈsɒlvɪt ʌt ˈɡleɪʃiɛm/

Amer: [ˌɛdʒəˈsteitəm ˌpɑtəˈsteitəm dɪˈsɑlvɪt ˈʌt ˈɡleiʃiəm]

Brit: [ˌɛdʒəˈsteɪtɛm ˌpɒtəˈsteɪtɛm dɪˈsɒlvɪt ˈɐt ˈɡleɪʃiɛm]

Aust: [ˌedʒəˈstæɪtəm ˌpɔtəˈstæɪtəm dɪˈsɔlvɪt ˈat ˈɡlæɪʃiəm]

Ave formosissima, gemma pretiósa! Ave decus virginum, virgo gloriósa!

diaph: /ˈeɪvi ˌfɔrməˈzɪsɨmə ˈdʒɛmə ˌprɨʃiˈoʊzə ˈeɪvi ˈdiːkəs ˈvɜrdʒɨnəm ˈvɜrɡoʊ ˌɡlɔriˈoʊzə/

Amer: [ˈeivi ˌfɔɹməˈzɪsəmə ˈdʒɛmə ˌpɹiʃiˈɞuzə ˈeivi ˈdikəs ˈvɚdʒənəm ˈvɚɡɞu ˌɡlɔɹiˈɞuzə]

Brit: [ˈeɪvɪ ˌfɔːməˈzɪsəmə ˈdʒɛmə ˌpɹiʃiˈəʊzə ˈeɪvɪ ˈdikəs ˈvɜːdʒənəm ˈvɜːɡəʊ ˌɡlɔːɹiˈəʊzə]

Aust: [ˈæɪvi ˌfɔːməˈzɪsəmə ˈdʒemə ˌpɹiʃiˈəʉzə ˈæɪvi ˈdikəs ˈvɜːdʒənəm ˈvɜːɡəʉ ˌɡloɹiˈəʉzə]

Ave mundi luminar, Ave mundi rosa! Blanziflor et Helena, Venus generósa.

diaph: /ˈeɪvi ˈmʌndaɪ ˈluːmɨnɑr ˈeɪvi ˈmʌndaɪ ˈroʊzə ˈblænzɨflɔr ɛt ˈhɛlɨnə ˈviːnəs ˌdʒɛnɨˈroʊzə/

Amer: [ˈeivi ˈmʌndai ˈlɵumənɑɹ ˈeivi ˈmʌndai ˈɹɞuzə ˈblænzəflɔɹ ˈɛt ˈhɛlənə ˈvinəs ˌdʒɛnəˈɹɞuzə]

Brit: [ˈeɪvɪ ˈmɐndaɪ ˈlumənɑː ˈeɪvɪ ˈmɐndaɪ ˈɹəʊzə ˈblænzəflɔːɹ ˈɛt ˈhɛlənə ˈvinəs ˌdʒɛnəˈɹəʊzə]

Aust: [ˈæɪvi ˈmandɑe ˈlʉmənaː ˈæɪvi ˈmandɑe ˈɹəʉzə ˈblænzəfloːɹ ˈet ˈhelənə ˈvinəs ˌdʒenəˈɹəʉzə]

History

Latin as traditionally pronounced by English speakers is part of the living history of spoken Latin through medieval French into English.

Three stages of development of A-L can thus be distinguished:

Stage I

Latin from the period when its orthography and grammar became standardized through to the pronunciation changes of Late Latin, while it was still a living language. Changes that took place in this period included:

Stage II

Latin spoken in the context of Gallo-Romance and French from approximately the 6th to the 11th-12th centuries. During this period, Latin became a primarily written language, separated from the ordinary spoken language of the people. While it escaped many of the changes of pronunciation and grammar of Gallo-Romance, it did share a few of the changes of the spoken language. This was for the most part a period of stability. Changes in this period included:

Stage III

Latin spoken in the context of English from the 11th/12th centuries to the present. This last stage provides the greatest and most complicated number of changes. It starts with the displacement of the native pronunciation of Latin under the Anglo-Saxon kings with that used in the north of France, around the time of the Norman conquest in 1066. The English and French pronunciations of Latin were probably identical down to the 13th century, but subsequently Latin as spoken in England began to share in specifically English sound changes. Latin, thus naturalized, acquired a distinctly English sound, increasingly different from the pronunciation of Latin in France or elsewhere on the Continent.

Some phases of development in this third stage can be reconstructed:

1200–1400

a [ɐ]:[ɐː], e [ɛ]:[ɛː], i [ɪ]:[iː], o [ɔ]:[ɔː]

1400–1600

1600–1800

1800–present

Other languages

Further information: Latin regional pronunciation

A similar situation occurred in other regions, where the pronunciation of the local language influenced the pronunciation of Latin, eventually being replaced with reconstructed classical pronunciation. In German-speaking areas, traditional Germanized pronunciation of Latin is discussed at Deutsche Aussprache des Lateinischen (German), with reconstructed classical pronunciation at Schulaussprache des Lateinischen (German).

See also

Alencar, Fabio Braga, 2008, A Regra Universal da pronúncia Inglesa, Brazil

References

  1. ↑ Campbell, F. R. (1888). The Language of Medicine. p. 59, Section I, Rule I.
  2. ↑ Campbell, F. R. (1888). The Language of Medicine. p. 59, Section I, Rule II.2.
  3. ↑ Campbell, F. R. (1888). The Language of Medicine. p. 59, Section I, Rule II.3.