Wikipedia talk:WikiProject Linguistics/Archive 17

This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

I don't mean to whine or provide baseless criticism of our articles on North American English, but there are things that need to be improved:

• We need to work on the phonemic side of the articles. Lexical sets aren't phonemes and neither are diaphonemes (nor are lexical sets diaphonemes).
• Again, we need to stop confusing diaphonemes with phonemes. Non-rhotic varieties *do not* have /ɑr/ in start.
• We're overusing phonetic transcription, especially when we're discussing phonemic contrasts. Misreading phonemic symbols is on the reader, not on those who edit the page (also, we don't need to treat our readers as incompetent - give them some benefit of the doubt).
• We need to specify when PRICE- and MOUTH-monophthongization result in a phonemic merger with /ɑ/.
• In varieties with the short-a split, we need to specify the phonemicity of [eə], because to me it looks like it could be considered a phoneme separate from /ɑ/ (perhaps belonging to the SQUARE class /ɛə/ in non-rhotic varieties - is bath the same as *bareth in traditional NYC English?)
• We need to specify what happens with /æ/ before /ŋ/. In varieties in which it merges with /eɪ/ before /ŋ/, rang should be transcribed /reɪŋ/ as this is a phonemic merger (yes, I know that /eɪ/ can't occur before /ŋ/, at least not in citation forms - alternatively, you could say that words with TRAP before /ŋ/ are automatically transferred to the FACE class). Also, isn't this merger universal in AmE?
• We need to specify whether START is merged with NURSE in those Canadian varieties that raise the latter vowel. If so, why don't we cover START-NURSE merger on Wikipedia? There's no mention of it anywhere.
• We need to deal with the phonemicity of NEAR, CURE, SQUARE, NORTH and FORCE (or, perhaps more often, NORTH/FORCE) in dialects with variable rhoticity, which would lose their phonemic status and become FLEECE, GOOSE, FACE, THOUGHT (or LOT/PALM) and GOAT (or THOUGHT/GOAT in the case of the NORTH-FORCE merger, with the phonemic identification being dependent upon the presence vs. absence of the cot-caught merger) as speakers restore the postvocalic /r/, perhaps with distinct allophones (but still allophones) before coda /r/. NEAR, CURE, SQUARE, NORTH and FORCE are kind of a Schrödinger's cat (not really, but you know what I mean) in certain kinds of AmE - they're both phonemic and non-phonemic. How do we cover that? I'd say: two vowel tables and transcribing both varieties simultaneously (e.g. /ˈskwɛə, ˈskwer/ in the case of NYC English).
• We need to stop saying that diphthongs (including FLEECE, GOOSE, FACE and GOAT) are fronted, raised, retracted, etc. and always specify which element of them is fronted, raised, etc.
• We need to stop saying that FACE etc. is laxed and specify that it's the first element that is laxed.
• We need to pay closer attention to FACE-monophthongization. I've heard idiolectal (near-)monophthongization from especially women from pretty much all around the country (though maybe that's an exaggeration), and those are speakers with strongly diphthongal GOAT.
• Is the FACE-FLEECE merger possible in any parts of the US? I've mistaken /e/ realized as [ei] (with a fully front, close-mid starting point) for /i/ several times. I've only heard women using this pronunciation. Surely there's a possible, variable merger in perception (not necessarily in production, though)?
• We need to specify the roundedness of the first element of GOAT (unless it's a monophthong - then, of course, we don't). We also need to deal with the o-schwa (schwo?) in window, fellow, etc.
• We need to specify the roundedness of NURSE and LETTER.
• In the case of non-rhotic varieties, do they merge NURSE with LETTER? In other words, are the second syllables of forward and foreword distinct or the same? Do speakers of otherwise non-rhotic varieties (or those accents that are variably rhotic) that have a rhotic NURSE distinguish forward from foreword by dropping the /r/ in forward?
• We need to specify what happens to intervocalic /ɪr/ and /ʊr/ (especially the latter) in all accents and how distinct they are from /ir/ and /ur/.
• We need to deal with CURE words in our articles and what happens with them in terms of *phonemic* mergers.
• We need to pay closer attention to the phonetic realization of SQUARE and the first vowels of MERRY and MARRY. Especially the first one is often closer to [e] and we don't seem to cover that.
• We need to specify accents in which there's a complete /t/-/d/ merger (manifested mostly [only? I'm not sure] as the same duration of the preceding vowel) in flapping environments. In those accents better probably needs to be transcribed /ˈbɛdər/ (or /ˈbɛdə/, if the accent isn't rhotic).
• Perhaps our coverage of fronting of /u/ and /ʊ/ could be improved and the phonetic qualities specified (including the quality of the first element of /u/ - remember that it's a diphthong like FACE and GOAT), especially if it involves unrounding in addition to fronting (as it so often does in the UK).
• If there are regions in which FACE and GOAT are monophthongal, are there regions in which FLEECE and GOOSE are monophthongal as in Scotland and South Africa? And I mean truly monophthongal, so that see and too (i.e. words with stressed word-final /i, u/) are consistently said with monophthongal vowels by the majority of speakers.
• We need to specify accents with the bunched /r/ and those in which /r/ is flapped after /θ/. Perhaps thrust should be phonemicized as /ˈθdʌst/ for those who flap the /r/? Guy Fieri pronounces his surname /fiˈɛdi/, with a flap, so maybe it makes sense to write it that way. Personally, I'd be perfectly happy with writing /ˈθɾʌst/ in accents where flapping of /r/ after /θ/ is consistent and where /t/ is fully merged with /d/ in flapping environments (though first we'd need to switch over to ⟨ɹ⟩ in the case of the rhotic for this to make logical sense: /ˈbɛɾəɹ/).

Kbb2 (ex. Mr KEBAB) (talk) 09:58, 20 May 2020 (UTC)

EDIT: Let me list sources that use ⟨e o⟩ for American FACE and GOAT right here, so that the list is easily accessible.

• A Pronouncing Dictionary of American English (1944)
• Handbook of the IPA (1999), as well as its predecessor The Principles of the IPA (1949)
• At least one JIPA article on AmE, American English: Southern Michigan (2003)
• Wells in his Accents of English (1982) writes GenAm FACE with ⟨eɪ⟩ and GOAT with ⟨o⟩, so he's inconsistent. Later, in Longman Pronunciation Dictionary, he switched over to ⟨oʊ⟩ for GOAT.

The original post is below.

(This is mostly regarding the ongoing discussions at Talk:Standard Canadian English).

I believe that FACE and GOAT should be written with ⟨e⟩ and ⟨o⟩ in our articles about North American English. The way they behave in that dialect group pretty much seems to mirror Northern Standard Dutch /eː/ and /oː/, which are mandatorily diphthongized to [ei ou] in most positions, yet they are realized as [eə oə] or [eː oː] before /r/ and as [eː oː] before /w/ (in the case of /eː/) and /j/ (in the case of /oː/). The reason they're transcribed with ⟨eː oː⟩ also has a lot to do with Belgian Standard Dutch, in which the closing diphthongs [ei ou] don't occur and the monophthongs are used in all positions.

In NAE (especially in the US), FACE and GOAT pattern with monophthongs (again, precisely as in Dutch), specifically FLEECE, GOOSE and PALM. FLEECE and GOOSE are variably diphthongized and PALM is monophthongal. Using ⟨e o⟩ would be more consistent with the other three tense vowels. /ɔ/, which I believe is also tense, is also monophthongal in GA (though diphthongal variants are found in e.g. New York - strangely, nobody seems to want to transcribe them with ⟨oə⟩ in phonemic transcription).

In some dialects they are monophthongal anyway. Plus, some diphthongal realizations sound almost indistinguishable from monophthongs (for instance, the allophone of FACE that sounds like [ei], with a close-mid starting point to me sounds like FLEECE when I don't pay attention to the context. Similarly, Canadian [ou], a back diphthong from close-mid to close sounds almost monophthongal to me).

In North American English the realization [eə oə] is either almost universal or at least widespread before coda /l/, as in 'scale' and 'goal', which are more commonly pronounced [skeəɫ ɡoəɫ] (or perhaps even [ɡoɫ]?), rather than [skeɪɫ ɡoʊɫ]. That's another argument to use ⟨e o⟩.

Then there's also the o-schwa in window, which can be monophthongal in any accent (AFAIK): [ˈwɪndo].

In dialects that have the father-bother, cot-caught and the north-force mergers (and I mean all three of them), the vowel in 'core' is usually identified as GOAT, which is a more economical analysis. For this specific purpose, ⟨o⟩ is a much better choice ([oʊr] is an impossible realization in most if not all dialects, unless we're talking about GOAT + /r/, with a /#/ before the consonant - but that's not the NORTH/FORCE vowel!)

I also propose that we retranscribe the NORTH/FORCE vowel with ⟨or⟩ in all dialects that have the merger, regardless of whether they have the cot-caught merger. The only exception to that could be New York English and Philadelphia English, where THOUGHT itself is very close. But in other regions it's (AFAIK) actually more open than NORTH/FORCE. This means that ⟨or⟩ is a better representation anyway, and by using it we'd achieve partial harmony with Australian, New Zealand and South African Englishes, varieties where ⟨oː⟩ is typically used for NORTH/FORCE. In RP, /ɔː/ has the same quality in all positions.

Remember that we universally use the symbols ⟨ɪ i ʊ u ɛ ʌ ɔ æ⟩ in our transcriptions of NAE, even where they represent values closer to e.g. [e ɪi ɵ ʉ æ ɐ oə a] (not that these specific values can be found in any specific accent all at once, I'm aware of that) or similar. I see no reason not to treat FACE and GOAT the same and use the broad transcription ⟨e o⟩. ⟨oʊ⟩ isn't that narrow anyway and the vowel is better transcribed ⟨ɵʊ⟩ or ⟨əʊ⟩ for a number of speakers, yet these symbols aren't used even in accents that centralize the first element. If ⟨e⟩ for FACE is an unnecessary deviation from the Gimsonian system, then so is ⟨oʊ⟩ for GOAT - it could be safely written ⟨əʊ⟩, as in RP. I know that it used to be a back diphthong in RP, but still.

The symbol ⟨e⟩ stands for DRESS in RP, that's true. But it also stands for FACE in Scottish English, a major dialect group in the UK, precisely where RP is spoken (not necessarily in that specific area, I mean the British Isles excluding the Republic of Ireland). Kbb2 (ex. Mr KEBAB) (talk) 10:21, 14 April 2020 (UTC)

A lot of this makes sense. My only concern would be whether this choice will make our transcriptions less readable to our audience. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 20:41, 14 April 2020 (UTC)

We should generally use ⟨e⟩ and ⟨o⟩ to refer to the FACE and GOAT vowel classes, and should be transcribed using slashes (e.g. /fes/ and /got/). Where monopthongization is discussed, dipthongal and monopthongal variants should be transcribed using braces (e.g. [feis] vs. [fes]) to show that we're discussing the phonetic realization of the broader vowel class. This would give us something like "Dialect X tends to have monopthongal variants of the FACE and GOAT vowel classes. While speakers of Language Y tend to pronounce goat as [goʊt], speakers of Dialect X produce a monopthong similar to [got]." — Wug·a·po·des​ 21:02, 14 April 2020 (UTC)

I'm down with using ⟨e, o⟩ in descriptions of NAmE so long as we're consistent (with all monophthongs transcribed as short, as in Kenyon & Knott). Nardog (talk) 22:44, 15 April 2020 (UTC)

It seems to me the phonetic arguments are besides the point. Instead, we should ask what the most common transcription in the relevant literature is. Using the most common transcription allows us to meet the expectations of readers familiar with the subject and not to mislead the others. I am no expert in North American English. It does not seem implausible to me that monophthongal transcriptions are more common, but apparently there are diphthongal transcriptions as well.

Another argument is our “diaphonemic” transcription system. Obviously, an article discussing North American English pronunciation must not be constrained by our “diaphonemic” system. However, in cases like ⟨eɪ⟩ vs. ⟨e⟩ there is a choice between a symbol that matches our “diaphonemic” system and another symbol that does not. In that situation, I believe we should prefer by default the symbol that matches our “diaphonemic” system – unless we are very sure that the other symbol is really more common. Using the symbol that matches our “diaphonemic” system is more helpful to our readers. It allows them to identify the sounds in the articles about North American English pronunciation with the transcriptions we provide e.g. of North American names. --mach 🙈🙉🙊 20:21, 16 April 2020 (UTC)

/e o/ for FACE and GOAT and /or/ for NORTH–FORCE in General American and Canadian make sense to me, though I don't know if it's justifiable based on sources. I dislike the transcription /oʊr/ because it suggests [owər] to me, which is very much not how most people pronounce it. (Similarly with /eɪr/ if it comes up, which suggests /ejər/; I prefer /er/.) As User:Kbb2 mentions, the offglide is less likely to occur in that position. However, that isn't much of an argument. Keep in mind though that I'm probably biased because I live in a region where these vowels tend to be monophthongal.

"North American English" might be over-broad; there might be dialects (such as Southern) ones in which face and goat should be transcribed phonemically as diphthongs. But the issue probably doesn't come up much because Wikipedia articles on these dialects usually don't go so far as to choose dialect-specific symbols for the vowel phonemes. — Eru·tuon 04:57, 24 April 2020 (UTC)

@Wugapodes, Nardog, J. 'mach' wust, Erutuon, and Wolfdog: So where do we go from here? Can anyone list sources that use ⟨e o⟩, apart from the ones already listed? I'd add that monophthongal [i u ~ ʉ] for FLEECE and GOOSE seem to be a L2 influence by definition (in Scotland, Wales, South Africa, Singapore, etc., just as in Germany and the Netherlands), especially in stressed syllables and even more so word- and phrase-finally. That's another argument to use the simple ⟨e o⟩ for FACE and GOAT. Kbb2 (ex. Mr KEBAB) (talk) 09:04, 20 May 2020 (UTC)

I certainly understand Kbb2's point. Yes, I grew up as an American English speaker and language-head who assumed FACE and GOAT were pure monophthongs (they were often called "long a" and "long o" respectively) though they are surely not, phonetically speaking, nor are they in most North American accents. Clearly, though, the overriding factor here (thank you J. 'mach' wust), is what the most common transcription in the relevant literature is. To protect against us relentlessly squabbling over every detail ("I like these two suggested notations but not that one") or to make our decisions here less vulnerable to later editors' upheavals or slippery slopes, shouldn't we simply lean on what the sources actually say? When we make decisions as a group, the best safeguards are not our collective hunches, but rather actual sources/scholars. This is a Wikipedia cornerstone. As far as I see, the most common phonemic transcriptions for FACE and GOAT are the diphthongal ones, among modern phonologists. So let's identify some! We have diphthongs used by the biggest name in North American English, William Labov, as well as his sometime-colleague Charles Boberg. We have them used by British linguist Peter Trudgill to apply to North American phonemes and even John C. Wells if you look at his more recent works since his 1982 Accents of English (such as the Longman Pronunciation Dictionary or Sounds Interesting; I'm not sure if he ever definitively writes about why/where he made the switch). And these are just four linguists acclaimed enough to have their own Wikipedia articles. Do we have any heavy hitters who do not use the diphthongs? Why would we transcribe sounds one way when the linguistics scholarship does it another way?

Kbb2: Here are some more specific responses to your comments above:

• FLEECE and GOOSE are variably diphthongized and PALM is monophthongal. True. Probably even usually diphthongized, though less so for FLEECE.
• /ɔ/, which I believe is also tense, is also monophthongal in GA It's probably about equally diphthongal or monophthongal: [ɔ̞~ɔ̞ə~ɒ~ɒə].
• though diphthongal variants are found in e.g. New York - strangely, nobody seems to want to transcribe them with ⟨oə⟩ in phonemic transcription True, but I'm confused why is this strange.
• In some dialects they are monophthongal anyway. Yes, but only a minority. Generally not in GenAm, the West, Midland, South, Northeast, eastern Canada, or younger/urban Canada in general.
• 'scale' and 'goal', which are more commonly pronounced [skeəɫ ɡoəɫ] (or perhaps even [ɡoɫ]?), rather than [skeɪɫ ɡoʊɫ]. I agree with your description of scale completely, but actually [ɡoʊɫ] for goal sounds perfectly fine/normal to me.
• Then there's also the o-schwa in window, which can be monophthongal in any accent (AFAIK): [ˈwɪndo]. I don't agree with this at all. Do you have any support for this? To me, [ˈwɪndo] certainly sounds Minnesotan.
• I also propose that we retranscribe the NORTH/FORCE vowel with ⟨or⟩ in all dialects that have the merger... The only exception to that could be New York English and Philadelphia English, where THOUGHT itself is very close If THOUGHT is so close, then NYC and Philly need not be exceptions, if we do end up agreeing on this system. See ANAE p. 124 for what Labov et al. call the "back vowel shift before /r/" and show "/ahr/ → /ohr/ → /uhr/", which we, in our current transcription, might render "/ɑ.r/ → /oʊ.r/ → /u.ər/". In other words, that middle one is FORCE. (Side-note: the more I think about it, the more I'm realizing how useful lexical sets are in general, often preferable to phonemes. 1) They are more reader-friendly for nonexperts. 2) They are less vulnerable to editors bickering over this or that choice of phonemic transcription.)
• ⟨oʊ⟩ isn't that narrow anyway and the vowel is better transcribed ⟨ɵʊ⟩ or ⟨əʊ⟩ for a number of speakers ...or, for a number of speakers, <ʌʊ> or <ɔu> or <ɔw> or <ɐʊ> or (in Labov's vein) <ow> or any number of other options spread across the country, sometimes not even consistent within a dialect. So we should probably go with whatever is used in a majority of sources or by the most respected sources, no?

Wolfdog (talk) 12:57, 20 May 2020 (UTC)

@Wolfdog: I'm not sure what "yes" means in this context. I've never said they were phonetic monophthongs, I'm saying that there's some (or more than "some") evidence that they are phonological monophthongs as in Northern Standard Dutch. It's an entirely different thing. At least in the case of GOAT, a pure monophthong sounds either strongly regional or foreign - exactly like a monophthongal [oː] when you're a native speaker of Dutch from Randstad.

We shouldn't agree on any changes based on sources that don't discuss phonemes or don't seem to deal well with phonemicity (when lexical sets are used instead of phonemic symbols, or when those lexical sets themselves are misapplied).

I just said why: because of the allophony before /l/ and because in accents that merge both cot with caught and north with force there's no [ɔ] outside of /r/ contexts (it doesn't appear anywhere but before /r/). It only makes sense to consider it to belong to the GOAT set /o/, but the thing is that before /r/, /o/ is never a closing diphthong. Also, it's just the monophthongal GOAT that sounds regional, monophthongal FACE is much less marked and I bet you don't hear it half of the time, unless you pay close attention ([ei], with a fully front and close starting point sounds very like a monophthong to me and I've mistaken it for /i/ more than a couple of times). Also, it's because FLEECE and GOOSE aren't transcribed with ⟨ɪi ʊu⟩ and those are just as likely to be diphthongal. Especially FACE can be almost as narrow as FLEECE in terms of tongue movement. Using ⟨e o⟩ makes for a more consistent transcription.

It's strange because it's another inconsistency.

By [ˈwɪndo] I mean [ˈwɪndö̞], with the final vowel being more front and lower than cardinal [o]. It's GenAM GOAT with deleted second element.

How so? Those dialects feature [oə] for THOUGHT, and schwa onglide before /r/ is very usual. IMO, we ought to use the phonemic symbol for THOUGHT in those dialects: [foɹ] sounds much closer to /f/ + THOUGHT + /r/ "faw + r" than /f/ + GOAT + /r/ "fo + r".

Being able to read and understand phonemic transcriptions is a fundamental skill for anyone that wants to read articles such as Boston accent, California English, etc. It's as essential as being able to understand English. If our reader doesn't know that /ə/ stands for an open vowel [ɐ] in the utterance-final position in English (as opposed to not being lowered in such manner in German), that ⟨ɑ⟩ can stand for a central vowel just as it can denote a back one or that /æ/ stands both for the open monophthong [æ] and (before nasals) the diphthong [eə] (well, at least in General American - but /eə/ seems to be phonemic in e.g. New York), they have no business reading those pages. There are limits to being noob-friendly and it can't be done at the cost of those who have done their homework. You're proposing that we accomodate those who can't distinguish phonemes from allophones, which isn't a reasonable proposition in the case of AmE as it has straightforward allophony in most cases. Perhaps there's an equivalent of WP:COMPETENCE that addresses the readers, rather than the editors - that's what I mean, though being able to work with phonemic transcription is a fundamental skill for those who want to edit those pages, too (I'm not saying that you're lacking in that area, I don't know that. If your preference is to avoid discussing phonemicity [for whatever reason] then that's fine, but you shouldn't expect other editors to do the same).

For instance, the Bostonian contrast between dark and dock is best shown using phonemic transcription as /ˈdak/ vs. /ˈdɒk/ (whatever symbols you want to use). This is much less cumbersome than having to write PALM–START vs. LOT–THOUGHT–NORTH (yeah, why omit the last one?)

That's why I'm asking for more sources that use ⟨e o⟩. Kbb2 (ex. Mr KEBAB) (talk) 16:56, 20 May 2020 (UTC)

"Yes" in my case means "Indeed" or any other transitional term equivalent to "It is true that...."

I didn't mean to sound like I was arguing one side or the other in the old phonological vs. phonetic debate; what I was arguing, in fact, was that sources/scholars should trump our own personal thoughts on the matter. In fact, our arguments should only be rooted in sources/scholars.

We shouldn't agree on any changes based on sources that don't discuss phonemes or don't seem to deal well with phonemicity (Where do you see that happening? They're certainly not happening among the scholars I mentioned) ...or when those lexical sets themselves are misapplied (Again, where do you see that happening? Are you suggesting the experts themselves are misapplying their own systems??)

The makes for a more consistent transcription argument, though great, is still trumped by the actual sources/scholars argument. Otherwise, we're in the land of original research.

Clearly, then, [ˈwɪndo] cannot be safely characterized as a GenAm transcription. ([ˈwɪndoʊ], on the other hand, can.)

In response to your How so? comment where you argue about the best way to transcribe four, why don't you just use actual sources/scholars to back up your arguments? It's very possible, for example, that Wells here agrees with you. Unfortunately, I don't have his works readily in front of me. Perhaps you can mine some sources.

The entire argument that Being able to read and understand phonemic transcriptions is a fundamental skill is merely your opinion. Other opinions exist, all logically valid yet conflicting. Certainly, articles could be edited to be more reader-friendly. One could just as easily argue the exact opposite of what you did and it would sound just as logical: "There are limits to using expert jargon and it can't be done at the cost of those who are new to the topic." See? Neither extreme will win this argument. At some point, we have to resign ourselves to the fact that Wikipedia is a place of compromise. How do we move forward in this compromise? At the very least... surprise, surprise: use actual sources/scholars.

Notice that your Boston point again goes into the realm of what is best instead of the realm of what "is actually done" by sources/scholars. You're in opinionland again.

Finally, your final sentence reveals that you seem to have your premise wrong. Instead of "here's the notation I use" (including a very thoughtful rationale behind it) and therefore That's why I'm asking for more sources that support it, you should turn your premise around in the other direction: we ought to start with the sources, which will then provide us whichever notation is typical. Wolfdog (talk) 21:01, 20 May 2020 (UTC)

@Wolfdog: Ok. Thanks for clarifying.

Both issues are present in Handbook of the Varieties of English, but they're not very pervasive. I can't point them out now, I'd have to go through that source. As far as your sources are concerned - I don't know whether that's the case.

Lexical sets are Wells's invention. I'm not saying that he's misapplying them. Also, there are limitations when it comes to lexical sets - for instance, there are no lexical sets for prevocalic /ɪr, ʊr, ɛr, ʌr, ær, ɒr/.

Maybe it can't. EDIT: Let's test it - listen to File:En-us-borrow.ogg, File:En-us-fellow.ogg, File:En-us-potato.ogg, File:En-us-sorrow.ogg, File:En-us-tomato.ogg, File:En-us-tomorrow.ogg and File:En-us-window.ogg. Are those pronunciations what you'd call General American?

Well, /ɔr/ is the most common way to phonemicize the merged NORTH-FORCE sequence, so there's hardly any need for me to provide sources. The question should be: why should we write it /ɔr/ in the case of New York and Philadelphia and /or/ elsewhere, and the answer is that because THOUGHT is lower in other accents and it tends to merge with LOT-PALM anyway. In the latter case, there's no phonemic /ɔ/ anymore and NORTH-FORCE is best phonemicized as /or/, so either way we end up writing /ɔr/ in some articles and /or/ in others (EDIT: ...as long as we don't implement my proposition of writing /or/ everywhere where there's no contrastive NORTH). Why shouldn't it be /or/ in New York? Because /ɔ/ is too close and the merged NORTH-FORCE vowel is the same or almost the same as THOUGHT (and much different from stressed GOAT) when it comes to quality.

The entire argument that "Being able to read and understand phonemic transcriptions is a fundamental skill" is merely your opinion. Other opinions exist, all logically valid yet conflicting. I think I'm gonna go to Talk:Planck constant and complain about not being able to understand the numerous mathematical equations (or whatever they are - I can't even call those properly) present in the article. Can we stop being ridiculous?

I find your consistent push for replacing phonemic transcription with other things (lexical sets, diaphonemes, etc.) to be disruptive and I'd like to stop discussing it here. You'll have a hard time convincing anyone (myself included) not to use it and it's not what this discussion is about anyway (are you really trying to tell me that "replacing lexical sets with phonemic transcription" is somehow going against the sources?). Let me repeat: being able to read phonemic transcription is a skill that's as essential as being able to understand English when it comes to articles like Boston accent. Throwing around accusations of being "opinionated" when I'm stating what the common cross-language practice is when it comes to presenting accents or saying that I'm perpetuating OR when what I'm doing is the standard practice of the International Phonetic Association (see any article from the Illustrations of the IPA series, many describe English accents) is not going to help you.

When to use phonemic IPA instead of phonetic transcription is another question. When you're presenting phonemic contrasts, then of course it's better to use phonemic slashes, per International Phonetic Association. In that way, readers who have done their homework and are able to distinguish phonemic and phonetic transcriptions are less likely to be confused. If it's going to make editing harder for anyone, then the correct course of action is to improve one's understanding of the International Phonetic Alphabet and how it's applied to English. Being able to use phonemic transcription as well as being able to differentiate it from phonetic transcription is, again, an essential skill for anyone who wants to edit articles such as Boston accent. WP:COMPETENCEISREQUIRED. At the very least, it's true when speaking generally. Of course any given edit may or may not involve using phonemic transcription - but that goes without saying, no?

As you can see, I've provided the list of references at the very top of the discussion. Kbb2 (ex. Mr KEBAB) (talk) 04:17, 22 May 2020 (UTC)

Regarding examples of the misapplication of lexical sets in the Handbook of Varieties of English, see e.g. the description of LETTER and COMMA in White South African English: Bowerman says that COMMA is "usually [ə], but may be as open as [ɐ] in Cultivated WSAfE; and also in Broad varieties close to Afrikaans English." To me this is a pretty clear misunderstanding of how lexical sets work - COMMA merely signifies the /ə/ phoneme that appears as the unstressed vowel in about, balance and comma regardless of its phonetic quality. Bowerman's wording suggests that he thinks that COMMA refers to word-final schwas.

Right above his description of COMMA, he says that LETTER is "[ə] is all varieties; very often omitted before another consonant: [ˈkɪtn̩] kitten", which is clear BS - it suggests that LETTER can contrast with COMMA based on vowel height, or that COMMA doesn't participate in syllabic consonant formation. In addition to that, the unstressed vowel of kitten is COMMA, not LETTER!

I'm really not a fan of describing those lexical sets that are merged separately (which really is pervasive in that book), especially when you then fail to copy-paste the entire description. We do need to be careful about which sources we follow.

On page 296, Gordon misapplies phonemic transcription, writing "/ɜ/ ~ /ʌ/" for STRUT and that "/ɑ/ is fronted to /a/". Kbb2 (ex. Mr KEBAB) (talk) 11:17, 22 May 2020 (UTC)

Let's just respond to all of your paragraphs with one sentence each: A published, respected work like Handbook of Varieties of English has more credibility on Wikipedia than you, me, or even the most unfailingly logical and high-minded genius in the world. It would be very surprising for Wells to misapply his own invention. Those sound files that end in the GOAT vowel indeed sound like GenAm to me. There's a need for you to provide sources pretty much whenever an editor asks; to quote Aeusoes1 in recent response to me wondering "Is a citation really needed?" when an editor asked for a citation to back up information that is widely known within the linguistics community: "Yes, even if it's obvious to us". I'm no longer really interested in the Philly/NYC debate (I was never saying you're wrong) -- you're missing that all I've been continually asking for is sources, and you've been predominantly giving back to me mere arguments. If the Planck's constant article is too difficult for you, you have every right to complain on the talk page without an accusation of being ridiculous. Sure, let's stop discussing lexical sets, even though you use them as often as I do -- I was trying to provide another possible alternative and had no idea it was so egregiously disruptive. I've been able to present respected sources/scholars that disagree with and diverge from what you believe to be the common cross-language practice, in reality some decent variety of different cross-language practices, and therefore, your absolute statements about only the former without regard for the latter reveals that, like any of us human mortals, you are in fact opinionated after all; also, any one of us is perpetuating OR when we don't back up our words with sources/scholars, and I'm glad to see that you are now, finally, beginning to bring up some actual sources: the very thing I requested all along. COMPETENCEISREQUIRED cuts both ways -- to such a notable degree, in fact, that an entire section exists on that page entitled What "Competence is required" does not mean. I'm legitimately glad that you provided the list of references at the very top of the discussion, which is why I publicly thanked you once you did, though to be fair to me you've only just done this since we were last corresponding -- again, though, this is wonderful, productive, and much-appreciated. Yes, Bowerman does seem to be wrongheaded or perhaps he simply left in a mistake that even he would recognize if he had just had a more astute proofreader; the Handbook is nearly 2500 pages long, chock-full with the writings of esteemed sociolinguists, and I doubt the whole book's validity should be called into question based on this, but I certainly see what you mean about lexical sets being misapplied in that chapter. I'm not sure I agree that Gordon's descriptions are misapplications (the meaning of his /ɜ/ does boggle me), though I do admit they could be better rewritten.

I expect you will hate this, but, still, Bowerman and Gordon (because they are published scholars in their field) have more credibility on Wikipedia than a system either you or I would design with no sources at all. That said, I value your sentiment that we do need to be careful about our sources. Let me reiterate on a positive note: though I wish you had done this earlier, I've appreciated you actually finding some specifics for me to look at, in your comment above. Wolfdog (talk) 16:15, 22 May 2020 (UTC)

@Wolfdog: I'm not calling into question the credibility of the entire Handbook but parts of it (again, not the majority of the book) that are obvious BS. Also, remember that neither of the examples that I've provided concern the phonemic transcription of FACE and GOAT in GA. Those are mere examples of how scholars themselves can make mistakes. The fact that they are mistakes is easily proven using other sources.

I hear them as [ˈb̥ɔɹö̞], [ˈfɛɫö̞], [pʰəˈtʰeɾö̞], [ˈso̞ɹö̞], [tʰəˈmeɾö̞], [tʰəˈmɔɹö̞], [ˈwɪndö̞], not only with a monophthongal final vowel but also with a stressed monophthongal FACE in potato and tomato. In these recordings: File:En-us-Cambodia.ogg, File:En-us-Oklahoma.ogg, File:En-us-Minnesota.ogg, File:En-us-Romania.ogg and File:En-us-Wyoming.ogg I hear stressed monophthongal GOAT and FACE (paired with unstressed monophthongal GOAT in the case of Romania): [kʰæmˈbö̞djɐ], [ˌö̞ʊkɫ̥əˈhö̞mɐ] (look at that: a monophthong in the syllable with primary stress and a diphthong in the initial syllable with secondary stress), [ˌmɪnɨ̞ˈsö̞ɾə], [ɹoˈmeniɐ], [waiˈö̞mɪŋ]. If they also are GA, you can see where this is going. It proves that what Wells said in Accents of English holds up to scrutiny in the 21st century: FACE and GOAT can be monophthongal in GA, at least when they aren't both word-final and stressed - although I don't know whether [ˈd̥e] for day truly isn't GA, given how narrow FACE can be when diphthongal (and how close can diphthongal FACE be for some women, almost [i̞i] - how isn't that a [variable] FACE-FLEECE merger?)

Again, you're putting words in my mouth. What I called ridiculous was that you think that those that want to read articles written in 5%-15% in IPA don't need to understand phonemic transcription or how it differs from phonetic transcription. What it is is treating our readers as if the majority of them were either unintelligent or unwilling to improve their knowledge of the IPA. Cut them some slack. I do use lexical sets and will continue to do so, and that's not what that part of my reply was about.

The cross-language practice I'm talking about is using phonemic transcription and strongly differentiating it from phonetic transcription.

Why are you talking about a system based on no sources? Let's stay on topic please. Kbb2 (ex. Mr KEBAB) (talk) 09:55, 23 May 2020 (UTC)

Thanks for your clarification about the Handbook. Again, I'm not sure why we're stuck on the phonetic monophongization of GOAT or FACE in American English; you keep encouraging us to stay on topic please, but I'm happy to stray in those directions if you want: yes, those recordings mostly sound like General American (though the pronunciations of Minnesota ((practically [mɪnɪˈsʊ̞ɾə]!)) and Romania sound like Northern regional accents to me, and Wyoming e.g. sounds like a slight diphthong to me) and just the very fact that you're asking me for my personal impression means we're drifting from sources anyway, which was my whole problem earlier. That said, GOAT and FACE certainly can be monophthongs in certain situations: e.g. for GOAT, of the File:Mid back rounded vowel.ogg type but not the cardinal File:Close-mid back rounded vowel.ogg type, which is marked and Upper Midwestern-like. More important to me, if we really want to stay in phonetics, is that broadly phonetic [oʊ] is less marked than broadly phonetic [o]: i.e. [mɪnɪˈsoʊɾə] sounds inarguably GenAm while [mɪnɪˈsoɾə] is marked assuming cardinal values and requires diacritics to bring it back to a GenAm quality (that cardinal sound is even a stereotype among Minnesotans themselves of how they pronounce their state's name; I can find a video if you want). Kbb2... come on... the reason I'm talking about a system based on no sources is because this is precisely what you were presenting up until May 20 or May 22, and even at that point you had the wrong idea of requesting sources to back up your notation rather than, as I've already said, start[ing] with the sources, which will then provide us whichever notation is typical. I'm glad you've since found some sources. We can put that to rest. As long as you have those to back you up and they're stronger (more seminal / more respected / more current) than the sources I've offered, your system can prevail. If you'd like to continue discussing phonetics, why don't we take that to one of our talk pages (since this page is being absolutely cluttered)? Wolfdog (talk) 11:03, 23 May 2020 (UTC)

I just want to nip something in the bud real quick without having to respond to everything else, but Kb linking to WP:COMPETENCEISREQUIRED was about a point about editors in general and Wolfdog seems to have taken it as a personal slight about their own personal level of expertise, which it was not (if this is not how Wolfdog read it, then I have no clue what COMPETENCEISREQUIRED cuts both ways is supposed to mean). — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 19:32, 22 May 2020 (UTC)

Nope, I did not take it as a personal slight (although, Kbb2 and I are indeed each feeling we are making some other personal slights) -- but, rather, just a reminder/suggestion. My point about "cutting both ways" is that WP:COMPETENCEISREQUIRED mentions not only that perceived disruptors need to have a certain level of competence, but also the editor who interacts with or calls out the perceived disruptor is expected to give some slack, be civil, improve the other's competence, etc. (or else they themselves have become disruptive). By the way, although I'm admittedly somewhat heated, I do believe/understand that Kbb2 is legitimately well-intentioned and just trying to better these articles. Wolfdog (talk) 22:02, 22 May 2020 (UTC)

@Wolfdog: I apologize if my tone was less than civil. Though I don't think that it's my role to improve other's competence. That's on them, obviously. I'm no teacher of phonetics anyway. Kbb2 (ex. Mr KEBAB) (talk) 08:20, 23 May 2020 (UTC)

Tl;dr but what do you mean, "where do we go from here"? As far as I can see, Aeusoes1, Wugapodes, me, Mach, and Erutuon all condoned use of ⟨e, o⟩ in articles about NAmE varieties, albeit with some caveats. I don't see any other "way to go" than to implement that. Nardog (talk) 00:03, 21 May 2020 (UTC)

Nardog: to be clear, actually: you, Aeusoes1, and Wugapodes condoned it with caveats. Erutuon certainly feels mixed. Mach and I did not condone it. And here's a very quick and easy-to-read nutshell of my argument: we need to use actual sources (and, AFAIK, the most relevant ones use diphthongs). To be clear, that includes Labov & Boberg's Atlas of North American English (2006) (or any works by those two), Wells's Longman Pronunciation Dictionary (2008), Trudgill & Hannah's International English (2008), and the Cambridge English Pronouncing Dictionary (2006). Wolfdog (talk) 11:14, 21 May 2020 (UTC)

That's what I mean too. There doesn't seem to be a consensus about this, and listing more sources than just Kenyon & Knott would certainly be helpful. Another source is Ladefoged (1999) "American English" in Handbook of the International Phonetic Association.

EDIT: [1] uses ⟨e o⟩ too, so there's an established practice in the JIPA to use those symbols in transcriptions of AmE. The American version of The North Wind and the Sun in the Principles of the IPA (1949) also uses ⟨e o⟩, exactly like the Scottish transcription that's right above it. So it seems that ⟨e o⟩ is the transcription of choice of the International Phonetic Association, as far as AmE is concerned, because the Principles of the IPA is the predecessor of the Handbook of the IPA (1999). Kbb2 (ex. Mr KEBAB) (talk) 12:02, 21 May 2020 (UTC)

There's no consensus in the literature on which one to use because no one cares as long as you're consistent and clear about the conventions. You can even make up your own system if you want, which is what Labov did, and his transcription system is still used by pretty much only him. The literature is just a bunch of house styles and author preferences, so it doesn't make sense to try and read some deeper meaning into them. Authors transcribe as specifically as they need to for the purposes of their paper, and we should do the same. All the conventions are used, and they're all fine. We should just pick one and move on. — Wug·a·po·des​ 01:38, 23 May 2020 (UTC)

A transcription system has been picked. This discussion is about changing it. I know that some people equate transcription systems to mere styles and think they are exchangeable. This point of view is not shared by all. Unlike style elements (e.g. color, font, or size), transcription symbols carry their own well-defined meaning. --mach 🙈🙉🙊 07:01, 23 May 2020 (UTC)

@J. 'mach' wust: Both ⟨eɪ oʊ⟩ and ⟨e o⟩ have a well-defined meaning in NAE dialectology, so I'm not sure what you mean. Kbb2 (ex. Mr KEBAB) (talk) 08:18, 23 May 2020 (UTC)

@Wugapodes: I know. What I meant is that we don't have a consensus here, among ourselves. Kbb2 (ex. Mr KEBAB) (talk) 08:18, 23 May 2020 (UTC)

I was wrong about Mach but this is indeed getting absurd. FWIW another work that transcribes FACE and GOAT as ⟨e o⟩ is Fromkin et al.'s An Introduction to Language, a highly popular textbook. ⟨e o⟩ are well established in American literature. Nardog (talk) 13:43, 23 May 2020 (UTC)

An editor wants the IPA in the article Phonics checked and is also asking if replacing the IPA in the article with something more accessible would be a good idea (Talk:Phonics#Changes related to the use of IPA). Nardog (talk) 17:32, 24 May 2020 (UTC)

Hi, could someone familiar with German phonology add the IPA transcription for Grassmann's name? lammbda^talk 08:07, 9 June 2020 (UTC)

 Done. Love —LiliCharlie (talk) 12:32, 9 June 2020 (UTC)

The terms "Aurat", "Arvad", "Avret", and "Awrath" may refer to: Women of Asian religious or cultural descent and identity.

Self nomination for AFD since article copy pasted to Draft:Aurat for incubation because IMHO current article title Aurat (word) is misleading and confusing leading to western systemic bias and stifling the article growth. Please find Detail reason at Wikipedia:Articles for deletion/Aurat (word)

I invite project members to review current and potential sourcing and weigh in on the AfD discussion. Thanks! Bookku (talk) 02:49, 10 June 2020 (UTC)

Members of the Linguistics WikiProject might be interested in a discussion for the article List of "-gate" scandals about which words with this suffix should be included on the list. The discussion can be found on the talk page Talk:List of "-gate" scandals#Add. Thanks for any input. Umimmak (talk) 20:22, 13 June 2020 (UTC)

I believe the list will be known as the Wikigate in a week or two. Love —LiliCharlie (talk) 20:35, 13 June 2020 (UTC)

LiliCharlie, that would imply a scandal involving Wikipedia (and it wouldn't be the first). I think it might rather become known as Gategate. Mathglot (talk) 01:37, 17 June 2020 (UTC)

Your feedback would be appreciated at Talk:Source language (translation)#Merger proposal. Thanks, Mathglot (talk) 01:37, 17 June 2020 (UTC)

The article Verbosity currently cites a number of authorities advising against wordiness, but seemingly none critiquing this advice. Any sources or other contributions would be appreciated. See Talk:Verbosity#POV Issues. Cnilep (talk) 02:39, 26 June 2020 (UTC)

I just made a first pass at reworking the article on the source-filter model to organize it better and attempt to explain the concept in a more accessible way. However, I believe it would benefit from other editors taking a look. It still needs a lot of in-line citations, and I don't have access to all of the original articles (for example, Chiba & Kajiyama) so I can place them. Does anyone want to check it out? Emflazie (talk) 19:38, 2 July 2020 (UTC)

@Emflazie: Send me an email with the papers you need and I'll send any pdfs I have. I'll also take a look at the article. — Wug·a·po·des​ 22:41, 2 July 2020 (UTC)

Wugapodes, thank you. This clarifies a lot. Emflazie (talk) 03:24, 3 July 2020 (UTC)

There's a long-standing merge proposal for Set phrase and Fixed expression, which might would benefit from some input from this project; both are short and almost unreferenced, and I wonder whether idiom might be a better target. The discussion is at Talk:Set phrase#Merge proposal. Klbrain (talk) 21:17, 14 April 2020 (UTC)

  Y Merger complete. Klbrain (talk) 08:38, 6 July 2020 (UTC)

Hi there is an RfC on the Cognitive-Theoretic Model of the Universe that may be of interest to this project. See: Talk:Cognitive-Theoretic Model of the Universe#Request for comment: on the notability of the CTMU in 2020 with sources published after 2006 and "unredirect" of this page to Christopher Langan - Scarpy (talk) 06:45, 8 July 2020 (UTC)

So, {{IPA-en}} is strange. It's used on three pages. Perhaps it was mass removed/replaced with {{IPAc-en}}? Why does it still exist, then, if deprecated? Nothing on the doc page either. The French version is used on 284 pages, odd, much more popular.[2]

It has no error checking. I came across this revision of Chewa language as it appeared from 20 May 2020 until I edited it just now.

With no parameters at all, it used to look like this:

English pronunciation: /{{{1}}}/

However, it's not really possible to call it without parameters and not notice it as an error, so the more common mistake is to try e.g. {{IPA-en||pron}}, which would show:

pronounced //

I figured we may as well get some use out of this template until we figure out what to do with the whole family of them, so to quickly fix Chewa language I made it so if the label is empty, only the {{small}} appears. So:

Really, what should happen is, an overarching Lua module covering all the {{IPA-xx}} templates should be made; so one module to cover {{IPA-tl}}, {{IPA-es}}, {{IPA-de}}, and so on, using a similar system to {{cite book}} and its ability to figure out the names of languages from internationally standard codes like nan-tw, ja, etc.

The same module can probably also cover {{IPAc-xx}} as much of the work is very similar, just the parameter layout is different. A good thing is Lua can take arbitrary numbers of template parameters, so {{IPAc-ja}}'s silly limit of 34 can easily be raised. {{IPAc-de}} supports 50. Definitely the lack of uniformity can confuses people, especially when 34, even 50 is not really that much. There's also the problem that some templates randomly use slashes and others brackets, and there's no logical reason for this that I know of. {{IPA-en}}, slashes. {{IPA-eo}}, square brackets. Of the two forms, most of them use brackets. English is the only major language I see with slashes.

It looks like Mr. Stradivarius and Nardog already got the ball rolling over at Module:IPAc-en. Module:IPAc-en/phonemes is cool.

So, without further ado:

Proposed game plan

1. Break the {{IPAc}} redirect
   1. First change all 8 transclusions to {{IPA}}, which will be left alone I suppose.
   2. ...perhaps eventually deprecate {{IPA}} and redirect it to {{IPA-}}, or {{IPA-}} to it. ┐(´ ∀｀ )┌
2. Overwrite Erutuon's Module:IPA, which isn't used in the mainspace and is abandoned since 2017;
   1. Give it two functions, cc, like IPAc-xx, and p, like IPA-xx.
3. Create {{IPA-}}
   1. Template just an {{#invoke:IPA|p}}
   2. To get the old {{IPA-en|ˈnjændʒə|pron}} output, now write {{IPAc|en|ˈnjændʒə|w<!--ith-->=pron}}
4. Add a function to Module:IPA, pc ("Old IPA-xx argument style compatible")
   1. Start replacing {{IPA-en}}, etc., with Module:IPA invokes: {{#invoke:IPA|lang=en|pc}}
5. Create {{IPAc}}
   1. Template just an {{#invoke:IPA|c}}. It calls out a lot to other modules like Module:IPAc-en.
   2. To get the old {{IPAc-en|ˈ|n|j|æ|n|dʒ|ə}} output, now write {{IPAc|en|ˈnjændʒə}}
   3. Add a |w= (with) here too to stop hacks like the one I did to Chewa language from proliferating
      1. So, to get {{IPA-en||pron}}{{IPAc-en|ˈ|n|j|æ|n|dʒ|ə}}, just do {{IPAc|en|ˈnjændʒə|w=pron}}
6. Create Module:IPAc-ja, Module:IPAc-de, etc. Work them into {{IPAc}}, the parent template, as we go.
7. Add a function to Module:IPA, cc ("Old IPAc-xx argument style compatible")
   1. Start replacing {{IPAc-en}}, etc., with Module:IPA invokes: {{#invoke:IPA|lang=en|cc}}

Discussion

Ping line

@Gonnym, LiliCharlie, Trappist the monk, Kbb2, Aeusoes1, and Nardog: This is all very involved, but things will work much more smoothly when done I believe, and there will be more uniform presentation. I'm happy to work on some of this, but I don't know if I can finish it all alone in any quick timeframe. What do you all think? Psiĥedelisto (talk • contribs) ^{please always ping!} 09:33, 9 July 2020 (UTC)

• I'm always in favor in reducing many identical language templates to one general one with a consistent style. This has recently been done with {{In lang}} which was the result of a TfD which deleted hundreds of near identical templates. --Gonnym (talk) 09:40, 9 July 2020 (UTC)
• I don't think your proposal fully accounts for the existing complexity/ununiformity of IPA(c)-xx templates, but I support consolidating the IPA-xx templates to one {{IPA}} (using Lua) and making a wrapper module for all IPAc-xx. Like I said in the previous discussion, I see the handling of whitespace and labels as the biggest problems with these templates. I have modules sitting in my sandbox but the lack of response to the last discussion discouraged me.
  (The use of slashes vs. square brackets isn't random at all. See Phonetic transcription#Narrow versus broad transcription, Help:IPA#Brackets. Help:IPA/English is the only key that uses a phonemic transcription, because readers of the English Wikipedia are expected to be familiar with the phonological system of English, and to account for the variety of accents they would have. The other keys use (rather broad but) phonetic transcriptions, because we can't expect them to be familiar with the phonology of non-English languages.) Nardog (talk) 10:03, 9 July 2020 (UTC)

  @Nardog: Wow, very interesting about the brackets, thank you! I learned something today. So, that's something we'd need to preserve then in the #invoke:. On another template I've contributed to, {{unichar}}, I did it by defining a parameter |br= expected to be a string of length 2, then just Module:String's substring stuff. So, {{unichar|40|br=[]|sans=y}}: U+0040 @ COMMERCIAL AT 🙌 About the complexity of the templates, for sure you are right, and it's a rough start, hopefully we will add to it as we go along if this goes from proposal to actually in the mainspace. Best, Psiĥedelisto (talk • contribs) ^{please always ping!} 10:41, 9 July 2020 (UTC)

  The choice between slashes and brackets would depend on the key, not the transcription. If we were to consolidate IPA-xx to IPA, then we would most likely have a data module storing the correspondence between languages, keys to link, and dialects, so the type of brackets could easily be dealt with there. It would look something like this:

  ["en"] = {
  	key = "Help:IPA/English",
  	slashes = true,
  	dialects = {
  		["uk"] = "British English",
  		["us"] = "American English",
  	},
  },
  ["sv"] = {
  	key = "Help:IPA/Swedish",
  	dialects = {
  		["fi"] = "Finland Swedish",
  	},
  },
  

  Nardog (talk) 10:48, 9 July 2020 (UTC)

  The dialects list should probably not duplicate any language module that already has that data. @Trappist the monk: is there one that handles dialects? --Gonnym (talk) 11:00, 9 July 2020 (UTC)

  No dialects list that I'm aware of. I presume that the above examples relate to IETF language tags: en-UK, en-US, sv-FI where UK, US, and FI are region subtags derived from ISO 3166-1. Module:Lang/data maps some lang-region tags to adjectival language names that link the tag to its en.wiki article but Module:lang doesn't attempt to create anything like 'Finland Swedish' from sv-FI. Do I remember correctly? Weren't you looking for an official source of country names in their adjective forms? Did you ever find such a source?

  —Trappist the monk (talk) 17:47, 9 July 2020 (UTC)

  You remembered right. I did look for something and ended up creating a module Module:Country adjective and basing the data on the list at List of adjectival and demonymic forms for countries and nations. Not sure if it useful here though. --Gonnym (talk) 01:30, 10 July 2020 (UTC)

  @Nardog: Sure, but could there not be a writer who can do narrow transcriptions of non-English languages? Many editors speak more than one language. So maybe we need both? Or is it more that the reader isn't likely to understand the nuances of a narrow transcription so it doesn't matter to mark one as such? Psiĥedelisto (talk • contribs) ^{please always ping!} 11:01, 9 July 2020 (UTC)

  We just write {{IPA|[...]}} for that. There's also {{IPA-all}}. The IPA-xx templates that link to specific keys under Help:IPA/ are exclusively for transcriptions that adhere to those keys, which exist to help readers figure out what each symbol means—otherwise readers would be utterly confused. See Wikipedia:Manual of Style/Pronunciation. Nardog (talk) 11:06, 9 July 2020 (UTC)

• (Also, IPAc-en supports labels too—they just have to be the first argument. (Except lang, which converts to nothing. It's a long story!)) Nardog (talk) 10:30, 9 July 2020 (UTC)

Is there a template I can use, or some wiki markup, to generate a non-underlined hyperlink? I'm working on a couple articles on proto-Semitic, which by convention uses underdots to symbolize emphatic consonants and underlines to symbolize interdentals. When wikilinking a single character, this looks pretty bad. The IPA template turns off link underlining, which is handy, but IPA notation isn't appropriate in context, as these usages require traditional semiticist transliteration. --April Arcus (talk) 18:40, 13 July 2020 (UTC)

@April Arcus: Template:Nounderlines. Nardog (talk) 21:39, 13 July 2020 (UTC)

There is a discussion proposing that part of the Phoenician alphabet article might need to be split off into an article that is to be named Canaanite scripts. Please comment at Talk:Phoenician_alphabet#This_page_might_need_to_be_split. Debresser (talk) 12:09, 26 July 2020 (UTC)

See https://en.wikipedia.org/w/index.php?title=Nabataean_alphabet&action=history . This edit is wrong for reasons explained at Talk:Nabataean alphabet#Problems in the table. Due to a combination of Coronavirus isolation and the stupid encryption protocol upgrade, I'm editing with a non-fully-Unicode-compliant tool, so I can't do it myself right now... AnonMoos (talk) 21:54, 26 July 2020 (UTC)

FYI I have opened a discussion about the rationale of the new dab page in Talk:Islamic language. –Austronesier (talk) 07:13, 28 July 2020 (UTC)

Could somebody look at Draft:Powari language. It's been kicking around WP:AfC for two years. We're trying to figure out if it's worth keeping, but needs a Subject-matter expert. -- RoySmith (talk) 13:35, 28 July 2020 (UTC)

I have collected some articles which have language- or linguistics-related links to DAB pages, where expert attention would be welcome. Search for "disam" in read mode and for "d" in edit mode, and if you solve any of these puzzles remove the {{dn}} tag and post {{done}} here.

• Bilingual lexical access  Done –Austronesier (talk) 18:03, 28 July 2020 (UTC)
• Crimea  Done Cnilep (talk) 04:10, 30 July 2020 (UTC)
• Jaranwala Tehsil  Done — kwami (talk) 18:45, 10 August 2020 (UTC)
• Komi alphabets  Done — kwami (talk) 18:45, 10 August 2020 (UTC)
• List of indigenous languages of South America  Done Cnilep (talk) 04:10, 30 July 2020 (UTC)
• List of Indo-European languages  Done Cnilep (talk) 04:10, 30 July 2020 (UTC)
• Negative verb  Done –Austronesier (talk) 14:16, 28 July 2020 (UTC)
• Panoan languages  Done Cnilep (talk) 04:10, 30 July 2020 (UTC)
• Phonology  Done Umimmak (talk) 14:21, 28 July 2020 (UTC)
• Preposition and postposition  Done –Austronesier (talk) 18:07, 28 July 2020 (UTC)
• Primitive Irish  Done Umimmak (talk) 14:21, 28 July 2020 (UTC)
• Tape language  Done –Austronesier (talk) 14:32, 28 July 2020 (UTC)

Thanks in advance, Narky Blert (talk) 14:03, 28 July 2020 (UTC)

Need more eyeballs at Third-person pronoun, which is an OR disaster. I'm particularly interested in the table in this section, especially the portion under the header, Gender-neutral singular pronouns. This entire portion of the table should come out; it is populated by hapax and other obscure science fiction trivia, and is completely unencyclopedic. Thanks, Mathglot (talk) 08:25, 4 August 2020 (UTC)

I have seen this page a while ago and closed it with mixed feelings of shudder and cringe, for various reasons. It's not about third-person pronouns in general. Much of it is about gender-distinctions, with a gender-distinction-normative bias: e.g. pronouns in gender-neutral languages are called "gender-inclusive", even if these languages never had an exclusion problem. Plus lengthy stray material that is not about third-person at all, e.g. the subsection Rapa; not to mention the non-notable hapax from hobby conlangs. –Austronesier (talk) 11:18, 4 August 2020 (UTC)

I moved it to Gender neutrality in languages with gendered third-person pronouns and removed all the sections on languages that didn't fit the topic, which left only Swedish and Norwegian and examples of the opposite trend it CJK. Please don't hesitate to try something else. — kwami (talk) 18:48, 10 August 2020 (UTC)

Wikipedia:Reliable sources/Noticeboard#etymonline could use input from this project's participants. Nardog (talk) 13:35, 13 August 2020 (UTC)

Kwamikagami edited {{IPA non-pulmonic consonants}} to show all click letters with velar/uvular symbols preceding them, saying ⟨ʘ ǀ ǃ ǂ ǁ⟩ only symbolize releases according to the the Handbook of the IPA, taking the fact pp. 20–1 of the Handbook show ⟨k͡ʘ k͡ǀ k͡ǃ k͡ǂ k͡ǁʰ⟩ to illustrate the sounds through examples from ǃXóõ and Xhosa. p. 10 of the Handbook says:

'Velaric' airstream sounds, usually known as 'clicks', again involve creating an enclosed cavity in which the pressure of the air can be changed, but this time the back closure is made not with the glottis but with the back of the tongue against the soft palate, such that air is sucked into the mouth when the closure further forward is released. The 'tut-tut' or 'tsk-tsk' sound, used by many English speakers as an indication of disapproval, is produced in this way, but only in isolation and not as part of ordinary words. Some other language s use clicks as consonants. A separate set of symbols such as [ǂ] is provided for clicks. Since any click involves a velar or uvular closure, it is possible to symbolize factors such as voicelessness, voicing, or nasality of the click by combining the click symbol with the appropriate velar or uvular symbol: [k͡ǂ ɡ͡ǂ ŋ͡ǂ q͡ǃ].

This doesn't strike me as saying the letters must be used with velar or uvular symbols, especially when the letters are seen by themselves in works, including the IPA Illustration for Sandawe.

Is it true that ⟨ʘ ǀ ǃ ǂ ǁ⟩ can only represent releases? How should our {{IPA non-pulmonic consonants}} be arranged? Nardog (talk) 08:27, 10 August 2020 (UTC)

I don't see that the Handbook says affricates "must" be written with two symbol either, they just give a couple examples written that way on p. 22. The examples are all we have to go on.

It's very common to leave out the accompaniment if it's <k>. However, it's less common to do so if there's a velar-uvular distinction to worry about.

Before the Handbook, the illustrations of the IPA implied that the clicks needed to letters but then wrote the tenuis clicks with only one. That was evidently cleaned up for the Handbook.

As far as the Sandawe illustration, note that the many of the illustration in the Handbook omit the tie bars from the affricates (Igbo even uses the tie bar bar for labial-velars but not for affricates!), but it's still proper to write them with.

When giving a prescriptive explanations, we should be careful not to take shortcuts that might confuse the reader. We can always explain common shortcuts and use them ourselves in the language articles, but when readers refer back to the IPA articles they should be clear on official usage.

For several years, there was a shift toward using the simple click letters for the complete consonant, with diacritics for anything else except for the velar-uvular distinction. Now usage seems to be swinging back. There's an upcoming phonetics volume on clicks coming out, and much of the transcription is explicit about both places, though typically using superscripts rather than tie bars (a convention also commonly seen for affricates and labial-velars). — kwami (talk) 08:45, 10 August 2020 (UTC)

The official IPA chart only says "Affricates and double articulations can be represented by two symbols joined by a tie bar if necessary" (emphasis added). I don't understand how that's germane even if what you say about clicks were true. Nardog (talk) 08:51, 10 August 2020 (UTC)

It's germane because it's exactly parallel. You object that the IPA doesn't explicitly require two letters for a click, but then it doesn't explicitly require two letters for an affricate either. So, should we assume that the second letter is optional? We can only go by the examples. The Chart doesn't give any, the Handbook does, and in the 'Guide to IPA notation' all clicks and affricates are written by two letters joined with tie bars. — kwami (talk) 08:56, 10 August 2020 (UTC)

It's possible to do all sorts of things with the IPA. It's possible to use ⟨c⟩ for an affricate, for example, but that doesn't mean we should define it as one. — kwami (talk) 08:56, 10 August 2020 (UTC)

What the IPA doesn't explicitly require for an affricate is a tie bar, not two letters. Nardog (talk) 09:02, 10 August 2020 (UTC)

It never says that two letters are required, and indeed in some of the illustrations affricates are written with only one letter (e.g. Korean, Sindhi).

Note also with Sandawe that they incorrectly transcribed the glottalized nasal clicks as ejective. They just adopted common conventions, they weren't being precise.

It's also quite common to use ⟨ɾ ɽ⟩ for laterals, e.g. in Indic languages, but that doesn't mean [ɾ ɽ] are lateral. — kwami (talk) 09:05, 10 August 2020 (UTC)

(edit conflict) But you can't possibly represent a double articulation with one letter, with a tie bar or not. So "if necessary" is clearly in reference to the use of a tie bar, not two symbols. I'm not saying you can't represent an affricate with one letter, just that a tie bar is only optional when representing an affricate according to the official chart. What the Igbo illustration does is actually a very sensible choice: [tʃ, dʒ] are in fact transitions from [t, d] to [ʃ, ʒ], while [k͡p, ɡ͡b] not so much, following closely the IPA Principle #4 (c).

By this analogy, which you brought up, then, the Handbook is saying that ⟨ʘ ǀ ǃ ǂ ǁ⟩ can indeed represent entire clicks. That's not quite the same as The Handbook requires them. What the Chart calls 'clicks' aren't consonants at all, just the releases, which is what you said at Template talk:IPA non-pulmonic consonants#velar vs uvular clicks. Nardog (talk) 09:25, 10 August 2020 (UTC)

You're reading a lot of your own opinions into the motivations of the writers of the Handbook. It makes just as much sense to write a labial-velar stop as ⟨kp⟩ in a language that doesn't contrast that with k + p as it does to write an affricate ⟨ts⟩ in a language that doesn't contrast that with t + s. And I don't see how you can possibly think that Principle 4(c) is any more relevant to one than to the other. — kwami (talk) 09:38, 10 August 2020 (UTC)

The question is whether you wish to be precise in your presentation of how the IPA works. In explaining the IPA, I feel that we should be precise. When being precise, [t͡s] and [k͡p] are one segment, [ts] and [kp] are two. [c] is a plosive, not an affricate. [ɽ] is central, not lateral. [ɨ] is a central vowel, not back. [k] is pulmonic, not ejective. [ǂ] is a click type, not a velar click. Outside pedagogy, it's fine to use the alphabet more broadly -- [ts] and [c] can both be affricates, [ɽ] can be lateral, [ɨ] can be a back vowel, [k] can be ejective, and [ǂ] can be a velar click. But you don't start off presenting them that way to people who are not familiar with how the IPA works. — kwami (talk) 09:35, 10 August 2020 (UTC)

As for the phrasing it is possible to symbolize factors such as voicelessness, voicing, or nasality of the click by combining the click symbol with the appropriate velar or uvular symbol, they're saying that it's convenient to transcribe features that have nothing to do with the rear articulation as if they were part of that articulation rather than of the entire consonant. The voicing and nasalization don't belong to the rear articulation -- that's only specified for uvular-velar, affrication, ejection and the like. The IPA transcription of clicks is weird, like writing labial-velars as *⟨k͡p, g͡p, ŋ͡p⟩. A lot of the variability and debate in transcription is related to this fact, that the IPA letters for clicks don't really fit in with the rest of the alphabet. I suppose that one might address this by transcribing clicks as e.g. ⟨k͡ǂ̥, ɡ͡ǂ̬, ŋ͜ǂ̃⟩, but I've never seen anyone do that. — kwami (talk) 09:55, 10 August 2020 (UTC)

The Chinese Wikipedia's IPA article uses the diacritics for voicelessness and nasalization to modify click symbols, see zh:國際音標#非肺部氣流音. Love —LiliCharlie (talk) 04:51, 12 August 2020 (UTC)

@LiliCharlie: I'm sure that's just copied from a version of our {{IPA non-pulmonic consonants}} before recent edits. What I want to know is the community's opinion on how that template should be presenting the links to articles about clicks. I for one think the articles themselves (most of which are unreferenced) are hardly notable and should be merged into just Bilabial click, Dental click, Alveolar click, Retroflex click, and Palatal click, so that then the template can simply have a row of ⟨ʘ ǀ ǃ ǂ ‼ ǁ (ʞ)⟩ much like the actual IPA chart. But even barring that, do we need both velar and uvular symbols, making the rows twice as tall? My understanding that the use of velar symbols has been far more prevalent, even if the actual posterior closure of such clicks may be more accurately described as uvular. Nardog (talk) 05:07, 12 August 2020 (UTC)

Following Ladefoged & Maddieson (1996:265–266) velar and uvular symbols are both required to account for a phonemic contrast in ǃXóõ. However Miller et al. (2007) say in their study of Nǀuu that "evidence suggests that the contrast between “velar” and “uvular” clicks proposed for the related language ǃXóõ is likely also one of airstream and that a contrast solely in terms of posterior place would be articulatorily impossible." Love —LiliCharlie (talk) 05:50, 12 August 2020 (UTC)

@LiliCharlie: The Cornell link is dead, here's the new URL: [3] –Austronesier (talk) 10:11, 12 August 2020 (UTC)

Good spot, thanks for pointing out and providing a working link, Austronesier. My outdated link was actually to the earlier 2007 version of the study that was submitted to JIPA where it was published in 2009. I've managed to find the 2007 version archived on WaybackMachine, but the 2009 JIPA version is even better and certainly more accessible. Love —LiliCharlie (talk) 17:07, 12 August 2020 (UTC)

Pace Miller, there's at least one language that distinguishes velar from uvular clicks without any airstream contour. You hear it in the vowel rather than in the release of the click. (Miller discovered when working on N|uu that in that language where a velar-uvular distinction had been posited, all clicks were uvular and the distinction was one of timing, e.g. [q͡ǂ] vs [ǂ͡q] -- that is, whether or not you could hear the uvular release. She suggested that all languages had uvular clicks only in this fashion, and that "velar" clicks simply had an inaudible back release. But it turns out that not all of them do -- some have only velar clicks, and some have both. Why she should say that such an easy distinction might be "articulatorily impossible" is beyond me. They're easy to articulate and the spectrograms are pretty clear, with e.g. a velar pinch after velar clicks.) But regardless, the question here is what is the IPA convention, not what is Miller's. Lots of people use the bare click letter for a tenuis velar click. And that's fine, if you're not sticking to strict IPA. The IPA itself did that when it introduced the Beech letters in 1923. But the 1999 Handbook -- the replacement for the 1949 Principles so they could accommodate the Kiel convention that had replaced the original click letters with the ones we see now -- doesn't take such shortcuts in the examples it gives.

(Side note, Sandawe and Hadza have only (somewhat backed) velar plosives and ejectives, and clicks at the same rear place of articulation. Some Khoe langs have both velar and uvular plosives and ejectives, and clicks at both places of articulation. If you were looking only at those languages, it would be natural to conclude that clicks are doubly articulated. But Xhosa has only velar plosives and ejectives and only uvular clicks. If you were to look only at that language, it would be natural to conclude that the uvular closure is part of the airstream mechanism, not a place of articulation. So there's plenty of reason for theoretical differences between phoneticians, which are reflected in how they choose to symbolize clicks.)

As for merging the articles, that would effectively be saying that click consonants aren't important enough to bother distinguishing. It's as if a French-speaker were to say that our articles on affricates should be merged into the corresponding fricatives because affricates aren't important -- after all, they don't occur in French and there are no IPA letters for them. — kwami (talk) 04:51, 14 August 2020 (UTC)

The reference is: https://www.sil.org/system/files/reapdata/17/94/43/17944307071856830681259992851645036622/Hajong_Phrasebook.pdf

Publisher: SIL

Context: A citation to this was flagged ([4]). The issue at the heart is the alphabet used for this language. The language is strongly associated with an ethnic group and there is an effort to preserve and advance it into a literate form. Currently a number of different scripts are being used by the speakers, who are distributed in Assam in India and Bangladesh. The scripts in use are Latin, and Bengali-Assamese script, and from among these both the alphabets, Bengali and Assamese, are being used.

The point here is, is this reference reliable for the claim that the Assamese alphabet is one of the alphabets used to write the Hajong language, from the examples in the book and from the sentence - "Each word in the word lists is written first in Roman script followed by Assamese script in brackets." (p.1)

Thanks! Chaipau (talk) 15:13, 4 August 2020 (UTC)

@Chaipau: Sorry, I have just noticed your question here. The answer is yes, it is just as reliable as the source (Ethnologue) for the use of the Bengali alphabet in the same box of the table. Both sources are published by SIL. But actually, I find this SIL source[5] most enlightening (p. 27), which is also cited in Hajong language. This survey indicates that some Hajong speakers actually are less sensitive to these details than Bengalis/Assamese would expect. –Austronesier (talk) 09:48, 14 August 2020 (UTC)

@Austronesier: You have gone ahead and responded to the tag as well. Thank you! Chaipau (talk) 11:50, 14 August 2020 (UTC)

Hey There,
many psychological pages have no etymology whatsoever
ie "panic attack" does not refer to Pan. Goddess Psyche is never mentioned anywhere
please upgrade the articles
thanks Linguists --Wittgenstein51 (talk) 19:08, 15 August 2020 (UTC)

Please go ahead and add what you deem to be missing. −Woodstone (talk) 07:33, 16 August 2020 (UTC)

Regarding templates within Category:Unicode charts, would there be a reason that the superscript numbers at the top could not be replaced by, for example, letters, to better distinguish them from article references? This would make them more clearly linked to the template notes they apply to. (Drmccreedy,BabelStone) CMD (talk) 14:52, 19 August 2020 (UTC)

@Chipmunkdavis: Personally, I find these less confusing/irritating than the notes/references in Help:IPA/English :) But sure, there is no reason not to convert them into something that better meets common expectations, even if not prescribed by MOS. –Austronesier (talk) 15:31, 19 August 2020 (UTC)

Your feedback would be appreciated at this discussion, which proposes the merger of four articles into Anti-LGBT rhetoric. Thanks, Mathglot (talk) 18:23, 22 August 2020 (UTC)

FYI: https://www.theregister.com/2020/08/26/scots_wikipedia_fake/ meta:Requests for comment/Disruptive editing on sco.wikipedia on an unparalleled scale. Visite fortuitement prolongée (talk) 14:53, 26 August 2020 (UTC)

Hello etymology friends. I often consider adding an etymology section to articles without one, but I'm never sure if that's acceptable. It's not clear to me that there's a consistent threshold, if you will, even for what one would imagine to be the most vetted topics. For example, Tree, Future, and March (music) don't have etymology info, but Animal, History, and March (month) all do. What gives? What is the policy/common practice/tradition for including etymology on a topic's page?

• Does it depend on how notable the page is?
• How important the topic is?
• How "obvious" and/or well-known it is what the etymology is?
• How attested the etymology is?
• Whether the origin is Germanic, Latin/French, or other?
• Whether the word is shared by other languages?
• How abstract the topic is?

• Should etymology be explained in a parenthetical in the lede? Or in its own section?

• Is there a policy at all?

CampWood (talk) 02:09, 30 August 2020 (UTC)

I have no idea if there is any guideline hidden somewhere, but intuitively I find an etymology section helpful if it explains how the concept described by the term developed, as in the case of History. For Tree, etymological information adds little to our understanding of what a tree is, so should be left out per WP:NOTDICTIONARY. –Austronesier (talk) 07:51, 31 August 2020 (UTC)

I'm just gonna leave a link to this discussion about using running numbering schemes for linguistic examples. The idea would be basically to have Wikimarkup support something like the LaTeX/linguex "\label" and "\ref" system. Botterweg14 (talk) 12:42, 31 August 2020 (UTC)