Wikipedia talk:WikiProject Linguistics

Welcome to the talk page for WikiProject Linguistics. This is the hub of the Wikipedian linguist community; like the coffee machine in the office, this page is where people get together, share news, and discuss what they are doing. Feel free to ask questions, make suggestions, and keep everyone updated on your progress. New talk goes at the bottom, and remember to sign and date your comments by typing four tildes (~~~~). Thanks!

We probably have too many vowel articles

Hello. I think that we have too many vowel articles, and at the same time too few to be consistent in the way we represent vowels on Wikipedia. I'll explain what I mean.

The article open back unrounded vowel currently covers four variants of [ɑ]: open back (close to cardinal [ɑ] or identical), open near-back, near-open back and near-open near-back. That's four variants of [ɑ] in one article. The article about the open front [ɶ] also covers open and near-open variants. This is how it should be: the vowel articles on Wikipedia aren't about cardinal vowels but vowels in world's languages that are close to any given cardinal vowel (except for vowels like [ɪ]).

Also, we don't cover all combinations of height/backness/roundedness in separate vowel articles on Wikipedia. That's another inconsistency. Here's my proposal, divided into parts:

Regarding true-mid vowels

(1) Merge the articles about true-mid vowels with those about close-mid ones:

• Mid front unrounded vowel with Close-mid front unrounded vowel
• Mid front rounded vowel with Close-mid front rounded vowel
• Mid back rounded vowel with Close-mid back rounded vowel

...and create separate sections for true-mid vowels that contrast with close-mid ones (or close-mid and open-mid ones) in those articles.

Reason: The true-mid [e̞, ø̞, o̞] rarely contrast with their close-mid and/or open-mid counterparts. There are obscure dialects or languages such as the Amstetten dialect of Bavarian or Kensiu that contrast three heights of mid vowels. Danish contrasts close-mid /ø/ with true-mid /œ/ (which is lowered to open-mid or near-open when in contact with /r/). These contrasts seem to be much rarer than those between close-mid and open-mid vowels, as in French, German or Italian. The IPA recommends that in cases like this (where the quality of a vowel is in-between two or more cardinal values), the simplest symbol be used unless there are convincing reasons (a phonemic contrast, traditional transcription in dialectology, etc.) to do otherwise. ⟨œ⟩ and ⟨ø⟩ are probably equally "complex", but ⟨e⟩ and ⟨o⟩ are simpler than ⟨ɛ⟩ and ⟨ɔ⟩ because they're ordinary Latin letters. Furthermore, the mid back rounded vowel is probably more of a variant of the close-mid [o] than just a vowel in-between [o] and [ɔ], because like [o] and unlike [ɔ] it can contrast with the (near-)open [ɒ] (which can be considered a subtype of [ɔ] as they sound so similar to each other) by height alone - see this article by Geoff Lindsey and our vowel charts (Geordie is probably no exception - the contrast between /ɒː/, /ɔː/, /oː/ and /uː/ is rather unstable). Lindsey's research might be a proof that essentially all (or nearly all) contrasts between rounded back vowels are subtypes of a /u–o–ɔ/ contrast, not matter the exact height, backness, length, etc.

Neither close-mid front unrounded vowel nor close-mid back rounded vowel cover *just* vowels in the close-mid front/back area but also their centralized counterparts (near-front in the case of [e] and near-back in the case of [o]). Also, compare e.g. File:Estonian vowel chart.svg with File:Finnish monophthongs chart.svg. Neither Estonian /e, ø, o/ nor Finnish /e, ø, o/ are exactly in-between cardinal [e, ø, o] and [ɛ, œ, ɔ] but closer than that (closer to the close-mid cardinals), it's just that the Finnish vowels (listed in the true-mid articles) are a tad more open than the Estonian ones (at least that's what the vowel charts suggest). This might suggest that there are some WP:CONTENTFORK issues with those articles.

Regarding near-close vowels

(2) Rename the articles about the mid-centralized cardinals [i, y, u]:

• Near-close front unrounded vowel back to Near-close near-front unrounded vowel
• Near-close front rounded vowel back to Near-close near-front rounded vowel
• Near-close back rounded vowel back to Near-close near-back rounded vowel

...and move fully front and fully back near-close (un)rounded vowels to close front unrounded vowel and close back rounded vowel, respectively (we don't list any fully front variants of the near-close near-front rounded vowel on WP and the distinction between front and near-front rounded vowels is mostly meaningless anyway). The near-close near-front rounded vowels typically transcribed with ⟨y⟩ (those in Dutch, French and Hungarian) should be moved back to close front rounded vowel and we should specify their height there.

Reason: The articles about [ɪ] and [ʊ] shouldn't cover peripheral vowels but only the centralized ones, which means that we probably should rename near-close front unrounded vowel and near-close back rounded vowel back to near-close near-front unrounded vowel and near-close near-back rounded vowel, which are the official names of the IPA symbols ⟨ɪ⟩ and ⟨ʊ⟩. Fully peripheral near-close vowels, at least of the unrounded front and rounded back type, are just variants of [i] and [u]. The older IPA practice was to prefer the symbols ⟨i, y, u⟩ over ⟨ɪ, ʏ, ʊ⟩ and use the latter when necessary. Also, when a language contrasts near-close front/back with close front/back vowels based on height alone, the near-close front/back vowels are almost never written with ⟨ɪ, ʊ⟩ but ⟨e, o⟩. Near-close front rounded vowel should probably be renamed to near-close near-front rounded vowel for consistency.

(3) Merge the articles about the near-close central vowels with those about close central ones:

• Near-close central unrounded vowel with Close central unrounded vowel
• Near-close central rounded vowel with Close central rounded vowel

Reason: [ɨ, ʉ] are the closest cardinal vowels, and the vowels listed in close central unrounded vowel and close central rounded vowel may or may not be the same as the corresponding cardinal vowels. "Near-close" is just slightly lower than fully close and those vowels have never been assigned separate IPA symbols, unlike [ɪ, ʏ, ʊ].

Regarding schwa-like vowels

(4) Merge the articles about (often non-contrastive) schwa-like sounds into schwa:

• Close-mid central unrounded vowel with Schwa
• Close-mid central rounded vowel with Schwa
• Mid central vowel with Schwa
• Open-mid central unrounded vowel with Schwa
• Open-mid central rounded vowel with Schwa

Reason: Scholars have simply not caught up yet with the latest reforms of the IPA. In English, ⟨ɜ⟩ is used for a vowel that can be open-mid, true-mid or close-mid (it's often just true-mid) in accordance with one of the former IPA definitions (or the former definition) of ⟨ɜ⟩. In Swedish ⟨ɵ⟩ is true-mid rather than close-mid. The symbol ⟨ə⟩ does *not* denote a specifically true-mid unrounded central vowel but just a schwa, a vowel in the general mid central area that can be rounded or unrounded. English schwa is an unrounded close-mid-to-open-mid vowel. Romanian and Sindhi schwas are open-mid, yet they're most typically transcribed with ⟨ə⟩ which is supposedly a vowel between close-mid and open-mid (that's what the current IPA chart suggests, very falsely in my view). IMO we're giving WP:UNDUE weight to something that simply isn't distinguished that strongly in the literature. WP:CONTENTFORK might also be an issue.

Also, should we ever do that, we should move the open-mid central unrounded vowels that are typically transcribed with ⟨ɐ⟩ back to near-open central vowel.

Check User:Kbb2/Schwa to see how the table in that article would (or could) work.

Regarding open front vowels

(5) Merge the articles about open front vowels into one article:

• Near-open front unrounded vowel with Open front vowel (an article that's yet to be created)
• Open front unrounded vowel with Open front vowel (an article that's yet to be created)
• Open front rounded vowel with Open front vowel (an article that's yet to be created)

Reason: A separate article for the open front rounded vowel might violate WP:UNDUE. It's an extremely rare vowel that doesn't occur as a phoneme in any language, it just happens to be one of the cardinal vowels (we already have an article about them). Plus, Danish [ɶ] (a canonical example of that vowel) is actually near-open, not open. Also, the distinction between [æ] and the front [a] isn't very clear as they sound awfully similar to each other. I think that as long as we differentiate [æ]/[a] from the central [ä] (which does sound different) that's good enough. [æ] should've never been assigned a separate IPA symbol by the way, and it was Anglocentric of the Association to do so. (EDIT: I've struck that out - it was an unnecessary personal comment). One of the reasons I'd like to see those three merged is that if we merged open front unrounded vowel with open front rounded vowel we'd have a discrepancy - the near-open variant would be covered in the case of the rounded vowel but not in the case of the unrounded vowel. Plus, the open front unrounded vowel might be transcribed with ⟨æ⟩ anyway.

Kbb2 (ex. Mr KEBAB) (talk) 17:07, 5 August 2019 (UTC)

I wholeheartedly agree that our coverage of vowels, and perhaps even consonants, needs to be improved. Simply put, every symbol on the IPA chart does not need its own dedicated article because not every sound is notable enough to warrant its own article. This is especially true of vowels because outside the cardinal vowels, there's no firm boundaries for their categorization. Further, the depth of coverage we can go into on each vowel is limited, making many of these permastubs, but we can give far more comprehensive and helpful if we organize our coverage by broad natural classes rather than particular phones. I'd like to propose a far more radical scheme of 5 comprehensive articles:

Front vowels

Back vowels

High vowels

Low vowels

Central vowels

These, with some obvious overlap, will cover describing the vowels in their domain. Vowels which have substantial coverage in their own right, like schwa can be spun out and summarized in these articles, but in general individual phones would not have their own articles unless they're notable. I think these articles should not include the tables of languages which have certain phonemic sounds, and instead we should have those lists at a dedicated page such as Phonemic status of ɜ or List of languages with phonemic ɜ. This will help divide our coverage between generalist and specialist articles, limit the number of pages that need to be maintained, and maximize use for our readers.

All this said, I still support Kbb2's proposal as an improvement over the status quo. Wug·a·po·des​ 17:26, 8 August 2019 (UTC)

I'm generally in favor of the proposed mergers but I would keep Mid central vowel. Schwa talks about phonological development rather than the phonetic sound(s), and the majority of its content is better merged with Vowel reduction or something rather than with Mid central vowel, and then Schwa can be turned into a redirect to Mid central vowel. Nardog (talk) 00:17, 10 August 2019 (UTC)

• Support merging articles on similar vowels. Not every sound in every spoken language deserves an entire article on its own, partially as many really aren't that notable. If they do get their own articles anyway, we will eventually just end up with hundreds of permastubs. Geolodus (talk) 15:04, 10 August 2019 (UTC)

Support per nom. —Yours sincerely, Soumyabrata (talk • subpages) 14:34, 30 August 2019 (UTC)

• (Mildly) support merging. The merger should not be used to dismiss well-sourced information accumulated in the individual articles, though. --mach 🙈🙉🙊 16:48, 24 September 2019 (UTC)
• Addendum: Oppose suggestions regarding open front vowels: I sympathize with the point of view that some of the IPA symbols may be questionable (I believe not only ⟨æ⟩ is Anglocentric and should never have been adopted by the IPA, but also ⟨ɪ⟩, ⟨ʊ⟩, and ⟨ʏ⟩). Nevertheless, Wikipedia is not the place for advancing such a point of view. I would much rather merge [ä] into [a], thus having one article per simple IPA symbol, and no additional articles with diacritics. --mach 🙈🙉🙊 19:18, 8 October 2019 (UTC)

@J. 'mach' wust: That's just a small portion of my argument. To quote myself, A separate article for the open front rounded vowel might violate WP:UNDUE. It's an extremely rare vowel that doesn't occur as a phoneme in any language, it just happens to be one of the cardinal vowels (we already have an article about them). Plus, Danish [ɶ] (a canonical example of that vowel) is actually near-open, not open. Also, the distinction between [æ] and the front [a] isn't very clear as they sound awfully similar to each other. I think that as long as we differentiate [æ]/[a] from the central [ä] (which does sound different) that's good enough. [æ] should've never been assigned a separate IPA symbol by the way, and it was Anglocentric of the Association to do so. One of the reasons I'd like to see those three merged is that if we merged open front unrounded vowel with open front rounded vowel we'd have a discrepancy - the near-open variant would be covered in the case of the rounded vowel but not in the case of the unrounded vowel. Plus, the open front unrounded vowel might be transcribed with ⟨æ⟩ anyway. I've now struck out that unnecessary personal comment of mine. Kbb2 (ex. Mr KEBAB) (talk) 04:27, 9 October 2019 (UTC)

I know what you have written. Whether or not you now retract a portion, my argument does not change: the IPA is the accepted standard in this field, and that we best represent the vowels by having one article for each simple IPA symbol. I think we should therefore discontinue the following articles (though without deleting the information they contain):

1. [ɨ̞] (already discontinued, though information has been deleted)
2. [ʉ̞] (already discontinued, though information has been deleted)
3. [e̞] (already discontinued, though information has been deleted)
4. [ø̞]
5. [o̞]
6. [ä]

I do not think a separate article for every IPA symbol is WP:UNDUE, since the IPA is the accepted standard in this field. To the contrary, I think what is WP:UNDUE is having separate articles for sounds that do not have their own IPA symbol but require IPA diacritics – especially when, as I fear, a majority of the sources never even uses these diacritics.

As to the discrepancy that you mention, it already presupposes the merger of [a] with [ɶ], which I would not support, and on top of that, an asymmetry in the open front vowels is inevitable since the IPA cover of the open front vowel is asymmetric (there is no rounded equivalent of [æ]). BTW, the canonical example of [ɶ] in my neck of the woods is Viennese German, though I have never researched it. --mach 🙈🙉🙊 12:28, 9 October 2019 (UTC)

@J. 'mach' wust: I don't think it's inappropriate to talk about information that has been deleted as if there was an option to retain it without changing the consensus that already exists regarding not listing more than one example per dialect. It was removed out of necessity.

We didn't discuss merging open central unrounded vowel with other articles. If, per current consensus, we merge the three open front vowels into one article, adding the open central one there could create some chaos, and I don't know a better way to get rid of open central unrounded vowel than to merge it with open front unrounded vowel. Perhaps open back unrounded vowel is the second best target for the merger. Dozens of millions of Americans think of their /ɑ/ when they hear the open central unrounded vowel. An argument against such a merger would be that only a minority of transcribers use ⟨ɑ⟩ for a central vowel.

We should add some Bavarian dialects to open front rounded vowel. I think that Traunmüller's Vokalismus in der westniederösterreichischen Mundart shows that the height of Bavarian /ɶ/ is variable and it doesn't have to be open-mid, as in Amstetten. It can be more open than that (then the same could apply to /æ/ and /ɒ/), but that could be just me misinterpreting the formant plots. I don't know. I'm out of my depth when it comes to formants, for the most part. Kbb2 (ex. Mr KEBAB) (talk) 08:47, 10 October 2019 (UTC)

I don't think it's inappropriate to talk about information that has been deleted as if there was an option to retain it without changing the consensus that already exists regarding not listing more than one example per dialect. – As I have explained in the very section you have linked to, there is no such consensus.

We didn't discuss merging open central unrounded vowel with other articles. – We are discussing it now.

If, per current consensus, we merge the three open front vowels into one article […]. – As this section shows, there is no such consensus either.

I don't know a better way to get rid of open central unrounded vowel than to merge it with open front unrounded vowel. – I think we should respect the sources: languages where sources use [a], [ä], [a̠], etc. should be added to [a], languages where sources use [ɑ], [ɑ̈], [ɑ̟], etc. should be added to [ɑ], languages where sources use [ɐ], [ɐ̞], etc. should be added to [ɐ]. --mach 🙈🙉🙊 09:34, 10 October 2019 (UTC)

Maybe someone isn't paying attention here, but there most certainly is a consensus on one example per language/dialect. That's not something that you can "explain" away or ignore. It's there, backed up by explicit discussion and a decade of implicit consensus through editing by scores of editors across dozens of articles. If one wants an exception to this for a particular case, they should have a clear and compelling reason for it. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 15:28, 10 October 2019 (UTC)

I do not know who pays attention and who does not, but what I have seen is that this implicit consensus has exceptions for languages with more than one quality of the respective sound; and that in the current discussion, two editors are against the deletions and one editor thinks the deletions are “hasty and premature”, whereas one editor is in favour of the deletions and one other editor tends to side with the latter – rather a case of WP:NO CONSENSUS. And then there is yet another implicit consensus those who pay attention may notice: the implicit consensus that some languages have had two examples in the articles for [e] and [e̞], for [ɨ] and [ɨ̞], or for [ʉ] and [ʉ̞]. I do not understand how this implicit consensus should be canceled out by a merger of these articles. How do you weigh one implicit consensus against another? Do you have any clear and compelling reasons? --mach 🙈🙉🙊 18:09, 10 October 2019 (UTC)

I've already provided a link that directs to where the explicit consensus is (which you have acknowledged). If you want to make a case to change how we do this, then make it. But you're wasting your breath and trying the patience of other editors when you say there isn't a consensus for one example per dialect/language when there clearly is. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 18:40, 10 October 2019 (UTC)

We have three points. Why do you only acknowledge one of them?

1. Your user subpage delineates the implicit consensus. If you read it carefully, you will notice that it allows exceptions in cases where languages have more than one quality of the respective sound.
2. The languages in question used to have two examples before the merger. If you read the above merger discussion carefully, you will notice that nobody mentioned that this should be changed.
3. If you read the below discussion carefully, you will notice that three out of five editors that have participated want to reverse the deletions until further discussion. Incidently, this is also in accordance with WP:BRD and WP:NOCONSENSUS, which kind of outweigh your user subpages. --mach 🙈🙉🙊 20:26, 10 October 2019 (UTC)

No, the user subpage delineates the explicit consensus and even links to the relevant discussions. It allows for a discrete number of exceptions, themselves based on explicit consensus for those particular cases. You generalizing from those cases goes against this consensus. Again, trying to come to a new consensus is fine. But calling what we've had for more than ten years something other than an established, explicit consensus is either gaslighting or sloppy reading. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 22:14, 10 October 2019 (UTC)

@J. 'mach' wust: WP:COMMONSENSE tells you that two examples per dialect should've been stated explicitly by me as a part of the merger. That is something I'd never do, because I was already aware of the consensus when I was posting the message. Are you really trying to tell me that you know better than me what I meant when I posted the proposal? Those users that are against the deletions are either unaware of the consensus or are against the deletions that aren't connected to the one example per dialect thing. Your message is an overgeneralization. Kbb2 (ex. Mr KEBAB) (talk) 00:08, 11 October 2019 (UTC)

If you put so much emphasis on the User:Aeusoes1/Phone tables consensus that it should overrule WP:BRD and WP:NOCONSENSUS and that everybody should know it, then why do you keep ignoring that it foresees exceptions for languages that have more than one instance of the respective sound?

You have merged [e] with [e̞], [ɨ] with [ɨ̞], and [ʉ] with [ʉ̞]. A number of languages have either sound. I fail to see any difference between this situation and the Basque or clicks situation: it is an unusual and highly notable situation. We should point it out to our readers, especially since we have sources and because we used to point it out to our readers before the merger, and especially since a majority of the editors on this page want to keep this information. --mach 🙈🙉🙊 05:16, 11 October 2019 (UTC)

This might sound really wild, but IMHO it would be an improvement to any of these articles to incorporate this kind of information in actual prose. I think that approach would be a better way of conveying the interesting outliers and representative tendencies of these vowels' relative distributions. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 05:34, 11 October 2019 (UTC)

I have thought of that, too. The downside is that we would loose the example words. After all, I would be very frustrated as a reader if an article mentioned the interesting information that a language had two qualities of the respective sound, but then only gave one example. I therefore prefer a combination: Use a common rowspan="2" prose description for both sounds, but still have two rows in the example columns, e.g. [1]. That is, if the sounds are allophones in complementary distribution. If however the sounds are phonemes on their own right, I would rather have them in different rows – similar to what we are doing _mutatis mutandis_ with the different phonemes in the examples for languages like English –, e.g. [2]. --mach 🙈🙉🙊 06:29, 11 October 2019 (UTC)

I'm not sure I understand. Why would prose need to omit example words? Look at the second paragraph at Spanish phonology#Consonants. There are numerous examples in the prose and they're used to illustrate the kind of thing we're talking about. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 18:12, 11 October 2019 (UTC)

Having examples within prose is certainly possible. However, it is not what we are currently doing in our “phone tables”, since these tables have three dedicated columns for the examples. I am all in favour of describing the sounds in the “Notes” column of the phone tables, but when it comes to providing examples, I think we better keep using the three dedicated examples columns. --mach 🙈🙉🙊 19:13, 11 October 2019 (UTC)

That's what I mean. There's information we want to convey and we're butting heads with the limitations in pigeonholing all information into tables. It's an encyclopedia and at some point we want to do more than just list examples and have actual encyclopedic content in the form of paragraphs. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 19:49, 11 October 2019 (UTC)

But for the time being, would the following mixed prosa and examples column format be OK to you for conveying the information that languages like Nothern Welsh or Maastrichtian Limburgish have two qualities of the respective vowels (two allophones in complementary distribution in the case of Welsh, two phonemes in the case of Limburgish)?

Language		Word	IPA	Meaning	Notes
Welsh	Northern dialects	llun	[ɬɨːn]	'picture'	Close when long, near-close when short. Merges with /ɪ/ in southern dialects. See Welsh phonology
		pump	[pɨ̞mp]	'five'
Limburgish	Most dialects	leef	[leːf]	'dear'	The example word is from the Maastrichtian dialect.
	Maastrichtian	bèd	[bɛ̝t]	'bed'	Typically transcribed in IPA with ⟨ɛ⟩.

Of course, they would not actually be in the same table. --mach 🙈🙉🙊 20:16, 11 October 2019 (UTC)

I think the rule should be one example per phoneme, and one phoneme per dialect. I agree with Aeusoes1 in that these pieces of information would ultimately be better presented in prose. Nardog (talk) 16:52, 12 October 2019 (UTC)

So you would have something like the following?

Language		Word	IPA	Meaning	Notes
Welsh	Northern dialects	pump	[pɨ̞mp]	'five'	Close when long, near-close when short. Merges with /ɪ/ in southern dialects. See Welsh phonology
Limburgish	Most dialects	leef	[leːf]	'dear'	The example word is from the Maastrichtian dialect. In the Maastrichtian dialect, this sound contrasts with a true-mid sound typically transcribed with ⟨ɛ⟩.

As I have already said, I think such a solution is very frustrating to readers: if there are two qualities of these vowels, then why do we only provide one example? --mach 🙈🙉🙊 19:08, 12 October 2019 (UTC)

Oh, sorry, scratch "one phoneme per dialect". What I meant to say is one dialect per phoneme, in that, as a general rule, there is no point in listing realizations of the same phoneme in multiple dialects if they all realize it with the same quality. Nardog (talk) 19:37, 12 October 2019 (UTC)

What I am talking about are only the cases where a vowel systematically has different qualities. As I have said before, I think it is interesting and notable information if a language distinguishes different qualities of a vowel that are so close to each other, whether or not they are allophones or phonemes. I think this is information relevant to our readers, and it is information that belongs with the respective vowel article because it illustrates the extent of possible differentiations. I am totally fine with such differentiations being mentioned in the “Notes” column, but I still think we must provide examples as a courtesy to our readers. It would be very frustrating to readers if we mention two qualities but only give an example for one of them (BTW, which one should we exemplify?). --mach 🙈🙉🙊 06:22, 13 October 2019 (UTC)

@J. 'mach' wust: I have fully reverted your reverts. You can't impose your WP:POV on those articles. We're supposed to reach the consensus here and then edit the articles accordingly. Your behavior is ridiculous. Kbb2 (ex. Mr KEBAB) (talk) 07:03, 13 October 2019 (UTC)

You are mistaken. Please read the policy WP:NOCONSENSUS: We are supposed not to make any deletions unless we have reached consensus here. --mach 🙈🙉🙊 07:50, 13 October 2019 (UTC)

Mach is correct. Again, it is BRD, not BRRD. Give the status quo precedence whenever a bold edit is challenged, especially if you're the one who went bold. Nardog (talk) 08:02, 13 October 2019 (UTC)

If a language contrasted close-mid and true mid, or two dialects of a language each realized phonemically distinct vowels as close-mid and true mid, then I think the inclusion of both would be warranted. Nardog (talk) 08:02, 13 October 2019 (UTC)

If I understand you correctly, you would include both qualities in a case like Maastrichtian because they are phonemes, and include them with example words in the appropriate example columns. But what about a language like Northern Welsh that has two qualities of a sound as allophones in complementary distribution? (As I have already said, I would include that information as well because it is relevant to the respective vowel, and I also would include examples because I think mentioning a sound without giving an example is pointless.) --mach 🙈🙉🙊 08:18, 13 October 2019 (UTC)

I'd say describe it in the Notes column, if at all. [ɨ] vs. [ɨ̞] is not a significant difference (precisely the reason we merged the articles). Nardog (talk) 08:43, 13 October 2019 (UTC)

I believe as Wikipedia editors, we should not be the judges of what is and what is not a significant difference. That is for the WP:SOURCEs to decide. --mach 🙈🙉🙊 09:21, 13 October 2019 (UTC)

I mean "significant" in the sense that it is enough to mention in the article (as a separate row in this case), nothing more. We make such judgments every day; it's called editorial discretion. Nardog (talk) 01:17, 14 October 2019 (UTC)

I don't feel comfortable making broad rules based on phonemes. These are phone tables, so we are talking about the actual sounds, not the abstract, language-specific conceptions of them. If we want to cover the incidence in Welsh of a phonemic contrast between two similar sounds (which I agree is notable), then we should do so in article prose, rather than in a table or a notes column. We're tying our hands by avoiding article prose.

It's actually kind of tiresome to see that every single person agrees that article prose would be an improvement, that it would be in reality the best solution here, but then no one wants to actually bell that cat. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 16:22, 14 October 2019 (UTC)

As I have said, I am all in favour of describing the sounds in prose, and I have twice posted proposals of how to do so, but both proposals have been met with utter silence.

And as I have said, I do not understand the point of describing two qualities of a sound in prose, but at the same time only allowing one example, or in other words, forbidding the other sound (which one?) to be illustrated with an example. That is how I understand your point of view, Aeusoes1, and it does not make any sense to me, so I sincerely hope I have misunderstood you. Therefore, I am asking you again: what do you propose we do? How would you include the Northern Welsh [ɨ] sounds or the Maastrichtian [e] sounds in the respective articles?

@Nardog: Have I understood you correctly: In the case of allophones like Northern Welsh [ɨ] and [ɨ̞], you would not mention them at all, not even in prose; but in the case of phonemes, you would mention them with examples? Just asking to be sure I have understood. (As I have said repeatedly, I would also mention allophones in prose, but with examples because I think they are interesting and unusual and noteworthy.) --mach 🙈🙉🙊 18:52, 14 October 2019 (UTC)

Take a look again at the Spanish phonology example I gave. That's how I think you can have example words in prose. That way you have examples and you're contextualizing everything in ways that you can't do (or can't do very easily, at least) in a table. When you do that, then the pressure to have exhaustive examples in the table is greatly reduced.

If you've got an idea for specific article prose, I recommend either BRDing it (still leaving the tables alone pending consensus) or putting a sample in talk space for discussion if you think it might be too contentious. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 19:40, 14 October 2019 (UTC)

So you want to get rid of the phone tables altogether. I disagree. MOS:USEPROSE does not mean we should never use lists. Lists or tables are preferrable when we are putting together numerous items of equal importance that are comparable in one or several ways (cf. WP:WHENTABLE), like the phone tables.

I believe that Spanish phonology#Consonants would benefit a lot if it were converted into a list (or a table). It is a list in disguise, consisting of numerous items comparable in several ways: What sound, what varieties of Spanish, what example? The one-sentence paragraphs indicate bad prose. If the information would be organized in a table, it would be much more readily understandable to readers. With the phone tables, we have the same sitution. --mach 🙈🙉🙊 19:02, 15 October 2019 (UTC)

Not sure how you got "replace tables with prose" but that is not my stance. We can still have the phone tables, but article prose would be a better focus of our energies if we're trying to talk about distribution or languages that have phonemic contrasts of similar sounds. As another example of article prose, look at palatal approximant.

And no, the Spanish phonology article would not improve if we converted article prose into a list or table. Your comment is a perfect example of how a list or table would oversimplify nuanced information. That all you see is a "list in disguise" (which is a joke, particularly for the second paragraph that I was pointing to in the first place) prompts me to question your capacity to even recognize the necessary nuance to produce the relevant article prose in the first place. If all you want is a list, all you'll see is lists. But we need something higher order than that. It might make sense to augment prose with a table, but that table wouldn't be able to house all the information in that section, even if the prose could do with some cleanup. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 19:32, 15 October 2019 (UTC)

In the case of allophones like Northern Welsh [ɨ] and [ɨ̞], you would not mention them at all, not even in prose; but in the case of phonemes, you would mention them with examples? What? No! I meant: When multiple sounds with slightly different qualities which are discussed in the same article (like [ɨ] and [ɨ̞], and [e] and [e̞]) are found in the same language (or dialects of the same language), only one row need be dedicated to them in the occurernce table unless the sounds belong to different phonemes. This does not preclude mentioning the other allophone(s) in the Notes column. And I'm not talking about what to do in prose at all. In prose you can do anything so long as it enhances the discussion of the subject of the article. Nardog (talk) 23:58, 15 October 2019 (UTC)

I am still not sure whether I understand you correctly. I see it is good that I have asked, so I will ask again: In the case of allophones, you would mention both sounds in the table, but give only one example? I think that is poor design and not helpful to readers. And how do hou decide what example to keep and what example to remove? --mach 🙈🙉🙊 04:49, 16 October 2019 (UTC)

Yes, give only one example. All sounds are allophones and are produced with a different quality in each and every phonetic environment. It is listing allophones of the same phoneme whose difference is so small they are covered in the same article that is not helpful to readers and impracticable in terms of deciding what to include. Keep whatever example that is considered a typical incidence of the phoneme.

As for Welsh [ɨ] and [ɨ̞ː] in particular, Welsh phonology treats them as different phonemes (which I assume is akin to positing different phonemes for [æ] and [ɑː] in RP-type English, which can be analyzed as short and long instances of the same phoneme just as plausibly), so that can justify listing both in separate rows. If we treat them as the same phoneme, then I don't think the difference is worth a mention even in the Notes column because, again, the difference is so small; if the difference is so significant the long one should be covered in the same article as [ɘ]. (By the way, the acoustic chart at Welsh phonology#Vowels shows both sounds at lower points than [i, u] and even [ɪ, ʊ], which suggests they are both better covered in the same article as [ɘ], provided the chart is representative of the "Northern dialects".) Nardog (talk) 05:45, 16 October 2019 (UTC)

All sounds are allophones and are produced with a different quality in each and every phonetic environment. – That is besides the point. It is obvious that there is an infinite continuum of phonetic realizations, and nobody has proposed listing each and every one of them. The point is that the reliable sources our material is based on single out some of them. They consider certain allophones worth mentioning and mention them, including examples. Our articles have had that information for many years. I still do not see any reason (or consensus) why we should now all of a sudden delete that information and deprive our readers of it. --mach 🙈🙉🙊 21:24, 16 October 2019 (UTC)

• Oppose merger of IPA letters. I'm not concerned with the merger of the others, per arguments similar to Mach's.
  That is, I'm not opposed to finishing (1), and I support the proposed moves for (2), bringing them into line with IPA description. (3) is already implemented. Some of the merged material might belong at reduced vowel. Oppose (4), except for mid central vowel > schwa, since the IPA set up this vowel specifically for a reduced mid-central vowel, i.e., schwa. (Though I haven't considered Nardog's objections.) But the other four vowels in (4) were set up as full vowels, with a defined location in vowel space and defined rounding. They are thus incompatible with schwa. Oppose (5) and instead, as does Mach, support merging open front unrounded vowel with open central unrounded vowel. The problem here is that the IPA vowels are defined by articulation, but no-one uses articulation any more as the defining characteristic of vowels. Rather, we nearly always use vowel formants, including articles in JIPA and the IPA Handbook. Using formants, even Daniel Jones' and John Wells' pronunciation of ⟦a⟧ is central, not front, and is much more open than ⟦ɑ⟧. On the other hand, as others have pointed out, [ɶ] in the few languages that have it is less open than ⟦a⟧ -- that is, it's the rounded equivalent of ⟦æ⟧. Merging the articles would badly obscure this discrepancy between the definitions of the IPA vowels in theory (hypothesized articulation) and their values in practice (demonstrated formants). Merging open front<>central unrounded vowel, on the other hand, would IMO be okay as the distinction is generally not made in practice. However, I think we should clarify in the lead of the merged article that there is some debate over whether ⟦a⟧ is truly a front vowel. Maybe we could move it to a name that reflects this ambiguity? The fact that people are chronically proposing an IPA letter such as ⟨ᴀ⟩ for the low central vowel (and being rejected) suggests that there is confusion as to the nature of ⟦a⟧. — kwami (talk) 03:43, 17 December 2019 (UTC)

@Kbb2: @Nardog: @J. 'mach' wust: @Aeusoes1: -- pinging some of you in case you're interested in my 2 bits, since it's been a couple months.

BTW, re. not liking ⟨ɪ, ʊ⟩ as an undue influence of English, these are actually quite useful letters throughout Africa, where using the ATR/RTR diacritics makes extended transcription difficult to read. ⟨ɪ, ʊ⟩ pair up nicely with ⟨ɛ, ɔ⟩ for the RTR vowels. — kwami (talk) 03:51, 17 December 2019 (UTC)

I suggest not only not combining IPA sounds that aren't already combined, but uncombining those which are. If you are trying to find information about ONE CERTAIN letter which is in an article where there are two or more sounds, it can be hard, hellish, and even downright impossible to find that information you need. For example, let's say, if all nasals were combined, and you wanted to find information about ɲ, it could take longer than it should. The article could also combine information, basically how the news expresses bias, by hiding information. It makes people like me who are trying to create a language feel like tossing their computer in a piranha-infested lake and jumping in with it. And then swim off a waterfall. (maybe a LITTLE exaggeration) --LinguistWorldbuilder43 (talk) 22:18, 25 December 2019 (UTC)

Articles with links to DAB pages

I have collected a batch of articles with language- and linguistics-related links to DAB pages which would benefit from expert attention. Search for 'disam' in read mode, and for '{{d' in edit mode; and if you solve any of these puzzles, post {{done}} here.

• Accusative case  Done — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 20:34, 10 September 2019 (UTC)
• Comparison of Japanese and Korean  Done
• Dagaare language  Done
• Great Vowel Shift  Done
• Hendiatris  Done
• Jukunoid languages  Done
• Kula language  Done
• List of Indigenous Australian group names  Done
• Logical form (linguistics)  Done
• Nheengatu  Done — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 16:01, 19 September 2019 (UTC)
• Nupoid languages  Done
• Ru (kana)  Done
• West Chadic languages  Done

Not all may be solvable, but every little helps. Thanks in advance, Narky Blert (talk) 19:51, 10 September 2019 (UTC)

Last ones done. I moved the Ngewin dab per primary use, and left an overt link to Middle Japanese per that source. — kwami (talk) 00:41, 23 November 2019 (UTC)

Request for information on WP1.0 web tool

Hello and greetings from the maintainers of the WP 1.0 Bot! As you may or may not know, we are currently involved in an overhaul of the bot, in order to make it more modern and maintainable. As part of this process, we will be rewriting the web tool that is part of the project. You might have noticed this tool if you click through the links on the project assessment summary tables.

We'd like to collect information on how the current tool is used by....you! How do you yourself and the other maintainers of your project use the web tool? Which of its features do you need? How frequently do you use these features? And what features is the tool missing that would be useful to you? We have collected all of these questions at this Google form where you can leave your response. Walkerma (talk) 04:24, 27 October 2019 (UTC)

Help in identifying a postalveolar fricative

Requesting help from the phonetics people out there. So, I've got the following description (with the text elsewhere making it clear the sound is a sibilant):

the tip of the tongue slightly touches the lower teeth, the blade appears to touch the upper teeth, and the front together with a part of the middle of the tongue is raised towards the hard palate.

It's actually labelled as "palato-alveolar", though I'm not sure if this meant what it means now (the text, published in 1963, is probably unchanged since the time it was written at the end of the 1930s), and when I try to replicate the pronunciation what comes out sounds to my non-trained ears more like [ɕ]. Any thoughts? – Uanfala (talk) 20:11, 27 October 2019 (UTC)

It's likely [ɕ], since [ʃ] usually doesn't contact the lower teeth. What's the source text? Wug·a·po·des​ 23:18, 27 October 2019 (UTC)

The source text, as in which work the extract came from? That's:

Bahri, Hardev (1963). Lahndi Phonetics : with special reference to Awáṇkárí. Allahabad: Bharati Press. {{cite book}}: Invalid |ref=harv (help)

– Uanfala (talk) 00:58, 28 October 2019 (UTC)

Page Edit Proposal: https://en.wikipedia.org/wiki/British_Association_for_Applied_Linguistics

Hi WikiProject Linguistics team.

First of all, conflict of interest statement, I am the Media Coordinator on the Executive Committee of this organization.

We'd like to update the main text of the following Wikipedia page: https://en.wikipedia.org/wiki/British_Association_for_Applied_Linguistics

A number of the details are out of date and we'd like to add more information about what our organization does. Would anyone be able to help me do this? — Preceding unsigned comment added by Baalmediacoordinator (talk • contribs) 05:48, 31 October 2019 (UTC)

Responded on your talk page. Wug·a·po·des​ 06:06, 31 October 2019 (UTC)

Etymology of chorisepala

Currently, I have a dispute with User:Gderrin about the etymology of the botanical epithet chorisepala. Gderrin writes:

• The specific epithet (chorisepala) is derived from the Greek choru meaning "separate" or "free" and -sepalus meaning "-sepalled".

I can not seem to convince Gderrin that Greek words usually do not end on -us (except for υ-stems in Greek (and with subsequent transliteration of -υς with -us), but -sepalus can not be a υ-stem) and the -us-ending is clearly a marker for being a Latin or a Latinized form. Moreover, I can not find any evidence yet in Greek lexica that choru would be a Greek word for "separate" or "free". It it a typo, or is it a uncommon transliteration of the Greek genitive χώρου (= "of space")? The specific source (Sharr, 2019) that he is using seems to err on a number of occasions (like stating that leprosus is Greek, despite the Latin suffix -osus). Wimpus (talk) 09:17, 18 November 2019 (UTC)

I do know that Late Latin leprosus is ultimately derived from Greek λεπρός (with the addition of the Latin suffix -osus) and that New Latin sepalum is probably influenced, according to Stearn's Botanical Latin, by Greek σκέπη, but that does not make leprosus nor -sepalus Greek, but merely Greek-derived. Wimpus (talk) 09:23, 18 November 2019 (UTC)

References

• Here's what the OED says,

chori-

(kɔərɪ)

before a vowel choris- (kɔərɪs),

a. Gr. χῶρι, χωρίς asunder, apart: used in botanical terms, as choripetalous (-ˈpɛtələs), a., having separate petals = polypetalous; chorisantherous (-ˈænθərəs) a., having separate anthers; chorisepalous (-ˈsɛpələs) a., having separate sepals; = polysepalous.

1880 Gray Struct. Bot. vi. §5. 244 Choripetalous is‥the most fitting name for a corolla the petals of which are separate.

— kwami (talk) 23:32, 22 November 2019 (UTC)

Thank you for the reference. This indicates that the first part is derived from "Gr. χῶρι, χωρίς" and not from "Gr. choru", that can not be found in Greek dictionaries. I do think that User:Gderrin would still be convinced that the first part is from "Gr. choru". Wimpus (talk) 23:40, 22 November 2019 (UTC)

Whether they're convinced is irrelevant, unless they can provide a better source. The OED entry is old, but dates from a time when lexicographers knew their Classics. It's not likely to be like the etymology of 'dog', where there may be new hypotheses, or of non-European words, where the OED tends to be sloppy. Greko-Latin technical terms are consciously designed, and we often have records. — kwami (talk) 00:45, 23 November 2019 (UTC)

Proposed rename of article regarding origins of Hungarian

Your feedback would be welcome at a discussion about renaming the very awkwardly titled, Obsolete theories of the Hungarian language relations. I've created a discussion at the TP, which is not an official RM because a rename seems uncontroversial to me, but more eyeballs would be welcome at the discussion to discuss alternate titles. Please add your thoughts at Talk:Obsolete theories of the Hungarian language relations#Title rename. Thanks, Mathglot (talk) 01:25, 4 December 2019 (UTC)

Definition of 'Language'

Contributors to this WikiProject may be interested in discussion at Talk:Language#Definition of 'Language'. I think changing the definition warrants input from more editors, but unfortunately don't have time to add much to the discussion right now. Cnilep (talk) 07:48, 10 December 2019 (UTC)

Do we only include "good" examples in phonetic sounds' occurrence tables?

Kwamikagami has been removing several examples from the occurrence table at Dental and alveolar taps and flaps on the grounds that they are not "good examples". According to them, Languages where a sound might be [ɾ] or might be [ɺ], or might be [ɾ] but might be a misanalyzed [r], are not good illustrations. I had never heard of this idea and I want know (or establish) the consensus on this. To me if the sound is in free variation or complementary distribution with other sound(s) that's all the more reason to include it because that illustrates the phonological status and relations with other sounds that sound can have.

(Pinging @Kbb2, Wugapodes, J. 'mach' wust, and Aeusoes1: as we've touched on similar points in #We probably have too many vowel articles and #Unmerge all of the vowel articles above.) Nardog (talk) 08:05, 17 December 2019 (UTC)

IMO, the point of these articles is to illustrate prototypical (i.e., IPA) sounds. Rather like cardinal vowels. That gives the reader a baseline comparison when reading about a particular phonology. If our illustrations are an incoherent mishmash, then they don't serve the reader well. And when a sound is common in the world, we can afford to be picky about the examples we choose.

If someone adds a language with only a trill to the flap article because they misread the source or don't know the difference, then yes, of course we need to remove it. In this particular case, what Nardog is actually objecting to is the removal of Japanese. Japanese /ɾ~ɺ/ is undefined for centrality. It therefor is not a good example of [ɾ], and IMO should not be included, because we have so many good illustrations. We need to cut off the thousands of possible illustrations somewhere, so why not cut off the bad ones?

(Japanese /ɾ~ɺ/ is not a good example of [ɺ] either, but for that sound we can't afford to be so picky. There are few well-known languages that would be good illustrations for [ɺ].) — kwami (talk) 08:26, 17 December 2019 (UTC)

The examples should be to the benefit of the reader in the first place, and not just to the benefit of having an article that is as broad as possible (especially in the case of common sounds). For this reason I believe it makes sense to primarily keep proto-typical (and well-sourced) examples, such as—in the case of [ɾ]—Spanish (where it has phonemic status), or old-school RP and Korean (where it appears as an allophone in well-defined contexts). Potentially ambiguous examples, including the iconic not-really-R-like [ɾ~ɺ] in Japanese might be in effect more confusing than helpful to a non-specialist reader who is familiar with Japanese. –Austronesier (talk) 15:00, 17 December 2019 (UTC)

Because a table is designed to steamroll over nuance, it makes sense that we stick to clearer examples. I think kwami's wording in the other thread sums it up pretty well: Languages that almost invariably have [ɾ] as the realization, like Arabic, are good examples. Languages where a sound might be [ɾ] or might be [ɺ], or might be [ɾ] but might be a misanalyzed [r], are not good illustrations.

As I said in the previous discussions, article prose could help us out here. We would be doing a disservice to our readers to not even mention Japanese at the [ɾ] article. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 16:08, 17 December 2019 (UTC)

Agreed, but that more general info might be even better placed in the main flap consonant article, along with one-contact trills. — kwami (talk) 03:00, 18 December 2019 (UTC)

I think our discussions about what is or is not a “good” example are pointless. The criterion should be relevant sources. When a relevant source deems an [ɾ] allophone important enough to be mentioned, then it should be important enough for us. --mach 🙈🙉🙊 06:08, 18 December 2019 (UTC)

@mach: What makes a source "relevant"? (I only know of "reliable sources", which are a necessary criterion for inclusion, but not a sufficient one.) Introducing the criterion of "relevance" is just as arbitrary as the criterion of a "good" (= illustrative) example, unless we discuss and finally agree about how we want to define it.

@Ƶ§œš¹: Agree, the prose part should be more inclusive and is a good place to mention more ambiguous examples and the potential (range-wise) overlap with similar sounds. –Austronesier (talk) 06:41, 18 December 2019 (UTC)

Don't get me wrong, I actually like mach's idea of "relevant sources", so I want to take it up and make a suggestion possible cut-offs. In the first place, I would eliminate all entries with sources which do not specifically deal with phonetics. General descriptive studies always have a phonetics/phonology section, but most field linguists and other descriptive linguists are certainly trained, but not specialized in phonetics. So in most cases we can expect accurate descriptions, but shouldn't exclude the possibility of close hits. Personal anecdote: I collected data in the field about an understudied language, and in the process of analysis decided that one of the allophones of /l/ is a retroflex lateral flap, because I thought that this would adeqautely describe what I heard, and what I produced myself to imitate it (correctly, according to my quite critical informants). Eventually, I had the oppurtunity to publish a short sketch of that language: my sketch was peer-reviewed, it passed, and now the world has a source about another retroflex lateral flap in a natural language. I am still pretty sure about my phonetic instinct, but frankly I would feel awkward if quoted in a phonetic overview article.

I would also exclude broad phonological studies which deal with phonetic details only in passing, for the same reasons. Not all phonologists are reliable phoneticians.

That leaves general phonetic sources, and also those specialized in indivdual languages. –Austronesier (talk) 10:03, 18 December 2019 (UTC)

I can see merit in preferring phonetic sources, but I don't know if it would be a good idea to keep such a high bar in all circumstances. If we consider this a matter of degrees of confidence, with phonetic sources at the top and broader linguistic sources under that, it might be a good idea to in general accept either while preferring the former. We would only exclude an example lacking the former if there is reason to doubt it, such as with rare sounds or contradictory sourcing. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 15:47, 18 December 2019 (UTC)

BTW, the reason I started paring down this particular article is that I was looking for an example of a dental tap, and saw that the one example we had was suspect. So, question, are there any languages with dental taps? — kwami (talk) 08:29, 24 December 2019 (UTC)

There are some mentions of dental taps in a number of general descriptive sources, but I also have found one mention in a phonetics textbook, viz. Rogers (2000), The Sounds of Language: An Introduction to Phonetics, p. 221: "Spanish contrasts a dental tap with a dental trill". This bold statement contradicts most sources that I know, which all describe an alveolar tap and an alveolar trill in Spanish. –Austronesier (talk) 13:18, 24 December 2019 (UTC)

Reichenau Glosses

Hey guys,

This is an article I wrote about Romance Linguistics. I'm wondering if you guys have any feedback on it, for example what needs to be improved?

Cheers, --Excelsius (talk) 05:33, 23 December 2019 (UTC)

(well, it's not entirely my work, but this is what it looked like before I touched it... yikes!)

Hindustani

Recent massive rewrite per the POV that Hindustani is extinct. The editor is confusing the official language under the British Raj (essentially Urdu) with the topic of the article, which is the language that MS Hindi and MS Urdu are standardized forms of. I don't remember who was in the discussion the last time this came up, so I'm mentioning it here. — kwami (talk) 00:17, 9 January 2020 (UTC)

New article Theory of language

I can't tell how much of this is original research (most seems reasonable) or whether this article should exist at all (should it be merged into another linguistics article?): Theory of language. Can someone take a look? Thanks. — MarkH₂₁^talk 02:36, 10 January 2020 (UTC)

A tough one. I'd want to hear other editor's thoughts. An overview of schools of linguistic thought such as structural linguistics and Systemic functional linguistics would be a wonderful addition, but I'm not sure how best to align this page with that scope. Pinging @Kbb2, Nardog, J. 'mach' wust, Aeusoes1, and Kwamikagami: since you all've been active recently. Thoughts? — Wug·a·po·des​ 04:39, 18 January 2020 (UTC)

Pinging @Austronesier: too. — kwami (talk) 05:18, 18 January 2020 (UTC)

I created the page to serve as a main article for Philosophy_of_language#Nature_of_language as well as a corresponding section in linguistics. I think it is quite basic knowledge. What is it that surprises you? I mean, can you not imagine that people have actually thought of what language is? In fact, there's quite a bit of the same discussion in language, too. I think the problem was that I wrote the article first with a few sources, and then started gradually adding more. So, maybe it just somehow came out of the blue. However, the fact will never change that theory of language, or the question of "what is language?", will always be a fundamental question in linguistics and philosophy of language. So merge it with... what? I'm actually quite curious to know which part of it you find so strange, to improve the article. Anyway, next time I add a new page I suppose I'll start with a stub. Weidorje (talk) 06:45, 18 January 2020 (UTC)

Links to DAB pages

I have collected another batch of articles with linguistics-related links to DAB pages where expert help 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.

• Egyptian language  Done
• Gallaecian language  Done
• List of Indo-European languages  Not done see RfD
• Tape language
• Ukrainian grammar  Done

Thanks in advance, Narky Blert (talk) 00:49, 18 January 2020 (UTC)

Regarding Tape, I'm not sure what the article means by 'labiovelar consonants' and will need to check out the cited source from the library to disambiguate. If someone can figure out the answer before next week, feel free to. — Wug·a·po·des​ 03:55, 18 January 2020 (UTC)

I don't have access, but the consonant chart has two m's. I suspect the first should have a tilde. The columns should also be defined, if you can look that up at the same time. And repeated confusion of the word "letter" with "phoneme", though any of us could clean that up. — kwami (talk) 05:16, 18 January 2020 (UTC)

IPA-x templates

Hey everyone. There was a recent discussion about a different group of language templates (the xx-icon family of templates) that had the same design approach as these - a template for each language. That discussion resulted in changing the design to one template that accepts as a parameter the language (see Template:In lang) - so in this example, instead of a template like {{IPA-ja}} it would be {{IPA|ja}}. For editors, the change is very minimal as instead of a hyphen they use a vertical bar (so exactly the same amount of characters), but the behind the scenes can now be maintained much more reasonably. Now if you want to apply a change to all templates, you need to update each individual template (currently ~289 templates), once consolidated, there is only one single template that needs to be updated. User:Trappist the monk has experience with this procedure as he was the one that merged the xx-icon set. Would love to hear any comments you have before moving on this in TfD. --Gonnym (talk) 11:38, 22 January 2020 (UTC)

Sounds good. This would require the {{IPA}} template to count parameters, as {{IPA|ja}} is currently a valid input yielding ja. (IPA without enclosing square brackets or slashes is frequently used in charts etc.) Love —LiliCharlie (talk) 12:10, 22 January 2020 (UTC)

Counting parameters is generally not a problem that causes much worry.

—Trappist the monk (talk) 17:05, 22 January 2020 (UTC)

We need to have a way of banning the usage of slashes when linking to phonetic (meaning: non-phonemic) guides. [esˈtaðos uˈniðoz mexiˈkanos] is only a valid transcription when you use the brackets; when you use the slashes, you need to write /esˈtados uˈnidos mexiˈkanos/ and that type of transcription isn't covered on Help:IPA/Spanish. Kbb2 (ex. Mr KEBAB) (talk) 12:27, 22 January 2020 (UTC)

Is there something in the 'text' of phonetic and phonemic transcriptions that can be relied upon to guarantee that transcriptions containing that 'something' will not be confused with the other? For example, is it ever possible for phonemic transcriptions (/.../) to be written with characters that are not basic ascii or space modifier letters? If any unicode character can be used in either transcription, it becomes very difficult or impossible for a machine to determine which of /.../ or [...] is the correct form.

—Trappist the monk (talk) 17:05, 22 January 2020 (UTC)

@Trappist the monk: I'd rely on the language code. For example, IPA|es|transcription should always yield Spanish pronunciation: [transcription] since Help:IPA/Spanish uses broad phonetic transcription. In general, AFAICS, hardly any of our Help:IPA/X guides uses phonemic transcription. The plain IPA template should, of course, allow both brackets and slashes - but that goes without saying. I don't think this can ever become a major issue, but it will be an additional thing to clean up and could create a problem that doesn't exist right now. Kbb2 (ex. Mr KEBAB) (talk) 18:37, 22 January 2020 (UTC)

This seems reasonable. Though what about IPAc templates like {{IPAc-pl}}? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 15:40, 22 January 2020 (UTC)

Different series of templates? As such I don't think that they should be discussed here but could be (probably should be) discussed in a separate discussion. A quick glance seems to indicate that, amonst themselves, they are mostly the same except for each language's translation table.

—Trappist the monk (talk) 17:05, 22 January 2020 (UTC)

It depends on the language. In some languages, the input and the output can have one-to-one correspondence, while in others the output may depend on surrounding sounds. To self-plagiarize, the IPAc templates are "kind of things of the past, as most of them were created before Lua modules became available, and even the Lua-based ones still inherit syntax from their non-Lua predecessors, when a module can take a full, undivided word or phrase as an input and convert it to IPA for languages with phonemic orthographies (Wiktionary has such templates for Spanish, Italian, Polish, Finnish, Japanese, etc.) or take a whole IPA transcription as an input and provide tooltips for each symbol. But that solution is also not free of problems as such a complex template would require skilled and committed editors maintaining it as changes are made to the key and bugs are found."

But the idea of a meta IPAc template, at least for languages with reasonably consistent input–output correspondence, is enticing. (I was working on it and that's when I shelved the overhaul mentioned below because I realized the enormity and complexity of the task.) Nardog (talk) 07:47, 23 January 2020 (UTC)

For the record, this topic arises from this discussion on my talk page.

—Trappist the monk (talk) 17:05, 22 January 2020 (UTC)

I was working on overhauling the IPA templates just over a year ago as I too saw a bunch of things that needed to be changed:

• Labels
  • In some templates, e.g. br returns "Brazilian Portuguese:", while in others e.g. fi returns "Finland Swedish pronunciation:" and langfi "Finland Swedish:". This should be consistent (towards the latter I think, given the default is overwhelmingly "<language> pronunciation:" with lang returning "<language>:").
  • In some templates the names of languages or dialects are linked while in others they are not. This should be consistent, while an option like |link=yes may be useful.
  • In most templates entering a text that does not match the code for a canned text (lang, local, etc.) as a label will have no effect. So editors have done things like {{small|Label:}} {{IPA-xx|...|}} to show customized labels. This is counterproductive. The template should be able to show customized labels by {{IPA|xx|...|Label:}}, or at the very least by e.g. {{IPA|xx|...|label=Label:}} for compatibility. (Perhaps the audio too should have a named parameter like |audio=. The template documentation currently recommends e.g. {{IPA-fr|o|-|Fr-eau.ogg}} to keep the default label—even though |3=Fr-eau.ogg will do the job.)
  • Currently there is no way to disable the reducing of the size of the label contra MOS:SMALLTEXT. An option like |small=no would be nice.
• /, ˈ, and ˌ, which are often used in transcriptions, trigger line breaks in undesirable positions. So editors have used {{nowrap|{{IPA|...}}}}, {{IPA|/{{wj}}ˈxxx{{wj}}ˌyyy/}}, etc. ad hoc, but this can easily be solved once for all by wrapping the entire transcription in <span class="nowrap">...</span> and all whitespaces within it in <span class="wrap">...</span> (inserting <wbr /> after whitespaces is another way but this doesn't work on Firefox or inside <blockquote>...</blockquote>).
• Currently the audio is added via Template:IPA audio link, which is redundant. I think the function should incorporated in the main template/module. (There is no justification why IPA-xx templates support appending an audio but {{IPA}} doesn't either.)
• There exist the categories "Pages including recorded pronunciations", "Articles including recorded pronunciations", and their respective language-specific subcategories, added by templates like {{Audio-IPA}} and {{IPAc-en}}. First of all, do we need these? If so, I think the new template/module should also use them (though only in certain namespaces), automatically figuring out which one to add.

See this for my attempt if you're interested (I know it's kludgy; I was still working on the premise that IPA-xx would still be in use, but I'm down with the proposed transition provided a bot owner is willing to fix the existing transclusions). Nardog (talk) 07:25, 23 January 2020 (UTC)

Request for citations for phonetic symbols for Unicode submission

Hi. I'm in negotiation with the Unicode committees to add support for the new extIPA and VoQS symbols, and am requesting a bunch of other unsupported phonetic symbols as well. I've gotten some feedback from LingList, but thought it would be worth asking here as well. I'm open to any system -- Americanist, Dania, whatever -- but my focus is extended IPA support, both obvious things that aren't on the chart, like superscripts, and extensions like those for Sinologists. The main gap in coverage is in superscript variants -- see Secondary articulation#Unicode support of superscript IPA letters for existing coverage of the standard IPA, but there's a lot missing from the Americanist set as well.

So, if you don't mind, I'll list the attestations I've dug up so far and ask you to add any others you know of or come across. Feel free to add additional symbols, but note that Unicode is no longer accepting letters with detached diacritics. (That is, if the diacritic doesn't connect to the base letter, they expect the font to handle it.) They also want a demonstration that a character is actually needed: a table of theoretical symbols won't be enough if they're never used for anything. But writing in a symbol by hand in your department's working papers should be enough -- that's how we got support for ⟨ꞎ⟩. It's nice to have a symbol in an explanatory table and in transcription both, if possible, but the transcription is more important.

Kirk (talk) 21:57, 24 January 2020 (UTC)

For example, see 'r with ascender' below. The ref in the TIPA manual is an unsourced chart and so by itself is insufficient to attest to use. Doke (1936) and (1938) are old and so do not demonstrate that the character is needed for digital typesetting. (AFAIK there is no conservation project for these languages that makes use of Doke's field notes, for example.) Dolgopolsky (2013) attests to rather minimal current use. Hopefully taken together these will be adequate.

full letters

• retroflex lateral flap ɺ̢ ( in SIL fonts)

  Rumsey (2017) Dependency and relative determination in language acquisition

• voiceless retroflex implosive ƭ̢

  Melaku Dissassa (1980) Some aspects of Oromo phonology, fn p. 10-11 [insufficient!]

  Bickford & Floyd (2006) Articulatory Phonetics, table 28.1 [just a table]

• retroflex click ǃ̢

  Julian Bradfield (2014) Clicks, Concurrency and Khoisan

  Bickford & Floyd (2006) Articulatory Phonetics

• lɾ (r with ascender)

  Fukui (2004) TIPA Manual, v. 1.3, p. 29 (#111)

  Doke (1938) Text Book of Lamba Grammar

  Doke (1936) An Outline of ǂKhomani Bushman Phonetics

  Dolgopolsky (2013) Indo-European Dictionary with Nostratic Etymologies, vol. I, p. 230, 494

• ⨎ (Beach letter for palatal click)

  Beach (1938) The Phonetics of the Hottentot Language

  Derek Elderkin (1989) The Significance and Origin of the Use of Pitch in Sandawe

superscript variants

missing completely

• ɖ
• ʡ
• ⱱ
• ɺ
• ø
• ɞ and ʚ
• ɶ
• ᵿ
• ɼ
• ƞ
• implosive, click and extIPA letters
• S, X, Y, Z [as wild-card letters, etc.]

minimal support

• ʀ

  Penhallurick (1991) The Anglo-Welsh Dialects of North Wales [list of symbols only]

• ɢ

  Wolfgang Kehrein (2002) Phonological Representation and Phonetic Phasing: Affricates and Laryngeals

• ˑ [half-long]

  Heselwood (2013) Phonetic Transcription in Theory and Practice

• ɮ

  Wolfgang Kehrein (2002) Phonological Representation and Phonetic Phasing: Affricates and Laryngeals

• ħ

  Edmondson et al (2005) A Laryngoscopic Study of Glottal and Epiglottal/Pharyngeal Stop and Continuant Articulations in Amis

• ɧ

  Riad (2013) The Phonology of Swedish, fn 12 [almost certainly inadequate on its own]

• ɫ

  Eric Bakovic (2005 Nov 16) Orange you glad I didn’t say the C-word?, Phonoblog, UC San Diego [a blog is unlikely to be sufficient]

• ɘ

  Caudal (2011) Contribution aspectuelle des temps et de la phrase, in Perspectives théoriques et empiriques: Sur l'aspect en anglais

• Q

  Fujimura & Williams (1999) Syllable concatenators in Japanese, Spanish, and English

some support

• q

  Menán du Plessis (2009) Unity Hypothesis for the Southern African Khoesan Languages

  Deborah Tooker (2012) Space and the production of cultural difference among the Akha prior to globalization

• ː [length sign]

  Heselwood (2013) Phonetic Transcription in Theory and Practice

  Boersma & Hamann (2009) Loanword adaptation as first-language phonological perception, in Loan Phonology

• ʜ

  Amanda Miller (2011) The Representation of Clicks, in he Blackwell Companion to Phonology [substituted with a full-cap H]

  John Esling (2010) Phonetic Notation, in The Handbook of Phonetic Sciences

  Miller-Ockhuizen (2010) A prosodic account of Juǀ’hoansi consonant distributional asymmetries

• ʙ

  Crowley & Lynch (2006) The Avava Language of Central Malakula

• ɽ

  Evans (1996) First – and last – notes on Wurrugu

  Upton, Parry & Widdowson (1994) Survey of English Dialects

• ɾ

  Penhallurick (1991) The Anglo-Welsh Dialects of North Wales

• ʈ

  Riehl & Cohn (2011) Partially nasal segments, in The Blackwell Companion to Phonology

  Harvey et al. (2013) Two types of pre-stopping in Kaytetye

  François (2010) Phonotactics and the prestopped velar lateral of Hiw

• ɬ

  McDonough & Ladefoged (1993). Navajo Stops, UCLA Working Papers in Phonetics 84

  Wolfgang Kehrein (2002) Phonological Representation and Phonetic Phasing: Affricates and Laryngeals

  Kehrein & Golston (2004) “A prosodic theory of laryngeal contrasts

• ʎ

  Wolfgang Kehrein (2002) Phonological Representation and Phonetic Phasing: Affricates and Laryngeals

  Maddieson (1981) “UPSID: Data and Index”, UCLA Working Papers in Phonetics 53

• ɭ

  Wolfgang Kehrein (2002) Phonological Representation and Phonetic Phasing: Affricates and Laryngeals

• ʟ

  Wolfgang Kehrein (2002) Phonological Representation and Phonetic Phasing: Affricates and Laryngeals

  Maddieson (1981) “UPSID: Data and Index”, UCLA Working Papers in Phonetics 53

• ʧ and ʤ

  Perry (2000) Phonological/phonetic assessment of an English-speaking adult with dysarthria

• ȵ

  Lu (2008) A Grammar of Maonan

• ɤ

  Bekker (2003) The Vowels of South African English

  Wetzels (2009) Nasal harmony and the representation of nasality in Maxacalí, in Loan Phonology

• ʏ

  Edward Flemming (2002) Auditory Representations in Phonology

• æ

  Akademisk Forlag (1974) Dialektstudier, vol. 3

  Andersen (1958) Fonemsystement i østfynsk: på grundlag af dialekten i revninge sogn

  Theo Homan (1975), Skíðaríma: An Inquiry into the Written and Printed Texts, References and Commentaries

• ɷ

  Penhallurick (1991) The Anglo-Welsh Dialects of North Wales

  Anderson (1987) A Structural Atlas of the English Dialects

• ɿ and ʅ

  Lee & Zee () Chinese Phonetics, in The Handbook of Chinese Linguistics

• F [as a tone letter]

  Grønnum (2013) Fonetik og Fonologi

Probably enough attestation

• ʢ
• C [as an IPA wild-card letter]