Jump to content

Talk:Latin phonology and orthography/Archive 1

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

Phonology?

This page is not a discussion of Latin phonemes; it's a discussion of Latin letters. (C and K are not different phonemes in Latin; they are two letters for the same phoneme.) Not to mention that the comment about K is wrong; classical Latin uses K only for Kalendae and Kaeso, and in both cases it survived probably because of its use in abbreviations.

Should the page be renamed? Etc...

--Tb 05:21 15 Jul 2003 (UTC)

Why is there no discussion of Latin dipthongs? No Latin phonology could be complete without them. MattH 03:38, 30 Jul 2003 (UTC)

See the previous comment. The page is mistitled. Basically, someone wanted a page that says "how each letter is pronounced", and they misused a fancy linguistic word to do that. There isn't any description of the difference between long and short vowels either; just as important a subject as diphthongs. (Minimal pair, for example, legit [short E, present tense], legit [long E, perfect tense].) I'll add fixing this page to my todo list. --Tb 03:45, 30 Jul 2003 (UTC)

Didn't know this was here. The problem you're going to run into is that there is no single phonology for Latin; it changed over time. You have at least three major systems: a "classical" system whose distinctive feature is vowel quantity reconstructed from the Latin use of Greek poetic metres, which creates issues with consonants mostly by allowing elision of final /-m/; a "vulgar" system, probably but not necessarily later in time, in which vowel quantity has been eliminated and replaced by mergers, with changes in vowel quality; and the ecclesiastical and other scholarly pronunciations of Latin used in singing, based largely on reconstructions of classical vowels, with contemporary Italian influencing the consonants. An article on Latin phonology, or even on the relation of Latin to the Latin alphabet, is going to have to address all of these issues. -- IHCOYC 04:07, 30 Jul 2003 (UTC)

The obvious choice is to handle classical Latin, which is the best-known. Ecclesiastical Latin doesn't really exist; traditional French pronunciation of Latin is wildly different from the more usually heard Italianate version, and the German is different still (but closer to the Italian). These are interesting to musicians, but little beyond them.

Vulgar Latin is several beasts: are we talking vulgar Latin from the Augustan age? Silver Latin? Or just early Romance? (vowel quantity, for example, was a real feature of the first two, not the last.) Vowel quantity is a feature of Jerome's translation (the vulgate); I recall reading a fascinating article which discussed Jerome's use of meter. I reject the notion that classical vowel quantity is known only by "reconstructions from the Latin use of Greek meters"; there are distinctive Latin meters, first off, plus there are good primary sources which directly identify vowel length, and there is good evidence from different evolution of the words in Romance depending on vowel quantity in the Latin form.

There is, in fine, a perfectly sensible beast called "classical Latin phonology", and we know a rather suprising amount of detail about it. It might be nice also to discuss Romance phonology, but we know vastly less about that. It might be nice to discuss late silver Latin phonology, but we know only a little about that, just a tad more than Romance plus guesses about things from classical pronunciation that we guess didn't change.

The final -m issue is no issue at all: it's a data point for God's sake! We know both from Romance reflexes, from ancient linguists, and from the poetic custom, that -m worked a nasalization on the preceding vowel in Classical Latin, and was essentially not an independent consonant. As a result, there is no special phonological rule about eliding -m in poetry; it's just the normal rule that final vowels in poetry are dropped when the next word begins with a vowel.

So I reject the idea that this is some impossible task: a reasonable summary of the conclusions of Vox Latina would be a perfectly sensible approach. --Tb 04:19, 30 Jul 2003 (UTC)

I suppose much depends on what you imagine the scope of the article is. If the purpose of the article is to inform people on how to pronounce Latin, it strikes me odd to claim that "Ecclesiastical Latin doesn't really exist." If people hear spoken Latin, that is the likeliest Latin they're going to hear. More importantly, if someone is called on to sing in Latin, that is the pronunciation they need to know. Don't want someone to get the notion that they should sing ". . . pax hominibus bonai woluntatis."
Not "woluntatis": "uoluntatis" or "VOLVNTATIS". —Chameleon 14:32, 8 May 2004 (UTC)
If we are talking about an actual article about the sounds of spoken Latin, some kind of diachronic approach seems unavoidable. Writing such an article doesn't strike me as an impossible task, only one that will need to address changes over time. My understanding in any case is that most of the sound changes that lead to proto-Romance were well under way by the first century A.D. --- IHCOYC 13:39, 30 Jul 2003 (UTC)

When I say "ecclesiastical Latin doesn't exist", I mean that there is no one beast called "ecclesiastical Latin" or "church Latin". The traditional French pronunciation--including singing--is very different from the Italianate one that English speakers are most familiar with. French singers of Palestrina traditionally sing it quite differently from the way Americans and Italians do. There have been English composers that intended their pieces to be sung using the Received Pronunciation, of all the crazy things! So there isn't "the pronunciation" if we are talking about "church Latin".

When you say "most of the sound changes that lead to proto-Romance were well under way" there are so many vaguenesses that of course that statement must be true. How do you distinguish proto-Romance from Romance? We don't even know what all the sound changes are between 1C BC and Romance! So what's "most"? We know some: like the rather early nasalization of -m; the dropping of initial H, and so forth.

We don't know exactly how latin D was pronounced (more like English and Italian, or more like Spanish?) At that level of detail, an article on Latin phonology is impossible. So, like I said, the article is mistitled. We need one on Latin pronunciation. It would focus on presenting the classical pronunciation, say a few nasty things about the british Received Pronunciation, describe Italian/English church Latin as being basically Italian pronunciation, give some history about a few of the sound changes in later Latin.

Oh, and there are composers who have composed pieces intending that they be sung with the classical pronunciation. We can have a good article, a perfect article, or a perfectly miserable article. We have the last, I propose to write the first. The second is not worth trying for. --Tb 16:23, 30 Jul 2003 (UTC)

I'm not sure that M as last consonant was silent eg. decem [dekem]. I was thought that M is not silent according Pekkala: Ars grammatica. --Hannu (talk) 15:50, 4 June 2008 (UTC)

I haven't read the work of Pekkala, but are you sure that he does in fact intend to describe the phonology of the classical period, and with enough details to touch on such minute aspects? Or is he dealing with post-classical pronunciation, or maybe giving recommendations to modern people how to pronounce classical Latin? In that case, it is not inconcievable that he simplifies it somewhat. For it is beyond dispute that the final -m of Classical Latin was at least peculiar in some way. I can adduce at least three arguments for this: 1) expressed testimony of antique writers discussing the strange sound of final -m. 2) how final -m does not hinder elision, thus in fact working as a vowel. 3) that final -m often is lost in inscriptions, and furthermore has disappeared in the Romance. Does Pekkala in any way address these arguments? Alatius (talk) 13:20, 24 June 2008 (UTC)

NPOV

About the changes to my last version. "Corruption" in linguistics is not a Point of View. When a form drifts away from a previous standard or norm, it is called a corruption of it. It's not like political corruption.

Nor is it POV to state that Italians, more than people in other countries, tend to consider their own pronunciation of Latin to be correct. This is just a fact. Indeed, given that Rome is in Italy, one would expect them to feel this identification with the Latin language; it would be strange if they didn't. The fact that Italian pronunciation has become the ecclesiastical pronunciation shows how widespread the belief is that the Italians pronounce Latin right. It is silly to change this into a statement about Italian pronunciation being "the most commonly heard pronunciation of Latin". Common where?

The point is not what pronunciation sounds nice, or is advisable to adapt while singing so as not to pronounce differently from other people in the choir, but rather that the Italians generally think the Ancient Romans actually went around talking like they do now —Chameleon 14:32, 8 May 2004 (UTC)

Though it's a POV I might be inclined to agree with, the very notion of linguistic corruption implies a falling away from a past standard. Since human language invariably changes, and there is no particular reason to believe that the language of the past is superior to that of the present, the very notion of "corruption" inevitably implies prescriptivism, and the current tendency of writers on language is to be descriptive rather than prescriptive.
If current Italians actually believe that the ancient Romans used their pronunciations, I'd want to see a reference for that. The standardisation of ecclesiastical Latin is an interesting subject in itself. It remains true that if a person hears a human voice using Latin, the odds are far greater that it will be pronounced in the ecclesiastical manner than in any other. A living tradition uses ecclesiastical Latin; no such living tradition uses any of the various reconstructions of Roman Latin. Smerdis of Tlön 00:52, 9 May 2004 (UTC)
So what? —Chameleon 10:28, 9 May 2004 (UTC)

Reverting last change

Someone has just added this bit

"and differently used lexicon. The Classical latin word for mouth is what Romance languages uses for cave."

The point of the sentence was to point out the main differences. If we add lexicon to the list, the sentence becomes pointless. Also, BVCCA meant cheek, not cave, so the example given isn't valid. —Chameleon 18:33, 21 May 2004 (UTC)

Are we sure?

Are we sure that semi-consonsantal I was pronounced /jj/ and not simply /j/?

Are we sure that the Greek consonants Χχ, Φφ and Θθ were, at the point they entered Latin, pronounced /kh/, /ph/ and /th/ rather than the traditionally accepted /x/, /φ/ and /θ/?

In any case, PH became confused with F, not P.

Chameleon 11:02, 19 Nov 2004 (UTC)

I've based my statement about jj on Sidney Allen's Vox Latina. He justifies this on the basis that:
  • Indo-European single j between vowels was generally lost
  • it's specifically mentioned by Quintilian
  • it results in a double consonant in Italian
  • the preceding syllable is heavy (as shown presumably by scansion or the position of the accent) even when the vowel is short.
as far as ph etc are concerned, I think (again based on Allen) that the Greek letters were still aspirated stops at the point that these combinations entered Latin (second century BC) although they may have become fricatives very soon after this. Apparently Φφ was represented in Latin first by P, later by PH (when an aspirated pronunciation was becoming fashionable in some Latin words) and eventually by F.
rossb

Proposal to move this page

I propose to move the page "Latin phonemes" to "Latin spelling and pronunciation" on the basis that it already covers more than just phonemes (as others have already mentioned) and I'm intending to add a section on light and heavy syllables, postion of stress, etc.

Any comments?

rossb 23:03, 28 Nov 2004 (UTC)

OK Chamaeleon 23:07, 7 Jan 2005 (UTC)

Ive now moved it. rossb 07:44, 28 Jan 2005 (UTC)

AE and OE

I had always been under the impression that <AE> and <OE> represented /ai/ and /oi/, rather than /ae/ and /oe/. I changed one of the instances in the article where it claims that they were /ae oe/ a few days back without giving it much thought, but apparently missed the other one, and now I'm not so sure that was the right thing to do anyway. I'd appreciate any more information anyone can give. Thanks. --Whimemsz 03:08, May 11, 2005 (UTC)

Reverted change about AE and OE

Most sources state that pronunciation of AE and OE shifted from /ai/ /oi/ to /ae/ /oe/ during second century BC. In the beginning I've also added a comment about "late Republic" as a time specification for the "Classical" pronunciation dealt with by the article.

Okay. Thank you! --Whimemsz 18:58, May 25, 2005 (UTC)

According to my high school teacher and Ars grammatica (published by Helsinki University) Latin ae and oe are e:, not ai and oi. I also read one of the links and there it is stated as above me comment. Please, do not change this fact for a while. I do like to check it. --Hannu 05:07, 2 August 2006 (UTC)

The article, from my understanding, describes classical Latin pronunciation but notes different pronunciations for later periods. According to the page on Vulgar Latin, the original pronunciations were /ai/ and /oi/. But if you want a few days to check the facts, I'm cool I guess. AEuSoes1 06:24, 2 August 2006 (UTC)
I have checked the facts. The spelling is for AE and OE e:. During Republic-period it was for AE /ai/ and for OE /oi/. This is quite well sayed in external link The Roman Pronunciation of Latin, by Frances Ellen Lord, in the Gutenberg Project. You could check it out if you want. I found it quite usefull and it also supports the sources what I stated in my alrier post.--Hannu 06:50, 2 August 2006 (UTC).
If it was /ai/ and /oi/ in the Republic period, then we should put it as those and mention a change. [e] is a Vulgar Latin pronunciation [1]
AEuSoes1 10:09, 2 August 2006 (UTC)
I would say that AE and OE are e: in the listing. The Roman Republic period thing should put as I put it in the first place. I don't think that outside education it was e:. To me it has been said that it was espeally during Golden Age of Latin (Cicero, Augustus, etc.) AE and OE /e:/.--Hannu 12:57, 2 August 2006 (UTC)

From An Introduction to Vulgar Latin [2].

Æ was originally written and pronounced ai, but through mutual attraction of its two parts it became œ, later : cæcus, cælum, quæro.
...
The regular change of æ to took place largely in Republican times in unaccented syllables.; in stressed syllables in the first century of our era and later. E for æ in dative endings occurs early: Corssen I, 687 ff. About the middle of the first century B.C., when Varro cited edus for hædus as a rural form, stressed æ was probably still a diphthong in the city but had become in rustic Latium; some hundred years later came into the city and pervaded the provinces: Lat. Spr. 465. Terentius Scaurus, in the first century, says that æ represents the sound better than ai
...
The spelling of e for æ was usual in unaccented syllables (as sancte) before the third century, in stressed syllables (as questor) from the fourth century on; it may be called regular by the fifth century...

AEuSoes1 22:27, 2 August 2006 (UTC)

Article split needed

This article could use a split. Calling it "Latin pronunciation" is not appropriate to begin with, since it implies that it's a pronunciation guide, which is not encyclopedic. I also suggest separating the orthography and phonology since the two are definetly not one and the same, something that goes for virtually all languages. I don't know if it's the most common method yet, but Latin orthography and Latin phonology seem like the most reasonable titles to me and have already been put to use with other languages.

Peter Isotalo 14:34, 14 July 2005 (UTC)

O as in Pot?

From the article: O /o/ (as in pot, but shorter)

Isn't the pronunciation of pot like [pɑt] or pawt? So wouldn't it be wrong to use pot for this example?

--Hecktor 10:13, 1 October 2005 (UTC)

The main problem is that the pronunciation of pot varies between British English and American English. Perhaps it would be better to specify the dialect of English that is meant. (Presumably it's British, but I'm not sure.) --Siva 23:09, 31 October 2005 (UTC)

Dental /t/?

In Spanish, Portuguese, French, Romanian and Italian, t and d are always dental, not alveolar as in English. Would it be reasonable to suppose then that the same was true in Latin? If so, it would be necessary to amend the article to prevent Anglocentric Anglophones from thinking that Romans pronounced their ts and ds on their alveola. --Siva 23:19, 31 October 2005 (UTC)

Not only in Romance langauges, but also in all the sister languages of Latin, /t/ is dental, viz, Slavic, Sanskrit, Ancient Greek, Old Persian, etc. So there is no harm in assuming that Latin /t/ must also have been dental; on the other hand, assuming Latin /t/ to be alveolar /t/ as in English, like it is pronounced in Britain and America, would be absolutely without any justification or logic. I have also added two links on /d/ as references for dentals.Cygnus_hansa 19:28, 19 December 2005 (UTC)

Thanks; that's what I thought too. --Siva 22:56, 25 December 2005 (UTC)

V

"(as in win) when short and unstressed before another vowel."

I am assuming that otherwise it is /v/ or /u/? Pronuncing words like "via" as /wia/ or /wja/ seems impossible.

- 81.15.146.91 19:38, 2 January 2006 (UTC)

/wia/ impossible? But that's precisely how "we are" is pronounced in Standard British English... -- Picapica 18:56, 12 February 2006 (UTC)
In class we always said /wia/. 82.139.85.33 22:55, 22 August 2006 (UTC)
/wia/ is correct. It is never pronounced as /v/, only /u/ or /w/. —Preceding unsigned comment added by 72.177.15.247 (talk) 09:24, 7 December 2008 (UTC)

V

I have reverted a recent change stating that V was pronounced /u/ and then describing it as a "blurred" u. Though consonantal I and V were more exactly semi-vocalic sounds, I find it safer to describe V as /w/ (especially since the // brackets denote a "broad", phonological transcription). This also avoids inconsistency with the rest of the article which continues to use /w/ and /j/. Maybe someone has a better solution? AlfredoM

short a, i, u

Is the pronunciation given for these vowels correct? I thought they were simply short versions of the long vowels [a:], [i:], [u:]! The pronunciation given for the short vowels seems to be based on English, not Latin. FilipeS 23:04, 12 April 2006 (UTC)

I agree entirely. There seems to be some inconsistency here, since later they are given as [a], [a:], etc. thegaij 15:41, 23 May 2006 (UTC)
Yeah exactly - I mean fine if we have a source, but what's given there looks English-based. I always thought they were just long vs. short. This REALLY needs to be right - it's extremely misleading if wrong IMO. Xipirho 03:04, 1 April 2007 (UTC)
Sidney Allen argues that there's no qualitative difference between long and short a, but that there is indeed a difference between long and short i and u. Short /i/ was similar to long /e:/, as seen by errors in inscriptions, and also developed to the same sound: /pira/, /ve:rum/ > Italian /pe:ra/, /ve:ro/. The same goes for u. Alatius 10:21, 1 April 2007 (UTC)
Then shouldn't the vowel chart reflect the qualitative as well as the quantitative differences?Szfski (talk) 12:47, 18 February 2009 (UTC)
In a phonemic transcription, you can get away with detouring a bit from phonetic reality. As long as it's between / slashes /, it's a good idea to keep it this way to be consistent with the literature on Latin phonology. — Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 14:35, 18 February 2009 (UTC)
No argument from me on that, but I'm of the opinion that, at least in the imperial period, there's no way to tell (with the exception of /a/) whether the phonemic distinctions were quantitative or qualitative to the Roman ear. I have no doubt that the good glorified schoolmarms whose prescriptions have come down to us taught their pupils that the distinctions were quantitative. But the errors in inscriptions (i for e, for example) along with the trends that vulgar latin is known to have followed suggest to me that the phonemic distinction was only quantitative for part of the history of classical latin, seeing as the colloquial register (which was definitely moving into qualitative territory as early as the 1st century) would almost certainly have affected the ways in which the forms of upper-crust, classical registers were realized on Roman lips. In fact, even in late Latin poetry (which as a rule tended toward archaism rather than innovation) one sees vowel quantity losing importance concurrent with nascent stress-based meters (the Pervigilium Veneris is the most famous example.) I don't mean to seem prickish, but the vowels in the chart strike me as overly simplistic at least, and verge on being downright incorrect.Szfski (talk) 17:18, 18 February 2009 (UTC)
I'd love to see the article reflect an academic conversation regarding that. If the vowels in our chart are overly simplistic, this is also true of literature I've seen about Latin pronuncitiation. — Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 17:36, 18 February 2009 (UTC)
See for example The Latin Language at the End of the Empire: A Phonological and Morpho-Syntactic Sketch, by Paul A. Gaeng (Modern Language Studies, Vol. 1, No. 1 (Feb., 1971), pp. 3-12 )- "In the course of the second and third centuries AD (and possibly even before that) new habits of pronunciation gradually became generalized. The accent came to be increasingly associated with stress rather than pitch, and this new feature completely upset the old rhythm based on quantity. St. Augustin notes that his fellow countrymen were no longer aware of the classical syllable quantity and hence they were unable to distinguish between long and short vowels."Szfski (talk) 23:53, 18 February 2009 (UTC)

R

is the latin rolling of the tounge more or less subtle then the Japanese pronunciation of the letter R?

The pronunciation of the R in Latin is somewhat controversial. It's probably an alveolar trill as it is in most Italian and most Romance languages. —Blurrzuki t - c 21:44, 13 August 2006 (UTC)
Japanese uses a flap for the R, not a trill. —Preceding unsigned comment added by 72.177.15.247 (talk) 09:31, 7 December 2008 (UTC)

Palatalization of G before A in French?

  • Palatalization of /g/ before /e/ and /i/, and of /j/, into //. French has a second palatalisation of /g/ before Latin /a/.

What would be some examples of this? FilipeS 20:32, 5 June 2006 (UTC)

I mean, of the "second palatalisation of /g/ before Latin /a/"?... FilipeS 15:15, 29 June 2006 (UTC)

O.K., a reference has been added, but how about some examples of this palatalization? FilipeS 16:03, 13 October 2006 (UTC)

Given that we don't show examples of the first, more important and widespread, palatalisation, it would seem inappropriate to show examples of the second (cantare -> chanter / gaudia -> joie) palatalisation. --62.49.68.79 22:36, 13 October 2006 (UTC)

I don't know about that, but thanks anyway for the example. I had never noticed or read about this palatalization! FilipeS 14:00, 14 October 2006 (UTC)

More about spelling

The article doesn't say much about spelling beyond touching on the i/j, u/v issues.

Modern English spelling, as recently as Shakespeare, was not very uniform. (i.e. The same word may be spelt different ways even by the same author in the same work. There was no "official" spelling.) But it is also very irregular. (i.e. The pronunciation of one word doesn't necessarily give you any clue as to the pronunciation of a similarly spelled word.)

(If there are more accepted words that linguists use in place of my "uniform" & "regular", please educate me.)

Modern Latin students, however, are presented not only with very regular spelling, but also very uniform spelling. (e.g. A Latin I text will never spell "cum" as "quum", & will likely not even mention it as a variant.)

I know that Latin spelling changed over time (i.e. a word may be spelt differently between classical Latin & medieval Latin), but I wonder how uniform it was within any particular era/mode.

Where did the uniform spellings found in our Latin textbooks & student dictionaries come from? Is it classical, medieval, or modern? Is uniform spelling found in authentic classical Latin, vulgar Latin, or medieval Latin?

I realize that this is a complex issue & that there are a lot of holes in our knowledge, but I'm just looking for any generalizations & accepted hypotheses as would be appropriate for this article.

I'm no expert, but I think Latin spelling was first standardized during the Carolingian Renaissance, and later during the Renaissance, when scholars began to bring out critical editions of classical texts. I'm not sure, though… --Siva 16:22, 12 July 2006 (UTC)
Classical Latin, being a literary language, had fairly consistent word spellings. Though Latin, being a more phonetically consistent language than English (in Latin, C is always pronounced one way, not four or five different ways!), had fewer potential ambiguities in how to spell words, less literarily "elite" writing (e.g., graffiti and inscriptions on tombs) tended to be increasingly inconsistent in spelling (often varying in terms of synonymous-sounding letters, such as whether to use C or K or Q in certain places and using variant (mis)spellings such as eks for the word ex), and, of course, bearing the many grammar and vocabulary differences of late and Vulgar Latin). -Silence 16:30, 12 July 2006 (UTC)
Wait a minute… if Classical Latin had such regular spelling, then why does Cicero use weird forms like "quom" for "cum"? And anyway, how do we know how Classical Latin was spelled? I thought that all we have of Classical Latin is somewhat corrupted medieval manuscripts, and that the only surviving writing samples from Roman times (e.g., graffiti and inscriptions) were written by people who tended to be inconsistent in the way they spelled words. --Siva 01:20, 15 July 2006 (UTC)
Note: I have no subject specific knowledge.
I is possible that Cicero pronounced "cum" as "quom" and therefore wrote "quom". Nothing wrong with that, in my opinion. Isn't it also possible that in the renaissance new editions of Latin works were standardized to a "greatest common divisor" Latin? Still nothing wrong with that, in my opinion. Shinobu 18:05, 19 July 2006 (UTC)
All right, so perhaps Cicero pronounced "cum" as "quom", and wrote it as such. But that still doesn't say anything about whether Classical Latin was spelled regularly or not. For all I know, it may have been the Renaissance editors who standardized Latin spelling, and there may not have been any standard spelling in Classical times. All I'm saying is that if the only evidence we have is in the form of medieval manuscripts (which may be corrupted), graffiti (written by less-literate people), and inscriptions (ditto), it's not possible to say with any certainty whether the Romans recognized any "standard" form of written Latin. This wouldn't be the case, however, if it turned out that, for example, the Romans themselves were concerned with proper spelling. Is there any evidence other than manuscripts, graffiti and inscriptions that suggests that the Romans had standardized orthography? --Siva 01:19, 20 July 2006 (UTC)
What nonsense! What proof do you have that Latin spelling was irregular? FilipeS
This[3] book. It gives actual examples of spelling inconsistencies. AEuSoes1 23:22, 5 August 2006 (UTC)
Although of course we don't know wether spelling was irregular or rather the pronunciation. "quom" (vs "cum") might be an irregular spelling, but it might also be a regular spelling of a non-standard (regional, perhaps?) pronunciation. Shinobu 10:13, 6 August 2006 (UTC)

the son of Latin

"Indeed, some have dubbed Castilian the son of Latin, and Portuguese and French the daughters of Latin, due to the alleged masculine and feminine sound of them, respectively. "

I deleted this from the page. The "some" should be properly referenced; and how a language sounds is subjective.

I don't think it should be there even if there were a source. Linguistic discussions of language history usually talk only about "daughter" languages. AEuSoes1 23:02, 28 July 2006 (UTC)

SU, GU

I think that in the article should be noted that SU and GU are /sv/ and /gv/ as in Suecia (svekia) and lingua (lingva).--Hannu 13:29, 6 August 2006 (UTC)

I thought the U was only /w/ in these situations before a vowel? Compare "sus", "sursum", "sub", "succedere", and "gustare", "guttur", "gutus", "gubernare". Shinobu 14:31, 6 August 2006 (UTC)
I'm pretty sure that in such positions V was /w/ but that it changed to /v/ sometime. I'm not sure when. AEuSoes1 23:47, 6 August 2006 (UTC)
I'm not an expert in Latin sound changes, so I'll leave that to other people. The point I was trying to make is that it's not always /w/, but often /u/, for example in the words given above. Shinobu 08:04, 7 August 2006 (UTC)
I think there was a misunderstanding. In Latin, when U comes before another vowel it represents a consonant, no matter the onset.AEuSoes1 08:10, 7 August 2006 (UTC)
So we agree. I thought that S, G bit was a bit strange. So I'm correct when I say "fuerit" /fwerit/ ? Which means that the article already mentions it (note 2 at the vowel table). Shinobu 08:45, 7 August 2006 (UTC)

I think there's a misunderstanding here. Consonantal V (ie the semivowel /w/) changed to /v/ in most cases, but not after S or G, where it remained /w/. Consequently when the moodern u/v distinction came into use, /w/ after S or G was written with a u rather than a v, although from the standpoint of Classical Latin there is no logic to this distinction. However V (and I) before another vowel can also be a normal vowel. Compare suādeō — "I urge" — /ˈswa:deo:/ and sua (feminine singular of suus, "his/hers" — /ˈsua/ — two syllables. And "fuerit" was definitely /ˈfuerit/ (three syllables). Compare also Iūlus — /iˈu:lus/ — three syllables, name of a character in the Iliad, with Iūlius — /ˈju:lius/ — "Julius", as in Julius Caesar. --rossb 23:26, 13 October 2006 (UTC)

Disputed tag

Could someone please add the relevant talkpage section to the {{disputed}} tag? I have noticed that anonymous disputed tags otherwise tend to stay in the article even after the dispute is long settled. Shinobu 08:34, 7 August 2006 (UTC)

V inconsistency with the Dutch article

een u wordt, behalve na een q of g die tot dezelfde lettergreep behoort, altijd als [oe] uitgesproken (ceterum = keteroem)

Translation:

a u is always pronounced as /u/, except after a q or g in the same syllable.

Note that this is different from the pronunciation rules here. Do we have a source for our guide? 82.139.85.33 23:14, 22 August 2006 (UTC)

http://books.google.com/books?id=_OzEl6nLsGIC&pg=PA4&lpg=PR4 for one. I'm sure there are also references in the Latin alphabet page. The Dutch article is simply wrong. AEuSoes1 02:53, 23 August 2006 (UTC)

Seems likely, yes. I'll notify them. Shinobu 23:13, 25 August 2006 (UTC)

inconsistent use of "semivowel" and "consonant"

An expert should decide whether "semivowel" or "consonant" should be used to describe the use of V and I in Latin; the article now confusingly uses both. --Espoo 07:51, 10 September 2006 (UTC)

Semivowels are consonants. AEuSoes1 08:56, 10 September 2006 (UTC)
I know, but to not confuse readers, the article should not use two different terms in talking about V and I in Latin. --Espoo 12:58, 10 September 2006 (UTC)

Ecclesiastical Latin in France

Greetings,

I live in France and have been to the traditional Latin Mass at several places, and can therefore tell you that here v is pronounced /v/ and not /u/. Why was my edit reverted?

Thank you,

Grumpy Troll (talk) 05:04, 22 September 2006 (UTC).

Because it's humanly impossible to pronounce, for example caput as */ka.pvt/. U can be used to indicate /v/ when it's in the onset of a syllable but that's done elsewhere as the paragraph indicates.

Don't use your personal experiences. Use sources. AEuSoes1 06:23, 22 September 2006 (UTC)

Actually, it's not impossible to pronounce. "Don't use your personal experiences. Use sources." Heh. Have I told you about the island of Krk yet? Shinobu 01:28, 23 September 2006 (UTC)
I'm afraid there are a number of problems with [ka.pvt]. Most importantly is the fact that [p] and [t] are voiceless while [v] is voiced. If you mention Krk because of the syllabic r I'm not impressed. Sonorants, a sound class that [v] does not belong to, can be syllabic even in English such as in finger, button, rhythm, and people.

Funny, I initially read your quoting me as a request for me to cite my sources. For what it's worth, here's a little ditty on why voiced obstruents don't form clusters with voiceless ones:

Kehrein, Wolfgang and Chris Golston. (2004). A prosodic theory of laryngeal contrasts. Phonology 21.3. 325-357. AEuSoes1 02:20, 23 September 2006 (UTC)
"while [v] is voiced": As is, by definition, any vowel. [v] as a vowel may or may not actually exist in some obscure language - I don't know. But to say that it's impossible to pronounce, well, that's false. Just try it and exercise on it for a while. Of course [p] and [v] are articulated at the same location, so such a sound-combination would probably not be stable, since it's easier to go [bv] or [pf], but that's something different. I don't really care whether there actually are languages that use [v] as a vowel, but when people start throwing words like "impossible" around... Shinobu 03:02, 23 September 2006 (UTC)
Well I've got some sourcing to back my claim up. You could also check out Vowels and Consonants as well as Sounds of the World's Languages. I suppose it is quite a bit harder to say something is impossible rather than implausible or unlikely. AEuSoes1 03:56, 23 September 2006 (UTC)
Implausible, yes. Unlikely, very. But not actually impossible, that's all. I think that for a sound to be impossible as a vowel it must block airflow - I can't imagine the [t] as a vowel. Of course this has nothing to do with "Ecclesiastical Latin in France", so perhaps we'd better wrap it up here. Shinobu 16:08, 23 September 2006 (UTC)
Please do excuse me for the hassle caused, for I just wanted to say that in France we pronounce, for example, lavábis me [lavabis me] and not [lauabis me], and super nivem [super nivem] and not [super niuem]. God bless you, Grumpy Troll (talk) 18:06, 23 September 2006 (UTC).
I thought as much :-) Over & out, Shinobu 18:41, 23 September 2006 (UTC)

Pitch accent? Force accent?

I don't see any discussion of pitch accent vs. force accent. Any particular reason for this? Bi 17:54, 27 September 2006 (UTC)

It's usually called a stress accent, and is covered. --InfernoXV 19:33, 27 September 2006 (UTC)
But what about pitch accent, which (if I read F. E. Lord right) was used by the Romans? Bi 07:58, 28 September 2006 (UTC)

Quality of vowels of Vulgar Latin

From the article:

Length of vowels was more significant and more clearly defined in Latin than in modern English. In Vulgar Latin, there was a difference in quality between long and short vowels (except a) in that short vowels were more open than long vowels ([ɛ] [ɪ] [ɔ] [ʊ] vs. [eː] [iː] [oː] [uː]), but this was less important than the differences in quantity (English also has phonemic length distinction, but quality has become more significant over time).

First of all, it needs to be clearer that this refers to the later (post-classical) stages of Vulgar Latin, not to Vulgar Latin since its beginning. Secondly, I would like to know what is the source for claiming that Vulgar Latin ever had the lax vowels [ɪ] and [ʊ]. FilipeS 15:40, 29 September 2006 (UTC)

I believe this is a source. I'd have to check it out again. AEuSoes1 21:19, 29 September 2006 (UTC)

Well, Heinrich Lausberg, Romanische Sprachwissenschaft makes no mention of the lax vowels [ɪ] and [ʊ]. His description of the vowels of Vulgar Latin is as follows (classical Latin is on the left, vulgar Latin on the right):

  • ǎ → a;
  • ā → a;
  • ě → ɛ;
  • ē → e;
  • ǐ → e;
  • ī → i;
  • ǒ → ɔ;
  • ō → o;
  • ǔ → o;
  • ū → u;

The notation for the vowels on the right is mine. Lausberg uses underdots and ogoneks instead of IPA.

Yeah, now that I've checked my source I see that it says the same thing. It's possible that the short high vowels went from near-close to close-mid. AEuSoes1 22:10, 30 September 2006 (UTC)

Something else, about this passage:

Length of vowels was more significant and more clearly defined in Latin than in modern English. In Vulgar Latin, there was a difference in quality between long and short mid vowels in that short vowels were more open than long vowels ([ɛ] and [ɔ] vs. [eː] and [oː]), but this was less important than the differences in quantity. English also has phonemic length distinction, but quality has become more significant over time.

Obviously, I don't disagree, but I am a bit uneasy with the broad statement that "In Vulgar Latin, there was a difference in quality between long and short mid vowels..." Wouldn't that depend on the time period being considered? I would like to see the article discuss when Vulgar Latin began to diverge from Classical Latin in this respect. FilipeS 13:14, 11 October 2006 (UTC)

I'm pretty sure that that's something we don't really know. It's possible that there was some sort of punctuated equilibrium wherein over the course of a few generations the emphasis on quality appeared and it stayed that way for a few centuries. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 18:44, 11 October 2006 (UTC)

Yes, that makes sense, but then it's misleading to word the article in a way that suggests Vulgar Latin was always different from Classical Latin, concerning vowel quality. On the other hand, as you note, there isn't much evidence either way, so it's also possible that there was a difference in quality between long E, O and short E, O already in Classical Latin! This happens in many natural languages with phonemic vowel length. FilipeS 10:11, 12 October 2006 (UTC)

Affricates?

I propose to remove the "Affricates" line from the table. If the letter Z did indeed represent /dz/ (which Allen for instance denies, saying that Greek Zeta had come to be pronounced /z/ by the time that the letter Z was introduced into Latin) this was not a true feature of Latin phonology. Moreover the affricate versions of C and G are much later than classical Latin, and do not even occur in all modern Romance languages. --rossb 14:53, 12 October 2006 (UTC)

That sounds fine. I put in the postalveolar affricates so as to discuss them below and reinforce that C and G were not soft in Classical Latin but simply glancing at the table can give the wrong impression. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 20:44, 12 October 2006 (UTC)

Allophones?

I don;t agree that the "soft" pronunciations of C and G in ecclesiastical Latin are allophones, for the following reasons:

  • phonetically they're pretty different from the hard pronunciations
  • ecclesiastical pronunciation is extremely close to the standard pronunciation of Italian, and the equivalent sounds are definitely not allophones in Italian (although they no doubt arose from allophones in vulgar Latin)
  • in both Classical Latin and later Latin there are plenty of examples of words (mostly Greek in origin, but naturatlised in Latin) starting with che-, chi- and chy- which using the ecclesiastical pronunciation would be pronounced with /ke/ and /ki/ so the hard and soft versions cannot be said to be in complementary distribution. --rossb 22:51, 13 October 2006 (UTC)
There are many languages with such allophony. It'd take me a few paragraphs to explain the philosophy behind their similarity. Italian has gone through quite a few changes from Latin and so citing it as a reason why Latin's occurence of postalveolars isn't allophony is unconvincing. Show me a minimal pair. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 01:14, 14 October 2006 (UTC)
A quick check of the Oxford Latin Dictionary reveals chaere (a greeting) which would form a minimal pair with cerae (genitive of cera, "wax". If you allow just parts of words there are of course large numbers of minimal pairs. But more fundamentally, i don't think I explained my point about Italian very well. Basically the "Roman" ecclesiastical pronunciation of Latin did not develop autonomously, it was the way Latin came to be pronounced by people whose native language was Italian. Its phonology is accordingly a subset of Italian phonology: all the phonemes of the ecclesiastical pronunciation are also Italian phonemes, so to postulate that variant pronunciations are allophones in ecclesiastical Latin when they are not in Italian seems very odd. --rossb 07:43, 14 October 2006 (UTC)
Your explanation of Italians speaking Latin makes me think of my second-language practices. When I speak Spanish, [ð] is an allophone of /d/ even though /ð/ is a phoneme in my native English. I am able to code-switch to another inventory. Granted, my English phonology makes it a fricative rather than an approximant, but the status as an allophone is still the same. It's possible that mediaeval Italians didn't do this, but it's also possible that they did.
I think we can just not have the word "allophone" for those two letters but keep the brackets. Your argument is going to need some sources before I'm convinced, but it's too much work (for either of us) to look up especially when you consider the payoff. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 08:13, 14 October 2006 (UTC)
OK I've removed the word "allophone" as agreed. --rossb 10:24, 14 October 2006 (UTC)

I can't speak about Latin, which I do not know that well, but, in Romance languages at least, I would agree that soft C and soft G are not allophones of hard C and hard G. For example, in Italian and French C and G are pronounced soft before E or I — but there are also words with hard C and hard G before E or I, except that those are spelled with CH, GH in Italian, and with QU, GU in French). FilipeS 10:22, 14 October 2006 (UTC)

The short A

I've seen it written in certain dictionaries (Oxford for one) that the correct pronunciation of latin short a, is as the 'u' in english 'cup' rather than the 'a' in 'cap'. Is this correct? 16th Oct 2006

As far as I know, there is no reason to believe that the short 'a' of Latin had a different quality from the long 'a'. FilipeS 18:05, 16 October 2006 (UTC)
Technically there's no exact equivalent but in some dialects of English that is the closest approximation. It's too inexact to be useful. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 18:07, 16 October 2006 (UTC)

Dark L in Latin?

I had never heard of such a thing. The claim should be properly sourced.

Also, which variety of English is it that uses the dark L always, "except when [it] precedes an /i/ or another /l/"? FilipeS 17:33, 29 October 2006 (UTC)

Sounded bunk to me. But according to An Introduction to Vulgar Latin:
It is likely that l before or after another consonant had...[velarization]...Before consonants, this formation led in some regions, sporadically by the fourth century but regularly not until the eighth and ninth and later (Lat. Spr. 476), to the vocalization of l to u... After consonants, this elevation, shifted forwards, brought about the palatalization of l in Spanish and Italian: clavem > kl'ave > Sp. llave, It. chiave.
...intervocalic l, except before i, also had [velarization]]... During the Latin period l regularly remained unchanged.
I don't know if Classical Latin (which is the page's emphasis) had this allophony but Vulgar Latin certainly did. As for varieties of English with l allophony, I think that the editor was oversimplifying the comparison. It could definitely be worded better. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 20:12, 29 October 2006 (UTC)

Now that you mention it, I've remembered this:

A similar process happened in Brazilian Portuguese and in Old French, resulting in [w], whence Modern French sauce as compared with Spanish salsa. (@ lateral consonant)

As you say, though, it could have been a characteristic of Vulgar Latin, rather than Classical Latin. FilipeS 21:12, 29 October 2006 (UTC)

I've looked up the reference in Allen and updated the article accordingly. I think there may be a confusion here between the normal versus velarised pronunciation which Allen attests (based on evidence from ancient writers) and the palatalization which seems to have occurred in later Latin. --rossb 23:01, 29 October 2006 (UTC)
Good work. Would it be better if we said "in the syllable coda" since that's what word-final and pre-consonant positions have in common? Or would that be too vague? Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 01:22, 30 October 2006 (UTC)
I'm still uneasy with this. It would be nice if it could be cross-checked in other sources. FilipeS 12:45, 21 November 2006 (UTC)

Phonetic transcription for QU

What makes the "U" in "quattuor" different from the "U" in "ueni"? FilipeS 12:45, 21 November 2006 (UTC)

The Q. In Latin, QV represented a labialized velar plosive /kʷ/. The V in VENI represents /w/, or /u/ in the syllable onset. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 20:28, 21 November 2006 (UTC)
The Q is just spellling. How do we know that the different spellings QU and CU corresponded to a different pronunciation for the U? FilipeS 23:00, 21 November 2006 (UTC)
No, Q is not just spelling. /kʷ/ is an entirely different phoneme than /k/ and does not equate to /kw/. /kʷ/ is a velar plosive with simultaneous lip rounding. We know this because of over a century of comparative linguistics. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 02:08, 22 November 2006 (UTC)
The article should explain what arguments comparative linguistics has for arguing that Latin QU stood for /kʷ/, rather than /kw/. FilipeS 12:55, 22 November 2006 (UTC)
No more than it should argue that C wasn't palatalized before front vowels or how we know which vowels are long and which are short despite the failure of the Romans to make such a distinction. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 19:07, 22 November 2006 (UTC)

Phonetic transcription in examples

Why are geminate consonants transcribed by doubling them, instead of using the lengthening mark? FilipeS 15:25, 24 November 2006 (UTC)

Are you asking why we do that or why the Romans did that? I imagine we do that because that's how geminate consonants are usually transcribed. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 22:49, 24 November 2006 (UTC)

If I'm asking about the phonetic transcription, then obviously I'm not talking about the Romans. Looking in Wikipedia: Gemination, I see that long consonants are sometimes transcribed by doubling the consonant in IPA:

In the International Phonetic Alphabet, long consonants are normally written using the triangular colon ː, e.g. [penːe], though doubled letters are also used (especially for underlying phonemic forms).

I hadn't realised this, sorry. It seems a little inconsistent to transcribe long consonants by doubling letters, and not do the same with the vowels, but who am I to argue with established tradition? FilipeS 00:19, 25 November 2006 (UTC)

The reason for this is partly that long consonants are believed to have been pronounced with the brief pause between them, so that the sound is thus made twice. In other words, a doubled consonant is also the point of a syllable break in speaking, so the transcription pretty much has to use doubled consonants rather than length marks. --EncycloPetey 08:36, 28 December 2006 (UTC)

No, [kk] and [kː] are simply alternate IPA notations, although long consonants can give the impression of being accompanied by a pause. FilipeS 22:43, 28 December 2006 (UTC)

Latin vowels in IPA

There has been previous dicussion about the inconsistencies in representing the "long" and "short" vowels of Latin in IPA symbols. I note that this article has made no distinction between length (sound quality) and length (duration). I have edited the section previously entitled "Vowel length" to read "Vowel duration" to make the distinction clearer. I think a great deal of the prior confusion comes from the fact that these two issues are often blurred together in textbooks so that their importance is lost. Even the latest edition of Wheelock's Latin does a very poor job of clarifying the difference. Fortunately, a set of audio CDs has been released to accompany the text as read by Mark Robert Miner. I am following a combination of what I hear in his readings and what is recommended in James Morwood's A Latin Grammar (Oxford Univ Press, 1999). Morwood gives the following:

a short, as in English hat
ā long, as in English father
e short, as in English pet
ē long, as in English aim (or, more accurately, French gai)
i short, as in English dip
ī long, as in English deep
o short, as in English pot
ō long, as in English mobile (or more accurately, French beau)
u short, as in English put
ū long, as in English fool

Although the short a description does not match the audio readings or other descriptions I've seen. Note in all cases that a different IPA symbol would be used for each word, and that this has nothing to do with vowel duration, which is represented separately by the pair of little triangles in IPA. --EncycloPetey 08:32, 28 December 2006 (UTC)

Any description of Latin that says the Latin short 'a' was the English 'a' in 'hat' (IPA [æ]), and the Latin long 'e' was the English 'ai' in 'aim' ([IPA [ej]) is dead wrong. The description you've just written above is Latin pronounced with a strong English accent. That's good enough for modern English learners, but not likely the original pronunciation. I will say it again: vowel length in Latin is not the same as vowel 'length' in modern English. Sorry about the inconvenience, but those are the facts. FilipeS 13:40, 28 December 2006 (UTC)
Not dead wrong, just using approximation as you say. I fully agree that vowel length in Latin and English is not the same. Long vowels in English tend to be diphthongs, which is not the case in Latin. Unfortunately, I have yet to see a Latin book that actually describes the vowels according to IPA conventions.
I maintain, however, that the system as currently described in the article is wrong. As I pointed out, I was not using the Oxford book exclusively. I have also used audio recordings by a professional classics reader. I have also relied on other sources that I did not elaborate upon. For instance, Tore Janson (A Natural Hisotry of Latin, p179) says that: "Right from the beginning Latin had five simple vowels: i, e, a, o, and u, each of which could either be lon or short. By long and short we mean just that: an extra degree of phonetic length coupled with a small difference in the way the vowel is articulated, much as in the contrast between bit and beat in English." Janson has explictly indicated both by example and explanation that there is a difference in the way long and short vowels were articulated, not just in terms of duration. The section on vowel pronunciation in Wheelock, coupled with the accompanying audio supports this interpretation as well.
If you have a text that says otherwise, could you please provide the relevant quote(s). I have been working on adding Latin pronunciations on the English Wiktionary for many months now, and if you have evidence that my methods are in error, then I need to know about it. I've looked in the Latin Wiktionary more than once for assistance, but that project has an IPA guide only for General American English, not for Latin, and there are only one or two Latin words given a pronunciation there (using ecclesiatical pronunciation). In any case, it would be nice to cite a source for the IPA pronunciations given in Wikipedia's article for the vowels. --EncycloPetey 20:35, 28 December 2006 (UTC)

Does Janson present any evidence to the effect that there was in Latin "a small difference in the way the vowel is articulated"? Because books on historical Romance or Latin linguistics typically assume no such difference. See the one here, for instance.

Unfortunately, that link doesn't work. The source is restricted to researchers, so I can't check the source.
You can find the book in a library, as I did. FilipeS 22:40, 28 December 2006 (UTC)
If you have acces to the book then why not provide the relevant info, rather than dangling the book out there inaccessibly. The book was published in German in 1956. I have little hope of finding it in any library that won't require a special four-hour trip to another part of the state. I'm loathe to make such a trip since I as yet have no indication that the book will contain useful information. --EncycloPetey 23:57, 28 December 2006 (UTC)
I had quoted the relevant parts above. FilipeS 17:31, 29 December 2006 (UTC)
You didn't quote anyhting at all. There is also no explanation of the sound in Classical Latin in the information you provided; there is merely reference to modern notation with macrons. I assume then that if this is the only part of the book relevant enough to mention, the book must not be worth looking into. --EncycloPetey 20:27, 29 December 2006 (UTC)

Several modern languages have vowel quantity distinctions without any perceptible quality differences. Examples: Hawaiian, Danish, Japanese.

That's fine, but it doesn't answer the question of what was done in Classical Latin. I understand that there is a difference between vowel articulation and duration; the question is what the Romans did, not what the Japanese are doing.
You've conveniently ignored my point about Friulian, below. FilipeS 17:54, 29 December 2006 (UTC)
I haven't ignored anything. Features of a single modern language (which are not backed up by any sources, I note) are hardly string evidence for pronunciation of vowels in a different language some 2000 years prior. --EncycloPetey 20:27, 29 December 2006 (UTC)

In particular, Friulian, one of the few Romance languages which have retained vowel length distinctions, shows no perceptible quality difference between short and long vowels, either.

And, frankly, don't you think it's too much of a coincidence that the Latin vowels would all just happen to coincide with vowels of modern English? FilipeS 20:46, 28 December 2006 (UTC)

But they don't coincide with the vowels of modern English. Modern English uses /æ/ for short a, and I haven't heard that sound in any Latin speaker's voice. Modern English does not even have /e/ as a vowel sound, it only has that sound in the diphthong /eɪ/.
You've heard Classical Latin speakers speaking live? Where?! FilipeS 22:41, 28 December 2006 (UTC)
Don't be dense. Either you're deliberately misreading what I wrote or you're not bright enough to be posting on this page. I will assume the former. --EncycloPetey 23:00, 28 December 2006 (UTC)
Don't be arrogant. I was merely pointing out, with a little irony, that you have no way of "hearing" actual Classical Latin, which has been a dead language for centuries. You will do best to listen to what the people who actually study it have written about it. Modern recordings are just guesses, and, worse, they're often done by people who settle for a lacklustre pronunciation which would stand out like a sore thumb, were there any native speakers still around to hear them. I expect Wikipedia aims higher than that. FilipeS 23:17, 28 December 2006 (UTC)
So what you're really saying is that experts who disagree with you are wrong. The modern recording I have is the accompaniment readings to Wheelock's Latin, one of the more prestigious Latin textbooks. The reader for the CDs is Mark Miner, who "is a professional reader specializing in Latin and Greek literature; his credits include recordings for the popular Athenaze Greek text and numerous performances throughout North America." The producer of the recordings is Richard LaFleur, who "is Franklin Professor of Classics at the University of Georgia and author of numerous books, including the revisions of all three volumes in the "Wheelock's Latin Series."" So do these people qualify as "people who actually study it", and on what basis are you dismissing their work? I have provided refernces in the form of two books and an audio recording. As yet you have provided no evidence in the form of citations of any kind from any person at all. If you want to provide contradictory evidence, then that's fine, but it should be in the form of cited evidence, not hand-waving arguments. --EncycloPetey 23:39, 28 December 2006 (UTC)
Your description disagrees with that of (many) other experts. And I don't give a damn about those people's pedigrees -- what is their evidence? FilipeS 17:31, 29 December 2006 (UTC)
You don't mean "pedigrees" (heritage), you mean "qualifications". And qualifications are precisely what are important in judging the value of an authority.
Which experts? What description do they give? It's ironic that you should be demanding evidence when you yourself have provided none. --EncycloPetey 20:27, 29 December 2006 (UTC)
The word "pedigree" was a figure of speech (and part of a paraphrased quote), but nevermind. Show me the evidence, if there is any. Arguments, observations, anything. Not opinions. FilipeS 23:35, 29 December 2006 (UTC)

It's not your (or my) speculation or opinion I'm seeking. I want solidly published research that explicitly assigns IPA symbols. As I said before, I haven't found a source that does this, though my research has been far from exhaustive. My sincere hope is that some knowledgable Latin linguists will be able to shed light on the situation. --EncycloPetey 22:19, 28 December 2006 (UTC)

I think it's strange that you would like a source with IPA to compell you to change your mind about the way Latin vowels were in Classic Latin when your own source does not include IPA. I believe An Introduction to Vulgar Latin may have a phonetic enough transcription (although not strictly IPA) detailing the shift from Classic to Vulgar although I haven't given it a good enough look to say either way. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 23:36, 28 December 2006 (UTC)
I'm using my knowledge of IPA against recordings. --EncycloPetey 23:39, 28 December 2006 (UTC)
The source you've cited does have some useful info (thanks!). I quote from section 165:
165. Originally, perhaps, long and short vowels were distinguished only by duration, the vowels having, for instance, the same sound in lātus and lătus, in dēbet and rĕdit... However this may have been, long and short e, i, o, and u were eventually differentiated, the short vowels being open while the long were close... That is, for the vowels of brief duration the tongue was not lifted quite so high as for those held longer.
This again supports my view that there is a difference in articulation of long and short vowels in Latin. Unfortunately, there is no time-related information as to when this is supposed to have begun. --EncycloPetey 23:48, 28 December 2006 (UTC)
Ahh, this indicates that we won't really ever be sure about Classic Latin vowels. The best method for determining the phonetic values of CL vowels is through the reconstructive method. Unfortunately, it was raining the last day that my library would be open before January 2nd so I didn't check out the book by Swadesh that may have been able to help illuminate the issue. I for one would rather be cautious about changing the vowels based on non-reconstructive sources (plus, the passage you quote is talking about Vulgar Latin, I've bolded where I see it indicating this). I really don't think that modern linguists making recordings, no matter how prestigious and knowledgeable they are, can be considered viable examples of Classic Latin speech. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 00:27, 29 December 2006 (UTC)
I'd say he's talking about Vulgar Latin there, not Classical Latin, as EncycloPetey had demanded. FilipeS 17:31, 29 December 2006 (UTC)
Probably, but the context is far from precise. We need a source that isn't so vague in its timeline. --EncycloPetey 20:27, 29 December 2006 (UTC)
Why should linguists be able to write authoritatively about sounds, but be unable to reproduce those sounds to any degree of accuracy? Surely even a hack could tell the difference between /ɛ/ and /e/? Miner's readings from the early sections in particular are those specifically designed to demonstrate Latin pronunciation with examples of long and short vowels. LaFleur calls the readings "careful and precise". Surely they can't be so totally incompatible with scholarship as you are saying? Why should a careful pronunciation based on research be any less accurate than a written description of the same sound? --EncycloPetey 00:53, 29 December 2006 (UTC)
The same reason that Arnold Schwarzenegger has lived in the United States for longer than I've been alive yet still speaks with a thick Austrian accent. Also, Miner could be producing Latin from a later period. It could be accurate but the ambiguity casts enough doubt in my mind that we should really be on the lookout for a better source. If FilipeS doesn't mind, I think we can keep the article as is (with your edits) for the time being and then change them back if we find a compelling reason to change it back. I would generally advise it the other way but it may prove to be a lot easier to convert back to the older representation of vowels than to restore your edits. I don't think it'll be too damaging either way. Wikipedia shall survive. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 01:26, 29 December 2006 (UTC)
Schwarzenegger hired a language coach to help him deliberately keep his Austrian accent after he began to lose it. Wheelock's book expressly states that the grammar, spelling, and pronunciation are all geared with a focus on Classical Latin. That said, I do agree that we should keep looking for more and better research to clarify the situation. If I am wrong in my understanding, then I would like to know so that I can change my views. I would also like to be certain that the best scholarship is reflected in the pronunciations of Latin words on Wiktionary, which is where I spend most of my time. --EncycloPetey 05:06, 29 December 2006 (UTC)

I strongly object to these changes. This article is about classical Latin, not Vulgar Latin. And, even in Vulgar Latin, when the pronunciation of short /i/ changed, it became /e/, not /ɪ/. And when short /u/ changed pronunciation it became /o/, not /ʊ/. The pronunciation EncycloPetey put in the article is clearly a clumsy English-accented fake. FilipeS 17:40, 29 December 2006 (UTC)

Personal attacks in the continued absence of any evidence hardly supports your claims. --EncycloPetey 20:27, 29 December 2006 (UTC)
What the? How is that a personal attack? If you take personally his criticism on what you've presented then you seem to identify a little too much with your sources. The Vulgar Latin quote (section 165) is talking about Vulgar Latin; if it's ambiguous out of context, the fact that the whole book is about Vulgar Latin should make it clear.
FelipeS, while you strongly object to the changes, I think you can discuss your objections a little more calmly. The only source you've provided doesn't use IPA and the only relevant quote that you've included from that source states that vowel length was "more clearly defined" than in English. That's true of Hungarian, but that's a far cry from saying that long and short vowels were identical phonetically. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 21:08, 29 December 2006 (UTC)
Fair enough. Might I suggest that EncycloPetey should do the same? I would like to make just three more remarks, for now.
  • First, EncycloPetey shouldn't be so quick to dismiss Lausberg's book without having even bothered to look it up (or any other book which doesn't fit his favorite description of Latin phonology).
  • Second, he shouldn't be so quick to dismiss the counter-example of Friulian. It is a Romance language, after all (unlike English, say). And there are other Romance languages with vowel quantity distinctions that are not followed by vowel quality distinctions, such as Lombard.
  • I'm trying very hard to not be ironic. I know I haven't succeeded entirely, but I want to make an extra effort. EncycloPetey, have you considered the possibility that perhaps -- perhaps -- the authors you are used to reading/hearing, which I'm going to assume are all English speakers, have a certain linguistic bias to interpret sounds in the way which more closely matches their own language (as we all do, no doubt)? In short, have you looked at authors or recordings from other linguistic backgrounds? Because I assure you that none of the descriptions of Latin I have read from actual linguists gave even a hint of a suggestion that (a) vowel length distinctions in classical Latin were provably accompanied by quality distinctions, for all five vowels; or (b) the vowel qualities in question were remarkably analogous to various sounds of the phonology of modern English. With all sincerity and without any irony I invite you to reexamine these assumptions. Regards. FilipeS 22:04, 29 December 2006 (UTC)
Reply to point 1: As both Aeusoes and I have pointed out, the bit you pulled from Lausberg's book says nothing about articulation of vowels in Classical Latin. You did say that you'd quoted the relevant information from the book, so if that's the only part you consider relevant, then the book has no useful bearing on the matter. I am accepting that you have (as you said) quoted the relevant portion. If you feel there is more of relevance in the book, then please provide that information. Otherwise, I have to accept your judgement that the unhelpful information you cited was in fact the most relevant information to be found in the volume.
Reply to point 2: Single languages in isolation or even two languages in close proximity do not establish a historically "deep" pattern. The information you've provided also address only the question of vowel length. No one is arguing that length wasn't an important feature of Classical Latin. The question is whether articulation differences accompanied the differences in length. If you look at the phonology section of the article on Friulian, you'll see that some long vowels are lengthened while others are differently articulated. The differences are regional and vary among the dialects. So, the evidence from Friulian leaves open the question of whether there were articulation differences in Classical Latin. What we need is a study employing the historical linguistic tool of cladistics to reconstruct ancient characteristics of the root language. I personally don't have ready access to a library that might allow me to find such a study, and it may be some time before I can make a trip that will afford me such access.
Reply to point 3: What you are asking me to do is speculate based on evidence I haven't seen. I'm sorry, but logically it only makes sense to have an opinion based on evidence I have experience with, and not on hypothetical evidence that may or may not exist. You have once again alluded to "descriptions of Latin that [you] have read from actual linguists", but have not provided any of those descriptions or named those linguists. Who are they, what precisely have they said, and where did they say it? --EncycloPetey 22:19, 29 December 2006 (UTC)
  • Regarding point 1, I admit that I did not quote enough from the book to settle this disagreement, but Lausberg does describe those vowels some pages earlier in the book. He does not use IPA, but nothing in his description indicates that he thought there was any difference in quality between long and short vowels in classical Latin. I could go to the library and quote some revelant passages, but (a) not until next week, at least, (b) that would probably violate copyright laws; (c) you probably wouldn't understand it, because the version available to me is a translation, not into English. My suggestion is that you look up either Lausberg's book or any other thorough study of classical Latin, which wasn't written by an author of the Anglo-Saxon tradition.
    Try me. Don't assume that Engish-speakers can't read other languages. I have formally studied Spanish, Dutch, and Koine Greek. I can read Spanish, Italian, French, and Romanian with reasonable ease, and can get through reading Dutch with a dictionary and patience. I can get by somewhat in German, but don't have a large enough vocabulary to not require frequently consulting the dictionary. If you want me to look in a book not by someone in the Anglo-Saxon tradition, do you have particular works in mind? --EncycloPetey 23:08, 29 December 2006 (UTC)
Lausberg's original edition is in German, and the translation I looked into is in Portuguese. But I expect there should be an English translation of his book, or at least a Spanish one. And, anyway, it doesn't have to be Lausberg. Similar books on the same topic should say the same. FilipeS 23:31, 29 December 2006 (UTC)
  • Regarding point 2, I must disagree again. What the page on Friulian says is that in some dialects long vowels are replaced with diphthongs. That's not the same as a difference of quality, in the context of our discussion (none of the IPA symbols you wrote above is for a diphthong), though your favorite Latin recordings probably use [ej] instead of [e:], because most native speakers of English have a hard time producing a pure [e:], and don't even realise it. The evidence provided by the living example of Friulian -- and Lombard, please! (and also some conservative dialects of French) -- is certainly not definite, but it still goes against the description you cited above, which claimed there was always a noticeable quality difference between short and long vowels in classical Latin. It's not definite proof, but it's one strike against that theory -- and where is the evidence for it?...
  • And regarding point 3, all I am saying is that you should keep an open mind, and double check with independent sources, rather than assuming the ones you are acquainted with are 100% accurate, just because they're the most accessible to you. FilipeS 22:46, 29 December 2006 (UTC)

There seems to be little consensus here. Since no one seems to have adressed what Allen, who primarily deals with "the educated pronunciation of Rome in the Golden Age", has to say on the matter in his Vox Latina, I'll quote the relevant passage: "There appears to have been no great difference in quality between long and short a, but in the case of the close and mid vowels (i and u, e and o) the long appear to have been appreciably closer than the short." He then goes on to the historical development, quotations from various authors (from around 2nd centuary AD, as well as evidence from older inscriptions where "e" stands for normally short "i", and "i" for long "e", etc. The way I interpret his conclusion, the short vowels are likely to have the qualities similar to /ɪ/, /ɛ/, /ʊ/, /ɔ/. (Sadly he does not use IPA.)

Now, I'm not saying that Allen should be regarded as infallible, but I would very much like to see possible refutations of his arguments. Sources that simply make no assumption of any quality distinction between long and short vowels in classical Latin are not really satisfactory unless that position is based on qualified arguments -- my point being that I see no reason for either position to be more valid a priori.

So, I would propose that the article (specifically the last example, "arma virumque...") gets updated to reflect Allen's opinion. That is, unless somone can quote a source that specify that the vowel qualities are indeed likely to be the same, and which base this opinion on some kind of arguments and historical evidence, in which case the article should be updated to specifically state that there is little scholarly consensus. Unless we can quote a source which refutes Allen's arguments, or which makes statements to that affect, I at least think that his conclusion should be included in the article.

A small revertion war came to place when I tried to make such an update, so I would like to call for some other opinions.

And, in any case, the article should be updated at least in some way, because as it is now, it is internally inconsistent, proposing one pronunciation in the table at the start and another in the final example. Alatius 17:42, 20 May 2007 (UTC)

It seems I said something which was wrong in my last edit summary, for which I apologize (that there were no open-mid vowels in classical Latin).
My problem is with the use of the symbols /ɪ/ and /ʊ/, which normally denote vowels that are not only more open than /i/ and /u/, but also more central (as in English). We can clarify this ambiguity by using /ɪ/ and /ʊ/ with the fronting diacritic and the retracting diacritic, or equivalently (my preferred solution) by using /i/ and /u/ with the lowering diacritic.
Specifically, I would prefer the following transcription:
[ˈaɾma wi̞ˈɾu̞mkʷɛ ˈkanoː ˈtɾɔjjai̞ kʷiː ˈpɾiːmu̞s ab ˈoːɾiːs
iːˈtali̞ãː ˈfaːtoː ˈpɾɔfu̞gu̞s, laːˈwiːnakʷɛ ˈweːni̞t
ˈliːtɔɾa mu̞lt i̞ll ɛt ˈtɛriːs jakˈtaːtu̞s ɛt ˈaltoː
wiː ˈsu̞pɛɾũː ˈsai̞wai̞ ˈmɛmɔɾẽː juːˈnoːni̞s ɔb ˈiːɾãː]
The difference between the long and the short vowels must not have been very great anyway, at least not in all periods, or the Romans would have devised different letters for each in their alphabet. Regards. FilipeS 18:15, 20 May 2007 (UTC)
Everything you say might hold. I have found only a couple of potential sources of good information since I last commented. The best of these is still not so helpful since it is a German translation of an English work as is rather old. I do have to disagree with your argument that "The difference between the long and the short vowels must not have been very great anyway, at least not in all periods, or the Romans would have devised different letters for each in their alphabet." The Romans never developed different letters for consonantal I or U, though scholars agree that those forms existed. There are plenty of languages that may serve as counterexample; it is not terribly unusual for a sungle symbol in a language to carry more than one pronunciation. This does not mean that your conclusion is incorrect, only that the argument you are using to support it is insufficient for the task. --EncycloPetey 19:34, 20 May 2007 (UTC)
Well vocalic and consonantal i/u were allophones, but you nevertheless have a point, considering that they also used the same five symbols for the long and the short vowels, even though they were ten different phonemes. FilipeS 19:45, 20 May 2007 (UTC)
In a diagram, Allen actually places short i and u within the /i:/-/a:/-/u:/ triangle, hence, I deduced, implying a more centered pronunciation as well as more closed, but whether this is a conscious decision of his, or rather based simply on a presumption that the laxness, so to speak, of the vowel automatically implied a centering as well, I do not know. Anyway, I quite like your suggestion to use the lowering diacritic, primarily because it seems inherently vague, and thus very well suited to our need. For, it seems we can agree, the lowering would have been gradual, and the lowering diacritical does not imply that we know exactly how far it had gone in the period we aim at describing.
I see you favour [ɾ] rather than [r]. While I don't really have a problem with that, I wonder though if not [rr], rather than a single [r] would be a better transcription of "rr", based on how the syllable division would fall between the r:s, and the length of the sound needed to make the syllable heavy in the verse.
As for the transcription of "ae", the idea seems to be that the change in spelling from earlier "ai" also reflects a narrowing of the last element. I would guess anyway that the resulting sound would be closer than that of short i, to warrant the change in spelling. Hence I am a bit wary of using the same transcription in both cases. On the other hand, I have no idea whether [ae] would be too bold. What about [ae̝]?
I removed the lowering diacritical in "Italia", based on a passage in Allen (Vox Latina p. 51 ff): "The Latin short i also may well have had a closer quality (more like that of the long ī) before vowels, to judge from the Romance development of Latin dies...".
Currently I would propose something like this (note that I also have added [ɫ] and [t̪], as well as introduced an [w] as the result of the elision in "multum"):
[ˈaɾma wi̞ˈɾu̞mkʷɛ ˈkanoː ˈt̪ɾɔjjae̝ kʷiː ˈpɾiːmu̞s ab ˈoːɾiːs
iːˈt̪aliãː ˈfaːt̪oː ˈpɾɔfu̞gu̞s, laːˈwiːnakʷɛ ˈweːni̞t̪
ˈliːt̪ɔɾa mu̞ɫt̪w i̞ɫl ɛt̪ ˈt̪ɛrriːs jakˈt̪aːt̪u̞s ɛt ˈaɫt̪oː
wiː ˈsu̞pɛɾũː ˈsae̝wae̝ ˈmɛmɔɾẽː juːˈnoːni̞s ɔb ˈiːɾãː]
By the way, what do you think of removing word spaces? Alatius 22:20, 20 May 2007 (UTC)
I wouldn't remove the word spaces simply so that it's easier to see how each word is represented. While we're discussing this, I'd like to bring up /slashes/ vs [brackets]. If we determine that any short vowel is phonetically different from its long counterpart, I think that we should still transcribe them phonemically the same. So it might be [iː] and [ɪ] but it's still /iː/ and /i/. This is not how the article currently represents it. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 01:55, 21 May 2007 (UTC)
Hello again. Interesting comments.
My very personal preference is to transcribe qualitative distinctions whenever they are phonemic, but to avoid making them when they are allophonic (except in special cases, such as when discussing dialects). In light of this, and given the sources we've all found, I favour marking the distinct quality of short E, I, O, and U (preferably as /ɛ/, /i̞/, /ɔ/, /u̞/).
However, I recognize that Latin is special. People are naturally curious about the details of its historical pronunciations (as far as they can be reconstructed). So I do not object to the more narrow transcriptions proposed here.
I'm a bit surprised, though, at Aeusoes1's transcription of the <LL> as [ɫl], with the first L velarized and the second not velarized!
I must also confess that I am not qualified to follow Alatius's argument based on verse metrics regarding /r/. Since <RR> and <R> were most likely pronounced as a trill and a flap, respectively, I would tend to favour the transcriptions /r/ and /ɾ/ for each.
As for the diphthongs <AE> and <OE>, I think these are going to be problematic. We're going to have to define "classical Latin" narrowly. According to what I remember reading, these diphthongs had probably already begun to be pronounced as /ɛ/ and /e/ by the early imperial period (1st - 2nd centuries), which most people would not hesitate to include in classical Antiquity... Another option might be to use one of the pronunciations, but add a note that the other was also used in other periods of Antiquity.
Alatius's suggestion that the unexpected spelling of these diphthongs (pronounced /ai/ and /oi/ in pre-imperial times, as far as I've read) may have been due to the qualitative difference between short and long is intriguing. I did not know that they had been spelled <AI> and <OI> in older Latin. But couldn't the change in spelling be equally well explained by a /i/ → /i̞/ sound change? FilipeS 19:31, 21 May 2007 (UTC)

To be honest, I've been warming up to the idea of using /ɪ/ and /ʊ/ for short I and short U. IPA symbols do not represent rigid positions in the vowel chart; they stand for a broad range. See for example the wide variation in their phonetic values in the dialects of Irish. The /ʊ/ of Connacht Irish is even a back vowel! Right now, I would be favorable to using these symbols consistently throughout this article, as well as at Vulgar Latin, provided that the article does not claim or imply that the short vowels of Latin were the same as their counterparts in modern English. FilipeS 20:56, 16 July 2007 (UTC)

And I continue to look for scholarship on the subject. I've found a German source (which turned out to be a translation from an English writer, unfortunately), and I have a Hungarian source (but the author seems to be an obscure one). We might end up expanding the discussion of vowels a bit to explain the marked difference in viewpoint between English and continental European scholars, if we can find enough of them to demonstarte a consistent difference. --EncycloPetey 21:10, 16 July 2007 (UTC)

An old source

I think I found a source, albeit a fairly old one, that gives a detailed account of vowle pronunciation "at the end of the Roman Republic" or the first century B.C.: Kent, Roland G. "The Sounds of Latin. A Descriptive and Historical Phonology" Language, Vol. 8, No. 3, Language (Sep., 1932), pp. 11-13+15-216. Here are some quotes:

page 45
SHORT AND LONG E. The short e was open in quality, and the long e was close, as in Eng. met and meet, or, more exactly, as in Fr. pièce and donné. This is shown by the statements of the Roman grammarians, and the manner in which the sounds developed into the Romance dialects. For both ĕ and the monophthong developing from the earlier diphthong ae produced in Romance an open e or its representative, but ē and ĭ gave a close e or its product in Romance:Lat. pĕdem gave Fr. pied, It. pid, Sp. pig, and Lat, caelum gave Fr. ciel, It. Sp. cielo; but Lat. sēta and bĭbit gave Fr. soie and boit, It. seta and beve, Sp. seda and bebe.
SHORTAND LONG 0. Like the sound just discussed, o was open when short and close when long, somewhat as in Eng. not (in the British pronunciation) and boat (if without the diphthongal vanish; strictly speaking, as in Gm. Boot). The essential evidence is the testimony of the Roman grammarians and the development of the sounds into the Romance dialects, where ŏ gives an open o, as in It. ruota and molle from Lat, rŏta and mŏllem, but ō and ŭ alike give a close o, as in It. sole and condotto from Lat. sōlem and condŭctum.
Short a and long a, unlike the other Latin vowels, seem to have been alike in quality and to have differed only in quantity.
page 46
SHORT AND LONG I. Short i was an open i, as in English pin, and long i was a close i, as in English machine. Such is the testimony of the ancient grammarians and of the Romance dialects.
Except that none of the extant Romance "dialects" has the Germanic vowel [ɪ]. FilipeS 18:48, 14 November 2007 (UTC)
SHORT AND LONG U. That u, like e, i, and o, had two sounds, an open sound when it was short and a close sound when it was long, appears from the testimony of the grammarians, scanty though it is, and from the development into Romance: Lat. condŭctum, sŭprā > It. condotto, sovra; but Lat. condūxī, condūcō > It. condussi, conduco. The two sounds were accordingly those heard in Eng. foot and boot. A Plautine pun on the pronoun and the cry of an owl shows that a was ū pure and not , as it is in modern English.

While it's old, it does attempt to give some evidence for its reasoning. To be thorough, Kent states on page 31 that "Among Roman grammarians may be listed those whose works are collected in the Grammatici Latini, edited by H. Keil, 7 vols. (Leipzig, 1855-80); the Corpus Glossariorum Latinorum, edited by G. Loewe and G. Goetz, 7 vols. (Leipzig, 1888-1901); and large portions of the works of Varro, Quintilian, Gellius, Festus, Nonius Marcellus, Martianus Capella, and Macrobius. Valuable passages occur also in the writings of Ennius, Accius, Lucilius, Cicero, Suetonius, Apuleius, etc." Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 06:36, 14 November 2007 (UTC)

He seems faaaar too quick to jump to the conclusion that the Latin short "i" and the Latin short "u" were just like their modern English counterparts, without any particularly compelling evidence. Again, no Romance language has the Germanic vowels [ɪ] and [ʊ].
I think we have a serious problem here, which is the habit that Anglophone authors have of (a) conflating vowel length with vowel quality, and (b) using different phonetic symbols for stressed and unstressed vowels. FilipeS 18:48, 14 November 2007 (UTC)
I think the next step here is to look at what the "Roman grammarians" said. I don't think we can say that the evidence is not compelling if we haven't seen it all. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 07:34, 17 November 2007 (UTC)
I was just commenting on the evidence which has been presented on this page. I don't have easy access to bibliography about this, unfortunately. FilipeS (talk) 15:49, 19 November 2007 (UTC)