Help talk:IPA/Japanese

Palatalized [ʲ] or approximant [j]?

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)

Maybe someone changed the recording, but the current one definitely does not have [kj] but [kʲ]. I have [kʲ] in my native language (Polish) and we don't use it before the mid back vowel, we use [kʲj] instead. The difference is very obvious to me. Mr KEBAB (talk) 19:17, 11 October 2017 (UTC)

Glothal stop and gemination

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

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

The article Mount Fuji tells us it's pronounced Template:IPA-ja. 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 Template:IPA-ja, or at the very least Template:IPA-ja. 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"?!

That has to be wrong. They sound nothing alike. — Preceding unsigned comment added by 67.80.164.161 (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)

Inconsistency in gemination

The way the key was put together allowed two (or three if you count nothing) ways to notate gemination, by ⟨ː⟩ (length mark) and by ⟨◌̚⟩ (no audible release). But in other languages that have gemination, such as Latin, Italian, Arabic, we just repeat the consonant. The use of length mark is allowed in Catalan and Tamil according to the respective keys, but this doesn't suit for Japanese when syllabification is taken into consideration (see Japanese phonology#Gemination). We should follow Latin etc. It's much simpler. Nardog (talk) 12:04, 4 October 2017 (UTC)

き as kʲ is inaccurate

It should be k, just as with く and け. ···日本穣 · 投稿 · Talk to Nihonjoe · Join WP Japan! 07:09, 12 December 2017 (UTC)

Well, name one scholar who says so. All consonants of Japanese are heavily palatalized before /i/, either phonetically or phonologically. Try saying just the consonant of き ki, and then あ a, う u, or お o. You will get きゃ/きゅ/きょ kya/kyu/kyo rather than か/く/こ ka/ku/ko.

In fact some linguists (e.g. Vance 2008, Akamatsu 1997) phonemically group the consonant of き ki with that of きゃ/きゅ/きょ kya/ki/kyu/kyo rather than with か/く/け/こ ka/ku/ke/ko. In other words, they analyze きゃ/き/きゅ/きょ as /kʲ/ + /a, i, u, o/, rather than as /kja, ki, kju, kjo/. Nardog (talk) 10:18, 12 December 2017 (UTC)

See, this is where you are showing your lack of knowledge. There is only one "consonant" in Japanese, and that's the syllabic ん. Everything else is a syllable. きあ does not sound like きゃ. It's not even close. The same with きう/きゅ and きお/きょ. They are completely different. ···日本穣 · 投稿 · Talk to Nihonjoe · Join WP Japan! 17:36, 12 December 2017 (UTC)

Of course Japanese has consonants. Every language has vowels and consonants. I think you're mistaking morae for segments. き is a mora (orthographically, a syllabogram), which is composed of a consonant and a vowel.

Here is what Okada (1999:118) says:

Consonants generally are strongly palatalized before /i/, as /mi/ [mʲi] 'body' ... /j/ affects the preceding consonant as /i/ does, and is itself absorbed, thus: /mjakú/ [mʲakú] 'pulse', /tˢja/ [tɕa] 'tea', /sjóː/ [ɕóː] 'prize', /kanjuː/ [kaɲ̟uː] 'joining'.

And you can demonstrate this by trying to say only the k part of き ki but not the い i part, immediately followed by, e.g., あ a. The resulting sound would resemble きゃ kya more than か ka.

And this is something no scientist would argue with. It is only whether the sound, which is phonetically, inarguably, [kʲi], is better represented as [kʲi] or [ki] in Wikipedia's own IPA-based transcription scheme for Japanese words that could be open to debate. Nardog (talk) 18:51, 12 December 2017 (UTC)

Apparently you speak for all scientists? Are you a scientist? Please don't try to bolster your argument by making wild and completely unfounded claims to authority which doesn't exist. If you say just the "k" part of き followed by あ, you get か, not きゃ. You never get きゃ. Your argument makes no sense to anyone with more than a rudimentary grasp of Japanese. ···日本穣 · 投稿 · Talk to Nihonjoe · Join WP Japan! 21:25, 13 December 2017 (UTC)

@Nihonjoe: Judging by the IPA and my experience with other languages and phonetics in general, he is right. You should start providing sources for your claims, otherwise it just looks like you're the one appealing to authority which doesn't exist (or not even that). This is no place for WP:OR. Mr KEBAB (talk) 10:33, 14 December 2017 (UTC)

I'm not a scientist but I can cite scientists. The quote above is from the book Handbook of the International Phonetic Association, published by Cambridge University Press. (Do these organizations' names signal anything to you?)

I also happen to have this newly published book 『日本語の音』 (Nihongo no oto, "Sound of Japanese") right in front of me, which says the following (p. 37):

拗音のうち，キャ行 [kʲ]，ギャ行 [ɡʲ]，キ゚ャ行 [ŋʲ]，ビャ行 [bʲ]，ピャ行 [pʲ]，ミャ行 [mʲ]，リャ行 [rʲ] には，[j] が付随する．それは，[k] [ɡ] [ŋ] の調音点が軟口蓋，[b] [p] [m] は両唇，[ɾ] は歯茎であるため，拗音（[j] があること）また母音が [i] の際には硬口蓋寄りの調音点になる．キ・ギ・キ゚・ビ・ピ・ミ・リについても精密表記では [kʲi] [ɡʲi] [ŋʲi] [bʲi] [pʲi] [mʲi] [ɾʲi] となる．これを口蓋化または硬口蓋化という．

Which, for the reference of those who don't speak Japanese, translates as:

Out of all yōon, morae beginning with ky [kʲ], gy [ɡʲ], ngy [ŋʲ],* by [bʲ], py [pʲ], my [mʲ], ry [rʲ] (sic) are accompanied by [j]. That is, since the place of articulation for [k] [ɡ] [ŋ] is velar, [b] [p] [m] are bilabial, and [ɾ] is alveolar, the place of articulation becomes closer to the hard palate in the case of a yōon (the presence of [j]) or when the [tautomoraic] vowel is [i]. Ki/gi/ngi/bi/pi/mi/ri also become [kʲi] [ɡʲi] [ŋʲi] [bʲi] [pʲi] [mʲi] [ɾʲi] in narrow transcription. This is called palatalization or hard-palatalization.

* 「キ゚」 denotes the nasal allophone of /ɡ/ known as 鼻濁音 bidakuon, which is produced by some speakers especially in formal registers. Ng is merely an ad hoc transliteration of mine.
(Hope my translation's serviceable, but I can't guarantee its accuracy.)

If you couldn't produce the sound like I told you, that is probably because you have already internalized the Japanese phonemes in your brain. (If that didn't work, I guess you can alternatively put a chopstick or something in your mouth and try to pronounce き ki and こ ko to figure out that the tongue touches slightly different points of the roof of the mouth, but I wouldn't recommend it because I don't want to cause someone to hurt their mouth or throw up.) After all, most speakers don't realize the difference between the p's in pie and spy until they put their hand in front of the mouth, or the one between the i's in 誓う chikau and 違う chigau until they put their finger on the Adam's apple. These minute―granted, minute―differences are hard to realize especially by native speakers.

Now that I think about it, I suppose your confusion comes in part from the fact that the entirety of き is in bold in the key. Notice only k is in bold in the transliteration. Obviously Japanese has only syllabaries and not an alphabet, so it is impossible to emphasize only a consonant and not the vowel that accompanies it, except ん or っ. き would be [kʲi], not [kʲ] as the title of this section suggests, according to the current key.

I can cite more if you want, but again, this is just one of the most basic characteristics of (most varieties of) Japanese that I don't think even Curly Turkey or Hijiri88 are going to disagree with. If you still think I'm "showing [my] lack of knowledge" or "making wild and completely unfounded claims to authority which doesn't exist", I don't know what to tell you. Nardog (talk) 21:07, 14 December 2017 (UTC)

@Nardog: I haven't been following this whole discussion throughout all the multiple fora, so forgive me if you have already acknowledged this somewhere yet, but the only really important part of Okada 1999 for our purposes is his first sentence: The style of speech illustrated is that of many educated Japanese brought up in Tokyo or other areas with similar pitch accent systems. This means that for the vast, vast majority of our Japan-related articles, the specific niceties of the phonetic transcription prescribed here are at best off-topic and anachronistic, and at worst arguably plain wrong. (So pretty much what CT says below, but more wordy and somehow less eloquent.) Hijiri 88 (聖やや) 11:33, 13 December 2017 (UTC)

And English fortis consonants are always aspirated in syllable onset, but it would be wrong to notate such in the IPA pronunciation guides to most article leads (e.g. [tʰəˈɹɒɾ̃o(ʊ)]). Please consider context—Wikipedia is not a linguistics journal, but a general layperson's reference work. Curly "JFC" Turkey 🍁 ¡gobble! 10:29, 12 December 2017 (UTC)

@Curly Turkey and Hijiri88: I'll respond to you in a moment including at WT:JAPAN. But for the moment I will say I concede [kʲi] is probably persnickety and [ki] is enough. I'll expound later. Thanks for your patience. Nardog (talk) 21:07, 14 December 2017 (UTC)

Edited the guide and existing uses of {{IPA-ja}} so the palatalization symbol is no longer used before [i]. The rationale for it was that it is phonemically more economical as I explained above, but I acknowledge き as /ki/ is equally defensible and to transcribe it as [kʲi] might have been overly detailed and less accessible. Nardog (talk) 17:07, 16 December 2017 (UTC)

Note for [ɯ]

@IPA editor: I'm not disputing your addition because I think it's inaccurate. It's because I find it too insignificant to include in a guide for layreaders. Also note that Okada (1991/99) is merely talking about Tokyo speech. Especially in and around Osaka, /u/ is distinctly back and rounded. At the other end of the spectrum, it can be as front as [ʏ]. There's just too much variation. This is why the note avoids going into phonetic details. And since [ʏ], [ʉ], [ɨᵝ] and [ɯ] are all acoustically very similar, [ɯ] does the job of giving the readers a target perfectly adequately. Nardog (talk) 14:34, 10 January 2019 (UTC)

OK, got it. IPA editor (talk) 17:15, 10 January 2019 (UTC)