Latin phonology is the system of sounds used in various kinds of Latin. This article largely deals with what features can be deduced for Classical Latin as it was spoken by the educated from the late Roman Republic to the early Empire. Evidence comes in the form of comments from Roman grammarians, common spelling mistakes, transcriptions into other languages, and the outcomes of various sounds in Romance.^[1]

Latin orthography refers to the writing system used to spell Latin from its archaic stages down to the present. Latin was nearly always spelt in the Latin alphabet, but further details varied from period to period. The alphabet developed from Old Italic script, which had developed from a variant of the Greek alphabet, which had developed from a variant of the Phoenician alphabet. The Latin alphabet most resembles the Greek alphabet that can be seen on black-figure pottery dating to ca. 540 BC.

As the language continued to be used as a classical language, lingua franca and liturgical language long after it ceased being a native language, pronunciation and – to a lesser extent – spelling diverged significantly from the classical standard with Latin words being pronounced differently by native speakers of different languages. While nowadays a reconstructed classical pronuncation is usually employed in the teaching of Latin, the Italian-influenced ecclesiastical pronunciation as used by the Catholic church is still in common use. The Traditional English pronunciation of Latin has all but disappeared from classics education but continues to be used for Latin-based loanwords and use of Latin e.g. for binominal names in taxonomy.

During most of the time written Latin was in widespread use, authors variously complained about language change or attempted to "restore" an earlier standard. Such sources are of great value in reconstructing various stages of the spoken language (the Appendix Probi is an important source for the spoken variety in the 4th century CE, for example) and have in some cases indeed influenced the development of the language. The efforts of Renaissance Latin authors were to a large extent successful in removing innovations in grammar, spelling and vocabulary present in Medieval Latin but absent in both classical and contemporary Latin.

Letterforms

This section needs additional citations for verification. Please help improve this article by adding citations to reliable sources in this section. Unsourced material may be challenged and removed. (June 2009) (Learn how and when to remove this message)

In Classical times there was no modern-like distinction between upper case and lower case. Inscriptions typically use square capitals, in letterforms largely corresponding to modern upper-case, and handwritten text was generally in the form of cursive, which includes letterforms corresponding to modern lowercase.

Letters and phonemes

In Classical spelling, individual letters mainly corresponded to individual phonemes (alphabetic principle). Exceptions include:

1. The letters ⟨a e i o u y⟩, each of which could represent either a short vowel or a long one. The long vowels were sometimes marked with apices, as in ⟨á é ó ú ý⟩, while long /iː/ could be marked with the special character ⟨ꟾ⟩.^[2] In modern times^[when?] long vowels are marked with macrons, as in ⟨ā ē ī ō ū ȳ⟩; sometimes breves may also be used to indicate short vowels, as in ⟨ă ĕ ŏ ŭ y̆⟩.
2. The letters ⟨i u⟩, which could either indicate vowels (as mentioned) or the consonants /j w/ respectively. In modern times the letters ⟨j v⟩ began to be used as distinct spellings for these consonants (now often pronounced very differently).
3. Digraphs such as ⟨ae au oe⟩, which represented the diphthongs /ae̯ au̯ oe̯/. In a few words ⟨ae oe⟩ also stood for sequences of two adjacent vowels.
4. Digraphs such as ⟨ph th ch⟩, standing for the aspirated consonants /pʰ tʰ kʰ/.

Consonants

Below are the distinctive (i.e. phonemic) consonants that are assumed for Classical Latin. Those placed in brackets have a debated phonemic status, and those preceded by a dagger are found mainly or only in Greek loanwords.

Phonetics

• Latin may have had the labialized velar stops /kʷ ɡʷ/ as opposed to the stop + semivowel sequences /kw ɡw/ (as in the English quick or penguin). The argument for /kʷ/ is stronger than that for /ɡʷ/.^[a]
• The former could occur between vowels, where it always counted as a single consonant in Classical poetry, whereas the latter only occurred after /n/, where it is impossible to tell whether it counted as one consonant or two.^[3] The labial element, whether or , appears to have been palatalised before a front vowel, resulting in / (for instance quī would have sounded something like listen^ⓘ). This palatalisation did not affect the independent consonant ?pojem= before front vowels.^[4]
• /kʷ ɡʷ/ before /u/ may have become /k ɡ/ by dissimilation. This is suggested by the fact that equus and unguunt (from Old Latin equos and unguont) are also found spelt as ecus and ungunt, probably indicating the loss of the labial element in this context. Another possibility is that such spellings indicate that /k ɡ/ in general were allophonically labialized to by a following /u/, such that writing a double uu was unnecessary.^[5]
• /p t k/ were less aspirated than the corresponding English consonants, as implied by their usually being transliterated into Ancient Greeks as ⟨π τ κ⟩ and their pronunciation in most Romance languages. In many cases however it was not the Latin /p k/ but rather /b ɡ/ that were used to render Greek word-initial /p k/ in borrowings (as in πύξος, κυβερνῶ > buxus, guberno), especially borrowings of a non-learned character. This might suggest that the Latin /p k/ had some degree of aspiration, making /b ɡ/ more suitable to approximate the Greek sounds.^[6]
• /pʰ tʰ kʰ/ were pronounced with notable aspiration, like the initial consonants of the English pot, top, and cot respectively. They are attested beginning ca. 150 BC, in the spellings ⟨ph th ch⟩, at first only used to render the Greek ⟨φ θ χ⟩ in loanwords. (Previously these had been rendered in Latin as ⟨p t c⟩.) From ca. 100 BC onward ⟨ph th ch⟩ spread to a number of native Latin words as well, such as pulcher and lachrima. When this occurred it was nearly always in the vicinity of the consonant /r/ or /l/, and the implication is that Latin /p t k/ had become aspirated in that context.^[7][8]
• /z/ was found as a rendering of the Greek ⟨ζ⟩ in borrowings starting around the first century BC. (In earlier borrowings, the Greek sound had been rendered in Latin as /ss/.) In initial position this appears to have been pronounced , and between vowels it appears to have been doubled to (counted as two consonants in poetry).^[9][10]
• /s/ was unvoiced in all positions in Classical Latin. Previously however Old Latin /s/ appears to have voiced to between vowels, ultimately turning to /r/. Cicero reports the family-name Papisius being changed to Papirius in the fourth century BC, which may give some idea of the chronology. Afterward there developed new instances of /s/ between vowels from sound-changes changes like the degemination of /ss/ after long vowels and diphthongs (as in caussa > causa), which Quintilian reports to have happened a little after the time of Cicero and Virgil.^[11]
• In Old Latin, final /s/ after a short vowel was often lost, probably after first debuccalizing to , as in the inscriptional form Cornelio for Cornelios (Classical Cornelius). Often in the poetry of Plautus, Ennius, and Lucretius, final /s/ did not count as a consonant when followed by a word beginning with a consonant. By the Classical period this practice was described as characteristic of non-urban speech by Cicero.^[11]
• /f/ was labiodental in Classical Latin but may have been a bilabial /ɸ/ in Old Latin,^[12] or perhaps in free variation with . Lloyd, Sturtevant, and Kent make this argument based on misspellings in early inscriptions, the fact that many instances of Latin /f/ descend from Proto-Indo-European */bʰ/, and the outcomes of the sound in Romance (particularly in Spain).^[13]
• In most cases /m/ was pronounced as a bilabial nasal. At the end of a word, however, it was generally lost beginning in Old Latin (except when another nasal or a plosive followed it), leaving compensatory lengthening and nasalization on the preceding vowel^[14] (such that decem may have sounded something like listen^ⓘ). In Old Latin inscriptions, final ⟨m⟩ is often omitted, as in ⟨viro⟩ for virom (Classical virum). It was frequently elided before a following vowel in poetry and lost without a trace (apart from perhaps lengthening) in the Romance languages,^[15] except in a number of monosyllabic words, where it often survives as /n/ or a further development thereof.
• /n/ and /m/ merged via assimilation before a following consonant, with the following consonant determining the resulting pronunciation: bilabial before a bilabial consonant (/p b/), coronal before a coronal consonant (/t d/) and velar before a velar consonant (/k kʷ ɡ/). This occurred both within words (e.g. quīnque may have sounded something like listen^ⓘ) and across word-boundaries (for instance in causā with ŋ, or im pace).^[16]
• /ɡ/ assimilated to a velar nasal ŋ before /n/.^[17] Allen and Greenough say that a vowel before ŋn is always long,^[18] but W. Sidney Allen says that is based on an interpolation in Priscian, and the vowel was actually long or short depending on the root, as for example rēgnum from the root of rēx but magnus from the root of magis.^[19] /ɡ/ probably did not assimilate to ŋ before /m/. The cluster /ɡm/ arose by syncope, as for example tegmen from tegimen. Original /ɡm/ developed into /mm/ in flamma, from the root of flagrō.^[3] At the start of a word, original /ɡn/ was reduced to n, and this change was reflected in the orthography of later texts, as in gnātus, gnōscō > nātus, nōscō.
• In Classical Latin, the rhotic /r/ was most likely an alveolar trill r in some positions and when doubled. Gaius Lucilius likened it to the sound of a dog, and later writers described it as being produced by vibration. /d/ was sometimes written as ⟨r⟩, possibly suggesting a tap [ɾ] (like the single /ɾ/ in Spanish).^[20]
• /l/ was strongly velarized in syllable coda and probably somewhat palatalized when geminated or followed by /i(ː)/. In intervocalic position, it appears to have been velarized before all vowels except /i(ː)/.^[21]
• /j/ generally appeared only at the beginning of words, before a vowel, as in iaceō, except in compound words such as adiaceō (pronounced something like listen^ⓘ). Between vowels, it was generally as a geminate /jj/, as in cuius (pronounced something like listen^ⓘ) except in compound words such as trāiectus. This /jj/ is sometimes marked in modern editions by a circumflex on the preceding vowel, e.g. cûius, êius, mâior, etc. /j/ could also varied with /i/ in the same morpheme, as in iam /jam/ and etiam /ˈe.ti.am/, and in poetry one could be replaced with the other for metrical purpose.^[22]
• ?pojem= was pronounced as an approximant until the first century AD, when ?pojem= and intervocalic /b/ began to develop into fricatives. In poetry, ?pojem= and /u/ could be replaced with each other, as in /ˈsilua/~/ˈsilwa/ or /ˈɡenua/~/ˈɡenwa/. Unlike /j/ it remained a single consonant in most words, e.g. in cavē, although it did represent a double /ww/ did occur in borrowings from Greek such as the name Evander.^[23]
• /h/ was generally still pronounced in Classical Latin, at least by educated speakers, but in many cases it appears to have been lost early on between vowels, and sometimes in other contexts as well (diribeō < *dis-habeō being a particularly early example). Where intervocalic /h/ survived, it was likely voiced^[24] (that is, ɦ).

Notes on spellingedit

• Doubled consonant letters represented genuinely doubled consonants, as in ⟨cc⟩ for /kk/. In Old Latin, geminate consonants were written as if they were single until the middle of the second century BC, when orthographic doubling began to appear.^[b] Grammarians mention the marking of double consonants with the sicilicus, a diacritic in the shape of a sickle. It appears in a few inscriptions of the Augustan era.^[25]
• ⟨c⟩ and ⟨k⟩ both represented /k/, whereas ⟨qu⟩ represented /kʷ/. ⟨c⟩ and ⟨q⟩ distinguish minimal pairs such as cui /kui̯/ and quī /kʷiː/.^[26] In Classical Latin ⟨k⟩ appeared in only a few words like kalendae, Karthagō - which could also be spelt calendae, Carthagō.^[27]
• ⟨x⟩ represented /ks/. In Old Latin, /ks/ was also spelled in other ways like ⟨ks cs xs⟩. The letter ⟨x⟩ was borrowed from the Western Greek alphabet, where chi ⟨χ⟩ stood for /ks/ as well. This was unlike the usage of chi in the Ionic alphabet, where it stood for /kʰ/, with /ks/ instead represented by the letter xi ⟨ξ⟩.^[28]
• In Old Latin inscriptions, /k/ and /ɡ/ were not distinguished. They were both represented by ⟨c⟩ before ⟨e i⟩, by ⟨q⟩ before ⟨o u⟩, and by ⟨k⟩ before consonats or ⟨a⟩.^[2] The letterform ⟨c⟩ derives from the Greek gamma ⟨Γ⟩, which represented /ɡ/. Its use for /k/ may come from Etruscan, which did not distinguish voiced plosives from voiceless ones. In Classical Latin, ⟨c⟩ represented /ɡ/ only in the abbreviations c and cn, for Gaius and Gnaeus respectively.^[27][29]
• ⟨g⟩ was created in the third century BC to distinguish /ɡ/ from /k/.^[30] Its letterform derived from ⟨c⟩ with the addition of a diacritic or stroke. Plutarch attributes this innovation to Spurius Carvilius Ruga around 230 BC,^[2] but it may have originated with Appius Claudius Caecus in the fourth century BC.^[31]
• The combination gn probably represented the consonant cluster ŋn, at least between vowels, as in agnus ˈäŋ.nʊs listen^ⓘ.^[14][32] Vowels before this cluster were sometimes long and sometimes short.^[19]
• The digraphs ph, th, and ch represented the aspirated plosives /pʰ/, /tʰ/ and /kʰ/. They began to be used in writing around 150 BC,^[30] primarily as a transcription of Greek phi Φ, theta Θ, and chi Χ, as in Philippus, cithara, and achāia. Some native words were later also written with these digraphs, such as pulcher, lachrima, gracchus, triumphus, probably representing aspirated allophones of the voiceless plosives near /r/ and /l/. Aspirated plosives and the glottal fricative /h/ were also used hypercorrectively, an affectation satirized in Catullus 84.^[7][8]
• In Old Latin, Koine Greek initial /z/ and /zz/ between vowels were represented by s and ss, as in sona from ζώνη and massa from μᾶζα. Around the second and first centuries B.C., the Greek letter zeta Ζ was adopted to represent /z/ and /zz/.^[10] However, the Vulgar Latin spellings z or zi for earlier di and d before e, and the spellings di and dz for earlier z, suggest the pronunciation /dz/, as for example ziomedis for diomedis, and diaeta for zeta.^[33]
• In ancient times u and i represented the approximant consonants ?pojem= and /j/, as well as the close vowels /u(ː)/ and /i(ː)/.
• i representing the consonant /j/ was usually not doubled in writing, so a single i represented double /jː/ or /jj/ and the sequences /ji/ and /jːi/, as in cuius for *cuiius /ˈkuj.jus/, conicit for *coniicit /ˈkon.ji.kit/, and rēicit for *reiiicit /ˈrej.ji.kit/. Both the consonantal and vocalic pronunciations of i could occur in some of the same environments: compare māius /ˈmaj.jus/ with Gāius /ˈɡaː.i.us/, and Iūlius /ˈjuː.li.us/ with Iūlus /iˈuː.lus/. The vowel before a doubled /jː/ is sometimes marked with a macron, as in cūius. It indicates not that the vowel is long but that the first syllable is heavy from the double consonant.^[22]
• v between vowels represented single ?pojem= in native Latin words but double /ww/ in Greek loanwords. Both the consonantal and vocalic pronunciations of v sometimes occurred in similar environments, as in genua ˈɡɛ.nu.ä and silva ˈsɪl.wä.^[23][34]

Vowelsedit

Monophthongsedit

Classical Latin had ten native phonemic monophthongs, five short /i e a o u/ and five long /iː eː aː oː uː/. Some loanwords from Greek had ⟨y⟩, which would have been pronounced as /y(ː)/ by educated speakers but approximated with the native vowels /i(ː)/ or /u(ː)/ by the less-educated.

Long and short vowelsedit

The short vowels /i e o u/ appear to have been pronounced with a relatively open quality, which may be approximated as [ɪ] [ɛ] [ɔ] [ʊ], and the corresponding long vowels with a relatively close quality, approximately [iː] [eː] [oː] [uː].^[c] That the short /i u/ were, as this implies, similar in quality to the long /eː oː/ is suggested by attested misspellings such as:^[35]

• ⟨trebibos⟩ for tribibus
• ⟨minsis⟩ for mēnsis
• ⟨sob⟩ for sub
• ⟨punere⟩ for pōnere

/e/ most likely had a more open allophone before /r/.^[36]

/e/ and /i/ were probably pronounced closer when they occurred before another vowel, with e.g. mea written as ⟨mia⟩ in some inscriptions. Short /i/ before another vowel is often written with the so-called i longa, as in ⟨dꟾes⟩ for diēs, indicating that its quality was similar to that of long /iː/; it was almost never confused with e in this position.^[37]

Adoption of Greek upsilonedit

y was used in Greek loanwords with upsilon Υ. This letter represented the close front rounded vowel, both short and long: /y yː/.^[38] Latin did not have this sound as a native phoneme, and speakers tended to pronounce such loanwords with /u uː/ in Old Latin and /i iː/ in Classical and Late Latin if they were unable to produce /y yː/.

Sonus mediusedit

An intermediate vowel sound (likely a close central vowel [ɨ] or possibly its rounded counterpart [ʉ]), called sonus medius, can be reconstructed for the classical period.^[39] Such a vowel is found in documentum, optimus, lacrima (also spelled docimentum, optumus, lacruma) and other words. It developed out of a historical short /u/, later fronted by vowel reduction. In the vicinity of labial consonants, this sound was not as fronted and may have retained some rounding, thus being more similar if not identical to the unreduced short /u/ [ʊ].^[40] The Claudian letter Ⱶ ⱶ was possibly invented to represent this sound, but is never actually found used this way in the epigraphic record (it usually served as a replacement for the upsilon).

Vowel nasalizationedit

Examples of nasalized vowels at ends of words and before -ns-, -nf- sequences

monstrum

---

mensis

---

infans, infantem

---

Problems playing these files? See media help.

Vowels followed by a nasal consonant were allophonically realised as long nasal vowels in two environments:^[41]

• Before word-final m:^[15]
  • monstrum /ˈmon.strum/ > ˈmõː.strʊ̃
  • dentem /ˈden.tem/ > ˈdɛn.tɛ̃
• Before nasal consonants followed by a fricative:^[16]
  • censor /ˈken.sor/ > ˈkẽː.sɔr (in early inscriptions, often written as cesor)
  • consul /ˈkon.sul/ > ˈkõː.sʊɫ̪ (often written as cosol and abbreviated as cos)
  • inferōs /ˈin.fe.roːs/ > ˈĩː.fæ.roːs (written as iferos)

Those long nasal vowels had the same quality as ordinary long vowels. In Vulgar Latin, the vowels lost their nasalisation, and they merged with the long vowels (which were themselves shortened by that time). This is shown by many forms in the Romance languages, such as Spanish costar from Vulgar Latin cōstāre (originally constāre) and Italian mese from Vulgar Latin mēse (Classical Latin mensem). On the other hand, the short vowel and /n/ were restored, for example, in French enseigne and enfant from insignia and infantem (e is the normal development of Latin short i), likely by analogy with other forms beginning in the prefix in-.^[42]

When a final -m occurred before a plosive or nasal in the next word, however, it was pronounced as a nasal at the place of articulation of the following consonant. For instance, tan dūrum tan ˈduː.rũː was written for tam dūrum in inscriptions, and cum nōbīs kʊn ˈnoː.biːs was a double entendre,^[15] presumably for cunnō bis ˈkʊnnoː bɪs.

Diphthongsedit

ae, oe, au, ei, eu could represent diphthongs: ae represented /ae̯/, oe represented /oe̯/, au represented /au̯/, ei represented /ei̯/, and eu represented /eu̯/. ui sometimes represented the diphthong /ui̯/, as in cui listen^ⓘ and huic.^[26] The diphthong ei mostly had changed to ī by the classical epoch; ei remained only in a few words such as the interjection hei.

If there is a tréma above the second vowel, both vowels are pronounced separately: aë ä.ɛ, aü a.ʊ, eü e.ʊ and oë ɔ.ɛ. However, disyllabic eu in morpheme borders is traditionally written without the tréma: meus ˈme.ʊs 'my'.

In Old Latin, ae, oe were written as ai, oi and probably pronounced as äi̯, oi̯, with a fully closed second element, similar to the final syllable in French travail^ⓘ. In the late Old Latin period, the last element of the diphthongs was lowered to e,^[43] so that the diphthongs were pronounced äe̯ and oe̯ in Classical Latin. They were then monophthongized to ɛː and eː respectively, starting in rural areas at the end of the Republican period.^[d] The process, however, does not seem to have been completed before the 3rd century AD, and some scholars say that it may have been regular by the 5th century.^[44]

Vowel and consonant lengthedit

Vowel and consonant length were more significant and more clearly defined in Latin than in modern English. Length is the duration of time that a particular sound is held before proceeding to the next sound in a word. In the modern spelling of Latin, especially in dictionaries and academic work, macrons are frequently used to mark long vowels: ⟨ā ē ī ō ū ȳ⟩, while the breve is sometimes used to indicate that a vowel is short: ⟨ă ĕ ĭ ŏ ŭ y̆⟩.

Long consonants were usually indicated through doubling, but ancient Latin orthography did not distinguish between the vocalic and consonantal uses of i and v. Vowel length was indicated only intermittently in classical sources and even then through a variety of means. Later medieval and modern usage tended to omit vowel length altogether. A short-lived convention of spelling long vowels by doubling the vowel letter is associated with the poet Lucius Accius. Later spelling conventions marked long vowels with an apex (a diacritic similar to an acute accent) or, in the case of long i, by increasing the height of the letter (long i); in the second century AD, those were given apices as well.^[45] The Classical vowel length system faded in later Latin and ceased to be phonemic in Romance, having been replaced by contrasts in vowel quality. Consonant length, however, remains contrastive in much of Italo-Romance, cf. Italian nono "ninth" versus nonno "grandfather".^[46]

A minimal set showing both long and short vowels and long and short consonants is ānus /ˈaː.nus/ ('anus'), annus /ˈan.nus/ ('year'), anus /ˈa.nus/ ('old woman').

Table of orthographyedit

The letters b, d, f, h, m, n are always pronounced as in English b, d, f, h, m, n respectively, and they do not usually cause any difficulties. The exceptions are mentioned below: