Jump to content

User:Nicodene/sandbox4

From Wikipedia, the free encyclopedia

Latin phonology and orthography is the system of sounds found in Latin and the system of spelling used to represent them.

Classical

[edit]

This part of the article deals with what can be deduced of the pronunciation of the educated classes around the first century BC. In this period the spelling system was largely phonemic. A notable exception is its rendition of vowels, which generally were not distinguished as long or short.[1]

Consonants

[edit]
Labial Coronal Dorsal/Glottal
plain labial
Nasal m n (ŋ)
Plosive voiceless p t k ()
voiced b d ɡ (ɡʷ)
aspirated {} {} {}
Fricative voiceless f s h
voiced {z}
Approximant l (j) (w)
Trill r

Consonants in brackets are disputed, at least as phonemes, and those in braces are found mainly or exclusively in loanwords from Greek.

Spellings and interpretations

[edit]
  • ⟨b p⟩ stood for /b p/.[2][a]
  • ⟨c g⟩ stood for /k ɡ/.[3][c]
    • They had no ‘soft’ variants before front vowels.[d]
  • ⟨d t⟩ stood for /d t/. They were likely dental.[7][e]
    • /d/ would often assimilate to a following consonant.[f]
  • ⟨f⟩ stood for /f/. It was likely labiodental.[8][g]
  • ⟨gn⟩ likely stood for [ŋn].[9][h]
    • Apparent minimal pairs with [ŋn] versus [nn] have been used to argue for a phonemic /ŋ/.[i]
    • It has been claimed that ⟨gn⟩ lengthened preceding vowels[11] but this seems doubtful.[12][j]
  • ⟨h⟩ stood for /h/.[15][k]
    • There was a strong tendency for the sound to be lost in medial position.[l]
    • ⟨h⟩ after ⟨p t c⟩ denoted aspiration and after ⟨r⟩ possibly nothing. See ⟨ph th ch⟩ and ⟨rh⟩ below.
  • ⟨i u⟩, as consonants, stood for [j w].[17][m]
    • Scholars disagree on whether these were phonemes or simply allophones of the vowels /i u/.[n]
    • Between vowels ⟨i⟩ normally stood for a geminate [jj]. ⟨u⟩ stood for [ww] in Greek words like Euander.[o]
    • In some cases ⟨i⟩ after a consonant stood for [ji].[p]
  • ⟨k⟩ stood for /k/. It occurred before ⟨a⟩ in a few words like kalendae.[22]
  • ⟨l⟩ stood for ‘dark’ i.e. velarized [ɫ] before other consonants, at the end of a word, and likely before the vowels ⟨e a o u⟩. In the geminate ⟨ll⟩ and before ⟨i⟩ it had a ‘clear’ quality, either [l] or [lʲ].[23][q]
  • ⟨m⟩ stood for /m/.[24][r]
  • ⟨n⟩ stood for /n/. It was dental or alveolar.[25][s]
  • ⟨ph th ch⟩ stood for the aspirates /pʰ kʰ/. They were found in borrowings from Greek. Sometimes they replaced Latin ⟨p t c⟩.[26][t]
  • ⟨qu gu⟩ stood for either /kʷ ɡʷ/ or /kw ɡw/. There is no scholarly consensus on the matter.[27][u]
    • The ⟨u⟩ in ⟨qu gu⟩ was apparently lost via dissimilation before a following ⟨u⟩.[v]
    • The ⟨u⟩ in ⟨qu gu⟩ may have been palatalised to a sound like the Greek ⟨υ⟩[w] (that is, like /y/).
  • ⟨r⟩ likely stood for an alveolar trill /r/. A flap /ɾ/ and a variable [r~ɾ] have also been suggested.[28][x]
    • ⟨rh⟩ was found in borrowings from Greek. It seems unlikely to have been pronounced differently from ⟨r⟩.[y]
  • ⟨s⟩ stood for /s/. It was likely dental or alveolar.[29][z]
  • ⟨x⟩ stood for /ks/.[30][ab]
  • ⟨z⟩ stood for /z/. It was found in borrowings from Greek.[31][ac]
    • Between vowels it stood for a geminate [zz].[ad]

Other theories

[edit]
  • Intervocalic /p t k/ might have been somewhat aspirated.[32][ae]
  • Intervocalic /b d ɡ/ might have been fricatives or approximants.

Phonotactics

[edit]
  • The following geminates could occur [...]
  • Syllabification [...]
    • Stress assignment [...]

Vowels

[edit]
The Classical Latin vowel-space per Allen 1978:47

Monophthongs

[edit]
Front Central Back
Close i
{y}

{}
u
Mid e o
Open a

Spellings and interpretations

[edit]
  • ⟨i e a o u⟩ represented two sets of vowels, the short /i e a o u/ and the long /iː uː/.
    • In the pairs /i(ː) e(ː) o(ː) u(ː)/ each short vowel was lax relative to the corresponding long vowel: ɛ ɔ ʊ] versus [iː uː].[33][af]
    • Sometimes the long vowels were indicated with apices, as in ⟨é á ó ú⟩. For long /iː/ there was a separate symbol ⟨ꟾ⟩.[34][ag]
  • ⟨y⟩ appeared in loanwords from Greek. It would have been pronounced [y yː] by the educated.[35][ah]

Post-Classical

[edit]

Important changes include palatalization and the loss of phonemic vowel length. For more information, see the articles linked above.

Modern Ecclesiastical

[edit]

An Italian rendition is described below. A notable difference from Italian proper is that it lacks the stressed vowels /e o/.[36]

Consonants

[edit]
Labial Coronal Dorsal
Nasal m n ɲ
Plosive voiceless p t k
voiced b d ɡ
Affricate voiceless t͡s t͡ʃ
voiced d͡z d͡ʒ
Fricative voiceless f s ʃ
voiced v
Approximant l j w
Trill r

Phonetics

[edit]
  • /s/ voices to [z] between vowels.[37][ai]
  • /t͡s d͡z ʃ ɲ/ lengthen between vowels, as in lignum [ˈliɲːɲum].[36]

Spelling

[edit]

⟨h⟩ is silent except in the words nihil and mihi, where it is stands for /k/.[38][aj]

The table below is based on descriptions and examples from the Liber Usualis.[39] For convenience the abbreviation ᶠᵛ is used for ‘followed by a front vowel’, i.e. followed by /i/ or /ɛ/. (Note that these vowels can be spelt in multiple ways.) Conversely, * is used for ‘not followed by a front vowel’.

sound spelling example sound spelling example sound spelling example
/m/ ⟨m⟩ mági /n/ ⟨n⟩ regína /ɲ/ ⟨gn⟩ regnum
/p/ ⟨p⟩ próut /t/ ⟨t th⟩ mártyr /k/ ⟨c* ch k q⟩ cáritas[ak]
/b/ ⟨b⟩ gubernátor /d/ ⟨d⟩ diéi /ɡ/ ⟨g*⟩ ego
/t͡s/ 00⟨t⟩[al] patiéntia /t͡ʃ/ ⟨cᶠᵛ⟩ caelum
/d͡z/ ⟨z⟩ zizánia /d͡ʒ/ ⟨gᶠᵛ⟩ génitor
/f/ ⟨f ph⟩ fílii /s/ ⟨s⟩ sápere /ʃ/ 00⟨scᶠᵛ⟩[am] descéndit
/v/ ⟨v⟩ vígor
/l/ ⟨l⟩ laetítia /j/ ⟨iᵛ~j⟩ alleluia
/r/ ⟨r⟩ dilígere /w/ ⟨uᵛ⟩ sanguis

Vowels

[edit]
Front Central Back
Close i u
Mid ɛ ɔ
Open a

Phonetics

[edit]
  • Length is affected by stress, as in Italian.[36][an]
  • /n/ assimilates to a following consonant. /m/ may do so as well, but variably.[36]
  • Unstressed ɔ/ are raised while unstressed /i a u/ remain unchanged in quality.[36]

Spelling

[edit]

Per the Liber Usualis.[39] Stressed vowels may be marked with an acute accent for didactic purposes.

sound spelling example
/i/ ⟨i y⟩ síccitas
/ɛ/ ⟨e ae oe⟩ exércitus
/a/ ⟨a⟩ grátia
/ɔ/ ⟨o⟩ constitútio
/u/ ⟨u⟩ excussórum

Pronunciation sample

[edit]

The following is a reading of The North Wind and the Sun:[40][ao]

  • Olim inter se Aquilo et Sol uter fortior esset certabant, cum viatorem quendam paenula amictum procedentem conspexerunt; atque ipsorum fortiorem existimandum esse consenserunt, qui efficeret ut viator ille paenulam deponeret. Aquilo autem vehementissime furere coepit; sed, quo fortiores flatus emittebat, eo artius viator se circumdabat paenula; tandem, viribus destitutus, propositum suum omisit. Tum Sol caelum clarissima luce illustravit; mox vero viator, calore victus, paenulam exuit. Itaque Aquilo, quamvis invite, confessus est solem esse fortiorem.
  • [ˈɔːli ˌmin̪t̪eɾˈs̪ɛ| ˈaːkwiloe̞t̪ ˈs̪ɔlː| ˈuːt̪e̞ɾ ˈfɔrːt͡s̪jo ˈɾɛs̪ːs̪et̪| t͡ʃeɾˈt̪aːban̪t̪| ˌkuɱviaˈt̪ɔˑɾem ˈkwɛn̪ːd̪am| ˈpɛˑnula ˈmik̟ːt̪um pɾot͡ʃeˈd̪ɛn̪ːt̪em| ˌkon̪s̪pek̟ˈs̪ɛːɾun̪t̪| ˈat̪ːkwe| ipˈs̪ɔːɾuɱ foɾˈt͡s̪jɔːɾem| eɡˌzis̪t̪iˈman̪ːd̪u ˈmɛs̪ːs̪e ˌkon̪s̪en̪ˈs̪ɛːɾun̪t̪| ˈkwi efˈfiːt͡ʃeɾe̞t̪| ˌut̪viˈaːt̪o ˈɾilːle̞| ˈpɛːnulan̪ d̪eˈpɔːneɾet̪‖ ˈaːkwilo ˈaˑu̯t̪eɱ| ˌveemen̪ˈt̪is̪ːs̪ime̞ ˈfuːɾeɾe̞ ˈt͡ʃɛːpit̪| ˈs̪ɛd̪ː kwo̞foɾˈt͡s̪jɔːɾes̪ ˈflaːt̪us̪ ˌemit̪ˈt̪ɛːbat̪| e̞oˈarːt͡s̪juz viˈaːt̪oɾ ˌs̪e̞t͡ʃiɾˈkun̪ːd̪abat̪ ˈpɛːnula| ˈt̪an̪ːd̪eɱ ˈviˑɾibuz ˌd̪es̪t̪iˈt̪uːt̪us̪ pɾoˈpɔːzit̪un̪ ˈs̪uˑu moˈmiːzit̪‖ ˈt̪umː ˈs̪ɔlː| ˈt͡ʃɛːluŋ klaˈris̪s̪ima ˈluːt͡ʃe̞il lus̪ˈt̪raːvit̪| ˈmɔkːs̪ ˈvɛːɾo| viˈaːt̪oɾ kaˈlɔˑɾe ˈvik̟ːt̪us̪| ˈpɛːnula ˈmɛɡːzwit̪‖ ˈiːt̪akwe ˈaːkwilo| ˈkwaɱvi s̪iɱˈviːt̪e̞| koɱˈfɛs̪ːs̪u ˈs̪ɛs̪ːt̪| ˈs̪ɔːlem ˌe̞s̪s̪efoɾˈt͡s̪jɔːɾem]

Notes

[edit]
  1. ^ Sturtevant notes that the testimonies clearly describe labial sounds and that traditional pronunciations of Latin agree on [p] for ⟨p⟩ and [b] for at least word-initial ⟨b⟩. In the first century AD inscriptions begin to show fluctuations between ⟨b⟩ and ⟨v⟩. Most likely this indicates a fricative value [β] for intervocalic ⟨b⟩, judging by the latter's Romance reflexes being fricatives, as in habēre > Romanian avea, Italian avere, French avoir. In Late Latin, according to Lloyd (page 239), even word-initial /b/ would have been allophonically [β] after a vowel, which would help explain the collapse of the b-v distinction in much of Romance.

    Testimonies per Sturtevant. Terentius Scaurus: ‘B harmonizes with p and m, since their origin corresponds, the mouth being energetically closed.’ Martianus Capella: ‘With the lips forced open by the impulse of the breath, let us utter b... The breath causes p to burst from the lips.’
  2. ^ Sturtevant states that ⟨bs bt⟩ were regularly pronounced [ps pt] judging by the testimonies as well as the fact that the alternative spellings ⟨ps pt⟩ are often found as well. Allen adds that ⟨b⟩ evidently underwent complete assimilation in cases like obfero, submoueo, obgero, subcingo, subripio - resulting in pronunciations like offero, summoueo, oggero, succingo, surripio. On the other hand he mentions the possibility of learned spelling-pronunciations featuring ⟨b⟩ as an actual [b], though still probably devoiced before voiceless consonants, as in o[p]fero or su[p]cingo.

    Cser's findings are as follows (pages 172–6). The /b/ in ob- would usually assimilate completely to a following /k ɡ f/ and sometimes /m/. It would devoice before /p t s/, although its spelling usually remained ⟨b⟩ before ⟨t s⟩. The /b/ in ab- never assimilated completely. It would devoice before a voiceless consonant. It had the following allomorphs: abs- before /k t/, as- before /p/, ā- before /m w/ (also /f/ in āfui), au- before /f/. The /b/ in sub- could assimilate completely to labials (especially /f/), velars (especially /ɡ/), and /r/. It had the following allomorphs: sus- before voiceless stops, su- before /sp/.

    Testimonies per Sturtevant. Quintilian: ‘In regards to the writing of prepositions, it is often inquired whether it is proper to observe the sound which they have in composition or when separate, as when I say obtinuit; for analogy demands b as the second letter, but the ears hear rather p.’ Terentius Scaurus: ‘Plebs, urbs, and pelops are also subject to dispute; Varro distinguishes them in such a way that he thinks that those words should be written with bs in the nominative which have the same letter in the genitive singular, as plebs plebis, urbs urbis, but those with ps which in the genitive of the same number end similarly in pis, as Pelops Pelopis. But it seems to me that both classes should be written with ps, since of these consists the letter ψ, which, as I have said, becomes either bis or pis in the genitive.’
  3. ^ Sturtevant notes that traditional pronunciations of Latin, and for the most part the Romance reflexes as well, are in agreement on the values of Latin ⟨c g⟩ except before front vowels.[4] In archaic Latin ⟨c k q⟩ tended to be in complimentary distribution, with ⟨c⟩ used before ⟨i e⟩, ⟨k⟩ before ⟨a⟩, and ⟨q⟩ before ⟨o u⟩. By Classical times ⟨c⟩ had generalized at the expense of ⟨k q⟩, which were relegated to more marginal roles. Allen adds that ⟨c⟩ originally stood for both /k/ and /ɡ/, as can be seen in archaic spellings such as ⟨virco⟩ for the Classical virgo, until a separate letter ⟨g⟩ was devised for /ɡ/ in the third or fourth century BC.

    Testimony per Sturtevant. Donatus: ‘...k and q seem superfluous to certain writers, who do not know that whenever a follows, k should precede, not c, and whenever u follows we should write q, not c.’
  4. ^ Sturtevant cites the following as evidence. Central Sardinian and Dalmatian have [k] for Latin ⟨c⟩ before front vowels. There are no Classical testimonies of ⟨c g⟩ standing for different sounds before front vowels. In Umbrian original /k/ underwent assibilation before /i e/, yet the resulting sound was not spelt with Latin ⟨c⟩. Inscriptions sometimes have ⟨k⟩ for ⟨c⟩ before front vowels, as in ⟨pake⟩. The Greek-style ⟨ch⟩ /kʰ/[5] spread to Latin words like centuriones > chenturiones. That Varro apparently cited anceps and ingenuus alongside ancora as examples of [ŋ][6] suggests that the following consonants were velar. Borrowings from Latin into Greek render ⟨c g⟩ before front vowels as ⟨κ γ⟩, as in Cicero > Κικερων. Borrowings into Germanic show /k/ for Latin ⟨c⟩ before front vowels, as in lucerna > Gothic lukarn or cellarium > German Keller. Not until the sixth century AD do Latin spellings indicative of affricate values for original /k/, or confusion of original /ɡ/ with /j/, occur with notable frequency. On the other hand, Sturtevant adds, it is possible that Latin /k ɡ/ had fronted allophones before front vowels, as is common cross-linguistically.

    Allen adds the following. Livy's ‘censuit consensit consciuit’ suggests an alliteration [k- k- k-]. Fronted allophones for /k ɡ/ would help explain why forms like celsus < cello apparently resist the usual backing effect of the following [ɫ] (as in pulsus < pello). Fronted allophones would also be consistent with the archaic tendency to write ⟨c⟩ before front vowels in particular.
  5. ^ Sturtevant notes that the testimonies from Terentianus and Victorinus clearly describe the tip of the tongue touching the teeth to make ⟨d⟩. While the same testimonies emphasize the tongue touching the gums to make ⟨t⟩, which is suggestive of an alveolar, they do not refer to the tip of the tongue in particular. This leaves open the possibility that both consonants were pronounced with the tip of the tongue against the teeth and its flat surface against the gums, where, says Sturtevant, ‘the more energetic articulation of t’ would explain the grammarians' greater emphasis on the gums in their descriptions for ⟨t⟩.

    Allen is skeptical on the grounds that the testimonies describe an alveolar contact for ⟨t⟩. On the other hand, he adds, the Romance outcomes suggest it was dental, and Roman grammarians, lacking a clear understanding of voice, may have seized on an insignificant or even fictitious difference to distinguish ⟨t⟩ from ⟨d⟩.

    Weiss states that ⟨t⟩ was probably dental, based on the Romance outcomes, and is skeptical of the difference in grammarians' descriptions of ⟨t⟩ and ⟨d⟩ for the same reason as Allen.

    Lloyd and McCullagh describe them as dental. Cser is uncertain. None of them seem to elaborate on the matter.

    Testimonies per Sturtevant. Terentianus: ‘But whenever the upper part of the tongue strikes the tips of the [upper] teeth and, with moderate curvature, the base of the teeth, it produces the sound of d and gives passage to the following vowel; for the sound of t, it is enough to strike with the surface of the tongue where the upper teeth have their inmost roots.’ Marius Victorinus: ‘But d and t, which, so to speak, are neighbouring sounds, are distinguished by the elevation and position of the tongue. For when it strikes at the same time the tips and the base of the teeth with its upper part, it produces the letter d. But whenever it is elevated and touches the place where the roots of the upper teeth are, it will produce t with the assistance of the following vowel.’ Martianus Capella: ‘D is formed by applying the tongue about the upper teeth... T is forced out by pushing the tongue against the teeth.’
  6. ^ Per Allen it is likely that the /d/ of ad assimilated completely to following consonants, other than [h j w m], such that for instance adsequor, adtineo, adripio, adfui, adpono, adgredior, adcurro would have been pronounced like assequor, attineo, arripio, affui, apponno, aggredior, accurro (all attested). He adds that a pun made by Plautus relies on adsum ‘I am present’ sounding like assum ‘roasted’. As with the assimilation of /b/, however, he mentions the possibility of learned spelling-pronunciations featuring ⟨d⟩ as an actual [d], per the testimony from Velius Longus, though it was still probably devoiced before voiceless consonants, as in a[t]sequor or a[t]pono.

    Cser's findings are as follows (pages 170–1). The /d/ in ad- was most likely to assimilate completely to a following consonant from the group /p t k/ and then, in decreasing order of likelihood, the groups /l r/, /ɡ s f/, and /m n/. He found it would not assimilate at all to /j w/ and was unlikely to assimilate to a following consonant cluster, especially one beginning with /k/, as in ad- + credere > adcredere. It was, however, likely to assimilate to a cluster of the type /sC/, such that the resulting */ss/ would automatically degeminate to /s/, as in ad + stare > astare.

    Testimony per Di Napoli. Velius Longus: ‘But we see that in almost all words beginning with p this preposition [ad] doubles the p, losing its d in the process, as in paret/apparet, ponit/apponit, pungit/appungit. Likewise the d is difficult to pronounce in words beginning with c when this preposition is added. So [c] is doubled in capio/accipio. In that vein Lucilius said atque accurrere scriba dne an c non est quod quaeras eque labores. He believed that it made no difference in spelling, but if one minds the sound, it is important that c be written and not d. If someone asks the same question about words beginning with t, let him know that attinet, attentus and adtinet, adtentus are written indiscriminately, as they should be, since they send the same sound to the ears, whichever letter is written. Words beginning with g produce no less a change in the pronunciation of this preposition. Sometimes g is doubled, with loss of the d, as in aggerat. There is also a context where d disappears without being replaced by another letter, as in aspicio, ascendo... There is also a context where d becomes l when the next word begins with the latter, as in alligere – not always, however, since we say adluere, adloqui, adlabi. Also some liked to pronounce and write abbibere with a doubled b and eliminated d. In this case Lucilius believed that it made no difference, saying abbibere non multum est d siet an b’.
  7. ^ Sturtevant opts for a labiodental according to the testimonies. On the other hand, he supposes that since the sound developed in most cases from Proto-Indo-European */bʰ/, it could have been bilabial in older Latin. In support of this he cites the apparent assimilation of /n/ to a bilabial in the early inscriptions ⟨im fronte⟩ and ⟨comfluont⟩ and states that ⟨m⟩ would be an unlikely spelling for labiodental [ɱ].

    Allen likewise gives Classical ⟨f⟩ as a labiodental according to the testimonies but questions the validity of the archaic inscriptions on the grounds that they could well have been representing [ɱf].

    Per Lloyd, Latin ⟨f⟩ is usually taken to have been a labiodental, as most of the Romance reflexes are. Nevertheless, he says, there could also have been a bilabial in light of the fact that Latin could not have inherited a labiodental as such from Proto-Indo-European, noting also that there is an inscription from Praeneste with ⟨Fercules⟩ for Hercules. He concludes: ‘It is very probable... that the pronunciation to be associated with the letter F was often [ɸ], possibly in free variation with [f], or with geographical variations’.

    Weiss and McCullagh opt for a labiodental, the former based on Marius Victorinus' testimony and the latter without elaborating.

    Gouvert notes that the scholarly consensus is that it was a labiodental and considers the evidence cited for a bilabial, namely the archaic inscriptions and the outcome /h/ in various Romance dialects, to be indecisive. He approaches the problem from the Proto-Italic side, based on the observation that Latin ⟨f⟩ developed from a three-way merger of initial */ɸ θ xʷ/. The most straightforward path for such a merger, he reasons, ends in a labiodental: *[ʍ] develarizes, thereby merging into */ɸ/, then the bilabial */ɸ/ merges with the dental */θ/ at their midpoint: the labiodental */f/.

    Testimonies per Sturtevant. Quintilian: ‘For that one also which is the sixth of our letters of a sound scarcely human, or rather with no sound at all, has to be blown through the interstices of the teeth.’ Terentianus: ‘Pressing the lower lip against the upper teeth you will sound this letter with a gentle breath, as if avoiding the Greek rough mutes.’ Marius Victorinus: ‘Putting the lower lip against the upper teeth, with the tongue withdrawn towards the high point of the palate, we shall pronounce f with a gentle breath’ Martianus Capella: ‘The teeth holding down the lower lip [form] f.’
  8. ^ Baglioni provides the following conspectus. Latin scholarship has come to the consensus that ⟨gn⟩ stood for [ŋn] on the grounds that it would be expected from the stop–nasal assimilations observed in cases like somnus, penna < *s(u̯)op-nos, *pet-na; that similar vowel-raising occurred in cases like dignus, lignum < *dĕk-nos, *lĕg-nom on the one hand and contingo < *contĕngo < *con-tango on the other;[10] that it would help explain the early reduction of word-initial ⟨gn⟩ to ⟨n⟩, as in nātus < gnātus, as well as the reduction of ⟨ngn⟩ to ⟨gn⟩ in compounds like ignosco < in-gnosco; that it is implied by inscriptional forms like ⟨dingnissime singnifer ingnes congnatus⟩; and that Plautus' pun ignem magnum for inhumanum would be difficult to imagine with [ɡn].

    Nevertheless, Romance linguists have converged on [ɡn] as the best explanation for the modern outcomes. Some have tried to reconcile the conflicting lines of evidence by theorizing that Latin ⟨gn⟩ was indeed originally [ŋn] but that at some point a spelling-pronunciation [ɡn] came into vogue among the litterati and later spread to the general population. Baglioni objects on the grounds that it is unlikely for spelling to have been so influential in an age that knew no printing press. He then goes on to argue that [ŋn] is no worse a starting-point for the Romance outcomes than [ɡn] and that it even seems more plausible in certain respects. For more, see the note for /ɡn/ on the page for Proto-Romance.
  9. ^ Sturtevant supports its interpretation as a phoneme. Lloyd prefers to regard /nn/ as /nː/ and implies that this invalidates the minimal pairs. McCullagh describes [ŋ] as ‘marginally phonemic’. Cser notes that it was in nearly complimentary distribution with both [n] and [ɡ], apart from pairs like annus/agnus and agger/angor, and describes it as ‘not a surface-contrastive unit’ but rather ‘a classic case of partial phonemic overlapping’ per Zirin.
  10. ^ Per Allen, this notion traces back to a passage in Priscian where he believes it is part of a late interpolation by someone who had misunderstood the work. Allen also points out that the ⟨i⟩ in dignus and lignum derives from an etymologically short vowel[13] and that Romance outcomes such as the Italian degno imply that the ⟨i⟩ remained short. Further, Allen notes that the grammarian Diomedes implies a short vowel in the first syllable of dignitas and the occurrence of transcriptions into Greek such a ⟨κογνιτου⟩.[14] He concedes that there are also some inscriptions with spellings that suggest the lengthening of /i/ (but not other vowels) before /ɡn/, for instance ⟨dꟾgne⟩. Allen supposes that the vowel-raising process that turned */ĕ/ to /ĭ/ may have in some cases continued, such that the resulting vowel was close enough to approach the level of /iː/, and hence the aforementioned spellings.
  11. ^ Sturtevant notes that it appears to have been a ‘weak’ sound from early on, especially between vowels, to judge by its loss in words like bīmus, nēmō, diribeō < *bi-himus, *ne-homō, *dis-habeō. Compare also ānser, liēn, meiō and their Sanskrit cognates haṃsas, plīhā, mehati. Conversely, an unetymological ⟨h⟩ was inserted in humerus (cf. the Greek ὦμος) and ahēnus (cf. the Sanskrit ayas or Gothic aiz). In poetry ⟨h⟩ was subject to elision and did not count as a consonant for the purposes of syllable weight. That the sound was [h] in particular is suggested by the fact that it was equated to the Greek rough breathing, that it was used after ⟨p t c⟩ to indicate aspirated consonants,[16] and that grammarians often referred to ⟨h⟩ as ‘aspiratio’ or ‘spiritus’. In Classical times the correct use of ⟨h⟩ was seen as a sign of education, and it appears to have been lost entirely in the popular speech of Pompeii. Judging by Quintilian's testimony, ⟨h⟩ was not pronounced in uehementer, comprehendere, mihi.

    Allen adds that Virgil had dehinc as a monosyllabic word. In light of the common loss of medial ⟨h⟩, he supposes that the unetymological ⟨h⟩ of ahenus originated as an orthographic device for marking hiatus. While ⟨h⟩ tended to survive in initial position, inscriptions already begin to reflect its loss or inappropriate insertion ‘by the end of the Republic’. The Roman grammarians reported uncertainties over whether there should be an ⟨h⟩ in humor, humidus, harena, harundo and they regarded uemens and prendere as acceptable forms of vehemens, prehendere.

    Lloyd: ‘H... by late Republican times had disappeared from ordinary speech...’ Weiss: ‘h was not pronounced in popular speech but only in the educated speech of the urban upper classes’. McCullagh: ‘it is rare word-internally between vowels and after consonants is only found in a few compounds (e.g. ad-hibeō)’.

    Testimonies per Sturtevant. Catullus: ‘Arrius said chommoda whenever he meant commoda, and hinsidiae for insidiae, and he hoped that he had spoken unusually well when he said hinsidiae with all his might [...] After he had been sent to Syria everybody's ears had a rest; they heard those same mistakes in mild and gentle form, and they were not afraid of such words thereafter, when suddenly comes frightful news: the Ionian waves, since Arrius arrived there, are no longer Ionian, but Hionian.’ Quintilian: ‘Those faults which are committed in pronunciation are judged only by the ear; though as to the aspiration, whether it be added or retrenched, in variation from common practice, it may be a question with us whether it be a fault in writing; if h indeed be a letter, and not merely a mark. The treatment of this sound has often changed with time. The ancients used it very sparingly even before vowels, as they said aedi and irci; and it was long afterwards withheld from conjunction from consonants, as in Gracci and triumpi. But suddenly an excessive use of it became prevalent, so that choronae, chenturiones, praechones are still to be seen in certain inscriptions; on which practice there is a well-known epigram of Catullus. Hence there remain even to our times uehementer, comprehendere, and mihi, indeed, among the ancient writers, especially those of tragedy, we find also in old copies mehe for me.’

    Testimonies per Allen. Anonymous: ‘H conrasis paululum faucibus... exhalat.’ Marius Victorinus: ‘H quoque adspirationis notam non litteram existimamus.’ Nigidius Figulus (quoted by Gellius): ‘Rusticus fit sermo si aspires perperam.’ Charisius: ‘...ortus sine adspiratione dici debere Varro ait... sed consuetudo... hortos cum adspiratione usurpauit.’
  12. ^ See the previous note. To the extent that ⟨h⟩ survived in this context, Allen supposes it was voiced. He provides no evidence for this but notes that English has [ɦ] for /-h-/.
  13. ^ Sturtevant points out that Latin poetry implies the existence of such consonants, with uolo for instance occurring as an iamb,[18] hence a disyllabic word rather than something like *u.o.lo In many cases the implied consonants correspond to Proto-Indo-European *[j w], to judge by cognates such as iugum, iuuenis and the English yoke, young. That they remained [j w] in Latin is corroborated by the fact that they were spelt with the letters ⟨i u⟩, which were also used for the vowels /i(ː) u(ː)/. Similarly they were transcribed into Greek as ⟨ι ου⟩ in Classical times. Evidence in favour of ⟨i⟩ standing for [j], and not some other consonant, includes the following. It changed to /i/ early on in cases like etiam < et i̯am. It could develop from /i/ in cases like insidantes (Ennius) or fluuorum (Virgil), a development also implied by Romance outcomes. Finally, Roman testimonies describe intervocalic ⟨i⟩ as a doubled consonant.[19] Evidence in favour of ⟨u⟩ standing for [w], and not some other consonant, includes the following. It combined with preceding ⟨a⟩ to form a diphthong in cases like fautor, naufragus after the following vowel was lost – cf. faueo, nauis. It was lost, apparently via dissimilation, before /o/ in cases like deorsum, parum < deuorsum, paruom. It developed from unstressed /u/ before a vowel in cases like solu̯o < *se-luo. In early poetry it still occurred as a vowel in cases like lārŭa (Plautus) and mīlŭos (Terence), and this apparently remained a possibility even in later times, as in silŭae (Horace). When borrowed early on into Proto-Germanic, it was approximated with the latter language's *[w], and not with *[β] or *[v], as in uīnum > Gothic wein, Old English wīn and uallum > Old English weall. Nevertheless, Roman inscriptions from the first century AD onward show a steadily increasing confusion of ⟨v⟩ and ⟨b⟩, implying the rise of a fricative pronunciation [β].

    Allen adds the following. That initial ⟨i⟩ before a vowel was a consonant is supported by its Romance outcomes, such as the Italian già, Old French ja [d͡ʒa] < Latin iam. That it was [j] as opposed to some other consonant is corroborated by conservative Romance outcomes like Spanish yace < Latin iacet and by early borrowings into Welsh like Ionawr < Ianuarius. That ⟨u⟩ stood for [w] is supported by cognates like uideō :: English wit (< Proto-Indo-European *[w]). Nigidius Figulus described uos, but not nos, as being pronounced with a forward protrusion of the lips. Varro gave an etymology for auris that linked it to auere. In the second century AD, Velius Longus describes consonantal ⟨u⟩ as involving some form of frication, which probably indicates [β]. In the fifth century AD Consentius appears to describe a survival of [w], among some speakers,[20] but the fricative pronuciation predominated to the extent that Priscian felt the need to explain when ⟨u⟩ should be written as opposed to ⟨b⟩ and vice-versa.

    Testimonies per Sturtevant. Cicero: ‘When Marcus Crassus was embarking his army at Brundisium, someone at the harbor who was selling figs imported from Caunus was crying "Cauneas". We may say, if we please, that Crassus was warned by him, "Caue ne eas", and that he would not have perished if he had obeyed the omen’. (That is, Cauneas sounded like Caue ne eas.) Quintilian: ‘... Is it in the power of every ear to distinguish accurately the sounds of the letters? No more, assuredly, than to distinguish the sounds of musical strings. But all at least who are grammarians will descend to the discussion of such fine points as these; whether any necessary letters be wanting to us, not indeed when we write Greek word, for then we borrow two letters from the Greeks, but letters that are proper to Latin words, as for example in seruus and uulgus the Aeologic digamma is needed.’ Consentius: ‘Some make the letter u too thin, so that when they say ueni you think a trisyllable is starting.’ Priscian: ‘...all nouns beginning with ui- are written with u, except bitumen and bilis (when it is equivalent to fel), and those that are compounds of the adverb bis, as biceps, bipatens, biuium.’

    Testimony per Di Napoli. Velius Longus, pages 32–3: ‘Among the vowels in our alphabet there are two letters which have the quality of both vowels and consonants, namely u and i. We shall briefly explain the difference... All syllables have a vowel at their core, whereas consonants are in an accessory position, such that changing the latter has no effect on the nature of the syllable. Just as it makes little difference whether we say ca or ba, it likewise makes little difference whether we say ia or ua. Since two adjacent vowels always make a syllable long, it follows that a syllable should become long from the addition of this letter [i or u] which is a consonant to the extent that it demonstrates that quality in verse. When I read ‘et iacit arrectae mentes stupefactaque corda’, the first foot et iacit is a dactyl, just as if I were to say et facit, substituting another consonant for the i in question. (It should be noted that at times this letter stands for two consonants, namely when it is placed between two vowels.) And so nobody will doubt that the verse ‘iam uitulos hortare uiamque insiste domandi’ is a proper one, iam uitu being a dactyl since the first syllable is long by position. It ends in the consonant m and is followed by u, which is positioned there as a consonant.’ Pages 39–40: ‘...but we should bear in mind that the letter u has the quality of a digamma not only in those words where it is pronounced with a kind of aspiration – as in ualens, primitiuus, genetiuus – but also in those where it is combined with q, as in quis.’
  14. ^ Sturtevant describes both as phonemes, comparing words like i̯am, adi̯iciō on the one hand and etiam, adiēns on the other. Allen does not comment on the matter. Lloyd states that their phonemic status is uncertain on the grounds that apparent minimal pairs like seruī/seru̯ī can be explained as a matter of syllable division. On the question of phonemicity he reports that Godel is undecided, Mariner is against, and Muljačić and Touratier are in favour. Weiss (page 67) indicates that ‘there is some debate’ on the matter and observes that [j w] were largely in complementary distribution with [i u]. As for the pairs like uoluī/uolu̯ī, he points out that there is an intervening morpheme boundary in the latter type but not the former: uol-uī/uolu̯-ī. McCullagh is confident that [j w] were phonemic ‘by the Classical period’ and adds a word-initial pair iambus/i̯am.

    Cser provides the following conspectus. ‘There is a tradition of analysing glides as position variants of the respective high vowels (see e.g. Hoenigswald 1949a; or more recently Ballester 1996; Marotta 1999; Lehmann 2005; Touratier 2005). While their environments are partly predictable, cases of contrast are far too numerous to be dismissed. Some of these cases can be explained away with reference to morphological structure... some cases clearly cannot: bel[u]a 'beast' vs. sil[w]a 'forest', q[w]i 'who/which' vs. c[u]i 'to whom/which', ling[w]a 'tongue' vs. exig[u]a 'small', co[i]t '(s)he meets' vs. co[j]tus 'meeting'...’
  15. ^ For the latter, Allen notes that the preceding syllable is scanned as heavy in words like Euander, Agaue, eoue and that these would have had [ww] in Greek as well.

    For the former, Sturtevant points out that several inscriptions feature the spelling ⟨ii⟩, that a syllable preceding intervocalic ⟨i⟩ always scanned as heavy, and that testimonies describe it as a doubled consonant. Allen adds that Italian outcomes like peggio for Latin peius have medial geminates as well and that Terentianus explicitly described the metrical effects of intervocalic ⟨i⟩. He notes that [jj] was variably lost before /i/ and cites Virgil as having reice as a trochee in one verse yet reicit as a dactyl in another, implying pronunciations like rei̯ce and re[jj]icit respectively. Doubling to [jj] apparently did not take place in compounds like bi-i̯urgus or quadri-i̯ugus, judging by their first syllable scanning as short. Allen adds that there is no reason to suppose that doubling took place after a prefix ending in a long vowel either, as in trā-i̯ectus or dē-i̯ero, but cites no evidence either way.

    Testimonies per Sturtevant. Quintilian: ‘Let [the grammarian] also know that Cicero approved of writing aiio and Maiia with double i; if this is done [the one i] will be joined [to the other] as a consonant.’ Velius Longus: ‘And in many particulars Cicero seems to have referred spelling to the standard of hearing; for when he thought that Aiiax and Maiia should be written with double i (certain writers observe that there is one i, if it possible to pronounce with one i, as we write). Wherefore that which would be pronounced more strictly and completely so, he thinks should be written with two i’s; so also Troiia, and all such words. Then the doubling increases, and in coiiicit we begin to write with three i’s, so that the first syllable is coi- and the following two -iicit. For if iacit is another word [i.e. if we compare iacit], a is substituted for i to carry the force of the vowel, while the first i remains, which has the force of a consonant.’ Priscian: ‘...the ancients used to double the same letter i and write maiius, peiius, eiius, which could not be pronounced unless the first i were spoken with the former syllable and the second with the following syllable, pei-ius, ei-ius, mai-ius; for, although a consonant, it could not be doubled in the same syllable; therefore it had to be pronounced in the same way as tellus and mannus. Whence they also wrote the genitive Pompeiii with three i s, the first two of which they understood as consonants, as if one should say Pompelli; for what kind of a syllable could be pronounced with three i s together?’
  16. ^ Note that the transcriptions given here are broad with respect to vowel qualities.

    Per Allen, the first syllable of compounds of iacio such as adicio, conbicio typically scanned as long, implying pronunciations like ad[ji]cio, con[ji]cio. Per Weiss (154–5), [j] was regularly lost before [i] early on, as in Plautus's obicias with a light initial syllable. Subsequently it was often restored – in pronunciation at least, if not in spelling – by analogy with conjugations where it had survived thanks to not having a following [i], for instance the preterite adieci. Weiss supposes that the restored [ji] underwent dissimilation to [je] judging by an inscription that features ⟨coniecant⟩ and two manuscripts that feature ⟨deiecit⟩ in the works of Virgil.[21]
  17. ^ The points of disagreement are as follows. 1) Sturtevant, Weiss, and Cser give Classical ⟨l⟩ as dark before all vowels except ⟨i⟩. Allen gives it as clear before all vowels but states that it had been dark before vowels other than ⟨i e⟩ in pre-Classical times. 2) Sturtevant and Allen give clear ⟨l⟩ as [l]. Weiss describes it as ‘palatal or apical’. Cser states it is highly probable that it was somewhat palatalized’.

    Testimonies per Sturtevant. Terentianus: ‘By pressing the opposite parts of the palate with its upper surface and vigorously opposing the sound of which it is itself the cause, the tongue produces a powerful sound within the mouth whereby may come to our ears the brilliant letter, to which our lists assigns second place’ Marius Victorinus: ‘Next will come l, which, with the tongue pressing the part of the palate where the roots of the upper teeth are, will send a powerful sound through the open mouth.’ Martianus Capella: ‘L with tongue and palate grows sweet.’ Priscian: ‘L has a triple sound, as Pliny thinks: thin when it stands second in double l, as ille, Metellus; full at the end of a word or a syllable and when it has a consonant before it in the same syllable, as sol, silua, flauus, clarus; intermediate in other words, as lectum, lectus.’ Consentius: ‘The Roman tongue has a correction to make in this matter also by way of distinction. For in some places the sound should be thicker, in others thinner; thicker when b follows, as in albus, or c, as in pulcher, or f, as in adelfi, or g, as in alga, or m, as in pulmo, or p, as in scalprum; but it should have a thinner pronunciation wherever a word begins with it, as in lepus, lana, lupus, or where in the same word the preceding syllable ends with this letter and the following begins with it, as ille and Allia.’
  18. ^ Testimonies per Sturtevant. Terentianus: ‘The third letter [m] may be said to moo within the closed mouth.’ Marius Victorinus: ‘With the lips pressed together m will sound like the mooing of cattle within the cavern of the mouth and the nostrils as well.’ Martianus Capella: ‘M is spoken with tight closed lips.’ Quintilian: ‘But the same letter m, when it ends a word and is in contact with an initial vowel of the next word so that it may coalesce with it, is written but scarcely pronounced, as multum ille and quantum erat; so that it gives the sound almost of a new letter. For it is not omitted but obscured and is merely a kind of a mark between two vowels to prevent them from combining.’ Velius Longus: ‘For we lack certain letters with pronunciation nevertheless demands; for example, when we say virtutem and virum fortem consulem Scipionem, you will find that what is virtually a foreign letter has come to the ears.’ Priscian: ‘M has a dull sound at the end of words, as in templum; a clear sound at the beginning, as in magnus; an intermediate sound in the interior, as in umbra.’
  19. ^ Sturtevant opts for a dental /n/ (as in French) based on Capella's testimony and parallelism with /d/. Allen gives it as dental or alveolar; Lloyd as alveolar; Weiss and McCullagh as dental. None of them elaborate on the matter.

    Testimonies per Sturtevant. Martianus Capella: ‘The tongue placed against the teeth makes the contact for n.’ Velius Longus: ‘Sometimes one should imitate the good taste of learned men, who have omitted certain letters for euphony, as Cicero who liked to say foresia, Megalesia, and hortesia without the letters n.’ Papirian (quoted by Cassiodorus): ‘One should write formosus without n in the second syllable, as arenosus, frondosus, aquosus, herbosus. But participles have n, as tonsus, tunsus, mensus, pensus. The ancients had no rule whether to write with n or without n; for they generally wrote tosus, tusus, prasus.’ Gellius: ‘In the same book [Nigidius Figulus] says: "Between n and g there is another sound, as in anguis, angari, ancorae,increpat, incurrit, ingenuus. In all these there is written not a genuine but a false n. For the tongue gives evidence that it is not n; for if it were that letter the tongue would touch the palate.’ Varro (quoted by Priscian): ‘As Ion writes, there is a twenty-fifth letter which they call agma, for which there is no character, but whose sound is common to the Greeks and the Latins, as in aggulus, agens, aguila, igerunt. In such words the Greeks and Accius write double g, others n and g, because in this matter it is not easy to see the truth. Similarly agceps, agcora.’
  20. ^ Sturtevant mentions that ⟨ph th ch⟩ first appear ‘in the latter part of the second century BC’ as renderings of the Greek ⟨φ θ χ⟩, which had previously been Latinized as ⟨p t c⟩. He explains the spread of ⟨ph th ch⟩ into native Latin words as a fad inspired by Greek.

    Allen argues that, while in some cases it may indeed have been a fad, in most cases it probably represented an internal Latin sound-change, as it nearly always occurred in the vicinity of /r/ or /l/, which is suggestive of a conditioned change.

    Clackson & Horrocks and McCullagh opt for the allophonic sound-change theory. Weiss opts for the fad theory.

    Testimony per Sturtevant. Cicero: ‘'Since I knew that our ancestors spoke so as to aspirate no sound but a vowel, I used to speak so as to say pulcer, Cetegi, triumpi, Cartago; finally after a long time, when the truth had been wrested from me by the clamor in my ears, I surrendered to the public in my habit of speech, and kept my knowledge for myself. Nevertheless I say Orciuii, Matones, Otones, Caepiones, sepulcra, coronae, lacrimae, because the criterion of hearing permits it.’
  21. ^ Allen is fairly certain of /kʷ ɡʷ/, Lloyd does not have a clear position, Weiss regards /kʷ/ as probable but /ɡʷ/ as uncertain, McCullagh considers the arguments on both sides inconclusive, and Csers leans towards /kw ɡw/.

    Per Cser, the (highly mixed) evidence is a follows. ⟨qu⟩ enjoyed a wide lexical distribution, while ⟨gu(V)⟩ was limited to a dozen or so words, where it was always preceded by /n/. The grammarian Velius Longus indicated that the ⟨u⟩ of ⟨qu⟩ was in some way different from /w/ in general. No geminate *⟨ququ⟩ is attested, whereas all (other) Latin stops are also found as geminates. Sequences of obstruent + glide are rare in Classical Latin. In poetry, whenever sequences of stop + glide occur in medial position, the scansion reveals that they are split across syllables, but this is never the case for ⟨qu⟩. Neither ⟨qu⟩ nor ⟨gu⟩ are ever followed by a consonant, unlike any (other) Latin stop, nor can they occur word-finally. The voicing contrast between ⟨nqu⟩ and ⟨ngu⟩ is not found in any (other) sequence of three consonants. Assimilation of the prefix ad- to a following ⟨qu⟩ is relatively rare, which is also the case when ad- is followed by a consonant cluster. The Proto-Indo-European predecessor of Latin ⟨qu⟩ is, in many cases, reconstructed as a single consonant */kʷ/, notably distinct from sequences of */kw/. Occasionally Latin /w/ scans as a vowel in poetry, when preceded by /s/ or /l/, but this is never the case for the ⟨u⟩ of ⟨qu⟩.
  22. ^ Per Allen, this is suggested by spellings like ⟨ecus⟩ and ⟨ungunt⟩, for equus and unguunt. Alternatively, he says, such spellings may suggest that /k ɡ/ were in general labialized by a following /u/. If so, it would have been unnecessary to write a double ⟨uu⟩ in this context.
  23. ^ Per Allen, this is reported by Priscian and corroborated by Classical transcriptions of Latin into Greek, for instance ⟨кυι⟩ for qui as opposed to ⟨κουα⟩ and ⟨κοα⟩ for Latin qua. Weiss, however, cites ⟨Κόϊντος⟩ for Quintus.
  24. ^ Sturtevart and Weiss gives it as an alveolar trill based on the testimonies and Romance outcomes. Allen gives it as an alveolar trill based on the testimonies, adding that Old Latin may have had other rhotic sounds as well, for instance a tap in forms like ⟨arvorsum⟩ for advorsum or some fricative during the development from Old Latin intervocalic ⟨s⟩ to the Classical ⟨r⟩. Lloyd at first states that Classical ⟨r⟩ ‘was doubtlessly an alveolar flap’ without elaborating. Later he adds that ‘it is likely that in Latin the word-initial vibrant was also a simple trill’ based on Romance outcomes. McCullagh gives /r/ as a trill and Cser states ‘it is uncertain whether it was a flap or a trill’; neither of them elaborate.

    Wireback argues that ⟨r⟩ was [r] initially and in other ‘strong’ contexts but a variable [r~ɾ] intervocalically and in other 'weak' contexts... (to be expanded).

    Testimonies per Sturtevant. Lucilius: ‘This cacophonous r isn't much different from saying in dog's language, "It's nothing to me."’ Persius: ‘Hereupon the dog's letter sounds through the nose.’ (That is, it makes a snarling noise.) Terentianus: ‘The next letter [r] shakes out a dry sound with rapid blows. Marius Victorinus: ‘R is scraped forth while the tongue puts breath into tremulous motion.’
  25. ^ Allen cites the non-etymological ⟨rh⟩ in the river-name Rhenus < Celtic renos.
  26. ^ Sturtevant entertains both the possibility of /s/ being dental (as it is French) and of it being alveolar. In the end he opts for a dental on the grounds that testimonies describe a sound made against the teeth or made by striking them. Allen, on the other hand, opts for an alveolar on the grounds that testimonies describe ‘a constriction behind the teeth’. Weiss and McCullagh opt for an alveolar; neither of them elaborate.

    Lloyd argues that /s/ could have been apico-alveolar based on its being so in various modern Romance dialects, especially in Iberia, and he cites Martinet as proposing that [s̺] dates back to Proto-Indo-European. Lloyd also argues that the Old Latin rhotacism would be better explained if the starting sound were apico-alveolar. He also points out that in most cases medieval Ibero-Romance /s/ was transliterated into Arabic with the spelling used for /ʃ/, although transliterations with the spelling used for /s/ are also attested. On the other hand Basque, which distinguishes /s/ from /s̺/, consistently shows /s/ in borrowings from Latin. Overall, Lloyd's assessment is that ‘the exact phonetic realization of (Latin) /s/ is not known for certain... it could have varied considerably from one region to another and over time without there being any difference in its phonemic status or in writing’.

    Testimonies per Sturtevant. Terentianus: ‘Then the final letters [in the list, i.e. s and x] are seen to cause a similar hissing against the teeth; still, just as for the former [i.e. s] the contact in the mouth is immediate [whereas for x it is preceded by the sound of c...] and it is formed behind the teeth, so it is smooth and brings to the ears an unchanging whisper.’ Marius Victorinus: ‘Then the final [letters in the list], s and x, will properly be joined. For with similar sound they hiss through a contracted opening; provided, nevertheless, that while the impulse of the former begins behind the teeth...’ Martianus Capella: ‘S makes a hiss by lashing the teeth.’ Cledonius: ‘S... is a hiss rather than a consonant.’ Cicero: ‘In fact, from those words whose last two letters were the same as in optimus they used to take away the last letter unless a vowel followed, a pronunciation which now seems rather boorish but was once the more fashionable. And so that stumbling block in versification which our modern poetrs try to avoid did not exist. For we said qui est omnibu’ princeps, omnibus princeps, and uita illa dignu’ locoque, not dignus.’ Quintilian: ‘And this, as I have said, was Servius' reason for taking away s whenever it was final and followed by another consonant; which practice Luranius blames and Messala defends. For they think that Lucilius does not retain final s when he says Aeserninu’ fuit and dignu’ locoque; and Cicero in his Orator records that the majority of the ancients spoke thus.’
  27. ^ Sturtevant and Allen argue this on the basis of ⟨-s-⟩ consistently being transcribed into Greek as ⟨σ⟩, not as ⟨ζ⟩. Cf. ⟨Καισαρ⟩ for Caesar. Sturtevant points out that intervocalic ⟨s⟩ has voiceless reflexes in Romanian and southern Italian, such that the voicing to [z] seen in languages like French likely dates to a later period.
  28. ^ Testimonies per Sturtevant. Terentianus: ‘A combination produces the double sound of x, because c and s by uniting their functions, check the voice, strengthen it, and compress it.’ Marius Victorinus ‘[X] is formed by the combination of c and s, whose place it takes and whose force it represents, as we shall be convinced by hearing.’ Martianus Capella: ‘X hisses out what c and s have formed.’ Diomedes: ‘X is a composite letter, which we call double because it consists of c and s.’
  29. ^ Per Sturtevant, the Greek ⟨ζ⟩ was originally rendered in Latin borrowings as ⟨s⟩ initially and ⟨ss⟩ intervocalically, as in sona, massa < ζώνη, μάζα. The Greek letter was then borrowed outright as ⟨z⟩ ‘in the second or first century BC’ and presumably assigned the Koiné pronunciation /z/. Still, Sturtevant says, a pronunciation like /d͡z/ appears to have prevailed in the Greek of southern Italy, which may explain why, once Late Latin happened to develop that sound via palatalization of [dj], it chose ⟨z⟩ to represent it.

    Allen adds that in some cases ⟨z⟩ was used to represent Greek ⟨σ⟩ with a following voiced consonant, as in ⟨zmaragdus azbestus lezbia⟩, a spelling apparently censured by Priscian.

    Of the cited sources only McCullagh indicates a point of articulation for Latin ⟨z⟩, namely alveolar, but does so without elaborating.

    Testimony per Sturtevant (1920 edition, pages 116–7). Velius Longus: ‘To me it seems that z was not foreign to the Latin language, since it is found in the Carmen Saliare, and I think that ζ is one thing, and σδ another, and that they do not have the same value or the same sound, but that they are pronounced according to the several dialects. For we know that the Dorians say μελίσδειν, and even παίζειν is spoken as παίσδειν in certain places... Finally, if anyone wants to investigate this letter by the natural method, he will find that it is not a composite letter, if only he tests it with an unprejudiced ear. For it can have one sound when written singly, another when doubled, which does not happen at all to a composite letter. For, write with one ζ and consult the ear; ἀζηχής will not sound like ἀδσηχής, but if ζ is doubled, ἀζζηχής will sound like ἀσσηχής. And clearly, whatever sound is added, while I prolong the sound of this letter, you will find that it is the same tone with which it began.’
  30. ^ Allen observes that a syllable followed by intervocalic ⟨z⟩ always scans as long and in addition cites the following testimony. Probus: ‘Z... quoniam duplex est facit positione longam’.

    Clackson & Horrocks and McCullagh concur on [-zz-] without elaborating.
  31. ^ So theorizes Allen, reasoning that this would explain why various Greek words with ⟨π κ⟩ were borrowed into Latin with ⟨b g⟩ instead of ⟨p t⟩, as in κυβερνῶ > guberno. He adds that the phenomenon is especially character of both pre-Classical Latin, as in Ennius's Burrus for Πύρρος, and post-Classical Latin, cf. πλαστα- > blasta in Appendix Probi) and ἀποθήκη > *botteca, the latter inferred from Romance outcomes like the Italian bottega. On the other hand Allen concedes that Latin /p t k/ were generally rendered in Greek as ⟨π τ κ⟩, rather than the corresponding aspirates, and that most Romance outcomes are unaspirated.

    Per Lloyd /p t k/ were ‘probably unaspirated’ based on the Romance outcomes, and per Weiss they were ‘not aspirated’ based on the Romance outcomes and transcriptions into Greek.
  32. ^ Such is the communis opinio of modern scholarship, as reported by both Cser and Leppänen & Alho. On the other hand, Calabrese argues that none of the pairs differed by quality and describes them as [i(ː) ɛ(ː) ɔ(ː) u(ː)]. This is a fringe theory, although it did manage to convince Weiss.
  33. ^ Per Sen, ⟨ꟾ⟩ is came into use around 110 BC and ⟨é á ó ú⟩ ‘towards the end of the Republican period’. In the second century AD there also appeared ⟨í⟩. All of these were used sporadically at most and went obsolete ca. 300 AD.
  34. ^ Per McCullagh, ⟨y⟩ is attested in Latin beginning in the first century BC. Previously Greek ⟨υ⟩ was adapted to Latin ⟨u⟩ in borrowings, and this remained the case through the Classical period in ‘vulgar’ registers. Allen indicates that, beginning in the second or third century AD, ⟨υ⟩ developed in some varieties of Greek to ⟨ι⟩. This is often reflected in Late Latin, for instance in the proscribed variant ⟨girus⟩ attested in the Appendix Probi (for proper gyrus).
  35. ^ The Liber Usualis describes /s/ as ‘slightly softened when coming between two vowels’. Canepari states that ‘[Italian Latin] rigorously has intervocalic s /VzV/’.
  36. ^ ‘In ancient books these words are often writen nichil and michi’ per the Liber Usualis. The spelling with ⟨ch⟩ is attested from Late Latin, where it reflects a learned adaptation of the Greek ⟨χ⟩, by then a fricative sound, to represent the lost Latin /h/ (Allen 1978:45).
  37. ^ Further examples: china, Kýrie, qui. Note that ⟨q⟩ occurs only with a following ⟨u⟩ /w/.
  38. ^ When followed by /j/ and not preceded by /s/ or another /t/.
  39. ^ Also ⟨xcᶠᵛ⟩ /kʃ/, as in excélsis /ɛkˈʃɛlsis/.
  40. ^ Canepari states that ‘length and the various vowel sequences also correspond to those found in neutral Italian’. Cf. his transcriptions [aˈbɛːɾe], [ˈrɔːma] habere, Roma.
  41. ^ Converted to standard IPA using the key provided in Canepari 2005:42–5. For simplicity, [ä] is represented as [a] and markers of tone are omitted.

References

[edit]
  1. ^ McCullagh 2011:85
  2. ^ Sturtevant 1940:173–4; Allen 1978:12–3, 21–2; Lloyd 1987:77, 239; Weiss 2009:66, 512; McCullagh 2011:84; Cser 2020:14, 172–6
  3. ^ Sturtevant 1940:165–9; Allen 1978:14–5, 22–3; Lloyd 1987:77, 137; Weiss 2009:58–9, 513; McCullagh 2011:84–5; Cser 2020:14
  4. ^ For which see the note below.
  5. ^ For which see ⟨ph th ch⟩ below.
  6. ^ For which see ⟨n⟩ below.
  7. ^ Sturtevant 1940:170–3; Allen 1978:13–4, 21–2; Lloyd 1987:35–6; Di Napoli 2008:41–2; Weiss 2009:58–9; McCullagh 2011:84; Cser 2020:13, 171–2
  8. ^ Sturtevant 1940:162–3; Allen 1978:34–5; Lloyd 1987:80; Weiss 2009:57, 59; McCullagh 2011:84; Gouvert 2016:§1
  9. ^ Sturtevant 1940:55; Allen 1978:23–5; Lloyd 1987:81; Weiss 2009:59; McCullagh 2011:87; Baglioni 2014; Cser 2020:15–6
  10. ^ Compare conscendo < *con-scando, without raising to /i/.
  11. ^ Allen & Greenough 2001:§10d
  12. ^ Allen 1978:71–3
  13. ^ As mentioned in the main note above.
  14. ^ For cognitum.
  15. ^ Sturtevant 1940:155–60; Allen 1978:43–5; Lloyd 1987:77; Weiss 2009:62–3, 66; McCullagh 2011:86–7
  16. ^ For which see ⟨ph th ch⟩ below.
  17. ^ Sturtevant 1940:140–7; Allen 1978:37–42; Lloyd 1987:78; Di Napoli 2008:32–3; Weiss 2009:67, 154–5; McCullagh 2011:87–8; Cser 2020:14–5, 17
  18. ^ This example has been taken from the 1920 edition of Sturtevant's work, p. 39.
  19. ^ For which see the note below.
  20. ^ Adams clarifies that Consentius' comment was in reference to Greeks, who were known for this kind of pronunciation in Latin.
  21. ^ The Codex Mediceus and the Codex Romanus.
  22. ^ Allen 1978:15; Cser 2020:14
  23. ^ Sturtevant 1940:147–50; Allen 1978:33–4; Giannini & Marotta 1989:235–40 Weiss 2009:117; Sen 2015:§2; Cser 2020:96–7
  24. ^ Sturtevant 1940:151–3; Allen 1987:30–1
  25. ^ Sturtevant 1940:153–5; Allen 1978:27–30; Lloyd 1987:81; McCullagh 2011:84; Cser 2020:13
  26. ^ Sturtevant 1940:157–9; Allen 1978:26–7; Clackson & Horrocks 2007:190; Weiss 2009:59; McCullagh 2009:86
  27. ^ Allen 1978:16‒7, 19–20; Lloyd 1987:77–8; Weiss 2009:58–9; McCullagh 2011:86; Cser 2020:§2.2.2
  28. ^ Sturtevant 1940:150–1; Allen 1978:32–3; Lloyd 1987:80–1, 244–5; Weiss 2009:62; McCullagh 2011:84; Wireback 2015; Cser 2020:49
  29. ^ Sturtevant 1940:160–2; Allen 1978:35–7; Lloyd 1987:80; Weiss 2009:60; McCullagh 2011:84
  30. ^ Sturtevant 1940:174–5; Allen 1978:45; McCullagh 2011:85
  31. ^ Sturtevant 1920:115; Allen 1978:46; Clackson & Horrocks 2007:190; McCullagh 2011:84, 86
  32. ^ Allen 1978:12–3; Lloyd 1987:80; Weiss 2009:58
  33. ^ Sturtevant 1920:15–36; Allen 1978:49–50; Lloyd 1987:73–5; Calabrese 2005; Weiss 2009:64; McCullagh 2011:88; Sen 2015:13; Leppänen & Alho 2018:2; Cser 2020:34
  34. ^ Allen 1978:64–5; Sen 2015:12
  35. ^ Sturtevant 1920:36–8; Allen 1978:52–3; McCullagh 2011:89
  36. ^ a b c d e Canepari 2021:189
  37. ^ Benedictines of Solesmes 1961:xxxviij; Canepari 2021:189
  38. ^ Benedictines of Solesmes 1961:xxxviij
  39. ^ a b Benedictines of Solesmes 1961:xxxvij–xxxviij
  40. ^ Canepari 2021:200–1

Bibliography

[edit]
  • Adams, James Noel (2003). Bilingualism and the Latin language. Cambridge University Press.
  • Allen, William Sidney (1978) [1965]. Vox Latina: A guide to the pronunciation of Classical Latin (2nd ed.). Cambridge University Press.
  • Allen, Joseph A.; Greenough, James B. (2001) [1903]. Mahoney, Anne (ed.). New Latin grammar for schools and colleges. Newburyport, Massachusetts: R. Pullins Company.
  • Baglioni, Daniele (2014). "Il nesso GN dal latino alle lingue romanze: Questioni aperte e prospettive di ricerca". In Molinelli, Piera; Cuzzolin, Pierluigi; Fedriani, Chiara (eds.). Actes du Xᵉ colloque international sur le latin vulgaire et tardif. Bergamo University Press. pp. 3–24.
  • Benedictines of Solesmes, ed. (1961). The Liber Usualis. New York: Desclee.
  • Brittain, Frederick (1955). Latin in church: The history of its pronunciation (2nd ed.). Mowbray.
  • Calabrese, Andrea (2005). "On the feature [ATR] and the evolution of the short high vowels of Latin into Romance". University of Connecticut Working Papers in Linguistics. 13: 33–78.
  • Canepari, Luciano (2005). A handbook of pronunciation. Lincom GmbH.
  • Canepari, Luciano (2021). Latin pronunciation & ancient & modern accents: Geo-social applications of the natural phonetics & tonetics method. Lincom GmbH.
  • Clackson, James; Horrocks, Geoffrey (2007). The Blackwell history of the Latin language. Oxford: Blackwell.
  • Clackson, James (2008). "Latin". In Roger D. Woodard (ed.). The ancient languages of Europe. Cambridge University Press.
  • Cser, András (2020). "The phonology of Classical Latin". Transactions of the Philological Society. 118: 1–218. doi:10.1111/1467-968X.12184.
  • Di Napoli, Marta (2008). Introduzione, testo critico, traduzione e note del De orthographia di Velio Longo (PhD thesis). Università Roma Tre.
  • Fortson, Benjamin W. IV (2004). Indo-European Language and Culture: An Introduction. Wiley.
  • Gilbert, Allan H (1939). "Mock accents in Renaissance and Modern Latin". Publications of the Modern Language Association of America. 54 (2): 608–10. doi:10.2307/458579.
  • Gouvert, Xavier (2016). "Du protoitalique au protoroman: deux problèmes de reconstruction phonologique". In Buchi, Éva; Schweickard, Wolfgang (eds.). Dictionnaire Étymologique Roman 2. Berlin: De Gruyter. pp. 27–51.
  • Hayes, Bruce (1995). Metrical stress theory: principles and case studies. University of Chicago.
  • Leppänen, Ville; Alho, Tommi (2018). "On the mergers of Latin close‐mid vowels". Transactions of the Philological Society. 116 (3): 460–83. doi:10.1111/1467-968X.12130.
  • Levy, Harry Louis (1989). A Latin reader for colleges. University of Chicago Press.
  • Lloyd, Paul Max (1987). From Latin to Spanish: Historical phonology and morphology of the Spanish language. Philadelphia: American Philosophical Society.
  • McCullagh, Matthew (2011). "The sounds of Latin: Phonology". In James Clackson (ed.). A companion to the Latin language. Blackwell.
  • Sen, Ranjan (2015). Syllable and segment in Latin. Oxford University Press.
  • Smith, Jane Stuart (2004). Phonetics and philology: Sound change in Italic. Oxford University Press.
  • Sturtevant, Edgar Howard (1940) [1920]. The pronunciation of Greek and Latin (2nd ed.). Philadelphia: Linguistic Society of America.
  • Ward, Ralf L. (1962). "Evidence for the pronunciation of Latin". The Classical World. 55 (9): 273–5. doi:10.2307/4344896.
  • Weiss, Michael (2009). Outline of the historical and comparative grammar of Latin. New York: Beech Stave Press.
  • Wingo, Elvis Otha (1972). Latin punctuation in the Classical age. De Gruyter.
  • Wireback, Kenneth James (2002). "On the metathesis of labials + /j/ in Hispano-Romance". Hispanic Review. 70 (3): 311–31. doi:10.2307/3247205.

Further reading

[edit]