Wikipedia talk:Manual of Style/Pronunciation/Archive 1

From Wikipedia, the free encyclopedia
Jump to navigation Jump to search
Archive
Archives
Archive I Archive II
Archive III Archive IV
Archive V

See also: Wikipedia talk:Manual of Style

What about pronounciation rules? Some pages (Guernica, for example) have a pronunciation guide. In the current format, these usually clutter the first (definition) sentence, and it looks ugly. A common practice in encyclopedias is to put the pronounciation in italic immediately after the word is first mentioned. In our case, we'd have to link it as well, so something like s{mpA:? Any other suggestions? Jeronimo 14:05 Oct 10, 2002 (UTC)

Including pronunciation is quite hazardous, I think. Whose pronunciation do you use? - Khendon 14:16 Oct 10, 2002 (UTC)

I don't find SAMPA the least bit communicative, but people have complained about the handmade phonetics, as in Guernica (which I added). I prefer the latter, as communicating more information to more people despite the issue raised by Khendon. Look at the SAMPA and concise SAMPA and see how much you get out of it. I don't think SAMPA is all that well known, either. Ortolan88
The problem with "handmade" phonetics is that there's no standard. Different writers will split different phonemes and transcribe them differently. I agree that SAMPA isn't very well-known. I would much rather have IPA, which is -- all British dictionaries use it. However, AFAIK, we can't display IPA in HTML. :( The advantage of SAMPA is that it is one-to-one correspondance with the IPA -- so when one day HTML can do IPA, it won't be hard to convert it. -- Tarquin
We can display IPA in HTML, using numeric character references. [1][2] But, it's A) ugly in source, B) not easy to type, and C) they don't show in all browsers; a lot of people don't have the fonts, or don't have them configured. --Brion 20:24 Oct 10, 2002 (UTC)
The "handmade phonetics" (as at Guernica, indeed) are ugly and often far from correct. Pronounciation should be denoted in a correct way, and an informal method won't do for that I guess. I don't care which method is used, as long as there's a page where I can find what the signs mean (you have to do that as well with your dictionary, so this shouldn't be a problem). If SAMPA's easy to display, let's go with that. Jeronimo
Just a whimper here, but at least one native Spanish speaker has edited Guernica several times since I put the "Gayr NEE ka" pronunciation in there, which is much clearer (and therefore more accurate) than "SAMPA [gernika]", which doesn't even show where the stress goes, one of the most important aspects of pronouncing Spanish. Ortolan88
SAMPA does support stress, through a mark placed in front of the stressed syllable: [ger"nika] [[User:Poccil|Poccil (Talk)]] 18:15, Jun 28, 2004 (UTC)

My opinion is that it should depend on what sort of word we're "pronounciating". If it's a difficult English word, then we should use handmade English "phonetic" spelling, taking care to include exactly the same dialectical ambiguity in our phonetics as in the word itself. For example, if there's a question about whether a certain /r/ should appear in the word, then we should spell it phonetically with an "r" whose possibility of pronunciation in a given dialect would match. OTOH, if it's a foreign name, then we should use IPA (or an ASCII equivalent such as SAMPA) to render the phonemes of a standard dialect associated with that person or place. — Toby 14:04 Oct 28, 2002 (UTC)

SAMPA/To Do has a list of pages with pronunciation guides on them, if we ever decide on a solution. Martin


Here's another new user question: How does one indicate pronunciation in an entry. Is it acceptable/desirable to do so? I haven't find any information about this. The example I have in mind is the Arkansas River and the city of Arkansas City, Kansas. The common pronunciation of Arkansas is like AR-kan-saw, but we Kansans do things a bit differently. The river starts in Colorado with the above mentioned pronunciation, but as it crosses into Kansas, the pronunciation becomes like ar-KAN-zus. The city name is also pronounced the latter way. The river reverts to the "normal" pronunciation when it enters Oklahoma. Thanks for any advice. Zeaner 00:03 Dec 18, 2002 (UTC)

Even if that's a somewhat vague answer: I think the criterion should be the relevance of the pronunciation. A Frenchman with only a basic knowledge of English asking a Londoner how to get to Beauchamp Place (near Harrods) will never be able to make himself understood, so it may be important to point that out. Also, in a bookshop it can be quite embarrassing if you have no idea how to pronounce names like Carl Hiaasen or Chuck Palahniuk. I don't know if similar misunderstandings could crop up if a tourist said AR-kan-saw instead of ar-KAN-zus. --KF 00:19 Dec 18, 2002 (UTC)

How to represent pronunciation is one of those issues that keeps getting discussed but never resolved. Take a look at Talk:Language and Wikipedia talk:Manual of Style (under the "Pronunciation" section) for some previous discussions. In your particular case, I think your information is very interesting, so I would include it just as you have here, with a homemade pronunciation guide. If the issue of how to represent pronunciation ever gets sorted out, someone can go back and change it. If not, it communicates what you want to say perfectly well. -- Stephen Gilbert 00:38 Dec 18, 2002 (UTC)

The absolute best way to represent pronunciation, of course, is to record a sound file (preferably encode as Ogg Vorbis) and upload it. --Brion

Is pronunciation of a printed proper name or words on Wiki a real issue? If you're from Oklahoma, you might pronounce "oil": all, or from New Zealand Oy-el. Dictionaries do help. Dialects vary all over the world. BF


For what it's worth, I've drawn up a proposal for phonetic spellings for English words that is much simpler than SAMPA, which appears now at Wikipedia:English phonetic spelling. It's there as a draft, mostly, for you to tear apart; I have begun incorporating it into a number of pages about which there was some recent debate at Wikipedia:Pages needing attention. -- Smerdis of Tlön 22:49, 18 Oct 2003 (UTC)

(Moved from Wikipedia talk:Manual of Style)


Phonetic spellings

On Wikipedia:Pages needing attention, Ruhrjung wished for SAMPA to replace all of the various non-standardized phonetic spellings used to indicate the pronunciation of words. I think he has a point. Jtdirl wanted no part of this, observing that SAMPA is much too complicated for use outside of a linguistic context. I think he has a point.

I have put together the rough draft of a proposed new standard, much simpler than SAMPA and based partially on English spellings, partly on the Italian or Spanish values given to the Latin alphabet letters, which appears now at Wikipedia:English phonetic spelling. I am posting about it here to call attention to it and invite everyone to tear it apart.

I will put it out in the open that my chief concern right now is to avoid the confusion that inevitably arises when the Received Pronunciation of British English is taken to be a universally accepted and understood norm. -- Smerdis of Tlön 22:45, 18 Oct 2003 (UTC)

Nohat would be a good person to talk to about this. I would recommend leaving in the standarized pronuciation guides so as not to irritate the linguists. Paullusmagnus 01:47, 19 Oct 2003 (UTC)


I think it is a very bad idea to invent our own system. We will eventually upgrade SAMPA to IPA. They are both international standards, and IPA is commonly used in dictionaries and encyclopedias around the world. -- Tarquin 09:38, 19 Oct 2003 (UTC)
I will cheerfully agree that in an ideal world, IPA would be the way to go. It would be entered easily, without having to have constant resort to a table of numeric codes. It would display accurately on each browser. Perhaps some markup language similar to the <math> label could be devised so that IPA could be easily edited and read, and this added to the software. This is well beyond my limited abilities.
Even IPA has some problems; you run into issues with too-precise descriptions of dialect forms and so forth. Most readers need not be confronted with an unusual character because this is the precise way IPA represents a standard English /r/. But yes, a subset of IPA would be ideal.
Until the millennium, we have fewer options. We can continue to struggle with systemless ad-hoc representations like tee-shogh or YUR-a-nus. These are not always helpful to non-native speakers. We can confront people with scarcely intelligible strings of SAMPA, which may be useful in a linguistic context but is hard on the general reader. Or we can devise our own system. -- Smerdis of Tlön 13:53, 20 Oct 2003 (UTC)
I think that SAMPA/IPA is incomprehensible to most people. (It certainly is to me.) But making up our own system can't be the best solution. Is there not some other system in use that's easier to understand than SAMPA? Axlrosen 15:42, 20 Oct 2003 (UTC)
A Google search for English phonetic spelling systems will yield dozens of proposals, all of which seem to be idiosyncratic performances, just like mine. None seems to be in wide use. Dictionaries that don't use IPA all have their own proprietary systems, most of which seem further to be based on diacritical marks --- hard to type or edit, even if used. A subset of SAMPA might be more useful, or at least more readable, with some standards (avoid non-rhotic transcriptions, don't worry about the quality of intervocalic /t/, a standard representation for syllabic liquids). But then it ceases to be SAMPA and becomes yet another idiosyncratic performance. The differences conveyed by SAMPA uppercase and lowercase letters (e.g. /sh/ is /S/, /dh/ is /D/ in SAMPA) are non-obvious to the neophyte, and a large factor in its unreadability. I fear that if a standard is desirable, we're on our own. -- Smerdis of Tlön 17:16, 20 Oct 2003 (UTC)

---

Moved from Wikipedia:Pages needing attention

For us with other mother tongues than English, the non-SAMPA pronounciation guides are rather hard to comprehend, as we barely remember where in the English speaking world people say so or so, and often can't know how a particular transcription is to be understood. Conversion to SAMPA is the requested attention.

Why not just add rather than convert? IPA would be just as good, or perhaps better.
The difficulty with IPA is that it is very hard to type, impossible to remember the codes, and not everyone is set up to display the characters. If all of these difficulties could be sidestepped, it would be ideal. The only sure way would be to upload a graphic with the IPA writing in it.
I've done some of the conversions. Someone needs to look over some of the Gaelic ones, and there are many more yet to do. I've included a link to SAMPA Chart on each one.
I can foresee problems. The SAMPA strings add serious difficulty even if they add precision. In antihydrogen, someone had edited that to show the pronunciation of bar as /ba:/ (!) This sort of thing should not be inflicted on the unsuspecting. -- Smerdis of Tlön 19:31, 18 Oct 2003 (UTC)

For most people internationally Sampa pronounciation guides are hell on earth. Few people can make head or tail of them. On something like wikipedia, where they come from different cultures that may used them or may never have heard of them, in which people have varying standards and experiences of english, SAMPA pronounciation guides if used at all should take second place to simple basic guides on words. But replacement of previous guides by SAMPA, or giving them priority is ludicrous when so many people don't understand them and won't use them. You might as well put in swahili or aramaic as SAMPA for all the use it would be to most people. FearÉIREANN 20:40, 18 Oct 2003 (UTC)

Although I love precision, I think IPA and SAMPA are too precise (or actually, too absolute). As a Canadian native-English speaker, I've noticed words in Wikipedia and Wiktionary that are indicated in British or US pronunciations, in a way I would never pronounce them. If the reader was completely unfamiliar with the word (exactly when the pronunciation guide is needed) this can be confusing or misleading. It would be clearer to me if there was an indication of the pronounciation's locale (e.g. British or US).
E.g. I interpret the Wiktionary logo's IPA as something like /"wIk Sən ri/. Sounds very English to me; I would say /"wIk Sə ,ne ri/.
The pronunciation guides in my grade-school dictionaries had the advantage that they referred to example words, and didn't indicate specific phonemes. Whether you spoke British, Canadian or US English, their indications showed how you would pronounce a word.
On the other hand, IPA or international SAMPA are great for showing a non speaker of a language how a word is pronounced in a specific dialect of the language. E.g. they can show me how a word is pronounced in Parisian French, or demonstrate to a German how something is pronounced in London English.
I still don't know which method I prefer to see in Wikipedia. I wouldn't mind seeing two or more methods of indicating pronounciation in one place.
Michael Z. 06:58, 2004 Aug 24 (UTC)
Well, I've learned a bit about IPA in the last few months.
A phonological transcription in IPA or SAMPA is a generic sort of transcription showing the phonemes, or sounds which carry significant meaning in a language. You can pronounce the phonemes your own way, whether you speak American, Irish, or Indian English (likewise in other languages). A phonological transcription is conventionally surrounded by slashes /.../.
A phonetic transcription is a very specific transcription, showing the phones. Phonetic transcriptions are used to compare regional accents or describe speech impediments. A phonetic transcription is conventionally surrounded by brackets [...].
Michael Z. 05:58, 2005 Jan 19 (UTC)
I haven't much to add to what I wrote initially. I don't have plenty of arguments or statistics. But I do sincerely contest that most people internationally would prefer transcriptions to English (to which variant of English?) of words which usually aren't transcribed, which I understand is your preference.
Further, one would expect that the non-English words were italized and the English transcriptions were not.
If wikipedia had been intended as an inter-English project it had been another mather, but as there are more non-native English speakers than there are native English speakers willing to read and write on other languages, we are stuck with a situation were English is both a lingua franca and a major mother tongue.
If the pages on matters Irish are intended mainly for native English speakers, I see no reason why not use transcriptions to English there, but is it really so?
--Ruhrjung 06:03, 20 Oct 2003 (UTC)
I have put forward a proposal at Wikipedia:English phonetic spelling for a simplified system that represents all of the English sounds while avoiding the complexities of SAMPA. You may wish to look it over. -- Smerdis of Tlön 22:05, 18 Oct 2003 (UTC)
Forget it! Yet another idiosyncratic system will only add to the confusion. Eclecticology 20:30, 2003 Oct 20 (UTC)

I'd prefer plain spelling of how to pronounce something in articles rather then the "squiggly lines" that is SAMPA. PMA 18:05, 4 Jul 2004 (UTC)

Pronouncing highlighted words

Hello, This is my first time writing at any website, so here goes... I am wondering if it is possible to include the proper/standard pronounciation of a featured (highlighted) word.

Specifically: I clicked on "cloture" from the cover page and was wondering if it is pronounced as it would be in French, or if it has been Americanized. (With the accent on the first or second sylable/hard or soft 'u'?)

(I know there is a dictionary here, as well,... but thought it might be useful to include a quick reference pronounciation so that you don't have to jump around so much; as I find this website so engrossing that I often spend WAY too much time semi-mindlessly grousing from one place to another...)

PS - I'm trying to sign this! Please bear with my/our newbieness! Sincerely, hotdiggittydave (and Lauria)

If you are logged in, you can sign using four tildes (~~~~). Angela. 14:25, 15 Jul 2004 (UTC)
Personally, I love when people include pronounciations; my only beef is when they use obscure systems for writing it. Please feel free to edit articles you know, and include pronounciations. Best wishes, [[User:Meelar|Meelar (talk)]] 15:11, 15 Jul 2004 (UTC)
What do you mean by "obscure systems"? What we definitely don't want is the proliferation of ad-hoc pronunciations such as "pruh-NUN-see-ay-shuns", since they simply don't work for non-English speakers, or those with regional accents, etc. The International Phonetic Alphabet and derivatives of SAMPA are much more useful since they attempt to establish some absolute sound values. SAMPA isn't perfect but until Unicode is totally pervasive it's probably the best we can do. Graham 06:02, 21 Jul 2004 (UTC)
P.S. It's "CLO-chure". HTH.

Thanks for the tips/clarification and Welcome!!! Keep up the great work :-) ```` -hotdiggittydave and Lauria


Pronounciation revert wars

Could someone point official wikipedia policy to 203.220.171.* (see [3], [4], [5] , [6]) and User:Darrien (see [7]). (I know I could do this myself, but I thought I'd bring it to the attention of you pronounciation guys!) There has been constant revert wars on Australia, Melbourne, Sydney for at least the past week.

User:Mark has pre-emptively warned them not to start doing it on Talk:Perth, Australia, and I don't blame him in the slightest! -- Chuq 12:54, 31 Jul 2004 (UTC)

Technology Solution?

Are there any developers here? One thing that would be nice, (if not to complicated to implement), would be something like this: the SAMPA form, (which is easier to type), could be standardized thoughout Wikipedia, which would be given its own special wiki-linking mechenisim, like //@U//, which the server would then visually serve as /əʊ/. Additionally, an individual editor could choose to add a more commonly understandable phonetic, as in //@U n'''''o'''''//. When the user clicked on the link, either:

  • a new window would open, or
  • a dynamic XHTML layer would pop up,

showing something like this:

  IPA : [əʊ]
SAMPA : /@U/

IPA symbols used:
  ə    about
  ʊ    put

SAMPA symbols used:
  @    about
  U    put

Common phonetic transliteration:
  no

In other words, PHP, Perl, whatever, could dynamically parse the characters used, and then display just the info the user needs to figure out pronunciation, along with an optional common version that (at least some) native speakers of English would more readily understand. What do people think?

P.S. Um... my example IPA and SAMPA values above were just sort of randomly grabbed off of WP's articles for them. I don't know if they're accurate. :) Please correct them if I'm off.

func(talk) 19:18, 7 Sep 2004 (UTC)

I'm not sure about the popup with explanations - it's nice, but unnecessary - but think that a tag like <IPA>@U</IPA> rendering as [əʊ] would be fantastic. And this could be done through a Wikimedia extension. --jnothman talk 05:30, 24 July 2005 (UTC)[]

Use of quotation mark for glottal stops, etc.

User:Susvolans replaced a paragraph about rendering the glottal stop in Hawaiian with new material. I have reverted to the older version. The new material was as follows:

The grave accent (`) is also used as a diacritical mark to indicate a glottal stop; however, except in article titles, the left quotation mark should be used for this purpose instead (e.g., Hawai‘i, not Hawai`i or Hawai'i). It is supported in almost all browsers, and can be used in Wikipedia. <!-- Get rid of the rest of the paragraph once the English Wikipedia supports Unicode throughout --> Do not type a left quotation mark directly into Wikipedia. It may display correctly on your browser, but not on others. Instead, use the HTML entity &lsquo; or the numeric form, &#8216; or &#x2018;. Wikipedia in English does not currently support left quotation marks in article titles; a grave accent should be used instead.

The difficulties are that, other than in Hawaiian, the glottal stop is normally indicated by a right quotation mark or a similar mark and the left quotation mark or similar mark mark indicates a pharyngeal consonant found in Semitic languages called there ‘ayan or ‘ain. The symbols resembling quotation marks are commonly used in Latin-letter transliterations of Arabic names and of Hebrew names and words in scholarly transliterations. Accordingly all renderings of glottal stop are not done by a left-quotation mark symbol. Also, two more technical symbols are available in Unicode: U+02BE MODIFIER LETTER RIGHT HALF RING and U+02BF MODIFIER LETTER LEFT HALF RING which are generally preferred by Semiticists for transliterations of Semitic words and names and which are already in use in hundreds of articles in Wikipedia for transliterations of Semitic names.

At the least the above needs to be reworded to recognize this. Personally, it seems to me reasonable to use &lsquo; for ( ‘ ) at the present time for the ‘okina character in Hawaiian. But that is itself not the proper character according to Unicode and Hawaiian government specifications. The proper character is U+02BD MODIFIER LETTER REVERSED COMMA. See the discussion at Okina and the external links there. If Susvolans' recommendation is accepted for Hawaiian, it should at least be mentioned in the recommendation that in future, when systems better support it, the proper ‘okina symbol should be substituted. Or, since of common browsers, only Microsoft Internet Explorer on Windows currently won't find the proper characters if they exist on any font on the system and insert them, should we perhaps now be using the proper character (as is being done in many Semitic transliterations) rather than wait only because Microsoft can no longer be bothered to improve their browser to the standard of other browsers?

Jallan 15:27, 13 Oct 2004 (UTC)

Regardless of MS's immobility, our service is ultimately to our readers, who almost all use IE, and aren't going to change browsers to see glottal stops on Wikipedia. If anything doesn't look right in IE, it's not an option. Derrick Coetzee 16:29, 13 Oct 2004 (UTC)
Almost all Wikipedia readers use IE? What percentage? I would hope they'd realize that if they use a browser that can't handle Hawaiian characters reasonably, their ability to read Hawaiian words will be impaired. Factitious 16:50, Oct 13, 2004 (UTC)
I think he's basing that on the general widespread use of IE. Do you really need statistics about how MS dominates in the browser wars? Unfortunately, most readers (i.e., users of IE) couldn't care less about Hawaiian characters. They come to Wikipedia and see an empty square or a funny looking character and conclude that Wikipedia is broken, not their browser. olderwiser 17:10, Oct 13, 2004 (UTC)
There are hundreds of articles in Wikipedia now that don't look right on most users' versions of IE, perhaps thousands. People who care about special characters generally don't use IE. There have been no guidelines in that area such as character sets for the Latin script, Greek script, Cyrillic script, mathematical characters and so on that must be used for any article in Wikipedia. Accordingly, for example, editors have used IPA characters for pronunciation indications and so forth. There are two extreme attitudes about this on the web (outside of Wikipedia):

1. Bleeding edge: Do things right from the beginning, using the proper Unicode characters and proper XHTML features and expect readers who care to use proper tools. The entire web will catch up soon enough. After all, using anything outside of Latin 1 or one of the other standard 8-bit sets on the web was considered rather daring only three years ago. Things move very quickly. So build for the future, not for the past. People too stubborn to upgrade when upgrades are free will just have put up with a few characters not appearing properly, but often characters they don't care about anyway.
2. Ultra-careful: Don't use any character or any feature that every browser at least as old as Netscape 5.0 onward doesn't support fully. You can always update the page later. For Wikipedia this would mean almost nothing outside of Latin-1.

Wikipedia obviously should generally be somewhere in the middle. But exactly where is probably very contentious. And the target is constantly moving. I expect this is why the matter has not come up in the style manual. It's too hot to touch and too complex. And the genie is out of the bottle already. Many articles, perhaps thousands of articles, exist using characters that don't display on all systems. Yet users and other editors aren't complaining signficantly. I don't recall ever seeing a complaint outside of an occasional individual complaint on a talk page, that was quickly overruled by others or where a compromise was easily obtained. Obviously, then, using such characters is an option, since it has been taken as an option with minimal or no opposition. There is no policy that every Wikipedia article must display to perfection on every system, regardless of browser being used and local fonts available. The problem tends to be self-correcting, in that those users who care about seeing particular special characters are also the ones who have their systems set up to view them properly (or are already used to seeing them displayed as boxes or question marks on other sites). The other users don't care, and so the boxes or question marks that appear don't matter to them. They just know that some characters are being used that don't display properly on their system, which often happens when languages other than their native language are displayed. Nothing odd there at all. All quite normal. Just what one expects. And editors using rare characters are generally restrained in their use.

Wikipedia could set a policy that, for example, until further notice, the normal left-quote and right-quote characters must be used for Semitic transliterations rather than the half ring characters, that dotted t for emphatic t in Semitic languages should be rendered by t. rather than by the proper Unicode character which has been used instead in many Wikipedia articles. But then who would do the fixing, expecially when even a year from now it might be generally conceded that those characters were no longer a problem?<br\><br\>Jallan 18:20, 13 Oct 2004 (UTC)
I pretty much agree with Jallan. I have no problem with restrained use of rare characters, so long as some equivalent that renders correctly in older browsers is also given. I like how this is handled in the Hawaii article, where the "correct spelling" with the special mark is noted and explained, but the more common English spelling is used through most of the article. olderwiser 18:41, Oct 13, 2004 (UTC)
We should be "bleeding edge" and warn people via {{SpecialChars}}. Chameleon 18:54, 13 Oct 2004 (UTC)
Actually, I like Chameleon's idea. Of course even Unicode is incomplete in some respects (such as lacking p-macron used in phonemic Hebrew transliterations), but I generally like using proper Unicode most of the time. But even I'm not perfect, and I find myself using &lsquo; for ‘okina. I suppose I am willing to bend on the half-circle/single-quote issue, going single-quote instead of the ultra-rare characters. As for other characters in Semitic transliterations, they are well supported in fonts such as Tahoma, so I'm comfortable using them. - Gilgamesh 02:17, 14 Oct 2004 (UTC)
I'm not even sure what we are talking about here anymore. If we're talking about providing a precise transliteration or pronunciation guide once in a article, I see no problem with using rare characters. But if you start using those rare characters at every instance throughout the article, that may be a problem. I mean, this is the English Wikipedia and if something makes words in an article illegible to a significant number of readers, that is definitely a problem. olderwiser 02:58, Oct 14, 2004 (UTC)

We were originally talking about User:Susvolans's quite reasonable attempt to replace some material in the style guide which is obviously outdated. Even the Hawaii article now ignores that recommendation. The difficulty was in part that Susvolans spoke of "glottal stop" rather than simply ‘okina and that opened up another set of issues, indeed the whole matter of special characters, however such characters are going to be defined. Also, if the style manual is going to be changed on that point, as indeed it should be, since the character proposed for ‘okina is in fact not the official character, there should be some discussion on the matter of whether people should go through articles containing Hawaiian forms and replace one incorect character (the typewriter apostrophe) by another incorrect character (the left quotation mark) or should immediately start using the proper character.

I don't think we should to try to define general usages here for all unusual characters. It's too big. What characters particular editors wish to use for transliteration depends very much on what they are trying to do and the particular languages they are dealing with and what users might expect ... and so forth. People dealing with Semitic languages should be dealing with issues having to do with transliterating Semitic languages, people dealing with English phonetics should be deciding how to render English phonetics, people dealing with Hindu languages should be deciding how to render Hindu languages, people working on articles on logic should decide what logical notation should be used, people discussing Chinese characters are the ones who can best decide which characters are most likely to display properly. If people working in any such field wish to set up standards for that field they should try to get consnensus themselves. Mostly that is what has been happening.

As to the current obsolete rule about how to render ‘okina, I have two alternate suggestions:

1. Simply remove it. It was an oddity anyway. We don't have similar rules for Semitic transliteration or Hindu transliteration or how to render Icelandic here. Those working on articles displaying Hawaiian forms can do as they wish, as is done with other languages.<br\><br\>2. Change to a modified version of Susvolans' recommendation that refers only to the ‘okina rather than glottal stop in general, which I believe is what Susvolans intended to do. The revised text might read:

To indicate the ‘okina (glottal stop) character in Hawaiian, the left quotation mark should be used (that is, Hawai‘i, not Hawai`i or Hawai'i), though in any article titles where the ‘okina should be displayed, the grave accent must currently be used as a substitute. (The proper Unicode ‘okina symbol is Unicode U+02BD MODIFIER LETTER REVERSED COMMA, but currently this is not available in many fonts and unfortunately display problems are likely to result on many systems if the correct character is used.)<!-- Get rid of the rest of the paragraph once the English Wikipedia supports Unicode throughout --> But do not type or paste a left quotation mark directly into the Wikipedia editor. It may display correctly on your browser, but not on others. Instead, use the HTML entity &lsquo; or the numeric form, &#8216; or &#x2018;.

I'm not pushing this exact text or even the exact recommendation. It is for those working on articles where Hawaiian forms will be displayed to decide whether ‘okina should now be rendered by &lsquo; or by the more correct &#699; (as in Hawaiʻi versus Hawai‘i). They can best balance current problems versus future benefits in using the &#699; character now or using &lsquo; now—knowing that in a year or two at most everything should be changed to the proper form.

Jallan 04:12, 14 Oct 2004 (UTC)

Read m:instruction creep. Restricting the character set other than on a case-by-case pragmatic basis is generally bad, so we should just cut back, possibly including parts of the preceding section. Susvolans 12:26, 14 Oct 2004 (UTC)


Agreed on attempting to restrict character set usage here. However we don't want people to get the idea that the single curly quotation mark characters are totally forbidden for all purposes. How about this:

Characters identical in appearance to left single quotation mark or right single quotation mark are used as letters in some languages and some Latin-letter transliteration systems, for example the ‘okina character in Hawaiian and for ’alef and ‘ayan or ‘ain in Semitic languages. For such purposes, use of the single left quotation and right quotation characters may be acceptable to produce unambiguous display on almost all systems even though these characters are not the recommended ones in Unicode for those purposes. The characters may also be used in discussions about the quotation marks themselves.<!-- Get rid of the rest of the paragraph once the English Wikipedia supports Unicode throughout --> If using left or right quotation mark for such reasons, do not type or paste a left quotation mark directly into the Wikipedia editor. It may display correctly on your browser, but not on others. Instead, use the HTML entities &lsquo; or &rsquo; or the correponding numeric forms: &#8216; and &#8217 or &#x2018; and &#x2019;. If necessary to represent such characters as letters in article titles, the normal straight apostrophe ( ' ) should usually be used in place of the right quotation mark and the grave accent ( ` ) in place of the left quotation mark.

This is general, rather than restricted to Hawaiian and makes it clear that the curly single quotation mark characters may be used in particular circumstances without committing to desireablity of doing so, as something for case-by-case discussion on talk pages and so on. I think it is useful to warn against attempting to use the curly quote characters directly rather than using the HTML entities.<br\><br\>Jallan 14:27, 14 Oct 2004 (UTC)
I agree with this, from Jallan: "As to the current obsolete rule about how to render ‘okina, I have two alternate suggestions:

1. Simply remove it. It was an oddity anyway. We don't have similar rules for Semitic transliteration or Hindu transliteration or how to render Icelandic here. Those working on articles displaying Hawaiian forms can do as they wish, as is done with other languages.<br\><br\> Maurreen 13:52, 17 Oct 2004 (UTC)

I have an idea. Editors of Wikipedia are encouraged to use English, right? However, there are different standards for English, including an independent Hawaiian English standard (which along with Hawaiian is the official language of the State of Hawaiʻi). I propose that Hawaiian English is a perfectly acceptable editorial standard for Hawaiʻi-associated subjects. And part of Hawaiian English is that it mandates a full transliteration of all Hawaiian language names and words, including ʻokina and kahakō in all situations. Provided that a user has the fonts (such as Arial Unicode MS), the {{Unicode|template}} can surround Hawaiian words or phrases so that even obselete IE will be forced to scour the installed fonts looking for one that includes ʻokina, and display individual ʻokina using that font. It is important to note that even English language Hawaiian state government documents and many newspapers and other publications use exact Hawaiian spelling in all situations, to the point where Hawaiian state government buildings have it indicated on their signs. If we are to use English in Hawaiʻi-related articles, why not use English as Hawaiʻi has standardized it? - Gilgamesh 20:51, 22 November 2005 (UTC)[]

Moved from Wikipedia:Manual of Style (pronunciation) 2004-01-20

Pronunciation guides

Why are people putting in those pronunciation guides? They are unworkable in the content of wikipedia. Such guides work where there is (i) recognition of what they mean, (ii) a broad experience of usage of them, (iii) relevant context. Most people writing international english for a non-academic audience run a mile from these things because they are not widely used in much of the world and so in many cultures completely incomprehensible, and because they pre-suppose a clear shared standard of english, which in Wikipedia's case cannot be guaranteed because while for some users it is a first language, for many it is a second or other language that they are not wholly fluent in. The sensible approach in a cultural context where there isn't the culture, comprehension or experience of these guides is to avoid unduly complex pronunciation formulae and explain the pronunciation in basic english of the sort all readers everywhere can follow.

On Taoiseach we are told the word is pronounced /"ty: S'Vx/. Even with a link attached, to many people worldwide it might as well be written in Aramaic for all the use it is to them. Previously, to recognise that many people don't have the practical experience of understanding complex pronunciation guides, they were simply told the office was pronounced tee-shoch (the och as in loch). That version could be followed easily by many people. /"ty: S'Vx/ to many would appear to be complete gobbledigook. FearÉIREANN 19:40, 18 Oct 2003 (UTC)

I agree with you, but I think that it would be easier still if a pronunciation file was attached. Then it would be even more universal (except for the deaf). I don't know how many times I've wanted to know the correct (or accepted pronunciation of a program or project, especially in the UNIX/Linux/GNU world). Dori 19:59, Oct 18, 2003 (UTC)
Many months ago, when I wrote the article for Abcdefghijklmnopqrstuvwxyz, I searched Wikipedia high and low for pronunciation guide standards. I was disappointed not to find any. Instead, I used the pronunciation guide from an American dictionary.
I don't know enough about linguistics to proposal which pronunciation guide standards to utilize for wikipedia standards. But I feel such standards should exist, and should be listed.
Pronunciation files can help, but they should not replace the written word. Such files without written symbols breaks the continuity of reading, and cannot be used by many users.
A metapage called Wikipedia:Pronunciations (or something like that) should be created by people who know what their doing in the subject. I'd imagine it would look like http://www.m-w.com/aschart.htm in form, but not necessarily in content. Kingturtle 21:50, 18 Oct 2003 (UTC)
The written pronunciation should stay. Perhaps I didn't make it clear, but I meant we should add sound bytes/clips for words, phrases, etc. to accompany such written guides. Maybe this could be a whole new site accompanying Wictionary and Wikiquote or it could just bee sound files uploaded to the pedias themselves. Dori 22:02, Oct 18, 2003 (UTC)
For example, have a look at Albanian_language#Pronunciation guide and Common phrases in different languages#Albanian (Albanian) that I just added. Dori 23:34, Oct 18, 2003 (UTC)
I think there;s no hard and fast rule, but generally they are helpful. Perhaps we could work on a standard for them... -- BCorr ¤ Брайен 19:52, 23 Oct 2003 (UTC)

Was it, then, wrong for me (before gaining a username) to append a "Pronunciation note" to Guillotine? My main worry lay in seeming too POV (it's difficult to find many views from orthoepy authorities these days, description having overcome prescription as the lexicographer's preference), but I presented Elster's view—which both amused and informed me—with an attempt at NPOV at the end.

Then there's the pronunciation scheme I used, which is Elster's own. Boldface and capital letters take care of primary and secondary stressed syllables, and, among other things, the scheme is comprehensible to most people. Chris Roy 21:35, 18 Dec 2003 (UTC)

[end of moved material Michael Z. 00:58, 2005 Jan 21 (UTC)]

Chart for editors

Is there an IPA in english chart for editors, that gives the HTML codes we need to use? -- Tarquin 11:08, 25 Mar 2005 (UTC)

Formatting IPA

Most of the pronunciation I've seen on Wikipedia uses the format:

(IPA: /.../).

I think this is preferable to "(pronunciation: /.../)", because:

  • It's shorter.
  • It explicitly names the transcription system (as opposed to SAMPA, "pronunciation guides", etc.).
  • The mouseover tool-tip text provides the full name "International Phonetic Alphabet". The successively revealed combination of abbreviation on the page, full name in mouseover, and link target page are self-explanatory for any reader, Wikipedia novice or old pro.
  • It's already commonly used.

I also think the link target should always be the same page. Choices that I've seen used are:

  • International Phonetic Alphabet
    • It's good to have the target page have the same name as the link text.
  • IPA in Unicode
    • Is this page still relevant? Now that we have template:IPA, should it be merged with the above?
  • IPA chart for English
    • Useful for English words and phrases, but I think consistency is more important, especially since the majority of IPA will probably be used for non-English words.


Alternatively, I'd be in favour of simply inserting the pronunciation /prəˌnʌn.siˈeɪ.ʃən/ inline on its own, like in this sentence. The slashes are suitable and distinctive enough to set it off; maybe colour can be used to emphasize the usage (I have .IPA{color:green;} in my user style sheet). Now that Template:IPA is being widely used, no one seems to be missing the removal of SAMPA. With one consistent format being widely used, it should be self-evident what it is, like in a dictionary or encyclopedia. Or maybe this should wait until IPA has been more widely deployed.

In any case, a short page about IPA should also be added to the Help: namespace.

Anyway, please leave your comments here, and in a few days I'll update the manual page if consensus mandates it. Michael Z. 2005-03-25 18:20 Z

I would suggest making a specific page in Help: for readers about using IPA -- rather like dictionaries have a couple of pages at the front (and like we already have for OGG sound files). It is not a good idea to use colour to hold information in a default stylesheet -- some readers may be unable to see there is a difference, or even see it at all. I agree with you on writing "IPA" in articles before the pronunciation, and linking it to a help page. As for deploying IPA, can we make a TO DO page of articles that need converting to IPA, either from SAMPA or ad-hoc pronunciation guides? -- Tarquin 12:43, 26 Mar 2005 (UTC)
I understand that some readers won't benefit from colour, but there's nothing wrong with adding it when many will. It's not carrying information on its own, it's a supplement to the /.../ and [...] notation for IPA.
Is it acceptable to link to a "Help:" page from the body of an article, or is that considered self-referential?
A good start is a google search for Wikipedia pages containing the word SAMPA (175 results). Even better, pages containing SAMPA but not IPA (102). Michael Z. 2005-03-26 15:24 Z
Colour may make things unreable for some. It's best to avoid it; green on white I think is a combination to avoid. OGG files link that had a link to a help page explaining how to listen to it (Wikipedia:Audio help), this is the same sort of idea: Wikipedia:Pronunciation help -- Tarquin 17:33, 26 Mar 2005 (UTC)
Good reminder not to assume anything about colour vision. I wasn't specifically promoting green, although it seemed attractive but not glaring to me, and looked high-contrast enough, but I can see how it might not be for some. It's probably best to review some resources about web design for colour blindness before introducing colour. I know there are some pages out there with information on the subject, but I don't have the references handy. Michael Z. 2005-03-26 22:48 Z

phonemics versus phonetics

  1. "Slashes should only be used if the distinction between phonemes and phones is important." I prefer the old convention (always use slashes for phonemes, brackets for phones), which is used in all the introductory linguistics/phonetics texts I've seen. Wikipedia is not a pure, hardcode phonetics site. If the International Phonetic Association gives us permission to use slashes, I say we avail of this facility. Most IPA transcriptions in Wikipedia will be phonemic not phonetic, especially as more pronunciations are added for proper names etc; hence the exceptionalism of brackets is a useful flagging device; following IPA recommendations gets us what advantage?
  2. On a related note, do we really need [ɹ] for /r/ in English transcriptions? I mean, I know phonetically it's [ɹ], but I'm talking phonemes here...though with the square brackets that might not be clear ;). I find this symbol very annoying. The Rhotic consonant article itself makes the point "In broad transcription rhotics are usually symbolised as [r] unless there are two or more types of rhotic in the same language. " Joestynes 10:29, 3 May 2005 (UTC)[]
I agree with you about using slashes. It's the established convention here, and I've seen it used elsewhere. Mixing it up according to the official recommendation will be confusing to readers who don't understand the nuance.
I have mixed feelings about ɹ/r. I definitely would apply the simplification only for purely English words, because readers may not know how to pronounce an "r" in another language or even be confused about an English loan-word from another language. Use of "r" in English pronunciations makes the IPA more accessible, but it also suffers from a similar inconsistency as the brackets/slashes convention that is recommended by the IPA standard. I think [ɹ] and [r] are allophones in just about any word in any national variety of English, so this shouldn't really cause any confusion. In a case where it's important to emphasize that a word has a rolled R, square brackets can be used to indicate a phonetic transcription. Michael Z. 2005-05-3 15:53 Z
after giving this a bit more thought, I would say that the "distinction between phonemes and phones is important" throughout Wikipedia, if only to keep it clear to readers who are new to IPA, so using slashes in most places is consistent with the official recommendation. Michael Z. 2005-05-3 16:05 Z
I disagree that slashes are the established convention. I could list many many pages that use square brackets instead. Further, the problem with slashes, as I explain on Talk:English alphabet is that using slashes makes a major implied assertion about the phoneme system of the language being described, which may well be disputed. In particular, it is a common convention in the phonetic transcription of English to transcribe offglides in the "o" and "e" vowels, and to transcribe length for "i", "u", "ɑ" and "ʌ" vowels. Nevertheless, the features of the offglides and the length are considered in many descriptions of English phonology to not be essential features of the phonemes; that is "o", "e", "i", "u", "ɑ" and "ʌ" could be transcribed thus, without any additional diacritics or symbols. Putting slashes around "oʊ" may thereby be an extremely insidious and subtle violation of NPOV policy, by making the assertion that the "oʊ" is a phoneme of English, and "o" is not. Using square brackets makes no such assertions. Secondly, and also significantly, square brackets are the preferred usage of the International Phonetic Association, as described in the Handbook of the International Phonetic Association. Nohat 03:20, 19 May 2005 (UTC)[]

Just looks like boxes in my browser

Whatever code is being used, it doesn't work in my browser, and that's a shame, since it's the single most often used browser in the world (IE). Sorry if you hate it, but Internet Explorer is the overwhelming standard, and to have all the pronunciation text look broken seems to be a real mistake.

"rhymes with"...

Given the strictures concerning different varieties of English, should we be adding "rhymes with" to articles? Many common words are pronounced in very different ways by different people (there's an amusing true story about regional versions of the B.B.C. news at the time of the Klaus Fuchs affair...). --Mel Etitis (Μελ Ετητης) 16:24, 18 August 2005 (UTC)[]

Template:IPA causes trouble in Safari

I'm using Apple's Safari 1.2.4 on Mac OS X 10.3, and this is what a selected portion of a paragraph from Wikipedia:Manual of Style (pronunciation) looks like:

For example the English word pull would normally be shown in a phonemic transcription as {{{1}}}, but if a more detailed phonetic transcription is also included, such as {{{1}}}, emphasing the aspirated allophone of {{{1}}} and the "dark" allophone of {{{1}}}, then the phonemic transcription could be {{{1}}}.

That's right, not a single IPA pronunciation, just a bunch of {{{1}}}'s. Is anybody else aware of this? —MICLER (Talk) 13:27, August 20, 2005 (UTC)

Fixed now. No idea why. Weird. —MICLER (Talk) 19:56, August 24, 2005 (UTC)

Sound file project page

I've been quite active in recording pronunciation files for titles of various articles, such as Dag Hammarskjöld, borscht and Joseph Stalin. I'm not sure if Spoken Wikipedia is the right forum to discuss these matters and I think this page should be reserved for transcriptions. I believe we would benefit from having a separate project page to discuss matters concerning pronunciation files. It could contain general guidelines and suggestions for how to record the files, what standard dialects to use and perhaps lists of articles that contain pronunciation files. Would anyone object if I started a separate project page for recording and uploading pronunciation (sound) files at the current redirect Wikipedia:Pronunciation?

Peter Isotalo 15:40, 1 September 2005 (UTC)[]

Using IPA to the exclusion of all other pronunciation systems

I very strongly object to the exclusive use of IPA as a pronunciation guide. For one, not everyone can read it. My workplace still uses IE 6.0, and a lot of the symbols come up as boxes.

More importantly, the exclusive use of IPA violates the style guide Wikipedia: Make technical articles accessible, which says "Articles in Wikipedia should be accessible to a general audience."

I can't speak for other parts of the world, but in North America, the only people familiar with IPA are linguists and some teachers of English as a second language. Using IPA to the exclusion of other pronunciation-representation systems means average users must go searching for an explanation of every symbol every time they want to figure out how to pronounce something.

Let's say I want to know how to pronounce "Illinois" from the article List of names in English with non-intuitive pronunciations. The IPA pronunciation given is complete unintelligible to someone unfamiliar with IPA. About the only symbols that are obvious are obvious are the consonants "l" and "n." In order to understand the rest, you have to find the IPA chart and keep going back to it. With a long word, it's extremely frustrating. The whole point of an encyclopedia is to be a quick reference.

The problem is that because the IPA is made for the entire world, many or most of its symbols have little similarity to the letters that make the sounds in English. For example, /j/ stands for the consonantal "y" sound, not the "j" sound. The long "a" sound is marked by the IPA symbol /e/.

I'm not trying to knock IPA, which looks like a wonderful creation. But the fact is, hardly anyone in America is familiar with it. Sure, we could expect readers to learn it before reading articles, but editors should be trying to make it easy for readers rather than expecting readers to do all the work.

Why not use the simple, easy-to-understand phonetic notation system taught to children and used in dictionaries aimed at the general public? (At least in America, although I can't imagine 6-year-olds being taught IPA anywhere.) Short vowels -- bat, bet, bit, bot, but -- are symbolized by no diacritic (a,e,i,o,u) or with a curvy mark (ă,ĕ,ĭ,ŏ,ŭ). Long vowels -- bait, beet, bite, boat, boot -- are symbolized by macrons (ā,ē,ī,ō,ū). The "a" in father gets an umlaut (ä). Different systems represent the "oo" in "book" differently, but one method is a u with umlaut (ü). A schwa, of course, is symbolized by an upside-down e (ə). All other sounds are as in English. I don't know what this system is called, but it's what we're used to in America.

(See [8] or [9] for examples of phonetic symbols used in American classrooms.)

The end result is something far easier to understand than IPA -- "il-i-noi'".

I think that not only should editors be encouraged to use the standard phonics symbols in addition to IPA, the style guide should require it. People who don't know IPA (like me) can't go in and add our own translations of the IPA pronunciation, because we might get it wrong. It's up to the people who know how the words are pronounced to make their edits accessible by adding the non-IPA translations. Mwalcoff 22:12, 25 September 2005 (UTC)[]

The majority of all IPA transcriptions are of names and terms in languages other than English. I'm not going to comment on which way to go when it comes to describing the small portion of English place names that have unintuitive pronunciations, but it's certainly a very secondary problem. I'm personally very skeptical to all the list-articles that concern themselves only with listing words with differing pronunciations. I would AfD them at once if I knew there was even a remote chance of getting rid of them.
As for IPA in articles that consist of something other than pure meta-debates, I would recommend that you check out articles like Moscow, Stockholm and Zürich (never mind that they have pronunciation files). These are the articles where transcriptions actually serve a valid purpose, and in these cases IPA is the only reasonable solution.
Peter Isotalo 00:10, 26 September 2005 (UTC)[]
My Random House College Dictionary has "mos kvä'", "stôk' hôlm" and "zŏŏr' ik" (actually, two o's with a single accent breve on top, Random House's symbol for "ü"). The Random House College Dictionary (now Random House Webster's College Dictionary) sells hundreds of thousands of copies every year. Are you saying it's "unreasonable?" Mwalcoff 16:28, 26 September 2005 (UTC)[]
Those all seem to be a mix between the English and other-language pronunciations. The IPA represents native sounds of those languages which are not found in that system: [mʌ'skva], [stɔkːɔlm], [ˈtsyrɪç]. (By the way, the representation for "zo͝or' ik" you suggest will always fail in some people's browsers, because Microsoft's Arial Unicode MS font mis-renders double-width diacritics, so there goes one of that system's touted advantages over IPA)
The problem with the so-called "standard phonic" systems is that they are not standardized in any reasonable way that we could use them. Furthermore, the systems used in American dictionaries are somewhat confusing in that they attempt to transcribe diaphonemes, resulting in the confusing situation where for every dialect there is at least one case where two symbols refer to the same sound. This is confusing and although IPA doesn't really solve this problem, it avoids this by not pretending to be diaphonemic. Nohat 06:41, 26 September 2005 (UTC)[]
OK, I don't know what a "diaphoneme" is, but I don't doubt that the IPA is a far superior system for linguists writing about pronunciation. The problem is that ordinary people (the kind who don't know what a "diaphoneme" is) don't know it. That's why Mirriam-Webster, Webster's New World, Random House, Encarta, Columbia Encyclopedia, etc., don't use it. What we should do is pick one of the "easy" systems to use in addition to IPA for the benefit of non-linguists. Mwalcoff 16:28, 26 September 2005 (UTC)[]
Nohat is refering to English phonemic diphtongs in words like "bite", "bait" and "boat". The "long" vowels here aren't just one sound, but two vowels pronounced as a glide. Passing them off as one sound isn't just confusing, it's incorrect.
And I can't help pointing out that your examples of transcriptions from Random House dictionary are quite far from accurate. The first vowel in Moscow is drawled and is more reminiscent of a schwa not an [o], the second vowel is a more frontal [a] which only exists as the first part of the diphtong in "bite" in (American) English; the last sound in Zürich is not a [k], but a voiceless palatal fricative; the two "o"s in Stockholm are not the same sound as in US English and the "h" has merged with the [k] to form a long consonant, even in very clear speech. The dictionary examples are just simplified transcriptions based on American English, not accurate descriptions of how these names are pronounced in the native languages.
Peter Isotalo 17:10, 26 September 2005 (UTC)[]
Actually, I was referring to the fact that American dictionaries have different symbols for the vowel in father (usually something like ä) and the vowel in cot (usually something like ŏ), which are indistinguishable in all but a tiny group of speakers of American English. Similarly, they have different symbols for the "w" sound in which and witch, which are only distinguished in a small set of dialects. This makes the systems confusing and the dictionaries usually don't provide much explanation for this. Nohat 23:11, 26 September 2005 (UTC)[]
So why do you think that no American dictionary uses IPA? This not being a real-time conversation, I'll have to answer my own question. It's because the editors of Merriam-Webster, New World, Random House, etc., are thinking about the needs of the general public rather than what they personally prefer. I'm sure that when communinicating on lexicological listservs or whatever, they use IPA (or SAMPA), but when it comes time to making a product, they swallow their pride and use something normal people can understand. Mwalcoff 19:24, 26 September 2005 (UTC)[]
American dictionaries don't use IPA solely because of tradition. All non-American English dictionaries as well as most bilingual dictionaries, American and otherwise, use IPA, so obviously IPA is accessible to non-linguists. You'll notice that no two American dictionaries use the same system, and the systems actually have far less in common than you think. Which system should we use? There have been proposals to invent or use similar systems on Wikipedia, but they have all been unsuccessful because they're nonstandard. See my comments at Wikipedia talk:Pronunciation (simple guide to markup, American). If you like, you can make a proposal for a particular non-IPA system to use, and I will explain why it is unsatisfactory. Nohat 23:11, 26 September 2005 (UTC)[]
Are you telling me that in the UK, they teach 6-year-olds the IPA? Or don't they teach phonics in the UK? Or do people learn the phonics symbols and then switch to IPA when they get older? I'm not doubting you, just curious.
Publishers, like all businesses, never do things just out of tradition. They do market research. If readers preferred IPA, they would use IPA. But they don't, because 99% of Americans have never seen IPA before.
This reminds me of the tedious discussions always going on on Wikipedia about metric vs. American measurements. Some people say that because they think metric is more scientific than American units, because only Americans use old-fashioned measurements and because it's easy to get conversion information, that only metric should be used on Wikipedia. Fortunately, the majority sentiment seems to be that there's nothing wrong with using both systems. If metric -- which most Americans have at least a little familiarity with and which is a lot easier to convert than IPA -- does not enjoy exclusivity on Wikipedia, than neither should IPA.
By the way, I have no idea just how widespread IPA familiarity is among the general population outside of the U.S., but when I taught English in the Czech Republic, none of my students ever used IPA to help with their pronunciations. They would record pronunciations using Czech letters, like "bejs-bal." My Czech-English dictionary, aimed at Czech speakers, uses IPA only for sounds that do not exist in Czech (e.g., "šæm" for "sham"). em' wäl'kôf 23:45, 26 September 2005 (UTC)[]
Regardless of the particular merits of IPA, seeing as how there is no standard alternative, I'm not clear what exactly the point of this argument is until someone makes a specific proposal for an alternative system. If there were a good alternative that would be familiar to Americans, I would advocate for it, but there is no such standard system. As such, I can't see supporting anything other than IPA. Nohat 00:23, 27 September 2005 (UTC)[]
I think we're reaching a new record of colons here. I believe that almost all of the pronunciation systems in use in American reference works share the use of macrons for long vowels, breve accents or no diacritics for short vowels, an umlaut for the "a" in father and an upside-down "e" for the schwa. Consonant sounds are always as in English. We would have to decide what to do with the other sounds -- whether to adopt Merriam-Webster or New World or whatever. But doing something like that would mean that American users would only have to refer to a chart for explanations of a few symbols rather than for a whole bunch of them. Mwalcoff 23:01, 28 September 2005 (UTC)[]
This description is a little overly simplistic I think. What about the vowels of boot, but, good, and caught and cot. Do we have separate symbols for the a of father and the o of cot even though for the vast majority of American English speakers, they are the same? What about the s of vision? What about the difference between teeth and teethe? There's no standard, so I can't see any reasonable way to just pick a system. Nohat 00:13, 29 September 2005 (UTC)[]
I guess we'll just have to agree to disagree before the indent runs off the page. I think an imperfect simple system is better than no simple system at all. You obviously feel differently. Mwalcoff 00:32, 29 September 2005 (UTC)[]
I don't think the system is as "simple" as you describe. It's just as complicated as IPA, just without being any kind of standard. Nohat 01:29, 29 September 2005 (UTC)[]

This summer I met a woman visiting the USA and Canada from Taiwan. She brought along a pile of English learning books aimed at Mandarin-speaking older children and adults, and I was quite surprised that they all used IPA for pronunciation. Michael Z. 2005-09-27 04:21 Z

Mwalcoff, you're still seeing things from a strictly US English perspective. Never mind that it leaves out a very large portion of the native English-speaking world. It completely ignores anyone who's not a native speaker of English. I'm very pro-American when it comes to language usage, but I can not accept a phonetic transcription system based solely on American traditions to describe pronunciation in languages other than English. And like I've already pointed out, the transcription of names and places in English-speaking countries is a relativly minor issue in this discussion.
Peter Isotalo 10:08, 29 September 2005 (UTC)[]

"My workplace still uses IE 6.0, and a lot of the symbols come up as boxes.": This is pretty much resolved, because the template {{IPA}} has been developed to work around MSIE's font display bugs, and is widely used in Wikipedia. So /ˌɪləˈnɔɪ/ appears as /ˌɪləˈnɔɪ/.

My Canadian Oxford Dictionary uses IPA, arguably the most respected English-language dictionary in Canada. As others have written, it wouldn't make much sense to use different systems for foreign-language terms and for English. I suspect the grade-school system would fail for many English terms where pronunciation is necessary (e.g., Scottish "loch", the examples in Canadian English), and for many foreign terms used by anglophones around the world (e.g., many French terms commonly used in Canada).

Some of the IPA sounds used for English are already used by older North American pronunciation guides, like the schwa and ash (letter). I suggest you start learning IPA with the IPA chart for English; it's really not more difficult than using the grade-school pronunciation guide which appears at the bottom of each page in older dictionaries. by the way, in IPA /e/ represents the short e sound in "bet". Michael Z. 2005-09-26 17:52 Z

I'm glad they're working on the IE IPA problem. However, according to the IPA chart for English you want me to memorize, /e/ only represents a short "e" in Australian English; in American English, it represents a long "a." Mwalcoff 19:24, 26 September 2005 (UTC)[]
I'm afraid I'm Canadian, so I think I say [e] in many words where folks from the U.S. say [ɛ]. It seems to me that the difference is distinguishable, but too subtle for me to notice unless I'm paying very close attention. Michael Z. 2005-09-26 23:29 Z

I count myself an amateur linguist, and have learned basic syntax, pronunciation, and grammer in about five languages. I own an encyclopedia of language, and I possess five or six dictionarys of English. I have very little idea how to promounce the IPA. It would be "nifty" to have a link to an IPA-to-.wav converter, if such a thing exists, but even that wouldn't work for people who can't or don't want to listen to it. I think "simplified syllables" (ill-uh-NOY) is the best compromise at this point. The "American dictionary phonetic alphabet" (ĭl'ə-noi') is still not basic enough for the non-IPA crowd, IMHO. --24.17.30.163 07:01, 19 December 2005 (UTC)[]

No help for newbs!

I'm not a linguist; I just wanted to show the correct pronunciation for "Fodor's" (the publisher) using the recommended IPA. After an hour of reading and trying, I gave up. It's not that I have a problem with (other) pronunciation keys, it's that the articles are written for those already versed in the subject. After I found simplified key, I either misunderstood or it was incomplete (couldn't find the "er" sound). When it came time to write the pronunciation into the article, I was totally lost!

I'm sure IPA isn't that hard, it's just that the existing resources are of no use for casual readers (especially those who don't have an hour to waste!). As it stands, it is useless for the casual reader. This isn't idle criticism; I'm more than willing to help out and take orders if it helps make IPA easier to understand. Just let me know how I can help. :) —Foofy 00:48, 11 November 2005 (UTC)[]

Suggestion for template tooltip

I like the fact that we use IPA, and read it quite confidently. However, I understand that the learning curve is quite steep. Here is a suggestion; concrete and implementable.

My dictionary has a brief and incomplete IPA guide at the bottom of each page. It looks like this, more or less:

ʃ she ʒ decision θ thin ʧ chip ʤ jar a cat ɑː arm ɛ bed

and so on, some 30 entries in total. Since they are at the bottom of each page, the pronunciations given with the dictionary entries are useful even for those readers who don't have the difference between ʌ and ɔː memorised.

In our current setup, it takes at least one mouse-click to get to that information. That's too far, and actually, most readers won't be helped at all by having been sent to IPA, there's too much information there. The difference between ʌ and ɔː is not readily apparent.

There are lots of ways of helping the reader along. Currently, the IPA template produces a tooltip with the unhelpful text "International Phonetic Alphabet". I want instead to have it display a short, incomplete English IPA guide, with its 30 entries, just as my dictionary. For 99% of the users, that guide will be enough to figure out the pronunciation. It's right next to each and every IPA pronunciation, and doesn't even take a mouseclick. Arbor 08:40, 24 November 2005 (UTC)[]

That sounds a great suggestion - I'm just not sure how it could be done. It would presumably require the tooltip to display the contents of a linked piece of text, instead of simply showing the name of the linked page as it normally does. I've experimented a bit and can't find a way of doing it, but I'm very much a novice at these things. Does anybody know whether it is technically feasible? (If wikisyntax were to support such a feature, I can see it being seriously useful in all sorts of contexts.)
Failing that, a single-click link to a very much simplified, bare-bones table as per your example would be better than nothing (a lot shorter and simpler even than IPA chart for English.) Anything which reduces the motivation to allow non-standard, ad hoc pronunciation guides is to be welcomed—I firmly believe we should stick to IPA exclusively but make it as accessible as possible. Vilcxjo 15:42, 23 December 2005 (UTC)[]

Brackets vs slashes

I think the situation with Cantonese pronunciation guides using IPA is in very bad shape in Wikipedia because of the recommendation to use brackets for phonemic transcriptions.

In Wikipedia, the convention is for [k], [p], [t] to represent unaspirated sounds when transcribing Cantonese, and [kʰ], [pʰ], [tʰ] to transcribe the corresponding aspirated stops. This looks pretty phonetic to me (esp. if our target readership consists of English speakers), but a random reader would have no way to know that this is phonetic, as elsewhere in Wikipedia, the use of brackets do not imply phonetic transcriptions. In articles where no transcriptions contain aspirated stops, the IPA transcriptions become ambiguous and thus rather meaningless. (This IMHO-bizzare convention is now codified into an article of its own, IPA (free style), which I find extremely misleading.)

I think that, in order to bring some sanity back to the pronunciation guides and to avoid misleading people, it is essential to change the guidelines to use slashes for phonemic transcriptions and brackets only for phonetic transcriptions.—Gniw (Wing) 08:41, 17 December 2005 (UTC)[]

Here's another vote in favour of slashes. --KJ 01:05, 10 January 2006 (UTC)[]

RFC

Having received no advice on Wikipedia talk:Requests for comment or Wikipedia:Village pump (policy) on the appropriateness of this course of action, I've decided to go ahead and place an RFC to expand the audience on this issue.

Current Wikipedia policy is that only the International Phonetic Alphabet should be used to represent pronunciation and that any other representations, even in addition to IPA, are discouraged.

The use of IPA and its superiority as an international way of representing pronunciation is not in dispute (at least not by me). The question is whether editors should be encouraged or allowed to use alternative forms of representation, such as pho-NET-ic SPELL-ing or a form of ə-mār'-ə-kən re'-pri-zen-tā'-shən as well.

The disadvantage of the exclusive use of IPA is that it is impractical for quick reference for most users, because (I would assume) most people are unfamiliar with IPA.

As a result, trying to understand how something should be pronounced can be quite frustrating. For example, the given pronunciation for Enver Hoxha is ɛnvɛɾ hɔʤa. Now, I have no idea how "ɛ" is supposed to be pronounced. I have to click on IPA and find the little symbol in a chart way down the page. I get to the page on the open-mid front unrounded vowel. The page explains that "ɛ" is pronounced like the "e" in "bed" as pronounced in "GA," that is, "General American." One down. Then I have to find "ɾ" on the IPA page. That links to alveolar tap. The description of the "alveolar tap" in English is not comprehensible to a non-expert, but I read that it's like the "r" in Spanish, which I understand. If I've downloaded the .ogg codec, I hear a guy saying "rah-ah-rah." I've now spent 5 minutes and can still only pronounce the first name. That's hardly how a reference site is supposed to work.

The alternatives are as follows:

  • Current official Wikipedia style — "IPA: /ɛnvɛɾ hɔʤa/"
  • Alternative 1 — "IPA: /ɛnvɛɾ hɔʤa/. Approximate English pronunciation: /EN-ver HAW-jah/"
  • Alternative 2 — "IPA: /ɛnvɛɾ hɔʤa/. American Heritage: /ěn'-věr hô'-jä/"

(Hačeks are used above because curvy breve accents are not available in the insert box.)

Currently, neither of the alternatives is permissible according to Wikipedia style. I know I am not the only person rankled by this. Other people have complained on Talk:International Phonetic Alphabet. Matthew White's WikiWatch blog makes the same point.

However, my attempts to press the issue on this page have follen on deaf ears. I believe that is because the people who frequent this page do not represent a good sample of Wikipedia users. They may be pronunciation aficianados who do not relate to the hurdles IPA exclusivity puts up for ordinary users. That's why I've decided to try to expand the discussion to a broader audience.

So how can we make pronunciations easier to understand for ordinary users? -- Mwalcoff 02:30, 19 December 2005 (UTC)[]

  • I agree with your assessment. I count myself an amateur linguist, and have learned basic syntax, pronunciation, and grammer in about five languages. I own an encyclopedia of language, and I possess five or six dictionarys of English. I have very little idea how to promounce the IPA. It would be "nifty" to have a link to an IPA-to-.wav converter, if such a thing exists, but even that wouldn't work for people who can't or don't want to listen to it. I think Alternative 1 is the best compromise at this point. Alternative 2 is still not basic enough for the non-IPA crowd, IMHO. --24.17.30.163 07:01, 19 December 2005 (UTC) --moving my comment, per comment below[]
    • Why did you start a new header? Look just two discussions up, and there's another discussion on the exact same topic, which you contributed to! --Improv 12:48, 19 December 2005 (UTC)[]
      • Because RFC guidelines suggest putting a description of the problem on the talk page. This heading was meant specifically to attract RFC readers and to explain why I submitted the RFC. -- Mwalcoff 23:29, 19 December 2005 (UTC)[]

This has been discussed at length before. The American Heritage method suffers from the same drawback as IPA; you need to use a guide to know what the symbols mean, so IPA is better since it has many other advantages. Do note that probably more English dictionaries now use IPA than any one variation of that other system. Even ad hoc praw-NUN-see-AY-shun guides have the same problem for non-native English readers (or is that pro-NAN-si-EH-shan?). We do need a simpler guide to IPA.

In many cases simpler IPA is acceptable, and more readable. I think that in the example [enveɾ hoʤa] would probably have been sufficient to convey general phonemic pronunciation (but I can't say for sure, because I'm not familiar with the name). Michael Z. 2005-12-20 00:24 Z

The question is not whether we should dump IPA in favor of something else. The question is whether we should allow and/or encourage the use of other systems in addition to IPA.
Let's take a look at the options:
IPA. Advantages — Most-accurate, international, familiar to some non-native speakers who have used it in language training. Disadvantages — completely unfamiliar to most people, not geared toward English, difficult to learn quickly, doesn't translate well to some browsers.
pro-nun-see-AE-shun. Advantages — Understood by all native speakers, very quick to use. Disadvantages — Inexact, can't represent foreign sounds properly, may be confusing to non-native speakers (although I would guess that a non-native speaker would figure it out quite quickly).
American system. Advantages — More exact than phonetic spelling, based on English so only a few symbols need to be learned, already somewhat familiar to many American users. Disadvantages — Unfamiliar to non-Americans, not as exact as IPA, no standard system among dictionaries due to copyright laws.
To me, it seems that the best solution is to use both IPA and one of the other two options, as is done on the page Illinois. -- Mwalcoff 00:55, 20 December 2005 (UTC)[]
  • This has been suggested over and over again, but the problems, as you note, are that alternatives to IPA are not precise, there are a lot of them, and they are confusing to people with different dialects. The solution that has come up every other time this has been discussed is to use IPA and only IPA. I think IPA-only is a good idea. I suspect if you get enough eyes, you'll find other people tend to agree. --Improv 07:12, 10 January 2006 (UTC)[]
I agree with Improv. In what way is IPA "not geared toward English"? Basic IPA is at least as intuitive for an English speaker as one of the so-called "American" or "common" systems. IPA can get confusing because it has so many more characters to accommodate the full range of speech (which the other systems can't do), but for most transcriptions of English a basic set of symbols is sufficient.
IPA is not completely unfamiliar to most people (it is used in many English dictionaries), and it is just as useful as the "common" systems are with a simple guide.
"Pro-nun-see-AE-shun"—"AE"? Is that like an old Engish æ, which is equivalent to IPA /æ/? Nope, it's ey, or ay, or maybe ei. Hm.
Do you suggest we use one system for English words and another for foreign words? How do we transcribe the the si in vision? The difference between Canadian and US pronunciations of house, or the Canadian pronunciation of franglais or Canadien? The ch in Bach, or Scottish loch? Michael Z. 2006-01-10 08:07 Z
  • I am in agreement with the OP of this section. IPA may be a better system but it's arrogant to think that the rest of the world should automatically and immediately learn another groups system because it's better. The metric system is superior to the US customary units and Wikpedia is saturated with both without issue. Some would say Esperanto is better langauge design but i don't see anyone hear foresaking English to go to a superior language. Fact is, the point of a pronunciation guide is to offer assistance. If Wikipedia doesn't offer the assistance to the majority of American readers (who assume make up the majority of English speaking people on en.wikipedia) that have never heard of IPA,, much less how to use it, then it's failing as an open free encyclopedia. Pattersonc 22:34, 27 January 2006 (UTC)[]

An additional disadvantage of IPA that hasn't been mentioned is that it's not abstract — a particular IPA symbol refers to a particular pronunciation. However, not everyone pronounces everything the same.

We could simply make them abstract, and declare that /e/ just represents whatever vowel the reader happens to have in the word "pay". The problems with this include that the reader has to learn Wikipedia's standard system of transcription (there's variation between dictionaries and linguists), that it isn't NPOV (which dialect's phones are we using?), and that it will be confusing given IPA's other, non-abstract usages (if I read a Spanish /e/ as my vowel in "pay" I'm going to be quite incorrect).

Or we could not make them abstract, but transcribe different dialects differently. This would be confusing too, if it isn't made clear. I've noticed some articles have used this system for transcribing Australian English, which means American readers (and Australian readers who've learnt the above method of transcription) are going to think that Australians pronounce "Melbourne" as if it was "Male-bun".

I would prefer the American system. It's abstract, and the choice of symbol to represent each abstract phoneme is not based on any particular dialect's pronunciation, but on spelling. And it's not entirely unknown outside of America: www.dictionary.com is a prominent user of it.

Lastly, a problem that is present is all the transcription systems: not only do dialects differ phonetically, they also differ phonologically. Mwalcoff transcribed "American" as ə-mār'-ə-kən above, but I would write it as ə-mer'-ə-kən. He has the Mary-merry merger, I don't. It's better for the reader if mergers are not represented — a Mary-merry-merging reader simply has to learn that 'ār' and 'er' represent the same sound — but this would mean editors with the merger can't write pronunciations in articles. --Ptcamn 00:00, 28 January 2006 (UTC)[]