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.
In the Middle Ages speakers of English, from Middle English onward, pronounced Latin not as the ancient Romans did, but in the way that had developed among speakers of French. This traditional pronunciation then 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.
Until the beginning of the 19th century all English speakers used this pronunciation, including Roman Catholics for liturgical purposes. Following Catholic emancipation in Britain in 1829 and the subsequent Oxford Movement, newly converted Catholics preferred the Italianate pronunciation, which became the norm for the Catholic liturgy. Meanwhile, scholarly proposals were made for a reconstructed Classical pronunciation, close to the pronunciation used in the late Roman Republic and early Empire, and with a more transparent relationship between spelling and pronunciation.
One immediately audible difference between the pronunciations was in the treatment of stressed vowels, in which the English version followed the sound changes that had affected English itself, the stressed vowels being quite different from their unstressed counterparts, whereas in the other two versions they remained the same. Among the consonants, treatment of the letter c followed by a front vowel was an obvious distinction. Thus the name Cicero is spoken in English as , in the ecclesiastical Latin as and in the restored classical Latin as.
The competition between the three pronunciations grew towards the end of the 19th century.
By the beginning of the 20th century, however, a consensus for change had developed. The Classical Association, shortly after its foundation in 1903, put forward a detailed proposal for a reconstructed classical pronunciation. This was supported by other professional and learned bodies. Finally in February 1907 their proposal was officially recommended by the Board of Education for use in schools throughout the UK. Adoption of the "new pronunciation" was a [|long] drawn out process, but by the mid-20th century, classroom use of the traditional pronunciation had ceased.

Illustrative survivals

The traditional pronunciation survives in academic and general English vocabulary:
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".
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 or, or. 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. For example, Actaeon is pronounced or . A diaeresis indicates that the vowels do not form a diphthong: Arsinoë .
The importance of marking long vowels for Greek words can be illustrated with Ixion, from Greek Ἰξίων. As it is written, the English pronunciation might be expected to be *. However, length marking, Ixīōn, makes it clear that it should be pronounced .
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. 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.
If more than two syllables precede the stressed syllable, the same rules determine which is stressed. For example, in Cassiopeia, 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. Therefore, the standard English pronunciation is pronounced .

Long and [|short vowels] in English

Whether a vowel letter is pronounced "long" in English or "short" 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 . In the middle of a word, a vowel followed by more than one consonant is short, as in Hermippe , while a vowel with no following consonant is long. However, when a vowel is followed by a single consonant and then another vowel, it gets more complicated.
  • If the syllable is unstressed, it is open, and the vowel is often reduced to schwa.
  • If the penultimate syllable is stressed, it is open and the vowel long, as in Europa .
  • If any other syllable is stressed, it is closed and the vowel is short, as in Ganymede and Anaxagoras .
Regardless of position, stressed u stays long before a single consonant, as in Jupiter .
Traditionally, English syllables have been described as 'open' when their vowel is long and they are followed by a single consonant followed by another vowel, and as 'closed' in the same environment when their vowel is short. However, it is debated how accurate this analysis is, as in English syllables tend to attract a following consonant, especially when they are stressed, so that all stressed syllables followed by a consonant are arguably 'closed'. Such following consonants are sometimes described as ambisyllabic. This effect is especially apparent in some dialects, such as RP, when the consonant in question is /r/, which affects the quality of the preceding vowel. None of this changes the patterns described in this article: The long-short distinction described above is maintained regardless. For example, the 'e' in Hera is long regardless of whether it is pronounced or in a particular dialect, or analyzed as open or as closed. American dictionaries tend to follow the former transcription, and British dictionaries the latter, so when the consonant 'r' is involved the rules for the English pronunciation of Latin words are more straightforward when using the conventions of American dictionaries.

Alphabet

Anglo-Latin 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 be used. These two digraphs respectively represent mergers of the letters ae and oe and are often written that way. 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 Greek sequences ch, ph, rh, th represent single sounds. The letters x and Greek z, on the other hand, are sequences of two sounds.

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:
Rh is used for Greek ρ at the beginnings of words, e.g. ῥόμβος > rhombus.
Rarely, k is used to represent Greek κ. In such cases it is always pronounced and never : e.g. σκελετός > skeleton not "sceleton".
Greek accent marks and breath marks, other than the "rough breathing", 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:
Examples:

Letters and sounds

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

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 ; medial chth and phth are pronounced and respectively, as in autochthon and naphtha.

Polyphony

The letters c, d, g, h, n, s, t and x have different sounds depending upon their environment: these are listed summarily below.
Lettercdghnstx
Underlying sound
Primary phonemes∅,
Secondary phonemes


The full set of consonantal phonemes for A-L is almost identical to that of English, lacking only.
Miscellaneous environments
Environments that condition the appearance of some of these phonemes are listed below:
Sound affectedSpellingEnvironmentResulting soundExamples
hbetween a preceding stressed and a following unstressed vowel∅cf. "vehement, annihilate"
hafter x∅exhibitor
nbefore velars and gincubator, fungus
sbetween two vowelsmiser, Cæsar, Jesus
sbetween a vowel and a voiced consonantplasma, presbyter
safter a voiced consonant at the end of a wordlens, Mars
xinitiallyXanthippe
xin the prefix ex- before a vowel or h in a stressed syllableexemplar, exhibitor

The change of intervocalic to 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
PalatalizationSound affectedSpellingEnvironmentResulting soundExamples
1twhen not initial, following s, t, or x, and before the semivowel iannunciator
usually changes to by Palatalization 3
2cbefore front vowels e, æ, œ, i, ycircus, census, Cynthia, foci, proscenium, scintilla, successor
2gbefore front vowels e, æ, œ, i, yGemini, regimen, algæ, fungi, gymnasium
3c, t
when not initial, before semivowel i and eacacia, rosacea, species, inertia, ratio
fascia, cassia
3xwhen not initial, before semivowel i and ecf. "complexion"
3twhen not initial, before semivowel i and ecf. "question, Christian, bestial, Attius"
3swhen not initial, before semivowel i and eAsia, ambrosia, nausea, Persia
4dwhen not initial, before open u, educator, cf. also gradual
4s, sswhen not initial, before open u, cf. "censure, fissure"
4xwhen not initial, before open u, cf. "luxury"
4twhen not initial, before open u, spatula
4swhen not initial, before open u'', 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 were degeminated, or simplified to a single sound. That is, bb, dd, ff, ll, mm, nn, pp, rr, ss, tt became pronounced. 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:
The simple vowels of A-L 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, by which they may be either open, semiopen, semiclosed, or fully closed.

Fully closed syllables

Fully closed syllables are those in which the vowel in the middle of the syllable is followed by at least one consonant, which ends or "closes" the syllable. Vowels in fully closed syllables appear:
Semiclosed syllables are closed, unstressed syllables that had been closed and 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.
Semiopen syllables are syllables that had been closed and unstressed, and 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, semiopen syllables are practically restricted to instances of s + obstruent, bl, and in some cases perhaps tl. Vowels in initial semiopen syllables may be treated as open for all purposes except for determining the value of u, which is still closed in semiopen syllables.
See further the section on initial unstressed syllables below.

Open syllables

Open syllables are those in which the nucleus is followed:

Primary stress

Stress is another characteristic of syllables. In A-L, it is marked by greater tension, higher pitch, lengthening of vowel, and 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.

Secondary stress

Secondary stress is dependent upon the placement of the primary stress. It appears only 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:
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 . 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, i, 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 ; historically semivowel e could also be distinguished from "long e". In current varieties of A-L, semivowels are pronounced in a variety of ways:
  2. * Most frequently as : labia, radius, azalea, præmium, cornea, opium, Philadelphia, requiem, area, excelsior, symposium, Cynthia, trivia, trapezium. In British Received Pronunciation, the prescribed pronunciation was once.
  3. * In some dialects or registers of English as, e.g. junior pronounced.
  4. * Merged with a following -es or -e ending, as in Aries, scabies.
  5. * They are usually deleted following the palatals,, and : Patricia, consortium, Persia, nausea, ambrosia, Belgium.
  6. *Occasionally a semivowel is retained after a palatal sound: ratio, sometimes Elysium. This type of pronunciation is an artificiality, as the sounds and resulted from an absorption of the original in the sequences,. The pronunciations with and 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.
  7. The consonant t changed to and then to before the semivowel arising from i: minutia, inertia, nasturtium.
  8. The sibilants and are usually palatalized before the semivowel:
  9. * > : cassia, fascia, species, militia
  10. * > : amnesia, ambrosia
  11. The vowels a, e, æ, and o in an open antepenult syllable become long if a semivowel appears in the next syllable:
  12. * radius, Asia, azalea, area
  13. * anæmia, chamæleon
  14. * genius, medium, interior
  15. * odium, cochlea, victoria

    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, 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 are sometimes retained in spelling preceding a vowel. In such cases the sequence is invariably pronounced as a simple vowel, sometimes i and sometimes 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 unstressed variants; and these variants would eventually also be altered based on 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 each have three primary variants: long, short, and reduced. Each of these may, in turn, exhibit allophonic variation based on features of its phonetic environment, including whether it is stressed, whether it is in an open or closed syllable, where it is positioned in the word, and what consonants are next to it. One of the most common environmental causes of vowel alteration is the presence of a following r. Vowels altered by a following "r" 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, 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 or by the end of the word. In addition, there is a variant of short a that only appears after a sound – chiefly found in the sound qu. This is a relatively recent phonetic development in English and Anglo-Latin, so it wasn't present in earlier stages of Anglo-Latin.
Short vowelsIPAType 1Type 2Type 3Type 4Type 5
apaxmantis, pallor, malefactorcamera, marathon, calculusanæsthesia, saturnalia antenna, magnificat
arpar, Marsargus, catharsisarbiter, Barbaraarbitrator, pharmacopœianarcissus, sarcophagus
erexsector, error, præceptor, interregnumGemini, Penelopememorandum, impedimenta pentathlon, September, spectator
æquæstorÆschylus, diæresisprædecessor, æquilibrium
œŒdipus
inilisthmus, lictor, cirrus, narcissussimile, tibia, antithesis, Sirius, deliriumsimulacrum, administrator, hippopotamusscintilla, dictator
ylynx, Scylla, Charybdischrysalis, synthesis, Thucydides, Syriasymbiosishysteria
erpervertex, Nervaterminus, hyperbolaperpetratorMercator, persona
ircircus, VirgoVirginia
yrthyrsusmyrmidon
w-colored aquantum
ononimpostor, horroroptimum, conifer, metropolispropaganda, operator October, thrombosis
w-colored arquartus
orcortex, forcepsformulacornucopiatorpedo

Exceptionally, monosyllables ending in es are pronounced with the rhyme, 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; they used to be pronounced with a short, as is still the case in the word "hypocrite" and hypochondria.
Prefixes may also behave in anomalous ways:
  1. The prefix ob- in unstressed syllables may be reduced to, even when it closes a syllable: cf. "obsession, oblivion".
  2. The Greek prefix en-, em- in a closed unstressed syllable may be reduced to, : encomium, emporium.
  3. The prefix ex- in an unstressed syllable may be reduced to,, despite always being in a closed syllable: exterior, exemplar.
  4. The prefix con-, com- is reduced to, 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": : post-mortem and cf. "postpone"; also thus in words in which post was originally a preposition 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. Hiatus may be original, or may arise from the deletion of h between a stressed and unstressed syllable
"Long" vowelsIPAType 1Type 2Type 3Type 4
aa, quacrater, lumbagoradius, rabieschaos, aorta, phaëthon
arpharosarea, caries
ee, reethos, lemur, Venusgeniusidea, creator
æCæsaranæmia, chamæleonæon, mæander
œamœba, Crœsusdiarrhœa
eiDeianira, Pleiades
er or ærserum, Ceres, ærabacterium, criterion, materia-
ii, pi item, Tigris, saliva, iris, horizonmiasma, hiatus, calliope
yhydra, python, papyrushyæna, myopia
oO, probonus, togaodium, encomium, opprobriumboa, Chloe, cooperator
orchorus, forum, thoraxemporium, euphoria

Reduced vowels

Reduced vowels appear in unstressed syllables, except for:
A variety of possible realizations are available for open, semiopen, and semiclosed initial unstressed syllables, including long, short, and reduced variants. Fully closed initial unstressed syllables are always short.
VowelIPAExamples
aamœba, anemone, ascesis
e*Elysium, emeritus, epitome, erotica
æ*ænigma
œ* œsophagus
iidea
yhyperbola, hypothesis
o**Olympus

* A "schwi". It generally conflates with in RP and with schwa in Australia. Directly before another vowel it may be the 'HAPPY' vowel.
** A rounded schwa. It has a w-like offglide before another vowel the way does, but otherwise in many dialects it conflates with schwa.
VowelIPAExamples
apapyrus, placebo, saliva, basilica
e*December, thesaurus
æ*Mæcenas, pæninsula, phænomenon
icriteria, tribunal, minutiæ, cicada
ylyceum, psychosis, synopsis, chrysanthemum
o**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 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, but by many speakers is still pronounced as a high front lax vowel, distinct from the derived from a, here transcribed. For such speakers, the first syllables in Demeter and Damascus are pronounced differently.
Unstressed o, also often transcribed, is by many speakers pronounced with considerable lip-rounding, here transcribed.
VowelIPAExamples
aaddendum, appendix, calliope, farrago
eellipsis, Ecclesiastes, erratum
iIllyria, cf. cirrhosis
ysyllepsis
ocollector, 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 or, regardless of whether they are in open or closed syllables.
VowelIPAExamples
adiabetes, emphasis, syllabus, diagnosis, melancholia
e*impetus, phaethon, malefactor, commentator, Alexander
i*animal, legislator
y*platypus, analysis, apocrypha
ohyperbola, demonstrator
Vrinterceptor, 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 ; it shares developments with the homophonous diphthong eu, which can however appear in closed syllables.
The sound in and its variants is deleted in various environments:
  • After palatal consonants, whether original or resulting from the merger of and the preceding consonant, in both stressed and unstressed syllables; e.g. : junior, Julius, Jupiter, cæsura, educator, spatula, fistula
After the following consonants when they precede u in an initial, final, or stressed syllable:
  • and : rumor, verruca, luna, Lucretia, Pluto, effluvium
  • , : super, superior, Vesuvius, and variably Zeus
In some dialects, particularly of American English, is deleted after all dental/alveolar consonants when they precede u in an initial, final or stressed syllable:
  • ,, and : duplex, caduceus, medusa, nucleus, lanugo, tutor, Thucydides
  • For some speakers, is pronounced following these consonants.
is not deleted in the following environments:
  • When u is the first letter of the word or follows : uterus, humerus
  • Following a vowel: Ophiuchus
  • Following labials : pupa, furor, nebula, uvula, musæum
  • Following velars : cumulus, lacuna, Liguria
  • When it is in an interior unstressed syllable not following a palatal consonant, remains after a single consonant even when it might be deleted in a stressed syllable: amanuensis and cf. "cellular, granular", for some speakers "virulent".
  • After a consonant cluster may or may not be deleted: pæninsula, cornucopia

    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 or the end of the word.

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 : Augustus pronounced as if "Agustus". Such words may be pronounced with the full value of the diphthong, however.
DiphthongsIPAExamples
aiAchaia, Maia, Gaius
auaura, pauper, nausea, autochthon, aurora, glaucoma, mausoleum
euneuter, euthanasia, zeugma
oicoitus, paranoia
ouboustrophedon
uialleluia, cuius
yiharpyia

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 and.

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.
LetterIPAExamples
acirca, fauna, mania, quota
eante, epitome, posse
æalgæ, larvæ, vertebræ
i, yalibi, Gemini, moly
oego, Pluto, torpedo
u situ

Words deriving from Greek long -ê end in unless assimilated, such as simile ending in .
In the words mihi, tibi, sibi, by an old tradition, the final i was pronounced like final e above.
A late and purely academic pronunciation distinguished final -ā from -a by pronouncing the former like "long a", : for instance, Oxford professor A. D. Godley rhymed Rusticā and "day". 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 had made all final a'''s "long", regardless of their Latin length.

Vowel before a consonant cluster

The second class consists of vowels followed by 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.
LetterIPAExamples
aclimax, phalanx
ebiceps, index
imatrix, phœnix
oCyclops
uexeunt, Pollux
ypharynx, oryx

Vowel before a single 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 : Cæsar, pauper, triumvir, Mentor, sulfur, martyr.
  2. All vowels are reduced to before l: tribunal, Babel, pugil, consul.
  3. Except sometimes before t, a is reduced to before any of this class of consonant: animal, memoriam, titan, atlas.
  4. All instances of u are reduced to 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. Final -es and -ies are alike pronounced. Final -eus, when derived from Greek -ευς -eus rather than from -εος -eos with the -os changed to -us in Latin, may be pronounced as a single syllable with a diphthong, or as two syllables with a long e followed by -us. 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 -eus ending is considered the penult, just as happens in derivatives ending in -ian – though the placement of the stress shifts. E.g. Ἰλιονεύς Īlioneus .
EndingIPAExamples
atmagnificat
atfiat
emidem, ibidem
emitem, tandem
enlichen, semen
esAchilles, appendices, fæces
iesrabies, species
etvidelicet, scilicet, quodlibet
eusPerseus, Nereus
eusPerseus, Nereus
impassim
iminterim
is*ægis, crisis, hypothesis, basis
it*exit, deficit
onicon, marathon
onbison, siphon, horizon
oschaos, pathos, pharos
otaliquot
otaliquot

* The vowel of -is, -it is a "schwi". It generally conflates with in RP and with schwa in Australia.
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. Such an ending is not found in English loan words or proper names.

Adjectives in ''-an'' and ''-ic''

English adjectives formed from Greek and Latin roots often end in a suffix -an or -ic added to the oblique stem, sometimes retaining a preceding thematic vowel. These produce generally predictable sound changes in the stem though, depending on its source or simply due to confusion, English -ean may be either stressed or unstressed. The first derives from the Latin adjectival suffix -ānus, though it has far wider application in English than in Latin. The other suffix, -ic, derives from the Greek adjectival suffix -ικος -ĭkos.
The fact that these suffixes are added to the oblique stem is relevant with words of the third declension whose stems end in a consonant that alters or disappears in the nominative case, as is apparent in such English noun–adjective pairs as Pallas ~ Palladian, Mars ~ Martian, Venus ~ Venerian, and indeed from non-adjectival derivatives of these words such as 'palladium'. Some roots have more than one oblique stem, in which case they may have more than one adjectival form in English.
;Stressed vowel plus -an and unstressed -ian
The forms of these suffixes depends on the first vowel of the nominal suffix in Latin or Greek. The resulting forms are carried over into Latin from Greek, and into English from Latin. English adjectives in -ean, where the e is inherited from the Greek, are stressed on the suffix, but there are also adjectives in -ean where the e originates in Latin, and those are not stressed.
For example,
  • Greek Samos > adjective Samios > Latin Samius > English 'Sā́mian'
  • Greek Periklēs > adjective Perikleios > Latin Periclēus > English 'Pericléan'
  • Greek Lēda > adjective Lēdaios > Latin Lēdæus > English 'Ledaean' or 'Ledéan'
  • Greek noun and adjective Ptolemaios > Latin Ptolemæus > English 'Ptolemaean'
  • Greek Argō > adjective Argōos > Latin Argōus > English 'Argóan'
  • Greek Sapphō > adjective Sapphōios > Latin Sapphōius > English 'Sapphóian'
Because the i of the suffix -ian makes an exception to the usual shortening of antepenultimate vowels, forms such as 'Samian' above are pronounced with a long stressed vowel.
Many Greek nouns ending in -o- take adjectives like those ending in -ē- instead, and some may take either form, in which case there may be two adjectives in English. An example is Sisyphus:
  • Greek Sīsyphos > adjective Sīsyphios > Latin Sīsyphius > English 'Sisýphian'
  • > adjective Sīsypheios > Latin Sīsyphēus > English 'Sisyphéan'
In the opposite direction, ē-vowel Aristotelēs produces 'Aristotḗlian' alongside expected 'Aristoteléan'.
The final iota subscript in Greek feminine words ending in -ῳ -ōi is frequently omitted in Latin, – if an 'i' appears in English it may be taken directly from the Greek, – but words that end in plain -ω -ō in Greek may also have Greek adjectives in -ōios if the 'i' was historically present.
;Unstressed -ean
There is another suffix spelled -ean in English, but this one unstressed. It corresponds to the unstressed Latin adjectival suffix -ĕus and does not derive from Greek. An example is Latin Herculēs > Latin adj. Herculĕus > English 'Hercúlean', alongside Greek Hēraklēs > Greek adj. Hērakleios > Latin Hēraclēus/Hēraclīus > English 'Heracléan'. The stress assignment of the 'Herculean' has been affected by 'Heraclean', so the mixed form 'Herculéan' is now also heard.
;Adjectives in -ic
The Greek suffix -ικος -ĭkos replaces the vowels of nominal endings apart from retaining a or ō, which will be stressed in English. Stress assignment of English follows the Latin, as though the suffix were -icus in English as well and the last syllable of the stem were the antepenult when it English it's actually the penult. Thus English 'Homeric' is pronounced with a short stressed penultimate e, when in other English words from Latin an e in that position would be pronounced long.
For example,
  • Greek Homēros > adjective Homērikos > Latin Homēricus > English 'Homĕ́ric'
  • Greek Aristotelēs > adjective Aristotelikos > Latin Aristotelicus > English 'Arístotĕ́lic'
  • Greek Alkaios > adjective Alkaikos > Latin Alcaicus > English 'Alcáic'
  • Greek hērōs > adjective hērōikos > Latin hērōicus > English 'heróic'
  • Greek ēchō > adjective ēchōikos > Latin ēchōicus > English 'echóic'
Thus 'Aristotelian' is pronounced with a long e, but 'Aristotelic' with a short e.
A number of Greek nouns ending in ē form adjectives in -aic, such as 'Cyrenaic'.

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:
  • the merger of f and ph as
  • the change in pronunciation of v to and of j to.
  • the merger of i and y as
  • the merger of e, æ and œ as
  • the change of non-initial, unstressed, prevocalic to
  • the loss of distinctions of vowel length
  • the palatalization of t to before

    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:
  • the palatalization of c and g to and before front vowels
  • the voicing of intervocalic s to
  • the fronting of u to
  • the restoration of the vowels and from

    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

  • The adaptation of the French sounds to English:
  • * was substituted for.
  • * the vowels were given the values a, e, i, o, u
  • * was substituted for in closed syllables
  • Stressed open penultimate vowels were lengthened, creating the short/long contrasts:

    1400–1600

  • Merger of unstressed open with
  • Non-syllable-initial, unstressed, prevocalic became
  • Lengthening of the first of two vowels in hiatus
  • Lengthening of e, i, or o in pretonic initial syllables
  • Diphthongization of to
  • Lengthening of vowels in open syllables before in the next syllable
  • Raising of and to and.
  • Degemination of geminate consonants
  • Palatalization of and before
  • Fronting of to

    1600–1800

  • Monophthongization of ai to and au to
  • Change of to in many words, restoring original syllabicity.
  • Change of fronted u to
  • Palatalization of before
  • Loss of distinctive vowel length, creating the short/long contrasts: a :, e :, i :, o :
  • Lowering and unrounding of short, to,
  • Former long i becomes
  • Fronting and raising of short a, long a, and long e to , creating the new contrasts: a :, e :, i :, o :
  • Beginning of vowel reductions to.

    1800–2020

  • Development of and to diphthongs and
  • Laxing of to , , and the latter to
  • Continued vowel reductions to .
  • Shortening of a in a stressed open penult syllable, e.g., data.

    2020–present

  • Reduction of and in words with free variation between the two vowels to

    Other languages

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 :de:Deutsche Aussprache des Lateinischen|Deutsche Aussprache des Lateinischen, with reconstructed classical pronunciation at :de:Schulaussprache des Lateinischen|Schulaussprache des Lateinischen.
OWIKI.org. Text is available under the Creative Commons Attribution-ShareAlike License.