Jump to content

Talk:Latin phonology and orthography/Archive 4

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
Archive 1Archive 2Archive 3Archive 4

Vowel pronunciation

The vowel pronunciation table lists two pronunciations for each vowel.

For example, for ⟨e⟩ it lists [ɛ] and [eː].

I do not dispute this, these seem to be the rules I remember, but no citations are given. So how exactly do we know (or guess) that these differences existed? And are only vowels which are long by nature pronounced [eː] or those which are long by position as well? If not, were those pronounced [ɛ] or [e]? — Preceding unsigned comment added by 82.139.82.82 (talk) 02:56, 17 July 2015 (UTC)

What you're asking about is vowel length. Every vowel in a word was phonemically long or short. Long vowels were sometimes marked with apices in Classical Latin texts, but not always. In a text without apices, a, e, i, o, u could either represent a short or a long vowel. Latin vowel spelling is ambiguous in a similar but less extreme way than English vowel spelling is ambiguous. Whether vowels in a word were long or short is determined in various ways: from poetry, from what Roman grammarians say, from historical linguistics or etymology, and so on.
"Long by nature" and "long by position" are terms for syllables, not vowels. A syllable was either long or heavy because it had a long vowel in it ("long by nature") or because it ended in a consonant ("long by position"). Vowels themselves were just long or short, nothing more complex than that. — Eru·tuon 15:31, 19 July 2015 (UTC)
You answered a different question than I asked with a bunch of stuff I already knew. — Preceding unsigned comment added by 82.139.82.82 (talk) 11:57, July 26, 2015‎
Please clarify your question. Are you asking how we know that short e was pronounced [ɛ] and long e pronounced [eː]? — Eru·tuon 17:40, 26 July 2015 (UTC)
Yes, that's the first half of my question, particularly the distinction between the [ɛ] sound and the [e] sound in [eː] (as opposed to e.g. [ɛː]). The second half of my question is what happens when a vowel (long or short) is part of a syllable that is long by position. A common case is the semi-vowel i. Scenario 1, short a: maius [m?jjus] Scenario 2, long a: traiectus [tr?j¹ectus] What vowel should be in the place of the two question marks? At ¹, should there be a second j? Is the syllable auc in auctor overlong in normal speech (as opposed to poetry)? Was a naturally long vowel shortened (and possibly changed in quality) when it was already long by position (to keep the syllable from becoming overlong)? Did a short vowel change length (or indeed quality) when long by position? Scenario 1, short o: montis [m?ntɪs] Scenario 2, long u: nuntius [n?ntɪus] Again, what goes in place of the question marks? And how do we know? I hope this clarifies the question, I don't really know how much more explicit I could be. — Preceding unsigned comment added by 82.139.82.82 (talk) 00:19, 7 October 2015 (UTC)
For information on how we know there was such a thing as a long and short vowel, and how we know how they were pronounced, read Allen's Vox Latina if you can, the chapter on vowels. This article simply summarizes what he says. Perhaps it needs to go into more detail, since you still have questions.
I'm still confused about your second question. Let's get rid of the misleading terminology "syllables long by position". The modern term "closed syllables" is better. All I can say is to repeat: there was a distinction between short and long vowels in closed syllables. Long vowels existed in closed syllables. People disagree whether the first vowel of nuntius was short or long, but a better example is āc-tus and fac-tus. The first had a long ā, the second a short a. Perhaps the long ā in āctus was slightly shortened, but it was still longer than the short a in factus, despite the fact that both vowels were in closed syllables. Probably the syllable āc- was overlong in everyday pronunciation. To find out how we know these two words had different vowel lengths, read Allen's chapter called "hidden length". — Eru·tuon 01:15, 7 October 2015 (UTC)

Pronunciation of S

Pronunciation of s between vowels as /z/ doesn't match what other sites say, including: http://www.canticanova.com/latin_pron.htm, http://www.ewtn.com/expert/answers/ecclesiastical_latin.htm, http://cmed.faculty.ku.edu/rehnotes/latin.html

You can search further on the internet but I haven't found any that say it should be /z/ between vowels, as it does here.

142.254.27.248 (talk) 20:54, 29 July 2015 (UTC)

Maybe it's from English ecclesiastical Latin? http://jandrewowen.com/en/english-latin
But I agree there isn't enough support for the assertion; let's remove it. — Preceding unsigned comment added by 82.139.82.82 (talk) 20:26, 13 October 2015 (UTC)
Actually, the EWTN source says s is "slightly softened" between vowels. That probably refers to voicing, though it's not phonetically precise terminology. The two other sources, however, specifically say it isn't voiced. So two out of three support unvoiced [s] between vowels.
Italian Ecclesiastical Latin would likely have [z] between vowels, because Italian does. — Eru·tuon 22:19, 13 October 2015 (UTC)
It quite does. Also, the first link that was given actually says "S is hard as in the English word sea, but is slightly softened when coming between two vowels. e.g. misericórdia" - which is a not-very-phonetically-sound way of saying "it's /z/", I'd reckon. I've readded /z/ as an alternative option to /s/ into the article, with a relevant citation from a book. I've also got another source that's awkward to cite because it involves a double negation, but look: it talks about very old Italian ecclesiastical pronunciation, and states: "The Italians, too, though they are given qualified praise for distinguishing between the sounds of s and c before e or 1, and — contrary to their practice to-day — for not giving a z sound to intervocalic s, do not escape criticism." (Latin in church: the history of its pronuniation - Page 29, by Frederick Brittain, 1955, emphasis mine). That's saying that "to-day", the Italians use /z/ for intervocalic ⟨s⟩, and that before, they were given praise for not doing so, which meant other nationalities did so. LjL (talk) 22:25, 13 October 2015 (UTC)

Vowel nasalization: recordings

Have they been made by a Portuguese speaker? "ẽ" sounds more like [eĩ] than [eɨ̃] (or whatever can be used to transcribe the diphthong that a "European nasal vowel" usually is). 195.187.108.4 (talk) 16:23, 28 January 2016 (UTC)

Close and mid vowel qualities

In Classical Latin as described by Allen, short i and u had a similar quality to long ē and ō. As CodeCat mentions, this vowel system probably reflects the Latin of Rome or of the western portion of the Empire, not all the forms of Latin across the Empire, given the developments of long and short e, i, o, u in the various Romance languages. We could postulate that Latin across the Empire had long and short vowel qualities similar to the ones reflected in Romance vowel developments, meaning that in some Latin dialects the long versions of e, i, o, u were similar to the short ones, and in some dialects they had height differences like the ones described by Allen for Classical Latin, but I'm not sure if this is reflected in sources or if it's just WP:OR. — Eru·tuon 22:50, 29 March 2015 (UTC)

The point of the table is to show phonemic distinctions. The height distinctions between short and long vowels were not phonemic, so they don't belong in the table. They only became phonemic once the length distinctions began to disappear, but this is post-classical Latin and by this point you can no longer speak of one "Latin" and therefore neither is there one single Latin phonology. CodeCat (talk) 23:43, 29 March 2015 (UTC)
I'm not sure what sources say about phonemicity, but it's well established that Latin vowels were distinguished by both length and height, whether phonetically or phonemically, even in the Classical period. The table originally displayed only vowel length, but I added height differences between long and short vowels because Allen displays them in his diagram. It's misleading not to present height differences, since it suggests that /ɛ eː/ and /ɪ iː/ were closer phonetically than /ɪ eː/, which is not the case. However, I should have posted here before changing the table. You can change it back if you want, until we reach an agreement.
I'm just speculating here, but merger is good evidence for height as a distinctive feature, at least before the merger. First two vowels are distinguished by both height and length, then they merge. The fact that /ɪ iː/ and /ɛ eː/ merged in Sardinian, but /ɪ eː/ merged in Italian, suggests that Sardinian and Italian had different phonological features for these vowels immediately before the merger, differences in the features of vowel quality. Perhaps the same was true of Classical Latin, the great-aunt of Italian Vulgar Latin, and arguably this is suggested by confusion between short i and long e: this suggests that phonologically long e was actually the long version of short i, rather than the long version of short e. But this is just speculation. Do you know of any sources that discuss whether height was a phonemic feature in Classical Latin? — Eru·tuon 02:40, 30 March 2015 (UTC)
In User talk:Benwing#Sicilian vowel system, Benwing remarks: "Early Latin apparently pronounced its short and long vowels with the same quality (this is noted in Ringe's newest book). Only later did quality differences appear, and quite likely simply never spread to Sardinia and North Africa." and "In the Harris/Vincent "Romance languages" it's claimed that Sardinian may have split off as early as the 1st century BC." However, this seems to mean that in the 1st century BC, which is the Classical (not Early) Latin period, and when Sardinian had already split (or was beginning to split) off, the quality differences could already have appeared. In this case, the spoken forms underlying Classical Latin in (Central) Italy could be thought of as already on the branch that led to Western Romance and Italo-Romance (Italo-Western Romance), and Sardinian is irrelevant, because it reflects the pre-Classical, Early Latin lack of quality distinctions. --Florian Blaschke (talk) 11:56, 12 July 2015 (UTC)
What you seem to be implying is that Sardinians didn't speak Latin as this page defines it. Which seems kind of nonsensical; even if they didn't participate in the earliest sound changes of the Romance languages, that doesn't mean they weren't still part of the Latin dialect continuum. Hence, if this page is to cover "Latin" and not "non-Sardinian Latin", we have to use a vowel notation that covers all the dialects. Using /e/, /i/ etc seems like the most obvious solution. CodeCat (talk) 13:49, 12 July 2015 (UTC)
Thinking about it more, there also seems to be a non sequitur going on here. The differentiation of the quality of short and long vowels does not necessarily have to lead inevitably to the merger of, say, short [ɪ] and long [eː]. So the fact that they didn't merge in Sardinian is no evidence that there was no quality distinction. They could have simply continued to exist as they were, and at some later date the short vowels merged into the long vowels. CodeCat (talk) 13:52, 12 July 2015 (UTC)
The Sardinians certainly did not speak Classical Latin, which is an artificially standardised language and does not reflect any spoken variety. Just as an example: There is no reason to think that /h/ was ever pronounced in popular Latin after the 4th century BC (Meiser points out that anser, which contains an etymological h, is not even spelt with h, even though Classical Latin orthography preserves various outdated traits, and usually retains the letter, thus indicating that h was lost not long after the establishment of an orthographic tradition, and diribeō indicates that the loss of h preceded the rhotacism, which allows us to date the change fairly precisely). As Benwing points out on his talk page, Sardinian (like other Romance languages) contains a couple of Pre-Classical archaisms too. So yes, the Sardinians most definitely did not ever speak the language as described in this article at any point as a vernacular language. It appears that Sardinian branched off spoken Old Latin (whose written form is perhaps less divergent from the spoken varieties) in fact – making Old Latin essentially Proto-Romance; at least I interpret "Early Latin" in Benwing's comment to really refer to what is usually called Old Latin. (However, the period preceding Odusia – ca. 240 BC – is frequently designated as Archaic Latin, or the preliterary period, so "Early Latin" may be used as an umbrella term for pre-Classical Latin here.) --Florian Blaschke (talk) 15:43, 12 July 2015 (UTC)
Interestingly, I have seen the suggestion that the rise of the vowel quality distinctions and the mergers found in Mainland (Italo-Western and Balkan) Romance could be due to the Sabellic substrate. --Florian Blaschke (talk) 15:51, 12 July 2015 (UTC)
Theoretically, as far I can see, it is very much possible to reconstruct Proto-Romance with a nine-vowel system /i ɪ e ɛ a ɔ o ʊ u/ without any quantity distinctions at all (/i/ < ī and ȳ, /ɪ/ < ĭ and y̆, /e/ < ē and oe, /ɛ/ < ĕ and ae, /a/ < ă and ā, /ɔ/ < ŏ, /o/ < ō, /ʊ/ < ŭ, /u/ < ū), and simply assume that Sardinian merged /i/ with /ɪ/, /e/ with /ɛ/ etc., but apart from the fact that that system looks at least a bit unusual typologically, the known facts about the history of Latin seem to make a system with quantity and no (or at best sub-phonemic) quality distinctions preferrable. But yeah, if we knew nothing about Latin, but only the medieval/modern Romance languages, this would be a viable solution. Anyway, it's not me who says that Early Latin didn't have any (sub-phonemic) quality distinctions (or that Sardinian proves anything either way); Ringe does. I was as surprised by it as you are, because I thought the quality distinctions were fairly natural (although my native German bias may colour my assumptions of naturalness here), and traces pointing to them attested even in Early Latin inscriptions (I thought I had read something about ĭ being occasionally found spelt as E in Wachter). --Florian Blaschke (talk) 16:04, 12 July 2015 (UTC)
Also, I'd like to remind people that Romanian (Balkan Romance) treats the back vowels the same way as Sardinian (but the front vowels like Italo-Western – this is the famous asymmetrical vowel merger of Romanian). But again, I'm not saying that this is in any way probative for the issue at hand, and I have no idea if Ringe uses that argument or what arguments he uses at all. I'm not even sure what book Benwing was referring to. --Florian Blaschke (talk) 16:27, 12 July 2015 (UTC)
Generally, however, I agree that (like in German) the most obvious and simple solution is to simply write /i/, /e/, etc., so Sardinian and North African Latin, as well as perhaps Balkan Latin and who knows what else (Corsican, Sicilian? Southern Italian? Maltese Latin? British Latin?) would in any case be covered. --Florian Blaschke (talk) 16:41, 12 July 2015 (UTC)
Whew, a lot of replies to wade through. I agree with both of you, CodeCat and Florian Blaschke, that using /a aː e eː i iː o oː u uː/ would be the simplest and would allow us to describe all Latin dialects, but I was working from W. Sidney Allen, and he only describes Classical Latin and the contemporaneous Vulgar Latin of Italy, not the Latin of other areas or times in the Empire. And he gives clear evidence showing that short i and long e, short u and long o, had basically the same qualities — evidence not from reflexes in Romance languages, but from Latin inscriptions and Roman grammarians in the Classical period — and evidence that short i and long i (and so on) did not have the same qualities. This evidence suggests to me that Classical Latin actually had seven vowel qualities, not five. These qualities could be written /a aː ɛ eː ɪ iː ɔ oː ʊ uː/ or even /a aː ɛ eː e iː ɔ oː o uː/. I like the latter system because it uses the same symbol for short i and long e, making it absolutely clear that the two have similar or identical quality. However, Allen was more of a classicist than a linguist and didn't attempt to create a phonemic notation.
I think the desire to phonemically transcribe Sardinian as well as Italian Latin is a little too ambitious. Representing both systems makes our phonemic system actually diaphonemic, like the Wikipedia IPA system for English, not truly phonemic. It's likely that Sardinian and Italian Latin had different vowel systems, with different numbers of vowel qualities, and therefore they need to be transcribed with separate systems. However, I don't know if any sources discuss this question based on the type of evidence that Allen uses (inscriptional evidence and evidence from grammarians).
Florian Blaschke, it would certainly seem to a German speaker (or even an English speaker) that it's natural for vowel length differences to be accompanied by vowel quality differences, but cross-linguistically it's by no means natural. For instance, Japanese has long and short vowels with the same quality. Probably the German and Latin pattern is more natural to a stress-timed language, while the Japanese pattern is more natural to a mora-timed language, but I'm just guessing. — Eru·tuon 17:38, 12 July 2015 (UTC)
I think you're mistaken to suggest that Classical Latin had any kind of pronunciation at all. I'm pretty sure that whenever someone read it out loud, they did so in their own dialect. CodeCat (talk) 17:57, 12 July 2015 (UTC)
Good point; excellent point, in fact. This means it would be preferrable if we made our representation of the pronunciation as abstract and high-level as possible, wouldn't it? Certainly, in the time we're talking about (1st century BC and AD), dialect differences foreshadowing the later major Romance subgroups (and some others) were at least incipient. (There would also be non-native pronunciations, of course, such as Gaulish-tinged Latin, but clearly this is not something we can reconstruct with any precision.) So, either we use a pandialectal representation, or we limit ourselves to a particular regional pronunciation of the written language (which at that point I think is actually comparable to a modern standard language in many respects), with the pronunciation of (Central) Italy being the most obvious solution. But showing /i/ and /eː/, /u/ and /oː/ as having the same quality already (which was certainly not the case in Sardinia, Africa and the Balkans) seems a bit extreme to me; [ɪ] is similar enough to [e] and [ʊ] similar enough to [o] that confusion is understandable anyway (although influence from other languages such as Sabellic may have exacerbated this). Similarly, in popular speech, /ae̯/ and /oe̯/ may already have become monophthongs by this time, but this was probably perceived as non-standard just like monophthongisation of /au/. --Florian Blaschke (talk) 18:19, 12 July 2015 (UTC)
CodeCat, you seem to be implying that Classical Latin is only a literary language and wasn't actually spoken natively by anyone. That's kind of true, but not completely. The writers of Latin works had a particular pronunciation, and some of their works were even made to be spoken, like Cicero's orations and poetic works. So, at the very least, Cicero's pronunciation and the pronunciation used by the writers of Classical Latin poetry count as Classical Latin pronunciation. More broadly, Classical Latin pronunciation could be defined as whatever pronunciation was used for the standard or high-class spoken Latin of the relevant period, the spoken Latin that followed the prescriptive conventions of written Latin. Whether readers from different places, social strata, and times happened to use another pronunciation is irrelevant. — Eru·tuon 19:11, 13 July 2015 (UTC)
But do we have any evidence that the lowering of /i/ and /u/ to almost the same or the same level as /eː/ and /oː/ was actually considered good standard and used by educated speakers of Latin, or is it equally or more likely that it was it a non-standard pronunciation that crept into inscriptions sometimes? For example, Pompeian inscriptions usually reflect popular speech habits and not educated language usage. Who made these inscriptions that show spellings such as sob for sub – were they highly educated, middle/upper-class citizens like Cicero? What do the grammarians say on this point?
For example, let's imagine that [sub] was a conservative, archaic or even obsolete pronunciation by the late 1st century BC, that [sʊb] was the pronunciation of the likes of Cicero and that [sob] was a more popular or rustic pronunciation (along with dropping of /h/, perhaps denasalisation, and monophthongisation of /ae̯/, /oe̯/ and perhaps /au/ – the diphthong /au/ is usually kept separate from /oː/ in Romance, however, and only in some individual words such as cauda and auricula merges with it). Then Cicero's would be standard, right? But can we be sure about what Cicero's pronunciation of /u/ and /i/ was like?
Modern Standard Dutch may be a close parallel for Classical Latin – the subset of traditional dialects on which the written standard is based are not very far from the standard language, but the standard does contain some striking archaisms such as the conservative pronunciation of ij, ou and ui, as rather narrow diphthongs [ɛɪ], [ʌʊ] and [œʏ], where many traditional dialects have more open diphthongs [aɪ], [aʊ] and [aʏ], which is now creeping into the standard, but was long dismissed as substandard. A Dutch person of a level of education comparable to Cicero would avoid the popular open diphthongs, and probably the devoicing of initial fricatives /v z ɣ/ too, even though it's pretty widespread now in the younger generation especially, as far as I'm aware. But I doubt that we can reconstruct Classical-era Latin sociolinguistic differences with this level of precision. --Florian Blaschke (talk) 21:34, 13 July 2015 (UTC)

I don't have Vox Graeca at the moment, but I believe Allen quoted some grammarians who wrote about the difference between long and short i and how short i was similar to long e, and the joke relating to cum nobis and conubiis was made by Cicero, if if I recall right (maybe the page should mention this). So I think this particular characteristic of the vowel system was apparently shared between the different layers of Roman society.

As to sociolinguistic differences, it's possible, given the difference in popular spelling of final vowels plus m, or medial vowels plus ns, nf, that popular pronunciation dropped nasalization completely, while the upper classes kept it, and that the upper classes continued pronouncing ae and oe as diphthongs rather than monophthongs for a longer time than the lower classes. It's likely, too, that long vowels were shortened more often among the lower classes, making long e more easily confused with short e, long o with short u.

Perhaps the fact that the upper classes knew which letter to write would suggest that they pronounced these pairs of vowels with more distinct vowel qualities, but it could also be due to better education, such that they, like better educated English speakers today, could spell better than the lower classes: they knew when a sound was supposed to be written with e and when with i, even though the two sounded similar. — Eru·tuon 16:34, 11 February 2016 (UTC)

Vowels

Okay why on earth are the Classical Latin vowel phonemes being described in both terms of quality and quantity? This page isn't just about Vulgur Latin and its weird prescriptivist assertions about this run against the general evidence that the distinction was quantity and that the shift to quality was part of the change to VL and should not be included as Canon Fact here. This is decidedly not what linguists discuss in regards to Classical Latin. Ogress 18:16, 16 March 2016 (UTC)

The differences in vowel quality the article describes are referenced by Allen, which is definitely considered an WP:RS, and corroborated by errors in Classical inscriptions. What citations are you bringing forth to substantiate your vehement point of view? I'm sure both views can be described in the article, if they're equally well-sourced. LjL (talk) 18:42, 16 March 2016 (UTC)

Nasalised vowels

Fellow editors, including CodeCat: respectfully, there is an entire section of this page about this feature. Ogress 17:54, 29 April 2016 (UTC)

Ah, I see I (and everyone else) was confused by the history of the edit. CodeCat is correct because the forms were reversed in a kind of weird transposition. Let me know if it looks right now? Ogress 18:03, 29 April 2016 (UTC)

cum nobis

"CVM NÓBꟾS [kʊn ˈnoː.biːs] was a double entendre for CÓNV́BIꟾS [koːˈnuː.bi.iːs]."

Are you sure it wasn't 'cunno bis'? --Excelsius (talk) 08:46, 23 March 2016 (UTC)

@Excelsius: No, I'm not sure. You can go ahead and insert that reading if you've got reason to suspect it's correct. Actually, it makes a little more sense. — Eru·tuon 21:00, 29 April 2016 (UTC)
Short u was confused with long o in quality, and short i with long e. But I don't think two long vowels or two short ones would have been confused with each other. After all, short vowels didn't undergo any mergers with each other in Romance, nor did any long ones. CodeCat (talk) 21:41, 29 April 2016 (UTC)

Lāvīnia can never fit the meter

I took the bold step to use Lavina instead, as Lavinia can never fit the meter - the only alternative would be la.ˈwi.niː.a.kʷɛ, and that seems like too many changes to me. Please go ahead and correct it if you think you know of a better option.2A00:1028:83D6:8E56:3011:5CF4:F39F:F644 (talk) 20:40, 29 September 2018 (UTC)

"Italianate" ecclesiastical pronunciation of vowels

I agree that sung Ecclesiastical Latin has /-tsi-/ rather than /-ttsj-/ (as in standard Italian) for -tiV-. I still have some doubts as far as vowels are concerned: I was taught e (not ae or oe) was always /ɛ/ and o was always /ɔ/. Nevertheless, I couldn't find a reliable source for this statement. Could you help me?.--Carnby (talk) 14:36, 15 September 2017 (UTC)

You might look at the Liber Usualis starting on page xxxv in the Introduction (http://www.ccwatershed.org/blog/2013/mar/19/1961-solesmes-liber-usualis-online-free-pdf/), which lists the correct pronunciation for Ecclesiastical Latin for purposes of singing and chant. I will note that I have to differ with the description of the pronunciation of h in the Liber; The experience that I and other Catholics in the US have had both before and after Vatican II was that the h was pronounced just as in English. My assumption is that the Romance countries drop the h because their native languages did (cf. Italian, Spanish, etc.), but that in English-speaking countries, the h was spoken because it is spoken in most versions of English.
William J. 'Bill' McCalpin (talk) 02:07, 7 December 2018 (UTC)

Change to Ecclesiastical Latin pronunciation

I have updated the pronunciation of the letter h. I first began learning Latin in the 1950s as a (young) Roman Catholic in Texas, then became an altar boy who memorized the Mass in Latin prior to Vatican II, then took Latin in a Dominican high school, then took Latin in a Catholic university (the University of Dallas), then lived in Rome. I checked with an American canon lawyer who has the same recollection. I will add more independent documentation as I find it.

William J. 'Bill' McCalpin (talk) 02:21, 7 December 2018 (UTC)

title

this article should be titled "Latin phonology" in line with every other article on different langauges' phonologies — Preceding unsigned comment added by Literally Satan (talk • contribs) 13:27, 27 January 2019 (UTC)

Pronunciation of 'suauis'

The Latin adjective SVAVIS originates Italian soave which is trisyllabic (so.a.ve). My guess is, it was prononunced /su'a:wis/ (not /'swa:wis/), that is, it is a counterexample. Please provide a cite or change the example.--79.44.191.36 (talk) 21:42, 10 September 2009 (UTC)

I had my doubts about that claim, too (being Italian it did seem weird). However, 1 and 2 --LjL (talk) 21:46, 10 September 2009 (UTC)
That guess is wrong, "suauis" was disyllabic, with both "u" being semivowels. Such semivowels were sometimes vocalized (e.g. disyllabic "sol-uit" shows trisyllabic variant "so-lu-it"), but I can't find any single instance of that for "suauis". In any case, ancient Latin poetry constitutes stronger evidence than modern Italian pronunciation. Mamurra (talk) 14:56, 19 October 2009 (UTC)
I don't think that Italian "soave" is trisyllabic: the "oa" is pronounced as a diphthong on one syllable. At least, that's how I pronounce it. It is true that the "o" has more of a vocalic character than an English "w", and it is perfectly possible that the first "v" in "svavis" also had a more vocalic character than our "w", but still falling well short of forming a separate syllable. Alternatively the phoneme has become more vocalic than it was. Either way, the evidence that "svavis" had only two syllables is overwhelming. JamesBWatson (talk) 10:49, 27 October 2009 (UTC)
Latin suavis is trisyllabic according to Lewis & Short – see suavis in Charlton T. Lewis and Charles Short (1879) A Latin Dictionary, Oxford: Clarendon Press – while it can be both according to Wiktionary. Italian soave is definitely trisyllabic, although, as every other Italian word, it normally undergoes synaeresis in poetry and becomes bisyllabic. 82.6.195.35 (talk) 02:49, 26 November 2019 (UTC)

Requested move 7 April 2020

The following is a closed discussion of a requested move. Please do not modify it. Subsequent comments should be made in a new section on the talk page. Editors desiring to contest the closing decision should consider a move review after discussing it on the closer's talk page. No further edits should be made to this discussion.

The result of the move request was: page moved. Andrewa (talk) 18:35, 15 April 2020 (UTC)

Latin spelling and pronunciation → Latin phonology and orthography – The article title is inconsistent with other article titles listed on {{Language_phonologies}} and {{Language_orthographies}}. Therefore, rename this article as "Latin phonology and orthography". Otherwise, split the article between "Latin phonology" and "Latin orthography". Soumyabrata stay at home wash your hands to protect from coronavirus 13:12, 7 April 2020 (UTC)

  • Support, on the grounds of consistency. Neutral on splitting. Length would justify it, but considerable effort would be necessary to separate the phonology and orthography, which are inherently linked concepts and the discussions of which are very intertwined in the article. Seoltoir22 (talk) 05:01, 12 April 2020 (UTC)
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
I am currently preferring splitting rather than renaming. Let me compose the splitted version in my userspace. --Soumyabrata stay at home wash your hands to protect from coronavirus 17:46, 19 April 2020 (UTC)

(I hope I´m doing this correctly). I have no linguistic knowledge but know Latin and Greek and I have often wondered if Latin "u" was pronounced similarly to Greek "o". E.g. Greek "Markos" and Latin "Marcus". After reading the phonology section it seems to me that the sounds are similar, is this correct? I suggest this question is adressed clearly by an author with linguistic knowledge. The Example of the Latin "U" sound is "put", but is this as in "to put a hat on" or "to put" in golf? Foto (talk) 09:09, 30 July 2020 (UTC)

Suffixal particles and stress

How do suffixal particles, such as -ve, -que and -ne interact with stress? Do they count as an extra syllable in the word, or are they separate? Rua (mew) 12:38, 2 November 2020 (UTC)

See [1]. Burzuchius (talk) 14:27, 2 November 2020 (UTC)

Extra V and VC make syllable heavy; what about extra C?

The current article contains the text:

"A syllable is heavy if it has another V or a VC after the first V. In the table below, the extra V or VC is bolded, indicating that it makes the syllable heavy."

In the table below it, also syllables extended with a C are indicated in bold, so I understand those syllables are heavy too. If someone more knowledgeable can confirm this, we can change the first quoted sentence accordingly to:

"A syllable is heavy if it has another V, another C or VC after the first V."Redav (talk) 00:21, 5 November 2021 (UTC)

Realization of vowels

Two unregistered users (23.241.195.116 and 77.246.196.74) have recently been attempting to change the realizations of short i and u to [i u] and long e and o to [ɛː ɔː] throughout the article. They do not cite a source for this, but that is Andrea Calabrese's reconstruction of the Classical Latin vowel system. As far as I know this is a 'fringe' view, and the majority of scholarship today rather upholds Sydney Allen's reconstruction; see e.g. Leppänen & Alho 2018 or Cser 2020.

Calabrese's reconstruction can be mentioned in the article as a minority view per WP:RSUW.

The Nicodene (talk) 18:47, 28 July 2021 (UTC)

Now a third unregistered user (198.52.158.114) has attempted to do the same thing. Is this page being targeted by fans of Scorpio Martianus or something? The Nicodene (talk) 05:41, 19 August 2021 (UTC)

It should be noted that Calabrese's reconstruction is, probably, a minority view because it looks at the subject through the lens of modern Romance languages and rustic/non-urban Latin, whereas most other research on the subject seems to lean on Germanic influence. As the paper points out, Ancient Greek <η> [ɛː] and <ω> [ɔː] for <ē, ō>, respectively. Additionally, the paper leaves this article's vowel system in place for Late Latin. Thus, I think it could be warranted to provide both systems until more research has been done.Michealin (talk) 22:24, 24 August 2021 (UTC)

I wonder from which time the inscriptions are that confuse long e and o with short i and u; Allen does not say. Are they from the classical period?
In any case, while the lax pronunciation of short vowels is believable for mainstream Latin, judging from the development to Romance, at least some accents seem to have been exceptions: in Sardinia, Corsica, parts of southern Italy, perhaps Sicily and Malta, and Northwestern Africa; presumably also the Balkans. This may have been due to Greek or Punic influence. --Florian Blaschke (talk) 21:56, 28 November 2021 (UTC)
According to Calabrese, in the first century BC, the innovative system (as well as the monophthongisation of /ae/ and /au/ to /ɛː/ and /ɔː/) was characteristic of rustic Latin (could Sabellian influence have been a factor?), while the conservative pentavocalic system /i ɛ a ɔ u/ plus quantity was characteristic of urban Latin, as well as of the southern varieties that gave rise to Sardinian, African Romance, and other Romance varieties mentioned. However, in the first century AD, the innovative rustic system began to spread into urban Latin, and became the norm in spoken Latin afterwards – except in the southern regions mentioned. Presumably, then, Cicero and Caesar still used the conservative system – but already in the first century AD, still in the classical period (!), the new system began to rival and supplant it. Sardinian etc. then preserve the old urban system. This also implies that the variety ancestral to Sardinian had become distinct from the mainstream of spoken Latin by the first or second century AD at the latest (and also that "Proto-Romance" does not actually postdate Classical Latin; instead, it may have been identical to the Latin spoken in Rome in the classical period if not even slightly earlier). --Florian Blaschke (talk) 00:37, 3 December 2021 (UTC)

The alleged rarity of the [ae] diphthong

Per WP:BURDEN, the burden of sourcing is on those who want to keep an unsourced statement. For this reason alone I'm going to remove it.

Gimson (2014) says that English /aɪ/ rarely reaches above the close-mid height, which means that [ae] is an appropriate narrow transcription. Das Aussprachewörterbuch (2015) says that [ae] is also an appropriate narrow transcription of German /aɪ/. We need to remember the distinction between phonemes and allophones - see here.

Please do not bring up the alleged tenseness of the non-syllabic element of [ae] (in German or other languages) in this discussion. It is pure nonsense and contradicts all phonetically explicit descriptions of German vowels. Non-syllabic elements of diphthongs are shorter and more weakly articulated than the other part of the diphthong, which by definition makes them phonetically lax (in fact much 'laxer' than /ɛ ɔ/!) and renders the distinction irrelevant (German and English /aɪ/ are phonological units on their own that behave like tense monophthongs in that they can occur in a stressed word-final position. But that's their phonological behavior. Here, we're talking about the alleged phonetic tenseness of the non-syllabic [e]). If [e] really were tense, there'd have to be a hiatus between the it and the preceding vowel: [a.eː]. This combination, if at all possible in any type of SG, doesn't occur in Northern Standard German but only western and southern dialects. Perhaps the word verehelichen [faˈeːəlɪçn̩] (NSG: [faˈʔeːəlɪçn̩]) can be said to contain it in those dialects, since ⟨ɐ⟩ denotes the exact same sound as ⟨a⟩ in Northern Standard German (see [2]), a statement that very likely applies to Western SG and perhaps some types of Southern SG as well. In fact, there are two hiatuses in succession there: [a.eː.ə]. Or is it perhaps [ae̯.eː.ə] for those speakers, with a quick [ae̯] diphthong in the first syllable? I don't know an answer to that to be honest.

[e] and [ɪ] are hardly contrastive in many if not most languages anyway - see e.g. [3]. and [4]. ⟨ɪ⟩ can be safely replaced with ⟨e⟩ in many types of transcription and the result would still be phonetically explicit (IPA symbols of vowels are not defined for tenseness, length or a precise amount and type of rounding and ⟨ə ɐ⟩ aren't defined for rounding at all), something that can't be said about ⟨i⟩ which often suggests a quality very similar to that of a syllable-initial palatal approximant.

If you transcribed the German word bitte with the symbols ⟨ˈbetə⟩, the resulting "literal" pronunciation (with every symbol read according to its canonical IPA value) would sound just like the native Northern German pronunciation (a more standard pronunciation would be written ⟨ˈpetʰə⟩, denoting voicelessness and aspiration or lack thereof), nowadays universally transcribed [ˈbɪtə] in German dialectology using a largely agreed-upon and well-defined set of IPA symbols. ⟨ˈbitə⟩, on the other hand, would sound like biete (especially when we're talking about Northern German, where tense vowels aren't really that long) unless you clarified that the first vowel is much more like a centralized cardinal [e]. With ⟨e⟩, you need no such clarification. English /aɪ/ varies freely between [ae] and [ai], or [ɑe] and [ɑi] for many speakers in England, with the variants ending in [i] being most common before a vowel. The same kind of variation is found in New Zealand English.

Wiese (1996) says that German /aɪ aʊ ɔʏ/ are best analyzed as /a/ + /ɪ/, /a/ + /ʊ/ and /ɔ/ + /ʏ/. But that still doesn't mean that the second element of /aɪ/ is not phonetically [e] (a non-syllabic [=a very brief and lax] close-mid front unrounded vowel), because it is. /aʊ/ is [ao] too, ending in a non-syllabic [=a very brief and lax] close-mid back rounded vowel, per DA 2015. Sol505000 (talk) 13:33, 4 July 2021 (UTC)

I agree. Certainly in my own speech and that of many other native speakers /aɪ/ is phonetically [ae̯], and /aʊ/ is [ao̯], simplifying the articulatory gesture somewhat, see Standard German phonology § Diphthongs.
Whether lax /ɪ/ is generally often realised as [e], and /ʊ/ as [o], especially in Northern Germany, I cannot say, but it definitely sounds plausible. This would only strengthen the conclusion that, apparently already in the classical period, the lax short vowels /ɪ/ and /ʊ/ had become more or less identical in height with the tense long vowels /eː/ and /oː/, with the (later) quantity collapse only consisting in the loss of the quantity distinction as a phonological contrast (and it becoming automatically conditioned by the shape of the syllable). However, this conclusion would seem to imply that the variety of Latin ancestral to Sardinian, where short /i/ and /u/ remained distinct in vowel height from long /eː/ and /oː/, had already become distinct at the point when these vowels had become identical in colour (though not merged yet) in mainstream Latin pronunciation – presumably, then, by the classical period. --Florian Blaschke (talk) 00:05, 3 December 2021 (UTC)
Yet the Latin of Dacia, colonized from 117 CE onwards, apparently had and retained a quality distinction between original /ŭ/ and /ō/. Moreover, graphical confusion of these phonemes is rare in the first three centuries CE. A better case could be made for /ĭ/ and /ē/, though even that is uncertain: Sardinia was only one of three regions that never lost the quality distinction, and the further back in time one goes, the more widespread the distinction must have been. Nicodene (talk) 09:22, 6 December 2021 (UTC)
It's possible that Dacia (or wherever the dialect that gave rise to Romanian was initially formed, which wasn't necessarily Dacia, another complication here) was initially substantially colonised from regions that preserved the old system. The asymmetrical Romanian vowel system – also found in one region in southern Italy! – may have arisen through contact between speakers of the more conservative pentatonic and the more innovative lax-tense system.
Already in the first century AD there is definite evidence for the innovative lax-tense system, see Calabrese's paper linked below. --Florian Blaschke (talk) 01:54, 7 December 2021 (UTC)
That there was a trend towards laxing /i u/ in the early Imperial period is not in question: what is doubtful, in my view, is that there were already stark regional divisions in the pronunciation of both phonemes, corresponding to future Romance isoglosses. That Sardinia, for instance, eventually merged /u/ with /ū/ does not prove that no laxing had occurred at all there; conversely, that Iberia eventually merged /u/ with /ō/ does not prove that, as early as the first century CE, the two were virtually identical by height. In speaking of laxing, we are not dealing with a binary choice between [u] and [o]- or even among [u], [ʊ], and [o]; rather we are dealing with a gradient of possible values. In all likelihood, trends leading in either direction fluctuated both diatopically and diachronically, even reversing at times, until the collapse of length contrasts made a merger with one of the neighbouring vowels difficult to avoid. What matters, ultimately, is what values /i/ and /u/ had in each region when vowel length collapsed, and in the period that followed, rather than what values they had in the first century CE. Nicodene (talk) 05:58, 7 December 2021 (UTC)
Considering that /ĭ/ and /ē/ as well as /ŭ/ and /ō/ respectively, per Allen and others, were already confused in the classical period, their colours must already have become very similar or even virtually identical that early, even though they were presumably still distinguished by quantity. This clearly did not happen in Sardinian, which shows no trace of tensing in /ē/ and /ō/ either – they are reflected as /ɛ/ and /ɔ/, the values reconstructible for Old Latin (sans quantity). Positing a lax-tense system for an early stage of Sardinian despite zero evidence for such a stage is unadvisable in view of Occam's razor. Therefore, the evidence, as laid out by Calabrese, strongly points towards an early split between the conservative pentatonic system preserved in Sardinian and the more innovative (initially "rustic", but as early as the first century AD also urban Roman) lax-tense system preserved in most of Romance. --Florian Blaschke (talk) 01:27, 9 December 2021 (UTC)
Cross-linguistically, it is the norm for vowel length distinctions to be accompanied by differences in quality, whether small or large. For short close vowels, this especially manifests as centralization relative to long counterparts; cf. Welsh, Czech, Fula, Egyptian Arabic, Kurdish, Urdu, Cantonese, Tlingit, to name only a few examples taken at random from Canepari (2007: §§17–21). Hence it seems unlikely that, from the Republican period to the time of the collapse of Latin phonemic vowel length (approximately the fifth century CE), there never developed a shade of difference between any short vowel and its long counterpart in Sardinia, Africa, or southern Lucania. That each pair eventually merged in these areas does not imply that, prior to the merger, they must have been identical, except in length.
Calabrese presents a false trichotomy; the only values he considers for /u/, for instance, are [u], [ʊ], [o]. That intermediate shades may be relevant, and that a degree of laxing (not necessarily as dramatic as, say, [u] > [ʊ]) can have occurred even in the aforementioned regions, seems not to occur to him as a possibility. He invokes southern Lucanian as a Romance dialect supposedly reflecting no vowel quality differentiation, but it has varieties where, in stressed position, reflexes of Latin /e o/ show metaphonic diphthongization, while reflexes of /eː oː/ do not (Loporcaro 2011: 113). I suppose one could invoke influence from 'mainstream' Neapolitan to explain this, except that one could equally invoke influence from Calabrian—with its Sicilian-type vocalism—to explain non-diphthongal outcomes in other varieties of southern Lucanian. For Calabrese's argument, this is, at best, a precarious peg to rest on.
——————————————————————————
The most modern, and comprehensive, treatments of evidence for Latin vowel changes that I am aware of are found in the works of the late James Noel Adams and Leppänen & Alho. The former, after extensively discussing the evidence, states by way of conclusion (2013: 67):
‘Signs of opening of ĭ are already to be found in the republican period. In the early Empire the tendency is most marked in final syllables. From about the third century AD evidence accumulates for shortening of long unstressed vowels (and for a complementary lengthening of stressed vowels, and not only in open syllables), not least in final syllables. The two developments that were a sine qua non for the emergence of the vowel system of most of the Romance languages, namely (i) opening of ĭ towards the position of articulation of ē and a corresponding opening of ŭ towards ō, and (ii) the loss of phonemic distinctions of vowel quantity, may have started in final syllables, or have been more advanced there early on. By the later period (notably of about the fifth century) we find numerous misspellings suggestive of (i) in other parts of the word, and evidence in grammarians and others for the loss of distinctions of quantity, and not only in final syllables. The evidence to do with the vowel system is consistent with gradual change, starting from opening of the front high vowel ĭ in the Republic and moving on to the undermining of the system of quantity and the opening of ŭ much later.’
In other words, evidence for the change /ĭ/ > ~[e] in all environments is found after about the third century (and not in all regions). Much later, we begin to see evidence of /ŭ/ > ~[o] in all environments; the considerable lag between the two changes is notably consistent with the Balkan Romance and Castelmezzano vocalisms. What the data are not consistent with is the claim that, in Classical times, there existed a precocious Italo-Western-like vowel system responsible for the outcomes of Standard Italian, Spanish, etc. As stated by Adams (loc. cit.): ‘The view advanced by Pulgram (1975) that there had long been a submerged vowel system of Romance type alongside that of Classical Latin receives no support from the evidence.’
The conclusions of Leppänen & Alho (2018: 21–22) are quite compatible:
‘As a result of three pre-Classical Latin sound changes (vowel weakening, monophthongisation, and long vowel chain shift), long vowels acquired peripheral (i.e. more close, tense) realisation and short vowels acquired nonperipheral (i.e. more open, lax) realisation. The strength of the peripherality-induced qualitative differentiation varied from dialect to dialect […] After CVL-loss, the first group of dialects with strong peripherality distinctions on both front and back axes underwent both mergers, the second group with prominent distinctions on front axis only underwent the merger of ĭ and e but not of ŭ and o, and the third group with weak distinctions on both axes underwent neither of the two mergers.'
Here CVL stands for contrastive vowel length.
There is, incidentally, a relevant detail that neither source mentions, namely that Sardinian does in fact show, as Pompeiian Latin also does, evidence for /ĭ/ lowering to mid-height specifically in verb endings. In the present indicative, the 2PL forms end in /-es/, not the */-is/ that one might have expected from Latin /-ĭs/; cf. Nuorese domáes, finíes < Latin domātĭs fīnītĭs (Pittau 1972: 107, 111). The same is also true of 2SG and 3SG endings in many third-conjugation verbs, such as facĭs, quaerĭt > Nuorese fakes, keret (op. cit.: 109, 115), although the Sardinian merger of Latin’s third conjugation with the second, which had /ē/ and /ĕ/ in the relevant endings, may be responsible for this. Cross-conjugational influence may have played a part in the Pompeiian phenomenon as well.
  • Adams, James Noel. 2013. Social variation and the Latin language. Cambridge: Cambridge University Press.
  • Canepari, Luciano. 2007. Natural phonetics and tonetics: Articulatory, auditory, and functional. München: LINCOM.
  • Leppänen, V., & Alho, T. 2018. On the mergers of Latin close-mid vowels. Transactions of the Philological Society 116. 460–483.
  • Loporcaro, Michele. 2011. Phonological processes. In Maiden, Maiden & Smith, John Charles & Ledgeway, Adam (eds.), The Cambridge history of the Romance languages, vol. 1, 109–154. Cambridge University Press.
  • Pittau, Massimo. 1972. Grammatica del sardo-nuorese: Il più conservativo dei parlari neolatini. Bologna: Pàtron.
Nicodene (talk) 05:00, 10 December 2021 (UTC)
Be that as it may, Occam's razor still favours an early split of Sardinian, which is also supported by Harris/Vincent, "The Romance Languages". --Florian Blaschke (talk) 18:14, 12 December 2021 (UTC)
From The Oxford Guide to the Romance Languages (2016: 270):
'Scholars from the nineteenth century onwards (e.g. Gröber 1884:210f.) claimed that, because of the early conquest of Sardinia, Sardinian conserved elements from archaic Latin. Yet the survey in Mensching (2004b) shows that this hypothesis cannot be maintained, since most of the relevant features (e.g. lack of palatalization of /k/ and /g/ before /e/ and /i/) were still common in Latin during the first centuries AD. It thus seems that the Latin of Sardinia was not significantly different from that spoken elsewhere until at least the end of the third century AD.' Nicodene (talk) 21:09, 13 December 2021 (UTC)

Latin phonology according to Cser (2020)

The consonant and vowel sounds given in the tables, as far as I can tell, are currently those in Allen's Vox Latina (Allen 1978). However, while Vox Latina is indeed a very important reference work in this field, it's a couple of decades old, and The Phonology of Classical Latin by Cser, which was published in 2020 and is already cited in this article, gives a slightly different phonology for Classical Latin. Now, I am no expert in this field, and I may be beating a dead horse here (if I am, sorry), but I think that, since it's been over 40 years since Vox Latina and that we probably know more about linguistics and the phonology of Classical Latin now than we did in 1978, it would not be a terrible idea to update the consonant and vowel tables to better reflect a modern understanding of the matter. Thoughts?

(DISCLAIMER: I am a fan of Luke Ranieri (i.e. the guy behind polýMATHY and ScorpioMartianus), who promotes (and uses) the phonology of Classical Latin as proposed by Calabrese (who, as per other messages on this talk page, is regarded as "fringe" by most professionals in the field). However, at the same time, I am also interested in historical accuracy, based on our current academic knowledge, and if Ranieri turns out to be wrong in some aspect, then I will not deny that in favour of fan loyalty. I'd still watch his videos, sure, but I'd keep that error in mind when I watch one of his videos that are relevant to the error.)

—MeasureWell (talk) 23:09, 9 March 2022 (UTC)

Could you elaborate on what differences between Cser and Allen you have in mind? Nicodene (talk) 21:32, 10 March 2022 (UTC)
@Nicodene: Actually, never mind – turns out I misread a part of the book. To elaborate: on p. 34 (Section 2.3 "Vowels"), Cser states that
"Grammarians' remarks and Late Latin developments indicate that the short vowels (with the exception of [a]) may have been lower than their longer counterparts, i.e. [ɪ ɛ ɔ ʊ]. This well-known phonetic detail will be disregarded in the representations throughout. Furthermore, vowel length will be indicated only where strictly relevant."
Clearly, I didn't read the part that I've italicised and put in bold, because on the next page, Cser gives a table of vowel sounds in Classical Latin which uses notation such as [e] and [eː] to represent the long and short vowel sounds, instead of [ɛ] and [eː], respectively. However, while reading through the book to try to get an answer together to your question, I found a couple of other notes that might be valuable for this article (and the main article on Latin in general), such as the question of labiovelars (i.e. [kʷ ɡʷ] or [kw ɡw]), nasal vowels, and some other things.
To summarise: The tables as they are right now are correct: I misread the book. However, there are a couple of other notes that may be relevant to Latin phonology and orthography and Latin that we could add to those articles. What do you think?
Apologies for the late reply!
— MeasureWell (talk) 01:16, 14 March 2022 (UTC)

Labiovelars in Classical Latin

Seeing as my previous comment didn't go down so well (due to my inability to read), let me be more specific this time. Regarding the labiovelars (i.e. kʷ and ɡʷ), Cser 2020 reviewed all the arguments presented for both sides and found that although the balance tilts slightly against there being labiovelars in Classical Latin (although not very convincingly, according to the text), the evidence is still generally inconclusive. Since we can all agree that Cser 2020 is probably the most recent large work on the phonology of Classical Latin (as per earlier discussions on this talk page), I think we should update this for sure, since the current content, as far as I can tell, reflects what's in Allen's Vox Latina. I'm not an expert in linguistics so I don't know how to reword it into one dot point, so I'll leave a link to the original paper here.

Cheers — MeasureWell (talk) 12:40, 6 April 2022 (UTC)

@MeasureWell Hi again. That uncertainty is reflected in the current version of the article, as the consonant chart shows the † symbol after both /kʷ/ and /ɡʷ/. Admittedly, there should be a note explaining what the dagger means here, as this is not a standard usage of it whatsoever.
The bullet point below would certainly benefit from incorporating the results of Cser's analysis, or simply substituting them entirely for what Allen says. If I find the time I will read through it all and make an appropriate summary. Nicodene (talk) 22:18, 9 April 2022 (UTC)
@Nicodene: Hello, and sorry for bothering you about this matter again. Yes, I mainly had the bullet point in mind when I wrote my original (comment? post?), since, like I said, I'm not a professional expert in linguistics (and, I'll be honest, Cser 2020 is a little bit dense (in terms of content) IMHO). I've just read through the section on labiovelars in Cser 2020, and he does actually bring up and look at Allen's arguments in favour of the existence of [kʷ] and [gʷ] when discussing the phonetic arguments relating to the subject, so we should definitely include that at the very least.
Also, I was looking at the main page while typing this, and above the consonants table it already spells out what the asterisks indicate, so I think that adding the note below the table is a bit redundant. Here's the text in question:
"Sounds in parentheses are allophones, sounds with an asterisk exist mainly in loanwords and sounds with a dagger (†) are phonemes only in some analyses."
There are quite a few other things in Cser 2020, but I think for now we can just focus on the labiovelars, then tackle the other content one at a time. Thanks for the reply! :) — MeasureWell (talk) 23:19, 9 April 2022 (UTC)
Right you are, for some reason I'd missed the note above the consonant table entirely. The redundancy is fixed now. Nicodene (talk) 23:30, 9 April 2022 (UTC)

nūllī̆us

Is there anywhere that discusses a macron and breve over the same letter, or how it's pronounced (for example ī̆)? Can someone please add it to the article?Jimhoward72 (talk) 22:56, 9 November 2022 (UTC)

as far as i know, it means reconstruction methods can't clearly determine whether it was long or short; so you may pronounce it the way you prefer, with a 50% chance of doing it wrong. Hyperbaton (talk) 16:30, 13 November 2022 (UTC)

'why split the column if you're going to merge the only cell anyway?'

@Nardog To reflect the fact that long and short a were not qualitatively distinct, unlike the other long-short pairs. With your un-merger, the table now seems to imply that short a was more front than the long counterpart. Nicodene (talk) 17:44, 25 September 2023 (UTC)

Long/short is a quantitative distinction. Tense/lax is the qualitative distinction. Then let's not even have a different order for the first two columns either. It's a mistake to imply anything by the order of columns or rows alone. Nardog (talk) 23:20, 25 September 2023 (UTC)
@Nardog I'm sorry, I really don't see any coherence in this comment. I don't know where to begin here. What, in anything I have ever written, could possibly be taken to mean that I don't know that vowel quantity is quantitative or that vowel quality is qualitative? What leads to the conclusion that 'we might as well not have a different order'? How is the information I have stated implied 'by the order of columns or rows alone' when it is stated multiple times in plain English, for example in the paragraph directly above the table? Nicodene (talk) 02:13, 26 September 2023 (UTC)
That's exactly the reaction I had to yours. I don't see any way in which splitting the Central column into short and long but merging the cell for /a, aː/ "reflect[s] the fact that long and short a were not qualitatively distinct, unlike the other long-short pairs", nor how "With your un-merger, the table now seems to imply that short a was more front than the long counterpart" (aside from the fact you seem to have decided that the order of short and long has to mirror the vowel space, which makes no sense to me). Both quality and quantity are not only represented by the IPA symbols alone but—as you point out—also discussed in prose. So why attempt to imply with the structure of the table any more than what the labels (<th>) say? Nardog (talk) 02:21, 26 September 2023 (UTC)
@Nardog Every other column split (uː | ʊ, etc.) reflects different positions in the vowel space. Having such a split for a(:) could mislead someone quickly looking up the Latin vowel inventory into believing that there was some more modest counterpart to the qualitative difference between, say, English /æ/ and /ɑː/. Granted, the fact that the same IPA symbols are used, on our table, for Latin a: and a should tell them otherwise, but it doesn't hurt to make things as not-misleading as possible when there's nothing lost by doing so.
I genuinely don't understand how arranging vowels, on a vowel inventory table, according to their position in the vowel space 'makes no sense' to you when that is the entire basic principle of how such tables work. You might as well ask me why /i:/ is on the top left and /o:/ on the middle right. Granted, not every table you'll find in the literature always reflects the vowel-space positioning in every respect, but the general idea of following the vowel space is not some absurd fringe notion. Nicodene (talk) 02:52, 26 September 2023 (UTC)
I can see the idea behind switching the order of the columns for front and back vowels to make a more symmetrical table, but a) if anything, that layout seems to imply that short and long allophones differed more significantly in centralization than height, whereas future Romance developments suggest the opposite was true b) there doesn't seem to be a good way to order the central vowels if you use the order long - short for front and short - long for back. Also, if we're concerned with inter-article consistency, compare Standard_German_phonology#Vowels, Dutch_phonology#Vowels, and English_phonology#Vowels, which don't switch the order of tense-lax or long-short axes for front vs. back vowels. I think using the order short - long for each column is best (compare how IPA consonant charts use the left-right axis for voicing as well as POA).--Urszag (talk) 03:42, 26 September 2023 (UTC)
A thought: if we want to represent the vowel space more directly in a visual manner, it might be better to include something like Allen's triangle-within-a-triangle chart in addition to presenting the phonemes arranged by distinctive features in a tabular format.--Urszag (talk) 03:46, 26 September 2023 (UTC)
I'd actually thought about removing ɪ/ɛ/ɔ/ʊ from the table outright and replacing them with i/e/o/u, to match the phonemic representations used elsewhere on this page. Ultimately though /iː ɪ eː ɛ aː a ɔ oː ʊ uː/ didn't seem that unreasonable a system to posit, transitional from purely quantitative (Old Latin) to purely qualitative (Late Latin ~ 'Proto-Romance'). Now that you mention it, though, none of the sources cited here present a phonemic /ɪ/, etc.
A vowel trapezium based on Allen would work nicely, I think. Nicodene (talk) 04:18, 26 September 2023 (UTC)
However it's not clear where one should place all the vowels in a trapezium. Perhaps we could simply reproduce Allen's chart as-is, without converting it to a trapezium at all. Nicodene (talk) 04:56, 26 September 2023 (UTC)
Yes, I agree that would be better. A trapezium feels to me like it would create a misleading sense of precision, so something that's basically a reproduction of Allen's chart in vector graphics seems like the best image.--Urszag (talk) 05:01, 26 September 2023 (UTC)

Transcriptions and the issue of "original research"

@YanisBourgeois: I want to apologize for the brevity of my comments so far. I've re-added the "Citation needed" label to the transcription of the Aeneid passage because I don't believe Allen p. 74 is a valid source for the entire, specific transcription currently in the article. Making our own transcription of this specific passage based on interpretations of statements Allen makes in general about Latin pronunciation isn't straightforward. In my opinion, this falls under the category of "original research" or "synthesis": Wikipedia jargon for "making inferences based on sources" (rather than reporting things that are stated directly by the sources). Wikipedia has rules against doing this (in general) because it is difficult to agree on exactly what inferences are valid to make.

For example, you commented that you inferred that final -im, -em, -um had the quality of short vowels but the quantity of long vowels. Some other people have inferred this too, but not everyone does: an alternative position is that final -im, -em, -um had the quality of long vowels. Before a stop consonant (a condition that is met in "virumque"), Allen mentions the alternative of final -m being pronounced as a nasal consonant preceded by a short vowel. Because of uncertainties like these, we can't say that Allen provided this specific transcription. Urszag (talk) 11:08, 5 November 2023 (UTC)

Indeed. The phonetic transcriptions in that section may be lovely illustrations, but they do represent Original Research (and would do so even if Allen were followed to the letter) and so are against Wiki policy for that reason alone.
To replace them, perhaps a lengthy transcription could be found in some reliable source and introduced here with a caveat such as ‘here is an example of how a passage in Classical Latin may have been pronounced’. And probably likewise with the passage rendered in modern Italo-Ecclesiastical, which can for example be replaced with the one from Canepari (I don’t believe converting his custom IPA to the standard one is Original Research, given that he himself provides the key). Nicodene (talk) 13:35, 5 November 2023 (UTC)

Tables of orthography

A great deal of this article is taken up by tables showing approximations of IPA sounds in English and sometimes other languages. I really don't see the need for this, and it isn't normal practice on Wiki. That sort of thing seems best suited for Help:IPA/Latin.

The tables do contain some phonetic information not found elsewhere on the page (for example, the treatment of Latin /kʷu/ and /wu/) but that seems best placed alongside similar information in the main 'vowel' and 'consonants' sections. Likewise the descriptions of Ecclesiastical pronunciation. Nicodene (talk) 16:06, 5 November 2023 (UTC)

Retrieved from "https://en.wikipedia.org/w/index.php?title=Talk:Latin_phonology_and_orthography/Archive_4&oldid=1222186378"