Help talk:IPA/French/Archive 1

Rhotic (ʁ)
Reece Llwyd has added this note. I don't doubt the information's accuracy, but considering that this page is designed to assist editors in transcribing French for Wikipedia articles, I think we should ignore the dialectal variations in the rhotic's pronunciation and use one for all of our transcriptions. Thoughts?— Æµ§œš¹  [aɪm ˈfɻɛ̃ⁿdˡi] 21:26, 9 August 2008 (UTC)
 * I agree. These "IPA for X" help pages (why were they moved out of Help: space?) are quick-and-easy guides to IPA transcription, not full-fledged discussions of languages' comparative phonology. —Angr 21:17, 11 August 2008 (UTC)

Nearest english equivalent
Is "German ‹r›" (how is this pronounced?) a commonly used term? I cannot find any mention of the term (or of "German R") on Wikipedia; the footnote points to the "Guttural R" article, which refers to it alternately only as a "French R". If there are no English words or loanwords that use this sound, perhaps the box should just read and point to rhotic or French R.--108.13.54.97 (talk) 02:12, 21 July 2010 (UTC)

Roundeds
I don't understand how the "rounded" end, way and see reflect their French counterparts, surely we can do better.


 * The ø page says that the closest English equivalent to a close-mid front rounded vowel is the open-mid central unrounded vowel, ɜ. I agree. (I can't imagine what dialect makes "way" sound like that!)


 * Open-mid front rounded vowel is more of a problem because ø and œ sound identical to me. But as this is just a cheat sheet, we could probably say that ɜ is a decent English approximation for either of these.


 * u and y could be better approximated in English with book and food. (No, it's not very accurate, but it conveys the difference slightly better, I think.)

And using vin blanc as an example just begs the question! But I don't have any particularly good alternatives. Nick (talk) 17:50, 23 July 2009 (UTC)


 * The problem with [ɜ] is that it doesn't work for rhotic dialects. But French [ø] is further back than [e], so it's not very precise to say it's a rounded [e] either. kwami (talk) 07:04, 25 July 2009 (UTC)

Schwa?
I know it's customary to transcribe e muet as a schwa. But that does give English speakers a bad accent, since e muet is rounded, and English schwa is not. Granted, [ə] is not defined for roundedness, so it can go either way, but mightn't it be better to use [ɵ]? kwami (talk) 07:04, 25 July 2009 (UTC)
 * I think that would be confusing. The rounding is barely noticeable at normal speech rates, and when it is emphasized, it's more like [œ] anyway. +Angr 08:49, 25 July 2009 (UTC)
 * And in any case, standard French practice is always to use [ə]. --seberle (talk) 17:45, 23 October 2009 (UTC)

Long vowels
Many French vowels can be long, but the chart only mentions ɛ as possibly being long. Why is ɛ singled out for mention and why are other long vowels ignored? The article should be consistent and include all or none. --seberle (talk) 17:39, 23 October 2009 (UTC)
 * Do the others have a phonemic distinction between short and long, as in faite vs. fête? +Angr 19:30, 23 October 2009 (UTC)
 * According to French phonology, a long contrast with short  is the only phonemic contrast between two vowels distinct only by quantity.  Other instances of long vowels are contextual allophones.  Not all dialects have  even if  shows up as an allophone of, which may be the source of seberle's confusion.  — Æµ§œš¹  [aɪm ˈfɻɛ̃ⁿdˡi]  23:30, 23 October 2009 (UTC)
 * Yes, that would explain it. Thanks. Now that I think about it, I do remember once reading about a possible distinction between maître and mettre, but I've never heard it. I suspect it is very rare in contemporary French. --seberle (talk) 00:48, 24 October 2009 (UTC)
 * I think it's effectively obsolete in Parisian, but it's still taught. kwami (talk) 02:12, 24 October 2009 (UTC)

The long /ɛː/ hasn't been used in standard French for about a century. It definitely doesn't belong in this chart, as the only dialect to still use it is Québec French, and perhaps a handful of metropolitan French patois. And it is *not* taught (maybe it is to foreigners whose French teacher uses a 19th century dictionary with ancient phonetics, but nobody knows about that sound in France for standard French). It is incorrect to say that "many French vowels can be long", as it is only the case for /ɑ:/ -- which is, as the note says, often replaced by /a/. It's a fairly recent evolution, but 99% people born after the 1960's replaced by as /a/. Same goes for the distinction between /ə/ and /œ/. As much as you still find /ə/ in current French dictionaries, it is more and more replaced by /œ/. Nobody under 30 knows there's supposed to be a difference. —Preceding unsigned comment added by 88.168.89.30 (talk) 07:54, 2 January 2010 (UTC)


 * I totally agree with my fellow countryman above about the long /ɛː/. It does not exist anymore, and there is no reason to add it here as an example of regional dialect (Québécois), otherwise you would have to add ALL the idiosyncrasies of ALL regional vairants of French. And if the person who keeps putting this archaic sound back on is still not convinced: just look at the "French phonetics" page in French and you'll see it's not on there. It's probably best to leave this pronounciation page up to native speakers, isn't it? —Preceding unsigned comment added by 194.51.215.62 (talk) 08:24, 22 February 2010 (UTC)
 * Maybe we should include "idiosyncrecies." We already have the nasalized mid rounded vowel and both open vowels.  Are there any others that we're missing? — Æµ§œš¹  [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi]  20:23, 22 February 2010 (UTC)


 * I have to disagree with those who have dismissed as belonging to a "regional" dialect. Quebec French is not any more a regional dialect than is French French.
 * I would like to correct a particular statement above that this phoneme hasn't been present in "Standard French" (by which the writer means Standard Parisian French) in over a century. The truth is, as late as the 1960s, the vs.  distinction was recorded in dictionaries such as Le Petit Robert. Also, those speakers in France who maintain the distinction include many speakers of what can only be described as Standard French, not patois as that poster disparagingly refers to it.
 * Regarding Canada and France, the situation is analogous to that of British and Australian English. One variety does have greater international visibility than the other, but neither is more "regional" or less standard than the other.
 * There are no Quebec-born native speakers of French who choose to speak with a Parisian accent. Just type "Standard Quebec French", "Standard Canadian French", "Standard European French" and "Standard Parisian French" into Google Books, and you will find many references. For example, this book on French phonology explicitly opposes "Standard Canadian French" and "Standard Parisian French" and says that Standard Parisian French can be regarded as one regional standard among many.
 * As an international project, it is important that Wikipedia not reflect a bias in favour of France. Referring to the French Wikipedia as an authority on these matters, as suggested above, would not be a good idea. The French Wikipedia is not immune from a bias in favour of France, which is where the majority of its contributors are from. 82.124.231.13 (talk) 09:21, 14 May 2010 (UTC)

Pronouncing õ
So I'm trying to learn how to pronounce Poisson, as in the Poisson distribution. The intro says it's pronounced [pwasõ], but there's no õ on this page. Is that just a stress symbol, and if so, why wouldn't it be the normal type? II | (t - c) 08:41, 7 December 2009 (UTC)
 * Somebody got lazy, I guess. I've fixed the article accordingly.  — Æµ§œš¹  [aɪm ˈfɻɛ̃ⁿdˡi]  09:02, 7 December 2009 (UTC)

/õ/ is technically wrong, it should be written /ɔ̃/, like in the chart. Then again, in modern French the sound is actually much closer to /õ/ (which is exactly how I do it, actually). To pronounce it, just nasalize a /o/ sound. —Preceding unsigned comment added by 88.168.89.30 (talk) 08:01, 2 January 2010 (UTC)

"Marginal consonants"?
I don't understand this section. As much as the /ŋ/ makes sense, why the others? They're only used for English words but French doesn't release those as affricates, and for us they are just successions of 2 consonants, nothing else. In English dictionaries they are separate entries in the phonetic section, but not in French. They should be removed from this page, otherwise one should start putting in all the common consonant successions native to French, like "tr", "pr",... which is of no interest since the individual consonants are already all in the charts. —Preceding unsigned comment added by 88.168.89.30 (talk) 08:13, 2 January 2010 (UTC)
 * That's a good point. If they're just sequences of a stop and fricative, and they're indicated orthographically, can we expect both editors and readers to get it right without listing them here? — Æµ§œš¹  [aɪm ˈfɻɛ̃ⁿdˡi]  21:47, 2 January 2010 (UTC)


 * I think 88 is right. Take for instance the Trésor de la Langue Française Informatisé. In their word search form they also have a phonetic input section (method 3 here, although I'm not sure the link works for you as it does for me), which allows entering a word using the sounds it's made of, by clicking buttons. There is a "NG" button there for the sound [ŋ], but none for our other marginal consonants. This is a clear indication that the authors of that dictionary considered those other consonants as simple sequences.
 * Also, the pronunciation given for words like tsunami, tzar, djebel, tchao in the same dictionary does not suggest that those consonants are to be understood in any way other than as sequences.
 * I would suggest removing those 4 consonants from our table, or at most mentioning them in a note. — Adi  Japan  06:53, 3 January 2010 (UTC)

stress
What's the policy on stress marks? Some articles have them (e.g. Arthur Rimbaud), some don't (Paris), and Nicolas Sarkozy has a note saying "Please do not add a stress mark: French is not stressed in this way".

(I hardly ever use this template myself, but it's a topic of general importance.) Lfh (talk) 20:49, 5 January 2010 (UTC)
 * French doesn't have lexical stress, so we shouldn't be marking stress on French words. +Angr 21:08, 5 January 2010 (UTC)
 * OK. Maybe there should be a note to clarify this somewhere, because quite a few articles do feature the stress mark. Lfh (talk) 21:59, 5 January 2010 (UTC)


 * There are two legitimate POV's here: true, French does not have phonemic stress. But it does have phonetic stress, and this is a phonetic transcription, not a phonemic one; readers approaches a pronunciation guide under the assumption that it will tell them how to pronounce the word. However, as Angr points out, the stress is not lexical, and it is words that we are transcribing, so adding stress implies that the stress is lexical. At least, that's the idea that people come away with: it's a common myth that French has word-final stress, and by marking coincidentally word-final stress, we reinforce that misunderstanding.
 * We have essentially the same problem in Engish: there is no lexical difference in English between primary and secondary stress. Nonetheless, we mark that prosodic difference, even though our English transcriptions are allegedly phonemic, because it's standard to do so in dictionaries. At least marking French stress isn't factually wrong, just misleading. kwami (talk) 02:42, 6 January 2010 (UTC)


 * I have yet to see a French dictionary (for natives or for non-natives, it doesn't matter) that marks the stress position in its phonetic transcriptions. So, verifiability-wise speaking, neither we should be marking the stress. I believe it wouldn't help readers pronounce more correctly anyway. On the other hand, we should give a short explanation about the stress in French on this guide page, for those who wonder why there is no stress marker in our French IPA's. — Adi  Japan  04:07, 6 January 2010 (UTC)


 * Yes, I've deleted stress from French transcriptions myself. I'm just acknowledging that those who advocate adding stress marks have a point, though if we go that route, we should be careful that we only stress phrases, not every word/name. kwami (talk) 05:43, 6 January 2010 (UTC)
 * I agree with not transcribing stress, but perhaps we should put something that indicates how stress works in French. Something like "Word stress is not distinctive in French.  In general, only the last syllable of the last word in a phonological phrase is stressed, unless it contains schwa, in which case stress falls on the penultimate syllable." — Æµ§œš¹  [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi]  08:17, 6 January 2010 (UTC)
 * We're not transcribing word-final e muet, though, are we? Être is, not , so there won't be any final schwas to worry about. +Angr 08:42, 6 January 2010 (UTC)
 * Aeusoes, there is at least one exception I can think of to your rule.
 * Les enclitiques ont un comportement accentuel comparable à celui des suffixes. L’accent final est en effet réalisé sur ces deux types de morphèmes. « Regarde-le » donne « Regarde-LE ». Source: 82.124.231.13 (talk) 12:18, 14 May 2010 (UTC)

Actually, Angr, there are a few articles which do have final unstressed schwa. I didn't tag them, but that's something we can delete in a future pass w AWB. — kwami (talk) 21:21, 14 May 2010 (UTC)

Links
Linking to a phone's article makes sense, but we shouldn't link to the article on vowel length or nasalization. Also, because the rhotic varies from dialect to dialect, we shouldn't link to an article for that, either. If we want these sounds to be explained to people, we can explain them in this article. — Æµ§œš¹  [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 20:21, 22 February 2010 (UTC)
 * Thanks.
 * "we shouldn't link to the article on vowel length or nasalization" - Why not? I can't see any harm. We could at least link the IPA symbols for nasalization and length, or "nasal" in the description.
 * "the rhotic varies from dialect to dialect" - Fair enough, but then we have to make it clearer in the article. I propose we change the note to say something like, "sometimes transcribed as [ʁ], sometimes as [ʀ]". Adding specific regions where this is the case would be best of course.
 * Finally, what about /a/ and /ə/?
 * Thank you 205.228.108.185 (talk) 00:30, 23 February 2010 (UTC)
 * If the meaning behind the links is "this is a page that talks about this specific sound" then linking to nasalization or vowel length varies with this. If a reader is to expect that meaning from each link, they won't necessarily know which links go to phone-specific articles and which go to general articles.  You're right that, in the description, we can use the links; I'm only in opposition to linking the specific characters.
 * You are right that we can modify the description for the rhotic. I don't know the specific distribution of which kinds of rhotics.
 * I took out the link to open front unrounded vowel because that can be front or it can be central but I remember now that that article talks about both as if they're the same, so it is okay to link to that article. However, French  is a vowel that has variable pronunciation depending on speaker ([{French phonology]] talks about this a bit).  — Æµ§œš¹  [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi]  02:07, 23 February 2010 (UTC)
 * Thanks, integrated, see how you like it.
 * "French is a vowel that has variable pronunciation depending on speaker" - it would be nice to be more specific and link to the phones.
 * Thanks 205.228.108.185 (talk) 04:29, 23 February 2010 (UTC)
 * I think I've clarified it a bit now. Part of the problem with French schwa is that it's unclear (at least from what I've seen) what exactly its phonetic nature is.  — Æµ§œš¹  [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi]  09:15, 23 February 2010 (UTC)

Brackets vs Slashes
A conversation about the representation of French words is taking place at Template talk:IPA-fr for anybody interested. — Æµ§œš¹  [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 23:57, 3 March 2010 (UTC)

Liaison, syllable breaks
This page should include syllable breaks ‹.› and liaison ‹‿›, shouldn't it? I'm not sure how to concisely describe liaison in a box, though... Also, should we encourage the use of syllable breaks like at French Wiktionary, or only use them when truly necessary, like with English IPA on WP? — ˈzɪzɨvə (talk) 06:12, 8 March 2010 (UTC)
 * I hesitate to incorporate syllable breaks if we can do without them, so we shouldn't use them everywhere. Do you have examples of where this might cause confusion?
 * Liaison can be used with the liaison template, which doesn't get a lot of use currently. — Æµ§œš¹  [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi]  09:20, 8 March 2010 (UTC)
 * I've only seen a syllable break used for diaeresis, as in Haiti, and it's useful there. Words with liaison have been resyllabified, as in en plein air, which as the phonetic realization is IMO probably best for a phonetic transcription. — kwami (talk) 06:04, 4 May 2010 (UTC)

AWB maintenance
I'm reviewing all transclusions w AWB. Not looking for much: illegal characters, more than one stress mark per prosodic unit [hasn't been working], N/_C, and nasal V/_N. Mostly I've found uvular trills; only a couple stress, length, & nasalization problems (such as <ã>).

A few you might want to check: [œ] to schwa in Degas; a [ɔ] left in a "Paul" somewhere (I can't find it now), Bujumbura (nasal vowel?), Niamey (palatal nasal?), Molenbeek-Saint-Jean (/ənb/?) & Neder-Over-Heembeek, Bajocasse ([j]?), Citroën Traction Avant (no liaison?), Staffelfelden & Illkirch-Graffenstaden (not French?), Mont des Arts ('des'), Marolles ('les'), Brunstatt ('un'), Kakabeka Falls (not French?), linguistic issues concerning the euro (cents: not a nasal V, [sɛnt], as if "cennete"?), Jean-Honoré Fragonard (liaison?), Drakaina (model) (just odd).

Let me know if there's s.t. else I should be on the lookout for. — kwami (talk) 06:08, 4 May 2010 (UTC)


 * Being prosodic, not lexical, I don't think we should be marking stress in French words at all. You're probably looking for these, anyway, but here are some (orthography-influenced) things that might be problematic:
 * for
 * word-final schwa
 * for
 * for
 * for
 * confusion of and
 * for.
 * In most of those cases I'm not sure AWB is particularly helpful, but I figured I'd chime in with some possible problems. — ˈzɪzɨvə (talk) 06:51, 4 May 2010 (UTC)
 * Yeah, thanks. Those would mostly need to be done manually. I've mostly been looking at phonotactically illegal combos, as I'm not good with personal and place names. For example, Bajocasse is transcribed with a [j]. Is that an error for [ʒ], or an irregular spelling? I have no idea.
 * Apart from transcription conventions, such as "[r]", most of the transcriptions are quite good, unlike say Spanish. Presumably that's due to the lack of much allophony in French.
 * As for the stress, people have disagreed on that, so I've left it alone, unless it's wrong prosodically. — kwami (talk) 07:57, 4 May 2010 (UTC)
 * I would agree with not marking stress on French words. I always take stress marks out of French words where I see them, and this page notably does not include stress marks among the symbols used in the transcription of French. As for Bajocasses, I have no idea whether it's pronounced with /j/ or /ʒ/, but I suspect Bayeusains has /z/ rather than /s/. I can believe there's no liaison in Traction Avant. Liaison doesn't happen between any word ending in a nasal vowel and any word starting with a vowel; the two words have to be in a certain syntactic relationship to each other. +Angr 10:17, 4 May 2010 (UTC)
 * Somehow I thought that nasal vowels triggered an epenthetic /n/ independently of liaison. Don't know what I was thinking of.
 * Added a few more oddities. — kwami (talk) 17:33, 4 May 2010 (UTC)
 * We could always add stress with a note that it's prosodic. I think people have a legit point that an English speaker is going to want to put stress somewhere, and the least we can do is tell them where it would be appropriate prosodically.
 * Stuff like /s/ vs /z/ I haven't even been looking for. Some langs (i.e. Portuguese) have a lot of transcription variants, and that's what I wanted to clean up first.
 * I've also left a few transciptions that were marked as Quebecois, as at French heraldry. — kwami (talk) 16:23, 4 May 2010 (UTC)

Arpitan
Just curious: Why are we including a guide to transcription on Arpitan, which is apparently an obscure name for Franco-Provençal? I can't imagine this being of anything but extremely limited use to only a few people anywhere. For comparison, we don't include a guide to Maltese pronunciation at IPA for Arabic, nor one for Ligurian at IPA for Italian, so why here? AlexanderKaras (talk) 23:19, 18 May 2010 (UTC)
 * Yeah, I wasn't a big fan of the change, but I declined to raise any objections at the time. It would seem more appropriate to split Arpitan/Franco-Provençal into its own transcription guide.  Even if it's kept here, the vowel section needs some work.  But eventually we're going to have to decide if we're willing to have one of these guides for every language (or phonologically-twinned cluster of languages).  We're now getting into languages with mere hundreds of thousands of speakers (though many languages with tens [even hundreds] of millions of speakers still lack such a guide), and I don't think we're prepared for 6000 IPA guides.  — ˈzɪzɨvə (talk) 01:12, 19 May 2010 (UTC)
 * Split? There's only one page that links here. Most of the Occitan transcriptions still link to the Catalan key. We could just link to the Occitan phonology section.
 * S#!+, I think I was supposed to fix that. Oops! — Æµ§œš¹  [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi]  03:45, 19 May 2010 (UTC)
 * BTW, I want to throw an idea out there: why not use the language code for every case where we're transcribing a specific language, no matter how obscure? "IPA-all" would then only be used for generic statements such as "[ɣ] is a velar fricative". Most of these IPA-xx templates would be simple redirects to IPA-all. However, there would be a couple benefits to this: (1) we could monitor how much a language is being transcribed, and create specific keys for the more popular and difficult ones (Thai is far more common than Arpitan!); (2) because we could link through the redirect, we could verify that all articles use a common system, even if there isn't a key for them; and (3) if we do create a key, the articles will be instantly linked, without us having to slog through the thousands of IPA-all and pron templates to try to find them all. Then it would just be a matter of going through and cleaning up which lede with "X pronunciation", "pronounced", "IPA", etc. — kwami (talk) 01:22, 19 May 2010 (UTC)
 * That sounds like a good idea. If you haven't already, you should propose it at Template talk:IPA-all.  — Æµ§œš¹  [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi]  03:45, 19 May 2010 (UTC)
 * I concur, that sounds like a great idea. What are your thoughts on what to do about Arpitan?  It sure doesn't seem close enough to Occitan phonologically to link to WP:IPA for Occitan, if that's what you meant, Kwami.  But then again my knowledge of it consists mostly of what's been shoehorned into WP:IPA for French... (And I've had WP:IPA for Thai on my radar for some time, when I'm up to a challenge.) — ˈzɪzɨvə (talk) 04:14, 19 May 2010 (UTC)
 * We could copy this over. I can't imagine there'd be any opposition. The only thing would be to tag the redirects somehow so that bots don't go about "correcting" them. Rather annoying when you have a specific link that will be of use some day, and a bot changes it to the dab page it's redirected to. Preventing that would be a handy trick to have.
 * No thoughts on Arpitan. Don't know enough to have a thought. Does seem rather bizarre to have to look French up under "A", though. — kwami (talk) 09:16, 19 May 2010 (UTC)
 * Do bots ever do that? If so, R with possibilities ought to be enough to let them know to leave them alone. Yet, in my experience it's human editors who don't know about (or don't understand) WP:NOTBROKEN who pipe out perfectly good redirect links to their targets. +Angr 09:28, 19 May 2010 (UTC)

I did a little research on Arpitan, and the best I could get was this orthography guide. This be an essentially unstandardized language, all of the following is based on what seems to be the best documented dialect, Savoyard: — ˈzɪzɨvə (talk) 07:34, 20 May 2010 (UTC)
 * The article says "Nous proposons de transcrire toutes ces consonnes, plus ou moins palatalisées selon les patois, par une combinaison avec ‹y› : ‹ly›, ‹ny›, ‹ky›, ‹ty›, ‹dy›, etc. Exemples: montanyi «montagne»; palyi «paille»; a rekyè «à l'abri du vent»; itye «ici»; dyablo «diable»"‹ly› and ‹ny› are shown elsewhere in the article to be, , but I don't know what ‹ky›, ‹ty›, ‹dy› are supposed to be. Are they just , , , and if so, is this phonemic?  Or are they  and ?
 * ‹h› seems to vary from to  to, so I'm not sure what to call it.
 * ‹r› seems to be an alveolar trill (dont on ne notera pas les différentes prononciations)
 * stress is contrastive, and on one of the two last syllables
 * I can't tell if what the article calls Les diphtongues (ou les suites de voyelles) are truly diphthongs or vowel+semivowel combos.
 * The moral of the story (apart from how difficult it is to search for phonology info on Arpitan) is that this is pretty different from French, and not much closer to Occitan, so perhaps throwing together its own IPA key would be the best way to justify extricating it from French. Here's the sum of what I've gathered:

I'm personally in favour of splitting into two different articles. I don't think being a closely related language justifies combining the two into a single article, not to mention that there are probably a lot of people who want help with French pronunciation and would just be confused by the inclusion of a different language that nobody has ever heard of. To wit, we have separate guides for Scottish Gaelic and Irish. I think that practice should be reserved for languages which are extremely similar in phonology, such as Swedish and Norwegian. AlexanderKaras (talk) 20:04, 20 May 2010 (UTC)


 * Why not merge Arpitan with WP:IPA for Occitan? — kwami (talk) 19:38, 22 May 2010 (UTC)

Requested move
IPA for Arpitan and French → — Arpitan now has its own page, and the Arpitan material has been removed from this page. — ˈzɪzɨvə (talk) 19:21, 22 May 2010 (UTC)


 * I'll move it. I don't see any reason for a formal request. — kwami (talk) 19:39, 22 May 2010 (UTC)

English equivalent for ɑ̃ is misleading
It currently gives "nasalized [ɑ]". [ɑ] is given as the vowel in "bra".

Each one of these sentences may perhaps be unexceptional on its own, but together they are very misleading. The English vowel [ɑ] in words like "bra" is in most accents far more open than the french [ɑ̃]. I would suggest that the closest oral equivalent of [ɑ̃], for most English accents, would be the vowel of strut. Grover cleveland (talk) 21:09, 1 June 2010 (UTC)
 * Huh? I've read that French is just a little forward from .  It certainly doesn't represent the vowel of strut acoustically to my ears. — Æµ§œš¹  [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi]  21:59, 1 June 2010 (UTC)
 * Also, the STRUT vowel is infamous for its wide variety of realizations across dialects. The Americans think the Brits pronounce love "lahv" and the Brits think the Americans pronounce it "lerve", and the Australians are different from both. So whose ʌ is French ɑ̃ supposed to be like? +Angr 22:13, 1 June 2010 (UTC)
 * Mine, I guess :) OK:  I admit that "strut" is not a good example.  However I still contend that the current text is misleading.  See, e.g. the French entry in the IPA Handbook:  "The position of the tongue is similar in [ɑ̃] and [ɔ̃], and the main articulatory difference is that [ɔ̃] has a greater degree of lip rounding." Grover cleveland (talk) 23:46, 1 June 2010 (UTC)
 * Ok, so what do we change it to? — Æµ§œš¹  [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 00:22, 2 June 2010 (UTC)

Phonetic nature of word breaks
This key (and the templates that links to it) encodes for phonetic, not phonemic, transcriptions. Perhaps we should extend this one further: does French phonology really use word-breaks? What with all the liaison, contracted forms, semivowel alternations, clitic pronouns, phrase-level stress, etc., I think you could make a strong case that we should transcribe phrases as one phonetic group:


 * Cour des aides as instead of
 * Président à mortier as instead of
 * Club de l'Entresol as instead of

Some reasons I feel this way:
 * 1) Liaison consonants resyllabify strongly with the following vowel (which word-breaks hide).
 * 2) It seems no less important to show the word-status of the contracted forms d'–, l'–, etc., than any noncontracted word.
 * 3) Some compound words are written as though separate words, but should not have a space in transcription (recent reforms have reduced this, with mixed success).
 * 4) This also rids us of the problem of marking phrasal stress leading to inappropriately marking stress on each word.

Would such a radical change be appropriate? Which is the weightier concern: the convenience of orthography-based word spacing, or representing the sound of French phrases more truly? — ˈzɪzɨvə (talk) 08:08, 29 November 2010 (UTC)
 * That makes sense. I argued something similar at Wikipedia talk:IPA for Catalan.  — Æµ§œš¹  [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi]  18:16, 29 November 2010 (UTC)


 * In theory, I agree. The problem is that many people have difficulty with the IPA. Keeping word spaces helps them parse it. Since the liaison mark indicates resyllabification, there isn't any actual inaccuracy in the current transcription. After all, the space isn't an IPA character, and it doesn't have any phonetic meaning. — kwami (talk) 07:16, 30 November 2010 (UTC)

W in oi digraph
It's important to note all the common cases where a sound occurs, so I noted the oi digraph as an occurrence of the w sound. Unfortunately neither of the letters really represents the sound, so it's a little misleading to bold either one of them. I bolded o, as the first in orthography as w is the first in pronunciation, and since it likely gave rise to the sound in the history of the language. What do others think: should we include this example, or should we not bold it at all? — Eru·tuon 00:58, 29 December 2010 (UTC)
 * That seems fine to me. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 01:04, 29 December 2010 (UTC)

Voiceless uvular fricative rhotic
Doesn't the voiceless fricative variant of the rhotic [χ] appear quite often in spoken French, particularly at the end of a voiceless syllable (vaincre, etre). It seemed particularly common in francophone Switzerland where I spent a lot of time as a kid, and I seem to only have heard the voiced trill in France. I feel like it's marked enough to be relevant, again at least in the Swiss dialect, but I wonder if you think it's marked in Parisian France and/or if you think it would be worth adding in the notes as a variant here. SamuelRiv (talk) 20:45, 1 April 2011 (UTC)
 * If it's an allophone of a normally voiced consonant that occurs in a voiceless environment, I wouldn't say it's particularly notable to mention here, though French phonology would be a good place for that kind of information. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 23:42, 1 April 2011 (UTC)
 * I guess the difference between a phoneme and an allophone, for a general language, would be defined by the phonotactics of the dialect itself. In this case, I am thinking that [χ] would be the dominant phoneme of at least this Swiss dialect, with the voiced form being the allophone. An example, a good determiner of this trait, would be the word raclette, which was distinctly [χaklɛt]. But honestly, this is a good thing to bring up against a published source. However, that can get messy - I read an essay on transcribing spoken versus musical French, and how the two would properly translate for singers of different languages, and the author there chose to transcribe both the spoken and sung rhotic as a flap [ɾ], presumably to emphasize the quickness of actual articulation. So... I don't know... is there a good standard or convention? SamuelRiv (talk) 01:49, 3 April 2011 (UTC)
 * I think the convention is what we're using right now. What you're talking about, by the way, is probably more accurately described as distribution of phones rather than phonotactics, since the latter refers to rules regarding phoneme placement, not that of allophones. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi]  02:40, 3 April 2011 (UTC)

english equivalent for "i" is wrong
on the actual page for this vowel the example given is "free", here the example is "bitter". I'm no expert, so I hesitate to edit. — Preceding unsigned comment added by 24.85.83.139 (talk) 02:26, October 14, 2011 (UTC)
 * The example for i was changed recently, and you're right, it's incorrect: French i is [i], but bitter has [ɪ]. I've corrected the example. — Eru·tuon 04:48, 14 October 2011 (UTC)

Made some modification

 * Added few more examples.
 * Perhaps someone could add final consonants are elided in many instances, but as already indicated here liaison "recovers" some of those omitted consonants when followed by a vowel sound in the same phrase, much like non-rhotic English dialects do with "r".
 * How could we use the syllable marker in French? Jɑυмe (xarrades) 21:47, 26 January 2011 (UTC)


 * I don't know whether the syllable marker is ever relevant for French. Does French ever distinguish sequences like [u.i] vs. [wi], [i.e] vs. [je], or [u.a] vs. [wa]? I can't think of a case where it would, but my French is kinda rusty. —Angr (talk) 22:59, 26 January 2011 (UTC)


 * French phonology mentions pays [pɛi] and paye [pɛj] as a minimal pair for full vowel vs. glide, so in transcribing pays you could use the syllable separator if you wanted to be extra disambiguatory. — Eru·tuon 23:22, 26 January 2011 (UTC)
 * Is disyllabic? — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi]  00:22, 27 January 2011 (UTC)
 * If it is dissyllabic this example might work for French where stress is a bit different, but not for other languages that contrast different types of stressed hiatuses, compare Portuguese pais [ˈpajʃ] (diphthong) with país [pɐˈiʃ] (hiatus) and pias (hiatus). Nonetheless, i wouldn't oppose to use it in French to ease the reader to see some possible hiatuses.Jɑυмe (xarrades) 01:20, 27 January 2011 (UTC)


 * Pays is disyllabic. Dictionaries give four different pronunciations, all of them with two full vowels: [pei], [peji], [pɛi], [pɛji]. The one given more weight currently is [pei], which is also what I hear when I listen to natives. See here. — Adi  Japan  03:36, 31 March 2011 (UTC)

I think moyen could be replaced by a phrase or short sentence... Something like tout est beau. What do you reckon? —Jɑυмe (xarrades) 15:44, 8 December 2011 (UTC)
 * Phrasal stress

Dental or alveolar /n/?
According to Fougeron & Smith (1993) French is dental. Why French is listed twice; with an alveolar nasal nous and dental nasal connexion ? Shouldn't French be deleted from the alveolar nasal article like Portuguese? Jɑυмe (xarrades) 15:56, 30 March 2011 (UTC)
 * Looks like someone added it to alveolar nasal last year. I've removed it. (comment added 01:20 GMT, 31 March 2011 by User:Aeusoes1)


 * I'd like to suggest/remind that widest transcription be preferable for articles on a language like French and leave the dental marker as perhaps what is listed for "standard" phonology, but not in word or syllable transcriptions. SamuelRiv (talk) 18:24, 8 December 2011 (UTC)

Add x
Title. And add uy. — Preceding unsigned comment added by 82.40.180.138 (talk) 10:21, 16 December 2011 (UTC)


 * AFAIK, neither /x/ nor /uy/ are phonemes of French. Angr (talk) 12:46, 16 December 2011 (UTC)
 * What about Puy-de-Dome, then ?Eregli bob (talk) 05:02, 1 September 2012 (UTC)
 * . — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ] 13:03, 1 September 2012 (UTC)

The value of the open vowel
According to the article, French phonology, the open vowel is central, not front. Why is it tagged a front vowel, not {IPAlink|ä|a}} ? Speaking about Fête's edit. --Mahmudmasri (talk) 23:50, 30 December 2012 (UTC)
 * I honestly know nothing about the question itself, but the user in question has been indefinitely blocked from fr.wiki & fr.wikt for introducing factual errors repeatedly.  Salvidrim!   00:18, 31 December 2012 (UTC)
 * It's hard to say either way. More than likely, the vowel's quality depends on speaker or it may depend on if the speaker in question has one or two low vowels.  — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ]  01:59, 31 December 2012 (UTC)
 * Once again, someone seems to be confusing vowel symbols with cardinal vowels. Although the cardinal vowel /a/ is a front vowel, the symbol /a/ stands for whatever openish vowel a language has that's not really back enough to be transcribed with /ɑ/. It covers a whole range of articulations, some of them central. Angr (talk) 09:53, 31 December 2012 (UTC)
 * Yep, not being rude but Fête must be overestimating his own understanding of the IPA to say that it is really a complete level at his user page. Portuguese varies between central and near-back, still it is transcribed with the symbol for cardinal vowels. In my personal opinion I find the marking of the IPA a as front ridiculous and absurd, it currently generates far too much confusion for people learning the alphabet like what we got here. But then the fact that I'm obsessed with the sounds of Portuguese and would love to transcribe it narrowly as  also strenghthens this opinion. lol 177.40.150.110 (talk) 15:37, 1 January 2013 (UTC)

di + vowel
Would like to see an indication on pronunciation of di + vowel. Many West African proper names (for people and places) are rendered in French with this spelling, but English speakers read it as a hard "d" with an extra syllable. Heard frequently in connection with the 2013 conflict in Mali that involved at one point the town of Diabaly (dee-ah-bah-lee, rather than like diable; the local Bambara pronunciation would actually be closer to jah-bah-lee).--A12n (talk) 16:39, 22 January 2013 (UTC)

pâte/bra equivalence
The article currently gives an equivalence for 'pâte' and 'bra' as an example of the 'a' sound (a). This is surely wrong. 'Pâte' has a short 'a', as in the English verb 'to pat'. 'Bra' has a long 'a' as in 'father' and 'after'. . Am I missing something? Span (talk) 04:36, 13 January 2013 (UTC)
 * For speakers who make a distinction between and, the latter is more back, like the vowel of bra, even if it isn't long. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ]  14:56, 13 January 2013 (UTC)
 * I'm not sure I understand your comment. Are you agreeing that there is no equivalence? Span (talk) 20:37, 13 January 2013 (UTC)
 * There never is crosslinguistically. These are approximations. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ]  20:55, 13 January 2013 (UTC)
 * The sound in 'pâte' is the same as 'cat' in English. Span (talk) 20:58, 13 January 2013 (UTC)
 * Depends on which dialect of English and which dialect of French. Most dialects, the vowel of cat is  or, even a front articulation of  is a weak approximation. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ]  00:22, 14 January 2013 (UTC)
 * I think the article is mistaken. It states that 'patte' sounds like 'pat' and 'pâte' sounds like 'bra'. This is misleading and unhelpful to readers as is.  Span (talk) 00:39, 14 January 2013 (UTC)
 * It's my understanding that there is dialectal variation among French speakers in this regard, including a number of speakers who do not make a distinction between pâte and patte. The speaker in the audio sample (who, by the way, sounds more like he's saying butt than pat to my anglophone ears) may be such a speaker. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ]  00:43, 14 January 2013 (UTC)
 * Maybe I'm missing something, but the vowel in English 'pat' is pronounced in pretty much every dialect of English, isn't it? I don't see how that's in any way equivalent to the  in French 'patte'. Maitreya (talk) 09:23, 22 February 2013 (UTC)
 * None of these vowels are. Again, they're approximations. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ]  16:03, 22 February 2013 (UTC)

Marginal phonemes
What about adding, and ?

Also perhaps, if I achieve consensus, among the main phonemes, as I think it is hugely inadequate to render yeísmo in every transcription since many people still use the lateral phoneme so it is still something reversible. If the person just doesn't like it, he/she would know that the true value of in his/her personal preference is. People often transcript cot-caught merger, non-rhoticity, Portuguese merger of unrounded mid vowel-onset diphthongs and Brazilian lowering of schwa in articles here, to me it seems like madness. The phoneme is still there for a lot of people, if not the most, their case is just dialectal, everybody will know about London's lack of coda, São Paulo's lack of final in infinite verbs or Spanish debbucalization and merger of final  and  when they travel there, using it in Wikipedia has no purpose at all. If there was, we would have vowel sets such as to transcript the RP/Cockney/Estuary English diphthong instead of using "American-centric". Lguipontes (talk) 06:14, 1 April 2013 (UTC)
 * Everything I've read about French phonology says that is rare among French dialects.  Do you have something that says otherwise? — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ]  13:06, 1 April 2013 (UTC)


 * Already at Google. I was not searching about it right now, but anyway, we know that, while yeísmo is happening since the start of the Modern Era in French, it still was there 40 years ago; also, the fact that American English doesn't have don't mean we will conflate this in the  set on the transcription. I will try my best, luckily English and Portuguese have a lot of French and I am used to Romance grammar. :P Lguipontes (talk) 16:56, 1 April 2013 (UTC)
 * Good luck on the search. If you find a ref for an article you think might help but you can't access, let me know and I'll try to get it. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ]  17:16, 1 April 2013 (UTC)


 * This 2005 Lithuanian research over French phonology states that "[a] part la sonante mouillée et l’expirée, le consonantisme du moyen français est celui du français moderne. L’élimination de la sonante  qui passera plus tard à  ne créera pas de relations nouvelles dans le système puisque la langue possède ce dernier phonème depuis des siècles. Il existe cependant des dialectes qui connaissent de nos jours la consonne mouillée , tels les dialectes du Midi et de l’Ouest de la France." at page 21, "[c]e qui caractérise le XVIe siècle, c’est essentiellement la réintroduction de consonnes qui s'étaient amuïes, et le règlement, partiel et provisoire, de la prononciation des voyelles finales. Ajoutons cependant quelques précisions: le  mouillé existe toujours, il se maintiendra [that is, will continue to be so] jusqu'au XIXe siècle (on prononce  et non ) ; les  sont «roulés», ils deviendront dorso-vélaires au XVIIe siècle." at page 28 and that "[c]omme le consonantisme moderne est déjà constitué en moyen français, il subit juste quelques retouches aux XVIIe et XVIIIe éliminant les derniers restes des caractéristiques d'autrefois, telle la mouillure du  qui passe à  dans le parler populaire parisien dès la fin du XVIIe, mais se maintient dans l'usage jusqu'au milieu du XIXe. Dans le système consonantique, il ne reste à partir du XIXe qu'une seule consonne mouillée—." (I believe they use this symbol for the same sound of Brazilian nh.) at page 35 even though "3) II existe des hésitations dans la prononciation  —  et  — : milliard [miˈljaːr — miˈjaːr]; panier [paˈnje — paˈɲe]; gagner [gaˈɲe — gaˈnje]" at page 41.


 * It seems that what previously were palatal or heavily palatalized lateral and liquids are for the most part non-lateral approximants and nasal approximants now, though some conservative speakers still have them the most ancient way (so IMHO at least for half of the French more closely fits Portuguese allophonies – just using sounds that demand less energy in articulation, and mastering liquids is indeed hard, not really a merger to the point you can neither hear the difference nor produce it – rather than Spanish yeísmo, seseo and English vowel and wh/w mergers, but we can't have the evidence as Portuguese is much more prone to both lateral and palatal minimal pairs than any other Gallo-Iberian AFAIK). They didn't mention Occitan or Arpitan interference (BTW they are so endangered now that I wouldn't expect) as one would expect from someone with a basis in Linguistics... that one was the best (most recent, complete, factually accurate, citing other people, not-my-native-language) I found and could understand (in 3 hours of research and 9 of sleep, I must admit). I could draw by it that it still is a phoneme, with a near-extinct presence in Oïl and Breton areas (true dying, if we are really to compare this with Spanish, only starting by the mid-1800s), a weak among the Provençals, and a mild presence in Corsican-, Arpitan- and other Occitan-speaking lands. Lguipontes (talk) 06:27, 2 April 2013 (UTC)
 * Yeah, none of that is surprising. I don't think there are enough dialects to make it worthwhile to change our transcriptions, though. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ] 11:58, 2 April 2013 (UTC)

"Vin" = "uh"?
I'm not a native French speaker, but the English equivalent given for ɛ̃ ("uh" as in "uh huh") seems way off. - ClaireJV (talk) 21:57, 17 April 2013 (UTC)


 * demain : [[File:Fr-demain.ogg]]

In Modern Parisian French, /ɛ̃/ becomes [æ̃]. 198.105.106.68 (talk) 13:14, 9 August 2013 (UTC)

Being a native speaker who has lived in different regions of France, I agree and would even add (though I'm no phonetician) that this is the standard Metropolitan French pronunciation, except for some parts of the South. [æ̃] or [ã], I'm not sure, but definitely not [ɛ̃]. Actually, all the descriptions of nasal vowels sound wrong to me, except [œ̃] which does exist but mainly in the South. And I'm pretty sure I have a standard pronunciation for these phonemes - I'm still talking only about Metropolitan French. As this is just my native, non-professional opinion, does anyone know of any source supporting this? Tanynep (talk) 09:22, 14 November 2013 (UTC)

Actually, I found a related discussion here. Tanynep (talk) 21:38, 2 June 2014 (UTC)

é = air (Australian accent)??
This seems highly odd for a number of reasons. First, while Australia is awesome and Australians are, to a man (and woman), quite excellent people, it's probably not that useful to provide a pronunciation example that requires knowing a particular sound in an accent with which comparatively few people around the world are intimately familiar. Second, and more importantly, I don't think that the Australian 'air' is anything like the French 'é'. A better example would be 'hey', which both sounds a lot closer to 'é' and has the additional benefit of sounding broadly similar in nearly every English dialect. -- Hux (talk) 02:38, 9 September 2014 (UTC)
 * Actually, according to Australian English, the vowel of air is [eː], which is indeed similar to French é, aside from being long rather than short. — Eru·tuon 02:03, 11 September 2014 (UTC)

Example: "Says"
Under sample English, it gives "Says" as an example for ɛ, what I used to call "Long A".

However, we routinely pronounce that word as"Sez". Better to list "Days" or "Ways" or even "Weighs" as the example. — Preceding unsigned comment added by 50.80.26.223 (talk) 18:13, 17 January 2015 (UTC)


 * Those would be closer to /e/. — kwami (talk) 22:02, 17 January 2015 (UTC)
 * You're missing the point. The example is supposed to have a word pronounced sez, rhyming with fez, not a word pronounced with ay, rhyming with weighs. The French sound being exemplified is [ɛː], not [eː] or [eɪ]. — Eru·tuon 01:19, 18 January 2015 (UTC)
 * Maybe we should use fez, then. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [lɛts b̥iː pʰəˈlaɪˀt] 02:24, 18 January 2015 (UTC)
 * Fez is a pretty rare word. Something like red would be better. (With voiced sound triggering long vowel in many English dialects, and no vowel shifts or allophonoic pronunciations that I know of occur before d, as opposed to n or l or r.) — Eru·tuon 03:02, 18 January 2015 (UTC)

Renault 4CV
More opinions requested at Talk:Renault 4CV. Is adding "CAT shu VOH" a useful addition to an article that already has IPA? One editor seems to think so and keeps re-adding. Andy Dingley (talk) 16:25, 18 November 2015 (UTC)

Tu like cute ?
I am not sure this is a good example. In both American and British english, "cute" seems to be pronounced like "cyute" ( as distinguished from "coot" ),  and the distinction in pronouciation which appears in "stupid" and "New York" does not seem to occur. Yet I never heard a French person pronounce "tu" like "tyu".Eregli bob (talk) 04:54, 1 September 2012 (UTC)
 * That's the closest thing we've got in English. It's either that or "no English equivalent" which IMHO isn't as helpful. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ]  13:04, 1 September 2012 (UTC)
 * If you're not clear on how it's an approximation, French u is [y], and the vowel in cute is [ju]. French [y] is a front rounded vowel. English [j] is an approximant equivalent of the front vowel [i], and [u] is a back rounded vowel (though actually it's almost central in many dialects of English). If you fuse the frontness of [i] and the roundedness of [u] into one sound, you get [y]. Therefore, [ju] is a good approximation for [y]. It's not an exact equivalent, but it's the best we have in English. — Eru·tuon 15:27, 1 September 2012 (UTC)
 * The American pronounciation of "tube" is much more similar to the French "tu", than "cute" is.Eregli bob (talk) 06:17, 2 September 2012 (UTC)
 * That vowel is much closer to that of French nous than of tu. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ]  14:13, 2 September 2012 (UTC)
 * IMO approximations are useful only if they would be understandable if used instead of the phoneme they approximate. I don't think it would be understandable if someone tried to speak French with [ju] instead of [y], though (although it probably would be if [œ] or [ø] was replaced by the approximations listed here). It might be better to give instructions on how to pronounce the sound in such cases ([y] is not that hard – it's just a rounded [i]). Double sharp (talk) 07:28, 20 November 2012 (UTC)
 * That instruction is less helpful than you think. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ] 13:48, 20 November 2012 (UTC)
 * Perhaps something like "pronounce [i], but with rounded lips", then? Double sharp (talk) 14:13, 20 November 2012 (UTC)
 * Same thing. Friends have told me that telling speakers to modify their ee vowel with rounding often results in a diphthong or an oo vowel.  To me, your measure of does-it-work-when-speaking-the-language is a slippery one, but is probably the most concrete measure we could have; I'd say it should make the difference between whether or not we say a sound an English sound "roughly" corresponds to that of a given language.  For French,  roughly corresponds with . — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ]  14:42, 20 November 2012 (UTC)
 * When teaching beginners, the technique I found worked best was sort of the reverse, and breaking it down into two steps: get them to say "oo" ("Ooh, I like that!"), then without moving their lips, try to say "ee". Really works, and you should see the look on their faces when that foreign sound emerges from their mouths. Any use here? Awien (talk) 00:07, 26 November 2015 (UTC)

pronunciation of o
for the pronunciation of o the English approximations of "go (Scottish English)" and "law (modern RP)" are given. But I found the second example of "law" confusing because in American English "law" has a much different sound than go and I was unfamiliar with the English Received Pronunciation of law, as I expect most Americans are as well. Perhaps another English word could be found to replace "law" that is pronounced nearly the same in the English RP as in American English. Maybe snow, low, tow, blow, or something. Mindbuilder (talk) 09:13, 29 July 2015 (UTC)
 * There's no such word I'm afraid. 89.72.244.110 (talk) 16:57, 22 April 2016 (UTC)

The English approximation currently says "go (Scottish and American English)". For American English, this is simply incorrect. It's the diphthong /oʊ/. Anrza (talk) 22:50, 25 September 2016 (UTC)
 * is the traditional transcription. The actual pronunciation is often monophthongal. You can see that is listed as an alternative under General American in the International Phonetic Alphabet chart for English dialects. — Eru·tuon 00:42, 26 September 2016 (UTC)

English approximation for "ʁ" as "loch?"
I think the English approximation for "ʁ" as the "ch" in loch is wrong, no? — Preceding unsigned comment added by 76.111.68.8 (talk) 03:06, 24 April 2014 (UTC)
 * Pretty spot-on for Belgium, actually, and probably the best we can do for the standard sound. — lfdder 03:12, 24 April 2014 (UTC)
 * Isn't it just a rolled R? I'm not an expert on Scottish English, but I am sure that's not correct. 88.107.156.109 (talk) 17:50, 14 April 2016 (UTC)
 * A rolled r is with the tip or blade of the tongue. The French R is at the back of the mouth. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [lɛts b̥iː pʰəˈlaɪˀt] 22:29, 14 April 2016 (UTC)
 * Isn't it, then, close to the Welsh double L sound (as in "Llanfair"), but substituting an r sound for the l sound? — Preceding unsigned comment added by 2602:30A:2C4A:1CB0:4DF7:916:EE79:A0BF (talk) 10:49, 30 November 2016 (UTC)
 * No. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [lɛts b̥iː pʰəˈlaɪˀt] 13:40, 30 November 2016 (UTC)

"o" seems to be two sounds merged into one
The "o" in a word like "côte" is much closer to the RP sound of "but." It is far from the rounded-o sound at the end of "bureau." — Preceding unsigned comment added by 47.34.162.72 (talk) 14:24, 9 July 2017 (UTC)
 * The word bureau is used as an example French word, not as an English equivalent. Is that confusing? Should we be using French example words that haven't been borrowed into English? — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [lɛts b̥iː pʰəˈlaɪˀt] 15:48, 9 July 2017 (UTC)
 * Since someone has found it confusing, it would probably not hurt to choose a different example word containing eau. — Eru·tuon 20:44, 9 July 2017 (UTC)
 * Yeah, we definitely shouldn't be using loanwords from a language that is the topic of the help page (any help page). Mr KEBAB (talk) 22:02, 9 July 2017 (UTC)
 * I disagree. The column is already ripe with French words, it is rather hard to misinterpret it the way the OP did. Words that have also been borrowed into English (and other languages) as examples help because they are more likely to have been heard (and recognized) by those who don't speak French as first language. Now, were loans like bureau in the English approximation column, that surely would be confusing. But not the other way round. Nardog (talk) 04:37, 10 July 2017 (UTC)
 * Words that have been borrowed into English are probably the worst examples we can use for just the reason you articulated, Nardog. Even if we don't consider someone mistakenly confusing the example column as an approximation one, people are more likely to believe the example to be pronounced more like the English loan than is really the case, than if it were an unfamiliar word. Their familiarity would thus steer them in the wrong direction. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [lɛts b̥iː pʰəˈlaɪˀt] 05:01, 10 July 2017 (UTC)
 * Hmm, by "heard (and recognized)" I meant when they hear French speech though. But obviously I'm the minority, so that's just my perspective I guess. Nardog (talk) 05:32, 10 July 2017 (UTC)

Move discussion in progress
There is a move discussion in progress on Help talk:IPA which affects this page. Please participate on that page and not in this talk page section. Thank you. —RMCD bot 16:16, 15 July 2017 (UTC)

The word mètre
In most cases, (older) French French and Canadian French agree on which which words have [ɛ] and which ones have [ɛː]. However, they differ on some words. Mètre is a poor example to illustrate [ɛː] because while for Canadians this vowel is indeed long, it is short for those speakers in France who maintain the distinction. This can be seen by comparing the entries for maître, mètre and mettre in the Littré dictionary online, where the pronunciations are given as mê-tr, mè-tr and mè-tr, respectively. The Canadian pronunciations are given in the Usito dictionary as [mɛːtʀ], [mɛːtʀ] and [mɛtʀ], respectively. Thus in France mètre is pronounced like mettre, whereas in Canada it is pronounced like maître. .24.50.171.67 (talk) 03:41, 14 October 2017 (UTC)

ə
there should be a note that /ə/ is often pronounced as /ø/ LICA98 (talk) 16:57, 25 January 2018 (UTC)
 * ✅ Nardog (talk) 18:41, 25 January 2018 (UTC)

Reality check
How does this guide actually work? How many dialects do we cover here and on which reputable sources do we base our transcriptions?

My questions are:
 * What about the distribution of the mid vowels vs. ?
 * What about the distribution of the short open-mid vs. the long open-mid ?
 * What about the distribution of the open vowels vs. ? Because we know that it's a mess in Parisian French.
 * What about the phonetic vowel length? Does it vary between dialects? We don't transcribe it here, I'm just being curious.
 * What about liaison? Surely it doesn't work the same in all dialects?

It seems to me that this guide is trying too hard to acommodate multiple dialects and it can't be turned into a diaphonemic guide akin to Help:IPA/English, which was my original idea. If that is the case, it might be time to turn this guide into the IPA guide for Parisian French and remove from the list. We can have a separate guide for other varieties, which probably still enjoy less prestige internationally and aren't as commonly taught to foreigners. Kbb2 (ex. Mr KEBAB) (talk) 12:14, 7 June 2018 (UTC)

We could actually add phonetic vowel length should we decide to limit this guide to Parisian French. Kbb2 (ex. Mr KEBAB) (talk) 12:40, 7 June 2018 (UTC)

mix up
"ʁ - regarder, nôtre - roughly like loch (Scottish English) but voiced, like 'gh' in Scottish Gaelic". A line or some cells missing it seems.  Tim riley  talk   19:22, 24 September 2018 (UTC)
 * That was not wrong, but indeed a weird note. Why the hell did you think it was a mix up? Nardog (talk) 19:48, 24 September 2018 (UTC)
 * Well, my Scottish accent isn't up to much (some unkind souls might say my French one isn't either) but I don't pronounce the French "r" like any "gh" phoneme I've heard. Happy to be corrected if I'm in error.  Tim riley  talk   20:17, 24 September 2018 (UTC)
 * How's your Scottish Gaelic? — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [lɛts b̥iː pʰəˈlaɪˀt] 20:24, 24 September 2018 (UTC)
 * Rather a good point. About as good as that of a vanishingly small percentage of our readers, I should think, so perhaps not a frightfully helpful cognate to cite.  Tim riley  talk   20:27, 24 September 2018 (UTC)
 * Agreed. It doesn't make sense to put non-English examples in the "English approximation" column. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [lɛts b̥iː pʰəˈlaɪˀt] 20:30, 24 September 2018 (UTC)

While I'm here, I have only been dabbling with French IPA hieroglyphics because I wanted to make the subject's pronunciation clear in the article on Ambroise Thomas, and if anyone more competent than I in IPA/French cares to look in there and check and if necessary amend my attempts it will be esteemed a favour.  Tim riley  talk   20:35, 24 September 2018 (UTC)
 * It looks good to me. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [lɛts b̥iː pʰəˈlaɪˀt] 22:13, 24 September 2018 (UTC)
 * Thank you. That's most reassuring. I'm much obliged.  Tim riley  talk   22:16, 24 September 2018 (UTC)

English approximation of [ɛː]
Is "Fairy" really the right vowel for [ɛː]? Doesn't seem right to me. Derekt75 (talk) 20:32, 28 June 2019 (UTC)

English "fairy" is diphthongal, while the French [ɛː] sound is not. Changed back to "red". Derekt75 (talk) 18:12, 3 July 2019 (UTC)
 * Red isn't a good example because it's short. Fairy has the SQUARE vowel, which is often long and sometimes has a slight diphthongal pronunciation, depending on pronunciation. Fairy is a much closer approximation than red. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [lɛts b̥iː pʰəˈlaɪˀt] 18:32, 3 July 2019 (UTC)

English approximation for [e]
The FACE vowel is diphthongal in most accents and its nucleus can be as low as [ɛ] or [æ] particularly in modern Southern British and Australian varieties, while KIT is monophthongal and can approximate or even be lower than [e] especially in North America (see e.g. the vowel diagrams for RP, GA, California, Inland North, and Standard Canadian). So KIT conveys the value of French [e] most efficiently. Nardog (talk) 12:29, 2 April 2019 (UTC)
 * I'm sympathetic to resistance to changing this from the FACE vowel. That's the vowel we choose for a lot of these guides. There are a lot of French words where [e] appears in open syllables, so using the KIT vowel feels very counterintuitive. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [lɛts b̥iː pʰəˈlaɪˀt] 21:34, 2 April 2019 (UTC)
 * Isn't counterintuitiveness what we want, though? English speakers typically use FACE for [e] or [ɛ] at the end of a word in another language, often without realizing it's diphthongal, sounding off to the speakers of that language. Citing KIT and DRESS as approximations for those vowels, even though they would be in violation of phonotactics if used word-finally in English, draws attention to this difference—which is beneficial to those hitherto unfamiliar with it. We shouldn't be using FACE for a monophthong except for languages with a three-way contrast of [i–ɪ–e] in the close-front region IMO. Nardog (talk) 11:41, 3 April 2019 (UTC)
 * Most of that makes sense, and I'm hand wringing about this because it's unusual to choose KIT as the closest approximation both at Wikipedia and in pronunciation guides in general. What if we specifically identified the GA variant of the FACE vowel as the closest English approximation? We do often focus on dialect-specific variants in cases like this. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [lɛts b̥iː pʰəˈlaɪˀt] 15:34, 3 April 2019 (UTC)
 * But again, GA FACE is diphthongal. We can do something like "grey", but I thought you were opposed to breaking up diphthongs like that (which I agree with). Are you talking about specifying the dialect of FACE in guides for languages where there is an [i–ɪ–e] contrast, or just in general (including French)? Nardog (talk) 16:58, 3 April 2019 (UTC)
 * Yeah, I am opposed to breaking up diphthongs like that. It's not a deal breaker for me that the GA FACE can be diphthongal, even when there isn't the three-way contrast that you talk about for other languages. We might even be able to identify a word that more often features a monophthongal pronunciation (does this occur before voiceless stops as in wait?). Maybe other people can chime in and share what they think. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [lɛts b̥iː pʰəˈlaɪˀt] 17:49, 3 April 2019 (UTC)
 * I am not extremely knowledgeable in this area, but the close-mid front unrounded vowel page lists MAY as being the [e] sound in English. The vowel sound in KIT is a near-close front unrounded vowel -- not even close in my Inland North accent, nor in any other accent in the U.S. that I have personally ever heard. -PanYaLin (talk) 18:55, 3 June 2019 (UTC)
 * I'm not knowledgeable, either, but "HIT" is simply wrong. "MAY", "DAY", "FACE", might be somewhat diphtongal, but at least the initial vowel sound is correct.  "HIT" is a completely different, incorrect, vowel sound.  I'm going to change it back to "DAY".  If there's a better word, great, but reverting to "HIT" is clearly inferior.  Derekt75 (talk) 20:13, 28 June 2019 (UTC)


 * I am opposed to breaking up, too. Please note that the realisations of mentioned in article Australian English phonology are , so English day is a confusing example (at least) for native English speakers from that continent. Love —LiliCharlie (talk) 20:35, 28 June 2019 (UTC)
 * Yeah, "DAY" isn't great. but it's better than "HIT".  If there's something better, I'd welcome the change.  just please nobody revert back to "HIT" again.  :-)  Derekt75 (talk) 15:56, 1 July 2019 (UTC)
 * Again, see the vowel diagrams at Received Pronunciation, General American, California English, Inland Northern American English, Standard Canadian English, etc. You haven't provided an actual argument based on phonetics or phonology. Why exactly is better than ? (Meanwhike, I'm changing it to kid because it might even be slightly closer to French [e] because it doesn't have the glottalization and clipping hit has.) Nardog (talk) 17:18, 1 July 2019 (UTC)
 * The vowel in question is /e/, not . In America, we call it the "long A" sound.  The links you provide only prove my point.  If you read the General American vowels page, you'll see that the vowel is almost always a diphthong to Americans, and it provides the examples of "lake", "paid", and "faint".  Use one of those if you prefer.  Most English speakers (maybe not Australians?) would think of the right vowel when provided the word "day" or "may".  The page Close-mid_front_unrounded_vowel uses the word "may" as the English equivalent, and I think that's good enough for me.  If you want something more monophthongal, I think the 2nd "a" in "alligAtor" would do the trick, but that's awfully long, and I really think most English speakers don't pay attention to the fact that their vowel shifts when they say the word "may".  Derekt75 (talk) 06:28, 2 July 2019 (UTC)
 * Symbols are not sounds. You said it yourself: the vowel is almost always a diphthong ... most English speakers don't pay attention to the fact that their vowel shifts when they say the word "may". French /e/ is not. And we want readers of the guide to pay attention to this difference, or else what's the point of a guide? Think of it this way: If [e] in a French word was replaced by a vowel in English, which one would French speakers find less marked, or ?
 * It seems to me most of the opposition to using as the approximation to French [e] has come from native English speakers. I want to hear from native French speakers, or at least native speakers of languages that permit an [e]-like vowel to end a word. If they also think  is better, I'll concede. Nardog (talk) 07:49, 3 July 2019 (UTC)


 * It's probably better for non-native speakers to err on the side than on the  side. The reason is simple:  is a frequent allophone of  in some well-known accents (e.g. of Quebec), while many speakers (e.g. of Southern France) always pronounce  as  in open and  in closed syllables, so that contrast is very weak, and listeners are prepared not to rely on it. My favourite choice would be to say that French  is like Australian English , and if English speakers from elsewhere pronounce English/​French  as  this is unkikely to hinder communication. — I think English  should not be used, or be used in cases like French  in paysage or  in oreille. Love —LiliCharlie (talk) 08:42, 3 July 2019 (UTC)
 * nardog, You could also try listening to any Frenchman. Here's a link to a Macron press conference:   He uses the  phoneme a number of times in his first handful of words:  mesdames, messieus les membres ... J'ai souhaité vous rencontrer devant les Français....  That last vowel in Français is held, and it's very distinctly .  None of those vowels matches .  I'm largely unfamiliar with the Canadian French accent, but I've heard the province Québec said enough times to know that the Québécois use  for the first vowel in Québec.  I'm going to change the word to match the word on the Close-mid_front_unrounded_vowel page.  Before you change it back, consider how many wikipedia editors have changed the word to a  sound.  As far as I can tell, it's just you.  Now consider how many have tried to make it a .  Derekt75 (talk) 18:01, 3 July 2019 (UTC)
 * I'd say we have a pre-existing consensus to use FACE, given that whenever a language contrasts two front mid vowels, we overwhelmingly tend to use the FACE vowel in our "English approximation" column for the closer one. Because this page is a little too high profile for us to experiment, let's just keep it as FACE until we actually get consensus for KIT.
 * I've never seen a French-language learning guide use KIT as an equivalent to French and have only seen them use the FACE vowel. Nardog, have you looked at such guides to see if any use something different than English FACE, either with IPA or English-orthography based respellings?
 * As I said before, it might make the most sense to identify a specific dialect that has a FACE vowel closest to French, that way our Australian readers won't be as confused. If GA is too diphthongal, we can always specify a dialect like Scottish. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [lɛts b̥iː pʰəˈlaɪˀt] 19:32, 3 July 2019 (UTC)

Should non-distinctive vowel length be shown in transcriptions?
It was pointed out to me elsewhere that the transcriptions are intended for people who have perhaps little knowledge of French. These readers can hardly be relied on to supply vowel lengths automatically. Older French-English dictionaries did include vowel lengths systematically. Perhaps that's less common now, though. 96.46.204.126 (talk) 06:17, 13 February 2012 (UTC)
 * We do that for Italian, so I don't have a problem with that. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ] 13:22, 13 February 2012 (UTC)
 * If transcription is phonetic [...], non-distinctive length should be shown. If it's phonemic /.../, no. In Italian, for example, vowel length is non-distinctive word-internally, predictable by stress and structure: long in stressed open syllable, otherwise short. The minimal pair fato 'fate' and fatto 'done' are phonemically /ˈfa.to/ and /ˈfat.to/, thus phonetically [ˈfaːto] and [ˈfatto] --47.32.20.133 (talk) 14:49, 10 July 2018 (UTC)
 * Yeah, transcription is phonetic. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [lɛts b̥iː pʰəˈlaɪˀt] 15:45, 10 July 2018 (UTC)
 * But how does vowel length work in French? Does it work the same in all major dialects? If so, there's nothing stopping us from transcribing it. Kbb2 (ex. Mr KEBAB) (talk) 17:53, 10 July 2018 (UTC)
 * I believe that we should transcribe Parisian French on Wikipedia, with phonetic vowel length. The distinction between and,  and  as well as  and  should be deprecated in favor of the former member of each pair. Exceptions to this could (not that they necessarily should) be made in the case of Quebec French as it is spoken in Canada. But I don't think we should transcribe Southern French, Belgian French or Swiss French (let alone African French) here. Parisian French is acceptable everywhere and it's the main variant taught to non-natives as well as the model used in dictionaries. Kbb2 (ex. Mr KEBAB) (talk) 07:32, 14 September 2019 (UTC)


 * The length distinction of vs.  is rarely indicated in recent leading dictionaries published in France. The same goes for the  vs.  distinction, which many Île-de-France speakers still seem to maintain, but with wildly varying distributions among speakers and in dictionaries published in the last one or two centuries. Conversely, the distribution of  and  follows orthography — in native words,  is spelled with $⟨un⟩$, and an opposing  is spelled otherwise — and is therefore much clearer; however it is maintained only by a minority of standard Hexagonal French speakers.
 * As to stress and combinatory vowel length: I don't think we should indicate them, as the former depends on sentence structure — what has been called the French "phonetic word" is an entire phrase, not a syntactic or orthographic word —, and the latter depends to a large extent on the former. (It is interesting to observe that French loans are usually pronounced with final stress on every single orthographic word in American English and Germany German, whereas the English and the German-speaking Swiss, who are more acquainted with the French language, prefer to treat them as having no particular stress pattern, so they apply their own Germanic rule of stressing the initial syllable.) Love —LiliCharlie (talk) 18:38, 14 September 2019 (UTC)

"French has no word-level stress" yet "grammatical stress is always on the final full syllable..."
The claim of no word-level stress is contradicted in the French phonology article that the Help links to, where it's made clear that word-level stress exists, and the position is described (albeit, alas, not immediately exemplified): "grammatical stress is always on the final full syllable (syllable with a vowel other than schwa) of a word." --47.32.20.133 (talk) 15:21, 10 July 2018 (UTC)
 * Word stress exists in French but is always rendered on the last syllable. Thereof, it would be pointless to denote it using IPA as every word always be denoted with a stress on every last syllable. --Bankster (talk) 17:37, 30 September 2019 (UTC)
 * Last full syllable. That excludes syllables whose nucleus is /ə/. Nardog (talk) 17:41, 30 September 2019 (UTC)
 * If it were just that, we would still want to denote it in our transcriptions, as we do with contextual allophones that are non-phonemic but also consistent and predictable. We don't indicate stress in French because it's phrase-level, so a word's stress will differ depending on if its position in a phonological phrase.
 * I mean, if someone doesn't like that reason, I can listen. After all, we tend to indicate transcriptions of words in languages with final-devoicing as if they are pronounced in isolation and we could do the same with French and stress. — Æµ§œš¹ <span title="Representation in the International Phonetic Alphabet (IPA)" class="IPA"> [lɛts b̥iː pʰəˈlaɪˀt] 18:18, 30 September 2019 (UTC)