Help talk:IPA/English/Archive 27

Archive 20

←

Archive 25

Archive 26

Archive 27

Archive 28

Inaccurate description of rhotic dialects

As a speaker of a common North American rhotic dialect, I don't find this to be true at all:

> Speakers of rhotic dialects (Irish English, North American English, Scottish English) do not distinguish between the vowels of near /ˈnɪər/, cure /ˈkjʊər/ and square /ˈskwɛər/ on the one hand and freerunning /ˈfriːrʌnɪŋ/, Q-rating /ˈkjuːreɪtɪŋ/ and dayroom /ˈdeɪruːm/ on the other. If you speak such a dialect, read /ɪər, ʊər, ɛər/ as /iːr, uːr, eɪr/. Tim314 (talk) 17:03, 24 February 2022 (UTC)

Please elaborate. Sol505000 (talk) 17:16, 24 February 2022 (UTC)

For example, I pronounce "cure" identically to the start of the word "curating", but I don't pronounce "curating" the same as "Q-rating". I'm not expert on IPA symbols, but "Q-rating" starts with a "K" something like the word "you" (a "Y" sound and then an "oo" sound sound like "boot"), whereas cure/curating start with a "K" sound followed by a "Y" sound and then something like the "er" sound in "her". There's none of that "oo" sound in how I say "cure". - 17:29, 24 February 2022 (UTC) — Preceding unsigned comment added by Tim314 (talk • contribs)

Edit: I mean to say "a K *followed by* the word "you" — Preceding unsigned comment added by Tim314 (talk • contribs) 17:32, 24 February 2022 (UTC)

That's the cure-nurse merger. The note could use a rewording. Sol505000 (talk) 17:37, 24 February 2022 (UTC)

I believe Tim314 may be making a larger point that the entire statement on the page rings false. For example, in American English anyway, day room /ˈdeɪruːm/ and dare room /ˈdɛər(r)uːm/ are not phonemically equivalent; nor are Q-rating /ˈkjuːreɪtɪŋ/ and curating /ˈkjʊəreɪtɪŋ/; nor potentially knee ring /ˈniːrɪŋ/ and nearing /ˈnɪərɪŋ/. Wolfdog (talk) 02:10, 25 February 2022 (UTC)

Yes, that's what I was trying to say. Thanks, Wolfdog. - Tim314 (talk) 04:09, 25 February 2022 (UTC)

An oft-cited pair is key-ring vs fearing. I agree the statement is misleading at best as far as NAmE is concerned. Just because there are no tautosyllabic /iːr/ etc. doesn't mean the speakers can't distinguish between /iː.r/ and /ɪr/. I don't know about Irish and Scottish accents though. Nardog (talk) 09:56, 25 February 2022 (UTC)

I think the note is trying to say that the normal English vowels occur before /r/, rather than having special rhotic vowels. But the examples all suffer from the syllabification issue you get with compounds, similar to the aspiration differences in /Vs+kV/ vs /V+skV/. — kwami (talk) 06:10, 25 February 2022 (UTC)

Perhaps this note is trying to (clumsily) refer to the phenomenon of NEAR-tensing, described by Professor Lindsey here:- https://youtube/5OULnCCvdk8 (just search for ‘English after RP with Dr Geoff Lindsey’ on Simon Roper’s YT channel if the link doesn’t work) Overlordnat1 (talk) 12:00, 25 February 2022 (UTC)

No, RP is not rhotic. Nardog (talk) 12:03, 25 February 2022 (UTC)

That’s true - a better example of NEAR-tensing would perhaps be nearer and freerunning, though I don’t think rhoticity would be an important factor to consider in that case. The idea that anyone pronounces the first syllable of dayroom like square apart from some non-rhotic people from Yorkshire seems rather fanciful though. Overlordnat1 (talk) 12:28, 25 February 2022 (UTC)

Nearer isn't pronounced [ˈniːɹə ~ ˈnɪiɹə] in any type of Southern English (AFAIK), including SSB/RP. It always has a vowel closer to cardinal [e], though this monophthong is conventionally transcribed with ⟨ɪː⟩: [ˈnɪːɹə]. [ˈniːɹə] is Australian, or Northern English/Welsh (if we're talking about non-rhotic accents alone). In Australia, monophthongal NEAR is as high as KIT, also in other positions. The transcription ⟨iː⟩ for FLEECE, which we use on Australian English phonology is, frankly, monumentally stupid. FLEECE is clearly a diphthong, it's NEAR that is a long close front monophthong (with or without the centering glide). Sol505000 (talk) 13:18, 25 February 2022 (UTC)

As Dr Lindsey says during the conversation about NEAR-tensing (approximately 29:00 to 39:00 into the video) the Southern band Coldplay have a song where ‘tear’ (not as in ‘rip’ but as in the liquid that is produced by tear ducts) is pronounced with two syllables but ‘teardrop’ is pronounced ‘tea drop’, so such pronunciations are indeed possible in the South. I agree with you and Dr Lindsey that FLEECE should really be thought of as flijs, so with a glide, arguably even a diphthong. Overlordnat1 (talk) 14:15, 25 February 2022 (UTC)

Not "tea drop". [ˈtɪːdɹɒp] - or [ˈteːdɹɒp], to be exact. The vowel is considerably lower than [i] and it's monophthongal, so that it wouldn't be interpreted as FLEECE by any native speaker from that area. [ˈtiːdɹɒp] is Australian (and not "tea drop" [ˈtɪjdɹɒp] either) and probably doesn't occur anywhere in the UK. Speakers of Northern England English and Welsh English would say [ˈtiːədɹɒp ~ ˈtɪjədɹɒp ~ ˈtɪːdɹɒp] but not [ˈtiːdɹɒp], let alone [ˈtɪjdɹɒp]. Sol505000 (talk) 14:24, 25 February 2022 (UTC)

OK, so I was a bit imprecise with my notation but saying ‘tear’ with either one or two syllables, whether on its own or as a syllable in a longer word, is possible throughout the U.K despite the standard IPA symbols used to reflect an RP pronunciation suggesting this is impossible. This point could do with being mentioned somewhere in the article. Overlordnat1 (talk) 15:45, 25 February 2022 (UTC)

kwami Indeed, the syllable break made the example not work. I'm still learning the IPA symbols, but I believe I say something like "near" \'niɚ\, "cure" \'kjɚ\, and "square" \ˈskweɚ\ for the single-syllable words (not so sure if \ɚ\ is right), but I say "free-running" \ˈfri:.rə.nɪŋ\, "Q-rating" \ˈkju:.reɪ.tɪŋ\, and "dayroom" \'deɪ.ru:m\ for the compounds, with the first syllable just the same as in "free" \ˈfri:\, "Q" (or "cue") \ˈkju:\, and "day" \'deɪ\. It looks like the latest edit removes the comparison between these words, which is probably for the best. Tim314 (talk) 21:35, 25 February 2022 (UTC)

Have we agreed that the sentence, as it pertains to rhotic accents, ought to be removed? I don't think it's useful to push hypothetical information like this unless we have a source solidly affirming it. Wolfdog (talk) 16:08, 25 February 2022 (UTC)

No, it needs to be rewritten. It's not "hypothetical information" to claim that the phonemes /ɪə, ʊə, ɛə/ occur only in non-rhotic dialects. Their historical rhotic counterparts are /iːr, uːr, eɪr/. In that sense the sentence is correct, it just needs to be reworded to account for e.g. the mary-merry merger, or the aforementioned cure-force merger. Sol505000 (talk) 16:19, 25 February 2022 (UTC)

@Sol505000: Oops, so sorry for the "bullshit". Maybe your insufferable highness can advise on some better wording. I'm done collaborating with you, where every conversation turns into a rude attack. Enjoy your petty victory. Wolfdog (talk) 16:52, 25 February 2022 (UTC)

What are you talking about? I thought Sol505000 had perhaps called your input "bullshit" and then deleted it after you took offense -- I was going to restore it (struck out) so that your response would make sense in context, but I can't find anything like that. — kwami (talk) 21:57, 25 February 2022 (UTC)

Shit, sorry for that. I had a tough day. I hope you'll forgive me. Sol505000 (talk) 19:13, 27 February 2022 (UTC)

It is of course a truly bizarre claim as well as an unsourced one, dayroom and square most certainly do not have the same vowel in anything like most rhotic accents (or non-rhotic for that matter) and when they do the vowel sound isn’t eɪ but i: in Scotland and e: in Ireland. The alleged full merger between ʊ and u: in Scotland and Northern Ireland is also wrong - it’s closer to being true in Scotland but some Scots (Sean Connery is one example) do in fact say things like lʊk for look instead of the standard Scots lu:k Overlordnat1 (talk) 23:42, 9 March 2022 (UTC)

@Overlordnat1: It's not about saying [lʊk] instead of [lʏk] but consistently distinguishing between the two vowels in a way that speakers of RP or GA do. In RP and GA, the distinction between /ʊ/ and /uː/ is phonemic in minimal pairs such as pulling /ˈpʊlɪŋ/ vs. pooling /ˈpuːlɪŋ/. That's the real test that gives away the merger, or lack thereof. In addition, pronunciations such as [ˈmʊd] for mood also give away the merger (especially if the [ʊ] is too short, as it is in Scotland before any stop within the same morpheme).

[lʏk] is a perfectly native SSB/Estuary pronunciation. The Scottish FOOT/GOOSE vowel is often not fully close, and so when it's short, it's likely to be heard as /ʊ/. /ʊ/ isn't a very common vowel in English (I mean there aren't that many words in which it occurs) and it can only occur before a consonant. I think the merger goes unnoticed in many words. Sol505000 (talk) 12:41, 24 March 2022 (UTC)

Minimal pairs like pulling and pooling and look and Luke definitely exist in Scotland but I’m not convinced a vowel merger is possible in ALL environments. The merger is nearly always a result of pronouncing /ʊ/ as /uː/ (or perhaps /ʏ/) rather than the other way around. I’ve never heard tooth pronounced as tʊth by a Scotsman, for example (though his pronunciation is very widespread, possibly even dominant, in Birmingham and South East Wales and can occur in Northern Ireland). I did find the pronunciation of juice as jʊs by both Scotswomen in this video [1] (‘Scottish Accent Comparison’ by YT user ‘Poison Pixie’ if the link doesn’t work) interesting though. Overlordnat1 (talk) 13:33, 24 March 2022 (UTC)

Minimal pairs like pulling and pooling and look and Luke definitely exist in Scotland Not according to Accents of English and other sources. The distinction, if it occurs at all, is haphazard and therefore not phonemic. RP and GA both feature a consistent distinction between pulling and pooling as well as look and Luke and other minimal pairs (and there aren't many) among all speakers. Sol505000 (talk) 13:38, 24 March 2022 (UTC)

I didn’t mean to suggest that these pairs are anything but homophones for a clear majority of speakers just that they are sometimes distinguished, so apologies for my imprecision and any confusion caused, but the way the article is written is a bit unclear, as it suggests that it’s possible that Scottish people could say things like tʊth instead of tuːth for tooth and it ignores the fact that the merger of the two sounds is in any case much rarer in Northern Ireland than in Scotland. Overlordnat1 (talk) 14:36, 24 March 2022 (UTC)

Poilievre

I know this is most likely neither here nor there, but I wonder whether the last name of Canadian politician Pierre Poilievre, which is reportedly pronounced "paw-li-ver" or "paul-ee-EV", is best represented diaphonemically with /ɔː/, /ɒ/, or /ɑː/. Given it corresponds to ⟨oi⟩ it might come from [wa~wɑ], in which case /ɑː/ seems appropriate—except, if [w] was elided because of the rounded quality of THOUGHT/LOT/PALM in Canadian English, is it really appropriate? It might also come from the realization of ⟨oi⟩ in whatever variety of Canadian French, which could be any of [we, wɛ, wɛː, wæ, wɔ, waj, wej, ɛ, e, ɔ] and more according to Walker. Alternatively one could ask what a speaker who maintains the contrasts is most likely to produce, but I don't know what that would be; those aware of the French pronunciation would probably interpret it as omitting /w/, which suggests /ɑː/, whereas others might interpret it as ignoring ⟨i⟩, which points to /ɒ/. And if there's a word of French origin where ⟨oi⟩ is pronounced as any of the three, that can settle it, but I can't think of one. Nardog (talk) 23:36, 23 March 2022 (UTC)

Same goes for pho. /fɑː/ doesn't even look like the father-bother merger as LOT doesn't occur in the word-final position. I'd change it to /fɔː/. The Canadian /ɒ/ sounds basically like your typical Northern English/Welsh THOUGHT [ɒː], except when it is short enough to be heard as LOT by speakers of those dialects. When in doubt, we could choose /ɔː/ as it is a free vowel and it doesn't look like transcribing the father-bother merger. On phonetic grounds alone, ⟨ɑː⟩ might be the worst choice out of the three as [ɒː] for /ɑː/ is ridiculed as "using /ɔː/ for /ɑː/", as in Broad South African mawk my words [ˈmɒːk mɑɪ ˈwøːts] regardless of whether an actual merger occurs (in South Africa, THOUGHT is an unremarkable [oː] also found in Southern England, Australia and New Zealand). In dialects with low THOUGHT, PALM is typically [aː], central or even front. I think Irish English might be similar in that regard, as long as the three vowels are distinguished.

I think that Poilievre would be simply /ˌpwɑːliˈɛv(rə)/ in other dialects. That points to /ˌpɑːliˈɛv/. Or, we could write it /ˌpɔːliˈɛv/ for the reasons explained above. Sol505000 (talk) 12:03, 24 March 2022 (UTC)

I'm at a complete loss as to what the stigma against certain realizations in South Africa has to do with this, but in any case I'm leaning towards /ɔː/ as well. As I said, I find it likely that [w] was elided due to the rounded quality of the merged vowel in Canadian English causing the onset to labialize anyway as a coarticulation effect. Nardog (talk) 07:25, 30 March 2022 (UTC)

The footnote on vowels before r

Footnote 13 currently reads "In non-rhotic accents like RP, /r/ is not pronounced unless followed by a vowel." It seems to me that this should be followed by "In rhotic accents, the preceding vowel is not pronounced long." The ":" after the vowels in the IPA renderings for "start", "force", and "nurse" confused me until I remembered the case of non-rhotic English. Largoplazo (talk) 12:23, 4 September 2022 (UTC)

See the note at the end of the section "Key". Just because a vowel is transcribed with ⟨ː⟩ doesn't mean it's invariably long. And quantitative vowel contrasts and rhoticity are not mutually exclusive. Hiberno-English, for example, is rhotic but maintains the contrast between /ɜːr/ and /ər/. Nardog (talk) 12:30, 4 September 2022 (UTC)

Ah, thanks. Largoplazo (talk) 12:38, 4 September 2022 (UTC)

Is sang a good example of the æ vowel sound?

Maybe I've been pronouncing it wrong? 96.61.176.120 (talk) 11:28, 28 September 2022 (UTC)

The diaphoneme is /æ/, but in General American the vowel is actually much closer to [eɪ] ("sayng", like "saying" but without the "i", i.e. as one syllable). But there are tens of millions of speakers of English (from the United Kingdom and Ireland, for example) who have the same vowel quality in "sang" as in "trap", "mat" etc. Sol505000 (talk) 14:46, 28 September 2022 (UTC)

Edit warring over the final schwa

Michael Bednarek (talk · contribs) is edit warring with me over the transcription of the final schwa, which diaphonemically is /ə/. Caloundra is clearly neither /kəˈlaʊndrʌ/ ([2]) nor /kəˈlaʊndrɑː/ ([3]). What is going on is a clear interference from German (which I assume is their native language) in which /ə/ contrasts with the short allophone of /aː/ (which is phonetically the same as /ər/, conventionally transcribed with ⟨ɐ⟩) even in the word-final position. English features NO such contrast among the short vowels, with /æ/ being confined to the word-internal position. In addition, the symbol ⟨ɑ⟩, with or without the length mark, is invariably used for the long/free vowel in spa, start etc. in IPA transcriptions of English, at least as far as dictionaries are concerned, whereas /ʌ/ cannot appear in the word-final position (just like /æ/) as it stems from the earlier /ʊ/ (speakers from, say, Liverpool or Newcastle upon Tyne would not say /kəˈlaʊndrʊ/!). The vowel is phonemically /ə/ regardless of the actual realization - especially if we're talking about diaphonemes.

I already tried discussing the issue, but to no avail. Sol505000 (talk) 14:55, 28 September 2022 (UTC)

I understand that IPA transcriptions are a "standardized representation of speech sounds", so the remark "regardless of the actual realization" strikes me as very odd. IPA transcriptions should give readers an indication how the name is actually pronounced where it's not obvious. Many Australian placenames have roots in Australian indigenous languages, where "/ʌ/ cannot appear in the word-final position' doesn't apply. In the case of Caloundra, the root of its final syllable is the Gubbi Gubbi word 'dha', so it's clearly not a schwa, as claimed. The misleading IPA transcription introduced by Sol505000 ought to be reverted. -- Michael Bednarek (talk) 03:13, 29 September 2022 (UTC)

I have already provided a citation that proves that speakers of Australian English cannot distinguish between /ə/ and /ʌ/ in the final position (and neither can speakers of other major dialects of English). ANY phrase-final (not necessarily word-final) /ə/ in English as pronounced by an Australian, a New Zealander or some Londoners will likely sound like the /ʌ/ of strut, so that the word order as pronounced by speakers of those dialects will sound like the German word oder [ˈoːdɐ]. The distinction between the final vowel in Caloundra and Yandina on the one hand and the one in sofa and letter on the other that you claimed does not exist. Phonemically, it is /ə/, which follows the historical distribution. Any transcription that links to this guide must respect that, due to the diaphonemic nature of this guide (meaning: the fact that it covers multiple dialects at once).

Also, the final schwa in Caloundra and Yandina triggers the linking R just like the one in sofa and letter. The phrases "Caloundra is" [kəˈlæɔndɹə.ɹɪz], "Yandina is" [jænˈdiːnə.ɹɪz] and "a sofa is" [əˈsəʉfə.ɹɪz] will be pronounced with an added [ɹ] before the initial [ɪ] of is just like in the phrase "a letter is" [əˈleɾə.ɹɪz], or subpoenaing [səˈpiːnəɹɪŋ], which is pronounced like "subpoena ring". This is true of most Australians and New Zealanders as well as most people from England and Wales. Note that the schwa is not lowered in this case. To never use the linking R in those contexts would be non-local (South African, for example). And /ʌ/ never triggers the linking R, it can only follow a prevocalic /r/ within the same morpheme (as in hurry /ˈhʌri/, which has the same /ˈCVCV/ pattern as honey /ˈhʌni/), just like the other checked vowels /ɪ, ɛ, æ, ɒ/. Sol505000 (talk) 07:09, 29 September 2022 (UTC)

Perhaps you should acquaint yourself with the nature of phonetic transcription, the difference between phonemic and allphonic transcription, what a phoneme is, and what complementary distribution is, by reading our articles, online courses, or books like the Handbook of the International Phonetic Association and A Course in Phonetics. Or at least you should look up words you think sound similar in dictionaries that document pronunciation and see if any of them ends with /ʌ/ (which is a checked vowel). Nardog (talk) 09:29, 29 September 2022 (UTC)

I agree that phonemic transcriptions are a pain in the backside but I don’t see how you can rightly say that the schwa and CUT/STRUT vowel are essentially in free variation for unstressed syllables in many accents, including ones with an intrusive R, and then go on to say that a CUT vowel is never used before the R in words like ‘subpoenaing’. I’m sure I’ve heard many people say it that way, in fact. --Overlordnat1 (talk) 12:30, 30 September 2022 (UTC)

I said neither of those things. Please don't put words in other people's mouths. Nardog (talk) 12:34, 30 September 2022 (UTC)

My comment was aimed at Sol505000 - I should’ve made that clear. --Overlordnat1 (talk) 12:54, 30 September 2022 (UTC)

I said that they are in a free variation in the word-final position. Not any other position (to my knowledge, intrusive/linking R blocks schwa-lowering like any other following consonant).

Also, the final vowels in Caloundra, Yandina etc. vary in height between [ɐ] and [ə] unlike STRUT. No Australian, to my knowledge, says [stɹət], rather than [stɹɐt]. A phonetic [ə] doesn't exist as a stressed vowel in AuE. It's heard as a variety (a non-local variety) of STRUT when it is used in that context, hence fush and chups. Sol505000 (talk) 13:02, 30 September 2022 (UTC)

I see what you mean but I would say that such final syllables can easily be pronounced as the STRUT vowel by many British people, even before an intrusive R. Also schwa is indeed used by many people where others would use a CUT or PUT vowel, even in stressed positions, but not in RP or Australian English. Overlordnat1 (talk) 14:27, 30 September 2022 (UTC)

Maybe. Still, Londoners and speakers of RP would not use [ə] in stressed syllables (so [stɹɐt], [stɹʌt] etc. disregarding the variation of consonants of course). Schwa in this context is used in Wales, (some?) West Country English, in Birmingham (when it doesn't merge with /ʊ/, which it can) and some East Anglian English. But this is entirely off-topic, the WP:DIAPHONEMIC system used in this guide does assume a phonemic difference between unequal and an equal, a difference that only occurs in the word-internal (meaning: non-word-final, so word-initial as well) position. The fact that a stressed /ʌ/ cannot end an English word means that it is a checked vowel like /ɪ, ʊ, ɛ, ɒ, æ/. And since /ə/ is an unstressed-only vowel (and one that can end a word, but that's circular reasoning of course), the phonemic identification is very easy: Caloundra and Yandina end in /ə/, not in /ʌ/. Also, bear in mind that a final /ə/ can also be lowered in accents without the foot-strut split. A Geordie would say [kəˈlæʊndɹɐ, janˈdiːnɐ] etc. while also saying [stɹʊt, ɹʊn, kʊp] etc. Sol505000 (talk) 14:34, 30 September 2022 (UTC)

I agree with much of that but all you have to do is go onto YouTube and listen to the Welsh singer Cerys Matthews sing her lines in the song The Ballad of Tom Jones that she recorded with the band Space and you’ll hear the clearest example possible of a schwa being pronounced as a CUT vowel when she says the word other with an identical vowel in both syllables. --Overlordnat1 (talk) 21:15, 30 September 2022 (UTC)

By schwa being pronounced as a CUT vowel I guess you mean phonemic /ə/ being pronounced differently from phonetic [ə]? "CUT vowel" is an abstraction just like /ʌ/ and can mean many different things depending on accent/speaker. In any case, your example very nicely illustrates Geoff Lindsey's point that is echoed by Sol505000 here: many (but not all) varieties of English use exactly the same quality for stressed /ʌ/ and unstressed /ə/.

To the editor who originally raised the whole issue, I can only say: is the final vowel of Caloundra in any way different of that in Canberra or Dora the Explorer? Whatever their phonetic realization may be, phonemically they are all /-ə#/ (as a merger of diaphonemic /-ə#/ and /-ər#/). And FWIW, phonetically, it's far from being a "real" [ʌ]. Aussie low vowels are never back. –Austronesier (talk) 21:45, 30 September 2022 (UTC)

Yes, the fact that they are the same phonemically doesn't mean that they are phonetically [ə]. Other can be [ˈɐðɐ] (or maybe [ˈɜðɜ], with rather less lowering) in all places in the UK that have the strut-schwa merger. This is called allophonic range. It shows that [ɐ] for final /ə/ is entirely unremarkable in most if not all major non-rhotic Englishes. An extreme lowering (all the way to [ä]) does sound regional, mostly cockney/Australian I believe. RP used to have it as well, but it has recently become regionally marked (the problem with that description is that it can be found in most parts of the UK - some accents have it, some don't. Scouse has fronting instead: [ˈʊðɛ]).

Still, the most local Welsh pronunciation in stressed syllables is [ə], or even a centralized [ɛ]. The latter used to render some words almost unintelligible for me, I really had to get used to it. It's only subtly different from DRESS (same with cockney /ʌ/, which sounds almost like /æ/. And since I can't hear the bad-lad split...) Sol505000 (talk) 23:52, 30 September 2022 (UTC)

My concern is how an IPA transcription can provide useful information to readers, i.e. how a word or placename is pronounced. I've been introduced to IPA when I was 13, and I get a good sense of a word's pronunciation from IPA (except that some Wikipedia entries contain unhelpful ephemeral and confusing details). I have read IPA transcriptions in various dictionaries, and of opera librettos in various languages, and I think I get a reasonable approximation of the sound so that a native speaker would understand me. I understand that in general, English words don't end in a short /a/, but these Australian place names are not strictly English – they are based on indigenous words and have retained parts of that phonology. 'Caloundra', and many place names like it, do not end in a schwa but in clearly enunciated full vowel /ɑ/ or /ʌ/. I dispute Sol505000's assertion that in a phrase like "Caloundra is" a linking R is triggered. Instead, a glottal stop between the words is used. I don't understand the logic of "ANY phrase-final ... /ə/ in English ... will likely sound like the /ʌ/ of strut" and then transcribing it as /ə/ – which also contradicts your assertion that "speakers of Australian English cannot distinguish between /ə/ and /ʌ/ in the final position". What's the point of IPA if not conveying pronunciation? Aside: abominations like "subpoena ring" or "law-r-and order" should be spurned.

Another example familiar to most Australians is Pitjantjatjara. Rules that English words cannot end in /a/ don't apply to indigenous words. Allowances have to be made for non-native words in an otherwise English-language environment. In short, maybe the template {{IPAc-en}} is unsuited in these cases and another should be used. But maybe I misunderstand the purpose of IPA transcriptions in Wikipedia; following the discussion above, their purpose seems decidedly not to provide "standardized representation of speech sounds" but a theoretical classification of sounds, disregarding actual realization. -- Michael Bednarek (talk) 02:36, 1 October 2022 (UTC)

I understand that in general, English words don't end in a short /a/, but these Australian place names are not strictly English – they are based on indigenous words and have retained parts of that phonology. 'Caloundra', and many place names like it, do not end in a schwa but in clearly enunciated full vowel /ɑ/ or /ʌ/. Check any dictionary and see whether such sounds are transcribed with any other symbol than ⟨ə⟩. You clearly don't understand how the English schwa works. Not to mention that on the recordings you've provided, the vowel varies from [ɐ] to [ə], exactly according to how Cox & Fletcher (those that you offhandedly dismissed as "theoretical linguists") describe the final schwa in AuE. Therefore, it is neither "clearly enunciated" nor "full" (and therefore absolutely not /ɑː/, which would be a long [ɐː]), it is variable because it is reduced. A clearly enunciated A in this position would belong to the /ɑː/ phoneme, which would be an invariably long (slightly clipped but not that much), open vowel (central [ɐː] in Australia and many other dialects, back [ɑː] in other dialects) as in spa. This would require treating Caloundra and Yandina as compound nouns Calound-Rar and Yandee-Nar, which they aren't (another problem is that the -Rar part would probably shift the main stress over to the last syllable).

So what I said about the interference from German is likely true. You hear Caloundra and Yandina as featuring a final full vowel [ä] because this belongs to the /aː/ phoneme in German, as if they were German words/neologisms Kelaundra and Jendina. The problem with that is that a final short [ä] is phonemically /ə/ in English, especially if it varies with [ə] (an Australian would not hear the difference between the German neologisms Kelaundra and Kelaundre, in fact native speakers of English are notoriously bad at reproducing the two schwas of German, which means that they can't distinguish gute from guter, which yields many grammatical errors in their speech). ⟨a⟩ and ⟨er⟩ is how you spell that vowel in English, and I think that all final schwas that are spelled with ⟨a⟩ are restricted to loanwords anyway (sofa etc.). This is where you get all the mute E's from in spelling, those used to be schwas. A long [äː] is phonemically /ɑː/. No other phoneme is possible in this context (you can substitute [ɐ] and [ɐː] for [ä] and [äː]. It's a slight difference in vowel height that isn't in any way meaningful in English... nor is it in German, with Oper being homophonous with Opa).

I dispute Sol505000's assertion that in a phrase like "Caloundra is" a linking R is triggered. Instead, a glottal stop between the words is used. No, it's not, at least not invariably. You're describing an artificial, comma-letter unmerged (thought-north unmerged, in the case of the phrase law and order) pronunciation that wasn't even adhered to in England when the avoidance of intrusive R was still heavily prescribed: [4], [5]. Australian English is basically old-fashioned cockney with slight differences among especially consonants, with the "intrusive" R (a misnomer for AuE, since there is a full comma-letter merger) not being one of them. A glottal stop can be used in both tuner amp and tuna oil, this doesn't mean that tuner is phonemically distinct from tuna (it only is when you keep them unmerged in all relevant contexts, as John Wells claims to do in his speech. Let me just say that a lack of merger OUTSIDE of the citation form (where the merger is present) strikes to me as very, very odd. It's like Lindsey said: it's suppressing unconscious allophonic rules of your dialect).

Another example familiar to most Australians is Pitjantjatjara. Rules that English words cannot end in /a/ don't apply to indigenous words. Yet Bauer 2007 (cited in that article) gives a transcription with a final schwa. Strange. Sol505000 (talk) 08:25, 1 October 2022 (UTC)

Again, it is simple: is the final vowel of Pitjantjatjara or Caloundra in any way different from the one in Canberra, Dora or explorer? Only if it is different from these in the context of an Australian accent, we will use a vowel other than /ə/. And well, it is the same vowel in all of these. It is true, we don't reproduce actual pronunciations (which is impossible given the latitude of different pronunciations of the same phoneme across various accents), but provide a transcription that will predictably produce the actual pronunciation in every standard accent, including the local one.

Without context, you couldn't even properly transcribe the first vowel in Melbourne. Only if you hear an Australian speaker say bell/hell against bill/hill, you will be able to safely assign the vowel of Melbourne (which pretty much sounds like the realization of /ɪ/ in most AE and BrE accents) to the first group. –Austronesier (talk) 09:52, 1 October 2022 (UTC)

I’ve occasionally heard Australians say ‘e’ as ‘i’ but that trend seems to be far more prevalent in New Zealand. Due to the Salary–celery merger (English-language vowel changes before historic /l/) I’d expect someone from Melbourne to say ‘Malbun’ rather than ‘Melbun’ or ‘Milbun’. Interestingly Robbie Williams from Stoke-on-Trent consistently says the word melody as malady when talking about his singing in interviews, so that’s an accent and a minimal pair that should be investigated by linguists looking into this merger (of course it would be original research if I were to add that to the relevant WP article). --Overlordnat1 (talk) 15:02, 1 October 2022 (UTC)

If IPA transcriptions claim that the final vowel of Canberra, sofa, Oceania, Caloundra sound like the initial sound of 'about' or the final sound of 'teacher', they fail to provide "a standardized representation of speech sounds". Anybody suggesting that in German 'Opa' and 'Oper' are indistinguishable ought to return their tuition fee. -- Michael Bednarek (talk) 05:15, 4 October 2022 (UTC)

Misspellings

I am no one special, however, I came across this article and it got my attention and interest.

Maybe I am missing something however, several words in the last edit by Nardog are misspelled in several cases where for example there are several locations with double letters such as "st" appearing in the word. Should these words not be corrected to the proper English spelling or is this to illustrate phonetic properties?

Thanks for any reply/correction on my part. — Preceding unsigned comment added by 72.142.67.86 (talk) 03:29, 20 October 2022 (UTC)

You seem to have misinterpreted how the mobile diff represents an edit. The red background means the characters were removed. Or do you still see misspellings in the live version of Help:IPA/English? Nardog (talk) 03:42, 20 October 2022 (UTC)

Non-recognition of bracketed vowel on the system for providing IPA in articles despite commonality in British English

The system for entering a bracketed schwa does not appear to be recognised by the system for providing IPA in Wikipedia articles although it is used in the Oxford English Dictionary to describe a common feature in British English in which it is normal to drop the Mid central vowel in words like "berry" and "bury" when they are used in suffixes even in Received Pronunciation e.g. the pronunciation for "Salisbury" is given as /ˈsɔːlzb(ə)ri/ (based on RP rather than the local accent as in the city's Wikipedia article) to indicate that the "u" is usually dropped although the inclusion of the schwa in the first place might indicate it is standard. I only discovered this during a recent edit to the Shrewsbury article in which I tried to include a bracketed schwa but the system would not recognise it in the IPA but it did in the pronunciation respelling. However, it proved more desirable to have the pronunciation respelling match the IPA and another user removed the bracket from the schwa in the respelling although it is normal to drop the "u" in "Shrewsbury". I looked back through the archive of this talk page but it is hard to find a previous discussion on this matter (the best I could find might be Help talk:IPA/English/Archive 5#Using Cambridge Pronouncing Dictionary. If it has been raised multiple times, it might be desirable to include it in the FAQ page (Help talk:IPA/English/FAQ) if this problem still cannot be fixed. Tk420 (talk) 21:28, 23 October 2022 (UTC)

@Tk420: If using IPAc-en, just separate the brackets with pipes: ˈ|s|ɔː|l|z|b|(|ə|)|r|i. — kwami (talk) 22:53, 23 October 2022 (UTC)

Thanks but would it be worth including this solution in the FAQ page to save this discussion from being restarted after it is archived? Judging by what I have seen in the archive pages to this talk page, this issue appears to have been raised multiple times or at least others appeared to have tried to raise it. Tk420 (talk) 21:34, 24 October 2022 (UTC)

If you're having a problem with a particular template, then the FAQ should be in the documentation for that template. — kwami (talk) 21:59, 24 October 2022 (UTC)

As illustrated by history in the key and explained in note 32, the diaphoneme to use there is /ər/, which represents /ər/ or /r/ before a weak vowel. Nardog (talk) 06:31, 25 October 2022 (UTC)

So why should we allow other transcriptions, as in Canberra (and now Shrewsbury etc.)? I thought MOS:PRON was clear that any transcription linking to a Help:IPA/XXX guide should agree with that guide. Sol505000 (talk) 09:56, 25 October 2022 (UTC)

We shouldn't. Nardog (talk) 10:00, 25 October 2022 (UTC)

I was previously unaware of the issue with note 32. To the untrained eye, including ə or ər seems to imply that it is usually pronounced. Those characters could be omitted if dropping those sounds is normal in the local accent or the standard accent of the topic's country but this does not always find favour with editors. Perhaps there should be a separate discussion on whether to include a bracketed vowel in this help page to indicate a standard syllable which is commonly dropped considering it is used by some dictionaries including the Oxford English Dictionary which many Britons see as authoritative. If a bracketed vowel is included, it could be inserted under Weak vowels. Then again, this feature could be explained under Dialect variations considering these vowels seem to be pronounced in American English (in General American anyway) where there are commonly dropped in British English. Tk420 (talk) 22:17, 25 October 2022 (UTC)

It's already been proposed at some point and the argument against transcribing it was that dictionaries do not agree on which schwas can be omitted. I don't think we'd want to follow any single dictionary on this. Sol505000 (talk) 00:01, 26 October 2022 (UTC)

Including /j/ in /njuː/ also seems to imply that it is usually pronounced to the untrained eye. It's just the general nature of this diaphonemic system. Nardog (talk) 04:44, 26 October 2022 (UTC)

Any local pronunciation that can predictably derived from the diaphonemic transcription doesn't have to be made explicit in a separate narrow transcription; it's redundant. What can we do to prevent a situation where tag-teaming of editors who only care about the transcription of "their" toponyms (but not about the functionality of this key in general) can produce a "local consensus" that overrides the principle a uniform and elegant transcription?

This also includes the sandwiched schwa in "Canberra". These parentheses only make sense if there is a group of speakers that generally keep sandwiched schwa (in "cemetery" etc.), but consistently drop it in "Canberra", and only in Canberra. Are there any such speakers? –Austronesier (talk) 18:34, 26 October 2022 (UTC)

Help with Molossia

Could you please "translate" this respelling transcription (moe-LAAHSS-eeyah) found in this website? Thanks in advance.-- Carnby (talk) 10:28, 29 October 2022 (UTC)

/moʊˈlɒsiə/. Nardog (talk) 10:48, 29 October 2022 (UTC)

U as in en-uk "student", "beautiful" ?

Maybe I'm missing something obvious but the vowel table doesn't seem to give a notation for the en-uk pronunciation of "student", "stew", "beautiful" (except in Norfolk, where it is famously pronounced Bootiful [buːt]). Can anyone rectify? 𝕁𝕄𝔽 (talk) 17:20, 6 November 2022 (UTC)

It's consonant + vowel, /ju:/. For student and stew, the /j/ is written with the preceding consonant. — kwami (talk) 18:55, 6 November 2022 (UTC)

TYVM. I really should make time to learn more IPA for at least English. --𝕁𝕄𝔽 (talk) 11:02, 7 November 2022 (UTC)

Some dialects e.g. in NYC do have a vowel [ɪʊ] in new. But it's more usual to transcribe it [ju:], e.g. in the OED. The advantage of that for us is that we can transcribe new as having a consonant /nj/ and a vowel /u:/, which means that we can tell our readers that /nj/ is the consonant in new, so that we don't need to change the vowel for people who pronounce it 'noo'. — kwami (talk) 17:55, 7 November 2022 (UTC)

How are the words chosen?

I notice someone has recently changed some example words to be shorter, e.g. "blockade" to "block" and "trustee" to "trust", only to be reverted by Nardog with the explanation that the words are "carefully chosen" to represent each sound. How exactly are these words chosen, and why is it more preferable, for example, to illustrate /ɒ/ with "blockade" instead of the shorter root word "block"? 2001:4453:532:1900:F4EC:6437:6CD2:4B (talk) 14:34, 19 October 2022 (UTC)

It illustrates that full vowels can occur in unstressed positions and that even a checked vowel can be at the end of a syllable, at least notationally (i.e. /blɒˈkeɪd, trʌˈstiː/). The aspiration following /k/ and the lack of aspiration following /t/ necessitate analyzing /k/ and /st/ as syllable onsets, which makes respelling difficult (consider blo-KAYD, truh-STEE). Nardog (talk) 14:45, 19 October 2022 (UTC)

Respelling didn't used to be difficult for those words. It's only been so since someone added a bunch of <Vh> variants, rather than keeping the transliterations invariable (blo-KAYD, tru-STEE). — kwami (talk) 18:36, 7 November 2022 (UTC)

No, the addition of h makes it less difficult, as tru is no doubt going to be read with /u(ː)/. Nardog (talk) 19:35, 7 November 2022 (UTC)

Not by anyone intelligent enough to check the key. That's why we have a key. And as you noted, if we don't follow a key, then blocade becomes blow-cade. If we keep dumbing it down, at some point it will get so garbled that we won't be able to reliably transcribe most words. — kwami (talk) 00:36, 8 November 2022 (UTC)

TBH, this is why I never liked the respelling that much. When it works, fine, but English orthography is enough of a mess that there is sometimes no unambiguous way to write the vowel you want. So you need a key. Except that given the nature of respelling (it looks like English spelling conventions are in force), people will try to read the spelling without looking at the key anyway. If you need to force people to look at the key, you might as well just direct people to the IPA page. Although maybe I think that way because my first English dictionary had IPA. Alas, Wikipedia:United States dictionary transcription is deleted. Double sharp (talk) 00:14, 10 November 2022 (UTC)

Syllable example

Text: /ˈtæks.peɪər/ taxpayer

Shouldn't it be?: /ˈtæks.peɪ.ər/ taxpayer Jj2many (talk) 18:49, 12 November 2022 (UTC)

No because /peɪ.ər/ is the only possible syllabification for /peɪər/. The example illustrates how syllable breaks are only used as needed. Nardog (talk) 19:13, 12 November 2022 (UTC)

Audio samples

In other languages IPA pages the symbols has a link to the relevant page of the phoneme. why don't we have in this article? 31.223.87.186 (talk) 10:45, 6 December 2022 (UTC)

Because we don't transcribe actual sounds. E.g. /æ/ is pronounced different ways in different accents, and we don't propose that any particular one is correct. — kwami (talk) 08:16, 7 December 2022 (UTC)

Chanukah is not /x/?

I don't think anyone says Chanukah as /x/, but rather as /χ/. They are different IPA characters. Hebrew uses /χ/, something like Scottish English uses /x/. Please either correct me or the article :)) JacobTheRox (talk) 10:04, 27 October 2022 (UTC)

I know of no language, let alone variety of English, that makes a phonemic contrast between [x] and [χ]. Besides, the precise phonetic value in the original language is irrelevant to how to transcribe the English pronunciation of a loanword unless English speakers realize that difference. Nardog (talk) 10:17, 27 October 2022 (UTC)

@Nardog: The distinction is very common in the Pacific Northwest, from Oregon to Alaska and into Siberia, but Jacob, even if it were [χ] in English, we'd probably transcribe it /x/ for the same kinds of reasons we use /r/. — kwami (talk) 08:43, 7 December 2022 (UTC)

Why doesn't the IPAc-en template accept all sounds in English?

Wiktionary's list of phonemes used in English has more sounds that the template won't accept, such as /ɐː/, /ɚ/, and /ʉ/. Why can't the template use them? They will help make more accurate transcriptions. 2001:4453:5C6:CB00:7410:E3D:B61E:5699 (talk) 08:24, 21 December 2022 (UTC)

Because this key is diaphonemic. The diaphonemes already encompass those "more accurate" sounds. Nardog (talk) 09:42, 21 December 2022 (UTC)

Consistency of IPA phonemes for American English dialect pages

For editors interested in the use of IPA on American English dialect pages, I'd love further thoughts and contributions on this discussion. Thanks! Wolfdog (talk) 02:03, 31 January 2023 (UTC)

IPAc-en for non-English names

I've seen some articles using IPAc-en to approximate non-English names for personalities who don't live in the UK/North America/Australia/New Zealand. I remember reading years ago (possibly from an editor who cited an MOS) that this was not permitted, and I agree with that regardless. Of course, it's completely justified for someone like Adrianne Palicki who appears to have a Polish name but is a US national, or Katie Couric for the same reason, but surely something like this – IPAc-en for Oleksander Usyk, a Ukrainian national – is absolutely a no-no. Could I get clarification? Mac Dreamstate (talk) 17:32, 11 February 2023 (UTC)

It depends not on the subject's origin or nationality but on whether an established anglicization exists. It's not uncommon for a historical or internationally famous figure to have an established anglicization of their name, and at that point the only prerequisite for its inclusion is verifiability. But if the point is to show the pronunciation in Ukrainian, you have to use {{IPA-uk}}, as explained in Wikipedia:Manual of Style/Pronunciation. Nardog (talk) 18:08, 11 February 2023 (UTC)

Why is "sing" included as an example of ɪ?

If ɪ is the vowel in KIT and history, why is "sing" included as an example? Even if there are some accents that pronounce sing that way (I can't think of any), including it is confusing. Is it meant to be singe? LBlakeW (talk) 11:48, 19 March 2023 (UTC)

The contrast between /iː/ and /ɪ/ is neutralized before /ŋ/ in most accents, with the realization being either or somewhere between the two, but it's traditionally identified with /ɪ/. That's the whole point of including it in the table: the realization/phoneme in specific accents may vary, but regardless the diaphoneme to use in this situation is /ɪ/. Try to find a dictionary that transcribes sing with the same vowel as fleece rather than kit; I know of none. Nardog (talk) 13:19, 19 March 2023 (UTC)

Huh, I speak American English from the middle of the East Coast, and I pronounce "sing" with exactly the same as "kit" and "history"; I didn't even realise that there were accents/regions where that was not the case, though now that I think about it that makes sense, so it's interesting to contrast our different experiences with the pronunciation of that vowel. Indigopari (talk) 18:53, 7 April 2023 (UTC)

Regarding Ç

Hi @Austronesier: (I'm not pinging Nardog, that would just be counterproductive), I'm hoping to solve the dispute regarding /ç/ before this escalates into an edit war, as you said. I feel that we need to include the phoneme as:

• It's present in a number of key dialects, such as Received Pronunciation and Australian English (see Voiceless palatal fricative) and is mentioned on the English phonology article (see note f). Sure, it may not be in General American, but if we went by that logic then we shouldn't be including /ɒ/ either.
• This article's table in question already lists various /-j/ consonant clusters, so there's no excuse to not include /hj/ (as /ç/ is often transcribed.) or /ç/. (while we're at it, why isn't /pj/ there?). If we want to deny /ç/ as an English phoneme, then we should get rid of all the other /-j/ clusters present in the table aswell.
• The diaphoneme section below the table, to my knowledge, does not make a single reference to /ç/. Granted, this section in particular is quite hard to read on my end and /ç/ is quite a hard phoneme to find as ctrl+f-ing and entering ç will result in all instances of c and not just ç, so if I am missing something just let me know.
• For the sake of this next argument, we'll assume the dialect this article is using is British English (or Received Pronunciation) and that diaphoneme and allophone mean the same thing. I bring up allophone because the Japanese counterpart to this article makes a distinction between /h/ and /ç/. you could argue that /ç/ serves a much bigger function/purpose and that would be totally valid, but I don't think we should be denying /ç/'s presence in English as it is very clearly a separate phoneme. I'll also state that I understand that English and Japanese are two very different languages that are not related to each other, I merely went for the one I understood a bit more about. Maybe I should've chosen Danish or German, but I guess too late now.

So those are my arguments, I don't want this to end up an edit war like my last dispute with Nardog did, so I hope we can resolve this issue quickly and come to an agreement or solution. Kind regards, Great Mercian (talk) 22:05, 4 April 2023 (UTC).

References

@Great Mercian: I have to start with a reminder about what our diaphonemic transcription intends to achieve. It's not a tool for an accurate phonetic transcription of particular phones that are realized in certain varieties of English, however common they may be. It is meant as a shorthand notation that allows to predicably derive the pronunciation of a word in all major varieties of English. Our diaphonemic /ɒ/ is realized as BE /ɒ/ and GA /ɑ/, /-ər/ as BE /-ə/ and GA /-əɹ/ etc. We write /oʊ/ even though many varieties have a centralized vowel in the first part, while other have a monophthong. And so on.

So if we were to add /ç/ to the inventory, it wouldn't be for the sake of an acoustic phonetic transcription of BE [ç] (which btw Wells considers phonemically to be better analyzed as /hj/[6]), but as a shorthand to transcribe the diverse realizations of the onset of 'huge' in all sorts of varieties. Clearly, /hj/ does the job just as well without adding an extra segment, a segment that NB – if added – only would occur in a very restricted environment. It's a matter of parsimony.

Note also that the selected /Cj/ clusters are here for a reason. It's the subset of clusters that shows structural diversity in the realization in various accents. Some varieties drop /j/ entirely, some fuse it with the preceding stop resulting in an existing segment (e.g. /tj/ merging into /tʃ/). The case of /pj/ is different; there is devoicing of /j/ following an aspirated stop ([pç]), but that's just allophonic variation from a structural viewpoint.

That said, we might consider to add a note to /j/ which mentions the realization as [ç] following /p, k/, and the fusion of /hj/ into [ç] in certain English varieties in natural speech.

A final observation (admittedly OR): I have noticted that British speakers pronounce /hj/ as [çː] under emphasis (as in 'that's really huge'). The same speakers don't have [fː] when they say 'that's really far', which clearly shows that [ç] is actually just at the end of a chain of possible pronunciations of a phonemic cluster ([hj] > [hç] > [çː] > [ç]). –Austronesier (talk) 20:33, 5 April 2023 (UTC)

I assume a case could be made for the inclusion of //hj//. There appears to be some diatopic variation, cf. Phonological history of English consonant clusters#Reduction of /hj/. I doubt that anybody uses the sign ⟨ç⟩ in phonemic transcriptions, though.

When we are thinking of the mouse-overs of pronunciations such as /hjuːm/ (as in David Hume), then it seems to me that /hj/: ‘h’ in ‘huge’; /uː/: ‘oo’ in ‘goose’ would be slightly more helpful than our current ‘h’ in ‘hi’; /juː/: ‘u’ in ‘cute’. --mach 🙈🙉🙊 21:45, 5 April 2023 (UTC)

I think the criterion for any sequence of phonemes to be its own diaphoneme should be that it maps to different combinations of phonemes across accents, as in //dj// > /dj, d, dʒ/, //ər// > /ər, ə/. I don't know of any major accent where what is /hj/ in RP/GA would not be analyzed as /hj/. In NYC it could merge with /j/, but the key is missing variations found in much more prominent varieties, such as the lack of NURSE mergers in Scottish English. Nardog (talk) 22:02, 5 April 2023 (UTC)

@Austronesier: after reading over your comment, I'm afraid it's only had the opposite effect. The case for the inclusion of /ç/ has been incredibly bolstered, especially in relation to it's function with /p/ and /k/ (and maybe /t/?).

Firstly, while I'll say that I wasn't disputing the overall function of the diaphonemic transcription we're using, nor was I debating how we were writing it, I will admit I hadn't a good understanding of it, I've never claimed to have a degree in linguistics. (also, I'll say that your use of a source "which btw Wells considers phonemically to be better analyzed as /hj/" had already been superseded by what I had said earlier "as /ç/ is often transcribed.", so I don't know why you brought it up again.)

Secondly, highlighting /pç/ and /kç/ seems to only strengthen the argument for /ç/'s inclusion (I'm taking aspirated stops (plosives is a better term) may also include /t/, though I was normally render that as /tʃ/). I was unaware of these occurrences until you pointed them out. even if we included /ç/ under the Marginal Segments table then that would at least be some form of acceptance for this phoneme. May I also state that many other IPA help pages do not include /-j/ clusters unless they're palatalised, perhaps that's my ignorance as I haven't really ventured outside of about six of them.

Thirdly, I think I need to stress this part as you seemed to have glossed over it in your response:

The diaphoneme section below the table, to my knowledge, does not make a single reference to /ç/.

If you are going to [indirectly] point me towards the diaphoneme section, you may at least include the one phoneme I was looking for, I even included sources in my edit for your benefit for when you potentially add it there.

Fourthly, and this is just a comment on your original research, I'm British (if you haven't guessed), I find myself using /ç/ quite regularly in speech without emphasis. In fact, in what is possibly a direct influence of Japanese, I'm finding myself using /ç/ before /i:/ in speech. You know, it's kind of amusing that you also chose to overlook my entire argument about the Japanese help page, I guess maybe you saw it as irrelevant and nor worth your time, but I took the time to write it out so I expect you read it and at least offer me a sentence discussing it. also I know this is literally fighting Original Research with Original Research (which is essentially looking in the mirror), but if you're allowed to use it then so am I.

Finally, I'll highlight some of the other arguments I made that you did not address so that I may receive a response to them.

• and is mentioned on the English phonology article (see note f).

OK it's only one, but still. From your response it seems you've hyper-fixated on one point and discarded the rest as probably not worth your time, I expect better in a response, especially if it's about a dispute with a single edit to an article that 99.99% of users and visitors won't see. as we have both made it clear (you in your edit summary and me here) we don't want this to devolve into an edit war, so no more beating around the bush.

I still believe we should include /ç/ in this article in some form as per my arguments in my original post and my response here. I was not expecting this to get as big as it did, and I hope you will take into consideration my other arguments. It's really late right now so I'm going to sleep.

Thanks, Great Mercian (talk) 02:16, 6 April 2023 (UTC)

Notes

"For the sake of this next argument, we'll assume the dialect this article is using is British English (or Received Pronunciation) and that diaphoneme and allophone mean the same thing" contains two erroneous premises, as I have clarified in the opening part of my answer, so I haven't really focused on the forth bullet point. Also, I don't think that my use of abbrivations turns this discussion into a "forum". Please have the "decency" not to resort to metacommunication when there's an actual topic to discuss.

Let me repeat: "I have noticted that British speakers pronounce /hj/ as [çː] under emphasis". Note the [ː]; my point is not about the existence of phonetic [ç] in BE which is beyond doubt, but the lengthening under emphasis which is not obsevered with other frictives in English (except in expletive [ʃːːː...] with diverse overt and euphemistic realizations).

I have little time to go into detail right now. For the rest, in the meantime I refer to User:J. 'mach' wust's and User:Nardog's replies. I'll explain to you later why using [ç] (not /ç/!) in Help:IPA/Japanese and Help:IPA/German is a completely different thing from adding /ç/ here. Cheers! –Austronesier (talk) 10:15, 6 April 2023 (UTC)

@Austronesier: as I've made it clear already, I don't have a degree in linguistics, I'm just using terminology as they are to my understanding, and to mean diaphoneme and allophone mean essentially the same thing, a different sound. and as for my use of slashes instead of square brackets, they're just what I'm used to using, you can just treat them like square brackets. Anyway, looking forward to the rest of your response. Great Mercian (talk) 15:09, 8 April 2023 (UTC)

I have only just noticed this discussion going on. It looks as if some of the earlier material on the topic is no longer present - is it accessible? RoachPeter (talk) 08:24, 6 April 2023 (UTC)

It looks like Great Mercian simply copied the Harvard citations (but not the full citations) from English phonology and Australian English phonology. O'Connor (1973) is available on archive.org, Cox & Fletcher (2017) on Google Books, and I assume you have access to Roach (2009). ;)

Or if you mean the context for the discussion, see the revision history of the page this talk page is about. Nardog (talk) 08:34, 6 April 2023 (UTC)

Thanks - sorry my query wasn't clear: I meant the latter. I've tried looking at the history of this discussion but I'm still struggling to understand why Great Mercian got into the argument. I would, of course, never argue that < ç > should be treated as an English phoneme, though I do think it's worth talking about why it would be a bad idea to do so. It seems to me that Great Mercian's argument shows a lack of understanding of the difference between phone and phoneme, and between phonetic representation and phonemic. To which add the difference between phoneme, phonemic and diaphone, diaphonemic. I think the various replies above have already made this point clearly, so I won't labour it. RoachPeter (talk) 15:42, 6 April 2023 (UTC)

Look, I never said I had a degree in linguistics. Great Mercian (talk) 15:02, 8 April 2023 (UTC)

The cluster /hj/ seems to have disappeared from the consonant table. Has it fallen victim to the above dispute? I think it ought to go back in. RoachPeter (talk) 11:28, 10 April 2023 (UTC)

The table has /Cj/ clusters only for situations where some major dialects "drop" the /j/. For example /tj/, /dj/ in /tju:n/ and /dju:n/ - most American dialects have /tu:n/, /du:n/. For /hj/, there aren't any major dialects where it's reduced to /h/. Indefatigable (talk) 14:38, 10 April 2023 (UTC)

See East Anglian English#consonants re yod-dropping. RoachPeter (talk) 08:50, 11 April 2023 (UTC)

I don't believe that cluster was ever listed in the table in the first place, at least not in recent times. Offa29 (talk) 14:50, 10 April 2023 (UTC)

I am comparing what is included in this list with what is given in English phonology#consonants, footnote (f). /hj/ (phonetically [ç]) is one of a pair with /hw/ (phonetically [ʍ]). Including /hj/ in Help:Ipa/English might help to prevent the sort of misunderstanding shown in the present discussion. RoachPeter (talk) 08:47, 11 April 2023 (UTC)

Giving any sequence of symbols its own row in the key would mean it's a full-fledged diaphoneme in our system. But as discussed in Help:IPA/English#Dialect variation, there are some distinctions which you might make but which this key does not encode, as they are seldom reflected in the dictionaries used as sources for Wikipedia articles. We might as well add /hj/ in that list, but not in the main key. Nardog (talk) 09:15, 11 April 2023 (UTC)

The diaphonemic system being Wikipedia's standard transcription

I'm wondering whether we can be a bit more assertive with Wikipedia's diaphonemic system across articles to prevent (or help quickly end) future debates. If this topic is of interest, contribute thoughts here please. Much obliged! Wolfdog (talk) 15:12, 12 April 2023 (UTC)

Help means for the "IPA-impared"

Is there a place in Wikipedia where an editor can request an IPA-knowledgeable editor's help in 1) inserting IPA in a page currently having none (see The_Ghan#Etymology) — or 2) verifying that an amateur's attempt at such is correct? I suppose initially such request should be made directly on the talk page for that article. If none forthcoming, then the novice editor could make a stab at it in hopes that if not done correctly, some reader of that very page will edit that attempt.Casey (talk) 13:20, 27 May 2023 (UTC)

You can go to Wikipedia:Reference desk/Language as the top of this page says. You may also use {{needs IPA}}. Nardog (talk) 06:48, 28 May 2023 (UTC)

Isn't [a] vs. [æ] phonetic?

Given that this is a key about diaphonemes, does there really need to be a bit about how some dialects pronounce the TRAP vowel lower than in others, given that it doesn't lead to the loss of any phonemic oppositions (except perhaps for TRAP-PALM in Scottish English, which, by the way, isn't mentioned)?

If the point is that some dictionaries write <a> while others <æ>, might I suggest saying the same for <əʊ> vs <oʊ>? Tyrui (talk) 04:25, 28 May 2023 (UTC)

that some dictionaries write <a> is indeed the point, and we already have a note about ⟨əʊ⟩ too. Nardog (talk) 06:49, 28 May 2023 (UTC)

I've checked again, and I'd like to note that I'm referring to the following paragraph in Dialect variation:

The pronunciation of the /æ/ vowel in most dialects of Scotland, Northern Ireland, Northern England and Wales has always been closer to [a]. Received Pronunciation has moved away from the traditional near-open front realization [æ] towards almost fully open front realization [a], and both the Oxford English Dictionary and the 2014 edition of Gimson's Pronunciation of English transcribe the vowel in lad, bad, cat, trap with /a/.

The note about [a] vs [æ] is fine, but this paragraph seems redundant. Tyrui (talk) 15:57, 28 May 2023 (UTC)

Myanmar

I think that the IPA transcriptions of English in Myanmar should use phonetic brackets with a broad phonetic transcription, not the diaphonemic slashes. Look at the note after the first word (Myanmar) and the Etymology section. The note should read:

Burmese: မြန်မာ [mjəmà]; UK pronunciations: [ˌmjænˈmɑː, ˈmjænmɑː] US pronunciations incl. [ˈmjɑːnmɑːr, ˌmjɑːnˈmɑːr]. As John Wells explains, the English spellings of both Myanmar and Burma assume a non-rhotic variety of English, in which the letter r before a consonant or finally serves merely to indicate a long vowel: [ˈmjænmɑː, ˈbɜːmə]. So the pronunciation of the last syllable of Myanmar as [mɑːr] or of Burma as [ˈbɜːrmə] by some speakers in the UK and most speakers in North America is in fact a spelling pronunciation based on a misunderstanding of non-rhotic spelling conventions. The final r in Myanmar was not intended for pronunciation and is there to ensure that the final a is pronounced with the broad ah ([ɑː]) in "father". If the Burmese name မြန်မာ [mjəmà] were spelled "Myanma" in English, this would be pronounced [ə] at the end by all English speakers. If it were spelled "Myanmah", the end would be pronounced [ɑː] by all English speakers. See other English pronunciations in § Etymology.

This looks better than the diaphonemic transcription followed by Note: Wikipedia's IPA conventions require indicating /r/ even in British English although only some British English speakers pronounce r at the end of syllables.

But the Etymology sections is trickier as the UK and US pronunciations are lumped together and would have to be separated first. Sol505000 (talk) 12:59, 11 June 2023 (UTC)

A few possible additions to the marginal segments

1. In descending order of frequency, many rhotic accents realize yeah, nah, and waah with a special centering diphthong(the most emphatic form in my own accent for instance is initially nasal and glides in a hook shape; something like [ẽ̠ɛa̟ɜ̠]), as opposed to the lax gah, meh, etc.
2. The interjection ew or eww is transcribed with /iu/ or similar in every dictionary that features it, and in some accents, a similar sound is found in whew, New York, and/or dude, among others.
3. In some variations of okay, -ing in rapid speech, and nightingale, a syllabic ŋ occurs.
I'm a newcomer and not 100% clear on what the bar is here, but these don't strike me as out of place among the marginals already listed. Célestine-Edelweiß (talk) 00:12, 19 June 2023 (UTC)

Just because they occur doesn't mean they should be added to the key. § Dialect variation lists non-marginal distinctive segments that are much more frequent than the ones you named. There's no reason to consider adding them unless there are a substantive number of articles that benefit from there being such diaphonemes. The marginal segments currently included are included because they appear in dictionary entries we cite. ([ŋ̍] in -ing and nightingale is just an allophone deducible from /ɪŋ/ and /ɪn/ respectively.) Nardog (talk) 00:23, 19 June 2023 (UTC)

I retract my [ŋ̍] suggestion, but unlike wholly, cider, and the like, 1 and 2 have no regular conditioning. If the glottal stop can be included due to uh-oh, I don't see why the interjections I listed don't count. Célestine-Edelweiß (talk) 00:37, 19 June 2023 (UTC)

See yeah and ew(should have included these links in the first comment, apologies) Célestine-Edelweiß (talk) 00:41, 19 June 2023 (UTC)

The glottal stop is included mainly for Hawaiian loanwords. Nardog (talk) 00:43, 19 June 2023 (UTC)

Link/embed ipa-reader.xyz

Apologies if this is a well worn topic, but I’ve wondered why the IPA notation in many articles doesn’t link to an audible pronunciation. I don’t have an issue with the notation itself, but the vast majority of readers are likely interested in guidance on how the word sounds more than an explanation of the phonetic notation. I’m sure there are practical considerations to any integration, but it seems like it’s a lot more steps to get to hearing the word than it should be, and many readers won’t find a site like ipa-reader or won’t bother. 24.241.32.226 (talk) 15:16, 17 December 2023 (UTC)

One reason is that this key is supposed to be dialect-neutral, though that hasn't deterred transcriptions occasionally accompanying audio. WMF is to develop software that will enable instant generation of audio from IPA, but it doesn't seem active at the moment so it'll likely take years. Once that becomes available, we might ditch the key in favor of dialect-specific ones (or who knows what happens). Nardog (talk) 21:08, 17 December 2023 (UTC)

Indecipherable

Ugh, I HATE that Wiki uses this IPA crap. I don't know why they can't show pronunciations like the dictionary does. It's so much more intuitive that this gobbledygook. IPA was created by drunken sailors. — Preceding unsigned comment added by 67.241.240.42 (talk) 20:16, 16 December 2023 (UTC)

Most english words on wikipedia use respelling as well, take for example Victor Wembanyama. It has the gibberish, and then a spelled-out more understandable version. If you see an article that should have one of these, add it!

For a more drawn-out explanation, this talk page has a question just like this in the FAQ section. Coulomb1 (talk) 01:56, 29 January 2024 (UTC)

ˈnɒnsᵊns, ðɪs ˈɡɪbərɪʃ ɪz ˈiːzi tuː riːd, ænd ˈdɛfɪnɪtli nɒt ˈsʌmθɪŋ juː ˌɔːtəˈmætɪkᵊli skɪm ˈəʊvə Local Potentate (talk) 09:10, 9 February 2024 (UTC)

Distinction between /æ/ in "trap" and "bad"

The latter is extended in most accents I've heard, so surely should include : like in /ɑː/ and /iː/. I'm not seeing it on the table. Mac Dreamstate (talk) 12:21, 24 February 2024 (UTC)

@Mac Dreamstate: Checking three reliable sources:

• dictionary.com shows /træp/; /bæd/
• Merriam Webster shows /ˈtrap/; /ˈbad/
• OED shows /trap/ (Br), /træp/ (US); /bad/ (Br), /bæd/ (US)

So although the "a" may vary in sound between English variants, the same sound is used within a variant for "trap" and "bad". Bazza 7 (talk) 12:50, 24 February 2024 (UTC)

Weird. To me they sound as different as /ɪ/ and /iː/. I'm not going to call into question anyone's linguistic credentials, but it's like someone's missing a trick somewhere. Mac Dreamstate (talk) 13:25, 24 February 2024 (UTC)

@Mac Dreamstate: Someone else might want to contribute here and provide some better references. I was, I confess, surprised by the sources. I think I know what you're alluding to; we might, from our seats on this side of the pond, respectfully call it a "drawl". Bazza 7 (talk) 14:12, 24 February 2024 (UTC)

It isn't a phonemic difference. A speaker's phonetic realizations might be transcribed as [træp] and [bæːd] but it's one phoneme, written /æ/. On the other hand, considering only my own (eastern US) speech, I don't see why, in contrast, the phonemes /ɑː/ and /iː/ are marked long. To me, the vowels in "beat" and "bead" differ in length in the same way as "bat" and "bad", so it isn't clear to me why the phoneme behind one pair would be explicitly marked long and the other one not. Perhaps there are enough examples of English around the world with long vowels in both "beat" and "bead" and short vowels in both "bat" and "bad" that it makes sense. Largoplazo (talk) 22:41, 24 February 2024 (UTC)

There is a phonemic distinction made in some accents of English: the Pronunciation of English ⟨a⟩#Bad–lad split. However this is not accounted for in the transcription system, presumably because it is not featured in most dictionaries. It is far from the only possible phonemic distinction omitted; cf. the horse-hoarse and fir-fur-fern distinctions in Scottish and Irish English. Offa29 (talk) 02:49, 25 February 2024 (UTC)

200 days ago? 50.231.115.70 (talk) 12:20, 18 March 2024 (UTC)

Coco Gauff

A minor dispute on whether IPA-en should be included for this tennis player. I maintain that Gauff is a very unusual and non-intuitive name in the English-speaking world, but the other editor disagrees. We have IPA-en for Jessica Biel, Alicia Silverstone and Joe Biden, so why should her article be any different? Mac Dreamstate (talk) 21:08, 14 February 2024 (UTC)

It's the normal pronunciation for those who have "cot-caught merger" plus "father-bother merger"... AnonMoos (talk) 16:56, 18 March 2024 (UTC)

@AnonMoos: But for those have neither merger (like me), it would be helpful. Is it /ɡaʊf/, /ɡɔːf/ or /ɡɒf/? Bazza 7 (talk) 09:09, 19 March 2024 (UTC)

This key is WP:DIAPHONEMIC. Nardog (talk) 10:14, 19 March 2024 (UTC)

@Nardog: I'm not sure how that's pertinent. As discussed elsewhere by Mac Dreamstate, my British English stumbled on that name: /ɡaʊf/ (gowf) came first, then /ɡɔːf/ (gawf), then /ɡɒf/ (gof).

I see that User_talk:Fyunck(click) reinstated the referenced original guide /ˈɡɑːf/ GAHF, which is just as well as it wasn't on my internal list of three distinct pronunciations. Bazza 7 (talk) 11:25, 19 March 2024 (UTC)

I was replying to AnonMoos. /ɑː/ is indeed the vowel for those with father–bother and cot–caught mergers, but it's a non-option for a diaphonemic notation. We need what those without the mergers pronounce it (the LOT–CLOTH split is also potentially relevant here, but given the subject is American—from the South no less—we could assume its presence; see the note at the end of the "Key" section). Nardog (talk) 11:32, 19 March 2024 (UTC)

@Nardog: My apologies for not paying attention to your accurate indentation. And thanks for your concise and understandable explanation. Bazza 7 (talk) 11:35, 19 March 2024 (UTC)

From what I can tell, American speakers without the cot-caught merger pronounce the name as /ɡɔf/, not /ɡɑf/. This makes sense as it corresponds to the spelling. I have therefore changed the pronunciation guide to this. Referencing Gauff's own pronunciation is useless as she has a PALM-LOT-THOUGHT merged accent, so her /ɑ/ phoneme corresponds to three separate Wikipedia diaphonemes. Offa29 (talk) 13:32, 19 March 2024 (UTC)

Double entry

Why is ɪ appearing twice, both under "Vowels" and "Weak vowels"? If we need two entries here, I would expect separate symbols (even if one is a modification of the other with a combining mark of some kind).  — SMcCandlish ☏ ¢ 😼  22:05, 29 April 2024 (UTC)

I believe the double entries for //ɪ// and //oʊ// are mainly there for historic reasons, back from the day when we were propagating our own idiosyncratic symbols for the weak vowel versions of the two. I have tentatively unified the symbols, keeping all the content. --mach 🙈🙉🙊 06:11, 1 May 2024 (UTC)

Colons for length symbols

In the 3rd bullet point of the Dialect variation section colons are used in place of length symbols:

Most speakers of North American English (with the exception of Eastern New England) do not distinguish between the vowels in father /'fɑ:ðər/ and bother /'bɒðər/, pronouncing the two words as rhymes. If you speak such a dialect, ignore the difference between the symbols /ɑ:/ and /ɒ/.

I think they need to be replaced. 2001:BB6:B84C:CF00:B1A9:DA55:640A:FC65 (talk) 20:13, 1 May 2024 (UTC)

Done. Nardog (talk) 22:39, 1 May 2024 (UTC)

IPA overwhelming

IPA is overwhelming, redundant, and not user friendly. If you use the basic latin sounds the phonics are all there and we all know them. No need to learn a whole new set of sounds that are extremely numerous and cumbersome. 136.143.149.206 (talk) 17:51, 2 June 2024 (UTC)

We don't "all know them", though. Your west coast US pronunciation will be different from mine. WP:RESPELL describes how simple pronunciation guides don't always work. For instance, I pronounce ""Mary", "marry", and "merry" differently, but know that some Americans don't. The same applies to "cot" and "caught". Some of my compatriots pronounce "aren't" and "aunt" differently, but I don't. Bazza 7 (talk) 18:55, 2 June 2024 (UTC)

Text on secondary stress

On the help page, we show both primary and secondary stress marks, yet we never define how we do (or don't) use those symbols in the diaphonemic system. I believe the last chat we had arriving at some consensus was here, where we agreed on WP to assign secondary stress only to a strong vowel preceding primary stress but not to a strong vowel succeeding it (i.e., following the British rather than American convention). It seems like it would be helpful to explain this, and even the concept of how secondary stress operates in English at all, if anyone can think of a concise wording for the concept. Wolfdog (talk) 12:53, 19 June 2024 (UTC)

Decided to BE BOLD. Wolfdog (talk) 13:54, 21 June 2024 (UTC)

Anki

We currently show the pronounciation of Anki as /ˈɒŋkiː/, which seems surprising to me. I am aware that /ɒ/ for the spelling ⟨an⟩ is not unheard of, especially in French loanwords with /ɑ̃/ in the original, but is Anki really pronounced like that? 187.245.68.84 (talk) 00:15, 23 June 2024 (UTC)

I expect /æ/ or /ɑː/ (especially from British and American speakers, respectively), but it's unsourced anyway. It should be sourced to an ad or a developer saying it. Nardog (talk) 06:01, 23 June 2024 (UTC)

/ˈɑːŋki/ is all I've ever heard in US English, so that's what was meant, most likely. I've corrected the IPA, changing /iː/ to /i/ because the nasal is always velar (so the vowel is weak, phonemically /ɪ/ according to Cruttenden, or /ɨ/ according to some others). Sol505000 (talk) 15:47, 23 June 2024 (UTC)

Youglish indicates that /ˈæŋki/ is the norm in BrE, which fits for the general pattern of nativisation of <a> in recent loanwords in BrE vs. AmE. Offa29 (talk) 15:59, 23 June 2024 (UTC)

Semi-protected edit request on 2 July 2024

change hw to ʍ 71.78.136.215 (talk) 00:43, 2 July 2024 (UTC)

 Not done for now: please establish a consensus for this alteration before using the {{Edit semi-protected}} template. Nardog (talk) 02:38, 2 July 2024 (UTC)

Inclusion of /ts/ as a marginal phoneme and removal of /ʔ/

/ʔ/ is an entirely paralinguistic sound and "uh-oh" is not a valid word to base the inclusion of a marginal phoneme around. However, seeing and /ts/ is a common marginal phoneme in words like "tsar" or "Mozart", including it would probably be valid. Plexus96 (talk) 14:36, 17 May 2024 (UTC)

/ʔ/ is included for Hawaiian loans. It's illustrated by uh-oh simply because it's one of the most common and intuitive ways to illustrate the sound; it doesn't mean it's only used in paralanguage.

/t/ and /s/ are already phonemes so there's no need to list /ts/ separately. Nardog (talk) 00:42, 18 May 2024 (UTC)

/t/ and /ʃ/ are already phonemes so there’s no need for /tʃ/ as well…? БудетЛучше (talk) 18:15, 21 June 2024 (UTC)

Because it's a phoneme in English if you ask just about any linguist. See English phonology#Obstruents for why. Nardog (talk) 18:17, 21 June 2024 (UTC)

/ts/ behaves more like a consonant cluster, rather than a phoneme. It doesn't appear word-initially, at least not regularly (see e.g. [7]), only word-internally and -finally. Compare this with German /ts/ which can easily appear in this position, as in zu /tsuː/ or ziemlich /ˈtsiːmlɪç/. Native speakers of English constantly mispronounce those as /syː ~ suː/ and /ˈsiːmlɪk, -x, -ʃ/. Sol505000 (talk) 15:39, 23 June 2024 (UTC)

@БудетЛучше 2001:44C8:4183:F40B:5071:B1FF:FE2F:A594 (talk) 17:06, 24 August 2024 (UTC)

thanks for the explanation БудетЛучше (talk) 17:33, 24 August 2024 (UTC)

"/ʔ/ is an entirely paralinguistic sound"

It's also a common allophone across most dialects of English, particularly for /t/ 167.206.19.130 (talk) 19:56, 21 June 2024 (UTC)

That's neither here nor there. We transcribe any allophone of /t/ as /t/ because this key is diaphonemic. Nardog (talk) 20:41, 21 June 2024 (UTC)

Edit warring at Richard D'Oyly Carte

There is edit warring against MOS:DIAPHONEMIC on that page. It's Melbourne all over again, with the 'local consensus' nonsense. Sol505000 (talk) 15:30, 4 July 2024 (UTC)

Nice canvassing. As you're the main edit warrior on the page, it's a bit of a stretch - and more than a little uncivil - to disregard the well-founded arguments of others as 'nonsense'. - SchroCat (talk) 15:34, 4 July 2024 (UTC)

You don't get to call me uncivil when you so blatantly disregard MOS:DIAPHONEMIC and this guide. Take a good look at the mirror, mate. You are the one ignoring a well-established consensus in the APPROPRIATE place, which is right here. Sol505000 (talk) 20:59, 4 July 2024 (UTC)

I’m not your ‘mate’, sunshine, and yes, I’ll call out incivility and idiotic edit warring when I see it. There’s a talk page for you to use if you’re able to do so without reverting to further incivility. - SchroCat (talk) 21:10, 4 July 2024 (UTC)

@Tim riley: has raised a valid point: Alas, readers familiar with the real-world IPA cannot be expected to guess that Wikipedia has its own esoteric phonetic system in which the pronunciation symbols mean "this unless you'd rather pronounce it that". Our transcriptions are still surrounded by single slashes /…/. Readers familiar with the IPA will know that this means a phonemic transcription. However, our transcriptions are not phonemic, but diaphonemic. We do not mark them as such, though. I thought last year we had reached a consensus that we should mark the diaphonemicity of our transcriptions by surrounding them with double slashes ⫽…⫽, see Wikipedia talk:Manual of Style/Pronunciation/Archive 11#Distinction between varieties of English. Are other steps required? --mach 🙈🙉🙊 11:48, 5 July 2024 (UTC)

To be fair, transcriptions meant to represent specifically GA or RP are themselves most often in effect diaphonemic - the former for assuming a cot-caught distinction(sometimes also the strut-comma and weak vowel mergers, which are mostly absent from the Inland North, Western Pennsylvania, New England, and Eastern Canada), the latter for assuming no l-vocalization, and both for assuming no /æ/ splits. Célestine-Edelweiß (talk) 01:29, 7 July 2024 (UTC)

Mach, perhaps Nardog knows. As for the edit-war situation, I've contacted a random uninvolved administrator (simply, the most recently active one), who can perhaps help the situation. Wolfdog (talk) 01:49, 7 July 2024 (UTC)

Done. I've also created {{IPA //}} as a shorthand akin to {{angbr IPA}}.

I've replaced {{IPA|/.../}} in the Key section where they were unambiguously referring to diaphonemes rather than phonemes as far as I could find, but it can be ambiguous sometimes. (E.g. should it be "/A/ is merged with /B/ in accent X", "⫽A⫽ is merged with /B/ in accent X", or "⫽A⫽ is merged with ⫽B⫽ in accent X"?)

Further fixes in articles where our diaphonemes are mentioned are welcome. Nardog (talk) 12:10, 7 July 2024 (UTC)

That’s great, thanks a lot! --mach 🙈🙉🙊 06:48, 8 July 2024 (UTC)

Undone per request. I suggest you start an RfC if you want to go through with this. Nardog (talk) 08:54, 8 July 2024 (UTC)

I have opened a request for comment to discuss the matter: Wikipedia talk:WikiProject Linguistics#RfC: Should we keep our non-standard use of single slashes to enclose diaphonemic transcriptions? Wikipedia talk:WikiProject Linguistics#RfC: Should we keep delimiting diaphonemic transcriptions with single slashes? --mach 🙈🙉🙊 21:38, 8 July 2024 (UTC)

I just suggested it at that talk page, but in case it's of more interest here: since it seems that //r// is so much more of a chronic objection than the fact that our key makes vowel distinctions that AmE doesn't, perhaps the problem is rather that "don't pronounce this letter that's clearly there" is a bigger mental block than "pronounce these two letters the same way". So perhaps it might help to change the symbols we use for the diaphonemes to //ɑː(r)//, //ɔː(r)//, etc., explicitly parenthesising the r? Double sharp (talk) 08:17, 11 July 2024 (UTC)

Using superscript ⟨ʳ⟩ has been suggested, e.g. here. That seems less intrusive than parentheses, and has precedence in CamGEL. Nardog (talk) 08:33, 11 July 2024 (UTC)

Even better, then, if there's a precedent. I'm happy to support superscripts if they get proposed. Double sharp (talk) 08:56, 11 July 2024 (UTC)

Either solution seems fair given the insistence of some to disregard well-established conventions over the mere sight of the /r/ symbol. I support that. ~ IvanScrooge98 (talk) 10:03, 11 July 2024 (UTC)

I also am content with either option. Something to consider for the implementer, though, is that we currently transcribe a word like bearing as b|ɛər|ɪ|ŋ in this system. If we made the change, we would want to avoid the output being something like bɛəʳɪŋ, since the r here in these types of cases is not optional even in non-rhotic accents but required. Wolfdog (talk) 14:12, 11 July 2024 (UTC)

That can be automated in Lua. Can be tricky, but feasible. Nardog (talk) 14:18, 11 July 2024 (UTC)

Nobody has intervened any more. Are we going to give ⟨ʳ⟩ a try or should we open a separate dedicated discussion before? ~ IvanScrooge98 (talk) 14:17, 21 August 2024 (UTC)

I would wait till the double slash RfC is closed. Nardog (talk) 02:15, 25 August 2024 (UTC)

i’m sure that replacements using a regex such as ([iɪʊuəɛɜʌɔæɑɒː])ɹ([iɪʊuəɛɜʌɔæɑɒː]) will suffice for finding R-liaison. БудетЛучше (talk) 18:11, 24 August 2024 (UTC)

Here's what really happened to initiate this discussion. There was a longstanding stable IPA on the Richard D'Oyly Carte page, and then User:Sol505000 edit warred there about it, but they did not get a consensus in their favor. So they brought the issue to this Talk page to see if they could get a different consensus here. -- Ssilvers (talk) 23:04, 11 July 2024 (UTC)

I have had disagreements with Sol505000 in the past, but we really should look at the situation and realize having a discussion about {{IPAc-en}} on the talk page of a specific article, when the template is used in countless pages, is out of place. ~ IvanScrooge98 (talk) 23:48, 11 July 2024 (UTC)