WikiProject Manual of Style	Manual of StyleWikipedia:WikiProject Manual of StyleTemplate:WikiProject Manual of StyleManual of Style articles
v • d • e This page falls within the scope of WikiProject Manual of Style, a drive to identify and address contradictions and redundancies, improve language, and coordinate the pages that form the MoS guidelines.

This talk page has a tendency to attract repeated discussions of the same themes. Please read through the list of highlighted discussions below before starting a new one: That the IPA is inaccessible to the layman, and that therefore for a quick guide to the pronunciation of Wikipedia entry words we should use pronunciation respellings such as proh-NUN-see-AY-shun. However, such respellings are inaccessible to the non-native English speaker, who generally is familiar with the IPA. There is wide agreement that respellings are acceptable, but should be used in addition to the IPA. To this end, any verbal description or "sound-alike" guide may be used, but there is also a standardized format at Wikipedia:Pronunciation respelling key. That for English words, the IPA should be specific to a particular national standard, and that the national pronunciations should be listed separately. However, listing multiple national pronunciations after every Wikipedia entry word quickly becomes unwieldy, and listing only one leads to accusations of bias. Of course, if a particular dialect or local pronunciation is relevant to the topic, as is often the case with place names, it may be listed in addition to the wider pronunciation. That the current "pan-dialectal" English convention at Wikipedia:IPA for English is arbitrary/unreferenced/original research, and is therefore invalid. However, this is a help key, not an article, and no more needs to follow Wikipedia's no original research policy than the pronunciation guide of any other encyclopedia or dictionary, nearly all of which use in-house conventions. That Latin or Greco-Latin words in fields such as biology, astronomy, mythology and medicine are improperly transcribed. This is a problem because of multiple conflicting conventions of how to pronounce these words in English. In general, the traditional English pronunciation of Latin is the most irregular and therefore the one that should be transcribed; the other conventions are straightforward and can all be covered, simultaneously, by including the Latin or Greek orthography, or Greek in Latin transcription. As long as Latin long vowels ā ē ī ō ū ȳ are indicated, readers will be able to pronounce the word according to the convention of their choice.

Archives
Archive 1, Archive 2, Archive 3 Archive 4, Archive 5, Archive 6 Archive 7

Contents 1 Even scholarly Greek etc. needs IPA 2 should we delete EnPR? 3 Proposed US School transcription 4 Compare the (older) EOD system 5 Audio recordings 6 =IPA criticism 7 Not indicating language 8 Do we need to tell people how to say George W. Bush?

[edit] Even scholarly Greek etc. needs IPA

The nerve of e.g., Comedy#Etymology waxing on and on:

The word "comedy" is derived from the Classical Greek κωμῳδία, which is a compound either of κῶμος (revel) or κώμη (village) and ᾠδή (singing): it is possible that κῶμος itself is derived from κώμη, and originally meant a village revel. The adjective "comic" (Greek κωμικός), which strictly means that which relates to comedy is...

I mean without IPA etc. we average English speakers feel left out. So I slapped a Template:cleanup-IPA on it. But in general, there should be a style guideline saying that it is not fair just saying things in a foreign language, no matter how erudite, without providing help for English speakers. Same for other language Wikipedias.

(True, one may say IPA itself is rather erudite.) Jidanni (talk) 06:08, 10 September 2008 (UTC)

Better is Bat

The word Chiroptera comes from the Greek words cheir (χειρ) "hand" and pteron (πτερον) "wing,"...

Jidanni (talk) 10:12, 12 September 2008 (UTC)

These look like they need transliteration, not IPA. Off the top of my head, I think it would look something like komodia, komos, komi, odi, and komikos. Already covered by WP:MOS#Foreign terms and WP:GREEK. —Michael Z. 2008-09-12 18:12 z

[edit] should we delete EnPR?

We have two templates, {{EnPR}} and {{EnPR2}}, which use AHD rather than IPA. I think that at the least they should be moved to {{AHD}}, but there are comments on their talk pages about deleting them entirely. Should we? If we keep them, we should mention them here. kwami (talk) 02:25, 2 January 2009 (UTC)

Isn't enPR distinct from AHD? JIMp _talk·cont 11:16, 6 January 2009 (UTC)

I couldn't find any differences. Based on the online AHD, it looks like the AHD might not distinguish secondary stress, but I think this is a font problem. (Words like rheumatoid are screwed up online.) Regardless, all the phonemic symbols are the same.

PS. Based on comments at Wiktionary, it seems that it may not be used the same as the AHD for any particular word, but there don't appear to be any differences in the system. kwami (talk) 17:44, 6 January 2009 (UTC)

Aren't we concerned that the American Heritage Dictionary's pronunciation system is proprietary, owned by Houghton Mifflin? Is there a similar established system which is public domain, for example from Webster's 1913 or an OED fascicle?

I see that we have a reference for the COD's “phonetic scheme” and system “without respelling” in Pronunciation respelling for English. If we could confirm that this is the system used in the 1911 edition, then it would be completely public domain. It looks like a suitable replacement (athough I'm not clear on the nature of the two different systems). —Michael Z. 2009-01-06 19:06 z

The enPR designation was developed on Wiktionary as a result of a vote and concern that we don't want a system or name that belongs to another organization. The AHD name is not ours to use. The system is in many ways similar, but does not copy the AHD's system. It draws upon Webster's, Random House, and other systems that use what is sometimes called "schoolbook" pronunciation in the US. If WP's table does not differ from the AHD's, then it should be updated, but under no circumstances may we name it after the American Heritage Dictionary. --EncycloPetey (talk) 01:45, 9 January 2009 (UTC)

I can see some sense in WP's using enPR—close ties between Wikipedia & Wiktionary are a good thing—but what we don't need is another transcription system. If we want something like AHD, let's use enPR, why reinvent the wheel? JIMp _talk·cont 11:56, 10 January 2009 (UTC)

I don't understand—are you suggesting we do or don't use the Wiki-originated enPR system? —Michael Z. 2009-01-10 17:53 z

I'm suggesting that if we use any AHD-like system, it should be Wiktionary's. Thus don't replace {{enPR}} with {{AHD}} but replace {{AHD}} with {{enPR}} ... or delete them both. JIMp _talk·cont 11:13, 11 January 2009 (UTC)

Okay, agreed. (I had thought that WP:AHD and WP:EnPR were two names for one “AHD-like” system). —Michael Z. 2009-01-11 17:13 z

AFAICT the systems are for all intents and purposes identical. Despite several requests, EncycloPetey has refused to give even one of the differences he claims between the two. If there are differences, they must be quite minor, such as deciding whether the vowel of sing is that of seen or sin, and so would merely be a local interpretation of AHD, not a separate system. EncycloPetey writes "we don't want a system [...] that belongs to another organization", but that's exactly what we're doing, unless we agree that AHD cannot own the system (below), in which case the sentence is meaningless. "The [EnPR] system is in many ways similar, but does not copy the AHD's system." Patently false. It might as well be a photocopy of the AHD. Any differences (if they exist) are in the implementation of the system, which are not spelled out in the description of EnPR. The symbols and their stand-alone values are identical. "It draws upon Webster's, Random House, and other systems"—if true, their input is minuscule. This is like scanning the EB, adding an article from World Book, and claiming that the result is not a copy of the EB.

However, AFAIK you cannot copyright a transcription system, because you cannot copyright a writing system. But that's my OR, since I don't know if transcription is legally writing. kwami (talk) 06:09, 12 January 2009 (UTC)

I doubt it's that clear cut. You can copyright your tables describing your own or someone else's writing system, but perhaps you can't copyright the fact that your system has x character represent y sound. But who says that you don't own your writing system if you invented it? And of course you can always have your lawyers send someone letters or take them to court, and the answers won't be clear until an agreement or verdict is reached.

But writing about AHD's system in an article is scholarship, while making use of it, under whatever name, is different. I would prefer to use something which is clearly free because it is out of copyright, or used by many publishers, or placed in the public domain by its creators. And something created by experts would be preferable to our own novel system. —Michael Z. 2009-01-12 17:36 z

International copyright law does not allow one to copyright a writing system. Tolkien's works are AFAIK under renewed copyright, but Tengwar and Cirth are not. Likewise, the Star Trek movies are under copyright, but the Klingon script which appeared in them is not. We cannot copy definitions out of the AHD, but there is nothing keeping us from using their transcription system. I don't think we even need to credit them, though of course since we are a reference it would best to do so.

The AHD rhotic vowels are counter-intuitive, so IMO it would be clearer to use the standard US school system, and stick to non-rhotic vowel letters plus ar for the rhotic vowels. However, that would mean a split between ourselves and Wiktionary. kwami (talk) 19:10, 12 January 2009 (UTC)

I think I see what makes this system unique to Wikipedia: it's "dynamic". That means we can change stuff as we go along—EncycloPetey just redefined /œ/ by splitting off /ë/. Of course, whenever we do that, the transcriptions of all the words with the relevant phoneme become incorrect, so "dynamicity" does not strike me as beneficial. It also means that we're not staying in sync with Wiktionary's enPR (the AHD stress marks and non-AHD foreign vowels differ from Wiktionary), and I had naively assumed that cross-reffing with Wiktionary was the whole point of having an enPR transcription in Wikipedia. kwami (talk) 20:07, 12 January 2009 (UTC)

What's the school system? Is there a reference?

Yeah, if it's dynamic then it's not a standard. Another argument for picking a system instead of developing one while it's already in use. —Michael Z. 2009-01-12 20:41 z

AHD is only useful for native English words and a few borrowed words from French and German. It is biased towards the pronunciation of English in America. It is not extendable because it is fixed upon publication. Wiktionary devised the concept of enPR as dynamic precisely because we were confronted with the fact that (1) we needed to represent sounds that didn't have symbols in AHD's system, and therefore potentially extend any system we had, (2) the name AHD belongs to another dictionary, (3) we had some symbols and notation in use that were not part of AHD's system. We could not use a standard US school system pronunciation because there is no such standard. Systems like enPR and AHD exist in Webster's and in other US dictionaries, and are threrefore sometimes collectively called "schoolbook" pronunciation in the US, but there is no standard US schoolbook system. The particulars of the system vary from dictionary to dictionary and publisher to publisher. --EncycloPetey (talk) 21:28, 12 January 2009 (UTC)

I'm not following. EnPR is also "only useful for native English words and a few borrowed words from French and German". Until you added ë today, the only symbol from another dicionary was a-dot, from Webster's. Why do you need to extend the system? Are you trying to make it universal? As for school systems, I was thinking of what they teach kids in US elementary schools, which use the same vowel letters before R as elsewhere. However, I think it's more useful to keep in sync with Wiktionary EnPR, which we've already diverged from. kwami (talk) 21:45, 12 January 2009 (UTC)

I believe elementary schools use whatever system is in their chosen dictionary, presumably one of the dozen or more represented in “pronunciation respelling for English”. Does the US even have a federal curriculum?

If EnPR is not the same as the AHD's system, then we should immediately remove any statement like “sometimes referred to as the AHD system”, which may infringe on a trademark (distinct from copyright issues).

If new English sounds are encountered, then it makes sense to add symbols for them. But once in use, the symbol for a sound should never change, because that could silently break the transcriptions in hundreds of articles.

EnPR is only intended for English, including sounds used in English for foreign words, right? Or are we planning on extending this to other languages? —Michael Z. 2009-01-12 22:43 z

We can't use EnPR for other languages, because the symbols are defined according to the reader's pronunciation. Any attempt to expand beyond what the reader can pronounce would become a real mess. That's what we have the IPA for.

If we revert the ë, which changed the value of œ, then the only difference from the AHD is the addition of a-dot. This is not a significant difference from the AHD, which for English words is identical to the EnRP. Adding one symbol to the AHD and giving it a different name is IMO plagiarism. We should call it like it is: AHD. kwami (talk) 02:21, 13 January 2009 (UTC)

I've deleted the unused a-dot and moved the key to Wikipedia:American Heritage Dictionary representation, where it belongs: We don't want to be constantly messing with the system, and failure to acknowledge that it's AHD is plagiarism. kwami (talk) 09:41, 16 January 2009 (UTC)

Now I'm confused. If you can't copyright a transcription system, then using it is not copyright infringement, right? (This is distinct from duplicating the dictionary's documentation, which would be.) But I presume that American Heritage Dictionary is a trademark. It's good form to credit the AHD as the source, but it may not be advisable just to slap their name on our system, which could be inferred as a some sort of direct contribution or endorsement. —Michael Z. 2009-01-16 20:12 z

That's a question for the legal folks, but it is their system, so calling it something else to avoid trademark seems like sophistry. Also "AHD" itself does not appear to be trademarked. As such, I don't see much diff between it and "IPA", but I'll ask at Wikipedia talk:Manual of Style (trademarks). kwami (talk) 20:47, 16 January 2009 (UTC)

According to HM's website,[1] only "American Heritage" is trademarked. kwami (talk) 01:09, 20 January 2009 (UTC)

So did the legal folks say that an abbreviation of American Heritage Dictionary doesn't infringe on the trademark American Heritage? —Michael Z. 2009-01-20 15:12 z

Haven't heard back. I've also written to Houghton-Mifflin to get their take on this. AFAIK, an abbreviation is not a trademark infringement. For example, anyone is free to use "Webster's Dictionary", since the trademark is only for "Merriam-Webster's". I think it's parallel to "IPA", which is (I believe) an abbreviation of a trademarked name but freely used. kwami (talk) 19:53, 20 January 2009 (UTC)

That's not really the same. “Webster's” passed into the public domain a long time ago. But I wouldn't just assume that one could get away with publishing a dictionary called Amer. Her. Dict'y or Merriam-W's Word Book without hearing from some lawyers. Our usage is somewhere in the grey area between these examples.

The IPA is a scholarly organization which “provides the academic community world-wide with a notational standard” to “promote the scientific study of phonetics”.[2] I didn't find any specific licensing information on their site, but they clearly want people to use IPA. Houghton–Mifflin is a for-profit publisher, we're not clear on the copyright and licensing issues here, and we don't know what the corporation's opinion of this would be. Contacting them is the right thing to do.

I also imagine that if we can use their transcription system with their name, there is an obligation to treat it as a standard and not alter it. —Michael Z. 2009-01-20 22:38 z

There's a big difference between claiming a name for one's own, and using the name to credit the proper owners. The key is legal, and the transcriptions are legal, so I can't imagine crediting AHD as the developers of the system would be infringement.

For all EncycloPetey's drama, it would be easy enough to make a user-friendly Wikimedia variant. The breves and macrons are ubiquitous; using French â for the low back vowel and maintaining phonemic vowels before r (like ēr for AHD îr—something which AFAIK none of the dictionaries do consistently) would avoid any plagiarism issues. But oh well. kwami (talk) 23:40, 20 January 2009 (UTC)

[edit] Proposed US School transcription

(cont. from previous discussion)

Here's a possible school-dictionary transcription. There's very little that's specifically AHD (only the ûr, actually) or any other dictionary, so there's no concern over copyright. Plus it's phonemic, and also browser friendly, only using the basic ISO range of Unicode plus schwa.

User:Kwamikagami/American dictionary transcription

kwami (talk) 12:49, 4 February 2009 (UTC)

Systems like the one above look "sort of ok" if seen in a list as above. Applied to real words, they become awkward quickly. To judge we should at least respell all the example words completely. I did a few to start. −Woodstone (talk) 14:46, 4 February 2009 (UTC)

Some more examples of how this works out. My guess would be that most people would get many of them wrong by using just intuition. −Woodstone (talk) 15:06, 4 February 2009 (UTC)

word	pronunciation
made	mād
believe	bēlēv
wine	wīn
hide	hīd
ball	bôl
low	lō
lure	loor
though	dhō
badge	băj
worse	wûrs
bear	băr bār

Wouldn't that be bear – bĕr?

Actually, it's \bār\, to people who make the distinction. (I've seen the slashes reversed in dictionaries. Don't know if that's an error, or a signal that this isn't the IPA.) kwami (talk)

How about a middle dot ( · U+00B7) for syllabification? It's more compact, and used in many dictionaries. There's also a dedicated hyphenation point ( ‧ U+2027), but I don't think we need to require it. —Michael Z. 2009-02-04 17:37 z

Yeah, I like the middle dot. (Best to stick to ISO characters, no U+2000 range.)

Woodstone, the point is that a couple hundred million people have been raised on systems like this. They're used in Webster's, Random House, & American Heritage dictionaries, among many others. I'm not expecting people unfamiliar with it to recognize it; it's intended for people who already know it and don't know the IPA—like the people who are constantly complaining that the IPA is gibberish to them, and why don't we use "normal" transcriptions like they see in dictionaries. The only thing that might throw people is the ûr, if they're not familiar with the AHD, and the dh, if they're not familiar with the Compact Oxford and a couple others. The point of the dh is to avoid soft formatting distinctions (like italics) that would get lost when copied & pasted, or non-ISO characters (like strike-through slashes) that won't display properly without the proper fonts. kwami (talk) 22:23, 4 February 2009 (UTC)

(I seem to remember ä, ö, ü from my grade-school dictionaries, but not â, ô, û.) I agree that it's best if the system relies on the character values, and not on any formatting. The nasals -ăⁿ, -eⁿ, -oⁿ, which use a superscript n (U+207F) are preferable to HTML <sup> elements or italics, which may be lost in a cut-and-paste or save as text. I also don't mind AHD's distinction using small capitals, which can remain within the ISO range. By the way, does a double under-tie work on a variety of platforms: c͜h, h͜w, n͜g, o͜o, o͜y, o͜w, s͜h, t͜h, d͜h, z͜h?

Frankly, no one learned “this system” – they only used it with a reference. We are including this to because it is comfortable, not because it's any easier than phonological IPA. We've documented at least 14 such systems (it occurs to me that they exist because they are proprietary—I presume each publisher developed their own to avoid the accusation of plagiarism, and that possibility a good reason for us not to rely on AHD's system). —Michael Z. 2009-02-04 23:02 z

The underties etc. are supported by all platforms, but not by all fonts, and many people don't have the proper fonts installed. I suggested on Wiktionary that they stop using SAMPA, and was told that they need to keep it for just this reason. Evidently the computer terminals in the UC Berkeley library do not support IPA! Pity someone trying to use a public computer at a less progressive institution. And people have complained that the Wiktionary oo-bar and oo-breve do not display properly.

I never learned 'ä' etc. in school, but my main concern was confusion with the German umlaut. People might expect 'ä' to be 'ĕ'. The 'â', on the other hand, is used with that value in French and several other languages, and parallels 'ô', which is used in several US dictionaries. 'ûr' is only used in AHD, but I couldn't think of anything better. ('ər' with a stress mark hits problems when people neglect to mark stress. And it's not always reduced, but has the same range as 'ŭ' vs. 'ə'.) kwami (talk)

By the way, is there a reason to stick to ISO and not use the Unicode range? There's some technical confusion between ISO and the rip-off Windows encoding anyway, and I don't believe the following characters are represented in ISO-8859-1 at all: ă, ā, ĕ, ē, ə, ĭ, ī, ŏ, ō, ŭ, ū. —Michael Z. 2009-02-04 23:18 z

I believe they are supported by enough fonts that nearly every computer will come with a supporting font installed. The same is not true of the IPA, higher Unicode ranges, or the Wiktionary rendition of AHD. kwami (talk) 23:24, 4 February 2009 (UTC)

All those letters display properly in my copy of IE. I tried a schwa-dot for schwi, but only got a box. Right now I have a small cap bar-I, <ᵻ>. Does that display properly for everyone? kwami (talk) 10:37, 15 February 2009 (UTC)

Both schwa + combining dot above (ə̇) and small cap I with stroke (ᵻ) work in the default configuration of Safari/Mac and Firefox/Mac. In Safari, the small slashed I matches the sans-serif capital, so it doesn't have the serifs. —Michael Z. 2009-02-15 19:58 z

Another editor said he had problems with it, so I subbed a cross-bar diacritic which he said he could read. I put it up at Wikipedia:United States dictionary transcription, with a template {{USdict}}. There are only a couple dozen articles linked to it, so it will be easy to update if we decide to change anything. kwami (talk) 01:43, 16 February 2009 (UTC)

This here: < ɪ̵ >? It fails to display on the Mac; details at Wikipedia_talk:United_States_dictionary_transcription. —Michael Z. 2009-02-16 16:46 z

Shouldn't we be deleting Wikipedia:Pronunciation respelling key now? The transcription domain is becoming overcrowded. The newly proposed scheme has a pleasantly consistent feel, but I am totally thrown off by <bare>=|bār|. That cannot be right. It should be |bĕr| or |băr|. And it's a pity that |tūn| makes it unusable for UK-en, where it would be |tyūn|?. −Woodstone (talk) 08:53, 16 February 2009 (UTC)

A lot of people use respellings, though most of them probably aren't very consistent. It probably wouldn't be too much trouble to switch over the ones that are linked.

/mĕr′·ē/ is 'merry', and /măr′·ē/ is 'marry', so how would we transcribe 'Mary' if not /mār′·ē/? Unless we want to have separate sets of vowels before /r/ than before other consonants.

/tūn/ works just fine for UK English. Dew is /dū/ and do is /dōō/; if you don't make that distinction, then just ignore it. kwami (talk) 09:20, 16 February 2009 (UTC)

Mary is different because it should be analysed as |mā′·rē|. Pronouncing a true |ā| before r in a closed syllable is near impossible. Can you think of any closed rhotic minimal pair with |ā| and |ǎ|? And should we advise on a bracket style? We should not overload /.../. −Woodstone (talk) 10:38, 16 February 2009 (UTC)

\ăr\ (Webster's uses back slashes) does not occur in closed syllables. Nor does \ĕr\, outside Scottish English. So English phonotactics prevents minimal pairs. Are you saying there is a four-way contrast, merry, marry, Mary, and mare?

You're arguing phonetic detail. Pronouncing a true \l\ at the end of a syllable is nearly impossible to many English speakers too, but that doesn't mean that dark el should be written with a distinct symbol. To me, at least, the \ā\ in \ār\ is as close to \ā\ in isolation as many other \āC\ sequences are. For instance, the a of bare is closer to the a of bane than it is to the a of bale, but I wouldn't argue for a special symbol for long a before el. Similarly, I'm happy to accept \bōl\ for bole, and \wôl\ for wall, though for me bole should have the \ô\, not wall. Likewise, the vowels of bide and bite are entirely different to my ears, but I'm not going to be confused by writing them both \ī\.

The OED and Random House both propose the same vowel for bare and mare that they do for Mary.

If we do choose a different transcription, I'd go with \ăr\, because AFAIK that won't cause any confusion, whereas \ĕr\ is distinct in Scottish, as well as in unassimilated French loans.

I'm not sure about using different slashes. People are confused enough with different stress-mark placement, without us getting dyslexic on the slashes. (The pipes are used for morphophonemic transcriptions, which these are not.) All forward slashes mean is that the transcription is phonemic; the symbols used for the phonemes are really irrelevant. You can use suits of cards, and the slashes would still be appropriate. kwami (talk) 11:03, 16 February 2009 (UTC)

This still shares many symbols with IPA, and many readers may have trouble differentiating the two in isolation. There should be some convention which unambiguously differentiates it from IPA. I'm okay with backslashes since there's a precedent, but we could also consider something with a clearer contrast. —Michael Z. 2009-02-16 17:00 z

Currently it's underlined (from the link, suppressed for the IPA) and has no set-off marks. That makes it look pretty distinct to me. kwami (talk)

Using "bare"=\bǎr\ (no need for "parry" anymore) would make the system clearer for me. That actually does not preclude "mary"=\mārē\. The vowels with diacritics are distinguished from normal spelling, and as such unambiguous. It would be more consistent to spell \ôy\ and \ŏŏr\. Especially \ow\ is problematic as it could easily be taken as the vowel in "low". How about using \âw\ instead? With these additions, no vowels can be mistaken for their normal equivalents, which can be most confusing in existing words. −Woodstone (talk) 21:48, 16 February 2009 (UTC)

The overuse of diacritics can also cause confusion. I don't think "oy" would ever be misread. I agree that ow is problematic, however, though several dictionaries use aw for our ô, so that's a potential problem with our using âw. As for oor, one could argue it is closer to ōōr because it's a long vowel. But since the distinction is neutralized before r, the point is moot.

What do people think of <ăr> (not *<ǎr>) and <âw> or <ăw>? kwami (talk) 01:36, 17 February 2009 (UTC)

On the screen, háček and breve can be indistinguishable from each other, so I would vote against adding <ǎr> to this system.

I haven't proposed using a hacek. kwami (talk)

I can't say much about the other suggestions, because for me merry and Mary are identical, and marry doesn't have to differ at all except to clarify the distinction from the other two – using the diphthong symbol \ā\ for any of these seems quite wrong. The vowel in bear and bare is the same, a short e and not any kind of a. —Michael Z. 2009-02-17 02:31 z

But we shouldn't be transcribing your dialect specifically. If we extended the same courtesy to others, the transcription would quickly become useless: We'd have to drop h, th, dh, ng, wh, final l, final r, half the vowels, and who knows what else. All of us are going to have pronunciations that differ from a common standard. Londoners might object to having an r in bare at all. We have three combinations that you pronounce [er], and need three transcriptions for them. All the reader needs to know is a key word that uses that symbol, and to pronounce the word in question that way. kwami (talk) 08:35, 17 February 2009 (UTC)

I changed Haumea, Makemake, and Eris from respellings to this transcription, and relegated it to a footnote. See how it looks.

A serious objection to the "âw" proposal: the whole point of this is to use letters and digraphs familiar from English orthography. /aʊ/ is not spelled "aw" in any common English word. We're basically left with "ow" and "ou", with or without diacritics. Likewise, V plus silent e typically is written with a macron, and this works as well for bare → ā as for made → ā. kwami (talk) 12:02, 17 February 2009 (UTC)

It does not work well for "bare", since that has a different sound (OED gives /bɛː/). Would you also like to spell "are"=\ār\, or would that be \âr\? Agree that \oy\ would not easily be mispronounced. For the sound in "how", using \ou\ without diacritics is dangerous because of "thou", "through", "though", "rough" etc. Similarly \ow\ is confusing because of "how and "low". At least adding a diacritic would avoid taking it for a "normal" spelling. So we are left with \ŏu\, \ŏw\ or \ŏŭ\. Perhaps the latter is best, because it uses only modified vowel symbols. That makes it "how"=\hŏŭ\, "low"=\lōw\ −13:53, 17 February 2009 (UTC)

Yes, are would be \âr\.

Like every other vowel in English, the pronunciation of \ā\ is affected by the following consonant. The effect of \r\ is especially great, but every consonant affects vowels to some extent. For example, the effect of a voiceless consonant on \ī\ is so great that it's marginally phonemic. For another, \sp, st, sk\ are pronounced exactly the same: they're all [s] followed by silence. They're distinguished purely by their effects on the following vowel. In my dialect, writer and rider are distinguished solely by vowel quality. But I have no trouble understanding a transcription of \rītər\ and \rīdər\. \l\ affects \ō\ to the extent that I would confuse it with \ô\. Etc. There are different effects in different dialects, so we can't go with a narrow phonetic transcription. Bare has the same vowel as Mary, according to Chambers as well as to the OED. In fact, both Chambers and the 5th ed. of the Concise Oxford use <ā> for both made and bare: <mād, bār>, despite the fact that the difference is especially strong in non-rhotic dialects.

<ŏŭ> is starting to overload the system with diacritics again. A third grader will understand <ow> (I mean literally: the system is taught in the first grade). Yes, low has that digraph with a different pronunciation, but the same argument could be made for every single unaccented letter in the chart, like Shaw's ghoti for "fish". And then there's allow with the same low sequence using our pronunciation. Dictionaries haven't had problems with this, or they would have come up with something else. If there's a particularly confusing word, such as bow \bow\ vs. bow \bō\, we can always be explicit with it: "bow \bow\, rhymes with 'cow'." kwami (talk) 22:58, 17 February 2009 (UTC)

I do not understand the resistance against "bare"=\băr\ or \bĕr\, which can be analysed as ă+r or ĕ+r coming much closer to the correct pronunciation than \bār\, especially when analysed as ā+r.

In the current proposal, \oor\, \oy\, \ow\ contain the only unaccented vowels. So you can hardly call it overloading when these acquire accents as well. For \ŏŏr\ that is just a matter of consistency with \ŏŏ\ (I cannot see a reason to leave off those accents). If you write "low"=\lō\ (not \lōw\), then having "how"=\how\ lacks logic. The most occurring spelling for the /aʊ/ sound is "ou", so a marked variant of that still seems best to me. Since \ŏ\ is already used for /ɒ/ (a sound somewhat similar to /a/), there is some logic in \ŏu\ or \ŏŭ\.

By the way, why call it explicitly US? It may be most current there, but it could actually be made an almost one-to-one translation of IPA. Actually it would be best if there were rigid rules to convert {{IPA-en}} to this one. −Woodstone (talk) 23:58, 17 February 2009 (UTC)

Do you have an example of where <ow> or <ou> would be confusing? I'd rather stick to the conventions people are used to unless there's a reason to change. People are used to <ow> and <ou>; <ŏŭ> is a novel combination that is likely to throw them. One more odd symbol that needs memorization.

The spelling \ow\ would be confusing in words like "out"=\owt\ which have a pattern that never occurs in English orthography. The reverse problem for "how"=\hou\ is less serious. −Woodstone (talk) 22:57, 26 February 2009 (UTC)

\ăr\ and \ĕr\ are short vowels. Bare has a long vowel. As I see it, the only appropriate long vowel is \ā\. That's also the decision reached by Chamber's & Oxford. Do you have a particular reason for abandoning it, other than the fact that the phonetic correlation isn't the best in your dialect? Because whatever you choose won't be a particularly good match in someone else's dialect.

We could leave the diacritics on boor, but they would be \ōōr\ (as in Chamber's & Oxford), since that is also a long vowel. Long vowels are lowered before \r\, but that is phonetic detail, like the difference between light & dark \l\. kwami (talk) 08:26, 18 February 2009 (UTC)

My perception is coloured by the fact that in my language distinction of /er/ and /ɛr/ is phonemic (not allophones). So any representation that annihilates the distinction sounds very wrong. I have the same problem (in other discussions) with /a/ and /ɑ/, an opposition found in only few languages. −Woodstone (talk) 11:21, 18 February 2009 (UTC)

For me the distinction isn't phonemic, but for those to whom it is, bare is something like a long [ɛːr] or [ɛər], and therefore as much like [eː] as it is like [ɛ]. /r/ lowers all vowels, so /oːr/ comes out [ɔər], /iːr/ comes out [ɪər], etc. You get the same schwa between long vowels and /l/, tho w/o the lowering: /iːl/ comes out as [iəl], /eːl/ as [eəl], etc. So /eːr/ as [ɛər] is completely regular. This is an orthography-based transcription that's familiar to kids, and usually isn't something that makes much sense to 2nd-language speakers. kwami (talk) 13:42, 18 February 2009 (UTC)

[edit] Compare the (older) EOD system

The Oxford Concise dictionary used to have the following system:

• ā ē ī ō ū (mate, mete, mite, mote, mute, moot)
• ǎ ě ǐ ǒ ǔ (rack, reck, rick, rock, ruck, rook)
• āṝ, ēṝ, īṝ, ōṝ, ūṝ (mare, mere, mire, more, mure)
• âr, êr, ôr (part, pert, port)
• ah, aw, oi, oor, ow, owr (bah, bawl, boil, boor, brow, bower)

−Woodstone (talk) 23:06, 26 February 2009 (UTC)

You left out moot, rook. kwami (talk) 03:38, 27 February 2009 (UTC)

[edit] Audio recordings

I was wondering if there was a project to add small audio samples to articles that would benefit from them. I mean an article discussing subtleties of pronunciation (like the one I read now, Hiberno-English) would be much easier to follow if I could listen to the examples. This would open these articles up to a much larger audience. --CyHawk (talk) 18:57, 5 May 2009 (UTC)

People are working on audio for many articles, but I don't know if there's an overall project. It would require having speakers of the dialect in question. kwami (talk) 07:35, 19 May 2009 (UTC)

[edit] =IPA criticism

My guess is that no more than one in ten thousand college-degreed Americans understands even what the International Phonetic Alphabet IS -- and that many fewer of those understands the IPA esoteric code [Yes, it IS a code, meant to be all-but-understandable by other than the IPA elite (obviously, a very small group. While IPA mavens may insist that ALL US college graduates PERMANENTLY memorize a very complicated code that is useful maybe once in every year, or two, or ten. It is quite stupid for any individual to do so --

  FOR 99.9% OF ENGLISH SPEAKERS, IT WOULD BE MORE ADVANTAGEOUS  

I think it would be EXTREMELY helpful if there was a link following the IPA pronunciation "language" that provided an audio version of the pronunciation. It would be great if we ALL knew perfectly KNEW, memorized, had no doubts AT ALL about the IPA pronumciation language. Some questions from an intelligent person,

Even most quite intelligent people have trouble understanding the IPA "language". And even if they DO, they will have the VERY

  LIMITED understanding of ONE --ONE, ONLY ONE pronunciation of the word in question 

It is probably a VERY nice thing for YOU (those who make policy and position decisions for Wikipedia) to have ALL definitions,

It could have these characteristics:

 * It could be triggered by a visual link like [audio] or [hear] (crude, I know; any reasonably well-known link technologies might be appropriate. * It could be encoded in mp3 (for maximum effectiveness over the installed worldwide base of computer users) or oog (if necessary for legal reasons). * For storage and bandwidth efficiency, it could be provided at speech-compatible bit rates    (20-60 kbps?). * The audio files would be completely different for each language, of course (Wikipedia    already provides this differentiation). * For widely-spoken, highly-diversified languages (there are many), the audio file would    contain multiple pronunciations in the various regional/social accents/dialects. * The order of the various pronunciations would be set by the population numbers using that    variation.    * For example, for English language versions of Wikipedia:      * "Middle American" English (known sometimes as "American Standard Stage English" would         be the first pronunciation;      * "Standard Stage British" would be the second in the order of audio snippets.      * Other regional and social dialects would also be presented in their order of popularity      * Other regional dialects of English in the US, the UK, Oceana, South Africa, English as          spoken in Oceana, South Africa, India, and other areas would be presented according to         the popularity of their usage. * The complexity of sampling, classifying, storing, and present 

BTW, the common pronunciation for "fuh-kweu" is —Preceding unsigned comment added by 76.212.205.151 (talk) 23:32, 11 December 2009 (UTC)

That sounds really difficult to pull off and enforce. You might want to check out this discussion about plans to utilize the mouseover feature. — Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 01:50, 12 December 2009 (UTC)

[edit] Not indicating language

I'm having a dispute at Joseph Beuys, which has the German pronunciation of his name, with an editor who argues that there is no such thing, as personal names are pronounced the same in all languages, and that we cannot name the language. So he changes the {{IPA-de}} template to {{IPA-all}} (arguing that the "all" means "all languages equally", not "all other languages".) He feels very strongly about this, whereas I feel just as strongly that we shouldn't imply that a pronunciation is English when it is not. Does anyone here agree with him? kwami (talk) 07:35, 19 May 2009 (UTC)

[edit] Do we need to tell people how to say George W. Bush?

I noticed the IPA template was in use on Bush's page. Seemed unnecessary to me, so I questioned its need here. Feel free to weigh in. Are there any guidelines anywhere on when not to include a pronunciation? Flowerparty^☀ 23:05, 29 June 2009 (UTC)

There are no real guidelines. Generally if you can expect the reader to find the relevant pronunciation in a readily accessible dictionary (i.e., don't expect them to have access to the OED or a medical dictionary), then we shouldn't include it here, per WP:dictionary. Personal names, however, can be a problem. Okay, in Bush's case everyone knows that his name is pronounced like the word bush, but in general that's not something that the reader can take for granted. What I've typically seen is that pronunciations for personal and place names that are written as common words are only provided when they aren't as written. However, sometimes there's a better-known name that isn't pronounced as expected, so when one comes along that is, that is itself unusual and needs to be pointed out. kwami (talk) 00:50, 30 June 2009 (UTC)

You lost me with that last sentence. Are you saying we only need to give a pronunciation for something like Cholmondeley? I'd agree with you that it's appropriate there. Flowerparty^☀ 01:19, 30 June 2009 (UTC)

I meant that sometimes a name is pronounced just like what it looks like, but that we still need to transcribe it. For example, if someone named Menzies pronounced their name "men-zeez", we would need to say that, because normally it's pronounced "ming-iss". In the case of Bush, I could imagine a non-native speaker trying to learn the presidents of the US, and being confused by which are pronounced as they're spelled, and which aren't. I don't feel strongly either way in this case. kwami (talk) 01:45, 30 June 2009 (UTC)

Gotcha, yeah it makes sense there. Is it likely that a non-native speaker would be trawling the English wikipedia to find out how to pronounce the name of each US president? Is this information not better suited to foreign language wikipedias? It seems to me that we're catering to minute contingencies by including a pronunciation everywhere on the off-chance that someone might find it useful. Another one that seems like overkill to me is with US states - every one has its name transliterated in IPA in the opening sentence. To me this is only necessary for Connecticut, Arkansas, maybe Michigan, possibly Illinois and a couple of others. Flowerparty^☀ 01:58, 30 June 2009 (UTC)

The non-native english speakers in my office frequent websites that have pronunciation guides for various names. I caught one of them using the Bush website for just that purpose some time back because he was having trouble with part of Bush's name. A good number of people can read english but have trouble speaking it and further there are various accents of the English language, so given that its not hurting anything I'm inclined to keep it. RTRimmel (talk) 03:08, 30 June 2009 (UTC)

With George Bush specifically, I think it's silly to have a pronunciation indicator. Neither his first or last name are words that have alternate pronunciations. He's also extremely well-known so that the overwhelming majority of readers will have heard his name pronounced.

When there's a name with an intuitive looking pronunciation that has a possibility of mispronunciation (such as Yankovic), I don't see a problem with a pronunciation indicator. There's probably a fuzzy line somewhere, but George Bush is definitely not near that line. — Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 03:40, 30 June 2009 (UTC)

Yeah. It's clearly unnecessary in this case. That said, as people have suggested above, it often is useful for personal names, and if someone's just been a bit over-zealous in this case, let's not sweat it. It causes very little harm. garik (talk) 09:05, 30 June 2009 (UTC)

Hm. Hush, lush, mush, push - native speakers of English can't help underestimate the difficulties others have with pronunciation. Enki H. (talk) 13:32, 30 June 2009 (UTC)

A fair point. I say err on the side of inclusion. garik (talk) 14:17, 30 June 2009 (UTC)

I didn't mean to spark off the same discussion in two places (my fault for picking such an irresistible section header, I guess). I think this discussion shows that we maybe should have a couple of sentences on this page indicating when a pronunciation is and isn't appropriate. Not sure we'd agree on a guideline at the minute though. Seems we have two approaches:

1. If unclear, assume that it's appropriate to give a pronunciation
2. If unclear, assume that it isn't

My feeling is that approach number 1, though well-intentioned, is misguided for our purposes. As Aeusoes has pointed out over there, it is outside the remit of an encyclopedia to act as English-language tutor to the unsteady learner. Essentially, wikipedia is not a dictionary. If we err on the side of inclusion we're bringing a lot of clutter into the opening sentences of articles that - while maybe not to a significant degree on its own, but these things are cumulative - does impede the reader's ability to read the sentence. This obviously isn't true of visitors to this page, but ( /nuːˈdʒɝːzi/ (help·info)) is meaningless to many. And anyway I don't think we should be asking ourselves "what's the harm?" - that kind of thinking leads to the inclusion of all kinds of whistles and gadgets and unnecessary trivia and decoration. Rather we should ask "what's the use?" If the answer is "it might help someone learning English", I may concede that that's a possibility, but we are not Rosetta Stone. I think we have to assume competence, and use the IPA sparingly. We should leave the pronunciation info for obvious words to wiktionary and to foreign language wikipedias. Flowerparty^☀ 16:43, 30 June 2009 (UTC)

Here's another point that's worth considering. If we include a pronunciation only for exotic spellings and for instances where the pronunciation cannot be intuited from the spelling (as with Yankovic), the presence of the IPA template nicely alerts the reader that the pronunciation is not standard. If we're going to include a pronunciation routinely for all names however (as seems to be the preference of some of the commenters above), then our readers are likely to start filtering out this information (in the same way I filter out the chinese script in articles on Chinese topics) and missing the non-standard pronunciations all together. Flowerparty^☀ 19:03, 30 June 2009 (UTC)

The problem with that logic is that as a native speaker of American English, I largly ignore the IPA anyway. I know how to pronounce virtually any word in the english language at a casual glance. However, I am aware of a signifigant number of non-native english speakers who, when reading articles on various subjects, cannot pronounce the words correctly and utilize the IPA to learn the correct pronounciation. Removing it from the article forces them to bop around the net for a correct pronounciation until they find it. I know you are making the argument that it adds bells and whistles and is meaningless to many, however the people who find it useful know what it is and adding a link to a short ogg file is as difficult as adding a picture to a page for a modern web browser. So in short, until MOS has a valid guideline as to what constitutes a requirement for IPA useage, I think we need to err on the side of including them because a solid percentage of our readers actually use them. Perhaps declaring that the prounciation is non-stardard in the IPA description would be easiest. RTRimmel (talk) 13:09, 1 July 2009 (UTC)

Right, that's why I'm suggesting we draw up some guidelines. You still haven't addressed why it's appropriate in an English-language encyclopdedia to routinely give a pronunciation key for English words. We could include a version of the article's title in semaphore or morse code and I bet more people would learn how to wave flags, but we don't because that's going beyond what's required of an encyclopedia. Flowerparty^☀ 18:51, 1 July 2009 (UTC)

Hammer has a picture of a hammer. I'm going to argue that everyone knows what a hammer looks like. Wheat has a picture of corn. I'm going to argue that everyone knows what corn looks like. Cat has a picture of a cat. And so on. Are we are going to argue why it is appropriate in an English-language encyclopdedia to routinely give a picture for common English words. No, because its what a modern encyclopedia is for and it is both what is required and expected of it. Adding in a prounciation guide is as useful to our readers as adding in those pictures. Unless there is a wikipedia based issue in terms of architecture or bandwidth, I argue to keep them. RTRimmel (talk) 19:29, 3 July 2009 (UTC)

Glossing over temporarily the apples-oranges problem, most of those pictures are combined with details that complicate more basic understandings of the subject or augment visually what is being described in nearby prose. So the picture of the hammer says it's a "modern claw hammer" something not obvious to those with only basic knowledge about hammers. The cat pictures illustrate social and biological aspects of the cats and the wheat pictures (other than the three that just say "wheat") show specific non-obvious details about wheat. We also have guidelines for images (WP:IMAGES and WP:CAPTIONS) that elaborate on what's appropriate.

However, images and pronunciation guides are two completely different things, certainly enough that we can come up with different policies on them.

It seems that you're arguing that "obviousness" should be the criterion that we use to measure the appropriateness of including pronunciation. However, another important factor is clutter within article prose. As mentioned above, putting pronunciation in article prose even when it's not an unusual spelling will prompt readers (especially those who have difficulty with the IPA) to ignore the IPA even when the word has a non-obvious pronunciation. — Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 20:15, 3 July 2009 (UTC)

You seem to be reiterating the first point of the FAQ at the top of the discussion page. I'd comment further, but it does a better job of summing up why we don't do this than I would. Aside from that, I'm simply arguing the opposite than obviousness. I'm arguing that if its worthwhile to include a picture for everything, its worthwhile to include a pronunciation guide for everything as well. To you it adds clutter, to me many of the pages are already cluttered beyond a reasonable level and so it doesn't significantly impact my reading of the page. RTRimmel (talk) 01:39, 4 July 2009 (UTC)

The FAQ item is in reference to criticism that we shouldn't use IPA at all because many people aren't familiar with it. I'm not arguing that. While there's a sound rejection of that argument, it doesn't mean we have to be dismissive of people who have difficulty with IPA. This is part of the reason editors are working on making language-specific IPA guides and making them accessible to readers unfamiliar with the IPA or with the particular languages.

In regards to clutter, it seems as though you're throwing your hands up in the air and saying "well, it's already a mess so we shouldn't try making things better." I think moving the pronunciations to the infobox (something I'm open to even with obvious pronunciations) will be beneficial to the casual reader while still being helpful to those you've argued would benefit from the pronunciation. If we overdo it in the prose, then it can cause many readers to gloss over pronunciations (which several people here have admitted they do) and nullify the point of having them in the first place. — Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 03:08, 4 July 2009 (UTC)

We could consider moving the IPA to the Infobox instead, to address the "desensitization" and cluttering concerns. Eg. right beneath the birthname. Enki H. (talk) 14:16, 1 July 2009 (UTC)

The language infobox allows for that option. That sounds reasonable. — Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 15:35, 1 July 2009 (UTC)

We could, but that's just shifting the clutter elsewhere. Most infoboxes are too busy anyway. Flowerparty^☀ 18:51, 1 July 2009 (UTC)

In effect, the infoboxes act as raw data that users can access if they choose to. I tend to ignore the infoboxes unless there's specific information I want to find (where is this language spoken? when was this individual born? when was this city founded? etc). Since the argument for the pronunciation is that some users find it helpful (much like other information in the infobox), and infoboxes aren't supposed to be uncluttered anyway, moving the pronunciation to the side to ease the clutter for the casual reader seems like a good idea. — Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 23:39, 1 July 2009 (UTC)

Although I'm guilty of some of the mess, I agree that multiple IPA pronunciations, or even names and transcriptions in multiple languages without the IPA, can be a distraction at the beginning of an article. The moons, planets, stars, and constellations now have an IPA slot in their info boxes, which I think is appropriate, but there's always footnotes.

One thing to consider with infoboxes: I've been cleaning up a lot of the entries with AWB (like people who transcribe sound-alikes and call it the IPA). This is possible in part because I can search within templates that contain "IPA". However, if the IPA is moved within an infobox, I don't know how to search them without sweeping up everything in the infobox, which means maintenance could become a problem. That's perhaps an argument for redesigning AWB, but that could take a while, if it ever happens. kwami (talk) 19:09, 1 July 2009 (UTC)

Couldn't the way around that problem be making it so that the pronunciation element of the infobox doesn't automatically assume IPA encoding so that users have to use {{IPA}}? It also allows for respelling pronunciations, which MOS says we can put alongside IPA. — Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 23:39, 1 July 2009 (UTC)

Yes, that would be good, because even within the IPA it would force association with the proper language. However, AFAIK it wouldn't solve the AWB problem unless options are added to that program. kwami (talk) 00:06, 4 July 2009 (UTC)

Infoboxes are not dumping grounds for trivia which has been moved out of the article body. The use of the lede for IPA text is well-established. Chris Cunningham (not at work) - talk 15:22, 11 December 2009 (UTC)

But pronunciation isn't useless trivia. It's actually vey common to put pronunciation in infoboxes. — Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 23:04, 11 December 2009 (UTC)

Not in my experience. However, the discussion above is not about adding it to the infobox so much as moving it to the infobox. That's what's inappropriate. Chris Cunningham (not at work) - talk 00:41, 12 December 2009 (UTC)

Well it is in my experience. You seem to be assuming that people wish to move it to the infobox because it's not notable enough when moving to the infobox reduces clutter in the prose. — Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 01:44, 12 December 2009 (UTC)

It doesn't concern me why people want to move it, but that they want to move it at all. Infoboxes should not contain unique information; the article body should be comprehensive. I don't have a problem in theory with IPA information being duplicated in infoboxes, but that's not what the discussion above seems to be about. Chris Cunningham (not at work) - talk 10:50, 12 December 2009 (UTC)

Well, I'm sorry to inform you that infoboxes often do contain unique information. Would you like examples? — Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 19:51, 12 December 2009 (UTC)

Then it's those examples which should be corrected. Chris Cunningham (not at work) - talk 18:14, 15 December 2009 (UTC)

It's very common for languages to have details about their classification in the infobox but not the article prose. This includes ISO language codes and the specific language family information. For example, Amuzgo, Awngi language, Atsugewi, Atikamekw language, Assiniboine language, Apurinã language, Albanian, and Äynu language. That's just the A's, too.

Still not convinced? How about some listed good articles? Like Mongolian language, Ottawa, Toki Pona, Wagiman, and Sinhala script.

Still not convinced? How about some listed featured articles? There's Gwoyeu Romatzyh, Mayan languages, Nafaanra, Nahuatl, Rongorongo, Swedish language, Tamil, and Turkish language.

You should see that this is neither uncommon nor unacceptable. — Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 19:27, 15 December 2009 (UTC)