Jump to content

Help talk:IPA/English/Archive 9

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
Archive 5Archive 7Archive 8Archive 9Archive 10Archive 11Archive 15

code switching

from a user's talk page addressed to Kudpung:

In order to continue, I need to know whether you have linking or intruding ar or not, something you've so far refused to say. Without that, I can't evaluate your claim. So, would you, speaking your best (snobbiest?), say "Worcester is home" (or any appropriate linking environment) with an /r/ before "is"? If you don't, would it be considered correct to do so? Then, take "Anglia is home" -- is there an /r/ there? If there is, is that considered correct? kwami (talk) 12:08, 18 January 2010 (UTC)

First and foremost, to set the record straight, the claims are not mine; several Wikipedians, among them the micromanagers of Wikipedia Projects, have suggested on this discussion and on other talk pages that there may be some weaknesses in the way the Wikipedia assumes that British place names should be transliterated into the IPA, and because they are probably not linguists, it has not been possible for them to express their concern in a way that the linguists here can understand and address in simple, non lingo-technical terms.

For the benefit of readers of this thread who are not linguists, and to answer the question in italics above: My best, snobbiest RP does NOT pronounce the linking r at the end of Worcester when the next word begins with a vowel. There are in fact several kinds of RP and they have all changed over the last 60 years since I started speaking and was educated in awfully rather posh schools. That 'posh' unlinked r is how BBC newscasters spoke in the 50s and how the Queen still speaks when reading from a prepared script - such as her speech at the opening of the parliamentary session, and her Christmas speech to the nation, and on other formal occasions. In more private circles and when on a walk-about, her speech has become a tiny bit more laid back, while two generations later, her grandchildren William and Harry speak much the same as any reasonably well educated kids of their generation. The accent of their father, Prince Charles, is still in many ways, more affected (snobbish) than that of his mother, the Queen. Contrary to some popular ideas expressed on this encyclopedia and elsewhere, nobody in Britain is in a hurry to speak with the same accent of the Queen. A form or RP, or regionally and socially devoid accent, is becoming exponentially widespread in the land, and regional accents among the younger genrations are becoming barely discernible.

However, in TESOL, for example, we expose our students to a wide variety of English accents to enable them to even distinguish words from the worst American slur, to perfectly fluent, but educated, native Indian English, and Thai and Chinese English gibberish (Tinglish & Chinglish). What I teach them is a more modern RP without the awful affectations of the 40s and 50s, but they very often have to learn the bulk of their pronunciation from either their indigenous non-native English speaking teachers, or from native English TEFLers who may speak with a strong American or British regional or cultural bias, and are not capable of code switching to any form of RP. Generally, the IPA is only of interest to those who need a pronunciation reference across several langauges. Anyone just learning one second language just has to 'listen and repeat' in the traditional way, with some help from the teacher in the physical aspects of producing some sounds that may be new to the learner. Very few second langauge learners will ever achieve even a near native pronunciation, it's rarely necessary, and their utterances will allmost always be coloured by traces of their own first language accent. You can be in a conference room full of mixed nationalities all speaking perfectly fluent French for example, but you will notice their origins from their accents. In TESOL, (outside the US), frequent use of the IPA is made because many exams, including TESOL certification, demand it. (just so you know, in my university I taught phonetics, phonology, morphology, and syntax in the Graduate School). Unlike some of the IPA specialists that contribute to the IPA articles in the Wikipedia, they are not expected to become perfectly fluent in their use of it. They will however, be able to look at a word of up to , say, three sylables, and recognise quite accurately how it will sound. One side effect of learning the IPA is that because it is a script, it can be an immense help in preparing learners how to use non Roman scripts. Musicians score quite well when learning the IPA because they have already learned to read a language in a non Roman script that represents sounds: that of music. Generally, the IPA is a working tool for teachers, linguists, and students learning several languages simultaneously. It does not demand total understanding of it as a stand-alone subject, any more than I speak fluent Thai, without a university degree in it, for running my business here. One of the problems in this discussion has been as far as I can see, is that the non linguists here have been brushed off and scared away by a lot of technical gibberish by IPA specialists who consider any non IPA experts as complete idiots, failed linguists, and trolls.

  1. It has been my suggestion that the IPA transcription of British place names should represent as closely as possible, the natural way that most speakers in that country would pronounce it, or at least most easily recognise it wen hearing it. This means that more attention should be given to the questions of Who are the readers of the IPA articles? and What are their needs? Too many editors think the only people who read the Wiki are the one who write it.
  2. It has been my suggestion that a Standard Global English is a fallacy. Alone in the United Kingdom,my homeland, I do not understand one word of very broad Scots. As a lexicographer, and author of an American-British bilingual dictionary and grammar, I can vouch (I can't supply citations here for obvious reasons) for the fact that there are hundreds of fundamental differences between General American, and modern, common RP.
  3. It has been my suggestion that where differences between AE and BE are great, then two pronunciations of a place name should be shown.
  4. It has been suggested by those who opposed dual entries that this would mean every pronunciation in every accent of English would also then have to be shown. That is, IMHO, a silly argument to use in opposition.
  5. Notwithstanding the current discussion, some editors have deliberately antagonised the debate and stretched the GF of participants by carrying out a programme of disruptive editing, by turning non rhotic place names into rhotic place names.
  6. Serious editors who started the discussion felt they were not getting satisfactory explanations to their enquiries. They either gave up trying or they took their discussion elsewhere.
  7. Finally, after all these weeks, the dicussion has been reviveed albeit through a particularly nasty echange, and we have a detailed explanation that could have been provided if Good Faith from its author had been demonstrated weeks ago.
  8. I partly agree with that explanation, but do not agree to implement it on the Wikipedia for the reasons stated in the preamble to this message.
  9. This thread is far from closed. Any suggestion to close this discussion prematurely is because the issue is too embarrassing to those who don't want to waste their time addressing questions from non IPA specialits. No consensus has been reached, and if it were it would represent my White House analogy.
  10. Because of constant incivility, I'm not going to answer any questions on the above. If you can be bothered to read it (and some of you have admitted that you are not interested in reading every post in detail), and you find it in any way helpful, please dicuss it amongst yourselves and leave me out of it.--Kudpung (talk) 22:02, 28 February 2010 (UTC)
Great! I think I finally understand what your objections are. If you had answered such simple questions when I'd asked them months ago, it would have saved a lot of frustration on all sides. I'll ignore your repeated and I can only assume knowing misrepresentations of what's going on here, and skip to the useful bits. (Yes, I've previously skimmed some of your missives, as I'd found them overly long and generally devoid of content. I'm glad we're back to the point where it's profitable to read what you have to say.)
It is true that our IPA conventions were not designed with the ESL student in mind. This is certainly worth discussion. How much should our transcriptions be targeted to the reader who knows English, and how much to the reader who's learning it? If we target the ESL student, should we then pick a representative dialect or dialects for WP, say just RP or just RP and GA, and ignore the other dialects of our native speaking readers? Should any of this be the job of the Special English WP maybe? Or is that inadequate for it?
We do use conservative RP as the basis for our pronunciations. For example, we maintain the hoarse, horse distinction, which is now merged among most RP speakers. This is because it's easier to make a merger than to undo one, if the distinction is not made in the transcription. Of course, this is also s.t. we can revisit.
We did use to have multiple transcriptions of a word, RP, GA, and Oz. It started to become a mess, and general consensus is that we should use a single transcription where possible. True, as you point out, there are numerous cases where this is not possible, as in the alt pronunciations of "graph", but it works the vast majority of the time. This is also s.t. that can be revisited, but I for one have no desire to go back to the mess that we once had. Of course, it would make a difference if we only used RP for English place names, but I can already see the objections to that.
We do, and always have, supported the inclusion of local pronunciations. No-one has ever had any problem with that, except in trivial cases. But although it's quite useful to know how locals pronounce the name of their town in their own dialect and accent, it's also useful to know how one should pronounce the name in one's own accent when talking with one's compatriots. If two Torontans are talking, and one of them mentions "Yawk", the other might not recognize that he means "York". Even if it were recognized, it might sound pretentious. So both are important, both are desired, and IMO where they differ, both should be included. Your POV would seem to be, based on your deletions, that only the local pronunciation should be provided. I find that unfortunate.
As for "Global English" being a fallacy, no-one has said that this is how people should speak. In fact, we've been careful to say just the opposite: the IPA is a key that enables the reader to decode the pronunciation of a word in their own dialect (within what we support; we have unfortunately not been able to support Scottish). Nowhere do we say that they should try to imitate our transcription as their pronunciation. If we've implied that, please point out where, and we'll change it. That is not our intent.
Once again, you say that we all need to carefully consider your opinions, and discuss them among ourselves, but that you have no intention of actually defending them or answering questions if we find them unclear. This is not due to any incivility, but has always been your approach. It's difficult to take you seriously as an editor with such an attitude. kwami (talk) 22:31, 28 February 2010 (UTC)
Based on this post in my talk page, I think your interpretation of Kudping's stance is correct.
While Kudping may not be returning here, I'd like to point out that Kwami has shown that using multiple systems actually requires a more thorough knowledge of the IPA and phonology than a single system. People complain enough about our using IPA without us making it more complicated. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 04:01, 1 March 2010 (UTC)
Coming from the opposite POV from Kudpung, I just got an objection that paradigm should have the "merry" vowel rather than the "marry" vowel, because that's how it's pronounced by most North Americans, and therefore by most English speakers.
And why haven't people objected to Hampshire having an /h/? kwami (talk) 07:14, 4 March 2010 (UTC)
The "paradigm" objection came from me. After having read the "Wikipedia:IPA for English" page, I think I have a better comprehension of what you're trying to do here, and I applaud your efforts. I am still concerned, though, that you may be fighting a losing battle against people who simply do not understand the problem, simply will not go read this page, and will continue to insist that various IPA transcriptions in Wikipedia are just plain wrong. And I'm not totally sure that readers should need to become dialect experts in order to understand pronunciation notes in WP — though I realize you've all been dealing with that question for a long time now, and I'm not sure if I have any useful suggestions that you haven't already thought of. Richwales (talk) 17:09, 5 March 2010 (UTC)

Perhaps the "Understanding the Key" section should always be completely visible

It's pretty clear and well-written, and might help with some of the confusions. Even though I come to this page constantly (mainly to cut and paste), I wouldn't mind seeing it every time. Grover cleveland (talk) 18:15, 1 March 2010 (UTC)

Okay, let's try that. I hid it assuming that it wouldn't be needed after the first visit.
Do you not have the list of IPA symbols under your edit window that you can click on to place IPA in text? kwami (talk) 21:44, 1 March 2010 (UTC)
Yes, but it's so hard to find the symbols that I need among all the jumble there that I find it much easier to copy/paste from this page. Grover cleveland (talk) 22:50, 1 March 2010 (UTC)
I'd even propose to keep the list of symbols itself hidden, with an instruction to read the "Understanding the Key" section first near the [show] link. ― A._di_M. (formerly Army1987) 16:12, 2 March 2010 (UTC)
Do you want to try that out? kwami (talk) 21:47, 2 March 2010 (UTC)
I don't think that's a good idea. If we're expanding the "understanding the key" box because we think people might be missing it, we don't want to compliment that by hiding the actual key since there will be people who miss it. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 23:07, 2 March 2010 (UTC)
Perhaps there is a way to set user preferences so that regulars can default to hidden while newbies default to visible? Grover cleveland (talk) 07:51, 3 March 2010 (UTC)

Does anyone understand this?

This phonetic alphabet uses characters, that I'm afraid most are unfamiliar with. I prefer the pronunciation provided by Merriam Webster over the IPA. Can we incorporate both into the articles?

For example: Washington Merriam Webster: "/ˈwȯ-shiŋ-tən" is easier for me to understand than IPA: "/ˈwɒʃɪŋ.tən"

99.73.184.21 (talk) 08:16, 8 March 2010 (UTC)

That depends on your education. For most people, the IPA is easier. But yes, the other can be added; it's just that few editors are going to bother, for the same reason they don't bother with pounds or inches in an article. kwami (talk) 08:28, 8 March 2010 (UTC)

Actually, despite my (colorful? to put it kindly) protests on the main IPA talk page, I would not argue that IPA is less suited to this purpose than some other form of pronunciation guide... wȯ-shiŋ-tən is really no easier to interpret than anything else: either you know what the silly glyphs stand for, or you don't.

Point is that, either way, some people are actually just going to have to look it up. The question I would raise is whether or not it would not also be beneficial to include, inline, the explanation of the symbols found on the IPA-EN key page (this thing, I guess), since a lot of readers are going to be forced to look it up anyway. (Otherwise, why is this page linked?)

I can concede that space and flow may be issues. In response I would suggest that maybe pronunciation doesn't belong inline in the first paragraph of an article to begin with; if there isn't enough room to do it right, there just isn't enough room to do it there at all.

Other than that, my biggest gripe about the guys running this IPA stuff is that the articles where I most want to see a pronunciation guide don't actually include it. :(

J.M. Archer (talk) 17:22, 8 March 2010 (UTC)

Actually, it is done right. That's like saying we shouldn't use kilometers "if we can't do it right" because some people have to look up what a km is. An encyclopedia cannot be responsible for the educational deficiencies of its readership, whether it's the metric system, IPA, standard abbreviations, big words, dates in the Common Era, spelling conventions from across the pond, etc.
We do have a mod to the template that enables hover-over keys, but it will be a while before we can roll it out. But if we don't have the IPA where you'd most like to see it, tell us, and if we can confirm the pronunciation we'll provide it. kwami (talk) 19:13, 8 March 2010 (UTC)
Whether or not a project like this is "done right" often depends on what exactly the goal is. If all you fellows are interested in is shoehorning a given set of characters into a given set of paragraphs, then--by all means--bravo, mission accomplished, this is the end of fighting in Iraq.
However, if the point is to help people to pronounce words they've not heard, then, I would argue...
Bah. You prally get my point by now.
And please don't be so disrespectful as to misrepresent my position. I did not claim that something other than IPA should be used. In fact, above, I stated the opposite. I only argued that it would be nice if information from the key (the pronunciation guide, found here, for the pronunciation guide) could be transposed to the article pages. I had actually been going to suggest the onhover thing you mentioned above, but I figured if anyone cared enough to implement it, someone would already have been working on it.
My point, as I felt I presented it above, was that if the stuff found here at this key is good enough to be found here at this key, it should also be good enough to appear in the articles themselves. Either it's accurate and effective or it isn't.
Oh, and my most recent annoyance was Aristeia, which I can't remember clearly from my classics courses. Tragically, the top zillion search results on Google are mirrors of the Wikipedia page, which does not include what I was looking for. Short of writing my old prof to ask how they pronounced it back when he was young (and Homer was in his sixties), I seem to be up a creek without a paddle. I don't remember any others at the moment, but I'll let you know if I run across any. :)
J.M. Archer (talk) 20:52, 8 March 2010 (UTC)
By the way, I feel that equating other pronunciation guides to non-metric measurements is utterly ridiculous; people develop an instinctive understanding of "miles and "gallons," but the average person (for whom newspaper articles are written at a junior high reading level) is probably just as confused by Webster's pronunciation symbols as by any other form of rocket science. J.M. Archer (talk) 20:56, 8 March 2010 (UTC)
Well, it depends on what you learned at school. A lot of US kids learn s.t. like the Webster's system in elementary school, so they think that's what we should use, not realizing that it's gibberish to most of the rest of the world.
I'll take a look at Aristeia. kwami (talk) 21:04, 8 March 2010 (UTC)
Until I get my OED back, I'm going to assume that it can be predicted by blending aristocracy and oresteia, so that's what I added. kwami (talk) 21:13, 8 March 2010 (UTC)
Yeah; I distinctly remember learning that Webster's gibberish in about second grade. It was only ever useful in second grade. Completely irrelevant to the usability issues that bug me about Wikipedia itself--and my contention with regard to pronunciation is about usability, not which pronunciation guide would be the more useful.
Thank you for looking into that funny Greek word. It's very useful, to the extent one can use it at all, in certain (rare) kinds of discussions--which happen way too often in certain lines of work.
J.M. Archer (talk) 21:21, 8 March 2010 (UTC)
I nearly edit-conflicted with you, Kwami. I went the other direction and tried to resurrect the Ancient Greek pronunciation, which, happily, seems to agree with your English one. This reminds me: we really need WP:IPA for Greek, ideally covering the Classical, Byzantine, and Modern phonologies. That will be quite a chore. — ˈzɪzɨvə (talk) 21:31, 8 March 2010 (UTC)
Ugh! It's not so much that it'd be a chore, but the Greek 'pedians who would edit war over it, claiming that Classical Greek pronunciation is an English conspiracy to deprive Greeks of their heritage. (You do know that Classical Greek was pronounced identically to modern Athenian, don't you?) Well, maybe those editors have given up on WP by now. kwami (talk) 02:25, 9 March 2010 (UTC)


CLOTH group

How are these words to be categorized? With THOUGHT (as in the U.S. and Conservative RP), or with LOT (as in most of the U.K., and I presume, the Southern Hemisphere)? 82.124.231.186 (talk) 14:49, 28 March 2010 (UTC)

Since we don't have a special letter for this vowel, we need two transcriptions. kwami (talk) 07:16, 29 March 2010 (UTC)
Does that go for BATH as well? I note that Bath, Somerset is simply given as /ˈbɑːθ/. Lfh (talk) 10:11, 29 March 2010 (UTC)
Yes. That is the local pronunciation, and should be labeled as such. (Any GA speaker giving it that pronunciation would sound pretentious.) There was a semi-serious proposal to use /aː/ for the BATH vowel, but given the low frequency compared to how often that's erroneously used for /ɑː/, it was thought best to avoid it. kwami (talk) 10:37, 29 March 2010 (UTC)
Ah, I see the explanation now, above the key. Lfh (talk) 10:54, 29 March 2010 (UTC)
Okay, that makes sense. 82.124.231.186 (talk) 14:13, 30 March 2010 (UTC)

Kilogram vs. omission

For me, the o in kilogram is a schwa, whereas the o in omission is not, even if reduced. The note seems to imply that this sound may be absent altogether in some dialects, but doesn't hint at the fact that those that do have the sound may disagree about which words fall into the omission category.

I suspect that a majority of North Americans would agree with me on these two particular words, and the Merriam-Webster would seem to support that. 82.124.231.186 (talk) 14:12, 30 March 2010 (UTC)

What vowel do you have in omission when reduced? Lfh (talk) 16:14, 30 March 2010 (UTC)
I don't know. Probably a centralized version of [o]. I'm willing to believe that the [ɵ] suggested by the table is accurate. My main point is that there's no rounding for me in kilogram. 82.124.231.186 (talk) 21:50, 30 March 2010 (UTC)
OK. Kilogram has been removed now anyway. Lfh (talk) 07:45, 31 March 2010 (UTC)

Pronunciation help

I don't know where to ask this, so I apologize if this isn't the correct place. Someone added a rough pronunciation to the ONEOK Field article (pronounced "wun ok"). I was hoping somebody here could provide a more accurate IPA pronunciation. Its pronounced "one oak" or "won oak" (its an Oklahoma based company). Thanks.—NMajdantalk 17:56, 6 April 2010 (UTC)

Oh dear, still...

It is salutary to note that all those previous contributors who had a degree of academic distinction and knowledge about this subject have long since left, driven to distraction by the repetitive, rambling and plain wrong assumptions/assertions/new interpretations of IPA by Wikipedia's 'editors'. Maybe it's time to wrap this one up. Fortnum (talk) 16:41, 26 February 2010 (UTC)

Do you have anything to contribute, or have you just posted on this page in order to tell everybody not to post on this page? Lfh (talk) 16:53, 26 February 2010 (UTC)
Merely an observation and suggestion, indeed one of the more useful contributions, I feel. I'm certainly not suggesting that people do not post on this page. People are quite at liberty to post comments on this 'talk' page, in much the same way as they are quite welcome to shovel sand from one heap to another, and then back again. Fortnum (talk) 17:16, 26 February 2010 (UTC)
If anything I've said has been in error, rather than simply at variance with your opinion, I'd like to know, as long as you're prepared to put it in polite terms. Lfh (talk) 19:46, 26 February 2010 (UTC)
Let's not feed the troll. kwami (talk) 20:05, 26 February 2010 (UTC)
Lfh, Kwami, this is not a 6th grade dispute. To anyone who has complained that the more intelligent discussion on this IPA topic has been spread around several user talk pages rather than here, I think Fortnum sums it up beautifully succinctly. it's just a question of some people being intransigent and not even prepared to entertain the idea that something somewhere might need some formal, structured discussion, significantly rewriting, or in the worst case scenario, abandoning completely. No one wants that to happen to something they have spent hours working on in good faith, and it takes a heck of a lot of courage to accept without pouting, getting upset, and throwing one's weight and/or authority around. The closing lines of this posting will summarise what I mean. I'm firmly convinced that if everyone were to approach this with as much good will as the the ones you have forced off this discussion, the problem would have been well on the way to being resolved. Some people seem to be forgetting that this is an encyclopedia - we are not writing articles for ourselves, we are producing a work of importance for the global community, and doing it voluntarily. However, I sense that many of the editors tend to think the whole thing is either a joke or a cheap Internet forum. When I suggested doing an RfC on the IPA issue, the idea was met with what looked like bad faith comments by an editor who looks as if he/she deliberately does some specific edits knowing that it will cause friction based on something that is already under discussion. Finally, it looks as if it could be a case for ANI or ARBCOM, but I don't go automatically running to seek protection behind Aunty Ani's skirts when someone, even an admin, 'accidentally' treads on my toes, like many do.
Look, I'm a linguist, but for a hobby I write Wikipedia articles about people and places in Worcestershire, Warwickshire, and Herefordshire, which ironically are among the targets for those possibly disruptive edits, and is the reason why I was dragged into all this in the first place. I'm also a lexicographer, but unlike others in these discussions, although the IPA is a daily tool in my work, it is not my major discipline; I write, among other things, dictionaries (published), and EFL textbooks (published) and the IPA interpretations and the studio recorded DVDs of both the American and English pronunciation used in our books are considered to be a fair representation of what we teach, and teach others to teach; that means RP and not Cockney or Geordie for BE; and General American and not Bronx, SAE or AAVE, for AE. By the same token, I would expect the IPA for American place names in the Wikipedia to be based on General American and not on British RP or some minority dialect of North America; and I would obviously expect the IPA for British place names in the Wikipedia to be based on RP and not on Brummy, Scouse, orNew England.
One thing is absolutely incontrovertibility sure however, (because enough editors have 'complained'), the IPA of British place names in the Wikipedia is being changed away from their standard pronunciation. Merely explaining this away in this MoS does not work if editors blatently refuse to implement it as it was probably intended, (vague, confusing explanation) and just continue to systematically go through British articles and unilaterlly change things according to their own interpretation.
I've tried to make it so clear, in long postings and in shorter messages with bulleted lists for clarity, that what I suggest the English Wikipedia and readers want, is an IPA transliteration that fairly represents the most commonly used form of pronunciation in the respective country and culture. I've tried taking part in the discussions above but I was edged out by craftily composed side issues used as smoke screens to hide from the the WP:NOT that the IPA in this encyclopedia has become (compare the IPA entry in the Encyclopedia Britannica). I stand accused (bordering on WP:CIVIL by an admin) that I don't bother - if that were true I wouldn't keep wasting my breath on this issue, and I wonder just how many really do care. If the Wikipedia IPAists remain intransigent, I won't however bother wasting my time getting more deeply involved, and they can believe they have a consensus.
In a nutshell, I am trying to resolve not one, but two clear issues: First, something that several editors contend is an error in the Wikipedia IPA transliterations, and second, the right for us to defend that opinion without risk of being assumed to be uncaring, unbothered trolls, and suffer possible disruptive editing that we can't complain about because they have been done by admins. It may be of interest to read THIS on this subject by User:Jamesinderbyshire. Too many editors on this encyclopedia project are more concerned with using the anonymity it affords to act completely stupidly, than in producing quality encyclopedic material.
To sum up what I have already stated sooo many times before, here and on other threads and article or project talk pages, and to answer Kwami's various questions once more that he maintains I have evaded, although they were often addressed directly on his and other users' talk pages:
  1. Both WP:IPAEN and International Phonetic Alphabet are in my opinion, in flagrant conflict with WP:NOT.
  2. WP:IPAEN would, in my opinion, be OK if Wikipedia would keep out of telling people how to use it. Don't ditch the article, just ditch the drop-down 'How To', and leave any OR out of it.
  3. International Phonetic Alphabet is, in my opinion, just simply far too exhaustive and goes way beyond the remit of an encyclopedia. If the article's author(s), feel(s) strongly enough about it, they should go into a huddle and write and publish a new book about it.
  4. Wikipedia:Manual of Style (pronunciation) is, in my opinion, inherently flawed because General American and RP are sometimes so fundamentally different that ESOL learners are totally confused, therefore at least both pronunciations should be shown where appropriate - this is in fact recommended, with instructions how to do it, but disruptive editing is being done instead.
  5. Some stats: WP:IPAEN edit count: one main editor 202, next editor 37; IPA: edit count same main editor 289, next editor 171, two of the other major contributors have retired from editing; Wikipedia:Manual of Style (pronunciation): edit count (same main editor) 44, next editor 25.
  6. I am suggesting that everything in the encyclopedia on IPA and pronunciation may not perhaps reflect a very wide diversity of authorship. However, in retrospect I suppose I could equally be accused of squatting all the Rhône wine articles.
  7. I don't have the slightest personal agenda in any of this, whatever some of my comments or those of other commentators might suggest.
  8. In Wikipedia's own official words: The current "pan-dialectal" English convention at Wikipedia:IPA for English is arbitrary/unreferenced/original research, and is therefore invalid.
  9. See THIS comment by User:Jamesinderbyshire.

--Kudpung (talk) 12:29, 27 February 2010 (UTC)

  1. Can you be more specific? Which parts of WP:NOT does WP:IPAEN violate?
  2. WP:NOTHOW specifically states that it does not apply to project namespace (i.e. when it's relevant to editing Wikipedia itself) and "describing to the reader how other people or things use something." In this case, WP:IPAEN instructs the editor how to indicate how words are pronounced and the reader how to understand the transcription. Both perfectly appropriate.
  3. That should probably be brought up at Talk:International Phonetic Alphabet. If it's too long, there are ways of dealing with it without removing information from Wikipedia.
  4. The key word is "sometimes." Those instances are so few that we needn't accomodate for them. The disruptive editing comes when people don't want to edit according to WP:IPAEN because they disagree with it but don't feel that they should try to change it (for whatever reason).
  5. Edit counts should also include those in the talk pages. In my experience, Kwamikagami has put forth a great deal of effort in creating consensus about use of IPA and indicating pronunciation and I don't doubt that his edits reflect the opinions of more than a few editors.
  6. Something else to consider is that silence is often taken for consensus. You may have primary authorship for Rhône wine articles, but it's not clear that your edits are controversial unless someone starts complaining.
  7. Okay
  8. It's already been said that WP:OR doesn't apply to project namespace. I would also like to appeal to WP:IAR. In so doing, I must explain why the rules of WP:OR should be "ignored".
    1. Most of the contrasts that this pronunciation guide reflects, independent of how they're represented (that is, which specific characters are actually used) can be found in many different and widely available dictionaries.
    2. Because they're found in third party sources, the WP:OR concern is one of WP:SYNTH, that is the guide indicates a set of contrasts that incorporates those of two more more different dictionaries but the actual synthesis is unique to (or originating from) Wikipedia and not implied by either.
    3. The synthesis itself does not take a great deal of expertise to understand or accept. If, upon looking up marry & merry in two dictionaries that represent different dialects, you get /ˈmæri/ & /ˈmɛri/ in one and /ˈmɛri/ & /ˈmɛri/ in another, then it's clear even to non-experts that the first dialect makes a contrast that the other does not.
    4. The synthesis also becomes even less significant when we go outside dictionaries and find sources that talk about dialectal variation (this may be something we can look into).
    5. Because these dictionaries choose one dialect over another, WP:NPOV concerns behoove us to actually conduct this synthesis.
    6. WP:NPOV concerns also behoove us to use characters in a way that doesn't favor one dialect.
I believe this is the logic of Wikipedia's diaphonemic representation. —Preceding unsigned comment added by Aeusoes1 (talkcontribs)

Kudpung, you claim you're answering my questions, while once again evading them. Also, as a trained linguist, you should know that several claims you are making are false. Why should "York" have an /r/ in RP? Why should "Warwickshire" not have an /r/? Until you're willing to support the claims you make, you're not engaging in honest discussion, but only standing on a soap box.

"Warwickshire" has a final /r/ in RP. If you disagree, we can discuss what phonemes and allophones are, though I really would expect that you should understand this. (I've tried to engage in this discussion with you several times, but you have adamantly refused.) "York" does not have an /r/ in RP. If you disagree, we can discuss that as well.

If, however, your objection is that we should transcribe names in the local pronunciation, then let's stick to that as a philosophical issue, rather than making up spurious linguistic arguments. So far the consensus has been that we use a diaphonemic transcription and add the local pronunciations where beneficial. I don't know any editors who would object to you adding local pronunciations to place names; the problem is when you delete the broader pronunciation that gives non-locals access to the article.

If you think the WP IPA is American cultural imperialism forcing itself on England (despite the consensus of English editors in crafting the IPA key), then please tell us what is American about it, since everything in it is found in English English.

If your objection is that we shouldn't have a diaphonemic transcription at all, then that is yet a fourth discussion, one that we've had several times, though we can always have it again.

If your objection is that the IPA should be accessible to ESL students, then that is yet a fifth discussion, and one that AFAIK we have not had. It is true that our English IPA conventions are designed for the native or near-native speaker rather than the ESL student. This is in contrast with the IPA for other languages, which does not assume any ability in the language.

If your objection is that we "shouldn't tell people what to do", but rather allow an idiosyncratic transcription for every article, then I don't think compromise is possible. People have a hard enough time following the IPA without it meaning something different each time they see it. We do need guidelines and standards. You're not going to see different transcriptions in the OED depending on who edited that particular word, but rather a consistent system for the entire dictionary. WP should be no different. kwami (talk) 23:28, 27 February 2010 (UTC)

Kwami, I don't see me mentioning York in this context, in this or any other discussion. Is this, with again accusing me of making up spurious arguments, another of your smoke screens?  :"Warwickshire" does not have a final r in RP. The only exception is when it links to a vowel. YOU have adamantly refused to acknowledge my insistences and those of other editors, and all you do is deliberately make disruptive edits which you know will cause polemic that will not take this issue forward to solutions. A point which you may still be possibly missing is that not everyone wanting a quick reference to the reasonably accurate pronunciation of a place name, wants to plod through all your lengthy lessons and contradictions on its use and still be left nonplussed.
Of course I'm standing on a soapbox - in spite of your personal innuendos, you're quite right! Q: Why do people stand on soapboxes? A: To get the listeners to sit up and absorb what is being said, think about it, then hopefully do something collective and positive about it. And that's the whole point isn't it? I'm asking you, the self appointed IPA gurus to listen to what people are telling you, and stop making a walled garden of your Wikipedia passion. Nevertheless, I'm not going to start contributing to your IPA or pronunciation articles. Even if when you open the gate, I could do so with some academic foundation.
There has been no consensus - except the one you admittedly consider to be that of the assent of a silent majority. That's great, so if ten Americans vote for me to become president and 359,999,990 don't go to the poll, I get to move into the White House, right?
Anything that introduces an intrusive r where there shouldn't be one in normal British English, makes it rhotic, which if I remember rightly in one of your other articles in this encyclopedia (I have noted it somewhere in my office), you clearly and unambiguously state that BE is not a rhotic language! To make it rhotic, in my ears, makes it sound American. You, sir, are an American, and have never even been to Britain, so I naturally, quite naturally, put two and two together and accuse you of linguistic hegemony. It's nothing personal, its's a logical conclusion. In fact the only other thing anyone knows about you here, apart from the fact that you have a truly excellent knowledge of the IPA - which doesn't mean you are also an expert on the sociolinguistic implications of its (mis)use - is that you come from the US. You clearly contradict yourself where you have on occasion stated that the Wikipedia's interpretation of the IPA is based on RP, while somewhere else you say that everything about it is American? Please make up your mind which side of the pond you are on, before accusing me anew of not being direct.
Yes, people do have a hard time using the IPA - the American people. In Europe and Asia the IPA is as common as Worcestershire Sauce, as is the European talent for being multilinghual Bit of a paradox really, with the USA being such a multi-ethnic nation.
I have never denied that Kwami is the major player on the Wikipedia in all things IPA & pronunciation - in fact I even pointed it out. I might have primary authorship of the Rhône wine articles (in fact I think actually Tomas_e does), but nobody has complained. In Kwami's case however, people are complaining, or rather not so much complaining but asking rather pointed, embarrassing questions, and getting disruptive edits to their articles in response.
Now, in the last 5 or so postings on this thread or on your user talk pages, I seem to repeating my Hyde Park Corner show, so please do not try to draw me into this thread again by posing more questions which good faith would force me to acknowledge and repeat my answers yet again. If you are all so keen to make any sense out of all this, then It's really up to you guys from the IPA to get this train crash of a discussion back on track. When you finally come up with some suggestions for some changes to your Wikipedia proprietary 'use of IPA' policy, and hold a straw poll on them, don't worry, I'll be watching, I may help with some suggestions for the poll question(s), and then and I'll come back to add one of two words: support or object. Unless of course it turns out to be yet another fiasco like the current talks about BLP, and citation templates..
--Kudpung (talk) 06:36, 28 February 2010 (UTC)
Ah, after months, you finally answered a basic question! Now maybe we can get somewhere! (I'll ignore your misrepresentations of what I've said as unimportant here.)
"Warwickshire" does not have a final r in RP. The only exception is when it links to a vowel. Since you're a trained linguist, I shouldn't have to point out that you've just agreed with me that Warwichshire has a final /r/ in RP. For the others in the audience, I'll explain:
The WP convention for transcribing English in the IPA is based on meaningful speech sounds known as phonemes. Now, let's take another language, as distance sometimes adds clarity. In Spanish and Italian, there is an en sound /n/ rather like English /n/ (we write these sounds between slashes to show they're meaningful speech sounds, not minor detail), but there is no independent eng sound /ŋ/ as in English sing. Therefore a Spanish or Italian speaker may have difficulty pronouncing the English word sing. However, the sound [ŋ] (we write phonetic details which are not independent sounds in brackets) does occur; whenever /n/ occurs before a /k/ or /g/ it is pronounced [ŋ], as in banco [baŋko]. That is, the speech sound /n/ is pronounced [ŋ] before /k/ and /g/, and [n] elsewhere. (More or less.) In linguistic terms, we say that [ŋ] and [n] are allophones of the phoneme /n/. Therefore, if we were to transcribe Spanish or Italian banco phonemically, as we transcribe English words, it would be /banko/ with phoneme /n/.
Now let's go back to Kudpung's example of Warwichshire in RP: before a vowel in ends in a sound [r] (let's not worry for the moment exactly how that [r] is pronounced; that's not important here). However, before a consonant or pausa (pause, end of a sentence, etc.) it ends in [] (that is, null, silence). This is somewhat more abstract that the Spanish & Italian example, but the idea is the same. We have two different pronunciations of the r depending on what follows, just as we have two different pronunciations of Spanish n depending on what follows. These two pronunciations are [r] and silence; they are two allophones of the English phoneme /r/. That is, in RP, English /r/ is pronounced [r] before a vowel, and [] before a consonant or pausa. Thus when we transcribe Warwickshire phonemically, we must write this final /r/.
You might ask, Why write r instead of nothing, since it's often silent? Because if we write /r/, you as an RP speaker will know that it's silent in certain contexts, but if we don't write it, you won't know whether there's an [r] sound before a vowel or not. (Of course, you can always go by the spelling, but English spelling is not always a reliable guide to pronunciation, especially in place names.)
For example, let's take the words bar and baa. One of them is pronounced sometimes [bɑː], sometimes [bɑr], depending on its environment. The other is always pronounced [bɑː]. Now, if we were to write bar /bɑː/, as Kudpung proposes, then you wouldn't be able to predict that it has an [r] sound before a vowel. You'd think it's just like baa. That is, the transcription would be missing information necessary for you to correctly pronounce the word. (I'm assuming of course that these words are unknown to you, just as place names are often unknown to people.) If however we transcribe bar as /bɑr/, then the pronunciation is obvious: You know that in your dialect, words like this (car, far, gar, jar, mar, par, tar, etc) drop their ars when not followed by a vowel. Therefore the transcription /bɑr/ is the correct one, as unlike the other it tells you how to pronounce the word in all circumstances. /bɑr/ is the phonemic transcription; [bɑː] is just the realization of the word in a particular environment, not a complete description of it. For the same reason, Warwickshire should be transcribed with a final /r/ even if we are only concerned with RP. Or in linguistic terms, final r is phonemic in RP. (This is not true for all non-rhotic dialects; some drop their ars entirely. But RP does not.) kwami (talk) 08:23, 28 February 2010 (UTC)
Kudpung has just informed me that he doesn't want to discuss this with me ever again. Since I'm part of the discussion here, I suppose we can now consider this thread closed? Unless perhaps Fortnum has something to add? I think I overreacted in calling him a troll, and I apologized on my talk page, as he does seem to be sincere. Fortnum, do you have anything to add? kwami (talk) 09:54, 28 February 2010 (UTC)
No, I'm happy with the apology, and thank you for it. As for the matter in hand, it does seem to have reached a conclusion, and I don't feel it profitable to add any more to it. Fortnum (talk) 13:47, 28 February 2010 (UTC)
No, "Warwickshire" in RP does not have a final /r/. What is being described above is linking r. AFAIK, a transcription of a single word does not include linking r, e.g. "car" in RP is /kɑː/, not /kɑr/. 92.40.12.2 (talk) 04:31, 22 March 2010 (UTC)
If you have linking /r/, then car is /kar/. If it were /ka:/, that would mean there is no linking /r/. (See allophone.) The question is whether any RP speakers actually have linking /r/; it appears that some do after certain vowels but not others. But regardless, convention on WP is to transcribe a word AFAP so that speakers of all dialects can pronounce it from the transcription. kwami (talk) 10:33, 22 March 2010 (UTC)
"Car" in RP is /kɑː/, not /kɑr/. Linking r (r liaison) is completely irrelevant when transcribing a single word, it definitely is not included in a phonemic transcription. AFAIK, linking r in RP is random; some people sometimes say it after some words and at other times they don't. Please name a dictionary which has RP pronunciations that include linking r. OK I can see that the convention on WP is to transcribe according to rhotic English and that non-rhotic English users just have to ignore the syllable final /r/, but let's not pretend that in RP car is /kɑr/ or Warwickshire has a final /r/. 92.40.211.182 (talk) 21:10, 22 March 2010 (UTC)
"when transcribing a single word" - you still don't know what an allophone is. Please read that article. kwami (talk) 00:57, 23 March 2010 (UTC)
My Oxford English-French-Spanish-German dictionary (which transcribes in RP with a few "US" pronunciations) transcribes car as /kɑː(r)/. It is not the case that linking or intrusive r is "random" it is the case that you don't know what the context is that triggers it. For many speakers, it is triggered by a following word that begins with a vowel. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 04:12, 23 March 2010 (UTC)
"Please name a dictionary which has RP pronunciations that include linking r." - I have a Collins French/English in which the English pronunciations are non-rhotic - and presumably based on RP - but linking R is included. That said, the linking R is indicated not with /r/, but with an asterisk, which is the same symbol it uses to show absence of liaison in the French section. Lfh (talk) 08:18, 23 March 2010 (UTC)
It is not completely "random" but it's not completely determined, either. Robert Plant sings the line "Valhalla I am coming" in "Immigrant Song" without an R on Led Zeppelin III but with an R on How the West Was Won, for example. But what Ƶ§œš¹ described is the environment where it's most likely to be pronounced. ― A._di_M. (formerly Army1987) 12:55, 23 March 2010 (UTC)
Kwami, what you write about the phonemic status of linking /r/ is not correct. Your account is fatally confused about the meaning of the term phoneme, and the difference between surface and underlying forms (or phonemic and morphophonemic representations; label according to theoretical taste). The whole discussion on this page seems to be driven by your non-standard interpretation of what a phonemic representation should consist of, and by a reluctance to engage with the standard academic consensus on this matter. C0pernicus (talk) 12:29, 20 April 2010 (UTC)
I don't see what you're talking about. What has Kwami gotten wrong? — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 15:20, 20 April 2010 (UTC)
I do - what's the problem?--Kudpung (talk) 07:20, 30 April 2010 (UTC)
The only thing I think Kwami has gotten wrong about linking-r is that it's the justification for our transcribing with r's in the syllable coda. We do it to represent the phonemic contrast in rhotic dialects, not to represent linking-r. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 20:43, 30 April 2010 (UTC)
Yes, he's certainly got something very wrong there, but that's not all.--Kudpung (talk) 19:30, 3 May 2010 (UTC)

Does a list of English words written in IPA exist?

Is there a list of basic English words written in IPA? The reason I ask is that I was just on the page for word I know how to pronounce and I saw the IPA. This made me realize that it might be easier to pick up IPA instead of looking at bare symbols and finding out which phoneme they represent, but instead to chunk a word, which is a string of phonemes, that I already know how to pronounce and I can store the IPA in my head. Does that make sense? THanks, --Rajah (talk) 05:56, 3 April 2010 (UTC)

Yeah, actually, it does make sense. Maybe we could transcribe the example words on this list into IPA? kwami (talk) 14:10, 3 April 2010 (UTC)
yes, that would be a good idea. Also, the new Template:IPAc-en template is great, we need a bot to go through and replace all Template:pron-en templates with that one. and also break the word up into phonemes. e.g. : {{pron-en|ˈsɜrf}} becomes /ˈsɜːrf/ (mouseover both to see the difference.) . The mouseover on individual phonemes is so awesome! --Rajah (talk) 04:14, 7 April 2010 (UTC)
That list is a good idea, but where to put it? The page is getting quite cluttered. Maybe a separate page, or a hover-over? Lfh (talk) 11:16, 7 April 2010 (UTC)
Most quality dictionaries (at least the ones from British publishers) include IPA transcriptions for their headwords. The Wikipedia on the other hand, is an encyclopedia, and is not intended to be an exhausitve, authoritative work on any subject. neither is istsupposed to be a handbook for use of the IPA, although the Wikipedia IPA and pronunciation articles may leave a reader with that impression. Most important before a bot does any mass changes, is to be absolutely sure that the Wikipedia is correct in its implementation of the IPA - something which however, is still very much open to debate.--Kudpung (talk) 07:27, 30 April 2010 (UTC)

rhotic diacritic?

We have ɝː, ɚ in some place names which are locally non-rhotic. Is this something that would be worth extending to other vowels, say ɑ˞ ? The /r/ seems to be the thing people most object to. (The /j/ after alveolars is perhaps just as objectionable, but much less common.) Of course, it might be a little silly to worry about the /r/ and not /h/ or /j/, but we could also mark them too, perhaps as /ˈʰɑ˞tfɚd/ Hartford, /ˌnʲuː ˈmɛksɨkoʊ/ New Mexico. This would only be for place and personal names. IMO, it would be a choice between introducing more IPA characters for people to have to learn, and having people object that "that's not how it's pronounced" (locally, that is). In the key we could gloss them "not pronounced in the local dialect". That would also reduced the number of redundant transcriptions. Or is our current approach of generic vs. local English good enough? kwami (talk) 21:07, 20 April 2010 (UTC)

It's also problematic in that it's not IPA. Superscript <j> indicates palatalization, not an optional palatal glide; superscript <h> indicates aspiration, not an optional glottal fricative. Similarly, [ɑ˞] is an [ɑ] with retroflexion, not one with a following rhotic. We've got enough trouble getting people unfamiliar with the IPA to acquire our system, we don't need to also make it difficult for people who are familiar with the IPA. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 00:11, 21 April 2010 (UTC)
Indeed someone recently slipped in /ɝː, ɚ/ again. In my opinion we are trying to define a system for transcription. We should not make it confusing by allowing all more and more alternatives. Writing the /r/ is good enough and allows to add a simple line that it is optional in many dialects. Let's keep it simple. −Woodstone (talk) 01:22, 21 April 2010 (UTC)
I think a line should be drawn at the symbols /ɝː, ɚ/, because sometimes the /r/ is not appropriate, particularly in names. --58.165.2.250 (talk) 12:40, 21 April 2010 (UTC)
I'm not sure what you mean by "particularly in names." Do you mean orthographic r is less likely to be present phonemically in proper nouns or do you mean that [ɚ] is more likely than [ər]. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 16:30, 21 April 2010 (UTC)
What I mean is that place names that may include a rhotic vowel in its pronunciation, but have a majority of speakers that are non-rhotic (ie, Australia, most of England, NZ, etc), it would therefore be inappropriate to use an <r>, so the use of the r-coloured vowels /ɝː, ɚ/ are an appropriate compromise between /ɜː, ə/ and /ɜr, ər/, because there is a need for rhotic vowels to be included, even when native speakers don't use them. The same can also be applied to the pronunciation of a non-rhotic speaker's name. --58.164.107.103 (talk) 08:25, 23 April 2010 (UTC)
I fail to understand why writing r-coloured vowels /ɝː, ɚ/ would be better than adding an r as in /ɜr, ər/ for a proper name of someone/thing from a non-rhotic area. What could be simpler than the general rule: "for non-rhotic dialects omit any r following a vowel in the same syllable". Not having these duplicate representations with special symbols makes applying IPA more straightforward, without losing accuracy or generality. −Woodstone (talk) 10:04, 23 April 2010 (UTC)
Mainly because a minority of editors gets really upset by transcribing the /r/, even though they don't mind /h/ and other differences. Take a look at Kudpung's subpage on the debate for an extreme example. They don't object so much to transcribing them as rhotic vowels. Okay, that's not an academic reason, more of a political one. kwami (talk) 10:09, 23 April 2010 (UTC)

Let's just be absolutely clear on this before anyone decides for me whether my argument is academic or political (which it is neither - it is a practical one and based on the real use):

There is growing concern that the IPA spellings of English place names, particularly those of the shire counties, and other names that end in an R that is generally not pronounced, are either not correct or do not represent the way in which the majority of British people pronounce those names. IPA Wikipedians have commented that the board-wide system they have designed and are in the process of implementing is phonemic and not phonetic, and that the r must be shown due to the fact that some speakers may introduce a linking r when the next word begins with a vowel.

Furthermore, I believe one of the confusions throughout this entire debate is the use of the word local, which may or may not have a slightly different connotation on different continents of the English speaking world, and may in fact be one of the root causes of so much conflict and misunderstanding in this entire issue. I also believe that the majority of readers a re neither interested in, or do not understand, the highly technical linguistic explanations they have been given.
Before any discussion takes place, it needs to established what is meant by local, regional, national, and global, as these words themselves appear to be being interpreted differently.[1]

Most likely in the United Kingdom they would mean:

  • local = in the city, in the immediate area surrounding the city, and possibly the rest of the county.
  • regional = the rest of a county that covers a particularly large geographic area, and its neighbouring counties.
  • national = the country where the language is spoken. In this case, England, where a neutral RP is more commonplace and/or widespread than say, for example, Scotland and Wales where their national accents a re the accepted educated accents of the majority.
  • global = worldwide, or in the case of this issue, the regions of the world where whre the two main versions (AE & BE) predominate, such as for example, The Philippines where AE predominates, and Southeast Asia and the Indian subcontinent where BE predominates.

We also need some qualification regarding the statement: ...a minority of editors gets really upset.... I don't think anyone gets upset by the transcription. To put it correctly, firstly the number of commentators equals or surpasses the number of major contributors to the various IPA articles, keys, and guidelines - all of which would appear to be primarily the work of one major editor; and secondly, people are not upset (yet) by the use, but are simply trying to explain that the prescriptive use practiced by the IPA author(s) may be red for a rethink. And thirdly, and most importantly,we must differetiate between Wikipedia editors (aka Wikipedians), and visitors to the encyclopeia who xanted to look something up, and then signed on to be able to suggest that the said prescriptions do not match the view they would expect. This is not insignificant, and should not be brushed of with a flick of the wrist.
--Kudpung (talk) 15:26, 24 April 2010 (UTC)

If I might, I'd like to offer one correction to "the r must be shown due to the fact that some speakers may introduce a linking r when the next word begins with a vowel": if that's the argument, then it's likely to be misinterpreted as many dialects produce an intrusive r that's similar to linking r; if we were accomodating for intrusive r (which is a lot more common) we would have to put a final r in transcriptions for nigeria and draw. The more agreeable (and consistent) justification for putting coda r's is that this marks a phonemic contrast that speakers of rhotic dialects still make. Just as with many of the vowel system we implement, there will be dialects that don't make such a contrast (i.e. most non-rhotic dialects) but we're trying to be inclusively accomodating. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 18:04, 24 April 2010 (UTC)
Thank you, Kudpung, for bringing this to a centralized location for discussion. You've heard everything I have to say, so I'll chime out now. kwami (talk) 18:49, 24 April 2010 (UTC)
Here's an example of an English name that shouldn't be transcribed with an /r/: Matthew Le Tissier. AFAIK, rhotic speakers wouldn't have an /r/ either. — kwami (talk) 00:54, 30 April 2010 (UTC)
The Tissier example is off-topic - the name is French.--Kudpung (talk) 07:29, 30 April 2010 (UTC)
It's not off-topic at all. The given pronunciation is how even English speakers of even rhotic dialects pronounce it. Who cares of the name comes from French? — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 07:51, 30 April 2010 (UTC)
Agreed. Matt Le Tissier himself is not French, and there are millions more English-speakers with names of French or other foreign origin. Are they all off-topic? And there are assimilated French words as well, e.g. sommelier, which even rhotic speakers pronounce without /r/. Lfh (talk) 08:12, 30 April 2010 (UTC)
...which even rhotic speakers pronounce without /r/. Gotta laugh this time ;) That's the best example yet of you IPA folks not reading what I wrote, and thinking that you have contradicted me, you have actually agreed with me! --Kudpung (talk) 17:14, 3 May 2010 (UTC)
Marvellous, it's always nice to agree. Lfh (talk) 18:31, 3 May 2010 (UTC)
Be nice if it were intentional, but it just goes to demonstrate yet again (sigh) that the IPA and pronunciation articles are dominated by a bunch of semi-intelectual clowns pretending to be linguists. What have they been smoking this time?--Kudpung (talk) 18:50, 3 May 2010 (UTC)
As I understand it, Kwami was pointing out that the current key distinguishes between final orthographic r's that are pronounced by no-one at all, and those that are pronounced by rhotic speakers only. And I was agreeing that "Le Tissier" is a valid illustration of this - i.e. not "off-topic". Lfh (talk) 19:14, 3 May 2010 (UTC)
Agreed. I don't get how "The Tissier example is off-topic - the name is French" could be construed to mean anything like what I and others have said in response. I'm getting the feeling that either Kudpung has himself not read what others wrote or that we are using the same words to talk about different things. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 19:40, 3 May 2010 (UTC)
Precisely, and there's the practical joke played on Kudpung that's backfired on you all by trying to be clever and re-invent the IPA for use in the Wikipedia to support a silly claim. So having said that, you can also now go ahead and agree that there are a lot more final r's in British and rest-of-the-world English that are never pronounced, and that it's time for the Wikipedia readers to get some accurate pronunciation guides, instead of a hypothetical construct that tries to cover too many, widely differing accents in one go, among which there are confusing examples that simply just don't exist.--Kudpung (talk) 19:43, 3 May 2010 (UTC)
In non-rhotic dialects there are many final (orthographic) r's that, essentially, are not pronounced. As intrusive r occurs even without an orthographic r in words like Nigeria and draw, we could even see this intrusive r as avoiding a certain form of hiatus. This phenomenon occurs quite frequently in utterances by speakers that exhibit them. Would you agree with this? — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 19:57, 3 May 2010 (UTC)
If Kudpung honestly thinks that our IPA conventions were drawn up to play a practical joke on him, I'm doubtful that rational debate will be of much use. If he doesn't think that, but says it anyway, I have the same doubts. — kwami (talk) 23:01, 3 May 2010 (UTC)
Kwami, the rational debate will have no success unless people read what other people post. Your comment above is totally off topic, wrong, and another dig. Are you now completely unable to follow the comments made by your henchmen too? If you are unable to comment sensibly in a mature manner why bother? The other thing is, why should I continue to respect Wikipedia civility guidelines, when you , as an admin to boot, don't give a hoot about them and still continue to harrass me?--Kudpung (talk) 00:48, 4 May 2010 (UTC)
When have you ever respected WP civility guidelines, except to tell other people to follow them? And how is calling you out on your nonsense "off topic"? You make it the topic. You come here and call us all idiots, claim other editors are my henchmen (I assume they're my henchmen and not the other way around because you first came into conflict with me), and accuse us of playing a great practical joke, simply because you have so far failed to present your case coherently. That's not a swipe at you, but an explanation of my frustration and why I've "twisted" your words. (I find it difficult to paraphrase someone or respond to their arguments intelligently when I can't follow those arguments.) You've made things so unpleasant around here that we now have other editors throwing up their arms in disgust and saying they've had enough. Why should anyone take you seriously? If you want the English IPA conventions to follow multiple national standards, fine, make a straightforward case for that. It's a simple enough case to make. — kwami (talk) 01:09, 4 May 2010 (UTC)
Kwami, Check the diffs to see at whom the editors are 'throwing their arms up at in disgust', and who does not observe policy, and why editors are afraid to defend themselves in this unpronounced 'r' business.--Kudpung (talk) 16:56, 8 May 2010 (UTC)

Please help with Sarah Jarosz

I noticed that the articles for both Sarah Jarosz, and for Philip Lynott (the second of which pronounced his surname as LYE-not, both need IPA help for their names. Jarosz already has a "sounded out" name next to her spelled name that someone else left behind. Could anyone take a look at the two of these musicians and see if you can improve with an IPA rendition of their last names? Thanks. --Leahtwosaints (talk) 17:39, 28 April 2010 (UTC)

Done. — kwami (talk) 18:29, 28 April 2010 (UTC)
So does that mean Jape are wrong (or joking) in the song "Phil Lynott", where they pronounce it "LINN-ott"? Lfh (talk) 18:46, 28 April 2010 (UTC)
Me, I have no idea. I'm just an IPA drudge. — kwami (talk) 23:57, 28 April 2010 (UTC)

Thank you greatly. You folks with IPA are a real blessing. Both sounded out sound correct. :) --Leahtwosaints (talk) 23:07, 30 April 2010 (UTC)

maintenance w AWB

Some of you may have noticed that I've been regularizing & maintaining the IPA with AWB. I've worked out some regex expressions that do a pretty good job - turns out there are a fair number of transcriptions with Cyrillic rather than Latin < a >, for example, which would make searches bafflingly difficult. Since it took a lot of head-scratching at bugging other editors to get the expressions figured out, I thought I'd post them on the talk of {{IPA-en}} and {{pron-en}}, in case anybody's interested in fixing the IPA for this or other languages for themselves. — kwami (talk) 00:01, 29 April 2010 (UTC)

Syllable split

Recently a lot of syllable splits have been indicated in the IPA renderings. I wonder if the way to do that has been discussed and what the scientific basis for it would be. One of the remarks pertaining to it in the article is in a footnote, converted to a table here. I wonder if these should be analysed as indicated in the third column:

one syllable two syllables alternative
our /ˈaʊər/ plougher /ˈplaʊ.ər/ /ˈpla.wər/
hire /ˈhaɪər/ higher /ˈhaɪ.ər/ /ˈha.jər/
loir /ˈlɔɪər/ employer /ɨmˈplɔɪ.ər/ /ɨmˈplɔ.jər/
mare /ˈmɛər/ mayor /ˈmeɪ.ər/ /ˈme.jər/

Woodstone (talk) 07:57, 30 April 2010 (UTC)

This is an example of a set of contrasts that not every dictionary encodes but that we accomodate for in our diaphonemic transcription. Help:IPA conventions for English is supposed to elucidate which dictionaries encode which contrasts, though it's not quite finished.
Your alternative of using semivowels in the onset to contrast the two may appropriate, but I notice that this then makes /a/ into a monophthong of English when it isn't as such. We would thus have to convert that into another low vowel phoneme of English (either /æ/, /ʌ/, or /ɑ/), the choice of which may depend on dialect. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 08:44, 30 April 2010 (UTC)
In the examples, the diphthongation of the vowel may be considered induced as a glide to the following semivowel. But my remark was meant in a bit more general sense behind all these (and more) cases. Many of the syllable splits I see being introduced en masse, seem more based on orthographic conventions of line breaking, than on a phonological principle. Line breaking would create break-ing, whereas in speech it is more like /bre.kɪŋ/. −Woodstone (talk) 15:24, 30 April 2010 (UTC)
I see what you're saying. I believe line-breaking is based on morphology, which it seems the syllable splits in question may also (arguably) be based on. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 19:57, 30 April 2010 (UTC)

BTW, I've been adding syllable breaks in names like Nuxhall /ˈnʌks.hɔːl/, so people don't misread the sh as /ʃ/. Likewise /t.h, n.k, n.ɡ/. I've also been more consistent with sequential vowels, adding a dot between all except after /ː/ and in the common ending /iə/. I know when I review other people's transcriptions, I'm never sure what "ng" is supposed to be, so this should hopefully clarify it. — kwami (talk) 18:03, 30 April 2010 (UTC)

Shouldn't /n.k, n.ɡ/ just about always be /ŋ.k, ŋ.ɡ/? In which case, do we need the syllable separator? On another note, I'm not sure how I feel about indicating one syllable break in a (trisyllabic-plus) word without indicating all of them. Not doing so seems to imply that the unbroken sequence is a single syllable. I know when I'm adding IPA for French, and feel a dot is necessary, I go ahead and mark all the breaks—which is much easier for French than for English (damn ambisyllabicity...) — ˈzɪzɨvə (talk) 21:37, 30 April 2010 (UTC)
Yes, that's the problem, they almost always are. So when they're /nk, nɡ/, as it Vancouver and Feingold, it's hard to tell whether they're really /n/, or if the transcriber just made a typo. (Okay, I guess I'm using the dot as more of a morpheme break there, but I think it gets the idea across.)
Personally, I'd go with Wells's ideas on syllabification, but I've had nasty fights with editors who insist that English is phonemically V.CV, never VC.V. And it's a hard call to make in many cases, even for Wells! (I'm not convinced myself that "mattress" is really /matr.ess/.) — kwami (talk) 21:59, 30 April 2010 (UTC)
Is Wells's idea of syllabification maximization of the onset? — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 22:15, 30 April 2010 (UTC)
I'm guessing from that example that it's maximization of the stressed syllable, though /tr/ is not an acceptable coda cluster otherwise, so I don't know...
Actually, I think I may have read Kwami's comment wrong, but in any case, my intuition is that English stressed syllables "pull" consonants towards them, but still have the same phonotactic restraints (more or less) as monosyllables. Thus marble /ˈmɑrb.əl/, but mattress /ˈmæt.rɨs/. — ˈzɪzɨvə (talk) 23:07, 30 April 2010 (UTC)

I should be done will all AWB-parsed transclusions of IPA-en and pron-en in a bit. If anyone sees non-canonical IPA that I've missed, please let me know and I'll try coding it in to my next pass with AWB. (On my list so far: syllable breaks and tense vowels before engma and ar.) — kwami (talk) 21:21, 30 April 2010 (UTC)

Oh, and here's an illustration of why we need to mark stress on monosyllables: Stow cum Quy. — kwami (talk) 22:01, 30 April 2010 (UTC)

Morphology is a conceivable base, but I think trying to sing the word on long notes brings out the phonological split better. That leads often to CV.C, or indeed a maximised onset. I could not possibly sing /mætr.əs/, but I must confess that /mæː.trəs/ isn't very musical either. The examples above with n.g show ambisyllabicity best, try /ˈsɪŋ.ŋɪŋ.ɪn.ðə.ˌreɪn/. −Woodstone (talk) 23:22, 30 April 2010 (UTC)
Wells presented it a lot as Xyzzyva said. I'd always felt that a following consonant became the coda after a short vowel, but not after a long one, but Wells makes a convincing case that it's a coda after either. Clusters as codas as well, and they're not ambisyllabic but simply codas, as in self-ish. The matr.ess thing comes about in an effort at parsimony of phon rules: tr behaves as an affricate, affricates become codas after stressed vowels like any other C, therefore tr is the coda. Well, a bit more sophisticated than that, but at that point I balked.
Ah, here we go: [1] It's a fun read. — kwami (talk) 08:07, 1 May 2010 (UTC)
So, going back to Woodstone's original post, it seems that if we're maximizing the stressed syllable then even if we interpret employer as /ɨmˈplɔjər/ with a semivowel, this system proposed by Wells would put the semivowel in the coda of the stressed syllable. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 17:13, 2 May 2010 (UTC)

Foreign a

I'd like to point out a problem with the table. For many Canadian speakers, baht and calm are not pronounced with the same vowel. Namely, calm is something like [kɒːm] (same vowel as cot and caught), while baht is [bɑːt], with a special vowel used in foreign words. (These words include taco, pasta, Mazda, etc., for those Canadians who don't pronounce these words with /æ/. However, while I can easily imagine taco, pasta, and Mazda with /æ/, this seems unimaginable for a word which has a graphical ah.) This vowel is in fact very close to British and American realizations of baht. It is the Canadian realization of calm that is different. I know that the following paper (to which I don't have access) talks about foreign a in Canadian English: The emergence of a new phoneme: Foreign (a) in Canadian English. 82.124.97.111 (talk) 15:07, 2 May 2010 (UTC)

Not an answer to your question, but I seem to have access to that paper, in case there is anything in it that you're wondering about (within copyright of course). Lfh (talk) 13:59, 3 May 2010 (UTC)

Thanks. I've managed to access the PDF. Don't know what the problem was.

It seems I may have been a bit off in my phonetic description of the vowels above.

In any case, for the purposes of the discussion here, the important point is that for many Canadians, there are words which belong neither to the PALM category, nor to the TRAP category, but to an intermediate one. The best example of this given in the article is the word plaza, pronounced by 38% of Canadians with the TRAP vowel, 18% with the PALM vowel, and 44% with something between the two, low-central for most of them. (According to the article, the PALM/LOT vowel of Americans is already low-central, whereas the PALM/LOT/THOUGHT vowel of Canadians is low-back.) An intermediate pronunciation was also produced in 41% of cases in each of the words lava and façade.

For 9 of the 20 words with "foreign a" tested, an intermediate vowel was more common among Canadians than was the PALM vowel.

This kind of pronunciation was not rare among Americans either, particularly for Colorado (27%) and panorama (23%). For panorama, it was more common than the PALM vowel (5%).

The article suggests that there may be a phonemic contrast between laggard/lager/logger, rack/Iraq(i)/rocky, stab/Saab/sob, sad/façade/sod, dally/Dalí/ dolly, and suggests the notation /ah2/ or /ahf/ (with f for foreign) for the intermediate phoneme.

Naturally, Britons already make these distinctions, but that is because they haven't merged the PALM and LOT vowels.

For the current IPA guide, the inclusion of baht and palm under the same heading is problematic for the likely 40 or 50 percent of Canadians who would pronounce them differently, and a smaller percentage of Americans. At a minimum, a note about the problem should be inserted.

But I would go further and propose using the symbol /aː/ or /äː/ to transcribe words like baht. Speakers who pronounce baht and palm with the same vowel sound could simply ignore the distinction. The choice of /aː/ (understood to refer to a low-central, rather than low-front, vowel) would have the advantage that it is phonetically accurate for both Americans and Canadians. 82.124.103.148 (talk) 06:47, 5 May 2010 (UTC)

newbie in edit war

An editor is giving contradictory pronunciations for Chichester, Pennsylvania and edit warring over any attempt to resolve them, including deleting my comments on the talk page. He's a newbie, so I don't want to have him blocked, so maybe s.o. here can speak to him? — kwami (talk) 03:10, 8 May 2010 (UTC)

/ær/ vs. /ɛr/ in names of living persons

In the article on Caroline Dhavernas, the transcription of the pronunciation of her first name was recently changed from /ʼkɛrəlɨn/ to /ʼkærəlɨn/.

Now, I understand (and, despite some earlier misgivings, am willing to accept) the general premise that /ær/ is intended to be a dialect-neutral rendering that should be equally meaningful to readers who have, or have not, undergone the "marry-merry merger" in their own speech. My concern in this case is that Caroline Dhavernas, herself, pronounces her own name as /ʼkɛrəlɨn/ (based on her self-introduction in the voice-over commentary for the pilot episode on the Wonderfalls DVD set). And I confess I feel pretty strongly that, when describing the pronunciation of a person's name, preference should definitely be given (where possible) to the way the person him/herself pronounces his/her own name — even if this pronunciation happens to reflect a specific regional dialect or is otherwise "nonstandard".

I know the /ær/ pronunciation is common in Montreal, where Caroline is from (i.e., the "marry-merry merger" is not characteristic of the English spoken in and around Montreal) — and, frankly, I was surprised when I listened to the way she pronounced her name in the DVD commentary, because I was mostly expecting her to say /ʼkærəlɨn/ — but she unquestionably did say /ʼkɛrəlɨn/, and no matter what people working on the IPA for English project are doing with this particular phoneme, I'm having trouble feeling good about consciously and explicitly transcribing someone's name in a manner that contradicts how they say their own name.

If necessary (and if it would make any difference to the argument here), I'm willing to try putting up an OGG or WAV file of Caroline Dhavernas saying "Caroline Dhavernas", so the rest of you can listen. It happens that my own speech has not fully incorporated the marry-merry merger, so I am very conscious of the difference and am pretty certain of what I heard.

Comments? Richwales (talk) 00:01, 13 May 2010 (UTC)

I agree with you, assuming of course she has the distinction. I had assumed this was due to the merger. If she makes the distinction but still says /ɛr/ then yes, IMO that's how we should transcribe it. It might be a good idea to add an inline comment to the coding ( < ! --- > ) mentioning this, so it isn't reverted later by some doofus like me who doesn't know any better.
BTW, what would you do with "Marylander (sometimes {{pron-en|ˈmærləndər}} MARR-lənd-ər)"? Since /ær/ does not appear before consonants, I suppose it would have to be /ɛər/, but I hesitate to change it. — kwami (talk) 01:00, 13 May 2010 (UTC)
Dhavernas spoke with a General American (US west coast) accent when talking as herself in the special features on the DVD. Assuming she had an anglophone Quebec accent as a child (essentially the same as the southern Ontario accent), this would suggest she completely "lost" her native accent and intentionally acquired General American speech — a dialect which has undergone the Mary-marry-merry merger (i.e., she doesn't make the distinction any more at all, even though it's likely she did as a child). I'm still not sure I go along with your reasoning that her pronunciation matters only if she makes the distinction with other words. As for "Marylander", I've never thought about that word before; I'm inclined to use /ɛər/, but I've never had a distinction between "merry" and "Mary" in my own speech, and I'd prefer to hear opinions from people who do have a full threefold "Mary-marry-merry" distinction. Richwales (talk) 02:13, 13 May 2010 (UTC)
I think either approach would be fine, but that we should notify the reader which one we're using. As it is, the key specifies that the transcription is diaphonemic, so IMO an unmarked transcription should be diaphonemic. For a local pronunciation (in place names) or a personal one (in personal names), we should alert the reader that that is what we have done, and not randomly use sometimes one and sometimes the other and leave it to the perhaps puzzled reader to decipher which one it is this time. — kwami (talk) 04:40, 13 May 2010 (UTC)
I've added a phonemic transcription in French. I'm dissatisfied with the French template, because it uses brackets, making the narrower [kaRɔlɪn davɛRnɑːs] more appropriate.
As for the English, I don't see the point of using /ɛr/. Either she never makes the marry-merry distinction, or she is pretending not to. It seems unlikely that she would normally make the distinction, but not for her own name. I think it should remain /ær/.
Also, I think replacing [ær] with [ɛr] deliberately seems unlikely. "General American" is broad, and a person who uses [ær] doesn't stand out by any means. 82.124.101.170 (talk) 08:24, 13 May 2010 (UTC)
Anon, do you think you could go to Wikipedia talk:IPA for French and make a case for adding the length marks? We've gotten flack for using a transcription system that is largely (but not completely) phonemic while using brackets but I don't know of very much French allophony.
Richwales, the logic goes something like this: In dialects that distinguish between /ær/ and /ɛr/, Caroline (a common name) is pronounced with the former. If Ms. Dhavernas doesn't make the distinction between the two, then her pronunciation with [ɛr] is not different from the usual pronunciation. If she does make a distinction between the two, then the pronunciation with [ɛr] is not reflective of the merry-marry merger. If it's not reflective of this merger, this would mean that any speaker who distinguishes /ær/ from /ɛr/ shouldn't pronounce her name in the usual fashion (that is, as the former) and we should transcribe it with /ɛr/.
The best thing to do to figure out if she makes that distinction is to check her pronunciation for the distinction in the same conversation that we hear her pronounce her name with [ɛr]. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 16:48, 13 May 2010 (UTC)

remove "a mission" as a sample for ə

I suggest that "a mission" be dropped as a sample for the pronunciation of ə. The pronunciation of the indefinite article "a" is rather variable and context dependent. The other two samples for ə, Rosa’s and comma, don't have that problem and should be sufficient. —Preceding unsigned comment added by 24.6.75.255 (talk) 23:35, 13 May 2010 (UTC)

It's meant to be a minimal set, to illustrate the reduced vowels. — kwami (talk) 00:45, 14 May 2010 (UTC)
I agree that the variably-pronounced indefinite article a is context dependent, but a mission gives a rather clear context. Outside of a stilted reading voice that produces "ey mission", how else would it be pronounced? — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 04:43, 14 May 2010 (UTC)

Template conversion proof of concept

If you have a moment, please feel free to take a look at User:Grover cleveland/IPATest. It's a little slow, unfortuately, but it seems to work mostly. Let me know whether this is something you think is worth doing some more work on. Cheers. Grover cleveland (talk) 05:15, 20 May 2010 (UTC)

Here's an sample of what the template does: given only the IPA-en format as input it now generates both RP and GenAm output. For example:
  • nɪər (diaphonemic)
The idea here is that perhaps we could get out of the diaphonemic squabble by generating both RP and GenAm output from the current diaphonemic input. The end result would be similar to a pronouncing dictionary like Wells's LPD. Many more examples on the link given above. Please -- any comments, positive or negative, will be welcomed! Demo is at User:Grover cleveland/IPATest. Cheers. Grover cleveland (talk) 06:09, 21 May 2010 (UTC)
I had a look and indeed it's intolerably slow. But aside from that, it shows that in many cases there is no difference. This should at least be filtered out to avoid duplicate transcription. In most other cases it takes a while to actually notice the difference. In all, I think the added value would be nil or negligible and create additional clutter. −Woodstone (talk) 08:25, 21 May 2010 (UTC)
If Kudpung et al's RfC comes about, this might be mentioned so that the community can discuss this as a feasable possibility. I suspect that most users wouldn't even care about the difference but the only way to really know is for an RfC, right? — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 16:05, 21 May 2010 (UTC)
Hi Woodstone: thanks very much for your input. Here are responses to some of the issues you raise:
  • Slowness yes it is indeed very slow at the moment, but I can work on fixing this. I have already had some ideas on how to improve the speed. Working with string manipulation using the tools Wikimedia makes available is definitely challenging -- rather like climbing a tree with one hand tied behind your back :) But I have had very little experience in working with these tools until now and I have already had some ideas on how to make things considerably faster.
  • Duplicates: definitely they can be filtered out. That will be easy (but see the next item).
  • Clutter: As some have discussed above, it might be possible to use user preferences to show only the accent (RP or GA) that the user selects in their user preferences.. If the user doesn't select a preference, or for IP users, we could either default to showing the current IPA-en format, or show both RP or GA, or do something else (not sure what right now).
  • Potential: If I get skilful enough in the templates, I have dreams of allowing the user to select respelling format instead of RP or GA, which should theoretically be able to be generated from the current input. Because of the difficulty of working with the Wikipedia String tools, however, I am not sure how practical this will be.
  • Motivation: While I am personally very happy with the current diaphonemic IPA transcription format, and I commend everyone who has worked on developing it, it is evident that some users are not so happy. The idea behind this little project is to allow users to see whatever format they prefer, while also allowing editors to continue to add pronunciations in the diaphonemic format. There are also some (don't know how many) users who object that the current format constitutes Original Research. Without taking any position on the merits of this objection, I think that it would be good to allow such users to choose to see either RP or GA (or both) formats, so that this objection would then presumably go away.
Cheers. Grover cleveland (talk) 19:00, 21 May 2010 (UTC)
Hi, Grover. I appreciate your attempts and realised the speed might be improved. But a while ago a similar debate raged about UK/US/International date formats. There it became clear that the vast majority of readers is not logged in, and so cannot have preferences. Guessing their preference is fraught with difficulties. Even if you might be able to guess their location, that does not imply a preference, especially outside places where English is the native language. So the only option would be to show both (all?) dialects or just the diaphonemic form. So either there would be (a lot of) clutter or no progress. The same effect crushes your dream of automated "other" representations. −Woodstone (talk) 20:30, 21 May 2010 (UTC)
Fair enough. Anyway, I have now changed the template so that the template converts to RP format only if the user has set "en-GB" as his/her "Internationalisation/language" user preference: otherwise it leaves it in diaphonemic format. Unfortunately Wikipedia does not provide a specific "en-US" language setting, which otherwise might have been appropriate to use for GenAm.
I appreciate your point about most users not being logged in. However, I suspect most of those who raise a ruckus about the current scheme are logged in users. Wouldn't it be useful if, when a non-rhotic user starts complaining about the IPA format being "American" or whatever, we could reply to him/her: "Just set your user preference to en-GB and you'll see everything in RP?" Grover cleveland (talk) 21:10, 21 May 2010 (UTC)
Same reasoning goes as for the date debate. Implementing this template for logged-in users, would hide to them what ordinary readers would see. However, entering pronunciations correctly is most likely dependent on logged-in users, which would be hampered if they do not see the same thing as most readers. −Woodstone (talk) 21:51, 21 May 2010 (UTC)
This idea seems worthy of further investigation to me. Martin Hogbin (talk) 21:53, 21 May 2010 (UTC)

A minor example of the problems hard-formatting dialectical differences can do: at germen, there is both a generic term and an English place name. Since the transcription is marked as phonemic, if we were to use /r/ in one but not the other, we would claim that the pronunciations differ, which would be false: any individual speaker would pronounce the two the same. — kwami (talk) 07:18, 24 May 2010 (UTC)

Kwami -- does this comment belong in another section? Grover cleveland (talk) 01:20, 25 May 2010 (UTC)
A method of automatically displaying the possibly most appropriate transcription system for non logged in users, would be regional IP identification. This has been mentioned before but appeared to receive little or no comment. It would seem to be the technically easiest solution to implement. One would assume that this being an encyclopedia, the majority of visitors are looking for information and not seeking to become Wikipedia editors. Thus a probable majority of visitors have no intention of ever registering and logging in. --Kudpung (talk) 00:09, 25 May 2010 (UTC)
It did receive a comment from Angr, which I hereby echo. It seems extremely crude to attempt to deduce someone's pronunciation from his or her IP address. In addition, I disagree with the claim that this would be "the technically easiest solution to implement". We would almost certainly need help from members of the Wikipedia developer team to make it possible, and their time is extremely scarce. Consider that even getting access to the "internationalisation/language" setting, a comparatively easy task, was only feasible because of a hack that a template expert told me about. Grover cleveland (talk) 01:20, 25 May 2010 (UTC)
Deducing an Internet user's location from his IP is a common feature of web design. It's used all the time.--Kudpung (talk) 02:44, 28 May 2010 (UTC)
I don't think they're arguing its feasability, just its appropriateness. What are some of the things websites do with this sort of deduction? — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 04:55, 28 May 2010 (UTC)
There are two kinds of feasibility:
  • If we had full access to the Wikipedia source code and configuration then, I agree, IP deduction would be trivial. However we do not have such access, so we would have to request help from the Wikipedia development team, which as I mentioned above, is extremely busy, and additionally might not be inclined to give us this kind of access. (I am certain they would not allow us to get IP information for logged-in users, because only CheckUsers are allowed that power).
  • Assuming that we do not have help from the Wikipedia development team, then I do not think that Kudpung's suggestion is feasible. I admit that I have not investigated this extensively, but I base this on the fact that even getting a user's preference in a template is not, in general, possible.
I should also reiterate what I said above, which is that I don't think that this use of IP addresses would be appropriate, even if it were feasible. Grover cleveland (talk) 05:34, 28 May 2010 (UTC)
Using IP to identify an Internet user's location is an extremely easy and common technical process. Wikipedia does it all the time, that's how we know the IP numbers of unregistered editors. It is clear however, that if one has have never used the Internet from outside one's native USA, one will never have noticed how website content changes according to one's location. Google, for example, automatically renders its page in the language of the country one is surfing from, while online shops, such as the Apple appshop, display the currency of the country one is in. One may also see notifications such as 'The service you have requested is not available in your region;', or 'This software is not available for download in your region;', or 'Please wait while we transfer you to our web page for your region.'
Wikipedia is more than a bunch of people creating artticles and arguing about policy. Whatever solution is eventually adopted it will have site-wide implications, thus it might probably not appropriate to suggest that Wikipedia's web development team cannot/will not have time. In fact it's precisely what they do. They do as they are told based on consensus and feedback.--Kudpung (talk) 00:36, 29 May 2010 (UTC)
There is little I hate more than websites trying to guess what format, language, currency and regional news I want to see. They get it wrong too often. I am not able to learn the language of all countries I travel to, nor do I want to be charged 100 euro for things that cost 100 dollar, or be barred from downloading on economic or political grounds. When I want local information I can go to a local site. When I am on a global site, I expect it to behave the same globally. −Woodstone (talk) 11:46, 29 May 2010 (UTC)