Appendix talk:English pronunciation

This page is for the pronunciation key, but how do we go about it with different accents, e.g. American, English, AustralianEdit

"This page is for the pronunciation key, but how do we go about it with different accents, e.g. American, English, Australian??"

Lots of luck. There's no way to determine a correct pronunciation, only what someone thinks is a correct pronunciation. Received pronunciation, which is often held as the standard, represents only a small percentage of native English speakers. The vast majority don't talk that way.

On another subject, I don't see much point to SAMPA, which is simply a corruption of the IPA. With all the IPA-capable fonts that are available today (which can be installed rather easily even on older machines), I can't imagine someone being too lazy to install an IPA font, and yet going to the considerable trouble to learn SAMPA. SAMPA is even more difficult to learn than IPA. It's the worst of all worlds. It contains a bunch of strange symbols that are even less intuitive than the symbols of IPA. The only advantage seems to be that it allows people to use out-of-date software.

If somebody wanted an alternative to IPA, some kind of respelling, which dictionaries have used for ages, would be a lot more useful, especially to native English speakers like myself. (and so begins my campaign to rid Wiktionary of SAMPA :-)) Bluelion 04:13 Jul 7, 2003 (UTC)


I don't think the merit of IPA is to be able to reproduce all the possible accents of English. If somebody reading the wiktionary can get an idea of how to pronounce those English words in any one of the accents and be understood, then it has been worth it to add it, if you ask me. I´m not going to add a placeholder for SAMPA anymore. I don't see the sense of it anymore either. I'm trying to add IPA to the entries I touch, but honestly I don't know whether I'm transcribing them the UK, US, AU, or CA way. I guess it is closest to US pronunciation. Feel free to change them if you see problems with them.Polyglot 07:42 Jul 7, 2003 (UTC)

The SAMPA is useful if you're on a computer where you don't have permission to install fonts such as in internet cafes which I for one use extensively. It's also useful for cutting and pasting directly into mailing list and newsgroup postings amongst other email uses where Unicode encodings often get lost.

A much bigger problem is chosing a particular IPA scheme. Almost every dictionary I've looked at uses a different variaton. For US English the situation is worse because I haven't been able to find an American dictionary yet which uses IPA at all! Each American dictionary rolls its own pronunciation scheme as European dictionaries used to do before the introduction of IPA.

Up to now I've been doing both IPA and SAMPA for my articles plus another I've been calling "AHD" simply because I didn't know enough about pronunciation schemes at the time and that was the first time I came across. Today I've been experimenting with a few articles which put all three representations on one line without naming them. First the "old" style, then IPA, then SAMPA. At first I was using my "AHD" fields only for US pronunciation but since discovering older British dictionaries using similar schemes, I've adapted it a little to work with both.

I think we need to standardise our own "old style" and IPA schemes in a way which can show systematic relationships between RP (British) and GenAm (US) with one system. I have seen dictionaries which achieve this quite well. The best is one designed for learners of English who may be aiming at either accent without having to publish two books.

The scheme illustrated on our "Pronunciation Key" page is a nightmare. I have never seen this scheme in any dictionary. It has left out length distinctions which most dictionaries choose to retain. It shows "bear" and "bait" sharing the same vowel! In what dialect does this occur? It uses both /e/e/ and /ɛ/E/ whereas every dictionary I've ever seen chooses one or the other. It shows /o/o/ (monophthongs) for "boat" and "nose" which actually contain diphthongs in both RP and GenAm. It shows both /oʊ/oU/ and /əʊ/@U/ as representing the same phoneme but gives different examples, thus implying those words have different phonemes! It shows poor and cruel as having different phonemes which I tend to agree with (possibly due to my own accent) but have never seen in a dictionary. It shows the vowel of "more" as a diphthong /ɔə/O@/ which in standard pronunciations on either side of the Atlantic is actually the same as the one in "saw" and "caught" here given as /ɔ/O/. It uses the ligature versions of /dʒ/dZ/ and /tʃ/tS/ (ʧ and ʤ) which the IPA has deprecated (though this is surely the most trivial of the problems I'm discussing here). It distinguishes between /l/l/ and /ɫ/5/ which no dictionary I know does, and sensibly since they are allophones. It contains /ɲ/J/ which is not an English phoneme at all but does occur in Spanish for example. It differentiates two types of /r/ - along dialectical grounds which makes even less sense to me than if it had included separate symbols for the two allophones of /r/ in English. It also incorrectly labels the syllabic consonant diacritic as being the syllable separation mark.

Phew!

What we really need is a table which shows one phoneme on the left side (in /slashes/) and each allophone or dialectical variant on the right side (in [square brackets]).

Another page illustrating the various systems used by the most popular dictionaries and relating them to our system would be very useful and I'll start work on it soon. Now where should such articles go in Wiktionary? Hippietrail 11:39, 26 Feb 2004 (UTC)

Hi Hippietrail,
I understand you have issues with the phonetic transcriptions. Please keep in mind that we don't only need phonetic transcriptions for English and all its accents. What I am most interested in is the transcriptions for other languages, so we certainly need to retain a way to transcribe sounds that don't occur in English.
The reason why it probably has become a mess is because we don't tend to copy dictionaries. I don't even consult them for phonetic transcriptions. So we used the information we could gather at various locations on the internet and in Wikipedia. That probably explains why we have deprecated symbols in there.
If you see a way to improve the situation, we would all appreciate it, if you want to have a go at it. Please bear in mind we also want to transcribe the phonetics of other languages.
The page you want to create could go into an appendix. Did you have a look at this page: Wiktionary_Appendix:IPA_Examples? I expanded that because it seems more useful to me. (I was trying to learn IPA with it) Polyglot 12:07, 26 Feb 2004 (UTC)

I suggest include machine parsed pronunciations in the definitions. You can follow the http://dictionary.reference.com/search?q=GNU example. No files stored, the server would offer to the user the pronunciation online.

Help with AHD?Edit

  1. u (en-us) boot, soon, lose u o͞o
  2. u into u o͞o
  3. uː (en-gb) boot, soon, lose u: o͞o
  4. ʊ put, foot U o͞o

From the first table, the above section appears. Where did those four "o?o" items come from? Shouldn't they all be "ōō"? Or is that syntax that dictionary.com recently added/changed? (They are aparently requiring a subscription to access pronunciations now.) --Connel MacKenzie 05:33, 20 Feb 2005 (UTC)

#1, #2, #3 should all be o͞o, which doesn't render for me, but ought to look much like ōō except with one long macron instead of two short ones. #4 should be o͝o, which also doesn't render, but ought to look like ŏ but with one long breve extending over two o's.
Oh, OK. That satisfies my curiosity, but I guess the rendering problem will remain for some time. --Connel MacKenzie 15:16, 20 Feb 2005 (UTC)
As a side-note, we shouldn't really call it "AHD" because ours a bit different to the American Heritage Dictionary's scheme. When I first started using this scheme I called it AHD in the belief that all American dictionaries used a standard system invented by AHD - this turned out to be wrong. I would prefer to call it something like "American dictionary style" but that's way too much of a mouthful ): — Hippietrail 13:19, 20 Feb 2005 (UTC)
Is it time to pick a proper term for this then? "ADS" might work. "ADPS" American dict. pronunciation scheme? --Connel MacKenzie 15:16, 20 Feb 2005 (UTC)
I think it probably is time. I should point out that while I was inspired by American dictionaries, it doesn't show just American pronunciation any more, particularly in the ways it handles optional rhoticity and unstressed final -i, but also the optional "y" sound before many "u" sounds. People may get the idea it's used just for American pronunciations if we base a name on that. I'm pretty sure that British and other dictionaries all used similar(ish) schemes before IPA was widely adopted in Europe and Commonwealth countries. Maybe "Pre-IPA", "Non-IPA", "Old Style"? It would nice to hear some other opinions.
Another issue I have is with introducing our own local jargon to Wiktionary. IPA and AHD are already pretty well known initialisms and so far we haven't invented any of our own. We would probably need some kind of glossary somewhere to explain such things. — Hippietrail 00:08, 21 Feb 2005 (UTC)

All kinds of problemsEdit

I've just got rid of the inappropriate "AHD" labels which were a legacy of my own poor understanding of how American Dictionaries do pronunciation guides. But there's a lot more wrong with this table.

Many dictionaries use variations other than in this table. At the same time there are other variations in this table which I do not know from dictionaries. If we're setting a standard set for people to use, we need to cull a bunch of symbols. If we're trying to describe what symbols are actually used so readers can interpret them, we need to include the missing symbols.

  • bed, bet → many dictionaries use /e/,/e/ rather than /ɛ/,/E/
  • fur, bird (US) → some dictionaries use /əːr/,/3:r/ rather than /ɝ/
    I'm pretty sure IPA /ɝ/ maps to SAMPA /3`/ and SAMPA /3/ maps to /ə/
  • boat, nose → I've never seen a dictionary which uses /oʊ/,/oU/.Actually I have seen /oʊ/,/oU/ as well as /o/,/o/ (and possibly /oː/,/o:/).
    Why is this symbol in the "Vowels" section and the "Diphthongs" section?
  • /tS/ and /dZ/ → The IPA ligatures for these symbols, "ʧ" and "ʤ" are deprecated and thus should not be given as primary. They should probably be in brackets.
  • /ɡ/,/g/ → The correct IPA symbol is "ɡ" which only has one variant unlike the usual non-IPA symbol which has two variants depending on the font. We should note this because it's not obvious.
  • /ɫ/,/5/ Isn't used in any dictionary I've seen. It's for narrow transcriptions used by linguists and maybe some advanced grammars.
  • /ɱ/,/F/ I've never seen this in English at all - certainly not in a dictionary!
  • /ɲ/,/J/ Is not an English phoneme. It exists in Spanish and other languages. Maybe some bilinguals might use it in English but nobody else.
  • /r/,/ɹ/,/ɻ/ → What a nightmare! Broad and narrow symbols all mixed up and/or separated at random - and the symbol for the "tap" version used by Americans and Australians between vowels is left out entirely! Some dictionaries may use something other than /r/ but it is by far the most common. In any case, we only need one symbol for English at the phonemic level. If we are describing the symbols a reader might encounter then we need also the missing ones and a much better explanation of what they all mean.
  • /ʍ/,/W/ → This is not British English. There are varieties inside and outside of Britain and America who use it, and who don't use it. I'll link the Wikipedia artice here when I do some more research. Here are a couple but there may be others: w:H-cluster reductions, w:Voiceless labial-velar fricative
  • /x/,/x/ → I need to check whether /χ/ is more appropriate. Only some dictionaries use it so I need to do some research.

Hippietrail 02:47, 30 May 2005 (UTC)

I'm now building a collection of IPA schemes used by actual print dictionaries over at User:Hippietrail/IPA/English. We can use that as a reference when deciding what to do here rather than relying on our memories and expectations. None of the big famous dictionaries are there yet and the format is still in flux, but it's a start. — Hippietrail 10:08, 31 May 2005 (UTC)
I agree with almost all of Hippietrail's comments, with a couple of exceptions listed below. A lot of my take on the best way of doing IPA on Wiktionary is purely cosmetic -- i.e., I have yet to use a computer (including my own personal mean lean IPA-slinging machine) where the higher-numbered IPA characters on Wiktionary don't come out as butt-ugly and often completely illegible. My arguments to avoid many of them don't reflect the way I personally prefer to transcribe (and not even the way I teach my students to).
  • on /e/ for the e of bed: yeah, some use it. Others use /e/ for the ay of day. Using /ɛ/ and /eɪ/ is unambiguous.
  • /g/ vs. /ɡ/. Actually the IPA gave up on this battle (in the 1990s, I think). Either is now OK. I personally prefer the specific guaranteed IPA /ɡ/. But in this case, my personal preferences would lead to butt-ugliness and illegibility.
  • ditto Hippietrail on [l̴], [ɱ], [ɳ]
  • mostly ditto Hippietrail on the Rs. We should have only one one. Much as I hate it, right-side-up /r/ is [Keffy grits his teeth] the least ugly here and the most likely to actually get used. I would even make the same arguments [gnashes his teeth wildly] for /ər/ over the more theoretically sensible /ɜ˞/, /ə˞/, or /ɹ̩/.
Keffy 00:26, 21 February 2006 (UTC)
What would be cool is a tag that would change SAMPA into IPA much like the (much crazier) <math> tag at Wikipedia. For instance, the code
SAMPA: /"r@Ubət/, /"r@U%bQt/
IPA: <sampa>en:/"r@Ubət/</sampa>, <sampa>en:/"r@U%bQt/</sampa>
is the SAMPA for robot, both U.K. and U.S., followed by the IPA as an image. Then there wouldn't be problems with people having to install certain fonts to see the characters, and this debate about which symbols to use, should consensus change, could be resolved immediately.
Alternatively, an X-SAMPA tag would fix the font problem, but wouldn't be as flexible on the other issues. Davilla 12:57, 21 February 2006 (UTC)
In an ideal world we would have support from the Wiki developers for some much-needed features that apply to dictionaries but not encyclopedias. Then we could have a user setting for which was preferred, an extension which converted IPA into a .png, etc. At the moment it is possible to let the user choose by creating templates that work with CSS and Javascript but there is no way to provide a setting meaning the user would have to choose on each and every page )-: The problems with images is you can't cut and paste them as text, the problem with text is that fonts are needed.
Another solution is for us to find a free font with IPA, prominently link to it and recommend users to install it, and put it in our IPA templates. A version of this solution is to use dynamic fonts but I believe they only work on Microsoft browsers - I don't know a lot about them. — Hippietrail 17:16, 22 February 2006 (UTC)

Missing ôEdit

I can't find the ô used in the pronunciation of 'door' in the English Pronunciation key. Could perhaps someone with a better understanding of these matters add it if it should be there?

Thanks! Erl 06:50, 3 May 2006 (UTC)

It's there now Ishwar 04:25, 15 May 2008 (UTC)

enPR transcription for [ʊə]Edit

I have been using [o͝ə], being as [o͝o] is the equivalent of [ʊ] and [ə] the equivalent of [ə]; however, this is less than ideal. Does anyone else have any suggestions as to which character we ought to use? † Raifʻhār Doremítzwr 12:11, 10 September 2007 (UTC)

Can you give an example of a word with this sound? (And keep in mind that enPR is intended to be transdialectic, such that enPR renditions of US and UK pronunciations should normally be the same.) —RuakhTALK 14:56, 10 September 2007 (UTC)
See cruel, wherefor both the UK and US pronunciations would retain the [ʊə] diphthong. † Raifʻhār Doremítzwr 23:24, 10 September 2007 (UTC)
This is wrong. That diphthong is not in the word cruel. cruel starts with the tense vowel [u] (like in the word shoe. RP has the diphthong while r-ful American dialects have [ʊr] = [o͝or]. Ishwar 04:24, 15 May 2008 (UTC)

Italics can be inaccessibleEdit

Using italics to differentiate th and th (for /θ/ and /ð/) may be an accessibility problem.

The distinction is irretrievably lost if the text is saved or copy-pasted as plain text (in any charset, including Unicode and ascii). It may be lost in text-only web browsers, mobile browsers, or other alternative devices, possibly including various assistive devices used by the handicapped.

The use of italics may be further problematic because wiki italics generated from dual apostrophes are rendered as an HTML <i> element, which is deprecated in some versions of HTML, and absent in strict XHTML. They may disappear or be transformed unpredictably if the data is moved into some systems which retain format.

We might consider borrowing a representation from another system than AHD. There are about several shown at w:Pronunciation respelling for English#Traditional respelling systems, including ð th̸ t̷h th: dh which require no formatting, and TH th [th] which have underscores.

It's possible that KH for the "ch" in "loch" has similar problems, but at least it is differentiated by capitalization, and set off in a <sup> element which is not deprecated (the nested <small> is deprecated). —Michael Z. 18:12, 27 March 2008 (UTC)

The <i> element exists in Strict XHTML 1.0 (see http://www.w3.org/TR/xhtml1/dtds.html#dtdentry_xhtml1-strict.dtd_i) at least. But, point taken. I think "dh" sounds good, personally. —RuakhTALK 23:14, 27 March 2008 (UTC)
I also think that dh is best. Mostly because it is easy to type. However, it is not a common symbolization, so perhaps there is some objection there. Definitely the italics is not preferred for the obvious fact that it is a stylistic modification of the font and not a difference in symbolization as far as the assignment of Unicode glyphs is concerned. Ishwar 04:20, 15 May 2008 (UTC)
I've seen "dh" used in a few places, but I think I've seen "TH" more often. --EncycloPetey 01:06, 17 January 2009 (UTC)

nasal vowelsEdit

These are necessary for Canadian English. IPA examples from the CanOD:

The simplest representation might be n in brackets – brāyô(n), kănădyă(n) frä(n)glā. Superscript n works nicely, but would get lost in a text-only copy – brāyôn, kănădyăn fränglā – unless we used the Unicode version – brāyôⁿ, kănădyăⁿ fräⁿglā. Michael Z. 2009-01-17 00:55 z

Using a parenthetical (n) may not be a good idea. A parenthetical symbol is usually regarded as optional for the pronunciaiton. I recommend adopting Webster's method, which is to italicize the n, which would still show up in a text-only copy. --EncycloPetey 00:57, 17 January 2009 (UTC)
Or. hmm... italics seem to be a point of debate in the above section. Did we decide they were a genuine problem? It is possible to use the floating tilde, as: kănădyă~ isn't it? --EncycloPetey 01:01, 17 January 2009 (UTC)
Yeah, right about brackets. The floating tilde works, but it looks like something from ASCII-only Sampa, and out of place in a scheme emulating the typographic standards of dictionaries.
Italics aren't unacceptable, but using them means that pronunciations may or may not survive reduction to text-only, or being copy-pasted, depending on the software. And after such an operation the failure wouldn't be easily detectable, so the results may be misleading. I'd much rather see this implemented in Unicode text.
Unicode does include a selection of superscript and subscript letters, allowing kănădyăⁿ (used by MWCD), kănădyăᴺ, or we can use capital N which will survive textifying, with the addition of style to make it look better—small capital N is used by several dictionaries, including AHD: kănădyăN, kănădyăN, or kănădyăN (the last form looks best, but does MSIE support the CSS?).
By the way, it's not just Canadian English which requires this; cf. lingerieMichael Z. 2009-01-20 17:37 z

/nɡ/Edit

If /ŋ/ as in singer is represented by 〈ng〉, then how do we represent the more clearly pronounced /ɡ/ in longer? Michael Z. 2009-01-17 01:00 z

I would use enPR: longˑgər. --EncycloPetey 01:03, 17 January 2009 (UTC)
That makes sense—I can't think of any situation where this doesn't fall on a syllable break. Come to think of it, just “lônggər” would work. Michael Z. 2009-01-20 16:42 z
The vowel is actually lŏnggər. kwami 11:16, 29 December 2010 (UTC)
Not among Americans who distinguish cot and caught. They have [ɔ], not [ɑ], before [ŋ] in almost all native English words like long, strong, wrong, etc. These words belong to Wells's CLOTH lexical set of words that have the THOUGHT vowel in (cot-caught-distinguishing) General American but the LOT vowel in RP. —Angr 12:11, 31 December 2010 (UTC)

Mary (e) Marry (æ) Merry (ɛ) Murray (ə)Edit

While many American dialects conflate the vowels in Mary Marry and Merry, not all do. My own Delaware Valley dialect distinguishes Marry (lax æ) and Mary (tense e) from Merry, but then, unlike Western dialects, rhymes Merry with Furry (ə) and Murray (ə). In rhotic dialects in New York City all four rhotic vowels are quite distinct, and the Merry vowel (ɛ) is the lax vowel of bed while the vowel of Mary (e) is tense. Given that I rhyme Merry and Murray (ər) I also rhyme Ferry with Bury - and the laughter that errupted when I spoke of the "Staten Island Furry" instead of the Staten Island Feh-rry" (which I had to learn to pronounce) makes me quite aware of the distinctiveness of these vowels and their disticness from the tense /e/ of Mary. The article does not reflect this four way distinction of mid-front rhotic vowels which is indeed a retention of the primitive situation, not an unconditioned split. The American vowel in hair pear there scary is most certainly a tense and not a lax one for any dialect I know of whether it distinguish two, three or four of the mid-front rhotic vowels. The lax one is found in berry, ferry and merry for those like rhotic New Yorkers who do distinguish it from Marry and Mary and Murray as well. I hope someone who recognizes this and who has a better grasp of the situation in RP can correct the page.

I don't pretend to know anything about US pronunciation, but in the non-rhotic English accent I speak (largely northern but slightly diluted with nearly 20 years of Somerset) and the various accents I hear, the vowels in the words that you mention are as follows:
Mary Hair Pair There Scary Marry Merry Ferry Bury Murray Furry Fury
ɛə a ɛ ʌ ɜː ʊ
so you can see that I also rhyme "ferry" with "bury", but not "merry" and "Murray". What the charts don't recognise is that there is a large difference in some pronunciations between northern and southern English, moor and is a good example, in southern England it is the monosyllabic /mɔː(ɹ)/ while in northern England it's /mu.ə(ɹ)/ (homophonous to mooer(one who moos)). Thryduulf (talk) 09:57, 19 June 2010 (UTC)
The comments above are mistaken – all contrasts are on the page. Mary = mâri, marry = mări, merry = mĕri, Murry = mŭri, furry = fûri. The pronunciation key does not show dialects that have merged these sounds.
One thing that is not on the page is the rare New England Short O (as in stone), which is a retention of a Middle English vowel that most dialects in the US and England have merged with another vowel. A similar retention is in East Anglia. The Kenyon & Knott pronunciation guide to American English lists this unmerged phoneme. I doubt any widely used modern American dictionary lists the words. Ishwar 22:23, 16 December 2010 (UTC)

ZEdit

Isn't there a difference of the pronunciation of z in zoo vs quiz? I thought zoo has a more vocal z whereas quiz is closer to a sharp s. I'm not a native English speaker so not sure if that's correct, please advise. --92.231.211.222 03:04, 28 January 2011 (UTC)

There's probably a phonetic difference, at least in some contexts, but phonemically speaking, they're the same sound. —RuakhTALK 04:08, 28 January 2011 (UTC)
I imagine you speak a language which, like most Germanic, Slavic, and Turkic languages, doesn't allow voiced sounds like z, v, b, d, g at the ends of words? In English, the z of 'quiz' is essentially identical to the z of 'zoo', or even closer to the z of 'zip' (the vowel has some effect). kwami 07:27, 5 February 2011 (UTC)

ăEdit

I'm having a hard time believing RP and GA would pronounce ă the same way. Shouldn't the American sound be wider and tighter than the British? -- Ke4roh 12:37, 13 June 2011 (UTC)

General American pronunciation is definitely IPA(key): /æ/; my understanding is that the Received Pronunciation is also IPA(key): /æ/, but that current posh British usage tends toward IPA(key): /a/. (Your comment seems to treat "RP" and "British" as synonymous, but they're really not.) —RuakhTALK 13:49, 13 June 2011 (UTC)
Nothing posh about it -- pretty much all southern accents use /a/ now, and /æ/ sounds (unless in American contexts) horribly old-fashioned. Ƿidsiþ 09:31, 9 May 2012 (UTC)
Right, but I think that in non-posh accents that's been true for a long time. I was just saying that current posh British usage is different from the "Received Pronunciation", that is, from the posh pronunciation of a hundred years ago. (But I see that w:Received Pronunciation claims that "RP" denotes the current standard, whatever that might be, so I guess my understanding of the term was off?) —RuakhTALK 17:18, 5 June 2012 (UTC)
I think that both accents have moved away from [æ]. Most British people now say [a] (this was always the case outside the south-east). Most Americans say [ɛə] or [ɛ:]. Dictionaries are slow to update. Clive Upton used [a] in the Oxford Dictionary, but this has not caught on across the world of linguistics. 94.0.180.226 13:44, 31 October 2012 (UTC)
I think you're misinformed about the American pronunciation. Most Americans use [æ], while some parts of America use [ɛə]/[ɛ:]/[eə]/[ɪə]. --WikiTiki89 14:01, 31 October 2012 (UTC)

Poor and tourEdit

Any reason 'poor', 'tour' and 'tourism' are listed as having the same sound in the IPA vowels list? Chicken7 (talk) 11:52, 2 June 2012 (UTC)

See Rhymes:English:-ɔː(ɹ). I hope that's enough explanation. Mglovesfun (talk) 12:38, 2 June 2012 (UTC)
I for one have the same vowel in 'poor', 'tour' and 'tourism', and it's not the same vowel as in 'pour' and 'tore'. —Angr 13:47, 2 June 2012 (UTC)
Ditto. —RuakhTALK 16:52, 2 June 2012 (UTC)
'poor' and 'tour' have different sounds in American English. 'poor' has the same pronunciation as 'pore'. GameKyuubi (talk) 04:24, 20 January 2014 (UTC)
For some Americans, yes, but by no means for all. —Aɴɢʀ (talk) 08:18, 20 January 2014 (UTC)
And for many, "tour" is the same as "tore". The "oor" sound is unstable in some dialects, and is merging with "ore". kwami (talk) 20:18, 20 January 2014 (UTC)
Except after palatal and palato-alveolar consonants, where it's merging with "err" (Surely you're jokingI'm not joking, and stop calling me Shirley). —Aɴɢʀ (talk) 20:39, 20 January 2014 (UTC)
Except in some dialects where even after palatals it is merging with "ore", where "sure" sounds like "shore" and "pure" is pronounced like "pyore". --WikiTiki89 20:46, 20 January 2014 (UTC)

ōrEdit

In the section for ōr, it gives both oɹ, ɔɹ for General American yet only ɔə(ɹ) for RP. There must be very, very few people left in England who use [ɔə] for FORCE words. John Wells says in this essay that NORTH and FORCE merged in the early 20th century. I can't think of any notable English people who have [ɔə] here. I suggest that [ɔ:] be placed in the box for RP, at least as an alternative. 94.5.242.218 18:37, 26 September 2012 (UTC)

Audio is missingEdit

Hi
Unfortunately the whole article is useless unless somebody adds the audio to it.
Just imagine anybody who doesn't know English.
How is that somebody supposed to learn from this article?!
--Jangirke (talk) 08:40, 13 October 2012 (UTC)

How could someone who doesn't know English use this website at all? —Angr 07:08, 14 October 2012 (UTC)
Hmm, reading English is fine for me since I'm 15 (2010) and writing it gives no more problem since I'm 17, but my pronunciation is still amazingly shitty (i rrehporrt yu, huehuehue brbr), and I can't really watch English TV shows without subtitles unless it is in very clear BBC English or it is a kids' show targeted at an international audience. The keys aren't a problem for me since I'm fluent in IPA, and Wikipedia helped me a lot to understand English phonology, but many people (perhaps most) won't get it. And as usual for wikiprojects, the English one has the best quality (or at least the best quality you can get in a language easy to master), so I learned English exactly for using them, as I am terrible at using searching engines to acquire knowledge (I think - with strong evidence, such doubt that I share with my parents - I have ADD, Asperger's or both, never told fellow Wikipedians before leaving out suddenly because that is no excuse) and the Portuguese ones many times let me down. 177.65.49.210 09:02, 21 January 2013 (UTC)
Angr, it's generally not that hard to browse around as a foreign speaker. You can easily find this page with only a rudimentary knowledge of written English, and Google Translate and the likes make it even easier. But even without regard to people who don't speak English (well), some audio wouldn't hurt.
Peter Isotalo 06:57, 22 January 2013 (UTC)
I'm also surprised that none of the sounds include an audio file. I know it's hard to do for every possible sound especially as some are controversial but at least do it for standard British English. Providing examples doesn't always help and non-native speakers reading this will only be helped if they know the words given as examples or if the equivalent IPA symbol is also used in their L1.--Xania (talk) 23:09, 12 January 2015 (UTC)

Does SAMPA still belong here?Edit

Last month's "news for editors" includes

27: SAMPA and X-SAMPA transcriptions are no longer included in pronunciation sections (vote).

So should this page still include SAMPA? If it stays in, at least there should be a prominent note to that effect ... IMHO. IANAn experienced Wiktionarian, so I will not make any such change myself.

I've taken the question to the Beer parlour. --Thnidu (talk) 21:02, 26 December 2013 (UTC)

Received PronunciationEdit

I'm a little confused by the vowels in city in the pronunciation section. It seems to use the same figure for both in RP, even though they are clearly pronounced differently. Can anyone explain why this is so? cwbr77 (talk) 16:43, 19 February 2014 (UTC)

In more old-fashioned RP, the vowels are the same. This is sometimes still represented in phonetic transcription, though ideally that should be secondary to showing the more up to date pronunciation; in other words, we should write IPA(key): /ˈsɪti/, /ˈsɪtɪ/ for the RP pronunciation. —Aɴɢʀ (talk) 16:48, 19 February 2014 (UTC)

dead linkEdit

please erase dead link to kynesthetic wheel in See also section. 77.27.97.158 16:12, 10 March 2014 (UTC)

PALM vowelEdit

I've added palm beside father to aid Irish readers. John Wells chose PALM as the lexical set keyword because many Irish English speakers have the THOUGHT vowel in "father". See here. Jnestorius (talk) 18:55, 20 September 2014 (UTC)

Dead LinkEdit

The link to the IPA 2005 chart PDF has moved to to here

Thanks, fixed. —Aɴɢʀ (talk) 18:25, 7 October 2014 (UTC)

Vowel of rosesEdit

How should we be representing the unstressed vowel of roses? Wikipedia uses a diaphoneme /ɨ/, which avoids the problem of having to write out separate forms with /ɪ/ for UK, /ə/ for Australia, and both /ɪ/ and /ə/ for the US. The two vowels show considerable variation, because it's not a particularly important distinction to make. See the problems this causes at s' for example, where the respelling and the IPA combined with context-specific forms and alternate pronunciations becomes a real mess. — ˈzɪzɨvə (talk) 17:25, 9 June 2015 (UTC)

We have been using both /ɪ/ and /ə/, depending on the preference of whoever is adding or modifying the pronunciation. I know we should probably be more consistent, but that is how it currently is. --WikiTiki89 17:56, 9 June 2015 (UTC)

IPA mark for secondary stressEdit

secondary stress (or sometimes tertiary stress), as in battleship /ˈbætl̩ˌʃɪp/

“Battleship” may not be the best example here as the diacritic to indicate a syllabic consonant, ◌̩, is very similar to the symbol to be demonstrated, which immediately follows it. Can we find a better word to demonstrate the secondary-stress mark? BlaueBlüte (talk) 23:45, 11 October 2015 (UTC)

  Done Changed to "battlefield". Equinox 23:46, 11 October 2015 (UTC)
Hang on, I see the problem you mean now. Will change again... Equinox 23:47, 11 October 2015 (UTC)

AuE and NZEEdit

I merged more of the cells for AuE and NZE. Several of the transcriptions are basically spelling the same sound: /æɪ/ and /æe/, /aː/ and /ɐː/, for instance. w:Australian English phonology and w:New Zealand English#Phonology give no indication that these vowels differ. Best to use the same transcription for both so as not to mislead. However, perhaps /ɐː/ could be chosen instead of /aː/.

If someone disagrees, go ahead and revert me and explain here. Presumably there was a rationale behind the previous choice of symbols. — Eru·tuon 23:05, 10 May 2016 (UTC)

Phonology vs phonetics for EnglishEdit

This may have been discussed elsewhere but I can't find it. If that's the case, I will be happy about pointers to that discussion.

In Wiktionary, in the pronunciation of almost all languages except English, we use a phonological representation of the pronunciation, and an appendix page to explain the chosen phonology, and eventual phonetic variations. For instance, the French word faire is transcribed as /fɛʁ/, and the page Appendix:French pronunciation explains that /ʁ/ can be pronounced in quite a few different ways. For English however, we seem to show always multiple pronuncations depending on the region, e.g. boat is shown as both /bəʊt/ and /boʊt/, for Received Pronunciation and General American. My question is: Why don't we use a phonological notation also for English? This would be easily feasible, for instance by following Wikipedia's system (https://en.wikipedia.org/wiki/Help:IPA_for_English). The system used by Wikipedia is able to represent English pronuncation faithfully as far as I can see, and also explains regional pronunciations. Is there a reason I'm missing, or is this something that is planned eventually? –Jérôme (talk) 12:09, 14 October 2016 (UTC)

I support a more unified approach to English pronunciation, but others don't seem to agree. See this discussion from a while ago. --WikiTiki89 15:56, 14 October 2016 (UTC)
@Jérôme: Wikipedia's system isn't phonemic; it's diaphonemic. It ignores the significantly different vowel systems of RP and GA (not to mention all the other dialects listed in the International Phonetic Alphabet chart for English dialects). Each symbol doesn't represent a phoneme, but rather a set of phonemes from each dialect that generally correspond with each other. Since you are making a proposal like this, you should understand that you are proposing a diaphonemic transcription system; Wiktionary already has phonemic transcription systems. It's frustrating that the Wikipedia system misleadingly uses slashes / /, which should mark phonemic transcriptions, when it should be using some other symbols to show that the system is not phonemic. — Eru·tuon 16:43, 14 October 2016 (UTC)
I don't believe that there really is much a difference between diaphonemic and phonemic. Our "phonemic" representations are meant to represent multiple sub-dialects as well. --WikiTiki89 16:53, 14 October 2016 (UTC)
I would very much prefer the diaphonemic approach and give users a choice in their preferences which rule-derived varieties to show them by default. Alternatively or additionally such rule-derived narrower transcriptions could also be chosen from a list of varieties (a pull-down menu?) on the current page. This approach would also be open to extension to less frequent native as well as non-native accents, especially accents used by second language speakers, but maybe also accents in use by foreign language speakers.
According to Wikipedia English is currently official in 67 countries, 27 non-sovereign entities and it is used word-wide, so it is no longer the language of only what are traditionally perceived as native speakers. —LiliCharlie (talk) 17:33, 14 October 2016 (UTC)
@Wikitiki89: Is that really true? The symbols labeled US might describe some variation, but not all. They don't describe, for instance, pronunciations where there's a phonemic split between tensed and lax vowels in can (verb) and can (noun), or where there's still a distinction between father and bother. It seems like the US symbols would be more correctly labeled General American. — Eru·tuon 17:59, 14 October 2016 (UTC)
@Erutuon: Are you talking about the symbols labeled US at Wiktionary or at Wikipedia? The main reason for the diaphonemic transcription at Wikipedia is to save space. Wikipedians wanted to provide pronunciation information for certain entries, but didn't want the lead of an article taken up with separate pronunciations for RP, Scottish, American, Canadian, Australian, and so on, especially since as an encyclopedia their focus is on the meaning of the term (and its history, cultural relevance, and so forth), not on lexical properties of the word itself. They wanted readers to be able to easily skim the pronunciation info and get to the meat of the article. But since Wiktionary is a dictionary, the lexical properties of a term are what we focus on, so it is no distraction for us to list multiple dialects' pronunciations of a word in its pronunciation section. —Aɴɢʀ (talk) 20:03, 14 October 2016 (UTC)
@Angr: I was referring to the GenAm symbols on this page, which are the basis for most of the pronunciations labeled US in Wiktionary entries. (Oops, they are not labeled as US in the Appendix, which is a good thing.) — Eru·tuon 20:08, 14 October 2016 (UTC)
Yes, many entries erroneously use {{a|US}} where they should use {{a|GenAm}}; likewise many entries use {{a|UK}} where they should use {{a|RP}}. I try to correct these wherever I encounter them, but it's like cutting off the Hydra's head: for each one I correct, nine more seem to spring up. —Aɴɢʀ (talk) 20:13, 14 October 2016 (UTC)
That's because there is no consensus about that. --WikiTiki89 21:06, 14 October 2016 (UTC)
Surely there's consensus that {{a}} is only to be used with accents, isn't there? There's no such thing as a single UK or a single US accent. —Aɴɢʀ (talk) 22:00, 14 October 2016 (UTC)
I agree, though I think instead of cutting off the Hydra's head a bot should be used to correct all instances of {{a|US}} and {{a|UK}} (as well as {{a|GB}}). —LiliCharlie (talk) 14:30, 17 October 2016 (UTC)

Hi everyone – I'll try to summarize the current situation, and my opinion. We are showing multiple pronunciations (e.g., GenAm, RP) for each English words, because a single system to represent all pronunciations would be too high-level, i.e., it would be diaphonemic. My question now is: Do we define anywhere the list of phonologies that should be used? For instance, we could define that every word has to have its pronunciation shown in GenAm and RP. In particular, I would then strongly suggest to list the phonemic transcriptions that should be used for all words. On the other hand, a solution would be to adopt a higher-level transcription, in which each word is transcribed only in a single way. I.e., we would use something similar to what Wikipedia uses, but it wouldn't need to be identical. I am of the same opinion as WikiTiki89, that having multiple phonemic transcriptions is counter-productive. That is, in almost all cases the GenAm and RP pronunciations can be derived systematically from a diaphonemic transcription without loss. Now, the main argument agains a diaphonemic transcription seems to be that the individual variants of English differ so much that the diaphonemic transcription would be misleading, if readers would interpret its IPA symbols litterally. That might indeed be a problem, but, in my opinion, only a minor one: It is the same problem that also any phonemic transcription has. I.e., the French transcription we use at the moment can be equally misleading, in that /ʁ/ may suggest a specific pronunciation of that phoneme. (On Wikipedia, I like a lot that you can hover the mouse pointer over a symbol and it shows an explanation such as "/m/ as in 'mouse'".) My suggestion for Wiktionary would thus be: Settle on a diaphonemic system (similar to the Wikipedia one or not), and additionally show GenAm and RP pronunciation if wanted. If editors add only a single pronunciation, I would however recommend that they write a diaphonemic one. BTW, I don't think "people have it easier to write down their own variant of English" is a strong argument; it would only be if people had an intuitive idea of the purely phonetic values of IPA symbols, but most people writing down the pronunciation of English words (native or not) will be looking up the correct symbols in the table of this page anyway, so the correspondance or not to the "pure" IPA value is moot–we should document on this page the main diaphonemic system used.–Jérôme (talk) 13:58, 17 October 2016 (UTC)

One more thing: As Wikitiki89 and LiliCharlie (and myself) favoor a diaphonemic system, there is still one detail in Erutuon's argument that is important: Some dialects of English make a distinction between phonemes that is not made in most others, such as the can/can example. We may solve this problem in two ways: (1) Use a diaphonemic system that includes such variations, or (2) Don't include these variations in the main diaphonemic system, but show them in additional dialectal pronunciations. In think (1) would interfere too much in most cases, and would recommend (2), at least for the given example. In general however, I think individual distinctions should be discussed. In any case, the diaphonemic system chosen should be given on this page, and it will make clear which distinctions are made and which not.–Jérôme (talk) 14:06, 17 October 2016 (UTC)

ū for /juː/Edit

The sound /juː/ is currently transcribed as the enPR:yōō. But this sound is commonly known as the long version of the u vowel; in addition, the dialectal variations of the "sole" ōō (RP:/uː/, GenAm:/u/, AuE:/ʉː/, WaE:/uː/) are distributed slightly independently of the dialectal reproduction of the "composed" yōō (RP:/ju̟ː/, GenAm:/u/, Aue:/jʉː/, WaE:/ɪu/).

I propose to use the letter u with macron (ū) for this sound. Thus the word "mute" (in which the silente-e lengthens the preceding vowel), currently transcribed as "enPR: myōōt" in its entry, would be transcribed as "enPR: mūt", as thus reflects both the development of the historical long-u, and the slightly independent distribution of yōō and ū.

PS. This suggestion also applies to the rhotic version of this two diaphonemes. E.g., ūr may be reproduced as /jɝ/ in GenAm and as /jʊə/ in RP, while oor may be reproduced as /ʊɹ/ in GenAm and as /ʊə/ in RP. 179.214.43.190 22:16, 15 October 2016 (UTC)
Return to "English pronunciation" page.