Wikipedia talk:Manual of Style/Pronunciation/Archive 5

Comparing fourteen apples to an orange

Which English dictionary respelling system is being proposed in favour of, or in addition to, IPA? There are fourteen different systems documented at Pronunciation respelling for English. Are any of these usable for MSIE/Windows without a font template like template:IPA, and without any HTML code? Below are my results from a quick test—more investigation would be in order:

A quick look at that page with MSIE on a vanilla Win XP system shows that the AHD, MWCD, NOAD, RHD, DPL, A, and COD systems display some broken characters even with class="IPA" applied to the table. Removing the class and previewing in the browser shows that only the WBO, NBC, MWO, Cham, and AB systems appear to work without the class applied. The AHD, MWCD, NOAD, RHD, MECD, and MWO systems use HTML code in the table.

Summary ("X" passes, "-" fails)

System         IPA  AHD  MWCD NOAD RHD  WBO  MECD DPL  DPN  NBC  MWO   A   COD  Cham  AB

Okay in MSIE    -    -    -    -    -    X    -    -    -    X    X    -    -    X    X
Okay with font  X    -    -    -    -    X    X    -    X    X    X    -    -    X    X
No HTML req’d   X    -    -    -    -    X    -    X    X    X    -    X    X    X    X

Most of the respelling systems give worse results in MSIE/Win than IPA (although there may be workarounds). The DPN system has the same technical requirements as IPA, and MWO works without a font specification but requires typing HTML code. Only the following systems have the advantage over IPA and the rest, because they seem to work in MSIE without requiring a font specification (like template:IPA), and without typing HTML code:

Are there any online references for these four respelling systems? Michael Z. 2006-06-30 17:06 Z

ARPABET, used by The CMU Pronouncing Dictionary, is described here. Nohat 17:39, 30 June 2006 (UTC)

I think most of the discussion here about respelling systems has focussed on the Help:Pronunciation respelling key that some Wikipedians originally developed for astronomical articles (based on the key in Robert Fagles' editions of classical texts). It's outside article space, and so shouldn't run afoul of Original Research charges. It's a proposed convention, like a hundred other WP conventions. However, an existing respelling system like one of those listed in Pronunciation respelling for English might also work. I suggest the main criterion for favouring one over another is absence of special characters, at least of any characters that create technical difficulties. Ideally, one that uses only the 26 standard letters of alphabet. Note that the Help:Pronunciation respelling key used the IPA schwa symbol - does any browser have problems with this?--Chris 17:42, 30 June 2006 (UTC)
I totally did not realize that anyone was talking about the astronomical system; I thought it was about the familiar dictionary keys we were used to seeing in grade school.
The usual culprit is MSIE/Win; Safari, Firefox, Opera, and even Lynx (!) all handle Unicode text better. I routinely use a Mac, so I can't check immediately—can you read the following schwa in MSIE/Win: ə? Michael Z. 2006-06-30 22:11 Z
I can. Schwa is used in Latin alphabets other than the IPA, so perhaps MS thought it worth covering. kwami 17:22, 30 August 2006 (UTC)

Wiki IPA for English

For use only for English words I like the idea proposed above of using IPA in principle, but converting it to use only latin letters. Whenever the IPA symbol is a latin letter, use it. When not, use a close relative. Here is a possible way of doing it, based on the Help:IPA for English article. For the consonants it's not difficult. For the vowels (only 5 in the latin alphabet), some diacritics are used. The colon (like "o:" or dash on top "ō"?) can be used for the long vowels, and implies as well their articulation variant. The ě is used for shwa and ǔ for the sound in "run".

IPA: English Consonants
IPA WIKI Examples
p p pen, spin, tip
b b but, web
t t two, sting, bet
d d do, odd
tsh chair, nature, teach
dzh gin, joy, edge
k k cat, kill, skin, queen, thick
ɡ g go, get, beg
f f fool, enough, leaf
v v voice, have
θ th thing, teeth
ð dh this, breathe, father
s s see, city, pass
z z zoo, rose
ʃ sh she, sure, emotion, leash
ʒ zh pleasure, beige
h h ham
m m man, ham
n n no, tin
ŋ ng singer, ring
l l left, bell
ɹ r run, very [1]
w w we
j y yes
ʍ wh what (some accents, such as Scottish)
x kh loch (Scottish), Chanukah (Yinglish/Yeshivish)
IPA: English Vowels
IPA Examples
RP GenAm AuE WIKI  
ɑː ɑ a: father
i i: see
ɪ ɪ ɪ i city
ɛ ɛ e e bed [2]
ɜː ɝ ɜː ǔr bird
æ æ æ ae lad, cat, ran [3][4]
ɑː ɑɹ a:r arm
ʌ ʌ a ǔ run, enough
ɒ ɑ ɔ o not, wasp
ɔː ɔ o: law, caught [5]
ʊ ʊ ʊ u put
u ʉː u: soon, through
ə ə ə ě about
ə ɚ ə ěr winner
 
IPA: English Diphthongs
IPA Examples
RP GenAm AuE WIKI  
or e æɪ ei (ey) day
ɑe ai (ay) my
ɔɪ ɔɪ oi (oy) boy
əʊ or o əʉ ou no
æɔ au now
ɪə ɪɹ ɪə i:r near, here
ɛə ɛɹ e:r hair, there [6]
ʊə ʊɹ ʊə u:r tour
juː ju jʉː yu pupil

Woodstone 19:13, 30 June 2006 (UTC)

Rather dangerous, I'm afraid: one would need to understand two phonetic alphabets: WIKI (English sounds) and IPA (other sounds) and become even more confused. Would we see a Turkish word represented by WIKI+IPA or by strict IPA when it contains only some sounds that occur also in English? Think about the problems that would be caused by either choice. -- SomeHuman 2006-06-30 19:56 (UTC)
Obviously the systems should not be mixed. The simplified form is only suitable for English words. Any foreign words would have to be expressed using real IPA. −Woodstone 21:00, 30 June 2006 (UTC)
An English approximation would be better and more realistic than a character set which fails to come out on many computers, including public ones. --Etaonsh 04:28, 11 August 2006 (UTC)

We used to use an established system for this purpose called SAMPA. Since English Wikipedia gained the capability to display UTF characters, SAMPA has been completely abandoned. I don't think anyone wants to reintroduce a non-standard replacement. Michael Z. 2006-06-30 22:14 Z

I still say Kirshenbaum is the best choice for this. It's much easier to just read than either SAMPA or straight IPA (the advantage over straight IPA is not having to memorize a bunch of symbols for which most of us have no names). --Trovatore 22:35, 30 June 2006 (UTC)
Wikipedia may have 'gained the capability to display UTF characters,' but it doesn't mean that we can all see them this end. --Etaonsh 04:28, 11 August 2006 (UTC)

New approach

Might someone know whether this can be realized (server side .php, user side JavaScript and possibly css, I guess):
Behind a word that could use help for its pronunciation, we would only see a single small symbol, its colour would indicate the first available option :

The cümbüş x•   xis a Turkish instrument.  each time you bring the cursor on the gray symbol (half a second), you hear the word (sound file – in this demo not functional);
click on the symbol and the colour of the symbol would indicate the next available option :
The cümbüş x •  xis a Turkish instrument.  hold the cursor on the dark blue symbol and read the IPA representation for the word (best keep the pointer moving just under the symbol);
click on the symbol and the colour of the symbol would indicate the next available option :
The cümbüş x  • xis a Turkish instrument.  hold the cursor on the clear blue symbol and read the sound-out representation for the word (best keep the pointer moving just under the symbol);
click on the symbol and the colour of the symbol would indicate the initial option is ready again.

Wikipedia's 'user preferences' might even allow to change this default order according to the personal preferences of the receiver (thus most often one would not even have to click once).
If one (or two) of the three is not available and thus has to be skipped, the next (user preferenced) available presentation is indicated with its matching colour. Note: each of the 3 dots would automatically be very light gray when its option is 'available', reddish if 'unavailable'; their order remains the above shown default order just as the colours stick with the presentation options, regardless user preferences, thus the appearance has the same meaning for everyone. This prevents needlessly pointing or clicking.

  • It solves a practical problem: a writer of an article is rarely able to add two or three sound representations, just any one of the three is enough and other users can afterwards provide the missing options.
  • A practical sample in case just the audio file is not available: a user who prefers sound-out above hearing and really does not want to bother with IPA, will see  x  • x; a user with default preferences intially sees  x •  xtill someone adds the sound file. While a writer only offered a sound-out, everyone sees  x  • x until someone adds one or both missing options.

SomeHuman 2006-06-30 18:24 / 2006-07-01 11:20 (UTC)

This is an attractive, cluttering-reducing idea, but probably a future project rather than something that would be the outcome of our current debate.--Chris 18:37, 30 June 2006 (UTC)
If the three-click-symbol is not too difficult to apply (both server side and user side handling, and server side creation from template), then what would the purpose of the current debate still be? -- SomeHuman 2006-06-30 18:45 (UTC)
There's still debate over whether a sound-out system is acceptable, and, if so, what its specifics should be. Once that is settled, then a system like yours -- which is simply an elegant way of present multiple systems - becomes discussible.--Chris 19:06, 30 June 2006 (UTC)
My approach makes 'sound-out' only debatable for its defenders: it is far easier to accept if not everyone has to live with it. -- SomeHuman 2006-06-30 19:22 (UTC)

Where do I get help?

Where do I go to request someone put the IPA pronunciation style on an article (specifically, Gallipolis, Ohio)? youngamerican (ahoy-hoy) 11:54, 10 July 2006 (UTC)

Is there a template? Hyacinth 08:23, 12 July 2006 (UTC)
You mean like Template:ConvertIPA? —Keenan Pepper 19:08, 12 July 2006 (UTC)

King of the Hill

Regardless of whether you think IPA is the greatest thing since sliced bread or a terrible disaster for WP, it's clear that it's highly controversial among WP editors. The MOS ought to reflect that. WP is not King of the Hill, where the people who get up first have the right to push everyone else down. Nareek 01:49, 10 August 2006 (UTC)

If you can offer an alternative or compromise that actually serves all the purposes to which Wikipedia puts pronunciation guides, you're welcome to do so. The current compromise I'm aware of is using IPA and a "rhymes with" or "PROnun" rendition simultaneously. IPA is necessary for accuracy (an important feature of an encyclopedia article), and the rhymes/PROnun is offered for those who can't read IPA. — Saxifrage 02:35, 10 August 2006 (UTC)
Yeah I'm not the biggest IPA fan but the compromise is fine with me. --Liface 03:03, 10 August 2006 (UTC)
To avoid King of the Hill fights, I suggested a new approach where every user can put his own preference (actual sound, IPA, sound-out) at top of his/her hill. Debates should then be restricted amongst people with a common preference, e.g. on having a slow pronounciation followed by a normal conversation pace in the sound file, others on the detail of IPA, a third group on choosing the kind of sound-out representation; it would soon become more constructive then. — SomeHuman 2006-08-10 03:07 (UTC)
That solution is technically interesting, but is of limited use when Wikipedia sees print. A preference-driven pronunciation style system is unlikely to be popular for the same reason that we don't have a preference-driven American/British English system in place. — Saxifrage 03:51, 10 August 2006 (UTC)
US/GB would require every page to have 2 complete versions, unthinkable unless some genius finds an automated translator technique (might as well do it from/towards all the world's languages then). Do you intend to have every page word by word rewritten in IPA or in PROnun, perhaps? User preferences already exist, for instance how a date is shown: it does not take different page content. The advantage of the three-click symbol is that it can be inserted by anyone who wants to deliver a pronunciation indication for an uncommon term, even (and most usually) only a single type. Other users can add one or both missing types at any later time without a problem. The defenders of IPA never suggested that every writer must either master IPA or stay away. The small symbol is not so intrusive as sound-out or as IPA; contrarily, delivering both (or just one type) in plain text would hinder normal reading regardless one's preferenced type, especially since many users will often have little need for an indication because they already know how to pronounce the specific term. — SomeHuman 2006-08-10 05:01 (UTC)
I skimmed and in my haste misunderstood your proposal. Yes, a BrE/AmE conversion would be much more problematic and isn't a good analogy.
Still, though, how would that work for a print version of Wikipedia? It seems that in deciding how to translate the symbol into plain text for a print edition we still have the same problem of what to include and how. — Saxifrage 05:25, 10 August 2006 (UTC)
A print-out would not show any representation (and certainly no sound files). Theoretically it could be done for IPA and PROnun, even with user preference, but the effort and overload would be out of proportion to the number of pages actually becoming printed. How often will a pronunciation indication on paper matter to 99% of the users? – One can always get back to the original article if the need arises. Articles become updated, let's assume improved, constantly; thus Wikipedia on paper seems a grotesque waste of both natural and intellectual resources. — SomeHuman 2006-08-10 15:44 (UTC)
I'm totally in support of your idea. --Liface 21:48, 10 August 2006 (UTC)
Part of the point of WP is to provide free encyclopedic content for anyone to use who has need of such--either online or off. Hence articles are supposed to be designed to work in either format. Nareek 17:07, 10 August 2006 (UTC)
Regardless, Wikipedia is aiming for a print version, though it's current target is a CD release. If a proposed standard for pronunciation would result in a preventable loss of information when put into print, I don't think I could support it. — Saxifrage 17:41, 10 August 2006 (UTC)
Having multiple systems in the soft edition will offer more choices for a hardcopy one. Whoever is formatting it can decide which ones to include. jnestorius(talk) 18:11, 10 August 2006 (UTC)
Doesn't it make sense to do it right the first time? On the first point I agree: multiple systems in the soft edition are valuable. They're included by the current defacto standard though. — Saxifrage 01:46, 11 August 2006 (UTC)
Standard? I thought that's something everyone uses a same way... and a reasonable one does not cause eternal debates. The three-click would for a writer be a simple template {{pron:||}}. A PROnun writer would put {{pron:||joom’-boosh}}, later someone creates a sound file (in a repository like now for images) and simply inserts the filename (assuming only normal characters would be allowed for filenames): {{pron:cumbus.wav||joom’-boosh}} and weeks or months later an IPA adept puts his part in the middle. Meanwhile the thing works as soon as options become available without giving a sloppy unfinished impression, and takes very little room in a text so one can easily read the content without being distracted. As for printing: all well-designed works offer more options online than in a book or print version, it's the strenght of the medium. — SomeHuman 2006-08-11 02:18 (UTC)
A reader who encounters a term he cannot figure out how to pronounce, can simply insert an empty template behind the word. A bot reports these (and incomplete templates) on some page and I'm sure some users will become active in adding the representation they specialize in. I do not see such possibilities with the 'de facto standard' or anything suggested so far. — SomeHuman 2006-08-11 02:28 (UTC)
You obviously have a lot of passion for this subject, and I must apologise that I just can't match that level of interest. A nice and polished version of that proposal might benefit Wikipedia. It doesn't seem to have gotten much discussion, though, and since I'm not much interested in pursuing this further, I have one suggestion to leave with: compile your original proposal and the clarifications you've made here and put them in one place (perhaps in a new section here, or a project-subpage of your userpage), and make mention of it on one of the Village Pumps. If it's given a thorough going-over by many interested parties with their various concerns (not every concern I can think of is covered by the proposal so far, as it understandably reflects your particular concerns), that might just result in neat solution and an end to these debates. — Saxifrage 05:01, 11 August 2006 (UTC)

IPA usage in Wikipedia articles is a disaster

It is horrifying that IPA is used as the pronunciation guide on Wikipedia articles, and more horrifying that it is encoded into this manual of style. It is safe to say that, statistically speaking, nobody understands it and nobody uses it. Claims that "it's more accurate" or "it's more international" or "we should lead the way" are hopelessly ivorytowerian. We should use simple soundalike guides. Tempshill 17:32, 16 August 2006 (UTC)

What you said. I am implacably opposed to the use of IPA for just the reasons you gave. However, in the spirit of compromise, I propose a two-state solution here, where IPA and "pro-nun" guides live side-by-side in peace; I'll even let the IPA come first as a convention, if that's what the pointy-headed types here demand. But I insist on good sounds-like guides to the non-academics among readers here.
Besides, a cursory glance at a few articles here shows just what I'm proposing as an evolving de facto standard here anyhow. +ILike2BeAnonymous 17:56, 16 August 2006 (UTC)
Please be civil and refrain from name-calling. — Saxifrage 18:00, 16 August 2006 (UTC)
Soundalike guides are inadequate for terms that aren't originally from English, and vary depending on the accent of the reader. Pronunciation is an encyclopedic detail that needs to be stored somehow. If you like, you are welcome to add soundalike guides alongside the existing IPA, which would maximize usefulness. — Saxifrage 17:58, 16 August 2006 (UTC)
Any objection to editing the guidelines to reflect that position?--CJGB (Chris) 18:20, 16 August 2006 (UTC)
Not at all! Never mind, it's already in there at Wikipedia:Manual of Style (pronunciation)#Other transcription systems. You might want to move stuff in that section around to make the allowability of additional systems come before the historical discussion. — Saxifrage 18:25, 16 August 2006 (UTC)
I very much object to IPA being the standard, and soundalikes being "allowed". This is absurd. By the rationale you have put forth above (in bits and pieces; I don't mean the preceding paragraph), we should just make it Wikipedia policy that all quotations from the Bible must be in Aramaic or Hebrew, because an English translation would alter the meaning slightly. Please re-examine your basic assumptions with regard to the function of this encyclopedia. Encyclopedias are meant to be used. Let's make our articles usable. IPA is not usable by 99.98% (and I am being generous) of the people who will ever read these articles. The ability of an article to communicate information is more important than encoding into the article an undecipherable fact for the benefit of future linguists. Tempshill 21:25, 16 August 2006 (UTC)
By all means add information to articles, but I, as you put it, very much object to the removal of information. Just because it is not useful to you and a hyperbolic percentage of the population doesn't mean it should be omitted, much less actively removed. — Saxifrage 21:38, 16 August 2006 (UTC)
Not exactly an objection, as such: I am very much in sympathy with the original poster and consensus, here, but I'm unconvinced that 'soundalike guides' are the whole answer, like Saxifrage. I feel we need a dictionary key like the IPA, only much simpler. Spelling reformers like myself spend a lot of time discussing spelling systems which better reflect the sounds represented in speech, and one of the 'ways in'to public consciousness we discuss is the use of our systems as dictionary keys. I, for example, use a system called 'Qixpel' [[1]] which makes use of some of the numeral keys to convey common sounds not represented by single Latin letters. --Etaonsh 18:35, 16 August 2006 (UTC)
I believe dictionary keys have already been discussed at some length. Note that IPA is a dictionary key as well. — Saxifrage 19:10, 16 August 2006 (UTC)
You would have to invent one first that is immediately understandable and useful, which is not the function of Wikipedia, fortunately. Tempshill 21:25, 16 August 2006 (UTC)
= 'Come back when you're established.' Sounds like the Conservative Party or something. :( --Etaonsh 06:54, 17 August 2006 (UTC)
  1. ^ Although the symbol r technically represents an alveolar trill, which is absent from most dialects of English, it is nevertheless widely used instead of ɹ in phonemic transcriptions.
  2. ^ Often transcribed /e/ for RP, for example in Collins English Dictionary.
  3. ^ Often transcribed /a/ for RP, for example in dictionaries of the Oxford University Press.
  4. ^ See bad-lad split for more discussion of this vowel in Australian English.
  5. ^ See low back merger for more discussion of this vowel in American English.
  6. ^ Alternative symbols used in British dictionaries are /ɛː/ (Oxford University Press) and /eə/.