Help talk:IPA/English/Archive 14

This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Archive 10

←

Archive 12

→

Syllabification - why is it used sparingly?

Based on one of the footnotes in the current version of this page:

Syllables are indicated sparingly, where necessary to avoid confusion, for example to break up sequences of vowels (moai) or consonant clusters which an English speaker might misread as a digraph (Vancouveria, Windhoek).

It seems that our preference is to not use syllabification. I generally find in most dictionaries that syllabification is present even where it's fairly obvious where to make the syllables. This seems like a good policy to me, since I see no downside to indicating syllables even when it's obvious where they should go, and as it is, users need to make an assumption about what the "intuitive" way to break up a word would be. What is the harm in adding syllabification to all IPA pronunciations? 0x0077BE [^talk/_contrib] 15:14, 27 October 2014 (UTC)

It can be an area of contention (e.g. is it Cali-fornia or Calif-ornia; blend-er or blen-der, etc.) where, if I recall correctly, even linguists may not all agree on the correct answer. It also doesn't normally add any new information. — Æµ§œš¹ _{[ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ]} 19:16, 27 October 2014 (UTC)

It would only add no new information if it's a one-syllable word. It may be information that people already know, but most people already know what a basket is without looking it up in the encyclopedia/dictionary, and we've got an entry for that and a million other mundane things. It seems pointless to avoid using a richer format if it's an area of contention when there are other solutions to that, such as including a section on the pronunciation if it's so contentious that linguists are debating how to represent it or if varies so widely that it can't be covered easily in the lead (see Missouri#Etymology_and_pronunciation). The syllable information is included in most dictionaries that I've seen, so presumably citations for this information is available (and it can be eschewed if no citation is available and there's some contention anyway). 0x0077BE [^talk/_contrib] 19:28, 27 October 2014 (UTC)

There's also the issue of compression. For instance, according to the Longman Pronunciation Dictionary, the RP pronunciation of raspberry can be either /ˈrɑːz.bər.i/ or /ˈrɑːz.bri/. How would you indicate that without transcribing the word twice? If you don't transcribe syllable breaks, you can simply write /ˈrɑːzb[invalid input: '(ə)']ri/ Peter238 (v̥ɪˑzɪʔ mɑˑɪ̯ tˢʰoˑk̚ pʰɛˑɪ̯d̥ʒ̊) 19:42, 27 October 2014 (UTC)

How about /ˈrɑːz.b(ə)r(.)i/? Alternatively, we could have a guideline that says that syllabification is optional and can be eschewed in cases where it is too complicated for a simple representation but not complicated enough to merit a separate section. All I'm seeing is edge cases where syllable representations would be difficult, not a reason why they're undesirable in well-cited, straightforward cases. And, again, the standard in modern dictionaries is to include syllables, and they've got much less space to work in than we do. 0x0077BE [^talk/_contrib] 19:58, 27 October 2014 (UTC)

Your basket example doesn't really hold water. Yes, there is information contained in what you might call narrower transcriptions, but we eschew narrow transcriptions for English pronunciations. This is why we don't encode for aspiration, glottalization, vowel lengthening, etc. Syllabification is normally one of those things that we don't need to mark because readers can get to the intended pronunciation (which is what our transcriptions are intended to convey) without them. — Æµ§œš¹ _{[ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ]} 20:27, 27 October 2014 (UTC)

If you had read the article, you'd know that baskets are not intended to store water - they are made of woven materials and are thus porous! But in serious response, I'm still not seeing why adding admittedly marginal but not-confusing information should be done "sparingly" when it is done by all the major dictionaries. People wouldn't know what to do with aspiration and glottalization information, but they can handle syllables. I'll also note that syllabification is marked in the complementary practice of pronunciation respelling on Wikipedia. It seems like it's marked everywhere except Wikipedia's IPA, so I don't see why we need to warn against it. 0x0077BE [^talk/_contrib] 21:24, 27 October 2014 (UTC)

Good points. It seems that there is a different thinking on the matter between the two camps. Other discussions on syllabification in the archives are here, here, and here. — Æµ§œš¹ _{[ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ]} 22:41, 27 October 2014 (UTC)

... "all major dictionaries"? The OED (full edition) does not see the need for separating words into artificial syllables. Dbfirs 01:31, 28 October 2014 (UTC)

I think /ˈrɑːz.b(ə)r(.)i/ wouldn't work, because it doesn't state explicitly that */ˈrɑːz.bəri/ is wrong (Cambridge EPD seems to allow /ˈrɑːz.br.i/ though.) Peter238 (v̥ɪˑzɪʔ mɑˑɪ̯ tˢʰoˑk̚ pʰɛˑɪ̯d̥ʒ̊) 23:09, 27 October 2014 (UTC)

There are two issues here, I think. One is about whether to mark syllable boundaries in WP transcriptions (I don’t know the answer to this), and the other is whether to show alternative pronunciations. As well as showing alternative pronunciations of the first syllable, both the CEPD and the Longman pronunciation dictionaries recommend either /bər/ for the second syllable or /br/(syllabic /r/), so /ˈrɑːz.bər.i/ and /rɑːz. br̩.i/ are both supported. I think it would become far too complicated for WP to give all the alternatives for a word. RoachPeter (talk) 08:44, 28 October 2014 (UTC)

Semi-protected edit request on 27 December 2014

This edit request to Help:IPA for English has been answered. Set the |answered= or |ans= parameter to no to reactivate your request.

42.117.105.205 (talk) 16:27, 27 December 2014 (UTC) napmucingiare.com

Not done: it's not clear what changes you want to be made. Please mention the specific changes in a "change X to Y" format. --I am k6ka ^{Talk to me!} _{See what I have done} 16:29, 27 December 2014 (UTC)

/ʍ/ vs. /hw/

Is there a reason /ʍ/ is /hw/ here? People without the whine-wine merger that I've talked to don't perceive /ʍ/ as two phonemes. Is it just for the convenience of non-whine-wine speakers? Charlotte Aryanne (talk) 18:31, 17 January 2015 (UTC)

The use of /hw/ is not for the convenience of speakers, but for phonologists to keep the analysis tidy and economical. Speakers who have the "voiceless w" in 'which' produce a single sound that is phonetically transcribed [ʍ], but phonologists don't want to add a consonant phoneme to the list of English consonants just on account of a small number of contrasts like 'whine - wine'. So the arbitrary phonemic representation /hw/ is used instead, even though there is no [h] pronounced. There is a similar issue with the sound at the beginning of 'hue', 'huge', which is phonetically [ç] but is treated phonologically as /hj/ despite the absence of a [h] sound in it. In theory (as Davidsen-Neilson once pointed out) you could do something similar for e.g. English /p, t, k/, transcribing them as /hb, hd, hg/, but nobody would accept that. RoachPeter (talk) 20:09, 17 January 2015 (UTC)

It's also less likely for lay readers to mistakenly confuse it with m. — Æµ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 23:42, 17 January 2015 (UTC)

"Fool" as an example of uː

Should this have a footnote explaining the fool/full merger, or perhaps remove it as an example word altogether, given the other words (goose, food etc) are good examples? — Preceding unsigned comment added by 106.68.221.236 (talk) 11:20, 27 October 2014 (UTC)

That wouldn't help those with the foot–goose merger. If your local dialect doesn't use uː then it's difficult to explain what it sounds like. Dbfirs 13:45, 29 January 2015 (UTC)

Request for Selena article

The article is (currently being) updated and am requesting someone who can add an IPA on the article Selena since I don't know how to do it. Thanks in advance, .jona ^talk 02:57, 3 February 2015 (UTC)

loir/lawyer—another bad example

lawyer can be both ˈlɔːjə and ˈlɔɪə, which can cause confusion. Why not settle for "employer"? 176.221.120.207 (talk) 16:24, 26 January 2015 (UTC)

Doesn't that have the same variance as lawyer? — Æµ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 18:34, 26 January 2015 (UTC)

Not according to LPD and CEPD. They state that lawyer can be pronounced with either /ɔːj/ or /ɔɪ/, but employer only with /ɔɪ/. — Peter238 (v̥ɪˑzɪʔ mɑˑɪ̯ tˢʰoˑk̚ pʰɛˑɪ̯d̥ʒ̊) 21:43, 26 January 2015 (UTC)

The OED says the same for British English. Is there any reason not to change the example? Dbfirs 21:51, 26 January 2015 (UTC)

I can't think of any. — Æµ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 22:33, 26 January 2015 (UTC)

Thanks. In the absence of any other comment, I've changed it. Dbfirs 12:38, 22 February 2015 (UTC)

sound files?!

IIRC, this page used to have links to the articles on each symbol. They should be put back. And there should be clickable sound files next to each symbol. --Espoo (talk) 10:20, 5 February 2015 (UTC)

You're thinking of Help:IPA. --SnorlaxMonster 10:51, 6 February 2015 (UTC)

Without them this page does not provide enough help for most non-native speakers. The problem is compounded by many pronunciation examples that are completely unknown to most non-native speakers. The reasons given by some for the unnecessarily difficult examples like dye, thy, nigh, thigh, rye, sigh, vie, wye would only be relevant to an introductory linguistics course that wants to familiarize the students with the concept of minimal pairs. Here we instead need words that are not unknown to most non-native speakers, for example do, the, no, thanks, run, say, van, we. --Espoo (talk) 10:22, 22 February 2015 (UTC)

I agree. This page is intended as a guide to help readers pronounce words using IPA on their pages, not to teach linguistics. This page should aim to use common words with unambiguous pronunciations, and use examples that demonstrate the different common spellings that correspond to the same phoneme. --SnorlaxMonster 10:39, 22 February 2015 (UTC)

We can't have recordings here, because the page says that "this key represents diaphonemes, abstractions of speech sounds that accommodate General American, Received Pronunciation, Canadian English, South African, Australian, and New Zealand pronunciations." In some cases, this means 5 to 10 (or even more) recordings for one sound, and explaining vowel splits and mergers. That's too much. — Peter238 (v̥ɪˑzɪʔ mɑˑɪ̯ tˢʰoˑk̚ pʰɛˑɪ̯d̥ʒ̊) 13:16, 22 February 2015 (UTC)

mayor/mare—bad example

Both are pronounced as one syllable in BE, and it's not about triphthong smoothing—cf. e.g. http://www.oxforddictionaries.com/definition/english/mayor with http://www.oxforddictionaries.com/definition/english/layer. Layer vs. lair would be a clearer example. (If it weren't for the confusing spelling, also prayer (person) vs. prayer (activity or text) would do). 176.221.120.207 (talk) 20:32, 6 January 2015 (UTC)

Yes, a bad example in British English where the OED records the same pronunciation. Lair and layer would be better. Should we change it? Dbfirs 22:00, 26 January 2015 (UTC)

Since no-one has expressed an opinion either way, I've changed mayor to layer (of eggs) since this is never (as far as I know) pronounced like lair. Dbfirs 12:37, 22 February 2015 (UTC)

Huh, according to the etymology notes in OED, any layer would suffice nowadays, but probably not several centuries ago...

Thank you. It would be nice if someone updated the relevant footnote as well. 176.221.120.207 (talk) 02:29, 9 March 2015 (UTC)

Does anybody in the world pronounce layer (of eggs) differently from layer (stratum)? If so, who? If not, why specify "of eggs"? J. D. Crutchfield | Talk 20:42, 10 March 2015 (UTC)

Yes, I've heard them pronounced differently, though perhaps only in some dialects. That's why I mentioned eggs, though perhaps it would be better to say "one who lays" to include bricklayers? Dbfirs 20:56, 10 March 2015 (UTC)

I've updated as suggested. Dbfirs 21:07, 10 March 2015 (UTC)

Proposal for new symbols

/aː/ for the BATH vowel. This was proposed before, and there seemed to be rough consensus in favour of it, but it doesn’t seem it was actually adopted.
/ɔ/ (no length mark) for the CLOTH vowel. The idea being that since /i/ is chosen to mean ‘dialectal variation between /iː/ and /ɪ/’, /ɔ/ can be used to mean ‘dialectal variation between /ɔː/ and /ɒ/’. However, /i/ is a weak vowel and doesn’t also occur at the start of a diphthong, so I’m slightly more wary of this one. On balance I think it would be better to adopt it.
/u/ (no length mark) for the vowel of situation and bedroom which varies dialectally between /uː/ and /ʊ/, likewise analogous to /i/. It was identified by Wells (if not by someone else before him) and is used in the LPD.

I’d also favour adopting /ɪ̈/ and /ʊ̈/ to replace /ɨ/ and /ʉ/ on the basis that they’re more intuitive. If I was unfamiliar with the IPA (or, more likely, familiar with the IPA but not with the system used on Wikipedia) and had to guess the meanings of /ɨ/ and /ʉ/, I’d guess that they had something to do with /iː/ and /uː/, not with /ɪ/ and /ʊ/. But my notation has the disadvantage of not being used by anyone else, as far as I know. I wish the IPA would adopt /ᵻ/ and /ᵿ/ officially! DavidPKendal (talk) 13:57, 23 June 2014 (UTC)

I'd like to make a few comments on the above. Concerning the /a:/ idea, it's true that this has been suggested in the past, and even used in some textbooks. It wouldn't be disastrous if it was adopted for WP, but I would imagine most WP writers would prefer not to get too far away from contemporary practice, and for British English at least, /ɑː/ is the symbol of choice for BATH. Regarding /ɔ/, I can't see what this would be needed for. It would be perfectly possible to stop using /ɒ/ for DOG and use /ɔ/ instead, but I don't think that produces any advantage. As a diaphonemic symbol covering the possible realization of a phonemic unit as either /ɔː/ or /ɒ/, it would have a different function. But do we need this? In the days when 'off' in old-fashioned RP might be pronounced /ɔːf/ or /ɒf/ it could conceivably have been used to represent both possibilities, and there could be regional accents where this free variation still exists, but otherwise we need both symbols. The reason for having /i/ to represent /ɪ/ and /iː/ is that when a close front unrounded vowel is unstressed the distinction between the two is neutralized; some speakers may pronounce 'happy' as /hæpɪ/, some as /hæpiː/and others as something in between.

/u/ as a diaphonemic symbol covering /uː/ and /ʊ/ in a way analogous to /i/ standing for /iː/ and /ɪ/ is well established and doesn't need to be argued for.

I don't think the IPA would have any view on the matter of /ɪ̈/ and /ʊ̈/ vs. /ɨ/ and /ʉ/. WP is using the latter pair to represent either/or choices (e.g. rose's/Rosa's), but for the IPA these symbols represent one specific vowel quality each. To make the symbols stand for two different possible realizations therefore goes against IPA practice. I can't see what is more "intuitive" about adding umlauts to [ɪ] and [ʊ]. It is hard to see how these symbols with their diacritics could have any phonetic meaning. Finally, I can't see why the IPA would want to adopt /ᵻ/ and /ᵿ/ officially - there would have to be a case made that some vowel qualities exist for which no symbols existed, and that these two would be the most appropriate choice. But what vowels would the symbols /ᵻ/ and /ᵿ/ represent? RoachPeter (talk) 19:30, 23 June 2014 (UTC)

Wikipedia’s transcription doesn’t represent RP. It represents a broad range of accents including but not limited to RP and General American. The RP CLOTH vowel is /ɒ/, whereas GenAm has /ɔː/. The RP BATH vowel is /ɑː/, but the GenAm is /æ/ (as well as many other British accents). Right now the guidance offered to editors for words with these vowels is to transcribe them twice — I’m trying to save that trouble.

/ɪ̈/ and /ʊ̈/ are the near-high central vowels — unrounded and rounded, respectively — and that’s what I’d like the IPA to officially use /ᵻ/ and /ᵿ/ for. [ɪ̈] and [ʊ̈] are actually phonetic realizations of these vowels for some speakers (including me), as are [ɨ] and [ʉ] in others. Full [ɪ] and [ʊ] also occur in careful speech, while in other accents they’re fully reduced to [ə]. That, I believe, is the reason we transcribe them with single symbols — because they are each a single diaphoneme. I merely wish to transcribe them with a symbol that is both (a) officially-recognized (¨ being an official IPA diacritic representing centralization) and (b) more likely, in my estimation, to lead a novice IPA reader to a correct pronunciation. This is not against IPA principles, it’s just using the IPA diaphonemically rather than phonetically, which should be no more controversial than using it phonemically as all dictionaries do. DavidPKendal (talk) 10:59, 24 June 2014 (UTC)

OK, I don't want to argue about the precise phonetic interpretation of the symbols [ɪ̈ ]and [ʊ̈] and [ɨ] and [ʉ] because it would get like the theology of counting angels on a pin-head. The essential point is that the IPA doesn't do diaphonemes, so there is no point in hoping the IPA would in some sense endorse a symbol that has more than one phonetic realization. If people want to use IPA symbols for other purposes that's fine, but this is not what the IPA does. RoachPeter (talk) 14:11, 24 June 2014 (UTC)

This is getting off-topic, but I’m not proposing that. I’m proposing that the IPA should officially recognize the symbols ᵻ and ᵿ with the ‘precise phonetic interpretation’ of a single sound each — those being the near-high central vowels. Precisely where ɪ̈ and ʊ̈ are on this IPA chart. Unfortunately, though the symbols are already in use by the OED (among others) for this exact use, I suspect the IPA would not approve the proposal much like they rejected ᴀ a few years ago. DavidPKendal (talk) 19:01, 24 June 2014 (UTC)

Thanks, I think I can see your point now. I was thrown by the fact that this was argued in the IPA for English context, whereas this issue is really what is, or should be, on the IPA Chart. I was also confused because the WP vowel chart you refer to contains vowels that are not on the official IPA chart. As you rightly point out, OUP have pre-empted the symbols ᵻ and ᵿ for their own purposes. RoachPeter (talk) 07:35, 25 June 2014 (UTC)

While we are on the subject of missing symbols, I was surprised to find /a/ missing. The OED has started using this transcription for the "lad, bad, cat, trap" vowel that it formerly transcribed as /æ/. This change reflects the gradual move away from 1950s RP towards a more open vowel by almost all radio and TV presenters in the UK. I appreciate that this might cause confusion for American readers and those whose native language has an even more open /a/. Dbfirs 12:39, 21 October 2014 (UTC)

There are certainly things in favour of adopting this idea. I would add the fact that the latest edition of the very influential book Gimson's Pronunciation of English, revised by A. Cruttenden, Routledge (2014), has also switched to using the /a/ symbol. I would strongly recommend, though, that if this change were to be adopted for WP, it should be used uniformly throughout all the relevant WP articles to avoid confusion. I think it would take a long time to make all the necessary changes. RoachPeter (talk) 14:44, 22 October 2014 (UTC)

Thanks for the support and the mention of the book — I'll have to read it. Wiktionary, in an informal vote, decided that it was not quite ready to make the change, partly because so many users had grown accustomed to the /æ/ symbol and partly because they didn't want to have separate pronunciations for American and British versions. I think the change will come, and an electronic medium ought to be in the vanguard, but the Wiktionary and Wikipedia community tends to be conservative, and will not be willing to make the switch to /a/ until the majority of published references have done so. At some time in the future, I expect that someone will write a bot to make the changes, but it will need to be supervised carefully, and probably needs to be co-ordinated across Wiki projects. It's certainly too big a task for me, so, until there is a big majority in support, and we organise lots of volunteers to assist, I think I'll just continue to make the mental substitution /æ/ -> /a/ for most British usage (except for the few who still speak 1950s RP). Dbfirs 15:33, 22 October 2014 (UTC)

My native language has both /a/ and /ɑ/ as phonemes. Both RP and US pronunciations of words such as cat are much closer to /æ/ than to either of them. All schoolbooks I have ever seen use /æ/. To me it would be utterly confusing to see /a/ being used instead of /æ/. −Woodstone (talk) 05:25, 23 October 2014 (UTC)

I agree that the American "cat" and the 1950s RP "cat" are both better transcribed as /æ/, but BBC English is gradually moving towards the northern English version of the vowel which has always been closer to your /a/ . Your schoolbooks reflect the older 1950s RP. The sound given in IPA vowel chart with audio is actually the long version /a:/ as in a fronted version of "father". Where can I hear your interpretation of /a/ ? Dbfirs 13:17, 23 October 2014 (UTC)

Since the vowel is definitely not pronounced /a/ in US, and /æ/ is still appropriate for RP, I see no advantage in switching to /a/. Doing so would alienate a large proportion of speakers. −Woodstone (talk) 16:57, 23 October 2014 (UTC)

I appreciate that a good proportion of readers are American, and a few others in the UK still use 1950s RP, so I agree that it's a bit too early to make the change. As I said above, I always make the substitution mentally myself. Wikipedia will have to wait until most other dictionaries catch up with the OED. I wasn't suggesting that we should put /a/ in place of /æ/, just add it as a valid vowel in the English list, because I've been using it all my life. Dbfirs 18:45, 23 October 2014 (UTC)

The list is meant to define a standard pan-dialectic set of phoneme representations. It is not meant to describe differences between dialects. So there should not be alternative choices for representation. Each (dia)phoneme should have a unique representation. −Woodstone (talk) 06:52, 24 October 2014 (UTC)

Yes, I realise that the symbol has to represent a wide variety of sounds, even amongst Americans, where I usually hear a diphthong (variations of /ɛa/instead of a pure /æ/). I accept that we'll stick with the /æ/ representation for simplicity. Dbfirs 08:29, 24 October 2014 (UTC)

In the absence of objections, I have added my three proposals to the chart. Could someone change Template:IPAc-en to recognize the symbols, please? DavidPKendal (talk) 23:15, 10 March 2015 (UTC)

Diaphonemic transcription in one of the best English grammars

Our diaphonemic transcription has gotten criticism in the past as being OR or SYNTH. I think that's been resolved, but just in case, you can see a very similar system in an excerpt of The Cambridge Grammar of the English Language here — kwami (talk) 21:31, 11 March 2015 (UTC)

Wow, thanks for that link! Reassuring indeed that Pullum et al. use a very similar system. DavidPKendal (talk) 23:00, 11 March 2015 (UTC)

Ooh, that could go at diaphoneme. — Æµ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 04:40, 12 March 2015 (UTC)

Variable vowels

We've talked about the variable BATH and CLOTH vowels before, but never implemented them, preferring to give dual pronunciations instead. Are we going to implement them now, or should they be removed from the key? If we are going to, we'll need to modify {{IPAc-en}} to accept them. Also, I changed the transcription of the CLOTH vowel to ⟨ɒː⟩. ⟨ɔ⟩ is a bad idea, because many sources use that for ⟨ɔː⟩. — kwami (talk) 23:55, 10 March 2015 (UTC)

I strongly support adding ⟨u⟩. If we want to add the CLOTH vowel, ⟨ɒː⟩ looks like the best symbol for it. — Peter238 (v̥ɪˑzɪʔ mɑˑɪ̯ tˢʰoˑk̚ pʰɛˑɪ̯d̥ʒ̊) 00:07, 11 March 2015 (UTC)

What would ⟨u⟩ be for?

One of the problems with this is that people have a hard enough time w the IPA w/o us adding more symbols. — kwami (talk) 00:12, 11 March 2015 (UTC)

For the back equivalent of ⟨i⟩. Having ⟨i⟩ but not ⟨u⟩ looks like a major inconsistency to me. — Peter238 (v̥ɪˑzɪʔ mɑˑɪ̯ tˢʰoˑk̚ pʰɛˑɪ̯d̥ʒ̊) 00:18, 11 March 2015 (UTC)

We'd thought about it, but it's so uncommon we didn't think it worthwhile. — kwami (talk) 01:20, 11 March 2015 (UTC)

I support this. I can only imagine the amount of work it will take to implement. — Æµ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 14:52, 11 March 2015 (UTC)

Taking a second look at the recent changes, it strikes me as odd to consider the difference between [əm] and [m̩] as the same as that between /ɪ/ and /iː/. The former is just an issue of phonetics, while the latter (and the others marked as variably pronounced) of incidence, right? — Æµ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 14:56, 11 March 2015 (UTC)

Unlike /aː, ɒː/, /i, u/ are an issue of phonetics. Wells describes the latter as neutral, by definition unstressed close (front/back, respectively) vowels occurring in positions in which the /ɪ - iː/ and /ʊ - uː/ contrasts are neutralized. But (as in case of the word 'roof'), the /u/ symbol can also be used for a stressed vowel, which is an environment in which the /ʊ - uː/ contrast is not neutralized in neither GA nor RP. — Peter238 (v̥ɪˑzɪʔ mɑˑɪ̯ tˢʰoˑk̚ pʰɛˑɪ̯d̥ʒ̊) 15:26, 11 March 2015 (UTC)

That's not how they're being treated here. We are told that /i/ is either promounced as FLEECE or KIT. Wells may argue that it's a neutralization, but there is definitely variation in whether speakers pronounce happy with the "tense" FLEECE vowel or the KIT vowel. Barred i is the same. Some people distinguish roses from Rosa's, with the former term having an unstressed KIT vowel and the latter having a COMMA vowel. Those are differences of incidence. With bottle, button, etc. they're differences of phonetic realization of the same underlying phonemic sequence. — Æµ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 07:02, 12 March 2015 (UTC)

Hmm, looks like you're right. So what do you suggest? — Peter238 (v̥ɪˑzɪʔ mɑˑɪ̯ tˢʰoˑk̚ pʰɛˑɪ̯d̥ʒ̊) 20:25, 13 March 2015 (UTC)

I suspect that anyone who knows [n̩] also knows that it's in free variation/variabily transcribed as [ə], so I think we can just remove the paragraphs referring to syllabic m, n, and l. — Æµ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 22:34, 13 March 2015 (UTC)

See conversation above. I added them to the chart because there were no objections. DavidPKendal (talk) 16:31, 11 March 2015 (UTC)

Parenthetical alternate pronunciations

Many of the sets listed under "... followed by R" include pronunciations in parentheses after the examples. It is very unclear what these are supposed to represent, and whether they apply only to the last example or all of the examples. Could someone explain what these are intended to represent, and ideally implement something on the page to clarify? /ɑr/ has one listed with "also", whereas none of the others do, so I would assume it is supposed to mean something different. --SnorlaxMonster 11:58, 13 March 2015 (UTC)

I think the parenthetical transcriptions are supposed to help in identifying the underlying sequence. So, for example, /ɪər/ in our system is a /iː/ followed by an /r/. I'm not a big fan of them. If you're confused by it, I say we remove them. — Æµ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 16:09, 13 March 2015 (UTC)

They explain why the rhotic vowels are listed where they are. I clarified that they're phonemic reps, but that's perhaps not the proper term for a diaphonemic transcription. — kwami (talk) 20:00, 13 March 2015 (UTC)

I get what you're saying, but I don't think anyone will be confused by their placement if those are removed. The format of the table (in particular, the title of the left-most column) really lends itself to readers already understanding that. — Æµ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 20:24, 13 March 2015 (UTC)

Okay, I think that's a bit better, but I have some thoughts. As I understand it this page is supposed to be an entry-level page for readers to understand how to pronounce words using Wikipedia's usage of the IPA. The term "phonemically" seems like jargon the average reader would not understand; I think this could be solved either with a wiki link or an entry in the bulleted "Notes" section below the page explaining what it means in this context. Secondly, I don't understand why this is in the "Examples" column; being there implies it is related to one or all of the examples, but it appears to be related to the sound itself. --SnorlaxMonster 14:14, 14 March 2015 (UTC)

As with the alternating pron, for considerations of space. Could be footnotes, but we already have a lot of those. — kwami (talk) 17:17, 14 March 2015 (UTC)

I changed "phonemically" to "=" and moved those parentheticals to the diaphoneme column. Two separate edits so you can revert just the latter if you don't like it. (Personally, I think it should be reverted, but I like the "=".) — kwami (talk) 17:29, 14 March 2015 (UTC)

BTW, we shouldn't start linking words because the undereducated might not know them. There's no end to it (barrow, cairn, and Sirius aren't common either), and that's what a dictionary is for. The two exceptions we have are "loir", which is truly obscure, and "Mykey", which isn't found in dictionaries. — kwami (talk) 17:44, 14 March 2015 (UTC)

Stratum

I don't see the problem with linking only the word stratum. It's a geological term, so it's not likely to be common knowledge, unlike other terms on this page. I think it's better not to link words in the examples section (since bold is being used for highlighting and linking upsets the formatting), but this is in a footnote, so that's not an issue. --SnorlaxMonster 17:46, 14 March 2015 (UTC)

Table of Reduced vowels

I mentioned a while ago that I don't think 'nasturtium' is a good example of a vowel "often dropped". I would say that there is never an /i/ vowel in this word in any dialect of English that I know. It's hard to think of any word in which an /i/ is simply optional in the pronunciation. I could suggest 'sentient', as both CEPD and LPD list two-syllable and three-syllable pronunciations, but if the /i/ is elided, the preceding consonant becomes /ʃ/ or /tʃ/. The only other example that comes to mind is that in many varieties of British English I have noticed an increasing tendency to pronounced 'create', 'creating' as /kreɪt/, /kreɪtɪŋ/, so this word could be transcribed with a superscript /i/ after /kr/; as far as I know, however, this pronunciation has not been noted in published work, so to use it here would presumably count as OR. The superscript /ə/ needs an example: the word 'bottle' is already used to illustrate plain /ə/, but would have done equally well to show how the schwa may be present or absent in the pronunciation of this word. RoachPeter (talk) 16:41, 10 March 2015 (UTC)

It's variable according to the OED. Will try to scan for others, but this isn't straightforward to do. (The OED search engine is horribly buggy.) — kwami (talk) 00:10, 11 March 2015 (UTC)

Yes, you're right, I've checked my edition of OED and it does give an optional /ɪ/. I find this very surprising - OUP's own Dictionary of Pronunciation and their Oxford BBC Guide to Pronunciation make no mention of this possibility, and to me the pronunciation of 'nasturtium' with an /ɪ/ or /i/ sounds like a caricature of a Victorian pedant. I shall be very interested to hear if you turn up any more instances. RoachPeter (talk) 08:44, 11 March 2015 (UTC)

Unfortunately, the online OED does not allow searches of pronunciations, and the CD-ROM is buggy, not allowing (AFAICT) searches for optional sounds. I've written the OED and am waiting for their reply. — kwami (talk) 21:29, 11 March 2015 (UTC)

OED tech help knows of no way to search for words like "nasturtium". — kwami (talk) 23:32, 20 March 2015 (UTC)

An even purer instance of an optional /i/ is towards the end of 'aluminium'. Upon elision, the preceding consonant is unmodified, but stress is shifted. −Woodstone (talk) 10:47, 11 March 2015 (UTC)

That would require two transcriptions, so doesn't work for us. — kwami (talk) 21:29, 11 March 2015 (UTC)

Elision of the second /i/ in aluminium is rare in British English (only /ˌal(j)ᵿˈmɪnɪəm/ in the OED), and is a different spelling in American English. The third edition of the OED has no optional /i/ in nasturtium, giving only /nəˈstəːʃ(ə)m/ for British English. Dbfirs 22:10, 11 March 2015 (UTC)

Example inclusion

It seemed to me that this guide was presenting possible spellings of sounds, so I added a few. But Woodstone reverted me, saying that the list is intended to show pronunciation, not alternative spellings. I'm not satisfied with this; not all the examples seem to be included with this purpose.

I can see that some examples are included because they show phonemic contrasts that don't exist in all dialects: among others, thy, thigh, die, tie, shy, phi, and vie; wye and why; marry, merry, and Mary; Sirius and serious. Some of these examples are uncommon words, used only because they create minimal pairs.

However, some examples appear to be included in order to illustrate spelling: sky and crack; equation, pleasure, vision, beige. If spelling were not the factor, I'm not sure why two words with three different spellings of /k/ would be included, and four different spellings of /ʒ/. Hence, if these alternative spellings are included, why not include others if they don't overpopulate the table?

Also, I think some of the examples I included have value from a phonological perspective: in particular, write (which could be replaced with wry for more parallelism) shows cluster reduction of /wr/, and finger shows an [ŋɡ] cluster. I'm not sure why on earth sing and singer, two words from the same root with the same pronunciation of ng, are included, but not finger. — Eru·tuon 20:11, 20 March 2015 (UTC)

The purpose of the page is to show the pronunciation of the chosen phonemic representations. There have always been examples of many consonant phonemes in initial, intervocalic and final positions, because these may be phonetically different. There has never been the intention to show all possible spellings for a phoneme. A attempt at the latter can found in English spelling. −Woodstone (talk) 12:23, 21 March 2015 (UTC)

If the intent is to include examples of the sound in phonological environments, why are sky and crack included, but no examples of /k/ between vowels and alone before a vowel, and why are three spellings of intervocalic /ʒ/ included? These example sets don't follow the rationale you describe. — Eru·tuon 20:13, 21 March 2015 (UTC)

You are right on both counts. Cases of too many examples result from other overzealous editors who include spelling variants. Too few from a mistaken close adherence to IPA. Since we do not use /k^h/ as a separate phoneme for English, words containing it should be rightfuly added at /k/. Woodstone (talk) 06:28, 23 March 2015 (UTC)

I'm glad we agree on that. Based on the phonological criterion, I think finger would be a useful example, since there is no illustration of the cluster /ŋg/. Write or wry may be useful, but less so because apparently only the Doric dialect of Scots pronounces the w. Let me know what you think. — Eru·tuon 06:42, 23 March 2015 (UTC)

I think including finger, will lead to more confusion than clarification, since many readers will overlook the subtle bolding difference compared to singer. This is a global key, not intended to be an exhaustive phonetic realisation table, so let's limit to frequent cases only. Woodstone (talk) 12:59, 23 March 2015 (UTC)

Optional sounds

I think /ᵗs/ is rather unnecessary. It's OK to sometimes give two pronunciations for one word, which I think would be better for the small number of words that this would apply to. — Æµ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 18:50, 22 March 2015 (UTC)

It's hardly even two pronunciations, the way BATH has. More just that some people don't have initial /ts/. — kwami (talk) 02:47, 14 April 2015 (UTC)

Small caps examples

I can't figure out why certain examples are in small caps, & always at the head of lists. How are such examples different from the others not so distinguished? Shouldn't the format be made consistent here? Denicho (talk) 20:53, 17 April 2015 (UTC)

@Denicho: The words in small caps are the traditional lexical sets of Wells. This is explained in a note below the table. — Eru·tuon 21:06, 17 April 2015 (UTC)