Talk:IPA chart for English dialects

From Wikipedia, the free encyclopedia

Chinese character "Book" This article falls within the scope of WikiProject Writing systems, a WikiProject interested in improving the encyclopaedic coverage and content of articles relating to writing systems on Wikipedia. If you would like to help out, you are welcome to drop by the project page and/or leave a query at the project’s talk page.
??? This article has not yet been assigned a rating on the Project’s quality scale.
??? This article has not yet been assigned a rating on the Project’s importance scale.
Notice The page English phonemes used to be a separate article, but is now a redirect to this article. Its talk page can be accessed at Talk:English phonemes

This is the talk page for discussing improvements to the IPA chart for English dialects article.
Article policies
Archives: 1, 2, 3, 4
Archive
Archives
  1. March 2005 to January 2007
  2. February to July 2007
  3. August to December 2007
  4. January to February 2008


Contents

[edit] CE

I really appreciate the additions to the IPA chart because RP, GA, and AuE were not adequate. I participate in a language forum with various people from around the world and the non-Canadians cannot imagine the Canadian monophthongs. It would be really helpful if someone could add a column for western Canadian (BC through Quebec) that includes the three monophthongal vowels that are common in this dialect. see toe hay

Thank you, Rindahl (talk) 00:49, 14 April 2008 (UTC)

How are they pronounced? — Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 02:24, 14 April 2008 (UTC)

I thought I'd answered this question but my answer doesn't seem to appear here. I guess I don't understand how to use Wiki discussion; please be patient with me while I try to figure this out. Anyway, the three Canadian monophthongs I refer to are see (i), toe (o), and hay (e). These are transcribed as diphthongs in other dialects. There is also a Canadian diphthong that isn't represented in the other dialects offered in this chart: eye (ai)Rindahl (talk) 02:01, 1 May 2008 (UTC)

I'm a little confused. Isn't that how it's pronounced in American English? — Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 03:28, 1 May 2008 (UTC)

You'll see on the GA (General American) list that these vowels are registered there as diphthongs. There may be some American dialects (possibly California) that share these monophthongs with Canadians. Rindahl (talk) 00:32, 5 May 2008 (UTC)

As a Korean speaker I'm very sensitive to changes in formants. (At least, I think I'm sensitive...) As such, the only time I've noticed see, toe, hay as monophthongs in the US was the "ee" in "really" being pronounced as "rilly" by some people. --Kjoonlee 10:05, 1 May 2008 (UTC)

Oh oh! I'm excited now. The "hay" sound! I've noticed before that "whale bone" (real bone of whales, not the baleen) is sometimes pronounced like "well bone." I also noticed that "fingernails" was pronounced like "fingernells" after watching Juno. --Kjoonlee 10:10, 1 May 2008 (UTC)

Well, in my dialect, there is a considerable difference between 'well' and 'wale'. I can understand that foreigners hear things differently. It's possible that the way I say 'wale' is the same way you'd say 'well'. This is why I'd like a Canadian section in the chart. I can't make sense of foreign pronunciations as represented in IPA unless I have a base of knowledge about how CE is represented in IPA. The Canadian English 'lay' is an exact homophone of the French 'les' and the Canadian English 'mow' is an exact homophone of the French 'mot'. When I refer to CE here, I mean the western dialect spoken from Quebec to BC. Nova Scotia, New Brunswick, and Prince Edward Island each have their own dialects and Newfoundland has a few dialects. Rindahl (talk) 00:32, 5 May 2008 (UTC)
Ah, "whale bone" I heard from two young girls from Toronto and "fingernails" I heard from Ellen Page, who's from Nova Scotia. --Kjoonlee 16:12, 5 May 2008 (UTC)
I don't have the DVD with me and I can't check for sure for some time, but "pain" seemed to be a normal diphthong. --Kjoonlee 10:13, 1 May 2008 (UTC)

If you listen to a Canuck say 'pay', you'll find that the vowel is a monophthong. The vowel moves somewhat before nasal consonants like n and m, but I think I'd still classify the vowel as a monophthong even in 'pain'. There is always a natural movement from one sound to the next. Rindahl (talk) 00:32, 5 May 2008 (UTC)

If you're talking about a specific dialect which is not very representative of Canada as a whole, then I don't think it fits in this chart. Maybe the Quebec English article might be a better place. --Kjoonlee 18:50, 5 May 2008 (UTC) As I said before, I am talking about the largest, most wide-spread (the standard) Canadian dialect -- the one spoken from British Columbia through Quebec. To repeat: Nova Scotia, New Brunswick, and Prince Edward Island each has its own dialect and Newfoundland has a few dialects, but the dialect generally referred to as CE is the dialect spoken by several millions of people, double the number of those who speak ScE and NZE combined and at least as many as speak AusE. But I see now that CE has been added to the chart and I am a happy man. Rindahl (talk) 11:34, 8 May 2008 (UTC)

I did it just for you. No, actually, it was a long time coming. Jack(Lumber) 15:50, 8 May 2008 (UTC)

Thank you, Jack.:^) Rindahl (talk) 04:17, 18 May 2008 (UTC)

Great! Now, can s.o. do SA? kwami (talk) 17:17, 8 May 2008 (UTC)

If by "SA" you mean South African English, it's already on the chart. Rindahl (talk) 04:17, 18 May 2008 (UTC)

[edit] Tripthongs

Is it worth adding another section for tripthongs? From what I understand, there are three main ones in Standard English:

  • Fire, hire, wire, etc. which take /aiə/
  • Cure, pure, etc. which take /juə/
  • Power, shower, etc. which take /auə/

Any thoughts on this? Epa101 (talk) 08:17, 21 April 2008 (UTC)

If by "Standard English" you mean RP, I'd say only /aɪə/ and /aʊə/ are triphthongs. /jʊə/ is a consonant followed by a diphthong. —Angr 17:05, 21 April 2008 (UTC)


Yes, most people agree with you on the j. However, I thought about this a bit more, and realised that you could add a schwa onto any of the R.P. dipthongs that do not already end in a shwa and get an acceptable tripthong, although some of these are rarely used. Here is my list:

  • /eɪə/ player, sayer, layer
  • /aɪə/ fire, wire, liar
  • /ɔɪə/ lawyer, employer, destroyer
  • /əʊə/ (lawn) mower, (hair) blower, lower
  • /aʊə/ hour, shower, power

If you look back at Joseph Wright’s work, which were early applications of IPA, all of the tripthongs there end in a shwa. In English Dialect Grammar, the only exceptions were /ɪaʊ/ and /ɪəʊ/, which were both confined to Scotland.

Any thoughts on whether the five tripthongs above are acceptable? Epa101 (talk) 22:16, 22 April 2008 (UTC)

You'd need to change them to rhotic schwas. Player is /pleɪər/ even in non-rhotic dialects. There's also the question of diminishing returns: A long list makes it more difficult to find the thing you're after, but these triphthongs are all pretty obvious once you know the diphthongs. kwami (talk) 01:40, 23 April 2008 (UTC)
See Talk:Received Pronunciation (we should move this discussion to one place). Such words often syllabify the schwa. — Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 06:10, 23 April 2008 (UTC)

I have been conducting the same discussion in two places. You caught me! Anyway, I suggest that it carries on, on this page. I have never heard that player is supposed to be said as /pleɪər/ in R.P. I did a quick check on http://dictionary.cambridge.org and Wiktionary, and both suggested that it would be said in the usual non-rhotic R.P. way. I think that the suggestion by Aeusoes on the R.P. page - that words such as "lawyer" and "player" involve a stress on the first syllable - is a more likely explanation for their omission. The R.P. article makes mention of two tripthongs at present whilst this article does not mention them. Should there be a standard that applies to both articles? I would suggest so for the sake of clarity. Epa101 (talk) 13:11, 23 April 2008 (UTC)

I doubt it's necessary. Anyway, the r is pronounced in player up, which means it's /eɪər/, not /eɪə/. The latter is found in maia. kwami (talk) 08:34, 24 April 2008 (UTC)
So much depends on what you mean by "the r is pronounced". In my Canadian dialect, the vowel+r digraph represents a distinct monophthongal vowel. It isn't correct to use IPA to transcribe the Canadian 'player' as /pleər/ because what you are presenting as ər is a monophthongal vowel -- perhaps (I'm guessing) this ɚ is the one. In RP, the er digraph would be pronounced as /ə/. My dialect uses a liquid r (rounded lips, bearing some relation to a w)in most places (as in 'Ralph') where RP would use a consonantal (flapped) r. Canadians use a consonantal r only when it is combined with a th digraph and I think the same is true for many American dialects. —Preceding unsigned comment added by Rindahl (talk • contribs) 02:33, 1 May 2008 (UTC)
RP doesn't use a flap in "Ralph", it uses an approximant. Some varieties of Scottish English use a flap or even a trill in "Ralph", though. —Angr 06:57, 1 May 2008 (UTC)
Captain Picard (actor Patric Stewart) uses RP, although Stewart is from Yorkshire, and he uses flaps/trills in syllable-initial Rs a lot. Anyway, I think Rindahl is saying that "player" ends with a vowel phoneme unless it's followed by a vowel. --Kjoonlee 10:25, 1 May 2008 (UTC)
Which means that it is required in a phonemic transcription. Though I personally have no objection to using <ɚ>. kwami (talk) 18:24, 1 May 2008 (UTC)

Yes, thank you. I did mean that a finial r is pronounced as a monophthongal vowel in my dialect as well as several other (primarily North American) dialects. This phoneme is in 'fur'. The same monophthong is sometimes represented by the digraphs 'er' and 'ir', but ar, er, ir, and or are sometimes pronounced as diphthongs and ir is sometimes pronounced as a triphthong. In my dialect 'ire' is a triphthong. It begins with something that is not quite æ and not quite ɑ (perhaps a is what we use) and followed by i which is followed in turn by the ur digraph. I probably could use ɝ or ɚ for what I call the vowel r but I don't know what these symbols represent because you don't have a Canadian section in your chart. In my dialect, r (unless it is in an initial position) is a monophthongal vowel even though it is always written as a digraph, and it is markedly distinct from a schwa. Think aboout the onomatopoeic sounds we use for animals; a cat's prrrrrrr uses a consonantal r and a dog's grrrrrrrr uses a vowel r. Thanks again. I hope somebody here can put up a Canadian section.Rindahl (talk) 20:18, 4 May 2008 (UTC)

Sorry, I have no idea what you mean by a consonantal vs. vowel r in the last example, or the difference between prrr and grrr. kwami (talk) 20:38, 4 May 2008 (UTC)

When you imitate the sound of a cat, don't you trill your r? Flapping your tongue against the top of mouth makes a consonant. The growl of a dog is pronounced in the throat (at least it is in my dialect) and this is a vowel. The two sounds are completely unrelated to each other and yet both are represented in our alphabet by an r. Rindahl (talk) 00:40, 5 May 2008 (UTC)

Somewhere above Kwami contrasted Maia and player in the context of a non-rhotic accent. I wondered if Kwami was making a point that non-rhotic speakers distinguish rhoticised vowels from non-rhoticised vowels in a way that's phonemic but only revealed in the linking r. As speaker of New Zealand English I have a non-rhotic accent and both Maia-up and player-up have an r sound at the word boundary. From this article it appears that this is called an intrusive r and is more common than not in non-rhotic accents. I'd argue that the linking/intrusive r is just a way for a non-rhotic speaker to get between two vowels that don't usually appear in combination, but can at a word boundary. In other words, it's phonetic not phonemic. (By the way I say /pleɪə(ɹ)/ and /maɪə(ɹ)/ so there are other differences between those tripthongs, but that's pretty much irrelevant to the issue I'm trying to clear up.) Ben Arnold (talk) 06:19, 9 May 2008 (UTC)

That's exactly what I was talking about. For you, then, the distinction has been lost. For other non-rhotic speakers, however, I believe it's maintained. kwami (talk) 06:40, 9 May 2008 (UTC)

[edit] Tore, tour examples

There have been a few reforms to the chart recently, mostly good ones. I have noticed that the English pronunciation for tour is now given with two forms. Do we need to use "tour" as our example, or could we not use a word that is a clearer example?

I know that some linguists have talked about /uƏ/ becoming less frequent but there are still lots of words where it is the only recognised pronunciation: cruel, jewel, jury, rural, endure, cure, lure, endure, pure. If you want words that sound similar as tore and tour do, how about door and dour or bore and boar? Epa101 (talk) 09:08, 12 May 2008 (UTC)

I have checked the Oxford Dictionary and "tour" only has the /uƏ/ as R.P. I also changed the vowel for the my group of words, as I noticed this whilst I was checking. Epa101 (talk) 15:40, 15 May 2008 (UTC)
My edit has been reversed. What sources claim that tour may be pronounced with /ɔ:(ɹ)/ in R.P.? Are these sources reliable? Why do the guides that you get to how to use IPA at the start of dictionaries always use poor as the example of /uƏ/? Epa101 (talk) 21:46, 15 May 2008 (UTC)
Note that the Wells link provided by Jack in the discussion on the Oxford Dictionary below uses /uƏ/ for poor as well. Epa101 (talk) 21:49, 15 May 2008 (UTC)

[edit] Conservative distinctions

I'm going to try to clean parts of this table up a little. In particular, I'll try to add Welsh English vowels.[1] - Gilgamesh (talk) 15:32, 14 May 2008 (UTC)

Okay, but you're making up a new wiki standard which links to but doesn't match anything else. kwami (talk) 18:42, 14 May 2008 (UTC)

[edit] English regional English

I hesitate to suggest adding even more varieties of English to the pronunciations, but... At present "English English" is represented only by RP – while this is is aspirational for many and widespread amongst affluent people (especially represented by the BBC, many actors etc), it is relatively uncommon or strongly diluted amongst most "ordinary" working people. There are many local varieties of English (often confined to quite small areas), but there seem to me to be three broad types which are serious omissions from the chart: Northern English, either Cockney or Estuary English, and rural Southern English English. To my ear these types are are at least as distinct as Scottish, American or Australian. Any thoughts...? --Richard New Forest (talk) 21:01, 14 May 2008 (UTC)

I think this page is getting increasingly difficult to read. I think we should go back to the format that International Phonetic Alphabet for English (which used to be a different page but now redirects here) had when it looked like this. —Angr 05:08, 15 May 2008 (UTC)
You say "getting" as in the addition of Canadian, Welsh, Scottish, Irish, and South African are cluttering up the information. Might we then remove some? Richard New Forest's suggestion makes me wonder if most Brits don't understand RP fairly well anyway (as it's the standard dialect). — Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 05:12, 15 May 2008 (UTC)
And Australian and New Zealand. I'd have no objection to removing everything except RP and GenAm, which is how the page originally was, but it will never stay that way. People will always come along and want to add a "me-too" column for their own accent. That's why I think the arrangement linked to above is better: an actual encyclopedia article with separate sections for each variety. The "quick reference chart" function of this page is no longer needed anyway, since Help:Pronunciation now fills that role and is the page all the pronunciation templates now point to instead of pointing here. —Angr 05:50, 15 May 2008 (UTC)
Then maybe this should redirect to English phonology. — Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 06:01, 15 May 2008 (UTC)
I'd be alright with that, as long as all these important varieties are represented there in IPA. It's still important to show how, on an inclusively international level, some phonemes are not yet merged, and indeed have merged in prestige accents only very recently within the past one or two centuries (such as [oː] vs. [oʊ], which is even still distinguished under certain circumstances in some dialects as far away as New Zealand). Since we are talking about English—the international intelligible language—we cannot say that the only polished English that exists in the world is that spoken on BBC and CNN. It cannot be said, for example, that there are no Welsh English polished varieties that phonologically contradict RP. The fact of the matter is, English is a giant 'me-too' language. Either we are inclusive of all the local prestige accents, or we treat the inevitability of English breaking into multiple languages, and treat RP, GA and so forth as different. This can't purely be a fantasy of English being a U.S.-U.K. language with everything else being irrelevant, as Wikipedia is a neutral point of view world-wide encyclopedia in English, without saying whether that is American English, Received Pronunciation, Hiberno-English, Welsh English, Scottish English, Canadian English, Australian English, New Zealand English, South African English, whatever... We cannot dismiss the whole world as patois when all sorts of perfectly valid posh accents exist—that would be brazen POV. - Gilgamesh (talk) 07:28, 15 May 2008 (UTC)
My understanding of this page's function is that it helps people with non-standard dialects understand our compromise system. If, as Angr says, Help:Pronunciation is all that we need then, yes, the information here can be put at English phonology and even expanded on.
I notice that no one has gone to Welsh English to put in the phonetic information. — Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 07:40, 15 May 2008 (UTC)
Yeah, I need to get around to doing that. I have the source, and I added a link to it in the references already, but it really needs to be entered into the article body. - Gilgamesh (talk) 08:16, 15 May 2008 (UTC)
Ugh, I just realized I don't know how concrete my references are. The reference was handed to me by someone on FreeNode #wiktionary. It's also the difficulty in references with many pages missing because it's a preview of a copyrighted work on Google Books. But what do I do? I don't really have practical access to shelf books. - Gilgamesh (talk) 08:24, 15 May 2008 (UTC)

Alright, I have an idea. To be internationally NPOV in regards to English, we need to capture the state of the global language. A broad transcription (compromise) should ideally be maximally distinctive, representing all the world's regional prestige accents (including the regions of England, Wales, Scotland, Ireland, and the former colonies and so forth), and should preserve distinctions until they go extinct in every prestige accent. I know that for a regional monolingual speaker, this might create a certain burden of having to learn distincts they never hear in their own ears on their own streets, but as English has so many prestige accents and the Great Vowel Shift is still on-going, this has become a necessity for the English language when treated on an international level. The unusual speech of one or two people might possibly be dismissable as non-notable without violating NPOV because it cannot be easily verified with references that back up its lasting prestige phonology. However, where prestige accents are attested to exist, even in places as small as Wales, we should honor what each of them distinguishes, and reflect that in the broad pronunciation. Since in English (because of our on-going Great Vowel Shift) this reflects the vowels (including vowel-liquid combinations) far more than the consonants, this is where broad transcription will spend the most time and effort. Prescribing pronunciations based on only one or two dialects violates NPOV. Fortunately, nonoriginal research on dialects already exists—we just need to gather it together. Some might say that the very act of gathering and organizing nonoriginal research as sources is itself original research, but where do we draw the line? Current RP-GA-only linguistic attitude continually brazenly violates NPOV, and that's just unacceptable. So, here's a very very tentative set of broad transcription IPA vowels—I do not suggest we use this in its exact form, but merely that adequate existing sources can verify and reference 99% of it, if not all of it: [ə ɚ ɨ ɫ̩   æ ɛ ɪ ɒ ʌ ʊ   ɑː eː (ɪː) iː ɔː oː ʊː uː   aɪ aʊ eɪ ɪʊ ɔɪ oʊ   ɑɚ ɛɚ ɪɚ ɔɚ ʌɚ   ɑːɚ eːɚ ɪːɚ (iːɚ) ɔːɚ oːɚ ʊːɚ (uːɚ)   aɪɚ aʊɚ eɪɚ ɪʊɚ ɔɪɚ oʊɚ]. The vowels in parentheses represent possibly dead distinctions; for instance, continued [ɪː] distinction is entirely conditional on whether or not it has gone extinct (whether the meet-meat merger is complete in all prestige accents—my brother reported still hearing it on the streets of Liverpool, though of course we cannot use such unconfirmed hearsay as a reference). But until we know that they are extinct, it might be responsible to still use them. For example, art as [ɑɚt] and are as [ɑːɚ]. Metadistinctions like [ʌ ʊ ʊː uː] help bridge accents that do not unround [ʊ] to [ʌ] (northern England, Ireland and so forth) with those that do, as the rounding accents correspond among themselves to [ʊ ʊ uː uː] while the unrounding accents have these as [ʌ ʊ ʊ uː]. - Gilgamesh (talk) 08:17, 15 May 2008 (UTC)

I agree that the page is packing a lot of information in – but I do think we need to have this information in a single table somewhere. It may not be all that easy to read, but it's not nearly as hard as chasing down the articles on each type of English, finding the appropriate sounds in the text and flipping backwards and forwards between them. Would it help to use the full width of the page for the vowels?
Perhaps we do need to think about the purpose of this page. Aeusoes says that it's to help people understand the IPA compromise pronunciations – but is that not the job of Help:Pronunciation? This is an article, not a help page. My feeling is that it should describe English as a whole. If regional types are not covered here, then where?
While we're on it, are there other major omissions? For example, surely we ought to cover West Indian and Indian Sub-Continent types, and some of the most distinct US accents (to my ear New York Dialect is far more distinct from GA than Canadian English is). Personally I don't think that "prestige" dialects are nearly as important or interesting as "real" ones – apart from anything else they gravitate towards RP/GA anyway.
A somewhat different point is that we seem to assume that consonants do not vary between dialects – it's a little odd to see, for example, Hiberno-English vowels covered without the consonant variations; the same would apply to Estuary English, with its T-glottalisation, th-fronting etc. If we did use the full width of the page for vowels, we could also include consonants in the same table. For most consonants the same pronunciation would be repeated across the table, but the variations would show up clearly (as currently for ɪ or ʊ). --Richard New Forest (talk) 09:37, 15 May 2008 (UTC)
Well, I didn't mean that there are no consonant differences. They're just not as big, and that's true. But certainly, document consonant differences too. As for omissions, most certainly. What we don't have here yet is data (notice the completely empty South African English column). The references surely exist—we just need to find them, little by little. - Gilgamesh (talk) 10:31, 15 May 2008 (UTC)
Oh, and by prestige accents, I didn't mean anything artificial. I just meant, from region to region, whatever sounds the most natural and mainstream of that region. I didn't mean snobbery of an exclusive elite. I mean, there's a difference between being educated and simply sounding like it with unnatural speech adjustments. An educated and clear person doesn't have to change their regional accent to be educated and clear. - Gilgamesh (talk) 10:39, 15 May 2008 (UTC)
By the way, I'm going to be leaving in a few hours until late Sunday or early Monday. If I don't give input during that time, then it doesn't necessarily mean I have nothing to say—I'm just not there. - Gilgamesh (talk) 10:50, 15 May 2008 (UTC)

I don't think that Northern English should be given a separate column. I live in York currently, and grew up in West Yorkshire. Speech here varies between old and young people a lot, and there is also much variation between different towns. There is no General Northern accent. Perhaps, there should just be a note that there is never a long /ɑː/ in bath, grass, demand, etc. and on the /ʊ/ vowel in cut, strut, etc. That does not seem worth making a column in its own right though. Epa101 (talk) 15:49, 15 May 2008 (UTC)

Well, as for Northern English, I wouldn't know very well personally. Maybe it deserves its own column, maybe it doesn't. - Gilgamesh (talk) 06:48, 16 May 2008 (UTC)
There are indeed many types of Northern English, and we can't include all of them – but they do have common features. I think it is both worth including, and possible to do. --Richard New Forest (talk) 09:04, 16 May 2008 (UTC)
I think Wells's Accents of English has a fairly detailed description of a Leeds accent as a representative of what he calls the "middle north" (also including Sheffield and Manchester but not of course Liverpool). That might be a place to look.--JHJ (talk) 11:47, 16 May 2008 (UTC)

[edit] Oxford Dictionary should be source for R.P.

I think that we should use the Oxford Dictionary as the basis for R.P. This is what has always been accepted as the authority in Britain. At the moment, the reference is for a 2004 essay by Roach. I don't think that one essay should be the basis for this system. The Oxford Dictionary uses large teams of experts whilst that essay was by just one person. I noted a few small mistakes with the table, and added in the O.D. as a source by which R.P. could be checked in the references. Epa101 (talk) 15:51, 15 May 2008 (UTC)

I have had a few more thoughts on this. There have been many additions to the chart recently. The way that the page is set up suggests that all recent additions are sourced in Roach, but did everyone who made them check that particular 2004 work by Roach before doing them? If not, the page misrepresents a source. This seems as another argument for using the O.D., as that is more widely available to everyone.

Secondly, I am not sure that the R.P. entry for bad, cat, lad, etc. is correct. My copy of the O.D. uses /a/. I shall check this first, but I thought that /æ/ is the old-fashioned form and is now considered regional to the south-east. I'll give this a bit more research. Epa101 (talk) 16:05, 15 May 2008 (UTC)

Oxford dictionaries use a unique transcription system, devised by Clive Upton; we use the more traditional, more common, and more widely accepted Gimson system (Roach basically uses Gimson's symbols). Many linguists, including J. C. Wells, have criticized the Upton system; see [2]. In English English dialects, the TRAP vowel may vary from [a-] to [ɛ]; Northern dialects normally have [a] or [a-]; the Queen has [æ̝]. Currently, in Southern dialects, this vowel is more open than it used to be (for example, in Estuary English it usually is [a], with the DRESS vowel often being [æ]), which is why Upton decided to use the symbol /a/. The substitution of /ʌɪ/ for /aɪ/ for the PRICE vowel is somewhat "bizarre," as Wells puts it. Jack(Lumber) 19:07, 15 May 2008 (UTC)
We don't use the Gimson system exactly: it uses /e/ for DRESS. As for TRAP, if both [a] and [æ] can be shown for Canada then why not for England?--JHJ (talk) 20:56, 15 May 2008 (UTC)
(edit conflict) Furthermore, keep in mind that different phonologists (for example, a Briton and an American) may choose different symbols to transcribe very similar sounds, or, conversely, they may choose the same IPA symbol to transcribe different sounds; this should be taken into account if we are to provide a meaningful comparison of different dialects. For instance, many British phonologists transcribe RP DRESS as /e/; many American phonologists transcribe GA FACE with the same symbol. Yet RP DRESS and GA FACE are very different; RP DRESS is, in fact, practically identical to GA DRESS, which is usually transcribed as /ɛ/. In order for our synthesis to make sense, we have to pick the most consistent transcription systems in use. Jack(Lumber) 21:01, 15 May 2008 (UTC)
As for Canada: Linguists usually use /æ/ for TRAP-BATH and /ɑ/ for PALM-LOT-THOUGHT, while the Canadian Oxford Dictionary uses /a/ and /ɒ/, respectively. Actual Canadian pronunciation is somewhat erratic, the exact phonetic quality varying from region to region, speaker to speaker, and even word to word, so the Wikipedia analyst decided to use both pairs of symbols. The Wikipedia analyst being Jack(Lumber) 21:10, 15 May 2008 (UTC)
The same applies to England, though, even within RP. If that's the way to go for Canadian English, I think the RP cells for TRAP, DRESS and SQUARE should contain both the Gimson and the Upton symbols. (Maybe NURSE and PRICE too.)--JHJ (talk) 21:15, 15 May 2008 (UTC)
For PRICE, it would probably mislead the reader into thinking that RP has some form of Canadian/American raising, which is untrue. By the way, ScotEng too has two different realizations of the PRICE diphthong--although not in the same environments as NAmEng; this should be noted somehow... Jack(Lumber) 21:22, 15 May 2008 (UTC)

Why are we favouring Wells over Upton on the "price" vowel? Was it written as /aɪ/ in older editions of the Oxford Dictionary? (bearing in mind any difference in notation system) I consider Wells to have a very unconventional definition of R.P. but, other than that, isn't the Oxford Dictionary the traditional benchmark of R.P.? Why is that Upton is chosen rather than Wells to write the introduction? Wikipedia policy is that we should go with the most reliable source, which seems to be Upton.

The symbols that we use are not particularly important, so the bad, cat vowel issue is fine by me. I have noticed that the pronunciation of "tour" as /tɔ:(ɹ)/ has been restored for R.P. This cannot be considered R.P. Tour has always been said to rhyme with poor, and poor is given as the test word for a /ʊə/ in practically every R.P. guide there is, including the Wells link that Jack Lumber posted, and in the most recent books such as Oxford B.B.C. Guide to Pronunciation and Ged Rid of Your Accent. Also, is the /tɔ:(ɹ)/ in Roach's work? He is the current reference for R.P. It just had /tʊə(ɹ)/ until fairly recently when someone inserted /tɔ:(ɹ)/ in there as well. It cannot have had both versions of the page in his work, so one of them must be wrong and I expect that is the current version. The same goes for the pronunciation of pure. If there is a modification of how to write /ʊə/ in different systems, then that is fine, but to say that there is more than one R.P. pronunciation of tour or poor seems as a very unconventional view to me Epa101 (talk) 21:39, 15 May 2008 (UTC)

P.S. I tried to put a [citation needed] tag on those pronunciations, but it made the whole table look unpresentable so I did not go ahead with the edit. Epa101 (talk) 21:43, 15 May 2008 (UTC)

Yes, PRICE used to be /aɪ/ in older Oxford dictionaries, including the second edition of the OED. Anyway, remember that this page is supposed to compare the phonetics of different dialects; preferring /ʌɪ/ over /aɪ/ would only mislead the reader into thinking that RP right = NAmEng right, while RP ride != NAmEng ride, when, if anything, it's the other way around. Upton's use of /ʌɪ/ for RP PRICE--I guess--is meant to show that the nucleus of the PRICE diphthong is similar to the STRUT vowel, which may be true for some Southern English accents, but it's false for NAmEng--note that we are using the same symbol for RP STRUT and NAmEng STRUT. J. C. Wells is arguably the most influential British phonetician; he devised the PRICE, STRUT, GOAT etc. notation, by the way. I don't believe that Upton's system is particularly representative of RP, either; I guess that it's just supposed to cover not just RP, but a more or less broad range of RP and near-RP accents of Southern England. After all, the pronunciation key of a dictionary is never supposed to teach you the "proper" phonetic values of the vowels, but is merely supposed to tell you how words are (phonologically) pronounced; for example, if I realize PRICE as [ɑɪ] and STRUT as [ɐ] (two realizations very common in Southern England), all I need to know is that /ʌɪ/ = [ɑɪ] and /ʌ/ = [ɐ], regardless of which vowel sounds the IPA symbols in slashes actually represent.
Re: poor: From Peter Trudgill, International English, p. 10:

A more recent, but by now also widespread development, is the loss of /ʊə/ and the merger of this diphthong [...] with /ɔː/. This latter change for some speakers has affected some words but not others, so that sure may be /ʃɔː/ but poor, /pʊə/. The current situation with respect to these vowels is something like this: [older speakers: paw, pore, and poor are all different; middle-aged speakers: paw = pore, but poor is different; younger speakers: paw = pore = poor = /pɔː/]

Jack(Lumber) 23:13, 15 May 2008 (UTC)

I went to read Roach's book in the library today, and made a note of the relevant bits.

  • Re: price vowel. I see what you mean, and agree with you now. Upton's description of the vowel seems to have also been used in much of the Survey of English Dialects work, where /ʌɪ/ is given for much of southern England and /aɪ/ is the normal pronunciation for the Yorkshire region. I know that this was done a long time ago, but I presume that the former was meant to represent the R.P.
  • Re: poor vowel. I think that Trudgill's view is the consensus, including the view of Roach. My point is that the table does not actually say this at present. I think that this vowel shift must have started a long time ago, as the 1975 clip of the Queen here shows her pronouncing "sure" so that it sounds as "shaw". Roach did note that many people in England merge the vowels but the important thing is the R.P. form he gives. As his is the work we use, it must be represented properly.
  • I also noted that the vowel added for the vowel in Engish is not mentioned in Roach's work anywhere. I shall put a referenced tag for this. Epa101 (talk) 11:49, 16 May 2008 (UTC)


By the way, to get back to my point about the O.D., I presume that there is more than just Upton who works on the pronunciations? In the large editions, every single word has a pronunciation listed next to it (there are odd cases where two pronunciation are acceptable, such as "again" or "either"). I am sure that there is a team of experts who work on them. I was always told that this was the benchmark of R.P. I believe that the B.B.C. works by it, as they collaborated with Oxford for their recent pronunciation guide. I know that figures such as Wells see trends in the south of England, such as l-vocalisation or t-glottaling, to represent trends within R.P., but I ought to point out that it would actually offend people in some of the Northern towns where R.P. is often spoken, such as Harrogate or (parts of) York. I suggest that pronunciations be checked against it but then transferred into the Gimson system. This might be a way of dealing with sounds that Roach did not mention in his 2004 work. Epa101 (talk) 12:17, 16 May 2008 (UTC) P.S. http://www.leeds.ac.uk/english/staff/pages/staffindex.php?file=upto Just confirmed that Upton was a fieldworker for the Survey of English Dialects. That is probably why he uses /ʌɪ/ for RP "price", as that is what the survey used. Epa101 (talk) 13:39, 21 May 2008 (UTC)

[edit] Dialects listed in alphabetical order

I've placed all the dialects listed in this chart in alphabetical order. I did this to help with clarity, much like the WP:Spelling article. :) --203.220.170.4 (talk) 03:15, 17 May 2008 (UTC)

Wouldn't it be better to group similar accents together so that we could do something like this:
IPA Examples
England
RP[1]		 United States
GA		 Canada
CanE		 Australia
AuE		 New Zealand
NZE		 Wales
WaE		 Scotland
ScE		 Ireland
IrE		 South Africa
SAE		 compro-
mise
IPA: English Vowels
æ æ/a æ,
æː[2]		 ɛ a a æ lad, bad, cat[3]
ɑː ɑ ɑ/ɒ aː ɐː aː ɑː ɑː father
ɒ ɔ ɒ ɒ ɔ ɑ ɒ not, wasp
ɔː ɔ oː ɒː ɔː ɔː law, caught, all, halt, talk
Or is that too much? — Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 03:42, 17 May 2008 (UTC)
No on first part of question, yes on second part of question.
If you look around Wikipedia you'll see that nearly all lists and charts of something have their elements listed in alphabetical order. I think it makes sense to do it in this way, especially when a list or chart has the potential to grow.
I alphabetised it not only for clarity and ease of use, but also to convey that not one dialect is better than the other. I was getting the impression that RP was first, GA was second, etc in importance, and down the end there was Irish and South African which gave me the impression that they were not that important as RP or GA.
It starts to get complicated when you start grouping dialects with similarities together, because many of these dialects listed here have both similar and dissimilar elements. But also, once placed in similar groupings, how do you order these groupings: in order of importance, regional order, no. of users. I think alphabetical order is simple and clean, and you won't (potentially) get someone's back up. --203.220.170.4 (talk) 04:26, 17 May 2008 (UTC)

[edit] "aw, aj, ej, ij, ow, uw" vs. "aʊ, aɪ, eɪ, i, oʊ, u" for GA

Hi all. I am a little confused... in my freshman phonology class at Rutgers University (New Jersey, USA), I was taught that the glides that follow the tense mid close vowels and close vowels in GA are [y] and [w]—which my professor carelessly said one day were exact Americanist equivalents for IPA [j] and [w]. I was also taught that these glides combine with [a] to form [ay] and [aw], which was I again told translate seamlessly to IPA [aj] and [aw].

I see, however, that consistently in almost all literature on GA I have been able to find on the internet (including every Wikipedia article I have read on English phonology), the glides that follow tense vowels are represented as [ɪ] and [ʊ], respectively. Is this just an alternative way of representing the same sounds, or are Americanist [y] and [w] not exact equivalents of IPA [j] and [w]? And if the latter is true, what is the scientific evidence behind this convention?

If [ɪ] and [ʊ] are just alternatives for [y] and [w] when representing glides, then perhaps we should show "Americanist-style IPA glides" (for lack of a better term) as alternatives for GA vowels—namely [j] and [w]. However, I have noticed that even Merriam-Webster uses a simplified version of the GA transcription proposed in this article, so I am beginning to question my phonology professor's casual claim that there is an exact correspondence between the two ways of representing the English tense vowel glides... --Rcgy (talk) 07:38, 20 May 2008 (UTC)

The difference depends on how precise you wish to be, and which features you consider relevant. See English phonology#Transcription variants. —kwami (talk) 07:53, 20 May 2008 (UTC)
OK, got it. Thanks. And so why are we not being consistent and putting down /iɪ/ and /uʊ/ for the tense close vowels? Plain [i] and [u] occur less frequently than their allophones [iɪ] and [uʊ]—are we really justified in choosing the less common allophones for the GA close vowels simply because they are easier to write? --Rcgy (talk) 08:09, 20 May 2008 (UTC)
Actually, they'd be /ɪi/ and /ʊu/. I think we had that at one point. You can find the discussions in the archives. —Preceding unsigned comment added by Kwamikagami (talk • contribs) 09:51, 20 May 2008 (UTC)
Interesting. Someone must have removed that section because I have searched for it and I don't see it. So, for future readers, we are basically saying that a modern, nitpicky listing of the 'tense' vowels of GA would be: /eɪ/, /ɪi/, /oʊ/, /ʊu/ where /ɪi/ and /ʊu/ are pronounced [i] and [u], respectively, in unaccented syllable final positions... which is why, for convenience, we are using /i/ and /u/ as the underlying phonemes... but we're leaving /eɪ/ and /oʊ/ because the glide is more salient in these two and omitting it in pronunciation would cause a perceptible difference to a native speaker... right? This is all very interesting, clears up many doubts; please correct me if any of my assumptions are wrong. Thank you! (And if anyone out there remembers exactly why we are no longer listing /ɪi/ and /ʊu/ for GA please share. Thanks again!) --Rcgy (talk) 00:53, 21 May 2008 (UTC)
It's mostly for the same reason we use <ʌ>; it's what the literature most commonly uses. — Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 03:28, 21 May 2008 (UTC)

[edit] Hong Kong & Indian English

User:169.231.10.131 inserted columns for Indian English and Hong Kong English. User:JackLumber deleted them, but I have restored them for the moment, pending discussion.

The summary for deletion was "...native varieties only, please". My feeling is that this is a very poor reason for the deletion, because:

  • Most of the dialects already included are not the native language of their area, and are not even necessarily the ancestral language of many of their users. How do the additional ones really differ?
  • Both Hong Kong English and Indian English have millions of users, including many first-language users.
  • WP is an international encyclopaedia, and we have to try avoid bias. Excluding these dialects is hardly achieving that (see WP:BIAS).

All this follows on from the earlier discussion about English regional English. If we are going to describe the English language as a whole, we do need to include some indication of the full range of dialects used.

If we are going to restrict dialects somehow (and of course we can't include them all) we ought to have some sort of objective criteria for inclusion. --Richard New Forest (talk) 16:56, 22 May 2008 (UTC)

To your statement and question "Most of the dialects already included are not the native language of their area, and are not even necessarily the ancestral language of many of their users. How do the additional ones really differ?" I would respond, the other dialects are the native languages of the speakers in question: Australian English is the native language of the people whose accents are described under "AuE", likewise for Canadian English, General American, Irish English, New Zealand English, Scottish English, South African English, and Welsh English. As for RP, I know it's fashionable to claim that no one really speaks RP as their native accent, but in fact I've met several people who do. (Still, it probably has more nonnative speakers than native speakers.) Indian English and Hong Kong English, on the other hand, are overwhelmingly nonnative. Including them here is a bit like adding columns for "English with a Spanish accent" or "English with a French accent", both of which also have millions of speakers. As for objective criteria for inclusion of an accent, how about "Has been used as the basis of a published pronouncing dictionary or similar pedagogical work"? That would bring this page back to just RP and General American (maybe AuE, I'm not sure). —Angr 18:29, 22 May 2008 (UTC)
To put it another way: EngEng, USEng, CanEng, IrEng, ScotEng, WelshEng, AusEng, NZEng, and SAfrEng (and Caribbean English) are the most "important" native varieties of English. If the table is to provide an IPA chart for each of these, we have to pick a particular accent as the reference model in each case (RP, General American, etc.) To be sure, this task may be a little daunting in some cases--in IrEng, for example, there is considerable regional and social variation. If we were to include the second-language varieties of English, or "New Englishes" as McArthur calls them (HongKongEng, IndEng, SingEng, PhilEng, WAfrEng, EAfrEng), not only would the task be overwhelming--the chart would be just impossible to look at. Not that it's particularly appealing right now, mind you. As unpopular as it may be, Angr's suggestion sounds reasonable, after all. Jack(Lumber) 19:22, 22 May 2008 (UTC)

Those arguments are good ones, but... It is not true that Indian English is just "English with an accent" – it may be that most speakers have it as a second language, but there are also very many first-language speakers, many of whom are descended from several generations of first-language speakers, and it is a well-developed dialect with its own vocabulary and other idiosyncrasies. It has arisen in exactly the same way as many other dialects – Welsh English for example is still a second language for very many speakers (in some areas for most speakers), and South African English is heavily influenced by speakers with other first languages. It's not so long since all non-RP dialects were seen as "wrong", including GM...

My point is that we need objective criteria – I think "important" is very open to personal opinion, and therefore bias. My feeling is that we should consider any dialect which is definitely English (not a creole or interlanguage etc), which is fully developed with its own accent and vocabulary, and which is distinct enough to be distinguished from others. I don't think the proportion of first-language speakers is important – a dialect should not be barred just because lots of non-native speakers have learnt it (or perhaps we'd be knocking out RP too). However I think the total number of first-language speakers is important.

I think it would be a great shame to restrict this page to too few dialects. Leaving aside the waste of the work that has already been done, the page would then fail to describe English as a whole, which I think it ought to do.

I don't know enough about Hong Kong English – if it is just "badly learnt" English with few first-language speakers, it should probably not be included. However, my feeling at the moment is that Indian English probably should be. --Richard New Forest (talk) 21:33, 22 May 2008 (UTC)


I am in favour of including Hong Kong English and Indian English in the list. In Hong Kong, the English language is used extensively as an official language of the Government and the Legislature, as the medium of instruction in most schools, as the working language in the business sector, and as the lingua franca for the communication between different ethnic groups. India has a similar situation -- also as a result of many years of British administration.

It is true that the English language is not the only language used in Hong Kong, because multilingualism is the official policy of the Hong Kong Government. But this is not an acceptable reason for deleting Hong Kong English from the list. As a matter of fact, a lot of other countries also use languages other than English -- the United Kingdom (English, Scottish Gaelic, Welsh and Irish), Canada (English and French), South Africa (English, Afrikaans, Zulu, Xhosa etc.), the United States (Spanish is official in New Mexico and Puerto Rico; French is official in Louisiana; Hawaiian is official in Hawaii). Why don't you go and delete those countries from the list also? Why should Hong Kong and India be discriminated? How do you define which accents are "important" and which ones are not?

Some folks tried to compare Hong Kong English and Indian English to "English with Spanish accent". I would like to point out that Hong Kong English, Indian English, British English, American English etc. are discussed in the article "Regional accents of English", while "English with Spanish accent" is discussed in "Non-native pronunciations of English". I don't think Hong Kong English and Indian English go under the same category as "English with Spanish accent". From what some people posted here, it just seems like they have some kind of prejudice / discrimination / racism against the people of Hong Kong and India. -169.231.10.131 (talk) 00:02, 23 May 2008 (UTC)

I'm leaning towards accepting Angr's suggestion as well. If we added a caveat of well-studied accents, this would allow the inclusion of Australian and maybe a few others but I'd recommend finding the research first and then adding the column. South African English has been up there for months and nobody seems to know how they pronounce their vowels. — Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 06:21, 23 May 2008 (UTC)
I don't think we should go back to just RP and GA: that would give an unnecessarily narrow view of the language. My view is that there's a case for including an accent if it is well enough described in a reliable linguistic source (e.g. Wells, Accents of English) and has notable differences from other accents already in the table. (Of course, "notable" needs to be made objective there.) I think it should be restricted to native varieties (there are other places for describing non-native ones) but I'm not sure that that does exclude Hong Kong and Indian English.--JHJ (talk) 07:52, 23 May 2008 (UTC)

Indian English would be a keeper, if we can find a source. I'm not sure there's a single Indian English, however. kwami (talk) 08:34, 23 May 2008 (UTC)

There isn't. It "varies quite considerably depending on the speaker's native language as well as on his or her educational background and degree of exposure to native English" (Trudgill & Hannah 2002:129-130). Native speakers, on the other hand, apparently tend to have a RP or near-RP accent; but these are far outnumbered by those for whom it is an additional language (ibid.). The article Non-native pronunciations of English should have a different title, because it really is about speakers of English as a foreign language. I mean, a Francophone Montrealer may be a second-language speaker of English much like someone from India whose first language is (for example) Hindi, with the caveat that the Quebecois are probably more exposed to native English than Indians, surrounded as they are by native speakers. Crystal (2003:358) distinguishes between the "outer circle", consisting of those countries where English has come to play an important role as a second language through a history of colonial contact (such as India and Nigeria), and the "expanding circle", made up of countries where the importance of English as an international medium has been recognized, but the language has received no special status (such as Japan and Brazil). So we basically have 1) native speakers; 2) speakers of English as a Second Language; 3) speakers of English as a Foreign Language. For the most part, Indian English falls into bucket #2. This doesn't mean we should leave it out of the table, however; askoxford.com makes the point that [t]he main regional standards of English are British, US and Canadian, Australian and New Zealand, South African, Indian, and West Indian. Within each of these regional varieties a number of highly differentiated local dialects may be found. But what should the reference standard for Indian English be? By the way, verifiable sources for Hong Kong English are not easily found--provided that Hong Kong English doesn't fail WP:N, to put it in Wikipedia terms.
As for RP: RP is not important per se, but because it is the reference model we have chosen to represent English as spoken in England--for lack of a better, more "neutral" or "universal" option. Jack(Lumber) 19:57, 23 May 2008 (UTC)

[edit] "Pour"

What's the source for this line, with its ɔa for GA and Hong Kong and its own symbol in the "compromise" column? In Wells's lexical sets, pour is a FORCE word, just like tore, boar and port in the row above.--JHJ (talk) 08:14, 23 May 2008 (UTC)

An error introduced by the HK column, which has been reverted several times pending correction. kwami (talk) 08:33, 23 May 2008 (UTC)
It's also shown as different in the Welsh column. I can't find any evidence in either the originally cited source or in this one (Rhondda) for words spelt with our having a different vowel from the /oː/ of FORCE. Nor can I find any evidence for words spelt with air having a different vowel from the /ɛː/ of SQUARE, so the row with hair and their is also questionable.--JHJ (talk) 12:43, 23 May 2008 (UTC)

[edit] Empty South African columns

Any reason why the columns for South Africa are empty? OhNoitsJamie Talk 03:51, 4 June 2008 (UTC)

Yeah, we haven't found anyone to fill them in. Are you volunteering? kwami (talk) 04:09, 4 June 2008 (UTC)

[edit] Move refs

I've moved the references from the header row of vowels to a more universal location, as the refs deal with both vowels and diphthongs, rather than have two(× two) references for vowels and diphthongs. I hope this does not upset anyone.
I realise both AusE and CanE are typical or common short forms for these two dialects — used Google to determine this — but for just this article I do not see any problem with having the first two letters of the dialects origin (ie, Australia, Canada, Ireland, etc) plus 'E' (for English), with exceptions of GA and RP, as codes. These codes are just for identification purposes for the tables. --203.220.171.54 (talk) 03:20, 7 June 2008 (UTC)

We could use the official codes en-AU, en-CA, en-IE, etc. —Angr 08:52, 7 June 2008 (UTC)
That could work. But how would you treat RP and GA with the official codes?
I was also thinking of using just the national codes (like the use of national flags in the tables), such as AU, CA, IE, etc, with GA and RP staying as is. As this article is about the main dialects of English only, the 'E' (in current codes) or 'en-' (in official codes) seam to be redundant. --203.220.171.56 (talk) 13:25, 7 June 2008 (UTC)