Jump to content

Help talk:IPA/English

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
For requests for transcription, see Wikipedia:Reference desk/Language.

Dew pronunciation is 'dyew' not just 'doo' and 'djoo'

[edit]

'Dy' exists and is standard, not just 'doo' (American)' and 'djoo' (technically, bad diction). ~2026-31781-5 (talk) 13:58, 15 January 2026 (UTC)[reply]

Nowhere do either "doo" or "djoo" appear on the page. In addition, in the IPA (which is what the page is about), "j" represents the initial sound of "you" and "yet", whereas "y" represents a vowel sound not present in most varieties of English, though you might be familiar with it if you know, for example, French ("fume") or German ("müde") or Dutch ("dus"). So the "/djuː/" that you see in the footnote does represent the pronunciation you were probably thinking of when you wrote "dy". Similarly, phonemically, "you" is represented as "/juː/" and "yet" as "/jɛt/". The letter "j" in English most commonly corresponds to the representation "/dʒ/", while "ʒ" itself corresponds to the "s" in "leisure" or "vision". Largoplazo (talk) 15:19, 15 January 2026 (UTC)[reply]

Enthuse is not a real word.

[edit]

Enthuse is not a real word. Find a better example. ~2026-15950-75 (talk) 19:43, 17 March 2026 (UTC)[reply]

If people use it, what makes it not a real word? Largoplazo (talk) 00:37, 18 March 2026 (UTC)[reply]
  • "enthuse". Oxford English Dictionary (Online ed.). Oxford University Press. (Subscription or participating institution membership required.)
  • "enthuse". Merriam-Webster.com Dictionary. Merriam-Webster. OCLC 1032680871.
  • wikt:enthuse
Seems like a word to me. ~ oklopfer (💬) 01:02, 18 March 2026 (UTC)[reply]

⟨œ⟩ in American dictionaries and correspondence in English accents

[edit]

It doesn't seem to be "merely a notational convention". In both MW and AHD, ⟨œ⟩ maps directly to either [ø] or [œ] (depending on the word, not the dictionary):

MW notes that it is often anglicized to NURSE (bird in their convention), but it is not some mysterious and unidentifiable vowel as this key makes it out to be. There seems to be a conflicting standard with "does not correspond to any vowel in any accent of English", as the French nasal vowels arguably do not either any more than this one does, they are all typically only found in loans; though on the contrary, /øː/ is well known to be the realization of NURSE in NZE (and several other dialects, see Roundedness#Roundedness in English), so the statement that it does not occur in any accent of English seems to just be provably untrue. ~ oklopfer (💬) 20:00, 23 March 2026 (UTC)[reply]

Unlike with /x, ɒ̃, æ̃/ (and to a less extent /ʔ/), I'm familiar of no description of English phonology or dictionary of English that counts /œ/ as a member of the phonemic inventory of English, even as a marginal one. AHD and M-W are not dictionaries of NZE; if NZE uses NURSE where rhotic accents don't, the corresponding diaphoneme is obviously /ɜː/. Which M-W already has a symbol for: ⟨ə̄⟩ (and AHD in theory ⟨û⟩).
Being reference works for laypeople and traditionally space-constrained, dictionaries of English often include pronunciations of non-English words that are not common among native speakers of English, using their symbols for English phonemes plus a few but not necessarily specifying that's what they are. So we get transcriptions like phonotactically illegal /luːvr/ and approximated \ā-ˈdēt\, even though everyone says /ˈiːdɪθ/. Since we have dedicated keys for other languages, we aren't and have no reason to be like them (MOS:PRON).
Each marginal diaphoneme besides /ɜː/ has known nativized alternatives. /x/ is /h/ initially and /k/ in coda. /ɒ̃, æ̃/ are /ɒn, æn/ or /ɒŋ, æŋ/. /ʔ/ is omitted (/həˈwaɪi/). /ɜː/ doesn't, so we say it's non-rhotic only. ⟨œ⟩ in American dictionaries is clearly not the NAmE counterpart to BrE /ɜː/ because it doesn't show up consistently where BrE dictionaries show /ɜː/ (see e.g. adieu, Henan, pho). So if we were to make it possible to convert transcriptions with ⟨œ⟩ to our diaphonemic system, it could only be done by adding another diaphoneme, not by extending the definition of /ɜː/.
But making it a diaphoneme would only make sense if reliable sources said native speakers of English (who are not speakers of languages with /ø, œ/) can and do adopt [ø~œ] in words some dictionaries transcribe with ⟨œ⟩. Are you familiar with any? M-W's "This vowel is often anglicized as ..." reads to me rather as an explicit acknowledgement that they regard \œ\ as not a part of English phonology they aim to describe. AHD and M-W also have ⟨ᵫ, ü⟩ for [y, ʏ]. Do you think we should add a diaphoneme for them as well? Nardog (talk) 03:34, 26 March 2026 (UTC)[reply]
I think we may be in disagreement about what counts as a diaphoneme at all.
I am not familiar with dictionaries that state /æ̃/ is a marginal phoneme in English. I have only ever seen /ɒ̃~õ/. I would appreciate some examples.
My NZE point was that the note in the key here says orthographic ⟨œ⟩ in American dictionaries does not map to any sound in any dialect of English; but in fact it does, as MW and AHD give the precise value, which we can then compare occurrence of across dialects.
I do not think we should be treating non-rhotic NURSE in foreign words as a distinct diaphoneme like the others, I would regard it as a possible native realization of //œ//, which is supported by what MW states, and also a realization of this separate foreign words diaphoneme, which is distinct from //œ// because of its cross dialectical patterning. I think that comparing adieu, Henan, pho to Goethe is an improper collapsing of the two (or more, since those all pattern differently in NAmE…) diaphonemes into one, merely because they are the same phoneme in some dialects (but not in others); the point of a diaphoneme is very specifically to avoid exactly this dialectical collapsing and not give any of them "undue weight". If a phoneme does not pattern consistently cross dialectically, then it is multiple diaphonemes, one for each patterning set.
I also think the part you've picked out from MW is too selective. The full explanation reads:

IPA [œ]. This vowel, which occurs only in foreign-derived terms and names, can be approximated by attempting to pronounce the vowel \e\ with the lips moderately rounded as for the vowel \u̇\. This vowel is often anglicized as the \ər\ of bird by those who do not drop their rs or as the corresponding vowel of bird used by those who do.
This symbol is also used to represent […] IPA [ø]. This vowel, which occurs primarily in foreign-derived terms and names, can be approximated by attempting to pronounce a monophthongal vowel \ā\ with the lips fully rounded as for the vowel \ü\. This vowel also occurs in Scots and thus is used in the pronunciation of guidwillie, mainly restricted to Scotland.

That says plainly that it occurs, anglicization is again merely a realization. Why else would the dictionaries include those phones at all in their descriptions, unless they actually were reliably seen as realizations? It's not like alveolo-palatal affricates or bilabial fricatives are listed anywhere in them. They single these vowels out because they are notable and consistent enough to warrant mentioning. ~ oklopfer (💬) 17:52, 26 March 2026 (UTC)[reply]
I actually do also think loan [y ~ ʏ] should be considered a separate diaphoneme. My proposal would be to do something like:
  • //œ//: Found in loans such as Goethe. Often nativized as /ɜː/ in r-less accents and /ɜːr/ in r-ful ones.
  • //yː//: Found in loans such as über. Often nativized as /uː/ or /juː/.
/ɜː/ in other cases of loans is highly problematic to treat as a single diaphoneme:
  • adieu, Peugeot (loans of /ø/) may correspond to /ɜː ~ (j)uː/
  • Villeneuve (loan of /œ/) may correspond to /ɜː ~ uː/
  • Möbius (loan of /øː/) may correspond to /ɜː ~ {ə~o}ʊ/
  • pho, Henan (loans of /ɤ/, without the tones) may correspond to /ɜː ~ {ə~o}ʊ ~ ə ~ ʌ/
Clearly they do not map consistently, neither from nor to, and in a few of these /ɜː/ is certainly not the dominant realization, even in non-rhotic accents. Conversely, in some of them /ɜː/ can show up in rhotic accents, like adieu and Villeneuve. /ɜː/ is a possible point of convergence for loans, not a diaphoneme in itself, and not necessarily governed by accent rhoticity either; it merely is more likely as a possibility in non-rhotic accents where /ɜː/ is already part of the phonemic inventory. ~ oklopfer (💬) 01:28, 28 March 2026 (UTC)[reply]
A diaphoneme = whatever symbol(s) given its own row on Help:IPA/English and shown as is when put into {{IPAc-en}}. I'm not talking about what analysis is theoretically sound, just what should happen to Help:IPA/English / IPAc-en, which is what this talk page is about.
If you don't think /ɜː/ should be its own diaphoneme, how do you think fauteuil /ˈfəʊ.tɜː.i/ (CEPD) should be represented using this key? Nardog (talk) 10:47, 31 March 2026 (UTC)[reply]
Does anyone actually pronounce it /ˈfəʊtɜːi/, though? (It took me a while to recognize this as a cognate of our (Polish) fotel). Sol505000 (talk) 11:49, 31 March 2026 (UTC)[reply]
I don't think diaphoneme is the right word to use then, because that's not how the term is used in phonology, and I imagine you know this. It very much is used in relation to how units are distributed and analyzed cross-dialectically (hence the dia- prefix).
I would still suggest //œ// for fauteuil, of which /ɜː/ is one possible realization, but other common realizations exist such as /ɔɪ ~ eɪ/. This avoids the erroneous non-rhotic subgrouping, as different dialects realize //œ// in a variety of ways, and very word-dependently. Right now the table implies a stability that simply doesn't exist. It's better for us to directly recognize that instability. ~ oklopfer (💬) 12:09, 31 March 2026 (UTC)[reply]
For clarity, my recommendation for //œ// is based on a combination of American dictionary usages (which regard both rhotic /ɜːr/ and non-rhotic /ɜː/ as subset realizations/nativizations of the vowel, alongside the many others (/oʊ, ɔɪ, uː, ʌ/ etc) depending on the word) as well as the actual phonetic etymology of the majority of the loans (pho and Henan are the only two examples we have gone over so far which are not close- or open-mid front rounded in their origin languages), for which some English speakers will still use the vowel, even if rare. The overarching category is historic close- or open-mid front roundeds, or in the case of pho and Henan, close-mid back or central unroundeds (I can see a stronger argument for keeping //ɜː// for these, but I have another suggestion below for this*).
If we note that /ɒ̃, æ̃/ are more often realized/nativized as /ɒn, æn/, I see no reason why the same logic should not be applied here. As I (misguidedly) mentioned on Template talk:IPAc-en, "/ɜː/ (non-rhotic only)" is far too narrow. Both rhotic and non-rhotic speakers exhibit a wide variance of etymological /œ/; likewise, regarding it as a binary of only /ɜː/ and /ɜːr/ is plainly wrong, as we have gone over above. As I also stated there, just as we use //r// for an overarching/blanket representation for various rhotic realizations, despite not being realized as [r] in most dialects, we can apply parallel logic here for //œ//.
* As an additional suggestion, rather than keeping a dialect-specific diaphoneme in the table (a contradiction based on the definition of a diaphoneme), I think it should be considered to add a parameter to {{IPAc-en}} to specify |rhotic=n, such that /-r/ is removed from display in vowels /ɑːr, ɛər, ɜːr/ etc, or otherwise using the language label (uk etc). In intervocalic examples, this parameter could either be ignored (Lua is very much capable of if-then-else statements for this), or simply split into e.g. ə|r, so that an example like perjury with |rhotic=n remains correct. If such behavior seems desirable, I'd even be happy to write the code for it and submit a template edit request once it is thoroughly sandbox tested. ~ oklopfer (💬) 16:24, 31 March 2026 (UTC)[reply]
Evidently the maintained separation of /ɜː/ from /ɜːr/ has also led to confusion by editors in usage of the template: https://en.wikipedia.org/w/index.php?title=Special:Search&limit=250&offset=0&profile=default&search=hastemplate%3A%22IPAc-en%22+insource%3A%22%C9%9C%CB%90%22&ns0=1
By my count, out of the ~94 articles using |ɜː as a parameter of {{IPAc-en}} (counted by filtering out |[ɜː, |/ɜː, link|ɜː, IPA|ɜː from the linked search):
  • 19 use |ɜː|r rather than |ɜːr
    • 3 use |ɜː|(|r|)
  • ~25 have orthographic ⟨r⟩ in the words they are transcribing, meaning by principle rhotic speakers would pronounce them as /ɜːr/, yet rhoticity is entirely ignored and only |ɜː is provided
    • Only one of these uses the |UK or |UKalso parameters, the rest treat it as universal
  • A singular but still notable example, perfume uses |ɜː for UK and (only) |ər for US, despite the fact that the stress for the word could happen on either syllable in AmE (therefore it should be using |ɜːr without a dialect parameter and |ər with |USalso)
    • Somewhat similarly, Perseus makes a dialect split between |ɜː|r|. and |ɜː|., but that one does not have a stress difference (therefore it should just be using |ɜːr|.)
That's ~48/94 (Perseus only counted once) using |ɜː problematically in regard to rhoticity. That should be particularly striking as it means more often than not, usage of the vowel for the template is erroneous.
Of the remaining bunch, only ~28 provide |US or |USalso. That leaves as many as 18 treating |ɜː as the only possibility cross-dialectically, when it assuredly is not. Several of those are foreign places and names with orthographic ⟨ö⟩, such as Jönköping, Königsberg, Wladimir Köppen.
Edit: I've since reduced most of these errors, apart from the ones with alternative vowel pronunciations. ~ oklopfer (💬) 02:07, 2 April 2026 (UTC)[reply]

Superscript or other form of marking for reduced sounds

[edit]

I already proposed this a while ago, but too little participation in the discussion resulted in no decision. I recurrently come across edits that misinterpret or more simply overlook this key, therefore using the symbols in a way that we do not intend as part of our conventions. Particularly, the /ər/ diaphoneme often ends up being replaced by /ə/, /ə(r)/, /r/ or /(ə)r/ depending on the specific case. In order to prevent having to run after such continuous non-conventional changes, I truly believe we should do something about the way we transcribe some of these multiple-character phonemes—namely /əl/, /əm/, /ən/, /ər/ and all the other /r/-ending vowels that can occur both prevocalically and preconsonantally. We could implement the automatic insertion of parentheses, replace baseline characters with superscripts, or some other unambiguous solution, idk. Of course there would be a need for technical expertise to implement the /-r/ cases in the module, as they would change depending on what follows, however I’d love to hear more opinions on this in the hopes we can find a more stable alternative to the current key. ~ IvanScrooge98 (talk) 22:02, 22 April 2026 (UTC)[reply]

(For reference, here’s the link to my original proposal and related discussion from last year: Help talk:IPA/English/Archive 28#Introducing superscripts. ~ IvanScrooge98 (talk) 22:49, 22 April 2026 (UTC))[reply]

If the problem is people writing variant forms, then surely the easiest solution would be to build correction features into the IPA template (if that's possible)?
If the problem is multiple-character phonemes being confusing, then I think parentheses seem like the best solution, considering (1) consistency with current Wikipedia usage (even though it isn't technically standard IPA) and (2) the possible confusion of superscript <j> with palatization (/θj/ /tj/ /nj/ etc.) (as mentioned in the last discussion). Also we could then have /(h)w/ for /hw/.
Finally, I'll repeat @Nardog's old suggestion to have the module automatically detect the surrounding sounds and modify the appropriate letters to prevent syllabic nasals coming after nasals (and other such cases), so that it's here. [citation unneeded] (talk) 01:01, 23 April 2026 (UTC)[reply]
Retrieved from "https://en.wikipedia.org/w/index.php?title=Help_talk:IPA/English&oldid=1355574091"