Help talk:IPA/French

amendments

- I find the sound in book more similar to the French o than story which resembles /ɔ̃/ in French
- lab in British English is pronounced /a/ rather than /æ/, hence would be a better example than trap
- the sound in mace (/ɛ/) in BE just isn't the same as in clé (/e/)
- monsieur & faisons resembles /ø/ rather than /ə/
Couiros22 (talk) 18:40, 5 September 2023 (UTC)

Firstly, I'm not a linguist or a fluent French speaker, so take this with a grain of salt.

I find the sound in book more similar to the French o

The sound in "book" is [ʊ], which is also a realization of /u/ in Quebecois (e.g. in «route»), so that's not a good example regardless of similarity, but either way, I disagree that it's more similar. The main thing is [ʊ] is centralized while /o/ is not, AFAIK.

story which resembles /ɔ̃/ in French

The "o" in "story" isn't nasalized in any way, so maybe you meant /ɔ/ like in «sort» ?

lab in British English is pronounced /a/ rather than /æ/, hence would be a better example than trap

/a/ isn't a phoneme in English, and British English isn't one thing, so are you talking about [a] as the standard realization of /æ/ in Received Pronunciation? I actually have that too in my Canadian accent, but "lab" and "trap" have the exact same vowel for me. Anyway, TRAP is the keyword for /æ/, so I'd rather stick with that regardless of whether some dialects can get a closer approximation. Some of the other approximations are so far off that a detail like this isn't terribly important IMHO.

the sound in mace (/ɛ/) in BE just isn't the same as in clé (/e/)

The sound in "mace" is /eɪ/, not /ɛ/ (as in "best"). I really have no idea where you got that, because in most British dialects, it's either a diphthong like [eɪ] or a long vowel like [e:] (e.g. Yorkshire).

monsieur & faisons resembles /ø/ rather than /ə/

/ə/ can be realized as [ø], maybe that's what you're hearing. That's actually mentioned in the page: In Mainland French, while /ə/ is phonologically distinct, its phonetic quality tends to coincide with either /ø/ or /œ/.

— W.andrea (talk) 22:48, 5 September 2023 (UTC)

- the sound in book is still by far the most similar to French o ; according to wiktionary story and sort are both pronounced /ɔ/

- in standard British English band (/æ/) ≠ bad (/a/) , hand /æ/ ≠ had (/a/) ; grændad.

- standard BE : mace : ɛɪ not eɪ

- so maybe we should have separate examples for /ø/ & /œ/

--Couiros22 (talk) 07:46, 6 September 2023 (UTC)

Please clarify:

story and sort are both pronounced /ɔ/

English doesn't have a phoneme /ɔ/ in the Wikipedia system.

standard British English ... /a/

As I said, those don't exist. Are you talking about RP?

ɛɪ not eɪ

Please use the correct notation.

Regardless, what are you proposing instead of "mace"? English doesn't have any sounds closer to the French /e/.

so maybe we should have separate examples for /ø/ & /œ/

We do. They start with «ceux» and «sœur», respectively.

— W.andrea (talk) 18:12, 6 September 2023 (UTC)

- ok I meant [ɔː], which is not the same as [o]

- in RP and standard English accent (except up north) 'grandad' is pronounced /grændad/

- ...while mace is pronounced /ɛɪ/ not /eɪ/

- so perhaps we should displace the examples under [ə] to [ø]

Couiros22 (talk) 20:22, 6 September 2023 (UTC)

/grændad/ ... /ɛɪ/ not /eɪ/ ... [ə] to [ø]

Sorry, I don't want to be rude, but that notation is still incorrect. Do you not know the difference between phonemes and phones?

Maybe part of the problem is you're confusing the forward slashes with backslashes? Those are often used in French, like on Wiktionnaire, for IPA-based transcriptions that -- I'm not sure how to explain it exactly -- sort of take a middle-ground between phonemic and broad phonetic.

— W.andrea (talk) 22:50, 6 September 2023 (UTC)

still comprehensible though (regardless of the notation) ...so could you please address each of the comments ? --Couiros22 (talk) 08:53, 7 September 2023 (UTC)

At this point I've lost track of what we're debating. Could you restate your argument? — W.andrea (talk) 18:19, 8 September 2023 (UTC)

- the French o phoneme's closest equivalent in English is [ʊ] as in 'book' (the o in 'story' corresponds to another French vowel sound : [ɔ̃], eventhough it isn't nasalised)

- the French a phoneme's closest equivalent in English (different to [æ]) is the [a] found in British English; however most of the time is only found in a minority of words (often before b,d,g, such as the second a in 'grandad', handbag' etc.) ; words like trap use the ordinary æ and so aren't a good example of equivalent

- 'mace' is majoritarily pronounced using [ɛ] rather than [e] across England , hence doesn't correspond to the e in French words such as clé etc.

Couiros22 (talk) 18:57, 8 September 2023 (UTC)

This is off-topic, but please be careful which message you're replying to. This is the third time I've had to fix the formatting. In this case it's not too bad cause it's just us two talking, but if there were more people involved, it'd be unclear who you're replying to. — W.andrea (talk) 21:32, 8 September 2023 (UTC)

the French o phoneme's closest equivalent in English is [ʊ]

I've already addressed this in my first reply. (... also a realization of /u/ in Quebecois ... [ʊ] is centralized ...)

words like trap use the ordinary æ and so aren't a good example of equivalent

I've already addressed this in my first reply. (TRAP is the keyword for /æ/ ... this isn't terribly important IMHO.)

'mace' is majoritarily pronounced using [ɛ] rather than [e]

Again, what are you proposing instead? Secondly, you're still presenting it as a monophthong, but it's not.

— W.andrea (talk) 21:46, 8 September 2023 (UTC)

- Quebecois is a cherry-picked example

- /æ/ is not used in French

- words like 'fit' 'did' 'insipid' etc.

Couiros22 (talk) 09:14, 9 September 2023 (UTC)

Please don't use edit summaries to discuss. After one revert, discuss on the talk page instead of reverting again. We can continue discussion here, but I've already addressed the points you brought up in your edit summary:

(trap) is /æ/, (grandad) is /a/

/a/ isn't a phoneme in English ... Some of the other approximations are so far off that a detail like this isn't terribly important IMHO.

— W.andrea (talk) 18:01, 12 September 2023 (UTC)

/a/ is a phoneme in English (at least in some varieties) ; here are some examples : North American English regional phonology --Couiros22 (talk) 20:01, 12 September 2023 (UTC)

I mean in the Wikipedia system, which is technically diaphonemic. — W.andrea (talk) 20:35, 12 September 2023 (UTC) edited 21:48, 13 September 2023 (UTC)

/a/ is also frequent in Britain let alone the North of England where even words like trap use this vowel

+ I don't think there's anyone in Britain who pronounces the second a in 'grandad' as /æ/ ...I think it's even impossible if it isn't slightly diphthongized e.g. in North American English /ɪæ/ — Preceding unsigned comment added by Couiros22 (talk • contribs) 08:42, 13 September 2023 (UTC)

/a/ isn't a phoneme in English ...

yes it is... in British English

Some of the other approximations are so far off that a detail like this isn't terribly important IMHO...

exactly, which is why I amended them right from the beginning... ; moreover there are many subtle nuances on the help page (particularly a/â ; this one is distinguishable in oral French)

from the edit summary

Does not apply to rhotic dialects.

so let's unbold the r then (as I did in some of my initial modifications, which you reverted without explanation)

- regarding book and story, I've also explained why the former is a better example (story just is too different, and strongly resembles another French phoneme ɔ̃)

- book being pronounced differently in a few very minor dialects isn't a good enough reason to invalidate this example --Couiros22 (talk) 09:09, 17 September 2023 (UTC)

Couiros22, please indent your comments correctly, you have been advised to do this earlier in this discussion, failing to do it is just rude.

BTW, there is no such thing as British English in the context of spoken language, as distinct from written English. I suspect you really mean RP or Received Pronunciation, remember there is a great variety of spoken dialects in Great Britain. Please use accurate terminology unless you wish to be reverted out of hand for arrant nonsense. - Nick Thorne ^talk 13:41, 17 September 2023 (UTC)

Also, please obtain consensus here before making your changes again. - Nick Thorne ^talk 13:48, 17 September 2023 (UTC)

Re Special:Diff/1175799536:

The first consonants in wet and one are the same sound. The latter is an utterly poor choice because o stands for both a consonant and vowel, /wʌ/ (another way to look at it is that /w/ is unwritten).

[ɥ] is a consonant. huit and Puy are [ɥi] and [pɥi], respectively, where [i] is the syllable nucleus. Huey is /ˈhjuːi/, which has two vowels in succession, and it is /uː/ that is more prominent. The insertion of Huey was a poor choice to begin with, so I suggest we restore the previous approximation, "between wet and yet", which is a straightforward description of what [ɥ] is—a labial–palatal approximant.

All vowels in trap and grandad are the same. Just look up the words in the OED, which uses ⟨a⟩ for BrE and ⟨æ⟩ for AmE. Our dialect-neutral IPA key for English represents it as /æ/, following the predominant notation before the 1990s. Oxford University Press has switched to ⟨a⟩ to represent the same phoneme in BrE in dictionaries aimed at native speakers, but this is just that—a different representation. It means the same thing as /æ/—just look up the word in a dictionary from any other publisher that uses IPA, or even OUP's dictionary not aimed at native speakers, such as the Advanced Learner's Dictionary.

The first vowel in hammock is the same as the ones in trap and grandad. This change makes no sense. As the accompanying footnote explains, [ɑ] in this key specifically represents the quality when not merged with /a/, unlike in contemporary Parisian French, so using anything other than /ɑː/ (which is the vowel in bra) would be inexplicable.

insipid for [e] is the only change of yours I find defensible, but people found it confusing so we'd better stick to /eɪ/, which is monophthongal in some varieties of English (e.g. Scottish).

bird and burn have the same vowel. Most varieties of English don't have front rounded vowels, so English has only /ɜːr/ to approximate [ø] or [œ]. Other keys note the suboptimal nature by prefacing it with "somewhat like". Denoting the difference between the two like "... but more open/close" is an option, but this is likely pointless because those who understand such terminology wouldn't need the approximations to begin with.

story is a far better approximation to [o] than book. The quality of the vowel in southern Britain has diverged from the conventional ⟨ɔ⟩ towards [o], see e.g. [1]. /ʊ/, the vowel in book, is not only higher but centered. /ɔːr/ (but without highlighting the /r/) also works for American accents, unlike /oʊ/ or /ɔː/ not followed by /r/, because it's not diphthongal and is little affected by the cot–caught merger.

off and cover don't have the same vowel. off is already close (most accents, with or without the lot–cloth split or cot–caught merger, have it around [ɔ]), and listing two is pointless and confusing.

Overall, your edits and comments show little understanding of the phonetics and phonology of English and French and the pronunciation of English words. Please do not edit the key without establishing consensus on the talk page any further. Nardog (talk) 15:25, 17 September 2023 (UTC)

BTW, there is no such thing as British English in the context of spoken language, as distinct from written English.

I meant for the phoneme in question, this is the case of every dialect in Britain, where the second a in graded and handbag are much closer to /a/ (except in northern England and Scotland, where even the first a resembles /a/.

The latter is an utterly poor choice because o stands for both a consonant and vowel

It's the best equivalent for loi.

"between wet and yet"

How can huit be anywhere near /j/ ?

insipid for [e] is the only change of yours I find defensible, but people found it confusing

I find mate even more confusing, which isn't pronounced as é or [I] in most British dialects.

story is a far better approximation to [o] than book.

...even in American English?

off and cover don't have the same vowel

I meant sort and off have the same pronunciation, so do pomme and cover

Couiros22 (talk) 16:23, 17 September 2023 (UTC)

so do pomme and cover is a fair point. /ɔ/ is centered in Paris and may sound close to Southern British/Australian STRUT [ɐ]. Sol505000 (talk) 13:41, 18 September 2023 (UTC)

Change "between yet and wet" to wheat

The French word huit, the ɥ sound, is very similar to the sound of the English word wheat (grass used for flour). Thoughts? ScienceOwl13 (talk) 01:54, 12 June 2024 (UTC)

The first consonant in wheat is voiceless for some speakers and is just /w/ so I don't see how it would be an improvement. Nardog (talk) 02:25, 12 June 2024 (UTC)

I agree. "Wheat" pronounced /wiːt/ is more comparable to «oui» [wi]. — W.andrea (talk) 17:38, 12 June 2024 (UTC)

Source cleanup

I propose several cleanups to some text and links in the source code of this article.

−

[[guttural R#French|uvular]]

+

[[Guttural R#French|uvular]]

−

[[Vowel]]s

+

[[Vowels]]

−

[[elision|elided]]

+

[[Elision|elided]]

−

[[e muet]]

+

[[French phonology#Schwa|e muet]]

See revision 1244706983 for details. Whn9695 (talk) 20:18, 8 September 2024 (UTC)

For guttural R and elision, go ahead, but it doesn't affect the rendered article since links are normalized to have the first letter capitalized.

For e muet, don't change it. The page e muet is a redirect that already goes to French phonology#Schwa, and if the link ain't broken, don't fix it. See also MOS:NOPIPE.

For Vowels, why? There's no need to do that and linking to Vowel saves an unnecessary redirect.

— W.andrea (talk) 22:30, 8 September 2024 (UTC)

The links to "consonants" and "oral vowels" are written with the plural s inside the brackets, so why not for "vowels?" 2604:3D09:2481:7D00:E1F2:F1A3:DABF:5911 (talk) 01:59, 9 September 2024 (UTC)

Ideally they would link to the singular, but it doesn't really matter. Again, if the link ain't broken, don't fix it. — W.andrea (talk) 03:05, 9 September 2024 (UTC)

[[Vowel]]s is preferable. See MOS:PIPESTYLE. Nardog (talk) 04:00, 9 September 2024 (UTC)

MOS:PIPESTYLE doesn't talk about redirects though (like [[Vowels]]). Instead, you could look at Template:R from plural, which says, This redirect link is used for convenience; it is often preferable to add the plural directly after the link (for example, [[link]]s). However, do not replace these redirected links with a simpler link unless the page is updated for another reason (see WP:NOTBROKEN). — W.andrea (talk) 17:19, 9 September 2024 (UTC)

/y/ examples

@W.andrea Please explain why you object to my edits. I have set out my reasons. The claim that RP's GOOSE "is firmly at the back" is very easily disproven.[2] Offa29 (talk) 22:40, 22 September 2024 (UTC)

I already explained: /u:/ is firmly in the back for RP, at least according to the chart at Received Pronunciation#Vowels.

You replied in an edit summary: No it isn't. Just read the caption. A "fairly conservative variety of RP" as of 20 years ago, i.e. an ultra-conservative verging on extinction variety of RP, is not today's mainstream RP. There are many, many, many sources out there that attest to the major and ongoing fronting of GOOSE in RP.

For me, RP is itself "an ultra-conservative verging-on-extinction variety" and I would rather call what's spoken today "Standard Southern British" like Geoff Lindsey. (But that said, I'm not too opinionated or well-read on this.) Either way, RP includes varieties from 20, 40, 60+ years ago, so you'd need to write "modern RP" or something like that. Or you could focus on more progressive varieties like Multicultural London English.

For the other example varieties I removed, I'm not familiar with all of them, but I skimmed their Wikipedia pages and they seem to have varying extents of GOOSE fronting, for example older Californian speakers don't have it at all.

So I'd rather have a shorter list of stronger examples, and if people want to know more, they can click the vowel link (y).

— W.andrea (talk) 23:38, 22 September 2024 (UTC)

First of all re RP: There is indeed a debate on what exactly does and does not constitute "RP", and what exactly to call the current mainstream prestige accent of Southern England.

There are essentially two schools of thought/sets of terminology: those, like Geoff Lindsey, who describe RP as an accent that has died out among all but the most elderly generations (the speakers for whom the vowel chart in the article you looked to would actually be an accurate representation of their vowel system, which is almost exclusively people born before c. 1950), and refer to the current mainstream prestige accent of Southern England as "Standard Southern British".

But there are many other sources out there that view RP as an accent that has simply evolved over time, as all accents do. They describe what Lindsey & co. would simply call RP as "conservative RP" and what Lindsey & co. would call SSB as "modern RP". For example, this paper on GOOSE-fronting in RP from the 1920s to the 2010s.

I don't think their is a clear consensus among linguists which terminology is preferable, as is evidenced by the fact that both are in fairly widespread use. I don't have a super-strong opinion on the matter myself, but would probably, like you, lean towards preferring Lindsey's terminology.

But Wikipedia generally uses the terms "conservative RP" and "modern RP", and does not generally refer to "SSB".

The Standard Southern British article you linked to redirects to a section of the RP article discussing alternative names. If you go, for example, to Wikipedia articles on a particular vowel which list their appearances in various languages and dialects, they will list "conservative RP" and "modern RP" as accents, for example the article for u lists "Conservative Received Pronunciation" and the article for ʉ lists "Modern Received Pronunciation" as accents with those vowels, and its a similar story with the other vowels that differ in realization between Conservative & Modern RP, or between RP & SSB, whichever terminology you prefer.

So my use of "Received Pronunciation" in the examples note was not me using my own personal preference, but simply attempting to conform to the established de facto standard on Wikipedia. I could admittedly have written "modern Received Pronunciation" instead.

I do not know whether using "modern RP" rather than "SSB" is explicit Wikipedia policy, and whether there have been any policy discussions around which is more appropriate to use, or whether the present situation has come about simply through inertia since most of the Wikipedia phonetics pages date from the mid-2000s when the term "SSB" was not in as widespread popular use as it is today. It might be worth having a discussion about whether to change the Wikipedia terminology if there has not already been one, but obviously that would have to be done in the main linguistics policy discussion page. While I would personally lean towards preferring the term SSB, I am doubtful that a consensus to change it on Wikipedia could be found as reliable sources vary and most discussions where this is the case usually end up at "no consensus" and the status quo remaining by default.

It is true that other accents in the London/Southern England area, such as Multicultural London English, are noted for having extremely fronted GOOSE vowels, even more than modern RP/SSB. "London English" is frequently listed as one of the most extreme examples of a fronted GOOSE vowel in general discussions of GOOSE-fronting among worldwide English speakers.

In general, the large majority of speakers in the London/South East England area, be they speakers of SSB, Estuary English, Cockney or MLE, have substantially fronted GOOSE vowels, ranging from [ʉ] to verging on [y] in their onsets. The only major exception is elderly RP speakers. With this in mind, it seems like a major omission that your trimmed list did not include any examples of accents from England at all.

Scouse and Mancunian are two more accents of England that are noted for having a particularly fronted GOOSE vowel, verging on [y] in some Scouse speakers, and in general most accents of England have fronting to a not insignificant extent (Yorkshire English being a notable exception), though it can be hard to track down precise descriptions for every single variety of English English that exists.

In conclusion I think we absolutely need to at the very least list some form of English English as an example, though I am open to discussion on precisely how we word it, whether to write "modern RP", "Standard Southern British", "London English", "Southern England English", or something else, and how far we go in applying it to English English in general, though if we don't than Scouse and Mancunian certainly merit a mention.

Re the other examples and how far we go with them, my instinct is to bear in mind that this is a help page, not an article. This page is aimed at "laypeople" not familiar enough with the IPA to be able to gather the pronunciation of a French word/name from the given IPA transcription alone. Our aim is not to perfectly describe French pronunciation, but to enable them to have some idea of roughly what it sounds like. There are many other examples on this and other languages' IPA Help page that will not be perfectly accurate for most people.

For example, on most articles we list an example of the NURSE vowel (without highlighting the <r>) as the example for /ø(ː)/ when that is certainly not completely accurate for most speakers of English, off the top of my lead only South Wales, New Zealand English and Geordie have actual NURSE vowels close to that, for most English speakers it is something in the [ɜ]~[ə] range. But is deemed an acceptable approximation to at least get the reader roughly close to what the foreign language's phoneme sounds like. (on this article, we list RP burn as the example for /ø/, and i don't know why we specifically list RP as it, conservative or modern, is not even the closest mainstream accent of English to [ø] for a NURSE vowel, and I would favour removing the (RP). But for most articles with an /ø(ː)/ phoneme, we simply list a NURSE word as an example without specifying any accent).

Even a less extremely fronted GOOSE vowel in the [ʉ] region is at least as good an approximation of French /y/ as RP bird is of French /ø/ or most English speakers NURSE vowel is of [ø(ː)] in general.

I think we have to remember that the laypeople who this article is aimed at are not going to be experts on the precise vowel qualities of other accents of English'. If we just list Australian, that is not much help to a layperson speaker of any accent of English other than Australian English, who will not be able to conjure up a mental image, well sound, of an Australian English GOOSE vowel and how/whether it differs from their own.

We need to give the laypeople reading the article the best possible chance of being able to determine whether examples that apply only to "some dialects" can reasonably be said to apply to their own dialect or not, as that is the only piece of information that will really be helpful to them. This page is not aimed at experts who are familiar with the qualities of the GOOSE vowel in other dialects of English, as they will be IPA-literate enough to know what [y] sounds like without needing an example.

To that end, I think that if an accent has a GOOSE vowel that is a somewhat acceptable approximation of French /y/, then listing it as an example is much more helpful to speakers of that dialect then just telling them that "it's like Australian English GOOSE", because they will not be able to picture what that sounds like and how/whether it is different from their own GOOSE vowel.

My criteria for determining whether a particular accent's GOOSE vowel is an acceptable approximation of French /y/ is whether they would perceive their own GOOSE vowel as being closer to French /y/ then French /u/. If they do, then it is helpful to indicate to them that their own GOOSE vowel is "somewhat like" French /y/ and that they can use it as a rough approximation.

Of course, exactly which dialects do or don't meet my criteria is not completely straight forward to determine, but it seems likely to me that the accents I listed do, and the accents I listed as counterexamples do not, though of course I am not a speaker of most of them.

For what it is worth, I was once told in an online discussion by a speaker of Western Pennsylvania English that they, upon hearing samples of French /y/ and /u/, perceived their GOOSE vowel as closer to French /y/. And my own native dialect is what you and Geoff Lindsey would call Standard Southern British, and I certainly perceive my own GOOSE vowel as closer to French /y/ than French /u/.

I do not think we need to list a "main example" as our aim is to make the advice on the page apply to as many speakers of English as possible, and Australian English is not even the most extreme example of a fronted GOOSE vowel. Offa29 (talk) 12:17, 23 September 2024 (UTC)

I'll reply more shortly, but for a start, adding some England varieties is fine with me, like MLE, Mancunian, and Scouse. — W.andrea (talk) 20:48, 23 September 2024 (UTC)

for example the article for u lists "Conservative Received Pronunciation"

Fact check: it just says "Received Pronunciation". — W.andrea (talk) 20:56, 23 September 2024 (UTC)

I've added just the three England examples you gave for now, though I maintain that most accents of Southeastern England (with the exception of Conservative RP), i.e. Estuary, Modern RP/SSB, Cockney in addition, have GOOSE vowels fronted enough to merit inclusion, but I await further discussion on this. Offa29 (talk) 22:49, 23 September 2024 (UTC)

If we just list Australian

I'd just like to point out that I'm not arguing for that. I didn't remove your list entirely, just trimmed it and moved Australian back out of the footnote. — W.andrea (talk) 20:14, 24 September 2024 (UTC)

Our aim is not to perfectly describe French pronunciation

I'm also not arguing for that. — W.andrea (talk) 20:23, 24 September 2024 (UTC)

that is the only piece of information that will really be helpful to them

I think this my main hangup to your argument. I'm one of those speakers who doesn't have a close equivalent, so for me, if I didn't have linguistics knowledge, reading that big list would just be a waste of time. Instead, I've been working on alternatives like adding at least a rough approximation (/ju:/) as well as an equivalent sound from another language (German) and next, adding audio.

Now that I've added some other options, if you want to go ahead and add more varieties (especially British), by all means. If I could just suggest, you might want to put "younger speakers", or something like that, for some of them.

— W.andrea (talk) 20:50, 24 September 2024 (UTC)

I have added back examples I feel there is sufficient evidence for, though I may make further alterations. I added the caveats of "modern RP" and "younger Californian". I considered making it "modern (but not conservative) RP" but decided that was too verbose. If you want to quibble with any of the reinstated examples then go ahead, but I will just note here and now than William Labov et al's Atlas of North American English pages 153-154 have data formant data showing the mean F2 of the GOOSE vowel (/uw/ in the book's terminology) to be fractionally higher (i.e. fronter) in the Midland than in the Southern region, implying that GOOSE-fronting in the Midland is on par with, or even slightly ahead of, that of the South. So I really can't see any reason to include Southern American English but not Midland American English. The Midland region did show a higher effect of a coronal onset on fronting, but the given example of too has a coronal onset! Offa29 (talk) 01:13, 26 September 2024 (UTC)

I concur with illustrating [u] with a word where a fully back allophone is commonly found, but I think it's a tall order for those uninitiated in phonetics (i.e. the very people who could use the approximation) to grasp the difference between co-allophones of the same phoneme that only subtly differ in backness. So I suggest we illustrate [y] with /juː/ (which I believe we used to at some point), though naming in addition some prominent accents in which /uː/ is close to it (e.g. California) couldn't hurt. Nardog (talk) 00:09, 24 September 2024 (UTC)

We certainly do use /juː/ as an example for some other languages that feature /y(ː)/ as a phoneme, for example Help:IPA/Standard German. While it does better convey that [y] is distinct from [u], and can apply to all accents of English, which is something I think we should strive to do whenever possible, I would query whether giving /juː/ as an example is an appropriate, given that it might indicate to the uninitiated reader that a yod is actually present at the start of French /y/, and that this is what distinguishes it from /u/, which it obviously is not. (The fact that the symbol is <y>, the letter they associate with the sound /j/, would not help). It's subjective and impressionistic, but I would question whether adding a yod which isn't there in French is a good approximation, and whether it is preferable to struggling to distinguish /y/ and /u/.

The help page for Help:IPA/Hungarian attempts to get round this by giving an English /juː/ example but then explicitly adding but without the initial 'y' sound. But if we do this, we're back to the problem you raised in the first place, of how they will distinguish it from the example given for [u].

But there is clearly inconsistency across the different language IPA help pages on how to approximate different languages phones/phonemes. Might it be a good idea to have a central policy discussion to agree on some guidelines, if not hard-and-fast rules, for what sort of English examples to give for certain phone(me)s Granted, phonemes transcribed with the same phonemic symbol across different languages will not always sound exactly the same, and there may be other considerations like how to best illustrate the contrast with other phonemes in a particular language, which will vary across different languages, but it still might be a useful starting point rather than having to have a bunch of discussions on the Talk pages for specific languages when there should still be a lot of commonality between the examples.

This doesn't apply to French, but one common practice on the IPA Help pages that I take issue with is the use of pre-fortis clipping to illustrate phonemic vowel length, by giving a pre-fortis English vowel as the example for a "short" vowel and a pre-lenis or word-final example for the equivalent "long" vowel. The Help:IPA/Danish page is particularly rife with this, for example beat for [i] and bead for [iː]. I think it is asking an awful lot for an ininitiated reader to notice that their vowel in beat is shorter than their vowel in bead, in fact I think it would be even harder for them to notice this that it would be for them to notice the difference between the fronted vs. back allophones of /uː/ that you raised, and it would be much better in these situations to just explicitly tell them that one vowel is "short" and the other is "long", for example "bee (short)" for [i] and "bee (long)" for [iː]. But it would be easier to have this discussion on a central policy Talk page rather than having to have it individually on every page that uses pre-fortis clipping to illustrate vowel length, which is a lot. Offa29 (talk) 10:54, 24 September 2024 (UTC)

I suggest we illustrate [y] with /juː/

I went ahead and did that. I also added a German equivalent.

(which I believe we used to at some point)

A bunch of other Help:IPA pages do, like Help:IPA/Finnish for example:

y   kesy   somewhat like cute

— W.andrea (talk) 19:26, 24 September 2024 (UTC)

Adding audio

I've drawn up a draft of the page with audio clips added: User:W.andrea/sandbox/Help:IPA/French. I made some additional changes too — details in the revision history.

I'm posting here to see if there are any major objections before I publish. For example, I'm not a native French speaker and I'm not sure if I picked the best examples for /œ/ vs /ø/.

I tried to put a canonical example first for each phoneme (e.g. /ɛː/ [ɛː] ) but also include variations (e.g. Quebec [~ɛɪ̯] ).

If you want to look at alternative clips, you can use Wiktionnaire, e.g. pain § Prononciation.

One of my reasons for doing this is to decrease the concern about accurate English approximations. For example, if most people reading the page can listen to clips of /y/, then we don't need to worry as much about which accents in English have a good approximation.

— W.andrea (talk) 18:01, 24 September 2024 (UTC)

It's also worth noting that the nasal vowels vary so widely between Europe and Quebec that they can overlap, e.g.

• (Europe) and (Quebec) are both [~ɒ̃]
• (Europe) and (Quebec) are both [~æ̃]

— W.andrea (talk) 18:15, 24 September 2024 (UTC)