Editing Latin phonology and orthography (section)

====Phonetics====
* Latin may have had the labialized velar stops {{IPA|/kʷ/}} and {{IPA|/ɡʷ/}} as opposed to the stop + semivowel sequences {{IPA|/kw/}} and {{IPA|/ɡw/}} (as in the English '''''qu'''ick'' or ''pen'''gu'''in''). The argument for {{IPA|/kʷ/}} is stronger than that for {{IPA|/ɡʷ/}}.{{efn|{{Harvnb|Allen|1978}} (p. 17) judges the evidence to favour {{IPA|/kʷ/}} and {{IPA|/ɡʷ/}}, while {{Harvnb|Cser|2020}} (§2.2.2) comes to the opposite conclusion. The relevant facts, per the latter, are as follows:<br /><br />{{angbr|qu}} enjoyed a wide lexical distribution, while {{angbr|gu(V)}} was limited to a dozen or so words, where it was always preceded by {{IPA|/n/}}. The grammarian [[Velius Longus]] indicated that the {{angbr|u}} of {{angbr|qu}} was in some way different from {{IPA|/w/}} in general. No geminate *{{angbr|qqu}} 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 can be split across syllables, but this is never the case for {{angbr|qu}}. Neither {{angbr|qu}} nor {{angbr|gu}} are ever followed by a consonant, unlike any (other) Latin stop, nor can they occur word-finally. The voicing contrast between {{angbr|nqu}} and {{angbr|ngu}} is not found in any (other) sequence of three consonants. Assimilation of the prefix {{lang|la|ad-}} to a following {{angbr|qu}} is relatively rare, which is also the case when {{lang|la|ad-}} is followed by a consonant cluster. The Proto-Indo-European predecessor of Latin {{angbr|qu}} is, in many cases, reconstructed as a single consonant *{{IPA|/kʷ/}}, notably distinct from sequences of *{{IPA|/kw/}}. Occasionally Latin {{IPA|/w/}} scans as a vowel in poetry, when preceded by {{IPA|/s/}} or {{IPA|/l/}}, but this is never the case for the {{angbr|u}} of {{angbr|qu}}.}}
* The former could occur between vowels, where it always counted as a single consonant in Classical poetry, whereas the latter only occurred after {{IPA|/n/}}, where it is impossible to tell whether it counted as one consonant or two.<ref name="Allen gm and gw">{{Harvnb|Allen|1978|p=25}}</ref> The labial element, whether {{IPA|[ʷ]}} or {{IPA|[w]}}, appears to have been palatalised before a front vowel, resulting in {{IPA|[ᶣ]}} or {{IPA|[ɥ]}} (for instance {{lang|la|quī}} would have sounded something like {{audio|La-cls-qui.ogg|listen}}). This palatalisation did not affect the independent consonant {{IPA|/w/}} before front vowels.<ref>{{Harvnb|Allen|1978|p=17}}</ref>
*{{IPA|/kʷ/}} and {{IPA|/ɡʷ/}} before {{IPA|/u/}} were not distinct from {{IPA|/k/}} and {{IPA|/ɡ/}}, which were allophonically labialized to {{IPA|[kʷ]}} and {{IPA|[ɡʷ]}} by a following {{IPA|/u/}} such that writing a double {{angbr|uu}} was unnecessary. This is suggested by the fact that {{lang|la|equus}} and {{lang|la|unguunt}} (from Old Latin {{lang|la|equos}} and {{lang|la|unguont}}) are also found spelt as {{lang|la|ecus}} and {{lang|la|ungunt}}.<ref>{{Harvnb|Allen|1978|pp=19, 20}}</ref>
* {{IPA|/p/}}, {{IPA|/t/}} and {{IPA|/k/}} were less aspirated than the corresponding English consonants, as implied by their usually being transliterated into Ancient Greek as {{angbr|{{lang|grc|π}}}}, {{angbr|{{lang|grc|τ}}}} and {{angbr|{{lang|grc|κ}}}}, and their pronunciation in most Romance languages. In many cases, however, it was not the Latin {{IPA|/p/}} and {{IPA|/k/}}, but rather {{IPA|/b/}} and {{IPA|/ɡ/}}, that were used to render Greek word-initial {{IPA|/p/}} and {{IPA|/k/}} in borrowings (as in {{lang|grc|πύξος}}, {{lang|grc|κυβερνῶ}} > {{lang|la|buxus}}, {{lang|la|guberno}}), especially borrowings of a non-learned character. This might suggest that the Latin {{IPA|/p/}} and {{IPA|/k/}} had some degree of aspiration, making {{IPA|/b/}} and {{IPA|/ɡ/}} more suitable to approximate the Greek sounds.<ref name="tenues">{{Harvnb|Allen|1978|pp=12–13}}</ref>
* {{IPA|/pʰ/}}, {{IPA|/tʰ/}} and {{IPA|/kʰ/}} were pronounced with notable aspiration, like the initial consonants of the English '''''p'''ot'', '''''t'''op'', and '''''c'''ot'' respectively. They are attested beginning c. 150 BC, in the spellings {{angbr|ph}}, {{angbr|th}} and {{angbr|ch}}, at first only used to render the Greek {{angbr|{{lang|grc|φ}}}}, {{angbr|{{lang|grc|θ}}}} and {{angbr|{{lang|grc|χ}}}} in loanwords. (Previously these had been rendered in Latin as {{angbr|p}}, {{angbr|t}} and {{angbr|c}}.) From c. 100 BC onward {{angbr|ph}}, {{angbr|th}} and {{angbr|ch}} spread to a number of native Latin words as well, such as {{lang|la|pulcher}} and {{lang|la|lachrima}}. When this occurred it was nearly always in the vicinity of the consonant {{IPA|/r/}} or {{IPA|/l/}}, and the implication is that Latin {{IPA|/p/}}, {{IPA|/t/}} and {{IPA|/k/}} had become aspirated in that context.<ref name="aspirate1">{{Harvnb|Allen|1978|pp=26, 27}}</ref><ref name="aspirate2">{{Harvnb|Clackson|Horrocks|2007|p=190}}</ref>
* {{IPA|/z/}} was found as a rendering of the Greek {{angbr|{{lang|grc|ζ}}}} in borrowings starting around the first century BC. (In earlier borrowings, the Greek sound had been rendered in Latin as {{IPA|/ss/}}.) In initial position this appears to have been pronounced {{IPA|[z]}}, and between vowels it appears to have been [[gemination|doubled]] to {{IPA|[zz]}} (counted as two consonants in poetry).<ref>{{Harvnb|Levy|1989|p=150}}</ref><ref name="Allen Z">{{Harvnb|Allen|1978|pp=46}}</ref>
* {{IPA|/s/}} was unvoiced in all positions in Classical Latin. Previously however Old Latin {{IPA|/s/}} appears to have voiced to {{IPA|[z]}} between vowels, ultimately [[Rhotacism|turning to]] {{IPA|/r/}}. Cicero reports the family-name {{lang|la|Papisius}} being changed to {{lang|la|Papirius}} in the fourth century BC, which may give some idea of the chronology. Afterward new instances of {{IPA|/s/}} developed between vowels from sound-changes like the degemination of {{IPA|/ss/}} after long vowels and diphthongs (as in {{lang|la|caussa}} > {{lang|la|causa}}), which [[Quintilian]] reports to have happened a little after the time of [[Cicero]] and [[Virgil]].<ref name="Allen s">{{Harvnb|Allen|1978|pp=35–37}}</ref>
*In Old Latin, final {{IPA|/s/}} after a short vowel was often lost, probably after first [[Debuccalization|debuccalizing]] to {{IPA|[h]}}, as in the inscriptional form {{lang|la|Cornelio}} for {{lang|la|Cornelios}} (Classical {{lang|la|Cornelius}}). Often in the poetry of [[Plautus]], [[Ennius]], and [[Lucretius]], final {{IPA|/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.<ref name="Allen s" />
* {{IPA|/f/}} was [[labiodental]] in Classical Latin but may have been a [[bilabial]] {{IPA|/ɸ/}} in Old Latin,<ref>{{Harvnb|Allen|1978|pp=34, 35}}</ref> or perhaps {{IPA|[ɸ]}} in free variation with {{IPA|[f]}}. Lloyd, Sturtevant, and Kent make this argument based on misspellings in early inscriptions, the fact that many instances of Latin {{IPA|/f/}} descend from [[Proto-Indo-European]] *{{PIE|/bʰ/}}, and the outcomes of the sound in Romance (particularly in Spain).<ref>{{Harvnb|Lloyd|1987|p=80}}</ref>
* In most cases {{IPA|/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 [[nasal vowel|nasalization]] on the preceding vowel<ref name="Lloyd 81" /> (such that {{lang|la|decem}} may have sounded something like {{audio|La-cls-decem.ogg|listen}}, i.e. {{IPA|[ˈdɛkẽː]}}). In Old Latin inscriptions, final {{angbr|m}} is often omitted, as in {{angbr|viro}} for {{lang|la|virom}} (Classical {{lang|la|virum}}). It was frequently elided before a following vowel in poetry and lost without a trace (apart from perhaps lengthening) in the Romance languages,<ref name="Allen -Vm">{{Harvnb|Allen|1978|pp=30, 31}}</ref> except in a number of monosyllabic words, where it often survives as {{IPA|/n/}} or a further development thereof.
* {{IPA|/n/}} and {{IPA|/m/}} [[Phoneme#Neutralization and archiphonemes|merged]] via assimilation before a following consonant, with the following consonant determining the resulting pronunciation: bilabial {{IPA|[m]}} before a bilabial consonant (e.g. {{IPA|/p/}} and {{IPA|/b/}}), coronal {{IPA|[n]}} before a coronal consonant (e.g. {{IPA|/t/}} and {{IPA|/d/}}) and velar {{IPA|[ŋ]}} before a velar consonant (e.g. {{IPA|/k/}}, {{IPA|/kʷ/}} and {{IPA|/ɡ/}}). This occurred both within words (e.g. {{lang|la|quī'''nq'''ue}} may have sounded something like {{audio|La-cls-quinque.ogg|listen}}) and across word-boundaries (for instance {{lang|la|i'''n''' causā}} with {{IPA|[ŋ]}}, or {{lang|la|i'''m''' pace}}).<ref name="Allen n">{{Harvnb|Allen|1978|pp=27–30}}</ref>
* {{IPA|/ɡ/}} assimilated to a [[velar nasal]] {{IPA|[ŋ]}} before {{IPA|/n/}}.<ref>{{Harvnb|Allen|1978|pp=23–25}}</ref> Allen and [[James B. Greenough|Greenough]] say that a vowel before {{IPA|[ŋn]}} is always long,<ref>{{Harvnb|Allen|Greenough|2001|loc=§10d}}</ref> but W. Sidney Allen says that is based on an [[interpolation (manuscripts)|interpolation]] in [[Priscian]], and the vowel was actually long or short depending on the root, as for example {{lang|la|rēgnum}} from the root of {{lang|la|rēx}} but {{lang|la|magnus}} from the root of {{lang|la|magis}}.<ref name="Vgn">{{Harvnb|Allen|1978|pp=71–73}}</ref> {{IPA|/ɡ/}} probably did not assimilate to {{IPA|[ŋ]}} before {{IPA|/m/}}. The cluster {{IPA|/ɡm/}} arose by [[syncope (phonetics)|syncope]], as for example {{lang|la|tegmen}} from {{lang|la|tegimen}}. Original {{IPA|/ɡm/}} developed into {{IPA|/mm/}} in {{lang|la|flamma}}, from the root of {{lang|la|flagrō}}.<ref name="Allen gm and gw" /> At the start of a word, original {{IPA|/ɡn/}} was reduced to {{IPA|[n]}}, and this change was reflected in the orthography of later texts, as in {{lang|la|gnātus}}, {{lang|la|gnōscō}} > {{lang|la|nātus}}, {{lang|la|nōscō}}.
* In Classical Latin, the [[rhotic consonant|rhotic]] {{IPA|/r/}} was most likely an [[alveolar trill]] {{IPA|[r]}}, at least 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. In Old Latin, intervocalic {{IPA|/z/}} developed into {{IPA|/r/}} ([[Rhotacism (sound change)|rhotacism]]), suggesting an approximant like the English {{IPAblink|ɹ}}, and {{IPA|/d/}} was sometimes written as {{angbr|r}}, possibly suggesting a tap {{IPAblink|ɾ}} (like the single {{IPA|/ɾ/}} in Spanish).<ref>{{Harvnb|Allen|1978|p=33}}</ref>
* {{IPA|/l/}} was strongly [[Velarization|velarized]] in [[syllable coda]] and probably somewhat [[Palatalization (phonetics)|palatalized]] when [[Gemination|geminated]] or followed by {{IPA|/i(ː)/}}. In [[Intervocalic consonant|intervocalic]] position, it appears to have been velarized before all vowels except {{IPA|/i(ː)/}}.<ref>{{Harvnb|Cser|2020|loc=§4.9}}. In footnote 206, he adds: "The evidence has been thoroughly assessed in the diachronic literature; see Sen (2012: 472–3; 2015: 15 sqq.), Meiser (1998: 68–9), Leumann (1977: 85–7)."</ref>
* {{IPA|/j/}} generally appeared only at the beginning of words, before a vowel, as in {{lang|la|iaceō}}, except in compound words such as {{lang|la|adiaceō}} (pronounced something like {{audio|La-cls-iaceo, adiaceo.ogg|listen}}). Between vowels, it was generally as a geminate {{IPA|/jj/}}, as in {{lang|la|cuius}} (pronounced something like {{audio|La-cls-cuius.ogg|listen}}) except in compound words such as {{lang|la|trāiectus}}. This {{IPA|/jj/}} is sometimes marked in modern editions by a [[circumflex]] on the preceding vowel, e.g. {{lang|la|cûius}}, {{lang|la|êius}}, {{lang|la|mâior}}, etc. {{IPA|/j/}} could also have varied with {{IPA|/i/}} in the same [[morpheme]], as in {{lang|la|iam}} {{IPA|/jam/}} and {{lang|la|etiam}} {{IPA|/ˈe.ti.am/}}, and in poetry one could be replaced with the other for [[metre (poetry)|metrical]] purpose.<ref name="Allen j">{{Harvnb|Allen|1978|pp=37–40}}</ref>
* {{IPA|/w/}} was pronounced as an approximant until the first century AD, when {{IPA|/w/}} and intervocalic {{IPA|/b/}} began to develop into fricatives. In poetry, {{IPA|/w/}} and {{IPA|/u/}} could be replaced with each other, as in {{IPA|/ˈsilua/}}~{{IPA|/ˈsilwa/}} or {{IPA|/ˈɡenua/}}~{{IPA|/ˈɡenwa/}}. Unlike {{IPA|/j/}} it remained a single consonant in most words, e.g. in {{lang|la|cavē}}, although it did represent a double {{IPA|/ww/}} in borrowings from Greek such as the name {{lang|la|Evander}}.<ref name="Allen v">{{Harvnb|Allen|1978|pp=40–42}}</ref>
* {{IPA|/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 ({{lang|la|diribeō}} < *{{lang|la|dis-habeō}} being a particularly early example). Where intervocalic {{IPA|/h/}} survived, it was likely voiced<ref name="Allen h">{{Harvnb|Allen|1978|pp=43–45}}</ref> (that is, {{IPA|[ɦ]}}).