This help page does not require a rating on Wikipedia's content assessment scale. It is of interest to the following WikiProjects:

Wikipedia Help High‑importance This page is within the scope of the Wikipedia Help Project, a collaborative effort to improve Wikipedia's help documentation for readers and contributors. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks. To browse help related resources see the Help Menu or Help Directory. Or ask for help on your talk page and a volunteer will visit you there.Wikipedia HelpWikipedia:Help ProjectTemplate:Wikipedia Help ProjectHelp articles High This page has been rated as High-importance on the project's importance scale. Languages Mid‑importance This page is within the scope of WikiProject Languages, a collaborative effort to improve the coverage of languages on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.LanguagesWikipedia:WikiProject LanguagesTemplate:WikiProject Languageslanguage articles Mid This page has been rated as Mid-importance on the project's importance scale. Writing systems Mid‑importance This page falls within the scope of WikiProject Writing systems, a WikiProject interested in improving the encyclopaedic coverage and content of articles relating to writing systems on Wikipedia. If you would like to help out, you are welcome to drop by the project page and/or leave a query at the project’s talk page.Writing systemsWikipedia:WikiProject Writing systemsTemplate:WikiProject Writing systemsWriting system articles Mid This page has been rated as Mid-importance on the project's importance scale.

Archives

Index Archive 1 Archive 2 Archive 3 Archive 4 Archive 5 Archive 6 Archive 7 Archive 8 Archive 9 Archive 10 Archive 11 Archive 12 Archive 13 Archive 14 Archive 15 Archive 16 Archive 17 Archive 18 Archive 19 Archive 20 Archive 21 Archive 22 Archive 23 Archive 24 Archive 25 Archive 26 Archive 27

This page has archives. Sections older than 90 days may be automatically archived by Lowercase sigmabot III.

@Mr KEBAB: I'm surprised you made those edits boldly. I have a few questions:

Can the archiphonemes /i, u/ represent [j, w] only before /ə/? So /ˈpwɑːtieɪ/ can only mean /ˈpwɑːtiːeɪ/ or /ˈpwɑːtɪeɪ/, but not /ˈpwɑːtjeɪ/? I'd like to know where you got the idea that the archiphonemes represent something different when preceding /ə/.

Is using /u/ preconsonantally something supported by multiple scholars? For Bugatti, CEPD has /ʊ/ and RDPCE has /ᵿ/ in RP and /ʊ/ in GA. I have a feeling we should limit the use of /u/ to prevocalic environments for the sake of simplicity. Nardog (talk) 12:55, 21 December 2017 (UTC)[reply]

@Nardog: I considered them pretty much harmless, so I just went ahead and made them. If you check LPD, you'll see that listing /iə, uə/ is consistent with listing /ər, əl, ən, əm/, with which nobody has a problem.

This compression to [j, w] doesn't happen in other environments. You're right in saying that poitier is transcribed with both /i/ and /j/. The latter is a US-only variant, and it's not even listed in RDPCE, only in LPD and CEPD.

I've removed Bugatti from the guide.

I think we should treat /i, u/ as proper phonemes that can only occur in unstressed syllables, as Wells does (at least he says that it's a possible analysis). Then we'd say that /ˈpwɑːtieɪ/ represents only [ˈpwɑːti(ː).eɪ] or [ˈpwɑːtɪ.eɪ], whereas /ˌkælɪˈfɔːrniə/ can represent [-iː.ə], [-ɪ.ə] or [-jə]. This makes the IPA less confusing. Mr KEBAB (talk) 13:22, 21 December 2017 (UTC)[reply]

@Mr KEBAB: Do you have a quote? I only have access to the electronic version of LPD3 and am not sure which part you're referring to.

I would like to see evidence that 1) /i, u/ can represent [j, w] (which I empirically know is true, but I haven't been able to find a reputable source that explicitly confirms it); 2) /i, u/ can represent [j, w] only before /ə/; and 3) this is supported and practiced by multiple scholars. Otherwise I don't think I would be able to approve of your edits.

/i, u/ as proper phonemes that can only occur in unstressed syllables Would that mean we would have six high vowels that can occur in unstressed syllables in total? And how would an editor be supposed to know which one to use when they only have access to a recording or a source that doesn't adopt such a convention? Nardog (talk) 14:37, 21 December 2017 (UTC)[reply]

@Nardog: Try Longman Pronunciation Guide -> Pronunciation Notes -> Compression. See also [1], [2] and [3].

Ditto.

We'd have to check recent books on English phonetics because, as you know, /i, u/ are relatively new symbols in English phonetics/phonology. I'll think of something.

You're misunderstanding me. I'm talking only about this guide. We're already using /i, u/ in our transcriptions. Mr KEBAB (talk) 14:53, 21 December 2017 (UTC)[reply]

@Mr KEBAB: Thanks. I'll see what I can find too.

As far as LPD goes, though, it seems the compression to a semivowel is indicated by the linking symbol ‿, not by the letters /i, u/ themselves. I'm still not convinced, including for #1.

Oh, so do you agree that /u/ need not be used for closed syllables? Nardog (talk) 15:09, 21 December 2017 (UTC)[reply]

@Nardog: About half a year ago you said that using ⟨‿⟩ is an overkill. Did you change your mind? If so, I'm for transcribing California /ˌkælɪˈfɔːrni‿ə/ (well, not exactly this word, since it has a straightforward pronunciation, but you know what I mean). It's a good solution and it's non-OR.

You know that I usually side with LPD when there are discrepancies between sources. But that's just me. It's probably wise not to use /u/ before consonants. Maybe we shouldn't use /i/ in that position too? The usual pronunciation of the first syllable of because is [bɪ] or [bə], but usually not [bi(ː)]. Transcribing because as /bi-/ is something I do disagree with LPD on. Mr KEBAB (talk) 15:23, 21 December 2017 (UTC)[reply]

@Mr KEBAB: I stand by that. But I can also get behind the idea that we choose to use /iə, uə/ as meaning varisyllabic, even if that was not a common practice in literature, so we can shorten notations like /ˌkælɪˈfɔːrniə, -njə/.

I think it's best to keep the use of /i, u/ before consonants to morpheme-final positions, e.g. re-, anti-. Otherwise we would invite a great deal of inconsistency. Nardog (talk) 15:38, 21 December 2017 (UTC)[reply]

@Maczkopeti: While I genuinely appreciate all the corrections you've made, please don't say that there's a consensus regarding the use of /iə, uə/. The discussion is ongoing. Mr KEBAB (talk) 15:06, 25 December 2017 (UTC)[reply]

These are not phonemes. I suppose they are listed in the table because /j/ gets elided in yod-dropping accents, but now that {{IPAc-en}} supports /dj, lj, nj, sj, tj, θj, zj/ as diaphonemes, I think what should be on the guide are not /juː, jʊər/ but those consonant clusters. That would make it easier to point out what's wrong with these kinds of good-faith but erroneous edits. Nardog (talk) 13:21, 21 December 2017 (UTC)[reply]

@Nardog: Sounds reasonable to me. I'm for it. Mr KEBAB (talk) 13:22, 21 December 2017 (UTC)[reply]

The sign groupings we choose to encode in {{IPAc-en}} do not make any claims regarding any specific phonemic (or diaphonemic) analysis. Otherwise, they would be original research.

That being said, there are certainly analyses of English that treat /juː/ as a phoneme (historic reasons and because it patterns with other vowel phonemes). And others that do not (because it does not trigger the prevocalic variants of the articles etc.).

That being said as well, I think that either of {{IPAc-en|n|uː|t}}, {{IPAc-en|nj|uː|t}}, {{IPAc-en|n|juː|t}} are fine choices. I see no need for prescriptivism and would go by something akin to MOS:RETAIN. --mach 🙈🙉🙊 19:27, 21 December 2017 (UTC)[reply]

I'm leaning toward being what mach is calling "prescriptive" (or what we might call "inflexible"). From the perspective of someone using the mouseover feature that gives example words, the transcription of {{IPAc-en|n|uː|t}} for newt would be incorrect for anyone who doesn't exhibit yod-dropping. Similarly, {{IPAc-en|n|juː|t}} would be confusing for those who do exhibit yod-dropping because it would guide them to a pronunciation outside of their own dialect. {{IPAc-en|nj|uː|t}} would accomplish the diaphonemic representation we're going for, IMHO. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 04:29, 22 December 2017 (UTC)[reply]

I agree with Aeusoes1 above. {{IPAc-en|nj|uː|t}} would give the person a choice to read it according to his own accent. Furthermore, this chart is not supposed to treat phonemes because of historic reasons (that would be the job of the English phonology page and the historical phonemes articles). The current system just makes it harder to interpret pronunciations for the average reader (the real purpose) while making it easier to interpret historical phonology (which is not what this page is for).--Officer781 (talk) 18:59, 22 December 2017 (UTC)[reply]

I also agree. Allowing {{IPAc-en|n|uː|t}} defeats the purpose of this guide and the mouseover feature of the IPAc-en template. Not completely of course, but it's not something that should be encouraged, let alone done. Mr KEBAB (talk) 15:38, 23 December 2017 (UTC)[reply]

Question: how often is the /j/ retained after /l/ that precedes it within the same syllable? Forms like /ɪˈljuːʒən/ sound very strange to my non-native ears. Mr KEBAB (talk) 20:58, 21 December 2017 (UTC)[reply]

Lute is listed in both LPD and CEPD as having /luːt/ as the most common RP pronunciation. So I checked absolute, Aleutian, allude, allure, dilute, elude, illuminate, illusion, lewd, lieu, lubricate, lucid, lure, lurid, lute, pollute, resolution, revolution and salute in LPD and CEPD, and only lure and lurid in LPD and Aleutian, allure, lewd and lieu in CEPD were given /lj/ as the first variant; i.e., interestingly, no words were given /lj/ as the first in both. An LPD poll reveals 58% of British speakers had /j/ in lure (but note when it's pronounced /-ɜː/ or /-ɔː/, /j/ is mandatory). Maybe the keyword for /lj/ too needs to be replaced. Nardog (talk) 15:13, 23 December 2017 (UTC)[reply]

@Nardog: So I could've checked it myself. I forgot about that feature of LPD/CEPD. Thanks. Mr KEBAB (talk) 15:25, 23 December 2017 (UTC)[reply]

/rjuː/ is long gone, so it may be a general tendency of English to drop /j/ after tautosyllabic liquids even faster than other coronals. (I recently heard someone say /ˈɛrjuːdaɪt/, which struck me as odd. But there are value, Matthew, etc. after all so not that strange I guess.)

And apparently LPD2 (2000) marked yod-coalescence pronunciations as non-RP, which LPD3 (2008) acknowledged as part of RP. So post-coronal /j/ in general seems to be disappearing in the UK too. Nardog (talk) 16:24, 23 December 2017 (UTC)[reply]

@Nardog: Yeah. The 2011 edition of CEPD goes a step further and recommends the coalesced pronunciations to readers who want to speak RP. Mr KEBAB (talk) 16:41, 23 December 2017 (UTC)[reply]

For what it's worth, I found the following in Daniel Jones' Outline of English Phonetics (1918, p. 67):

Usage varies in words in which l is initial or preceded by an unstressed vowel; thus lute, absolute are pronounced ljuːt, ´æbsəljuːt by some, and luːt (like loot), ´æbsəluːt by others; the forms with j are generally recommended by elocutionists, but the forms without j are, if anything, the more usual in ordinary speech, at any rate in the commoner words.

So apparently the pronunciation with /j/ was already old-fashioned 100 years ago. Nardog (talk) 22:31, 20 February 2018 (UTC)[reply]

There exist loanwords such as föhn, Möbius, Peugeot, and pho, in which a vowel (usually a front rounded one in the native language) is pronounced as the NURSE vowel in non-rhotic accents, even though it is not followed by ⟨r⟩. In rhotic accents,^[1] such a vowel is pronounced as either:

1. another vowel such as /uː/ (Betelgeuse) or /ʌ/ (pho), sometimes even violating the phonotactics;
2. the marginal vowel /œ/ (found in American dictionaries, though I'm not sure how much this is meant to be descriptive of English speakers' actual production as opposed to the pronunciation in the original language^[2]);
3. a lengthened schwa-like vowel, just like the non-rhotic NURSE; or
4. the usual rhotic NURSE vowel, as in many speakers' pronunciation of Goethe and hors d'oeuvre.

And so long as what /ɜːr/ as a diaphoneme tells readers is "Pronounce this /ɜː/ if you're non-rhotic, /ɜ(ː)r/ if you're rhotic", it wouldn't be accurate to use /ɜːr/ in cases #1–3. One might say we can still write e.g. "UK: /fɜːr/, US: /fʌ/", but that doesn't really solve the problem because it is highly unlikely that e.g. a rhotic UK speaker would pronounce the word with a coda /r/ instead of improvising a non-rhotic /ɜː/ or opting for the US variant. It should be able to write "English pronunciation: /fɜː/, English pronunciation: /fʌ/".

CEPD18 includes at least 42 instances of case #3^[3] and 17 of case #1,^[4] and LPD3 includes at least 52 instances of #1,^[5][6] counting only the first variants in each accent. In fact, there are already at least a dozen articles which could use /ɜː/,^[7] a bigger number than that of the articles currently using /ɔɪər/ or /æ̃/. So why don't we add /ɜː/ as a marginal segment? Nardog (talk) 14:37, 7 January 2018 (UTC)[reply]

@Nardog: I don't know if we should add it, but what I'm curious about are US-specific pronunciations with /ɒ/ which doesn't exist as a separate phoneme in General American and in the vast majority of regional accents of the North America. See American and British English pronunciation differences, for example. Pronunciations such as /ˈkɒzməs/ are listed as American and they're not. /ˈkɑːzməs/ is American.

The same applies to UK-specific (more like Anglo-Welsh-specific) pronunciations with coda /r/, which for the majority of speakers isn't there even phonemically and mantaining it in speech is non-standard. After all, here UK usually stands for RP/Estuary, no? Northerners wouldn't even think of saying /ɡrɑːs/ etc. and that's what dictionaries prescribe.

Perhaps it's about time we split the pronunciations into GA and RP? If we do that, transcribing words in both accents should be mandatory in all articles as long as they're sourced. We can think about other accents if there are dictionaries/wordlists to back up the IPA.

I'm also curious about how many people actually use the mousehover feature. If we dropped it and just used IPA-en as we do for other languages, would that make a negative difference? I think it wouldn't. Many readers have admitted to me that they either don't use the mousehover feature, they don't understand our diaphonemic system or even both. Mr KEBAB (talk) 14:54, 7 January 2018 (UTC)[reply]

Please stay on topic. ;) Abandoning the diaphonemic principle is such a big change with so many ramifications throughout Wikipedia that we cannot possibly hope to discuss it here.

Nonetheless, you raise a good point: One thing WP:RHOTIC (or WP:PRON as a whole) or this guide don't specify is what to do when IPA(c)-en is used explicitly for a pronunciation in a specific dialect. Should it be e.g. "US: /ˈkɒzməs/" or "US: /ˈkɑːzməs/"? I agree it should be the latter. Shouldn't it be "English pronunciation: /ˈkɑzməs/", you ask? I'd say no. I know length isn't phonemic in GA, but it'd be way too much to ask average readers (and editors) to remember /ɑː/ and /ɑ/ are equivalent but should be differentiated based on dialectal difference. In fact LPD and CEPD do use /ɑː/, /iː/, etc. even for GA, instead of supplying a distinct set of symbols for each accent (RDPCE's approach). And we could prescribe essentially what LPD/CEPD do when IPA(c)-en is used to describe a pronunciation in a specific accent, so readers and editors wouldn't have to be bewildered by so many symbols. That would be another rationale for adding /ɜː/. Nardog (talk) 16:03, 7 January 2018 (UTC)[reply]

Not to throw a wrench in here, but I've only ever heard fellow North Americans use the pronunciation /ˈkɒzmoʊs/ (or, for most U.S. citizens, yes, /ˈkɑːzmoʊs/). Just listen to just the first minute of this interview to hear two Americans' pronunciations of the word, one being the lifelong scientist Neil deGrasse Tyson. Two Americans (of three) on Forvo also use this pronunciation. Although it's just my impression, I'd wager that the schwa pronunciation is an older one that is receding. Wolfdog (talk) 16:22, 7 January 2018 (UTC)[reply]

The example is from American and British English pronunciation differences, which covers that variant too. Nardog (talk) 16:37, 7 January 2018 (UTC)[reply]

I'm having trouble understanding the problem with UK: /fɜːr/, US: /fʌ/. What do we want this hypothetical rhotic UK speaker to do that this wouldn't accomplish? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 19:56, 7 January 2018 (UTC)[reply]

To not pronounce the /r/. And don't forget there's also case #3. CEPD has a bunch, so we are currently unable to cite it accurately with {{IPAc-en}}. Merriam-Webster also includes variants with \ə̄\, i.e. [ɜː].^[8] Nardog (talk) 20:30, 7 January 2018 (UTC)[reply]

I think I see what you're saying. Because we're using "UK" to mean "non-rhotic" it might confuse rhotic UK speakers into producing the wrong pronunciation. The easiest solution might be explanatory footnotes. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 23:55, 7 January 2018 (UTC)[reply]

Not just that. There are also words that reliable sources say are pronounced with /ɜː/, not /ɜr/, in GA. In addition, Mr KEBAB has suggested that when IPA(c)-en is used to illustrate a pronunciation explicitly in one dialect or the other, we not follow the diaphonemic principle and use e.g. /ɑː/ in place of /ɒ/ for US pronunciation. I echo his suggestion, and should it be accepted and incorporated into MOS:PRON, we would not be able to notate the UK pronunciation of words with the NURSE vowel with IPAc-en without adding /ɜː/ as another diaphoneme. Nardog (talk) 02:58, 10 January 2018 (UTC)[reply]

I feel like we might need some investigating on the presence of /ɜː/ as a non-native phoneme in rhotic dialects of English (akin to use of front rounded vowels or uvular rhotics in English) that goes beyond just the transcriptions found in some dictionaries. Also, is it even a good idea to use IPc-en if it's not a diaphonemic transcription? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 15:29, 10 January 2018 (UTC)[reply]

@Aeusoes1: Well, that's difficult when it comes to marginal segments isn't it. But listen, for example, to people pronouncing Goebbels. Some rhotic speakers do indeed pronounce it with non-rhotic /ɜː/ or some equivalent of it, but it eventually comes down to how much one is acquainted with the donor language and how much one is willing to approximate the pronunciation to it, as with any other marginal segment. The fact many people simply have a syllabic R in Goethe and hors d'oeuvre and other vowels in Betelgeuse, pho, etc. suggests non-rhotic /ɜː/ in rhotic speech may perhaps only be present during a transitional period in which the word has not yet completely assimilated into English (which makes it a marginal segment). Nevertheless, isn't making it possible to write e.g. "English pronunciation: /fɜː/" enough of a reason to add /ɜː/ for you? I don't quite see how a footnote can be an easier solution.

Describing a pronunciation in a specific variety using IPAc-en doesn't automatically make it not diaphonemic. Rather, e.g. "English pronunciation: /ˈkɑːzmoʊs/" is a transcription of a pronunciation in one variety using our diaphonemes, within the confines of what the diaphonemes allows us to do. /ˈkɒzmoʊs/ wouldn't be accurate because GenAm doesn't have /ɒ/. /ˈkɑːzmoʊs/ using IPAc-en helps readers identify what phoneme each symbol corresponds to more easily than e.g. /ˈkɑzmos/, which requires them to learn a whole separate set of symbols for that specific dialect. Nardog (talk) 15:43, 16 January 2018 (UTC)[reply]

Sorry, but it seems like you don't understand how this works. Describing a pronunciation for a specific variety that way does, indeed, make it not diaphonemic (KEBAB's suggestion was to do away with the diaphonemic system). Under the logic of our diaphonemic system, if we transcribe cosmos as /ˈkɑːzmoʊs/, we're saying it's pronounced with the PALM vowel, when it should be the LOT vowel. For General American, these have merged, but the point of the diaphonemic system is that we mark contrasts don't appear in all varieties. That's why the phrase "a transcription of a pronunciation in one variety using our diaphonemes" doesn't really make any sense. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 17:15, 16 January 2018 (UTC)[reply]

@Aeusoes1: I see what you're saying, but IPAc-en does already have the built-in prescripts "UK" and "US" and is used for pronunciations in one dialect or another in many articles. And who/what says one can't use the diaphonemic system, the point of which, yes, is to mark contrasts that don't appear in all varieties, to describe a pronunciation in a specific variety, and not mark contrasts that don't appear in that variety―especially when the pronunciation of the word(s) in another variety is also provided?

(Mr KEBAB said "/ˈkɑːzməs/ is American", with ⟨ɑː⟩ not ⟨ɑ⟩, so I'm not sure if I agree with your assessment of his statement. His ultimate suggestion was indeed to do away with the diaphonemic system, but the insinuation I got there (in his first paragraph) is that we should use /ɑː/ in place of /ɒ/ as far as GA goes should we use the diaphonemic system to notate it. This he can clarify.)

Consider, for instance, what to do when describing the General American pronunciation of lieutenant. Should it be /ljuːˈtɛnənt/ because those who pronounce lieu /ljuː/ might pronounce the /j/? But such people would most likely pronounce it /lɛfˈtɛnənt/. /luːˈtɛnənt/, which describes a pronunciation in one variety using the diaphonemes, signals the pronunciation better than the awkward, theoretically diaphonemic /ljuːˈtɛnənt/ without making the readers having to learn an additional set of symbol/phoneme combinations. Nardog (talk) 18:10, 16 January 2018 (UTC)[reply]

The built-in prescripts are for cases when UK and US pronunciations are the result using different diaphonemes, such as your lieutenant example, not for cases like cosmos where it's just a case of differing vowel inventories or phonetic nuances. So the accurate use of the diaphonemic system with your example would be something along the lines of:

Lieutenant (US: /ljuːˈtɛnənt/, UK /lɛfˈtɛnənt/) ...

The problem with using IPAc-en, which is for diaphonemic transcriptions, for transcriptions that aren't diaphonemic, is that it then undermines the work done to get readers to understand the transcriptions. We should be consistent and using dialect-specific pronunciations like /ˈkɑːzmoʊs/ is being inconsistent. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 21:19, 16 January 2018 (UTC)[reply]

@Aeusoes1: I disagree. When a source says /luːˈtɛnənt/ is the modern General American pronunciation of lieutenant, there is nothing that indicates the first syllable is underlyingly /ljuː/. Historically, yes, it may have been the source of the current realization /luː/, but unless some source says it's still pronounced with /j/ in some variety in the US, "US: /ljuːˈtɛnənt/" wouldn't be an accurate diaphonemic transcription. (I don't understand why cosmos is not a case where "UK and US pronunciations are the result using different diaphonemes". In your suggestion it would be something like "UK: /ˈkɒzmɒs/, US: /ˈkɒzməs, -moʊs/".)

Anyway, we've gone off on quite a tangent. What do you think of /ɜː/ as an additional marginal diaphoneme? I don't see why "the easiest solution might be explanatory footnotes". Nardog (talk) 16:06, 18 January 2018 (UTC)[reply]

I think the lj in lieutenant is a bit more debatable, but you're right that this is a minor point. I think you get my point regarding why there are built-in US and UK prescripts. The way you've indicated the variant US/UK pronunciations is, I think, just right (my comment about cosmos was more with the vowel of the first syllable).

In answer to your question, I still don't like it. For one, no one has put forth academic research on this, so that doesn't help us on understanding who does this, why, and why dictionaries have been putting this in their pronunciations. Is it a botched attempt at pronouncing a front rounded vowel in borrowed words? In addition, the symbol ⟨ɜː⟩ would be about as confusing as using ⟨a⟩, since that's a common symbol used for pronunciations to mean something slightly different. An explanatory footnote is the ideal choice with these words because there is too much variation within and between the words. Goebbels might sometimes be pronounced with this sound, but at other times with the vowel of nurse. Pho will possibly have this sound, but may also be pronounced with the vowel of goat. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 17:54, 18 January 2018 (UTC)[reply]

I was led here because, as as speaker of northern English (though well south of the Scottish border), I was confused by Wikipedia's insistence on /ɜː/ appearing as /ɜːr/ when I would never think of pronouncing the r. My pronunciation is largely non-rhotic (close to RP), but with just the occasional slight rhotic hints, so I would appreciate the addition of /ɜː/ in IPc-en to avoid confusion for those of us who have even a slight British rhoticism. I'm not an expert on IPA, so I will leave it to you experts to make a decision. Dbfirs 08:19, 3 February 2018 (UTC)[reply]

There are a number of problems with your proposal. Without getting too technical, putting /ɜː/ as separate from /ɜːr/ would mean that they were two distinct vowels. Even if there were dialects that made a contrast between the two, if the contrast is not encoded in dictionaries, we wouldn't have a way of verifying transcriptions. Even if we could verify this contrast in transcriptions, it would not be a good idea to represent one of those vowels as ⟨ɜː⟩ because that symbol is commonly used for the vowel of words like nurse for non-rhotic pronunciations and would likely confuse readers.

I'm not sure what you mean by "a slight British rhoticism" though. Do you contrast two vowels like this or do you switch back and forth between rhotic and non-rhotic pronunciations so that the same word can be pronounced in two different ways? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 14:56, 3 February 2018 (UTC)[reply]

No, I don't contrast the two vowels. I usually pronounce them both as əː (long schwa, slightly less open than ɜː), and in words such as nurse, I might, sometimes, for emphasis, sound the r either partially or fully. For words such as milieu, adieu, cordon bleu etc I would never use a linking R when followed by a vowel, as I would if the pronunciation were /ɜːr/. Dbfirs 18:45, 3 February 2018 (UTC)[reply]

That's interesting. So you have linking R for a word like miller but not for milieu? Or are you saying that there are words that you vary in your pronunciation but that there are other words that you still pronounce non-rhotically even when you are speaking in this emphatic fashion? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 19:04, 3 February 2018 (UTC)[reply]

Yes, the pub The Miller of Mansfield would have a linking R in nearly all British accents, but in the milieu of the battle would have a slight hiatus, avoiding a linking R except by those (too many in the UK) who introduce an intrusive R where none should exist (e.g."Victoria-r-and Albert Museum"). My mention of partial rhoticism is probably a red herring -- I'm not too far from the border with Scotland where speakers are fully rhotic and where readers would be very confused by /ɜːr/. Dbfirs 19:37, 3 February 2018 (UTC)[reply]

In what way would such speakers be confused? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 01:51, 7 February 2018 (UTC)[reply]

Regarding Mr KEBAB's edit, there is a variation in the pronunciation of ⟨ea⟩ in words like "idea", "diarrh(o)ea", "theater"/"theatre" and the aforementioned "Korea". In OED, these are noted as /ɪə/ for UK and /iːə/ for US. It seems that for UK there's a reduction to a non-rhotic /ɪər/, while some US dialects also seem to reduce the /iː/ in a similar manner to /i/. In theory, there could be a similar stressed variant of /uə/ too, but I've yet to find an example for that.
--maczkopeti (talk) 13:10, 7 March 2018 (UTC)[reply]

/i/ (the happY vowel) is just a non-phonemic notation of the high front vowel found in unstressed and morpheme-final or prevocalic positions, traditionally identified as belonging to the KIT phoneme (see [4][5]). So there could be no such thing as a "stressed /i/" as far as our diaphonemic system goes.

In most US dialects there is no distinction between /iː/ and /i/ (i.e. no appreciable difference in segmental quality, although some phonologists might still identify the happY vowel as belonging to the KIT phoneme). It is true that in RP the stressed vowels in words like idea, Korea, theatre, real may be realized as the same as NEAR, but they obviously cannot in GA. And in accents like RP where this can happen, the second segment in any diphthong before /ə/ can also be elided with the schwa losing its syllabicity (e.g. [faə̯] fire), suggesting [ɪə] in the aforementioned words is the result of this type of diphthongization ("smoothing") from /iːə/ rather than originally having the NEAR vowel (Wells 1982:240 says This can be interpreted as evidence in favour of analysing FLEECE and GOOSE as underlyingly diphthongal, /ɪi, ʊu/; Smoothing then has its usual effect of producing a monophthong with the phonetic quality of the starting-point of the underlying diphthong). So in our diaphonemic system they are best transcribed as /iːə/ (meanwhile, this does constitute yet another reason not to transcribe them as /iː.ə/, with a syllable break). Nardog (talk) 15:26, 7 March 2018 (UTC)[reply]

I agree with Nardog. /iːə/ is the best choice for this. Can we expect non-rhotic speakers to read this correctly without guidance? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 17:04, 7 March 2018 (UTC)[reply]

Confused Christian t (talk) 15:10, 2 April 2018 (UTC)[reply]

@Christian t: About what? Mr KEBAB (talk) 15:16, 2 April 2018 (UTC)[reply]

Per the recent edits, particularly this addition, I'm not sure if we have a consensus on that. It seems from this conversation back in December, people seemed to accept that this is a stylistic choice. I'm fine either way, but I don't want cycle of reversions every three to six months on the matter. Anyone care to make their case? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 18:09, 26 March 2018 (UTC)[reply]

I'm also confused by Kwami's edits. As far as I know, no dictionary which uses IPA puts stress in monosyllabic words, so to do so on Wikipedia is WP:UNDUE ("we have done for decades" is not a valid reason—nor true, it's idiosyncratic at best). No stress in a monosyllable does not mean unstressed, it just means neither stressed nor unstressed, as far as lexical stress is concerned, because in order for a syllable to be stressed there needs to be another syllable that is unstressed or less stressed. And we do, in the vast majority of the cases, record lexical stress, not stress in running speech—in running speech lexically stressed syllables are unstressed and lexically unstressed syllables are stressed all the time. There's no reason for it to be there. Nardog (talk) 18:39, 26 March 2018 (UTC)[reply]

We have in this recent thread an example of where the previous determination may have left too little guidance. If we decide that it's a stylistic issue, we also don't want editors going back and forth over it, even if they aren't intentionally warring. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 03:57, 27 March 2018 (UTC)[reply]

Thanks Aeusoes1! I must say I am slightly confused.
It was me who had revised Kwamikagami's edit here, for the same reason that Maczkopeti gives here, based on the convention followed by Wiktionary. If it is true that "no dictionary which uses IPA puts stress in monosyllabic words" then I wonder why Wikipedia does it differently? So far I haven't seen any rationale for why the unusual convention was adopted in the first place (actually I am not convinced yet that it actually has been the convention at all, regardless of what Kwami says).
Citing Mr KEBAB: "Editors' preference. Remove the stress marks if you want, both variants are correct." (from here). Obviously there is no consensus that this is true. --Renerpho (talk) 13:42, 28 March 2018 (UTC)[reply]

@Renerpho: I stand by that. As long as you're transcribing words in isolation rather than phrases, both variants are correct as there's only one way to realize stress in a monosyllabic word (more accurately: a monosyllabic phrase, because stress in words said in isolation is both simple word stress and phrasal stress). When you're transcribing phrases, you must place stress marks in a correct manner, otherwise your transcription is incomplete. Mr KEBAB (talk) 14:23, 28 March 2018 (UTC)[reply]

@Mr KEBAB: Right now, the article says "Some dictionaries use a stress mark to distinguish two-syllable from one-syllable words, but on Wikipedia we mark all stressed syllables." Should that be removed then? --Renerpho (talk) 15:04, 28 March 2018 (UTC)[reply]

@Renerpho: I have no strong preference. I'd rather not transcribe stress in isolated monosyllables, but it's not a very big deal to me. Mr KEBAB (talk) 15:13, 28 March 2018 (UTC)[reply]

If you pronounce a lexically unstressed monosyllable in isolation, it will take the prosodic stress of the phrase. But that doesn't mean there's no lexical distinction. There is a phonemic lexical distinction of stress in English monosyllables, and to transcribe a phonemically stressed word without stress in a phonemic transcription is just as wrong as transcribing it without vowels. True, I doubt there are many cases where the IPA trascription in the lead is going to be for an unstressed monosyllable, since nearly every WP entry is for a noun, but leaving out the stress is still technically incorrect (and would be critical in discussing the variable stress of pronouns or auxiliary verbs, for example). And when we transcribe a phrase, the distinction is certainly relevant (e.g. /ˈprɪns əv ˈweɪlz/).

BTW, I have used a bot to add stress to all IPA-en transcriptions in WP that didn't have it, unless the words were truly unstressed. And that was accepted -- I did it more than once. So we certainly had consensus at the time.

My main problem with giving false information for our convenience, under the assumption that the reader will understand it's false, is that in general they won't. The last time I was in a discussion about this, there were users who argued that such words actually were unstressed because dictionaries didn't mark stress for them. If many of us don't understand the convention, how can we expect our readers to? And the convenience of omitting phonemic stress is minimal, so why do it? — kwami (talk) 19:00, 28 March 2018 (UTC)[reply]

@Kwamikagami: Oh, that's what you mean. I suppose you have a point there. I guess it is one of the best examples, when it's /ˈɪt/ the vowel is relatively front [ɪ] and when it's /ɪt/ the vowel is prone to being centralized to [ɪ̈] or it even merges with [ə]. Mr KEBAB (talk) 06:24, 30 March 2018 (UTC)[reply]

Yes, and since those vowels are now transcribed the same, we rely on the stress mark to distinguish them (though of course there are cases where there is a full unstressed /ɪ/ -- we no longer have any way to indicate that). — kwami (talk) 06:46, 30 March 2018 (UTC)[reply]

@Kwamikagami: TBH I'm not aware of a single dictionary that would get reduced /ɪ, ʊ/ right, so removing them was pretty much the only option we had. Mr KEBAB (talk) 12:43, 31 March 2018 (UTC)[reply]

Even Longman? — kwami (talk) 17:45, 31 March 2018 (UTC)[reply]

Can Lao-Tzu really be pronounced /ˈlaʊˈdzʌ/ or is that just another invention of certain dictionaries? I strongly suspect it's the latter. Mr KEBAB (talk) 17:57, 28 March 2018 (UTC)[reply]

It could be the closest theoretical English approximation of the Chinese, rather than an actual English pronunciation. You certainly see things like that a lot. But, while marginal, you do occasionally see stressed 'short' vowels in words like pho and meh. (Thus the joke in the restaurant name "Pho King Delicious".) So it wouldn't be unheard of. — kwami (talk) 19:06, 28 March 2018 (UTC)[reply]

There is no /ʌ/ in Standard Chinese phonology. Pronunciation would be something like /t͡sɨ/. Not sure if I'd call /dzʌ/ a close approximation. --Renerpho (talk) 21:56, 28 March 2018 (UTC)[reply]

"Lao-dzer" in a non-rhotic accent would be closer, or maybe a syllabic /z/, but otherwise /ʌ/ is the closest I can come up with. — kwami (talk) 06:51, 30 March 2018 (UTC)[reply]

Thanks for the responses. IMO /ʌ/ isn't very close to the original Chinese sound. In comparison with it, it's very open and lacks the alveolar friction. /uː/ is a better approximation. Mr KEBAB (talk) 12:42, 31 March 2018 (UTC)[reply]

That may be true, but aren't we going with what is actually cited in dictionaries or used by English speakers? Wolfdog (talk) 19:17, 2 April 2018 (UTC)[reply]

@Wolfdog: I just responded to Kwami, nothing more. I'm not challenging any dictionary. Mr KEBAB (talk) 19:30, 2 April 2018 (UTC)[reply]

Right. I agree that it seems to be strangely derived. Wolfdog (talk) 19:33, 2 April 2018 (UTC)[reply]

I think we should replace them with ordinary ə|l, ə|m and ə|n for a few reasons:

• əl, əmand ən look the same as ə|l, ə|m and ə|n. There's no way to distinguish them unless you use the mouseover feature. How many users/readers of Wikipedia do that? Probably not that many. This renders this distinction completely irrelevant to those who don't use the mouseover feature.
• For speakers that vocalize their /l/'s /əl/ in treacle or bubble is the same as /l/ in mail or oil. Cockneys would pronounce these [ˈtɹ̥ekʊ, ˈbɐbʊ, ˈmæɪ̯ʊ ~ ˈmæʊ̯, ˈoɪ̯ʊ] (forgive me if any of these particular words isn't used by cockneys, I don't know their rhyming slang very well).
• For many speakers of Estuary English the distinction between [əm, ən] and [m̩, n̩] (and [əl] and [l̩], as pointed out above - EE is watered down Cockney after all) is no longer relevant. Older speakers (of RP in particular) perceive the presence of a phonetic schwa in words such as button as sounding childish, but for younger speakers it sounds normal (Gimson (2014:171)).
• Dialects vary somewhat as to where exactly [n̩] can be used. RP and GA don't allow it in London, but the Cardiff dialect does (Collins & Mees (1990:90), Wells (2008)).

Even if we do remove əl, əmand ən from the IPAc-en template, we should still mention syllabic [l̩, m̩, n̩] on this page as they are rather important and auditorily distinct allophones.

Sources:

Thoughts? Mr KEBAB (talk) 15:44, 11 April 2018 (UTC)[reply]