Talk:International Phonetic Alphabet/Archive 2

From Wikipedia, the free encyclopedia
Jump to: navigation, search



Does anybody know anything about the history of the IPA [where A in this case stands for Alphabet, not Association ;)]? I'm wondering who decided on these symbols, and when, and how. leigh 01:41, Feb 24, 2004 (UTC)

There are a few books about the history of the IPA & much criticism. I will look up some sources for you (dont ask when though...). The IPA has gone through many edits thoughout its history -- many of the old charts are rather different from the present one. It would be interesting to compare them (maybe I should do this). The IPA has an obvious Eurocentric slant to it, most obviously in the choice of symbols. The IPA is essentially a creation of the British, although other European countries had influence in its development. There is a distinct Americanist tradition (sometimes APA, i.e. American Phonetic Alphabet) that is a bit different from the IPA. The Americanist transcription system was created mostly by missionaries & anthropological linguists to describe the indigenous American languages. The Americans encountered many sounds not present in Europe, and as a result the IPA was influenced significantly by the Americanist system. The Europeans didnt have access to as many languages & their various sounds. Colonists in Africa and the like gradually contributed to changing the IPA in order to account for the sounds in these langs. As you might have guessed, politics definitely had a role in shaping the chart. I am interested in learning about its history, too. Just havent had the time. Cheers! - Ish ishwar 09:09, 2004 Dec 17 (UTC)

Pronunciation guide

There is an English pronunciation chart for SAMPA. It would be helpful to have a similar chart here. Is that possible? (Dbenbenn)

For stylistic reasons, I removed the "questions" section with this question:

  • Are there any good guides to learning to pronounce all of the sounds in the alphabet?

A better approach would be to ask the question here on the talk page, then include a link in the article if someone comes up with a pronunciation guide.

Speech segments in relation to phonemes

"The general principle is to use one symbol for one speech segment..."

Is a 'speech segment' a phoneme? If it is, that would be a better word, and should probably be linked. --Spikey 04:25, 25 Dec 2003 (UTC)

I'd like to repeat Spikey's question, above. Does "speech segment" mean phoneme? If so, why not put it in the article?

No, certainly not a phoneme, since phonemes can have several possible realisations in practice, each of which would require a different symbol if you want to represent the sound. IPA is the International Phonetic Alphabet, not a Phonemic Alphabet.
It can be used to represent phonemese if you aren't interested in the exact realisation, but I think that's not the underlying purpose.
A speech segment may possibly be a phone, though. -- pne 10:09, 17 Jun 2004 (UTC)

In re Greek

I deleted mention of [y] being pronounced like Greek υ because that letter (ypsilon) is pronounced [i] in Modern Greek—except when it's pronounced [f] or [v] in certain digraphs. All in all, it's a very poor example. —Tkinias 18:48, 31 Mar 2004 (UTC)

New tables

tt. has some nice updated tables we might want to use. -- Ævar Arnfjörð Bjarmason 00:27, 2004 Dec 1 (UTC)

IPA Website

It's a website tha t shows how to pronounce the IPA symbols.

modified Latin letters correspond to a similar sound

Hi. I take a little issue with this paragraph which reads:

"Letters that have shapes that are modified Latin letters correspond to a similar sound. For example, all the retroflex consonants have the same symbol as the equivalent alveolar consonants, except with a rightward pointing hook coming out of the bottom."

I don't find this to be really incorrect, but I find the statement plus example to be a little misleading in regards to the systematicness of the alphabet. Among the nasal, retroflex, ejective, and implosives there is common visual element tying together these consonant classes, the other classes do not have anything like above tying them together. So, the relationships between the specially created IPA characters and the Roman alpha characters are unpredictable.

When describing the IPA it should be noted that the connection between the sound and the shape of the symbol is not regular. This contrasts with other phonetic alphabets that do have systematic correspondences.

So in other words the "modified Latin letters" do correspond to non-modified Latin letters with similar sounds, but nature of the correspondence is different for most every correspondence.

Cheers! --Ish ishwar 07:56, 2005 Jan 17 (UTC)

In contrast, however, sounds corresponding to modified Latin letters are all related in some way to the sounds corresponding to unmodified versions of those letters, so saying that they correspond to a similar sound only "sometimes" is simply false. On balance, misleading is better than false, but feel free to clarify the relationship between Latin letters and the modified letters in IPA. Perhaps examples that exemplify the non-sytematicity of the relationship instead of or in addition to the retroflexes? Nohat 08:36, 17 Jan 2005 (UTC)
My edit was due to my initial interpretation of the first sentence which was something like: "modified Roman letters correspond to a similar sound, i.e. modified Roman letters are related to each other". I think your restatement is clearer. Peace - Ish ishwar 19:47, 2005 Jan 17 (UTC)

IPA reader software??

Is there any IPA reader software or like? or any screen readers that can read IPA? TIA --Rrjanbiah 06:12, 25 Jan 2005 (UTC)

This isn't exactly what you were looking for, but I just wrote a quick little Java application that puts up a clickable IPA chart. You click on a symbol and it plays that sound. I took the IPA chart and all of the sounds from the Wikipedia article.
One thing though: Although it's Java, the application isn't yet web-friendly. For one thing, because I could only figure out how to get Java to play wav files, I converted all those oggs to wav, meaning that the application's total size -- including class files, wav files and the one png for the chart -- is about 13 MB. Adding insult to injury, I didn't feel like rewriting my code to deal with HTML lookups, so I just put the whole thing in as a jar. So basically, you have to download 13 MB just to use this program.
I don't have it uploaded right now, but if there's an interest I can put it up somewhere. It'd be open source, of course, so if someone wants to play with it and make it more web-friendly, that'd be great. But before I go through all that effort, I just wanted to post here to see, basically, if anyone cares :)
(One final note: I'm not a real programmer, I just play around with simple stuff. So if you do take a look at my code, please don't deride me for how bad it is.) -- Eleusinian

no extIPA or VoQS

The article is sorely missing the charts of the extensions to the IPA (extIPA), the symbols used to transcribe disordered speech (by speech pathologists), and the Voice Quality Symbols (VoQS). Consult Ball et al. (1995) for more info.

- Ish ishwar 00:30, 2005 Jan 29 (UTC)

For us Americans...

For many Americans (maybe people elsewhere too), it would be nice to compare IPA to the Merriam-Webster pronounciation guide (of course, they use something different in print than what is on their website...). Then there are the characters many U.S. schoolchildren use in their early years ("long a" ā, "short a" ă, etc.). I suppose these systems are probably too simplified to be of any use for linguists, but it just might help regular folks trolling around Wikipedia to understand IPA. Frankly, it always strikes me as odd whenever I see IPA or SAMPA linguistic marks in Wikipedia, but this is just because I'm totally unfamiliar with it and it's different from what I grew up with. User:Mulad (talk) 17:06, Mar 3, 2005 (UTC)

The relation is mutual: for most non-Americans, all those legacy pronunciation writing systems like Merriam-Webster are completely incomprehensible. A comparison chart would be definitely helpful. EJ 11:21, 21 Apr 2005 (UTC)
How many different systems of phonetic transcription are there besides Merriam-Webster? Peter Isotalo 11:37, Apr 21, 2005 (UTC)
Something similar to the Webster usage is taught in Usonian elementary schools. At least the five basic vowel glyphs, and oo, with either breves or macrons, plus schwa, ow, oy, are commonplace. The various systems usually only differ in how they treat the other low back vowels, voiced vs. voiceless th, etc. I can add something for the basic forms, where they're mostly in agreement. kwami 06:18, 2005 May 29 (UTC)
Actually, I don't know where to put it. How about here, and someone can move it to someplace more useful if they like? The basic idea is this: this is where English long and short vowels ended up after the Great Vowel Shift. The main exception is ū, which represents what happened to French [y], since Old English ū came to be spelled ou, ow under French influence. So, the short vowels are what you hear in bit, bet, bat, bot, but, and the long vowels are in bite, beet, bate, poke, puke. The oo vowels are what you hear in foot, food.

Here's the elementary-school system. BTW, the stress marks go before after the accented syllable:

IPA Usonian short vowel IPA Usonian long vowel
[æ] ǎ [ei̯] ([eː]) ā
[ɛ] ě [ɪi̯] ([iː]) ē
[ɪ] ǐ [ɑi̯] ī
[ɒ] ǒ [əu̯] ([oː]) ō
[ʊ] oo-breve [ʊu̯] ([uː]) oo-macron
[ɐ] ("[ʌ]") ǔ [jɨu̯] ([juː]) ū
[au̯] ow [ɔi̯] oy

Of course, IPA [ɐ] is almost invariably written with the wrong symbol, [ʌ], when transcribing English. (Vietnamese has an [ʌ], English does not.) Also, [ɒ] will be [ɑ] for many speakers. This and [ɔ] are where you need to go beyond the simple breve-macron convention, with symbols like ä and a, o with an over-dot. Since I don't control these vowels, it's a little hard to convert the full Webster system to IPA. Here's my best shot at the system used in Merriam-Webster's Third New International Dictionary, leaving out things I can't make heads or tails of, such as a, aa, and ai, all of which are supposed to represent the a in bag for some speakers; and the diphthongs, which are straightforward. (When I say "a-dot", I mean an a with a single dot atop it, which isn't supported by Unicode.) :

a, e, i: as the letters with breves above. ā, ē, ī, ō: as above. Schwa: [ə] (if unstressed), [ɐ] (if stressed); schwa-macron and əi: [ɚ] for r-droppers; schwa-dot: [ə] or [ɪ] in unstressed syllables; ə: the following consonant is syllabic; ər: [ɚ] or [ɹ̩]; ä: [ɒ]; a-dot: [ɑ]; d‧ : a flapped t or d; hw: [ʍ]; k: [x]; n: the preceding vowel is nasalized; ŋ: [ŋ]; o-dot: [ɔ]; œ: [œ]; œ-macron: [øː]; ü: [uː]; u-dot: [ʊ]; u-e ligature: [ʏ]; u-e ligature with macron: [yː]; yü: as ū above; sh: [ʃ]; ch: [tʃ]; j: [dʒ]; th: [θ]; th: [ð]; zh: [ʒ].

--kwami 08:21, 2005 May 29 (UTC)
"a" with a single dot above it is supported in Unicode. Its HTML entity is ȧ (ȧ). Denelson83 09:15, 14 July 2005 (UTC)

How to vastly improve the value of all phonetics articles to non-experts

It's easier than you might think. We have sound files for the different sounds of the IPA, why don't we simply link to these directly in articles?--Pharos 07:50, 7 Mar 2005 (UTC)

It would sound like Stephen Hawking on a bad day. It's hard to handle allophones. Joestynes 01:17, 8 Mar 2005 (UTC)
I Think it's a pretty decent idea; hearing the Pronunciations would be really helpful.

Actually, if you go to the page for any particular sound, there is a recording of Peter Ladefoged making that sound. See e.g. velar nasal, voiceless retroflex fricative. Nohat 04:47, 8 Mar 2005 (UTC)

He kind ducked out on the voiceless dorso-palatal velar fricative and I've added a sound file I've recorded, though it might be a bit too cheerful to quite fit in. I'll make one that fits with the rest of Peter's recordings later. I've also recorded an almost complete Standard Swedish phonology over at Commons.
Peter's done a very thorough job, but I think his recordings would be easier to take note of with the audio template. - karmosin 23:27, Mar 8, 2005 (UTC)
Well, he didn't make those recordings for Wikipedia. I copied them from [1]. The copyright notice seemed to indicate use on Wikipedia was OK. Peter Ladefoged is one of the world's foremost experts on phonetics. See, for example, his books: [2] Nohat 23:36, 8 Mar 2005 (UTC)
Oh. And I thought he was just a wikipedian! Guess I still have a lot to learn about linguistics. But what do you think about applying the template to the phonetics articles? - karmosin 23:47, Mar 8, 2005 (UTC)
I haven't seen that template before. I'm not opposed to its use. Nohat 11:14, 9 Mar 2005 (UTC)
Sorry to be a killjoy but is this legitimate? Clearly a link to Ladefoged's excellent site is fine but is it OK to have copied his sound samples to the other pages? The "title page" simply says, "All the material on this site is copyright" [3] although there may be other statements I have not found. Has anyone asked Ladefoged if his sounds can be copied to Wikipedia? He'd probably be delighted! Thincat 11:33, 9 Mar 2005 (UTC)
When I first copied them there was a comment to the effect of these files can be used for any noncommercial purpose. The audio files should probably be properly tagged. I'll try to find the notice soon. Nohat 11:35, 9 Mar 2005 (UTC)
[4] is the page as of 5-June-2003. It said "... material may be used in any way, provided that it is acknowledged. It should be cited as: UCLA Phonetics Lab Archive, 2003." It was unclear which usage statement went with which content, so I assumed his book publishers weren't claiming copyrights to the recordings. You are right, though, we should probably just ask for permission explicitly. Nohat 11:53, 9 Mar 2005 (UTC)
Surely we should acknowledge Peter Ladefoged's copyright over these sound files? At present when you click on an image you are taken to a page giving the image's history and copyright. For sound files I think there should be an (i) [information] icon between the [external link] icon, and the ? [help] link. CS Miller July 3, 2005 18:27 (UTC)
OK, the links now lead to the Image: link for the sound files, which will at least let you see the copyright notice. CS Miller 21:14, July 11, 2005 (UTC)

misspelling on diacritic chart

Hi. "apital" is misspelled. should be "apical". — ishwar (SPEAK) 20:59, 2005 Mar 18 (UTC)

Join with IPA in Unicode?

There's been a suggestion that we use the material we have at IPA in Unicode here. Could there be a way to use both the tables as well as the images of the charts in the same article? Peter Isotalo 16:31, Apr 17, 2005 (UTC)

I tend to keeping them separate. Quite some people will not be able see the IPA font properly, because they have computers, browsers or settings that are not compatible with the IPA font. They will still see the IPA page, with its graphical pictures, straight from the IPA handbook. The separate IPA in Unicode gives the same information. I use this page often in a separate window to cut and paste from. I am afraid merging the two would give a rather cluttered article. −Woodstone 18:59, 2005 Apr 17 (UTC)
There is now a merger notice on the IPA Unicode article, put there by someone who believes that the tables should be removed from this article and placed as links (to avoid clutter), as IPA support is nearly universal. I've been tempted to start the merger myself. You might want to comment there. kwami 07:21, 2005 Jun 3 (UTC)

I'm not sure what the point of the "article" IPA in Unicode is....... It should be merged. --Menchi 07:46, 3 Jun 2005 (UTC)

I've been meaning to merge them for months, but I was worried it would be controversial. — Chameleon 08:38, 3 Jun 2005 (UTC)

Note to "description" section

Nohat sorted the consonants alphabetically. I guess he did not notice they were sorted more logically as in the IPA chart. I also do not understand his remark about complex German vowels. Ask any German to say a, e, i, o, u and you will get a perfect rendering of the corresponding IPA symbols. I am inclined to revert. −Woodstone 21:41, 2005 Apr 28 (UTC)

The order that they were in would only be meaningful to someone who already knows about IPA and what the letters stand for, which is not the intended audience of the sentence. Putting them in alphabetical order makes it easier for someone who is unfamiliar with IPA to note whether or not a letter they're thinking of is in the list. As for German vowels, the reason I removed the comment is because in written German, the letters a, e, i, o, and u often do NOT correspond to the sounds [a], [e], [i], [o], and [u], quite like they do in Italian or Spanish. For example, in the very common German words es, mit, and von, the vowels do not correspond to [e], [i], and [o], respectively. So saying that the IPA vowels are like German vowels would be misleading at the very least. Nohat 04:36, 29 Apr 2005 (UTC)

The first o in Italian "rosso" does not sound like [o] either, but more like the o in German "von". Similar for e in Italian "mensa" and (but less clear) for a in "carro". For i and u there are hardly any deviations both in German and Italian. Only in Italian these differences are usually not phonemic (but consider the contrast with "caro") whereas they often are in German. −Woodstone 09:24, 2005 Apr 29 (UTC)

Help - Genghis Khan pronunciation

The Genghis Khan article has a non-standard pronunciation guide - apparently of the Mongolian pronunciation so I'm not sure exactly what sounds it is supposed to represent. Could someone help, please? SteveW | Talk 13:36, 30 Apr 2005 (UTC)

Unfortunately I don't have any Mongol stuff with me, but no one else has answered this, so I'll give it a shot. I can only go by the Cyrillic, but I believe that's reasonably phonetic (actually, it's not a very good system, but I don't believe there are too many problems with this name). A reasonable approximation would probably be ʧiŋgis haːn. The h and final n might actually be velar, and the ch might be like Mandarin j, but that should at least be recognizable. kwami 06:44, 2005 May 29 (UTC)

Little IPA assist

Doing some disambiguation work and saw a request for someone who knows IPA to help on Talk:Mi'kmaq -- I don't, can someone here give him a quick assist? Thanks — Catherine\talk 14:14, 11 Jun 2005 (UTC)

Offered some time ago, no takers. kwami 21:37, 2005 July 22 (UTC)

New, compressed chart

In my view the new, compressed chart is not an improvement:

  • the font used is a rather primitive typewriter font
  • the character size is quite small and thus difficult to read
  • the whole chart will not fit on the screen of many users, so they will have to scroll both vertically and horizontally
  • what was wrong with the original IPA chart?

I propose to revert, but I will snatch a copy for personal use first. −Woodstone 11:14, 2005 Jun 16 (UTC)

Well, please, leave the links to the new chart so other people can print it out as well. Thincat 12:45, 16 Jun 2005 (UTC)
Well, judging by the way it looks, since I'm the one who created the chart and the fonts used in it, I'd say yeah, it doesn't look very good like that, so I'll revert my change and just provide a link to it. I created it so it could fit on a computer screen all at once, and I chose a size of 1024x768 because that's the screen resolution that most computer users have their monitors set to, according to my own research. Denelson83 19:48, 16 Jun 2005 (UTC)
hi. i think that your idea is good — it is more convenient to see it in one screen (& it appears on one page in printed material). what if you change your fonts (to the liking of youself & others)?
i will mention one thing about the font: the IPA has historically encouraged adherence to their typeface, more or less. while your font currently does not deviate from their character shapes in general, your shapes of < w > & < ʍ > have rounded angles which differs from the IPA character. but of course, you dont have to agree with the IPA. and apparently the IPA is not being as strict as before (e.g. it accepts either < g > or < ɡ > now). peace — ishwar  (SPEAK) 20:34, 2005 Jun 16 (UTC)
I actually take a little pride on that version of the IPA chart, since I created those fonts on my own, and they allowed me to construct the chart using a colour depth of one bit. Besides, my chart is only an "at-a-glance" version, so the font doesn't have to look exactly the same as that used by the International Phonetic Association. Denelson83 20:39, 16 Jun 2005 (UTC)
Small follow-up: I decided to take your advice, Ishwar, and I made the < w > & < ʍ > symbols a little more straight. Denelson83 01:19, 19 Jun 2005 (UTC)

Deviation from real IPA charts

In my view an article about the IPA (alphabet) should be the one from IPA (association). The recent changes seem to deviate from the IPA publication, such as adding rows (lateral flap, and is nasal stop just a typo?) and columns (alveolo-palatal, epiglottal). These sounds and symbols are recognised by IPA, but not part of the main table.

Splitting the voiced/unvoiced pairs into separate boxes in the table obscures the pairings, especially at the fricatives. −Woodstone June 29, 2005 20:34 (UTC)

Hrm. Internet Explorer doesn't seem to be following the border specifications correctly. It looks right in Firefox. I will experiment a bit and see if I can get it to work correctly in IE and Firefox. Nohat 30 June 2005 00:01 (UTC)


I'm adding many Bambara words to the Wiktionnaire. And I'd like to give an indication of tone. So now I just do what seems right, but I'm not sure. See for instance bananku (I added this one to the English Wiktionary as well). Maybe some expert can tell me if this is right? I.e. "banan" should be low, "n" means nasalisation and "ku" is normal (or high, Bambara has only 2 tones). Guaka 4 July 2005 19:00 (UTC)

Answering on your talk page. kwami 10:27, 2005 July 14 (UTC)


Now that the graphic charts are gone, there are two pages of IPA that uses Unicode (why so?). Can we have an older version of this (IPA) article with the graphics moved to something like IPA_in_graphics or something? --Dara 06:52, July 11, 2005 (UTC)

The IPA in Unicode article will soon be deleted and redirected to here. The graphics links should work. The whole reason for getting rid of the IPA in Unicode article was that there was no substantial reason for having two articles (at least not according to anyone who commented on the move). kwami 03:06, 2005 July 13 (UTC)

Well executed move

You've done a very fine job merging IPA in Unicode with this article. Much praise to all of you who helped out.

Peter Isotalo 22:08, 13 July 2005 (UTC)

Should we put a redirect on the main article, then? I assume the page history stays there? kwami 01:24, 2005 July 14 (UTC)

talk: IPA in Unicode

I've added a table for the consonants I put together a while ago - I believe the old table is incorrect in a number of places, though it does resemble the "official" IPA layout more closely. Maybe it'd be best to change my table to conform to the official table again (i.e. unshade some areas, and possibly remove the epiglottal consonants again, depending on what exactly pulmonic actually means)?

Uhm, there's another table at Talk:International Phonetic Alphabet, and another one at eo:SAMPA.

Prumpf 17:42, 22 Nov 2003 (UTC)

I just cleaned up and merged the pulmonic consonants tables (but I had forgotten to log in: User: is me ;).

Prumpf: pulmonic consonants are those that involve exhalation, as opposed to the clicks (no breath) and implosives (inhalation). No idea about ejectives :/ —Tkinias 19:42, 8 Apr 2004 (UTC)

I'm not sure I wasn't happier with my table layout, to be honest. My original plan was to have wiki links for all the sounds (well, at least those that actually occur), including the ones without symbols, at some point, and having two of them in the same table entry still strikes me as a bad idea (particularly since we currently fail to black out the voiced glottal stop).

Prumpf 01:18, 9 Apr 2004 (UTC)

I've reverted to putting the epiglottals where they belong, in the regular consonant chart. The official IPA chart doesn't have them there because they're relatively new symbols, and it was decided that adding a new column would create too many difficulties with typesetting the table. Plus at first there was some debate about whether any languages used them distinctively (they do), and therefore they were put in the "other" table provisionally while being debated. However, they're perfectly ordinary pulmonary consonants, and leaving them off the chart just causes confusion. There's no need for it on a web site, where we have room for another column or three. It would make more sense phonetically to put the postalveolars (which are not simply postalveolar) or the glottals (which often don't behave like consonants) in the "other" table than to have the epiglottals there. kwami 09:01, 2005 May 29 (UTC)

I added some CSS statements that allow this page to appear correctly in my IE. I assume this is a good thing, although a hacky solution at best. Sorry for donig it in so many separate edits. -Chinasaur 22:29, 1 Jul 2004 (UTC)

Thank you for this page

A big thank you to all who contributed to this page! I use it very often and it already saved me lots of time. - Mark Dingemanse (talk) 11:14, 28 Oct 2004 (UTC)

Should the {IPA} notice be included on this page, since it refers people to this same page if they have display problems?? (Perhaps a manual copy of the template notice, minus the self-reference, and a manual inclusion in the proper category....) [[User:CatherineMunro|Catherine\talk]] 18:08, 5 Nov 2004 (UTC)

Other symbols

  1. kp͡
  2. k͡p

The above shows 2 placements of the tide character as seen in the "Other symbols" section. We are using number 2.

It's a combining double inverted breve.

Number 1 looks much better than the right number2, but the right number2 seems to be the standard (I've checked the Unicode char. table). Which should we adopt??

(I've taken the liberty of restructuring your query: do you mean the tie character?) I say go with the one that actually appears to work, which is number 1. That character is described as "COMBINING DOUBLE INVERTED BREVE" which kind of implies that it combines with maybe two preceding characters, although that might just be my brain exploding. --Phil | Talk 18:06, Jan 7, 2005 (UTC)

[I took the liberty of removing the ins/del formatting above]

I think the problem is that some fonts render double-width combining diacritics in the wrong place, but they can render them correctly if put in the wrong place. Michael Z. 20:56, 2005 Jan 7 (UTC)

A selection of double-combining modifiers, set in your default font:

  • 1. letter-letter-modifier
    • kp͝
    • kp͞
    • kp͟
    • kp͠
    • kp͡
    • kp͢
  • 2. letter-modifier-letter
    • k͝p
    • k͞p
    • k͟p
    • k͠p
    • k͡p
    • k͢p

The following are formatted so that if you don't have the named font installed, then the text will show up in Courier New or Courier.

Arial Unicode MS

  1. kp͡
  2. k͡p

Lucida Sans Unicode

  1. kp͡
  2. k͡p

Lucida Grande

  1. kp͡
  2. k͡p


  1. kp͡
  2. k͡p


  1. kp͡
  2. k͡p

On my machine (both Safari 1.2.4 and Firefox 1.0, on Mac OS X 10.3.7)

  • #1 is rendered correctly by Arial Unicode MS
  • #2 is rendered correctly by Courier New, Lucida Grande, Code2000 and Gentium.

In MSIE/Mac 5.2.3, they all show up as question marks.

In a stock Windows XP installation and MSIE/Win 6.0, they all show up as box characters. Lucida Sans Unicode comes on this machine, but I guess it doesn't include these characters.

I don't actually know which is supposed to be correct, but most of my Unicode fonts seem to like #2 (letter-modifier-letter). Michael Z. 21:18, 2005 Jan 7 (UTC)

Judging by the way the double combining modifiers are illustrated in the Unicode Chart (PDF, page 2), Arial Unicode MS is rendering these incorrectly. The correct way to code these is #2: "letter-modifier-letter". Michael Z. 05:39, 2005 Jan 8 (UTC)
On Windows XP+Firefox 1.0 all #2's are unbalanced; the #1's look fine. On Windows XP+IE6.0.29 (SP2), only the Arial part is working, but again, #2 is bad, #1 looks fine. So for my configuration, it seems that letter-letter-modifier would be best. mark 08:38, 8 Jan 2005 (UTC)
It sounds like Arial Unicode MS is the only one of these Unicode fonts that you have installed, and Lucida Sans Unicode doesn't include these characters. Do you know if you downloaded Arial Unicode, or if it came with some software? To bad it's buggy and can't display these glyphs correctly. Michael Z. 22:06, 2005 Jan 8 (UTC)

Link Individual Phonemes

Shouldn't the individual phonemes in the tables be linked to their articles (and possibly also directly to their sound files)?

--Joe Llywelyn Griffith Blakesley (c)(t) 03:33, 2004 Dec 24 (UTC)

Does anyone have objection to linking the individual sounds to their respective article? Peter Isotalo 22:10, Apr 17, 2005 (UTC)
I linked all the sounds of the main chart to the articles for those sounds to see how it works. Peter Isotalo 20:34, Apr 18, 2005 (UTC)

Move Page Contents?

Shouldn't the stuff on here be moved to International Phonetic Alphabet?

This is not specifically about Unicode at all (as I thought) but just happens to be written in the character set. All OSs and browsers worth talking about now support Unicode. It is the de facto and official standard for WWW pages now. There seems no reason to not move the stuff here to International Phonetic Alphabet.

In fact I found it very annoying going to International Phonetic Alphabet and finding images (that didn't tell me much, that I couldn't search in or copy from &c, that took time to download, &c). It is the images that sould be removed from International Phonetic Alphabet and moved somewhere else (if they are really nessecary).

--Joe Llywelyn Griffith Blakesley (c)(t) 03:48, 2004 Dec 24 (UTC)

  • Oppose. IPA in Unicode is such a useful reference here. I refer to it all the time; it's quite handy. If we had to find this table buried somewhere 70% of the way down a much longer article, I'd cry. QuartierLatin 1968 21:12, 23 Jun 2005 (UTC)
  • Agree. If. The 'if' is that the article must still be convenient for Wikipedists like QuatierLatin (and myself) to clip and paste from. That is, Unicode tables should be at the top of the article. Images should be thumbs for those that need them. (And, pace Joe, there are a lot of people who do.) IPA History should be broken off to a new article, since each of the two present articles is getting too big. That said, there is a reason for this being a separate article, so I don't care much either way. kwami 2005 June 29 00:16 (UTC)
  • Support. Having history contents here just because it's written in Unicode is not right. We now have a lot of IPA's written in Unicode everywhere, (i.e. most of languages, most of phonetic topics, etc.) so the main article for IPA having Unicode characters seems fair. --Puzzlet Chung 29 June 2005 13:54 (UTC)

Spurious contrast between Latin and IPA

In my browser, the characters in the table that happen to be in the Latin character set appear in an ordinary weight, while the ones that are unique to IPA are heavier, as if they were bold. The contrast is quite distracting. This is probably due to a misconfiguration in my browser (which is Firefox). Does anyone else see this contrast? What can be done about it?

ACW 21:25, 26 Jan 2005 (UTC)

IPA is probably handled by a different font, unless your main Latin font happens to be a huge one containing IPA (Arial Unicode MS, Gentium, or similar). This second font is apparently heavier. One way to fix it is to set the font for the IPA range to one that closer resembles your normal font (if Firefox has this option at all), another is to change your normal font to a larger one. Jordi· 22:29, 26 Jan 2005 (UTC)


Is there a standard collation sequence for IPA Unicode?

(In other words: if you have phonetic transcriptions from multiple languages stored in a database system, and you want to sort them, what symbol comes first?)

I suspect it's just by code point. IPA itself doesn't have a "native" collation order. Gwalla | Talk 05:46, 20 Feb 2005 (UTC)
Unicode has collation charts, but they don't include IPA. Going by code point is probably not useful, because that would sort all of the plain-Latin characters before the specialized IPA ones, instead of putting related characters together.
If the context permits, I would collate according to the original text that was transcribed into IPA. I don't know if the Unicode collation charts indicate how to deal mixed writing systems, though. Michael Z. 2005-02-20 06:40 Z

The symbols are actually officially numbered by the IPA. Here's an official IPA chart from 1999 that even includes symbols not recognised by the IPA (but ironically mentioned by them): [5] [6]-- 23:17, 12 August 2005 (UTC)

unfamiliar symbols

What does these Unicode-IPA-symbols relate to?...

  • ɩ - U+0269;
  • ɵ - U+0275;
  • ɷ - U+0277;
  • ɼ - U+027C;
  • ɿ - U+027F;
  • ʅ - U+0285;
  • ʆ - U+0286;
  • ʇ - U+0287;
  • ʓ - U+0293;
  • ʕ - U+0295;
  • ʖ - U+0296;
  • ʗ - U+0297;
  • ʞ - U+029E;
  • ʠ - U+02A0;
  • ʬ - U+02AC;
  • ʭ - U+02AD

Some I could guess what they should mean, but why aren't they included in the official IPA-chart?... Are they older symbols that have been removed, because they weren't useful?... Are they in use although they are not mentioned?... Or does this depend on a disaccord between the IPA and the Unicode Consortium?...

I'm working on the chart de:IPA in Unicode, where I want to list all Unicode-IPA-characters with their description... Someone who knows anything about that?... --Primordial 10:42, 26 Apr 2005 (UTC)

Some are older symbols that were removed. Others are symbols for "disordered speech" (like lisping), which aren't in the basic IPA chart. Gwalla | Talk 01:13, 26 Apr 2005 (UTC)
Thx for that!... Do you know any sources (URLs, books) to find out exactly what these symbols are/were standing for, or rather could you tell me which symbols you know the description of?... --Primordial 10:42, 26 Apr 2005 (UTC)
I've seen /ʆ/ user to describe "щ" in Russian phonetics (the final sound in borscht. I've inquired as to its validity at the talkpage, but have received no answer thus far. Does anyone know if this is used in modern Russian phonologies? If not, what IPA characters correspond to it? Peter Isotalo 13:01, Apr 26, 2005 (UTC)
I don't know russian so well, but polish... To me the final sound in "borscht" sounds exactly like the /ś/ in polish, which in IPA is transcribed as [ɕ]... --Primordial 14:49, 26 Apr 2005 (UTC)
I think that character is an obsolete IPA palatalized consonant (as are a few others with a curl at the bottom). It is equivalent to /ʃʲ/. Michael Z. 2005-04-27 22:08 Z
Going by the Unicode code chart (PDF) here:
  • ɩ - Latin small letter iota, semi-high front rounded vowel, obsoleted in 1989 in favor of ɪ (latin small capital I)
  • ɵ - small barred O, rounded mid-central vowel/rounded schwa
  • ɷ - Latin small closed omega, semi-high back rounded vowel, obsoleted in 1989 in favor of ʊ (latin small upsilon)
  • ɼ - latin small R with long leg, voiced strident apico-alveolar trill (Czech ř), obsoleted in favor of (r with raised diacritic)
  • ɿ - latin small reversed R with fishhook/long leg turned iota, apical dental vowel, used by Sinologists (not standard IPA), IPA spelling (z with syllabic diacritic)
  • ʅ - latin small squat reversed esh (actually ɿ with retroflex leg), apical retroflex vowel used by Sinologists, ʐ̩ (syllabic ʐ)
  • ʆ - latin small esh with curl, palatalized voiceless postalveolar fricative, normally spelled ʃʲ
  • ʇ - obsolete symbol for dental click
  • ʓ - latin small ezh with curl, palatalized voiced postalveolar fricative (ʒʲ)
  • ʕ - pharyngeal voiced fricative (this is on the chart, under "phar.")
  • ʖ - obsolete symbol for lateral click
  • ʗ - obsolete symbol for alveolar or palatal click
  • ʞ - proposed symbol for velar click, withdrawn by the IPA in 1970 (velar clicks being judged impossible)
  • ʠ - voiceless uvular implosive (I suspect you'd use ʛ̥ for this)
  • ʬ - bilabial percussive (smacking lips)
  • ʭ - dental percussive (gnashing teeth)
Some other uncommon IPA symbols are found in the combining diacritics block (PDF). Gwalla | Talk 01:35, 28 Apr 2005 (UTC)
Thank you very much indeed!... That was exactly what I wanted to know!... --Primordial 16:29, 28 Apr 2005 (UTC)

Suffixes for comparatives and superlatives

common -> commoner -> commonest? sounds wrong to my non-native english speaker ears... I thought it'd be common -> more common -> most common? (clem 18:49, 27 Apr 2005 (UTC))

The rule is that -er and -est can only be added to words that have a single metrical foot. Common is a single foot, and the American Heritage Dictionary gives commoner, commonest [7]. Nohat 04:11, 28 Apr 2005 (UTC)
Can but don't have to be. More/most common are also acceptable, and quite ... er, common. Monosyllables generally take -er/-est, except for fun, which for some reason a lot of people use exclusively with more/most. kwami 02:03, 2005 May 29 (UTC)
The fun restriction generally derives from the prescriptive claim that fun is a noun; in fact it would have to be "the most fun", unless of course you were talking about fun in general; As a child, most fun happened outside.Nohat 28 June 2005 06:52 (UTC)

Accent transcription

I know that Unicode currently does not support proper tone and accent notation, but I was wondering if anyone could give me some suggestions on what to use instead. I want to make fairly narrow phonetic notation of the Swedish grave and acute accents (also known as simply accent 1 and accent 2). This is not just a matter of secondary stress as in English, but rather of a different type of word accent that stresses syllables in two distinct manners. In certain minimal pairs most variants of Swedish differentiates between the two accents that are not merely a shift of the stress from one syllable to another, but rather of the actual phonetic pattern. Swedish phonologies transcribe this with the characters for high and low tones, but this doesn't seem to work for all types of IPA characters, and Swedish has quite a lot of vowels that have to be written with IPA.

Any suggestions on how to solve the problem? Peter Isotalo 14:26, May 1, 2005 (UTC)

Do you wish to use tone diacritics such as acute and grave? If so, the problem isn't with Unicode, but with font support. Unicode would require thousands of extra letters for all the possible vowel+tone+phonation+nasalization possibilities. Instead, a decision was taken to rely upon combining diacritic marks, which a properly designed OpenType font would render properly. Unfortunately, there are very few OpenType IPA fonts. One's on sale online for $100, but I don't know how good it is. Non-OT fonts like Arial Unicode do a half-assed job, but are better than nothing. The other option is to design your own font. FontCreator can't handle OT, but you can put the combinations you want in the Private Use area. FontLab will create OT fonts, but is quite expensive. Neither of these will work online, but are great if you're printing something. kwami 02:34, 2005 May 29 (UTC)
I don't know wherein the problem lies, but the tone diacritics work fine when they are used in a table that sets the same fonts as the IPA template. See Template talk:IPA for further discussion.
Peter Isotalo 12:36, May 29, 2005 (UTC)

Non-standard IPA table

changing presentation

Adding an extra column for epi-glottal fricatives seems like a fairly reasonable idea, but I am very skeptical of double notation for some of them as approximants. This is not what this article is for and it does not seem to be supported by how the IPA structures it tables. It's certainly relevant to point this out in the articles for the individual sounds, but please consider not adding double notations of this sort to a standard table. It's really just supposed to be a text rendition of the PDF-files that are officially sanctioned by the IPA.

Peter Isotalo 12:41, May 29, 2005 (UTC)

Maybe we could put them in parentheses? Usually when you use a symbol for something other than its cardinal value, you need a diacritic. Voiceless nasals, for example, require the ring. However, the IPA specifically sanctions the use of voiced fricative symbols for approximants without any diacritic at all. Since no known language distinguishes them in the dorsal region, you will frequently see dorsal fricative symbols used this way, in Spanish and Arabic, for example. Most people looking up something in this chart aren't going to read all the fine print. They would simply see the symbol in the fricative row and wrongly conclude that it's a fricative. This causes a lot of confusion, and I don't know any other simple way of dealing with it than making it overt in the chart itself. And as far as I know, this is the only case where symbols of one cell are used to represent other cells without modification, so it's not like the chart will fill up with duplicate notations. kwami 20:00, 2005 May 29 (UTC)
I don't think this is what the article was intended for. We're not here to brush up on perceived errors of the IPA, but only to describe their terms. If we start making these kinds of adjustmens to the tables, we're not being encyclopedic anymore.
It's not a perceived error of the IPA, just something that isn't obvious from their chart. The IPA intentionally made the symbols ambiguous. Since this is official IPA usage, we should cover it here. That's exactly what an encyclopedia is for. For example, in the 1999 edition of the Handbook of the International Phonetic Association, the IPA chart for Hebrew has two pharyngeals, a fricative [ħ], and an approximant [ʕ]. The [ʕ] is precisely in the cell where it has parentheses around it in this article. It officially belongs there. kwami 04:51, 2005 May 30 (UTC)
Oh, just noticed: [ʕ] is defined as "Reversed glottal stop: Voiced pharyngeal fricative or approximant". kwami
I'm really starting to wonder whether this really is an encyclopedic article. As far as I can tell it's merely a convenient way of rendering the IPA-tables in text format for those who need it. I have myself have had much practical use for it, but Wikipedia is not a usage guide. I think we should consider simply merging this article with International Phonetic Alphabet and include both the text tables.
Peter Isotalo 20:26, May 29, 2005 (UTC)
Yes, this article is exactly that. Several people have suggested merging the two articles, but others have said that would be a mess. I don't much care either way, but I don't see a problem with having two pages, depending on whether or not your browser supports the IPA. It's duplicated by design, not disorder. And if it helps Wikipedians write and edit their articles, isn't that of value? What's the point of merging the articles, if that ends up hindering the expansion of Wikipedia? (If it would hinder it.) kwami 04:51, 2005 May 30 (UTC)
I strongly disagree about any ideas of expansion for its own sake. Any expansion should be qualified and follow our policies. That's what the "Not a general knowledge base"-policy is all about, and that goes for all the other restrictions and demands of quality. If it's only intended to be useful to editors, it can be set up at Wikipedia:WikiProject Phonetics or some place like that. Our normal set of articles should not be meta-tools for editors. Keep in mind that the vast majority of our users are merely passive readers.
That makes sense. kwami
As for your modifications of the IPA-table, I think you need to show some sources that support your claims. You still seem to be refering to actual applied usage of IPA on actual languages, not the symbols themselves. If this is just your own extrapolations (even if they are reasonable) and you want to comment the individual sounds, you can do this in their respective articles. Just click on the links in the table. This article, however, is not the appropriate place for it.
Peter Isotalo 14:10, May 30, 2005 (UTC)
My source is the Handbook of the International Phonetic Association: A Guide to the Use of the International Phonetic Alphabet, Cambridge University Press, 1999. It's probably on the shelf of nearly every phonetician in the English-speaking world, and I'd assume it's published in other languages as well. Although there may be other more detailed documents, this is pretty much taken to be the defining document of how to use the IPA. And of course applied usage belongs in an encyclopedia article. If the IPA defines [ʕ] as being either a fricative or an approximant, people shouldn't have to dig through the individual articles to find that out. Hardly anyone will, because they won't see any reason in our main article to do so. And if our table shows it as only a fricative, then the table is in disagreement with IPA usage and needs to be amended. The table put out by the IPA is inteded for phoneticians, and so can afford to be incomplete or ambiguous where someone with the proper training doesn't need the details. But our audience will not usually have such a background, and we need to be more explicit. If all we're going to do is replicate the published IPA charts, then there's no need for a Wikipedia article at all — we can just provide a link to the IPA website. kwami 19:27, 2005 May 30 (UTC)
I'm well aware of the contents of the IPA handbook and I know that it does not contain the tables you're proposing now. If you're using it to extrapolate your own versions of the IPA tables, you're dabbling in original research alternatively writing a usage guide; both are contrary to policy. If you want to describe the uses of IPA, please make contributions to one of our many phonology articles or one of the many article for individual sounds, but please don't start making new tables just because you feel they are more "correct". Not even if you were a real phonetician (I'm an amateur myself) would I accept it because this is simply beyond the scope of Wikipedia. I'm going to suggest that this article be merged with International Phonetic Alphabet. I frankly don't think the fears of alleged messiness are all that well-motivated.
Peter Isotalo 21:22, May 30, 2005 (UTC)
Sorry, I'm not following you. Which new tables am I proposing? I've added some Ext-IPA symbols, and have started adding historical charts, but I'm not sure what you're objecting to. kwami 22:20, 2005 May 31 (UTC)
Rereading this, I think I understand. I'm not doing original research. What I've done comes out of the IPA Handbook. But there's no reason to slavishly follow the layout of the official tables, when those were designed for hardcopy and compromises were made in order to keep everything legible. Online, we can adjust the size of our fonts and scroll around, so there's no reason to relegate normal pulmonic consonants to the "other" table, especially when the ones so relegated are there because they aren't important to European-language speakers. So I moved the epiglottals and the lateral flap home. These leaves only the doubly articulated consonants and the alveolo-palatals in the "other" table. I've contemplated moving the alveolo-palatals, but that would require a double row for post-alveolar fricatives, since adding another column (as they did in older IPA charts) would be misleading.
As for duplicating the dorsal voiced fricatives in the approximant row, that ambiguity is stated specifically for the pharyngeal fricative/approximant on p 177 of the Handbook. Peter Ladefoged, who's on the IPA Committee, says the same is true for the other dorsal voiced fricatives. Hardly original on my part. And on p 20 of the Handbook it states that the pharyngeal is usually perceived as an approximant, not a fricative! kwami 08:33, 2005 Jun 3 (UTC)
You know, the legibility characteristics of text don't really change just because it's not printed. If a table is confusing in HTML, it's just as confusing on paper. You're not following the IPA handbook, but rather reading the texts on how to apply them and superimposing that information on the table and adding a lot of superfluous information intended merely for phoneticians. Please keep the tables the way the IPA intended them. Again, make these comments in in the articles of the individual sounds.
Peter Isotalo 09:48, Jun 5, 2005 (UTC)
Adding an extra column to a hard-copy page of fixed width may require reducing the font size to get it to fit. Adding an extra column to a web page may just mean that a scroll bar will appear. So yes, the legibility does change when you change the medium.
Secondly, it's silly to clone the Handbook when we have a different audience. Putting ordinary pulmonic consonants under "other" is confusing. Putting approximants in the fricative row is confusing. You're a phonetician, so perhaps you take it for granted. But for a general audience, we need greater clarity than the standard charts provide. People often aren't going to go to the articles for the individual sounds, and even when they do, they'll forget what they read and will need a straightforward summary on the chart. But if you have specific complaints about something I've done that makes the article more confusing, please let us know. kwami 10:30, 2005 Jun 5 (UTC)
I am not a phonetician and neither are you. That's exactly why I deem it inappropriate for either of us to start fiddling with the standards. I'm not trying to assume what peoeple need or needn't know about these tables, but by what the IPA has recommended and I believe it to be a lack of editorial humility to do otherwise.
And I have made specific complaints; making double entries of some of the fricatives in the approximant row is confusing, adding comments that belong in individual articles is confusing and simply adding extrapolated comments is confusing. Very minor comments of a technical nature is acceptable, but not using alternative terms and the likes. Can't you just add a disclaimer saying that details of the characteristics of the individual sounds are more complex than they might appear and that they will be described in detail in their respective articles? I'm sure you can do a very good job of describing the phonetic subtleties there rather than just piling minor comments in this article.
Peter Isotalo 12:16, Jun 5, 2005 (UTC)
I don't think it's a lack of humility to take information from the IPA and present it to the public. And I think it's irresponsible not to address our audience. The IPA charts were made for phoneticians, and that does not include our audience. The IPA placed ʕ in the fricative row for historical reasons, but now specify that it is to be used for both fricative and approximant. Phoneticians understand this. The general public does not. Someone just learning the IPA will be bewildered when the chart says one thing, but common usage is something else. It is our job to clarify such hidden assumptions. If we only parrot info from existing web sites, why bother writing an article at all? Why not just supply a link to the IPA website?
However, I see your point about the comments. I'll see about referring people to the specific articles. kwami 19:02, 2005 Jun 5 (UTC)

hi. I like the new additions. The charts are now displayed in a more logical layout. The current IPA chart that appears in various publications is a result of various political & historical issues. Presentations in general phonetic textbooks make their own phonetic charts, modifying IPA's layout (actually they are not really departing from IPA's layout but rather a common practice layout).

I am especially glad that the extIPA is being expanded (maybe we can then add the VoQS chart, too). I also note that there is no general article on phonetic notation. Maybe more advanced issues can be indicated there (and also alternatives to the IPA can be introduced). peace — ishwar  (SPEAK)

Hi Ishwar. What do you think about puting the alveolo-palatals in the main chart? They really belong there, but since both the alveolo-palatals and the palato-alveolars are post-alveolar in place, this would require placing four symbols (two rows) in one cell. Would this (with an explanatory note) be more confusing than leaving them in the "other" chart, as if they were somehow unusual? After all, the only reason they're there is because they aren't common in Western Europe, where the IPA started. The IPA chart even had a separate alveolo-palatal column for several decades, but that I find confusing, as it suggests one is further back in the mouth than the other. (At least the retroflexes are further back for a few languages, such as Tamil, so a separate column is more easily justified.) If we move the alveolo-palatals, the only consonants left in "other" will be the doubly-articulated approximants and that odd Swedish fricative, which truly belong there. kwami 23:58, 2005 Jun 5 (UTC)
Putting them on the main chart makes more sense to me. And other people do this too. (of course, some would argue that they dont really need to be there since they may not exactly be a unique place of articulation). I note that some, for instance Catford, do describe these as being slightly behind the post-alveolar region. Catford calls them, alternately, dorso-prepalatal or lamino-prepalatal, where his artic. place order is (from front to back): alveolar, post-alveolar, pre-palatal, palatal, etc. But, then Ladefoged & Maddieson call them palatalized lamino-post-alveolar (vs. the domed apico/lamino-post-alveolar, i.e. palato-alveolar). To give a specific language example, Akamatsu's (1997) manual on Japanese phonetics describes a voiceless laminodorso-alveolopalatal fricative where his alveolopalatal region corresponds to both of Catford's post-alveolar and pre-palatal regions (maybe Akamatsu is more in agreement with Ladefoged & Maddieson's palatalized except than Akamatsu does not suggest that the palatalized component is secondary).
But at any rate, if we can consider retroflex a distinct place of artic., why cant we call alveolo-palatal a distinct place of artic. as well — they are both perhaps marginal "places". (I wouldnt recommend leaving them off the chart since many would miss them.)
So what is my answer? I guess that having two rows in a cell may be confusing to some who think of (traditional) places of articulation as unitary things rather than an interaction between an active & passive articulator. Putting alveolo-palatal in a separate place of articulation, although perhaps somewhat misleading, may be easier & less jarring to the IPA's standardized layout. And, as noted above, some descriptions consider them to be between post-alveolar & palatal. So maybe following the older IPA is better (?). — ishwar  (SPEAK) 19:31, 2005 Jun 7 (UTC)
I must say that I found the old system to be very confusing, so my first instinct is to avoid a separate alveolo-palatal column. But my confustion was probably due to the lack of clarification for the rather useless labels. I'll put them in, with a note that it is primarily the shape of the tongue rather than its position that distinguishes [s, ʃ, ɕ, ʂ]. kwami 07:10, 2005 Jun 8 (UTC)

unicode chart vs. graphics

The Unicode page is easier to amend than the graphical charts, and I notice it already has a few differences from them. Perhaps the Unicode tables should be put in the main article, and the graphics can be linked as an alternative version of the tables. On the other hand, the vowel table doesn't replicate all of the features of the graphic. Can that be done using MathML or something? Michael Z. 2005-05-30 21:27 Z
I agree. I don't think we can redo the vowel table with MathML though (and probably shouldn't anyway, since MathML support is rare). Instead, we should just use an image of the vowel chart alongside the text version. Gwalla | Talk 22:54, 30 May 2005 (UTC)
I thought Wikipedia offered most browsers a GIF rendering of the equation. If possible, this would probably still make it more flexible than an image file.
Ah, you mean TeX, not MathML. MathML is one of the ways MediaWiki can render TeX equations in <math>...</math> markup, as is PNG (not GIF). AIUI there is a MediaWiki extention for rendering TeX IPA, but I don't think it's capable of doing arbitrary layouts, and at any rate it's not installed on Wikipedia. Gwalla | Talk 02:28, 1 Jun 2005 (UTC)
Another possibilty is to format the table with CSS, and give it the slanted lines as a background image. This would be the best of both worlds; Unicode text in a table that looks like the original image. One of these days I'll try messing around with that. Michael Z. 2005-05-30 23:29 Z
I've asked the person who objected to the merger in the main article to join this discussion. Personally I think the articles should be merged. But another problem is with the tone markings. Since not a single standard font has Unicode support for contour-tone diacritics or tone letters, that jpg is all any of us have to go on. (If anyone knows where to get a truly Unicode-compatible IPA font, please let us know where to get it!) kwami 08:33, 2005 Jun 3 (UTC)
Might I suggest you visit here and test the MediaWiki extension WikiTeX which includes TIPA support? --Phil | Talk 15:59, Jun 3, 2005 (UTC)
No font with Unicode support has contour tones because the Unicode Standard proper does not even include them. Denelson83 21:45, 13 July 2005 (UTC)
That's a bit like saying Unicode doesn't support tone at all, because it doesn't have separate codings for every vowel-tone combination. A decision was made that sequences of level tones should be rendered as contour tones by the font, just as |a´| should be rendered as <á> by the font, and therefore no special encoding by Unicode was necessary. However, the only font I'm aware of that does this is LaserIPA, and I'm not willing to shell out US$100 for it. (It doesn't look like it even does all the rising and falling combinations - just three each - let alone peaking and dipping tones - it looks like it has just one each!) kwami 01:38, 2005 July 14 (UTC)


Started adding charts from 1887 and 1932. It'll take a couple days. kwami 07:35, 2005 May 30 (UTC)

hi. will be very cool additions (you beat me to it!). peace — ishwar  (SPEAK) 07:55, 2005 Jun 3 (UTC)
Okay, I know the 1932 vowel chart looks bad, but at least it's up. I'll get back to it and the diacritics soon. (It's not exactly the published chart, since I integrated some of the vowels that were left for footnotes in the original.) kwami 06:07, 2005 Jun 10 (UTC)

Moved to International Phonetic Alphabet

Everything from this article has been moved. Does someone want to transfer the talk page as well? I don't know how to go about that, but don't want to turn the article into a redirect until it's done. kwami 03:12, 2005 July 13 (UTC)

talk: merged IPA article

So, how long are we going to have two pulmonic consonant charts in this article?

Or should I just be bold and take one of them out? Denelson83 5 July 2005 15:18 (UTC)

Moving one of them to here: (Woodstone July 5, 2005 15:59 (UTC))

  Labial Coronal Dorsal Radical
Blab. Ldent. Dent. Alv. Palv. Apal. Ret. Pal. Velar Uvular Phar. Epig. Glot.
Plosive p b   t d ʈ ɖ c ɟ k ɡ q ɢ   ʡ   ʔ  
Nasal stop m ɱ n      ɳ ɲ ŋ ɴ  
Trill ʙ   r            ʀ   *  
Tap or Flap * * ɾ      ɽ         *  
Lateral Flap   ɺ      *        
Fricative ɸ β f v θ ð s z ʃ ʒ ɕ ʑ ʂ ʐ ç ʝ x ɣ χ ʁ ħ ʕ ʜ ʢ h ɦ
Lateral Fricative ɬ ɮ * * *    
Approximant * ʋ ɹ      ɻ j ɰ (ʁ) (ʕ) (ʢ)  
Lateral Approximant   l      ɭ ʎ ʟ    
  • Asterisks (*) mark reported sounds that do not (yet) have official IPA symbols. See the articles on the appropriate manner for ad hoc symbols found in the literature.

But now, several consonants are missing (i.e. ʡ, ʜ, ʢ, ɕ and ʑ). Denelson83 5 July 2005 20:40 (UTC)

  • Now fixed. /ɺ/ was also put back. Denelson83 5 July 2005 20:51 (UTC)
It looks to me like there is a bug in the wiki-software. The history seems to be screwed up. I find some changes I made in versions before I did them. The list you added back was already missing quite long. Mysterious. Anyway, it looks ok now. −Woodstone July 5, 2005 21:10 (UTC)
Added the epiglottals and alveolo-palatals back in too. kwami 03:07, 2005 July 13 (UTC)
I do not believe they belong in that table, as I do not think those consonants are pulmonic, especially the epiglottal consonants. Denelson83 04:10, 13 July 2005 (UTC)
They are perfectly ordinary pulmonic consonants. The epiglottals were once considered allophones of the pharyngeals (they were "lower" pharyngeals); they were put in the "other" table while their status was discussed, and left there to avoid the typographic problems of adding another column to the table. The alveolo-palatals used to be in the official IPA chart, but were later dropped. They were dropped mainly for esthetic reasons, because they're the only symbols in their column (for the same reason, the lateral flap isn't normally placed in the main chart). This is a problem of bias; other symbols are left in similarly sparse rows or columns because they're found in Western European languages.
By leaving these symbols out of the main chart, we perpetuate the idea that they're somehow unusual, and mislead people like yourself (and I'm sure many others). The only thing unusual about the epiglottals is that they aren't very common cross linguistically, but actually, when measured in the laboratory, many languages that are supposed to have pharyngeals turn out to have epiglottals instead, so they're not all that rare either. There doesn't seem to be contrastive voicing in epiglottals (the voiced "fricative" is always an approximant, as far as I know), but that's true for the pharyngeals and glottals as well (ʕ is usually an approximant, and ɦ isn't really the voiced equivalent of h). That is, they have some ideosyncracies, like all places of articulation, but aren't otherwise special. kwami 04:57, 2005 July 13 (UTC)
This article, however, is about the International Phonetic Alphabet, not the topic of phonetics in general, so maximal accuracy would be accuracy to the IPA, not accuracy to phonetics. Certainly the points you bring up merit mention, but that does not mean that we should change how the IPA is presented. To do so would be to mislead our readers in how the IPA presents their alphabet. Nohat 05:14, 13 July 2005 (UTC)
But I am advocating maximum accuracy to the IPA. The IPA is presented however is most convenient for the purposes at hand. One chart isn't superior to another; there is no one "official" chart. There are plenty of charts published by the IPA were <c> represents an affricate, or <ʕ> is in the approximant row, or [w] is in the velar column, or where there is an extra column for labial-velars. Our job is to present information as clearly as possible. The main chart on the IPA website is not as clear as possible, and should not be taken as a rigid standard. That's why textbook introductions usually stray from their presentation. If all we're going to do is parrot the Association, then we should just link to their website, and not bother with a Wikipedia article at all. But if we're going to explain the IPA, and how it works, which I think we should, then we need to tailor the presentation to our audience. kwami 06:18, 2005 July 13 (UTC)
But there is one official chart. It's the one published by the International Phonetic Association. I protest the inclusion of the modified version of the chart here! If you want to make your own version of the chart, put it on a user subpage or something. This is the International Phonetic Alphabet, not the Kwamikagami Phonetic Alphabet. Nohat 20:58, 13 July 2005 (UTC)
There are many official charts. The choices of what was included in the main chart had as much to do with typesetting and tradition as they did with phonetics. The chart on the IPA web site is biased toward Western European languages, due to the provenance of the IPA as a method for transcribing Western European languages.
No, there is only one official IPA chart that is published by the IPA. Any other chart is on its face not official. All official IPA publications use the one official IPA chart, which is the one that is posted on their web site; the one that is published in their Handbook, etc. There is only one official chart, and any other chart that claims to represent the official chart is wrong. Cases where actual phonetics differs from the way the chart is constructed should be explained in the article. They should NOT, however, be dealt with by making changes to the chart and sweeping the discrepancy under the rug under the guise of being more sensible to our readers. Nohat 21:21, 14 July 2005 (UTC)

Now this table has taken a liberal turn, I am going to include the w's in the bilabial column. −Woodstone 16:30, July 13, 2005 (UTC)

Wrong move. You have now put /w/ into the space in the table where /β̞/ would go were it not in the diacritics table. The pulmonic consonant table is reserved for singly-articulated consonants. /w/ is doubly-articulated. Denelson83 20:05, 13 July 2005 (UTC)

The IPA handbook has w in that position in the specific table for Hausa. It has it in the velar column for several other languages, which is equally wrong by your reasoning. The alveolo-palatals are doubly articulated as well. They were added in their own column without your objection. The bilabial approximant postion is quite suitable for w, because that is the major articulation point. Most non-specialist people would be quite surpised to find w as a velar. −Woodstone 20:57, July 13, 2005 (UTC)

If we want to be strict, then [ʂ] should be listed as postalveolar, and both [ʃ] and [ɕ] should be removed. They are all postalveolar, but [ʂ] is apical, [ɕ] is laminal (often palatalized, or slightly so), and [ʃ] is domed (often labialized, or slightly so). By the way, none of them are doubly articulated: as far as is currently known, doubly articulated fricatives do not occur in human language. (By the way, laminal consonants often cover more than a single traditional place of articulation: denti-alveolar [usually considered "dental"], alveolo-palatals in many languages, even dental-palatals. This is a common characteristic of laminal consonants in general.)
We should also remove the glottals, since they have no point of articulation (only phonation). They really are "other"; the 'voiced' glottal 'fricative' is neither modally voiced nor a fricative. But there's no reason to remove the epiglottals or the lateral flap, except to follow the Western bias of the IPA. (Why should one, but only one, of the voiceless approximants, [ʍ], get its own symbol? Because it's found in English, and English was one of the founding languages the IPA was designed for. Why should [ɧ] get its own symbol, when so many other phonetic oddities around the world don't? Because it's found in Swedish, and there are a lot of Swedes on the IPA committee. Etc.)
As far as putting [w] in either the bilabial or the velar column, that's a convenience. In fact, many charts put it in both! (Much as "c" is used for [tʃ] as a convenience, and even placed in the palatal plosive cell with that value — a practice that we certainly don't want to immitate.) But it's clearer for our readers to have [w] in a separate chart. There's a good phonetic reason for doing that, unlike the epiglottals or lateral flap.
In case my intent doesn't come across, I'm not actually advocating the removal of the glottals or [ʃ]. Just trying to put things in perspective. And this isn't some ideosyncratic interpretation; I'm following Peter Ladefoged the best I'm able, and he's the editor of the IPA Journal. He clearly describes all of these (well, except the glottals) as normal pulmonic consonants, though each has its own peculiarities. kwami 10:02, 2005 July 14 (UTC)


Let's settle this with a poll. Do you want the official IPA chart in this article, or do you want the comprehensive version as suggested by kwamikagami? Vote by leaving your signature below the appropriate section.

Official IPA chart

I see two possible tacks for this article to take: (1) is to present the IPA as the IPA presents the IPA, reproducing each section of the official chart and then describing. (2) is to present the IPA in a different way, one that completely ignores the current construction of the official IPA chart, and to present all the symbols using some other system, grouping them in a way that is consistent with the contemporary phonetic literature. What I do not see as reasonable, however, is to attempt to do both: present the IPA in the format the IPA presents the IPA, but also to make whatever changes we deem appropriate to make the IPA chart "look like" what the current status of the phonetic literature says about phonetics. We inherit all the awkward baggage of the IPA chart and its somewhat obstuse layout and construction without the benefit of it actually representing what the IPA chart actually looks like. I think we should do both (1) and (2), on separate pages.

My proposal is to move the HTML-Unicode table recreations of the Official IPA Chart to another page, say IPA Chart (This is what the IPA calls the chart, per the foreword of the Handbook). That page will contain just the official IPA chart, without any additions or enhancements, with any appropriate annotations we may wish to add.

This page (International Phonetic Alphabet), then, can explicate the entire alphabet in a format which is completely divorced from the IPA chart, grouping the symbols in whatever way we Wikipedians as a group decide upon is the most accurate and accessible format, and I don't think closely approximating the Official Chart is either of those things. Nohat 00:00, 15 July 2005 (UTC)

That sounds quite reasonable. We already have links to the published chart, so I wasn't worried about that info getting lost. But I've come across too many people who were needlessly confused by its inconsistancy to want to exacerbate those problems. Given the links we already have, I couldn't see any benefits to simply duplicating the chart here, but could see advantages to clarifying its layout.
This would, in effect, take us back to the IPA in Unicode page, which I believe was created to have quick access to the chart for editing purposes.
We had had some discussion there for how to best present an IPA chart. What kinds of changes would you suggest?
kwami 00:38, 2005 July 15 (UTC)
my 2 cents:
I guess I dont know really what to say. In order to understand phonetics, we need a deeper understanding than of just the IPA chart. A general article of this nature does not exist in Wikipedia (although parts are scattered piecemeal throughout). So there is a question of whether to include this information here or to not include it (or how much to include). I dont know if we should represent this information in such a radical way (as Nohat's #2) and call it the IPA (or even relate it to the IPA).
There is a third option which is not such a radical departure as you suggest but rather a slight reorganization of the IPA chart layout that is a little more meaningful and less biased. I think that this is the intention of kwami's work. It is only a slight adjustment, essentially
  • (1) moving appropriate symbols to the main chart,
  • (2) indicating sounds for which there is no IPA symbol,
  • (3) indicating the multiple values of the fricative/approximant symbols.
I dont think that this is misrepresenting the IPA. We can create an article about the IPA and present its symbols in a chart in meaningful way that differs from an "official" chart but which is consistent with the general IPA layout and IPA usage. Indeed it may be that information about IPA could be presented in a way that is better than the way the IPA presents itself to this particular audience (this is the internet after all).
So I guess that I am saying that I agree that #2 should be written but without need to reference the IPA at all and that I have no problem with #3. peace – ishwar  (speak) 00:52, 2005 July 15 (UTC)
I'm glad you like the idea. I have been trying to think of the best way to deal with the conflicting problems of presenting the IPA accurately, but also authentically. I think we may be onto the crux of the matter.

As for the separate page with the IPA chart, I would suggest that we simply replicate the IPA chart using wiki-syntax and css styles as closely as possible, with the exception that the sections are presented linearly in the article, rather than all together in one blob, as in the official chart. Something akin to the way IPA in Unicode was in this [8] revision.

As for how to do it on this page, I would suggest keeping the general divisions that the official chart does: consonants, vowels, suprasegmentals, and diacritics. However, the general problem with the IPA Chart is that it is too dense and compact to provide all the information about each symbol that we would like to provide. Currently, the consonant chart is a rather elegant way to display the manner and place of articulation of the consonants, with a kind of sloppy work-around for voicing, and there is simply no room for any more information.

I would recommend instead construction tables with one row for each symbol, providing all the information necessary about that symbol. We can make each manner of articulation a separate table, laying out all the symbols for that manner of articulation, and explicitly naming them in the text. I would definitely advocate against making wide tables that make browsers on small screen scroll. Perhaps something like this:



Pulmonic consonants



These are the IPA symbols that represent nasal consonants.

symbol sound IPA Name IPA number Unicode Name
m bilabial nasal Lower-case M 114 LATIN LETTER SMALL M
ɱ labiodental nasal Left-tail M (at right) 115 LATIN LETTER SMALL M WITH HOOK
n alveolar nasal Lower-case N 115 LATIN LETTER SMALL M
ɳ retroflex nasal Right-tail N 115 LATIN LETTER SMALL N WITH RETROFLEX HOOK
ɲ palatal nasal Left-tail N (at left) 115 LATIN LETTER SMALL N WITH LEFT HOOK
ŋ velar nasal Eng 115 LATIN LETTER SMALL ENG
ɴ uvular nasal Small capital N 115 LATIN LETTER SMALL CAPITAL N

Of course, the columns are just examples, we could do something totally different if desired. This way, we can put e.g. the epiglottal fricatives in with the rest. Nohat 01:19, 15 July 2005 (UTC)

New IPA chart

Simply linking to the IPA website, and certainly including images of their publication, is sufficient for representing the de facto IPA chart. But that's geared toward linguists, and lots of lay people become needlessly confused by its compromised layout, as is clear from several of the debates above. (And this from people who know enough about the IPA to edit this article - imagine what most of our readers must be going through.)

I like Nohat's suggestion above. Although I don't know what kinds of debates we might get into if we completely abandoned the Association's layout, I imagine some informative subjects may come up, and we might all learn something. (At least I hope to.)

If people don't like Nohat's suggestion, I still think this article should be on the IPA, and not on the IPA Chart. kwami 00:38, 2005 July 15 (UTC)

Trapezoidal vowel chart

Take a look at what I just cooked up. This is the IPA's style of the vowel chart, created using div tags and CSS. This looks correct in my browser (Mozilla Firefox). If you have your text size set to anything other than 100%, this chart will not look correct on your screen. You may judge how good or bad this looks by comparing it with the graphical version. -- Denelson83  01:54, 16 July 2005 (UTC)

IPA vowel chart
Front Near-​front Central Near-​back Back
Blank vowel trapezoid.svg
Paired vowels are: unrounded • rounded
This table contains phonetic symbols, which may not display correctly in some browsers. [Help]

IPA help • IPA key • chart • Loudspeaker.svg chart with audio • view base table • view

It looks beatiful on my browser (also FF), except that the back vowels are aligned vertically.

However, the 100% text size is a fatal flaw: lots of people customize their browsers, and we should not assume how they view the web. A good design is both browser and end-user independent.

However, I do notice that text sizes less than 100% are not a problem. Since you've made the chart viewable at 50%, can't you extend that to 200%?

Also, it would be a good idea to label all the rows and columns, since we are also addressing the linguistically naive. kwami 02:14, 2005 July 16 (UTC)

Like so? And by the way, I set this vowel chart to a specific size because MediaWiki does not know what text size you use in your browser, thus it does not have the capability to dynamically resize the background graphic accordingly.  Denelson83  02:33, 16 July 2005 (UTC)
That is better, but the high back vowels still stack vertically even at 100% (I'm probably using a different font than you are), and at larger sizes, other back vowels do as well. Then they start drifting off the edge of the chart, but at larger sizes than before. (I don't know what the practical limit is for text sizes that people browse with.) kwami
Okay, I have attempted to optimize this chart so it displays well at larger text sizes. But remember, I can't make it look perfect at all text sizes. It's quite a delicate balancing act. You can try adjusting it yourself so that it looks better for you. It's at Template:CSS IPA vowel chart. Anyway, what do you think of the chart now?  Denelson83  03:43, 16 July 2005 (UTC)
I can now increase the text size (Ctrl+) twice from 100% (whatever percentage that gives me), and it still looks good. With a third increase the vowels pop off the background, but they still line up properly. kwami
Does not look good in Safari. The aperture labels are are stacked up on the left and don't align with the trapezoid. You need to specify explicit cell heights for those cells. Nohat 06:24, 16 July 2005 (UTC)
Okay, I've implemented your suggestion. How does it look now, Nohat?  Denelson83  06:56, 16 July 2005 (UTC)
Still looks wrong. I tried fiddling with it and you can get it to work if you specify the cell heights in pixels, but I couldn't get it to align quite right, and I don't have time right now to make it work. Since the trapezoid image is a set number of pixels high, you should be able to calculate how many pixels high each cell should be. Nohat 17:33, 16 July 2005 (UTC)
I fixed it now. It looks marvelous in Safari and in Mac Firefox. Of course it doesn't display correctly at all in Mac IE5, but that's to be expected because it doesn't have any Unicode support. Nohat 18:26, 16 July 2005 (UTC)

I forced the font to IPA by the template {{IPA| }}. It now looks good on IE6. Nice work (I have not figured out how you did it). The only disadvantage is that in many language specific the old style is used, which is much easier to edit. Are the other editors expert enough editors to change them all? −Woodstone 12:24, July 16, 2005 (UTC)

Good idea applying the {{IPA}} template to all the symbols in the chart. By the way, it took me quite a bit of looking through the W3C's CSS specifications, plus a lot of trial and error to put this together.  Denelson83  12:57, 16 July 2005 (UTC)

Semantically this is quite horrid. Without CSS it should be like the old table. ¦ Reisio 15:35, 2005 July 16 (UTC)

You're always welcome to do better! kwami 19:14, 2005 July 16 (UTC)
BTW, despite the other fixes, it still displays correctly on my FF browser at two increases of text size, and acceptably at three. kwami 20:00, 2005 July 16 (UTC)

Okay, I've drastically changed this vowel chart to give it more exacting specifications. I also put in support for browsers set to larger text sizes by creating an article that contains a double-sized version of the same chart. What do you think now?  Denelson83  05:38, 18 July 2005 (UTC)

Haven't checked the large version, but the regular one now displays correctly in FF at three size increases above 100%. I imagine that covers the vast majority of people, but really don't know for sure. kwami 05:56, 2005 July 18 (UTC)

There. Now I've applied another significant improvement to this vowel chart. Right now, I'm kicking myself for not getting this idea sooner. The sizes of the divs that contain the actual vowel symbols have been redefined in terms of em space units, which do change with browser text size, instead of percentages, which don't. Now, the vowel chart will look clear, although still slightly mis-aligned, even if you put it through about five or six text size increases. How's that?  Denelson83  13:55, 18 July 2005 (UTC)

I don't know why you think FF covers the vast majority of users. However, your latest version displays well on IE6 now at all scales. Congratulations! That should surely cover the majority of readers. −Woodstone 15:44, July 18, 2005 (UTC)
I think he meant the vast majority of FF users.  Denelson83  16:36, 18 July 2005 (UTC)
Yes, I meant that a 3+ size should accommodate most people.
The latest change doesn't seem to make any difference in FF: at 4+, the letters overlap and are illegible. However, this may be due to limitations of Wikipedia, for the text is constrained into a fixed column that doesn't allow space for 4+. 3+ looks good, though, as it did before. kwami 18:08, 2005 July 18 (UTC)
I can increase the text size six times in FF and the vowel chart still looks good. Did you get that 3+ figure right after you pressed [Ctrl]+[0] in FF?  Denelson83  18:14, 18 July 2005 (UTC)
Yep, starting from 100%. Perhaps we have different viewing prefs. In that case, anyone who needs to magnify more would have their prefs set accordingly. (And anyway probably has a magnification bar rather than magnifying the entire page.) kwami 18:31, 2005 July 18 (UTC)


I frequently run across articles with haphazard pronounciation keys. Some of these I've rendered into IPA. Others I avoid as I am unsure the correct transliteration. What I want to know is if there is a WikiProject that addresses this problem. If not, one could be created similar to Wikipedia:Translation into English. -Tydaj 20:45, 16 July 2005 (UTC)

Oh, look what I found: {{cleanup-ipa}}. -Tydaj 20:56, 17 July 2005 (UTC)

pharyngeal stop

Here is a paper & related video by Jerold A. Edmondson that claims that there are word-final pharyngeal stop allophones of an epiglottal stop phoneme in Amis:

happy reading – ishwar  (speak) 22:58, 2005 July 17 (UTC)

Sounds like it's an "epiglotto-pharyngeal stop" (in their words). I think Ladefoged's point was that if there is occlusion in the pharynx, the point of contact will be epiglottal before anything else, which this doesn't contradict.
Seems to me (and I may be talking out of my ass here) that the pix indicate a third radical place of articulation, and not really a simultaneous epiglottal stop and pharyngeal approximant as Edmondson's ad hoc transcription [ʕ͡ʡ] would suggest. A "pharyngeal stop" as generally conceived (and which L argues against) would be the (active) root of the tongue occluding the (passive) pharyngeal wall. That's never been observed, so far as I know. In a prototypical epiglottal stop, the (active) aryepiglottal folds contact the (passive) epiglottis. What we have here seems to be an (active) epiglottis occluding a (passive) pharynx.
The "epiglotto-pharyngeal fricative [ʜ͡ħ]" has epiglottal and pharyngeal components, but it's not clear to me whether they're sequential (or even coarticulated) fricatives, or epiglotto-pharyngeal in the same sense as the stop. The article doesn't make that clear either, though there seems to be an assumption that it's the latter. If confirmed, this (or at least the stop) would require another column in the IPA chart - assuming they ever assign symbols to it, which they may not.
Thanks for bringing this up. Shall we create an entry for epiglotto-pharyngeal consonant? kwami 00:16, 2005 July 18 (UTC)

Tone diacritics for non-latin vowel symbols

One of the ways in IPA to indicate tones is by diacritics, like acute (á) and grave (à). On symbols that are represented by latin letters, these can be obtained from the edit bar below the edit box. However, many vowels are represented by special IPA characters like [ɔ ɑ ɛ ɯ]. Does anyone know how to create accents on top of these? −Woodstone 22:30, July 21, 2005 (UTC)

Yes, you use the "combining diacritic marks", which are a dedicated section of Unicode. With most fonts, like Doulos, this will be a little sloppy, and they won't line up perfectly. There are OpenType fonts, which display this combinations properly (with a professionally designed glyph for each combination), but the only one I know of for the IPA is LaserIPA, which will set you back $100.
The "Interactive symbol selector" in the external links at the bottom of the article gives you a clickable IPA chart which you can use as a keyboard. When you click on a diacritic, you get the combining form. kwami 23:31, 2005 July 21 (UTC)

Thanks, very good resource, but we're out of luck: it does not support contour tones (like â and ǎ). −Woodstone 14:19, July 22, 2005 (UTC)

Open a Word document; 'insert symbol' (in the 'Insert' menu); when the window opens, choose a font that support the IPA; scroll down the 'Subset menu' (top right on a PC) until you get to 'Combining Diacritical Marks'; in Arial Unicode, the 3rd glyph is ô (here on an o) and the 13th is ǒ. For more complicated contours, you'll need an OpenType font that supports them, and Arial Unicode ain't it. Even if you get one, almost no one online will have something similar to read it with. Unicode is weird here: they gave separate codings to the sample contour tones on the IPA chart, but not any of the others! Granted, that's supposed to be handled within the font, but unfortunately the IPA isn't a priority for font designers. kwami 19:40, 2005 July 22 (UTC)

In my browser IE6 and with the IPA forcing template, the combining diacritics seem to work well (though less nice on wide symbols like in ɯ̂). Thanks again for the pointer. You might have a look at Thai language where I did some IPA work. −Woodstone 22:14, July 22, 2005 (UTC)

This article is too long

I think it is time to split it into sections. --Grouse 22:31, 28 July 2005 (UTC)

Since we'd want to keep the consonant, vowel, and diacritic charts in a single article, what do you suggest we move? Obsolete glyphs? Names? I don't see much else we could remove. kwami 23:09, 2005 July 28 (UTC)

Alveolo-palatal fricatives

I have a problem with what these symbols represent. According to what I've read about it, it's different from the palatalised postalveolar fricatives. But as an example of a language in which they are used, Polish is listed, while Russian isn't. I speak Russian, which makes polish extremely easy for me to pronounce. The languages are so similar (phonologically, not grammatically) that you can replace the Polish Latin alphabet with the Russian Cyrillic alphabet and cover every phoneme without a letter to spare (with the exception of nasalised vowels and /w/). In the Russian language, palatalisation can occur with any consonant, with the exception of the voiced postalveolar fricative (Russians can pronounce it, it just doesn't occur in the language). It seems Polish does the same, and every palatalised consonant is pronounced the same in both languages with the exception of these pesky alveolo-palatal fricatives (according to linguists). But when I hear people speaking polish, or listen to the sound samples in the Polish language article, the Polish voiceless "alveolo-palatal" fricative is pronounced exactly the same way I, and every Russian I know, pronounce the Russian voiceless palatalised postalveolar fricative (Needless to say, the voiced variant is pronounced the same too). I understand how the palatal plosive is different from the palatalised alveolar plosive, but the rest of the palatals seem to be nothing more than the palatalised form of other consonants. For example: Palatalised alveolar nasal and palatal nasal; palatalised velar fricative and palatal fricative; palatalised alveolar lateral approximant and palatal lateral approximant. The alveolo-palatal fricatives only differ from the palatal fricatives in that the front of the tongue is raised in order to articulate in what sounds exactly like the postalveolar region. So basically it's an articulation occuring while the back of the tongue is raised up against the palate. That's the definition of palatalisation, so it seems ridiculous to claim that it's different. Would somebody please tell me why I'm wrong? If I'm right, then it seems that there isn't enough collaboration with linguists outside of France and England within the IPA. -- 12:29, 14 August 2005 (UTC)

As far as I know you're right about the alveolo-palatals (and maybe the velars). This is briefly discussed in Postalveolar consonant, but the alveolo-palatal articles need to be cleaned up to conform with it. (Maybe something also needs to be done with the velar article?) My understanding is that alveolo-palatals are palatalized postalveolars, while palato-alveolars (most commonly called simply postalveolars) are partially palatalized postalveolars. In some languages there are non-palatalized postalveolars, which are written [s̠]. That is, there are three degrees of palatalization in the postalveolar region, and a palatalized postalveolar is a alveolo-palatal: [ʃʲ] = [ɕ]. (If something different is claimed, it might be a cluster [ʃj].)
Part of the reason the postalveolars and dorsals are given separate symbols and names when palatalized may be because the primary articulation is more affected by palatalization than other places are. For example, a palatalized labial [fʲ] is clearly co-articulated: there are two distinct closures at the lips and the palate. But you cannot have a palatalized velar in the same sense: Both places are dorsal, and therefore cannot be separately or independently articulated. They do not involve separate articulators the way a palatalized labial does. Although I cannot be sure, I believe that if you palatalize a velar you get a true palatal: [xʲ] = [ç]. (Either that, or it's not truly palatalized, but merely the consonant cluster [xj], or maybe the pre-velar equivalent of [xj] or the post-palatal equivalent of [ç]?) Is this true for Russian? (Are you a native speaker?) Velar is the place immediately behind palatal. Postalveolar is the place immediately in front of it, so perhaps there are similar complications involving palatalization.
However, I believe you're wrong about the nasal and lateral. The problem is that there are not any (or at least not many) languages that contrasts palatalized alveolars with palatals, except maybe for fricatives. So people get sloppy in their descriptions. However, the distinction is noticeable cross-linguistically. For instance, in Madrid the sound ñ is a palatal [ɲ], while in Mexico City it's a palatalized alveolar [nʲ]. This is part of what distinguishes a Mexican from a Castillian accent. Palatal laterals are rarer, but it seems likely that the same distinction could be found.
Also, I understand that the voiced alveolo-palatal fricative does occur in Russian, though it's rare. When found, it's written double-zhe, as in жжешь [ʑos̠] thou burnest. kwami 23:30, 2005 August 14 (UTC)
Thank you for clearing that up. I've always noticed that the Russian postalveolar fricative is far less palatalised than French and English, but the same symbol was always used so I began to assume that they're the same. Is the subscript macron used to represent depalatalisation? If so, then why do you use 's' instead of 'ʃ'?
You are right about the palatalised velar fricative. It doesn't exist as a phoneme in Russian, but does exist as an allophone of the voiceless velar fricative. When it is followed by the vowels /i/ and /je/, it becomes /çi/ and /çe/ respectively. In modern Greek, the same thing happens, even with the voiced velar fricative. I believe that you are correct to believe that the articulation becomes solely palatal. And yes, I am a native speaker (specifically of the Ukrainian dialect).
I understand what you're saying about the nasal and lateral palatals being different from mere palatalised variants, but when I try to pronounce the nasal palatal without the Russian bias of alveolar co-articulation, I seem to be pronouncing a palatalised velar nasal. Now that I understand how they differ from their co-articulatory counterparts, it really bothers me that so many people incorrectly use them. I think the problem is that people who learn the IPA in order to describe a certain language don't adequately understand it the way a professianl linguist does.
Yes, you're right about the voiced alveolo-palatal fricative. I did not even think of that. When I was trying to think of an example, all I could come up with was French loan words like 'jury' (жюри) in which the soft vowel does not palatalise the previous postalveolar fricative.-- 01:59, 15 August 2005 (UTC)
[s̠] means a retracted/back alveolar. That is, postalveolar (but without being palatalized, since [s] is not palatalized). [x̠] would be a post-velar (that is, somewhere between a velar and a uvular). Etc. Ladefoged transcribes Polish sz as [s̠], and I've seen it for Russian ш as well.
When you say the Russian ш is less palatalized than English sh/French ch, is it also less labialized? That's what I understand.
Yes, I can see how a palatal nasal might feel like a palatalized velar. (It doesn't to me, quite, but then [ŋ] is phonemic in English, and palatalization is not.)
Careful when reading phonemic transcriptions. Since a phoneme is an abstract concept, you can use any symbol you like. In practice, poeple tend to compromise between a broad phonetic transcription for the most universal allophone, and the convenience of using common symbols. So, for example, even in official publications the IPA may use <c> for [tʃ]. You could use <q> if you wanted, or the numeral <4> (for "phoneme #4"?), or even a picture of an eyeball. So you can't read too much into the symbols that are used. Also, if people don't consider a distinction to be important, they often won't bother with it. If you use <c> for [tʃ] and call it a "palatal stop" (there's actually an example of this in the IPA Handbook!), then you're certainly not going to bother with the difference between English and Russian! It's all very subjective, even among professional linguists. What you want is a detailed verbal phonetic description, down to the minutest detail. Then you can select the elements you're interested in. Unfortunately, it's difficult to know how careful the phonetician has been. kwami 02:32, 2005 August 15 (UTC)
Ooh, I didn't realise that it was a retraction sign, now I understand. I never really thought about labialistion for the postalveolar, but now that I think about it, there is a very slight labialisation in English (I don't know about French though, since I don't speak it). It always bothers me when Americans try pronouncing my Russian short name (Sasha) because the slightly palatalised postalveolar fricative followed by the schwa sounds very feminine (which is probably why it's a popular female name in America), so I prefer them to call me Alex.
It always upsets me when I go from one source to another when I'm trying to learn a language and the pronunciation is never consistent. This is especially true of the vowels. Isn't there some place that has thorough thought put in the use of IPA, or do we have to create it ourselves?-- 11:16, 15 August 2005 (UTC)
Many phoneticians attempt to keep to a standard description. Unfortunately, they tend to work at cross purposes, with the languages they're familiar with as the standard. I understand that Peter Ladefoged trains his students with cardinal IPA vowels, which don't sound exactly like the vowels of any real language. However, it's rare for phoneticians to mark explicitly how the vowels of their language differ from the cardinal vowels, and they tend to use the nearest IPA symbol unless they're seriously divergent.
Vowel charts are much more useful. If the vowels are lined up with the lines of the chart, you can assume they're not precise, but if they're scattered all over it's likely they include more detail than the IPA symbols, even with diacritics. A thorough phonetician will measure vowels adjacent to various consonants, measure multiple tokens from multiple people, and create separate charts for men and women. You will sometimes see charts where the vowels are represented by large oval outlines; these typically include all tokens within one standard deviation of the mean for each phoneme. There may even be separate ovals (or distinct marks for the vowel points) for vowels after [p, t, k], etc. However, this is a lot of work, so few go to the effort. Also, the ovals are often replaced by points at their mean when the chart is published, so you don't see which vowels are variable, and which ones overlap. And when you do find good, complete data, it's usually only for a few languages, and may not include the ones you're interested in. kwami 19:53, 2005 August 15 (UTC)
I may have been mistaken to suggest that the palatal fricative is the same as the palatalised velar fricative. The Russian palatalised velar fricative is in fact co-articulatory. It just sounds very similar and is easy to switch from one to the other while speaking quickly. The distinction is so slight that a Russian speaker would not even notice the difference.
As for the Russian/Polish (possibly Slavic) postalveolar fricative, I'm not sure if it is correct to use the symbol 's̠'. The Russian postalveolar fricative is actually retracted further than the English postalveolar fricative. The difference is actually in the shape of the tongue. In Russian, the tongue is much more apical, which is why there is a lack of palatalisation. In English, the tongue has a laminal position, in which the tongue nears the palate producing the effect of slight palatalisation. I'm not sure if this is a fact, but it is definitely the case when I personally experiment by comparing the languages I speak.--Potupchik 20:00, 16 August 2005 (UTC)

Font issues

I'm having a lot of trouble seeing unicode characters in this website. I use "Arial Unicode MS" but the characters are always invisible, and the only way I can see them is by copying and pasting them. Some of them even display incorrectly (boxes) even though I know that some of those characters are supported by this font. It's only this website in which I have this type of problem. I'd appreciate any advice, thanks.

I also have trouble with the subscript linking bar (absence of a break). It shows up correctly, but when I use it with other characters, it displays a box.--Potupchik 20:11, 16 August 2005 (UTC)

Template Issues

What's going on? The template seems to be translating the IPA template as {{{1}}} ... --Chris S. 11:38, 20 August 2005 (UTC)

It's fixed now -- Tim Starling 15:41, August 20, 2005 (UTC)
Here, maybe, but not in all articles. ¦ Reisio 16:03, 2005 August 20 (UTC)
Don't feel obliged to keep the identity of these articles a secret. And note that action=purge should fix it. -- Tim Starling 00:58, August 21, 2005 (UTC)
They seem fixed. :) ¦ Reisio 01:10, 2005 August 21 (UTC)

Help! What is going on with the IPA template? In the tenth (?) paragraph under the heading Description there is a sentence which goes like this: "For example, in some dialects the English word pretzel in a narrow transcription would be [ˈpʰɹ̥ʷɛʔt.sɫ̩], which notes several phonetic features that may not be evident even to a native speaker." In Internet Explorer I now see the transcription as three letters (p, t and s), one period (.) and nine identical squares instead of just a number of the IPA pronunciation symbols. This change seems to be going on all over the Wikipedia, in places which very recently looked quite fine.

Due to Mediawiki changes, the old code in Template:IPA was being broken. Template:IPA has been updated and should now continue to work right. For details, see Template talk:IPA#Font declaration has been moved to monobook.css.

Any places where this technique has been used for a table or div on a page will have to be changed manually. Short explanation: put in class="IPA". For details, see Template talk:IPA#Font declaration has been moved to monobook.css. Michael Z. 2005-10-4 15:58 Z


The article should have some discussionof the EXTENT that IPA is used. For example, as mentioned i\on this discussion page, it is largely unknown among lay people in the U.S. Also, U.S. linguistics working on American Indian languages rarely use it. Nor is it necessarily taught in the U.S. in courses on U.S. and general linguistics.

Split the pulmonic consonant table?

I think the pulmonic consonant table should be split into two tables, one below the other. One table should show the "front", or labial and coronal consonants, and the other should show the "back", or dorsal, radical and glottal consonants. I believe it might look better that way, and also be more manageable.  Denelson83  00:51, 11 September 2005 (UTC)

IPA always has the table as one block. Splitting it would obscure the unity of the rows. What do you mean by more manageable: for the reader or the display? I few versions ago I had condensed the width of the columns by splitting the headers more. For example writing alveolo-palatal (two lines) as alve-olo-pala-tal (four narrower lines). The table looks more compact that way. I do not know if it is possible to have text sideways (text running upwards), that would look better still. −Woodstone 08:47:43, 2005-09-11 (UTC)
BTW, "alveolo-palatal" really is a double, not a single, articulation. It doesn't belong in the main table. And the table columns really should be the same width.  Denelson83  16:38, 11 September 2005 (UTC)
Yes, there is no alveolopalatal column in the chart published by IPA. They have them in a separate chart for double articulations. As before I am for sticking to the IPA version. However there has been quite a debate on this specific column. You may want to revive that. Meanwhile I will add more soft returns to enable narrower columns. Making the columns equally wide is difficult in view of the many browsers, fonts and screen sizes. It might turn out really ugly on some combinations. −Woodstone 16:51:17, 2005-09-11 (UTC)
Actually, they're coarticulated, but not doubly articulated. The problem is that the "postalveolars" are also coarticulated; the true postalveolars are the retroflex consonants. At one time both palato-alveolars and alveolo-palatals had columns in the chart. The most justified approach phonetically would be to put both in the coarticulated chart (I need to change the name), but I think people would have a fit if we did that. (It would help with your formatting issues, though!) The reason the former are still in the main chart is a combination of historical residue and their importance in the western European languages (which are related, of course), just like leaving out the epiglottals. kwami 05:51, 2005 September 12 (UTC)


Surely to goodness this should be a featured article? -- ALoan (Talk) 00:28, 30 September 2005 (UTC)

WP self-ref

International Phonetic Alphabet#See also includes

  • IPA is preferred in Wikipedia articles, however.

(which BTW is good "news for everyone here in the castle", thanks for relieving us finally of the unlearnable SAMPA blots on articles). Is this not a violation of our advice to "avoid self-references [to WP]"? IMO it belongs on both this talk page and (especially prominently) on talk:SAMPA and someplace in the WP:MoS, but not in this page's article.
--Jerzyt 16:11, 9 November 2005 (UTC)

Simple Question

How many letters are there in the IPA? I think that many casual readers would like to find this information in the lead. Seabhcán 12:28, 14 November 2005 (UTC)

As in just standalone characters (i.e. the consonant and vowel symbols), or including diacritics and prosodic symbols? If you are referring to only the standalone characters, there are 107. Taking the diacritics and prosodic symbols into account, that number rises to 158, and that's not counting the various contour tone posibilities.  Denelson83  20:26, 14 November 2005 (UTC)
How many sounds can the alphabet represent? Seabhcán 09:30, 15 November 2005 (UTC)
Hundreds of thousands, I'd guess. A lot would depend on what you allow as possible sounds. kwami 11:19, 15 November 2005 (UTC)
How many "Segment (linguistics)" can it represent? Seabhcán 11:46, 15 November 2005 (UTC)
If you do not count combinations of vowels and tones, then I'd guess tens of thousands. (I was assuming 150 tones.) Don't really know, though. kwami 00:15, 16 November 2005 (UTC)
Actually, considering only the standard IPA, there are about 100 segmental symbols, and 30 diacritics for them. Assuming (and this is just a guess) that on average each segmental symbol can sensibly take an average of 3 diacritics at a time (they can take more, but not all combinations are possible), that would make 100×30³ = 2,700,000 segments. So lets just say "millions". kwami 00:22, 16 November 2005 (UTC)

/dɑgz/ or /dɒgz/?

Chameleon uses rounded vowels, while I do not. I suspect he speaks with Received Pronunciation. Perhaps we should represent both points of view here?  Denelson83  19:06, 28 November 2005 (UTC)

I don't think it's worth it. It's just an illustrative example. When I wrote it, I used my pronunciation, which is [ɑ]. However, there's no reason we can't use RP [ɒ]. When transcribing a language I generally try to keep as many phonemic distinctions intact as possible, so I'm sympathetic to Chameleon. Either way will work, but trying to accommodate both will just make the passage difficult to read. kwami 19:52, 28 November 2005 (UTC)
I don't think either it's worth it, so I've searched another sample that doesn't suffer from this ambiguity: pets and beds. I like the dogs-cats sample, and I second Kwamikagami's reasoning, but we don't need a controversial sample if there can be uncontroversial ones. -- j. 'mach' wust | 20:03, 28 November 2005 (UTC)


I'm redoing the images of the consonant table, to take care of the shading discrepancy, and while I'm at it, adding the new symbol. kwami 21:03, 5 December 2005 (UTC)

Okay, I'm posting the pulmonic consonant chart first, since I haven't finished the rest. (I want to add more contour tones, since the Doulos SIL font now supports tone letters.)
I've changed the layout to an alternate version sometimes seen when teaching historical linguistics. Stop rows remain adjacent, but the plosive, fricative, and approximant rows are now also adjacent. (In order to do this, I separated out the lateral rows at the bottom.) There are four benefits to laying the chart out this way:
  • Lenition processes (plosive → fricative → approximant) are now easy to follow by moving down the chart;
  • The raising/lowering diacritics now correspond better to moving up & down the chart (trills need to be below fricatives for this reason);
  • The symbols that are linked to signify affricates are now adjacent;
  • The symbols that officially stand for both fricative and approximant no longer need to be duplicated.
I've also entered the "glottals" as both fricative/approximant, though of course they're really neither and properly belong in an 'other' chart.
Comments? kwami 00:53, 6 December 2005 (UTC)
Okay, full chart is up. More examples of contour tones, and diacritics classified by function. One possible error: the upstep & downstep symbols perhaps shouldn't be raised this high. kwami 01:52, 7 December 2005 (UTC)
The new chart could use some anti-aliasing.  Denelson83  03:02, 7 December 2005 (UTC)
It was created with Acrobat's automatic filtering. Here's a version with up filtering instead. kwami 07:26, 7 December 2005 (UTC)

What's the source of the "Light grey letters are unofficial extensions of the IPA"? ― j. 'mach' wust | 12:43, 7 December 2005 (UTC)

These are the ad hoc symbols found in the various articles and linked from the hypertext chart, where they're marked by asterisks. I'm trying to keep the two charts in sync. Some, such as the labiodental plosives and the strident diacritic, are found in Maddieson & Ladefoged's SOWL; others are from publications on languages such as Toda & Iwaidja that have extra-IPA phonemes. These aren't Americanist or anything, just IPA transcriptions where an extra symbol or two was needed. (I once created an extended Keil IPA font with some of these symbols for a phonologist writing on Toda, so there is a call for this.) I need to remove the hook-w. It's included in the new Doulos IPA font and I stuck it in at the last minute while preparing the chart, but now that I review it I see that the proposal for the labiodental flap suggested using an advancing diacritic on the labiodental for the bilabial rather than a separate symbol.
The cells that are linked in the hypertext chart but blank in the PDF chart are ones I've never come across symbols for, or which I've only seen indicated with diacritics, such as the voiced lateral fricatives. (Archi is illustrated in SOWL with a raised small cap el for a voiced velar lateral fricative, for example, so that hasn't been included.) kwami 19:08, 7 December 2005 (UTC)