Help talk:IPA/Japanese

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


In IPA [j] section, "yakusha" is a correct example, "kyū" is not. There's no [j] in "kyū", there is only one consonant [kʲ] (or maybe even [c]). And there is nothing about palatized consonants in Japanese. — Preceding unsigned comment added by (talk) 06:29, 20 April 2012 (UTC)

Please do not delete[edit]

The intent is to create this page as a pronunciation guide for Japanese terms which have an IPA pronunciation. The intent is that {{IPA|[[Help:IPA for Japanese|/aɪ piː eɪ/]]}} will link to this page in the same manner that {{IPA|[[Help:IPA for French|/aɪ piː eɪ/]]}} links the IPA pronunciations of French words to the Wikipedia:IPA for French page.

There are some aspects of Japanese pronunciation that makes the English IPA key inappropriate for accurate transcription. Some examples are the significance of vowel length and the existence of the non-English consonant /ɸ/.

We only need one fact to show that the English IPA key inappropriate: it's not English. Note, however, that vowel length is significant in Australian English & that [ɸ] is just an allophone of /h/. JIMp talk·cont 11:03, 11 April 2009 (UTC)

Please edit[edit]

I had to make some difficult decisions involving when and how to represent phonemes and allophones. You may note my choice of /ɽ/ to represent "r," as per the convention found in the IPA Handbook.

In general, I feel it's the best policy to keep a symbol in rather than leave it out, so that it will be here for a reader to reference if an editor should choose to use a narrow transcription. On the other hand, should we make transcription more standardized by restricting the numbers of options editors have for transcribing each sound?

I'd like to suggest leaving the "Under Construction" bar up for a couple weeks to encourage everyone to discuss and edit the symbols as much as possible. I don't know about you, but when I see a pronunciation guide, I feel like I'd be messing up the system everyone uses if I changed it.

First, these keys are meant for the reader, not the editor. Comments about what format the articles should follow should be here on the talk page, not in the intro to the article.
Generally these templates are designed to help us standardize our transcriptions. IPA-en, IPA-fr, IPA-ru, IPA-ko, and IPA-de all do this. However, Japanese has the difficulty of two phonemes, r and u, which do not have IPA symbols. I added the second common transcription for r. There's also a problem with tone, since that is so regionally variable.
I think it would be a good idea to choose one symbol for u and r and modify the articles to match. [ɽ] isn't too bad for the r, I guess. For u, we could use a diacritic such as [u͍]. (It is rounded, or at least labialized, so [ɯ] really isn't appropriate.) kwami (talk) 23:40, 11 May 2009 (UTC)
Wikipedia:Manual of Style (Japan-related articles) prescribes Hepburn romanization for transcribing Japanese, so this page is not so much an aid for editors in making transcriptions as it is an aid for readers in pronouncing already transcribed Japanese. -- Meyer (talk) 06:03, 12 May 2009 (UTC)

Suggest move to "IPA for Japanese (Hepburn Romanization)"[edit]

I suggest the page be moved to the name above since it provides an IPA mapping not for the Japanese language itself but for one of several alternative romanizations of Japanese. I think this page will be of great value to WP since although we've standardized on Hepburn romanization, there are still ambiguities in how Japanese words are pronounced. -- Meyer (talk) 04:22, 2 March 2009 (UTC)

I don't understand this comment. The IPA doesn't depend on the system of romanization used. kwami (talk) 22:50, 11 May 2009 (UTC)
It is dependent in this case because the article arranges the IPA sequences and provides examples assuming Hepburn romanization. If we assumed, for example kunrei-shiki romanization, some of the consonants would have to be reordered and examples replaced in order to make sense.
I think there's two divergent purposes this article can serve, and we need to be clear what we're about:
  1. Supplement the article Japanese phonology with an IPA key to the pronunciation of natural Japanese phonemes. If this is what we're trying to do, I think the article needs to be restructured based on Japanese phonology (kana).
  2. Provide an IPA pronunciation guide for the (Hepburn) romanized Japanese found throughout the English Wikipedia. If this is our goal (which the article as it stands suggests is the case), then I think the name change and some additions to the text are in order to clarify what we're about.
Personally, I don't think there's much value in 1. because the phonology article already covers the information and because the typical Wikipedia reader will not encounter much Japanese text, will rarely need to make a cold reading without romanization being provided, and the problem of pronouncing kanji.
But even if there is interest in pursuing both purposes, I think they are best explained on two separate pages, so I stand by my name change proposal for the current page. -- Meyer (talk) 05:37, 12 May 2009 (UTC)
But we're not arranging the key per Hepburn, we're arranging it per the IPA. We currently illustrate it with Hepburn, but we could always illustrate it with other systems as well. And anyway there's no necessity that we romanize everything we transcribe in IPA. kwami (talk) 06:37, 12 May 2009 (UTC)
IPA-order of the charts is not the most useful arrangement if we assume this page will be used mostly by readers trying to clarify the pronunciation of a romanized Japanese word rather than trying to puzzle out an IPA transcription. Hepburn romanization is our only practical choice for examples here since it's prescribed in the WP MOS for Japanese articles. It is unfortunate that is subject to mispronounciation by readers not already familiar with Japanese, although that is a problem common to all romanization systems.
I'm going to defer to your greater knowledge of phonology and give up the idea of renaming the page, but we definitely need to clarify that the examples are based on Hepburn so that readers can make sense of them, which I will do momentarily. -- Meyer (talk) 08:20, 12 May 2009 (UTC)
Ah, I see where you're coming from now. No, this wasn't intended to be a guide to Hepburn, it was intended as a guide to the IPA as used to transcribe Japanese words. It's part of a series: IPA for French, IPA for German, etc. kwami (talk) 09:03, 12 May 2009 (UTC)

"ː" and "seppuku"[edit]

Is there an authoritative source for this usage? To me, the "double sound" in "seppuku" versus "iie" and "doumo" are completely different types of sounds and it's misleading to use "ː" to represent both. The romanization of both sounds with double letters is just an accident of the Hepburn romanization system.

That being said, we still need one or more IPA sequences for "little-TSU + consonant" combinations. -- Meyer (talk) 04:26, 12 May 2009 (UTC)

[ː] is used for length. It's standard IPA for both geminate consonants and long vowels. kwami (talk) 06:29, 12 May 2009 (UTC)
Actually, I take that back. [ː] is appropriate for "double" consonants, which are long. However, it is not appropriate for "long" vowels, which aren't actually long, but doubled. Itō should (could) be [itoo], not [itoː]. Is that something we want to change? kwami (talk) 00:55, 22 May 2009 (UTC)
I'll remove "doumo" as an example for [:]. Beyond that, it's a question of how thorough we want to be in providing IPA representations of Japanese phonemes. Do we illustrate every phoneme, just enough to cover every morpheme in Hepburn romanization, or some other subset? -- Meyer (talk) 01:43, 22 May 2009 (UTC)
Based on keys for other languages, I'd say we want to indicate all phones that an English speaker is likely to hear as distinct. Phonemes aren't necessary, since they are theory dependent and don't tell the reader how to pronounce the word. I don't think we should have /Q/, for example. kwami (talk) 02:04, 22 May 2009 (UTC)

Another complication: I added aꜜ as a downstep for Japanese pitch accent. There are a couple other ways it could be transcribed: ˈa, as if it were stress, and á, as if it were regular tone. However, all of these present a complication when used with long vowels. For example, in dōmo, the downstep occurs after the first mora, and the downstep breaks up the vowel from the length sign: doꜜːmo. However, if the downstep occurred after the second mora, the stress mark would break up the sequence, and the tone mark would have to be placed over the length mark. I can't think of any such words offhand, though. Is this an acceptable transcription? kwami (talk) 11:05, 23 May 2009 (UTC)

Sorry, I'm getting to this kind of late. aꜜ shows up on my display as a and a box. What's it supposed to look like? — Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 02:38, 15 July 2009 (UTC)
This display shows a vertically-centered dot, though that may not be correct as I other IPA characters are not being displayed correctly. Have you set your browser to display Unicode characters? On MS Internet Explorer, follow menu path Display > Encoding > Unicode. -- Meyer (talk) 04:53, 15 July 2009 (UTC)
Yeah, my browser shows unicode characters. I think my font is just a little behind compared to yours. Is the dot above or below the a? — Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 06:15, 15 July 2009 (UTC)
On my display, the dot's vertical position is about 80% of the lower-case a's height. -- Meyer (talk) 06:22, 15 July 2009 (UTC)


I don't think Japanese actually has this phoneme, which represents a blending of the "n" and "g" sounds as in English "ring", etc. I believe Japanese words whose romanization includes "ng", are always pronounced as two phonemes, "ɴg" (necessarily preceeded and followed by vowels), as in "ringu" ("rin-gu", not "ring-u"). Again, show me, or better cite in the article, an authoritative source for using "ŋ" for Japanese and I will shut up. -- Meyer (talk) 05:43, 12 May 2009 (UTC)

Just about any source on Japanese will tell you this. We're not saying it's a phoneme, any more than [ɸ] and [ts] are phonemes in [ɸutsu]. kwami (talk) 06:31, 12 May 2009 (UTC)
I have seen a couple of documents on the web that also suggest this and the [:] usage above, though of unclear authority. Though both these run counter to my layman's sense of what the language sounds like and how best to represent it, I'll again defer to your knowledge of phonology. We should still cite where it comes from, though. Can you add references to some of those Japanese sources? -- Meyer (talk) 08:25, 12 May 2009 (UTC)
I think this key is based on the IPA Handbook, but I'll leave it to the author to add that if s/he likes. kwami (talk) 09:04, 12 May 2009 (UTC)

I've read that the pronunciation [ŋ] for 「ん」 is disputed (and I sure have trouble accepting it), but [ŋ] definitely exists in the Japanese language. The /g/ in 「が」、「ぎ」、「ぐ」、「げ」、「ご」 becomes [ŋ] when it is in non-initial position in many (most?) dialects, and also in the particle 「が」. Acidtoyman (talk) 23:43, 27 March 2011 (UTC)

Yes, [ŋ] exists; /ŋ/ doesn't. JIMp talk·cont 16:32, 7 May 2012 (UTC)

Palatal n[edit]

Kwamikagami removed the following entry for the reason, "palatal n not used in phonology article."

ɲ nihon knee

Ironically, I nearly did the same thing a few hours earlier, but decided against removing it because it was documented in Japanese phonology.

In between my and Kwamikagami's visits, at 19:42, 14 July 2009, Aeusoes1 made several edits to Japanese phonology, including removal of the table where palatal n usage was documented, for the reason, "some cleanup; rm tangential orthography information."

Did Aeusoes1 throw out a baby with the bath water? Palatal n in Japanese is still mentioned in article Palatal nasal. Should it be restored both here and in Japanese phonology? -- Meyer (talk) 05:46, 15 July 2009 (UTC)

You're starting the same conversation in multiple places. Let's keep it at Talk:Japanese phonology. — Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 09:36, 15 July 2009 (UTC)
"Kwamikagami removed the following entry for the reason, 'palatal n not used in phonology article.'"
"Ironically, I nearly did the same thing a few hours earlier, but decided against removing it 'because it was documented in Japanese phonology.'"
I spent a semester in Japan doing an independent study project on allophonic variations in Japanese, including the "doubling phoneme" and allophonic palatalization. I will participate in some of the discussions on this page when I find my notes so I can back up my statements with citations. (I was in error for not putting my references in the article when I created it.) However, I don't need to cite a source to convince people of the sensibility of this remark: We should not be using other Wikipedia articles as a basis for verification! Minetruly (talk) 20:53, 12 April 2010 (UTC)
In these "IPA for xx" keys, which are only intended to allow readers to get the gist of a transcribed pronunciation, we tend to use a quite broad transcription, unless an allophonic distinction is salient for an English speaker. So we generally have [ŋ] for languages in which it is not phonemic, but don't bother with things like [ɲ], because accessibility is a concern. In the lang phonology articles, on the other hand, we go into a lot more detail, and use a correspondingly narrower transcription. kwami (talk) 23:50, 12 April 2010 (UTC)

I think if we're going to include the alveo-palatal consonants (which are technically allophones in Japanese) we might as well include the (alveo-)palatal nasal. It would fit with the other allophones of /n/, which are [ŋ] and [ɴ] (which English doesn't contrast either). AlexanderKaras (talk) 01:35, 27 May 2010 (UTC)

It only occurs before /i/ and /j/, where it's scarcely audible to English speakers, so it's just another bit of technical jargon our speakers need to learn, without any benefit. Especially when our English example is "knee"! [ŋ] and [ɴ], on the other hand, are clearly audible. Also, if we have it here but not in the transcriptions, that would imply that it doesn't exist in those words when it does.
We could indicate that /m/, /k/, and /r/ have palatalized allophones too, but what's the point? It would do nothing to help the reader. — kwami (talk) 02:50, 27 May 2010 (UTC)

I don't really agree with that. Since we're making a guide for non-native speakers, we're not making a purely phonemic transcription, but a narrow one that more accurately captures the pronunciation of the language. Otherwise, why include vowel reduction in our guide for Russian? Besides, I don't think including [ɲ] is any more confusing to the average reader than [ɸ]. If it helps, I provided the example of "onion," which is probably a much better example than "knee." AlexanderKaras (talk) 01:51, 28 May 2010 (UTC)

I don't think Kwami is arguing for a phonemic transcription, rather he's saying that a broad phonetic transcription of Japanese geared towards native English speakers doesn't necessarily need to bother with allophony that English speakers aren't going to perceive anyway. This is why Russian has vowel reduction factored in.
With that said, wouldn't words like にゃく (nyaku), にゅうじょう (nyūjō), and にょう (nyō) contend with the generalization that the palatal nasal only appears before /i/ or /j/?

Those words begin with /nj/ (the maximal syllable structure in Japanese is CjC). But I'm in favour of including [ɲ] because of the articulatory differences from /n/. If nobody else is, then never mind. AlexanderKaras (talk) 05:10, 30 May 2010 (UTC)

English Example for a[edit]

It seems that this example is rather misleading. Aye is pronounced as the diphthong /aɪ/ and very few English speakers unfamiliar with phonology would be able to extract the initial /a/ of it (while those familiar with phonology would know it from the IPA anyway).

Now, consider a common word like run, which (in Australian English at least) is pronounced very similarly to [a]. I propose this be used as the English example either in place of or supplementary to the current one. D4g0thur 16:41, 28 October 2009 (UTC)

The approximate example could be modified. The vowel in call isn't the closest to [a]. Man, father are closer. --Mahmudmasri (talk) 20:24, 12 January 2011 (UTC)
Father is often used for this vowel in other languages. Man isn't as good. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 20:33, 12 January 2011 (UTC)

For the 'a' in father, is it like the a in fat, or the a in call? As in, is it an 'æ', or an 'ɔː'? For me, call and father have the same vowel, an ɔː, but I know the Japanese should be closer to an æ. magentafeelings (talk) 17:34, 28 March 2013 (UTC)

It's an approximation, though if you have merged those vowels you may pronounce the vowel in a way similar to the Japanese. — Ƶ§œš¹ [ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ] 18:07, 28 March 2013 (UTC)


Why was my edit reverted concerning the Japanese downstep? I only replaced the character because it doesn't display on everyone's browser. AlexanderKaras (talk) 06:06, 31 May 2010 (UTC)

Two reasons: this key should match the transcriptions; we should stick to the IPA. This may need to be an exception, but we should at least discuss it. — kwami (talk) 07:07, 31 May 2010 (UTC)
I only used the symbol found in the article on downstep. I think this may be an issue because the current symbol doesn't display on any of my broswers.AlexanderKaras (talk) 02:47, 2 June 2010 (UTC)

About Japanese /o/ and /e/[edit]

I don't think the Spanish /o/ is the same than the French /o/. French (not all dialects though) distinguishes mid-close /o/ and /e/ and mid-open /ɔ/ and /ɛ/. I think Japanese /o/ is the same than Spanish, Romanian and Greek /o/, same concerning Japanese /e/. They both are mid vowels, neither mid-close nor mid-open vowels. And Japanese /a/ is central as most of the Romance languages. Jaume87 (talk) 16:02, 14 July 2010 (UTC)

You are right. I'll fix the examples. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 16:45, 14 July 2010 (UTC)
But i don't think this is a proper example, you need to specificy it is in code /ou/ as in American English, without pronouncing /u/. Otherwise a British who reads this will think the Japanese /o/ is pronounced as /əu/.
Is /o/ from American English /ou/ a close-mid vowel closer to French /o/?
Jaume87 (talk) 17:27, 14 July 2010 (UTC)
I don't like code, either. If we want to say "Spanish o" then readers might go to IPA for Spanish and see code there as well. cold or core might be better for both. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 18:22, 14 July 2010 (UTC)
I think would be better between boy /ɔɪ/ and cold/code /oʊ/. And Japanese/Spanish/Romanian e between ray /eɪ/ and met /ɛ/.Jaume87 (talk) 19:00, 14 July 2010 (UTC)

Japanese 'g', different pronunciations[edit]

I'm not comfortable making the change to the official page because I'm not a linguist, but when I was taking Japanese in college, my professor taught us that the 'g' syllables (が・ぎ・ぐ・げ・ご)are sometimes (not always) pronounced with a nasal quality, close to ŋ (velar nasal). Particularly, I remember him showing us that pronunciation with が used as a conjunction (その映画をまだみません、見てほしいです。) Typically, at the beginning of the word, the sound was a hard 'g'.

Is this a minority dialect or standard Tokyo Japanese? I think it deserves mention. (talk) 21:53, 27 July 2010 (UTC)

Japanese phonology says it's a feature of "many dialects" though there's also some variation. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 22:07, 27 July 2010 (UTC)
I saw the examples of dialectical differences with word hage in Japanese phonology article. In standard Japanese most probably it would indeed be /hage/, however there is a very often used conjunction daga ("though") and it's always pronounced /daŋa/, if one would want to represent it's sound in Hepburn, the best way to approximate the sound would be dan'a, well of course it won't be written that way because it's composed of a da-kana and a ga-kana. And the /g/ indeed sounds like /ŋ/. The above user mentions the ga conjunction and daga is a derivative of the ga conjunction. Maybe it's speakers' way of differentiating between "ga" as a subject marker and "ga" as a conjunction ("though, but")? Samarkandas valdnieks (talk) 02:35, 9 October 2011 (UTC)

Japanese /r/[edit]

What is it? Is this sound really retroflex, or just postalveolar? (There is no symbol for a postalveolar flap in the IPA, so ‹ɽ› has to do). I'm inclined to think it's just a retracted [ɾ] because that fits better with the phonology (where there are no retroflex consonants after all) and because it parallels our use of ‹ɺ› for a (post)alveolar lateral flap. AlexanderKaras (talk) 01:43, 16 March 2011 (UTC)

Despite using a retroflex symbol, Okada (1991) says that it's postalveolar, not retroflex. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 03:09, 16 March 2011 (UTC)
It sure feels postalveolar. Certainly not subapical. — kwami (talk) 00:57, 8 May 2012 (UTC)

Dispute over english example for /t/[edit]

So, I can't find any authoritative source for the Japanese pronunciation of 食べる (taberu, the Japanese example). However, I will assume that /t/ is the correct IPA for the initial consonant. If this is the case, it's easy to find a reliable source that give English pronunciations in IPA. I would propose the Oxford English Dictionary. Now I will grant you that they give different pronunciations for atom in British and U.S. English. They do for table as well, but not for the initial consonant. I would propose tan as only a single pronunciation is given /tæn/. It is also used by the OED for their example of /t/. Now if /t/ isn't the correct IPA for the Japanese sound, we should change that. I don't claim to be an expert in Japanese phonology, but if the Japanese "t" needs an explanation because there is no good English approximation then we should add one like we did for /ɸ/. Just using stable really doesn't solve the problem. I know it's a project page, but still the examples should be taken from reliable sources rather than arguing over which editor's ear is better for hearing his or her own dialect of English. ---- Selket Talk 18:25, 20 April 2011 (UTC)

The problem with tan is the same as the problem with table. I think you're confusing phonetics and phonology here. English /t/ (in slashes) is a group of sounds (or allophones) dependent on context. In word initial position, it is an aspirated alveolar [tʰ], that is, there's a large puff of air that comes with the release of the closure between the tongue and the alveolar ridge (this, by the way, is not something that differs between American and British varieties of English). At the end of a word (and before other consonants), /t/ is either a glottalized and unreleased alveolar [t] or it is a glottal stop. For a number of speakers, including those in the United States, the /t/ in words like atomand butter, where it falls between a stressed vowel and an unstressed one, it is an alveolar flap, much like the r in Spanish cara. Finally, for our consideration, the /t/ that follows an /s/ in the same syllable, such as in still or stable is an unaspirated voiceless alveolar plosive. It's very similar to the aspirated form, but without the puff of air when the articulators separate.
The Japanese /t/ also has a range of pronunciations depending on context, the major ones we indicate in this pronunciation key. The allophone in question, [t̪] is an unaspirated voiceless dental plosive (we omit the diacritic, which is superfluous in normal transcription), which is most similar to the allophone of /t/ in stable. Both are voiceless, unaspirated, and pronounced with the crown of the tongue. The alveolar ridge is also very close to the teeth, meaning that the place of articulation is very similar between the two.
So, the long and the short of it is that stable, because it features an allophone of English /t/ that is closest to Japanese [t̪], is a better example than words like table or tan. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 22:56, 20 April 2011 (UTC)
The problem with stable as an example is the blend. I provided a source that states the t in table, tan, and stable are all pronounced the same, although it also classifies the "st" in stable as a single phoneme, much like "ts". If you have a source that the t in stable is not aspirated and the t in table is, then I'll differ to that. Otherwise, I've shown a source that says they're pronounced the same way. -- Selket Talk 23:38, 20 April 2011 (UTC)
The source you provided indicates that table and stable have the same phoneme, it doesn't say that they are the same phonetically. That /t/ is aspirated in certain contexts is something any linguistics 101 student can tell you. But you can check out, for example, Silverman, Daniel (2004), "On the phonetic and cognitive nature of alveolar stop allophony in American English", Cognitive Linguistics, 15 (1): 69–93 Æµ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 01:21, 23 April 2011 (UTC)


We have

IPA Japanese example English approximation
, ʑ jibun, gojū jeep, garagist

The second "g" in "garagist" is meant to approximate [ʑ] with English /ʒ/. That works for me, a US native, but /dʒ/ is common in many Englishes. (Wiktionary) I'm changing this example to pleasure, which AFAIK has only [ʒ] in all dialects. (Wiktionary, OED) --Thnidu (talk) 21:12, 13 October 2013 (UTC)

I think that was the point. The Japanese sound varies between zh and dzh just as English garage does. — kwami (talk) 07:01, 14 October 2013 (UTC)

Narrow transcription problems[edit]

Using very narrow transcriptions with obscure diacritics is not an appropriate solution for describing pronunciation of any language. The standards here are supposed to actually be used in articles that have no relation to linguistics. The transcriptions [i̥], [w͍] [u͍̥] and to some extent the tone drops and nasals are pretty obvious detail overkill. They make it more difficult to write out pronunciations and are bound to be misunderstood or confused. The u͍̥ is especially problematic since it's very likely going to be confused with a basic [u].

Anyone reader who wants to get down to specifics can read Japanese phonology. Details and fined points can easily be summarized in comments here, as is already done in the notes.

Peter Isotalo 17:13, 17 January 2015 (UTC)

Done, but I didn't touch the voiceless vowels. Also, I think something needs to be done about [ɽ~ɺ]. We can list the flap allophones separately, or just use one symbol in all situations. Peter238 (talk) 04:14, 26 December 2015 (UTC)
I very much agree with this. Frankly, I doubt if there are more than 17 people in the world who: (a) can read the obscure IPA symbols approximating Japanese, (b) need more information than is in the Hepburn romanisation, and (c) cannot read Japanese anyway. There was a little discussion about this here: Project Japan talk. In fact the list of proposed IPA pronunciations is full of little errors -- for example, the new page Sayonara wa Emotion created by the user who asked about these has a wrongly extended 'e' at the beginning of emōshon. I could correct it, but I cannot see that this in any reasonable way adds to the value of the page, but I hesitate just to remove it, without a bit of consensus. Should the Japan MOS say anything about this? Imaginatorium (talk) 09:02, 3 April 2016 (UTC)

Palatalized [ʲ] or approximant [j]?[edit]

I just noticed that the Tokyo article has long indicated the pronunciation as [toːkʲoː], with a palatal diacritic [ʲ] that's not listed on this page. As palatalization would affect countless transcriptions, I think we should come to a consensus whether to transcribe it as [ʲ] or [j]. Japanese phonology is ambiguous on the matter, I don't know enough about Japanese to chime in myself, and someone eventually needs to go through the transcriptions and standardize them. — ˈzɪzɨvə (talk) 17:34, 31 July 2015 (UTC)

Sounds like a [kj] sequence to me, and that's how I usually see it. — kwami (talk) 04:02, 1 August 2015 (UTC)
I just realized, [ʲ] is indeed necessary not just because of Yōon /kj, mj, rj, gj, bj, pj/ but especially because /ki, mi, ri, gi, bi, pi/ are also palatalized and to notate them as [ki, mi, ɾi, bi, pi] would be wildly inconsistent with the narrowness of the rest of the transcription such as devoicing and nasal vowels (if we didn't need them we might as well only use Romanization). In addition, [kj] and [kʲ] contrast phonetically anyway. Nardog (talk) 22:04, 5 May 2017 (UTC)

Problems of clarity[edit]

I am not an expert in phonetics, though I do speak Japanese. A couple of points which seem dubious to me.

(1) ɲ has the example niwa. Is this really palatalised? I know that Japanese speakers generally cannot "do" a 'yi', but I can, and it is quite different from a plain i. (In other words, I can say "nyiwa", which I believe is palatalised, and not "niwa".)

(2) I believe "compressed" is the technical term for what happens to the lips in Japanese wa-wi-we-wo, but explaining it by saying "lips open wider" is very confusing. "Open wider" normally means "further apart", whereas what this is trying to say is that the lips are stretched horizontally, and form the 'w' by moving vertically. I suggest wording like "lips open vertically, rather than horizontally". Comments please.

Imaginatorium (talk) 09:13, 3 April 2016 (UTC)


What animes do you know? PatienceMichel (talk) 21:26, 11 September 2016 (UTC)

That's not related to this article. (talk) 06:22, 11 August 2017 (UTC)

Glothal stop and gemination[edit]

Don't you guys think that the entrance on ʔ/っ (sokuon is a little vague or not thorough enough at least? It not only lacks in examples, but also does not mention the whole idea behind the gemination (jap. 長子音, chōshi-in) occurring in "nm" clusters, as in 專門 (せんもん) being pronounced as [semʔmonꜜ] or [sem:on] (with doubled /m/ separated with a glothal stop) or in 最悪 (さいあく) being [saj:aku] with doubled /j/ also separated with a glothal stop. I guess it might not be gemination but rather something sandhi-ish (jap. 連声, renjō), but that's not my area of expertise anyway. I just think it's worth including.
Vegeta391 (talk) 14:57, 13 February 2017 (UTC).

Move discussion in progress[edit]

There is a move discussion in progress on Help talk:IPA which affects this page. Please participate on that page and not in this talk page section. Thank you. —RMCD bot 16:16, 15 July 2017 (UTC)

"j" as affricate[edit]

The article Mount Fuji tells us it's pronounced [ɸɯꜜdʑisaɴ]. This surprises me (and doesn't seem to accord with what's written in this help page). Why the affricate /dʑ/ and not simply /ʑ/?

OTOH if it is correct, then what's written on this help page needs some correction or elaboration. -- Hoary (talk) 01:42, 16 July 2017 (UTC)

@Hoary: It should be [ɸɯꜜʑisaɴ], or at the very least [ɸɯꜜ(d)ʑisaɴ]. As the footnote explains, /z/ is highly variable, so it should be [z, ʑ] intervocalically and [dz, dʑ] otherwise in our reasonably broad yet non-phonemic system, unless to illustrate the presence or absence of the affrication among certain speakers. There are dialects that retain one or both of the contrasts between [z] and [dz] and between [ʑ] and [dʑ], but Standard Japanese isn't one. See Labrune 2012:64 and Yotsugana for more. Nardog (talk) 17:42, 16 July 2017 (UTC)

ɾʲ is approximated by "dew"?![edit]

That has to be wrong. They sound nothing alike. — Preceding unsigned comment added by (talk) 06:24, 11 August 2017 (UTC)

It is not a mistake. The alveolar tap [ɾ] occurs in major varieties of English only as an intervocalic allophone of /t/ and /d/ (see Flapping). But as far as I know flapping in English doesn't occur before [j], e.g. hit you usually becomes [hɪʔjuː] or [hɪtʃuː], but not [hɪɾjuː]. [ɾ] essentially is a really short [d], hence dew [dj]. If dew sounds nothing like Japanese /rj/ to you, that's probably because you pronounce it the same as do (yod-dropping).
But upon second thought, guardian may be better. There /d/ could become [ɾ] and /i/ could become [j], and it's less subject to dialectal biases. Nardog (talk) 06:57, 11 August 2017 (UTC)