Talk:Japanese phonology

From Wikipedia, the free encyclopedia

This article is part of WikiProject Phonetics, an attempt to build a comprehensive and detailed guide to phonetics and descriptive phonology on Wikipedia. If you would like to participate, you can edit the article attached to this page, or visit the project page, where you can join the project and see a list of open tasks.

Contents 1 Consonants 2 moved article 3 Japanese O 4 back semivowel 5 Approximate English vowels 6 touou wo oou 7 what palatal fricative? 8 Phonetically long vowels 9 Nasal vowel 10 L and R 11 Interesting idea I got from a book 12 Question 13 Desu

[edit] Consonants

At present, this page lists /ɴ̩/ as one of the consonants. While widely acknowledged as a phonetically occuring allophone, it is not, to my knowledge, given as a phoneme in the literature. If I am mistaken, could someone provide references, please? If not, I will take it out.

Moreover, syllabicity (indicated by the vertical line under the symbol) is not normally treated as phonemic, and including it here is definitely a non-mainstream view of Japanese phonology. Accordingly, I have removed it.

Finally, I have added ɸ in the chart, as it is treated as phonemic by the majority of modern scholars. I also aligned ɸ and h so that they would be in the voiceless side of their columns. Godfrey Daniel 00:21, 22 March 2006 (UTC)

I believe that Shibatani treats it as a phoneme or archiphoneme. I dont remember what Timothy Vance says, but I would guess that he mentions something about this. Removing the syllabic diacritic is fine — I dont know how the IPA marks moraic consonants. My original exclusion of ɸ from the phoneme inventory was in keeping with an older analysis — feel free to update the page, I dont know the current literature anyway. peace – ishwar  (speak) 17:22, 22 March 2006 (UTC)

As I recall, Shibatani treats it as the archiphoneme N, not the concrete, place-of-articulation-specified honest-to-goodness segment ɴ. Vance also includes discussion of N (not ɴ), but, wily fellow that he is, never tells you exactly what he thinks.

In IPA usage, syllabic/moraic consonants are marked as you had it, but whether a given segment is syllabic or non-syllabic is not normally treated phonemically. Rather, it comes from a combination of syllable structure, the position of segments in a syllable, constraints on what segments may be syllabic, timing systems, and more. In fact, the IPA rescinded a symbol designed to represent the Japanese mora nasal when they realized that it was really just a regular nasal with a mora, and not a special articulation.

Since those were the references you used in coming up with the table (and it's a beautiful table!), I'll change ɴ to N, and put it in the placeless category. Later, when I have time, I'll add the competing analysis without N and Q (alluded to elsewhere on the page). Godfrey Daniel 20:37, 22 March 2006 (UTC)

Made the change. In doing so, I noticed that there was no row for affricates, so I added one, as well as its members. While the argument can be made that affricates are allophones of stops, this only works in older varieties of Japanese. Vance includes affricates in his analysis of the modern language, as do most other scholars, so I believe we should include them here as well. Godfrey Daniel 20:53, 22 March 2006 (UTC)

I finally got around to changing all the ɴs to Ns, as well as making a few more changes. Godfrey Daniel 21:13, 4 April 2006 (UTC)

[edit] moved article

OK, I moved the article here, in accordance with the consensus reached in the brief discussion at Talk:Japanese_language#Proposal_to_move_Japanese_language.23Sounds_to_Japanese_phonology Tomer ^TALK 13:02, May 29, 2005 (UTC)

Thanks for moving it! However, it needs a lot of work...

Yes, it's good, but it needs work to be better. I'll do some work on it as time allows. Godfrey Daniel 00:21, 22 March 2006 (UTC)

[edit] Japanese O

The Japanese 'O' is described as having a simmilar pronounciation to the O in the english word 'Go'. I've always been taught that it is closer to the O in 'Clock', or something in between the two.

The wovel in clock is much more like to a Japanese A. Japanese O is similar to the wovel in law.

That depends on your dialect of English. The vowel in clock is only close to Japanese A for Americans.. but even in American dialects, the vowel in go is very often a diphthong like [oʊ] or [əʊ], so isn't all that close to the Japanese [o]. --Ptcamn 19:52, 24 August 2005 (UTC)

I doubt there are any easily explained examples of non-dipthonged o's that could be used to compare with Japanese. It's probably simplest just to explain that it's similar, except not dipthonged.  freshgavinΓΛĿЌ  13:52, 22 February 2006 (UTC)

In a lot of the dialects it sounds like the oo in poor or o in pore (in my dialect those two sound the same O.o). -Iopq 07:00, 5 May 2006 (UTC)

[edit] back semivowel

Is Japanese w pronounced like a Spanish intervocalic g, as the article implies? To me they sound quite different... Maybe a /w/ (possibly with the less rounded diacritic) would be more accurate...--Army1987 21:26, 9 August 2005 (UTC)

Yes, you could do that, but as there is no IPA symbol that indicates lip compression the sound must be described. Choosing a given symbol in this case, does not explain what the sound is. Using [w] with a less rounded diacritic is also somewhat misleading. I think it is is better to use [ɰ] following Akamatsu & Ladefoged so that it will not be confused with [w] (which is the way it is symbolized in phonological works and transliteration). peace – ishwar  (speak) 21:58, 20 September 2005 (UTC)

Hi Ishwar,

I didn't see this until after I had changed the symbol to /w/. I'll tell you why I think <w> is better, but go ahead and revert if you wish.

1. ASCII letters are commonly used for their convenience or for their use in the orthography. Therefore people don't expect them to be precise. But when you go to the trouble of using a special IPA letter like [ɰ], people expect it to be correct.
2. Japanese /w/ is coarticulated, which is captured by [w] but not by [ɰ].
3. If you use w with a diacritic, as w̜, people know there's something funny going on, because it's so bizarre, and are more likely to read the explanatory notes.
4. If you want to be precise, perhaps [ɰ͡β̞] would work. But I personally would be quite confused by a simple [ɰ].

Also, do you have any idea if this is what the Iroquois 'unrounded' w is like?

kwami 07:24, 27 September 2005 (UTC)

hi. Although you have good points (capturing the coarticulation is persuasive), I think I prefer ɰ . If you write a bizarre symbol, maybe someone will take the trouble to read about it carefully as well, maybe more so than with w + diacritic. Plus w is so commonly used to indicate the rounded sound. I think that [ɰ͡β̞] is pretty good; I guess you could also use [ɰᵝ]. Akamatsu (coming from a pedogogical perspective in addition to descriptive) is very opposed using w (he also uses other symbols like p’ instead of pʲ as he wants to avoid an interpretation that this represents [pj]). But I dont really have a strong opinion either way. I'll just leave it the way it is (will have to make adjustments though). Maybe someone else will have an opinion. peace – ishwar  (speak) 04:58, 29 September 2005 (UTC)

As of 10/10/06, this article is using both ɰ and w̜, in different places. Shouldn't we pick one, whichever one, and use it? 24.159.255.29 20:59, 10 October 2006 (UTC)

[edit] Approximate English vowels

Japanese vowels can be approximated in English as follows:

/a/ between cap and cup

/i/ as in fit

/u/ as in cook

/e/ as in get

/o/ as in dog (British English)

The above was deleted from the main page without being moved onto the talk page. I think they're more or less all dialect-independent and short vowels, except "dog" which doesn't have a close equivalent in American English. I suggest putting these back in. --DannyWilde 03:54, 29 September 2005 (UTC)

There's always a dialect you didn't consider when designing these. Here, New Zealand, South African, some northern English and/or Scottish (as I understand it) are some I can think of that won't work with these. Further, many colloquial varieties of English often use a centralised vowel in cook; for some of them, boot might be nearer anyway. You might say "but they can just look at the IPA transcriptions", but so can anyone (and the diagram), and get a more accurate understanding as well. Also, British English dog is quite a bit more open than the Japanese vowel. If you want a non-technical description, provide recordings of a Japanese speaker saying the Japanese vowels, and let the reader make their own conclusions about what's closer. —Felix the Cassowary (ɑe hɪː jɐ) 07:59, 29 September 2005 (UTC)

I'd guess most of those speakers would be familiar with, let's say, so-called "British English", though. As for making a recording, I don't know how to do it unfortunately. Until there is a recording, why not keep these? --DannyWilde 10:54, 29 September 2005 (UTC)

There's also, I think, some American dialects that will have trouble with the descriptions (the Northern Cities/Great Lakes area in particular, I think, but various others exist). As for whether the speakers are familiar with other dialects, I'm not sure how reliable that is particularly when expressly stating one dialect and not the others (for instance, it'd be reasonable to assume that the variation in Japanese is equal to or greater than the variance in English); and speakers knowledge of the pronunciation of vowels in different dialects will of course vary (frex, I don't think of my STRUT as being particularly different from British, but mine is almost exactly the vowel intended anyway, whereas between cup and cap will probably get me some weird, contorted and undesirable vowel). Anyway, this is an encyclopædia article and not a tutorial. Also, it seems only fair that if we include a list of approximations in English, we should do that to for French and Dyirbal (the audience is not just English-native, but also EFL who don't necessarily have as extensive an article on Japanese in their own Wikipedia).

To make a recording, you get a speaker (native or with an indistinguishable-from-native accent), a digital recording device (e.g. computer+mike, high-quality MP3 player with record). Record them saying it, then upload the file to the Wikimedia Commons, then make a link to it. There's some pseudo-instructions at Wikipedia:Media, or if you get that far and can't work it out, give me a yell.

—Felix the Cassowary (ɑe hɪː jɐ) 12:47, 29 September 2005 (UTC)

I don't have such a device. Why can't you do it? --DannyWilde 00:58, 1 October 2005 (UTC)

I don't have a native speaker :) (I assumed you did, or at least a good-enough accent, when you said you didn't know how to; if I'm wrong, I apologise.) (—Felix the Cassowary (ɑe hɪː jɐ))

Sorry, I really don't have any equipment. I have some native speakers, though, so I'll try to get some recordings made at some point. --DannyWilde 03:33, 1 October 2005 (UTC)

My idea to resolve this conflict is to make a table of "approximate vowels" with a heading for each dialect. I've added the British ones, and the American "go" and others can add the other ones. --DannyWilde 00:58, 1 October 2005 (UTC)

I'm still not sure of what advantage it offers. It's not encyclopedic. It'd go down very nicely in the Japanese Wikibook, though. —Felix the Cassowary (ɑe hɪː jɐ) 01:13, 1 October 2005 (UTC)

The technical level of the "Japanese phonology" article is very high, representing a lot of skill and knowledge, so I understand that putting this kind of "approximately" table does not exactly fit with the other contents. However, I have taken into account your opinion about problems with dialects and other languages, and I also think that for some people, who don't know a lot about phonology, this table will be the most useful and understandable part of the page. I hope the current form is at least tolerable to a majority of editors and readers. --DannyWilde 03:33, 1 October 2005 (UTC)

I'm not keen in general on this sort of thing, for essentially the same reasons as Felix the Cassowary. But, if you really want it, could you refer to Received Pronunciation rather than the vague and general term British English? I'm not very familiar with Japanese, but I have heard its /a/, and it's really got nothing whatsoever to do with the vowel that I (a British native speaker of English) use in cup (though cap is very close). The same sort of thing probably applies to American English, though someone more familiar with that will know the details.--JHJ 12:13, 1 October 2005 (UTC)

I don't have any objective way to measure which of the vowels in "cup" or "cap" is nearer to the Japanese vowel in "kappu", but it might be worth noting that when transcribing English to Japanese, the vowel in cap is often transcribed using a yōon, as in "kyatto" (Japanese version of "cat"), whereas the transcription of "cut" in Japanese is just "katto". Japanese people find it very hard to tell the difference between these two sounds and katakana versions of English words use the Japanese a for both. I've been listening to Japanese people saying Japanese "a" every day for twelve or so years, and my own guess is that the Japanese vowel is actually closer to "cup" than "cap". But, as I said, I have no objective way to measure it. Thank you for your edits to the page. Please feel free to improve it in any way you can. --DannyWilde 05:24, 13 October 2005 (UTC)

My point is really that these vowels (the ones in cap and cup) are quite variable across English accents, which is why I think descriptions like this should be clear about which variety of English they're referring to. If you speak RP, then your description makes sense, but here (northern England) most people have a vowel in cup that is probably closer to a Japanese /o/ or even /u/ than an /a/. (See foot-strut split.) I've seen far too many descriptions of vowels in non-English languages as "like English cup" or similar, when they invariably turn out to be some [a]-like vowel that sounds to me like the vowel of cap. --JHJ 16:35, 13 October 2005 (UTC)

This is a difficult issue; I would say that only one out of the five English sounds suggested above (e as in "get") is really the closest match to the Japanese sound in what we North Americans like to call a "neutral" accent. In fact in Japanese language textbooks and websites I've only once seen this particular collection of sounds suggested as the best approximate equivalents (I remember since it seemed so unusual). 221.189.192.63 20:03, 2 May 2006 (UTC)

[edit] touou wo oou

This doesn't seem particularly unnatural to me, although I'm not a native speaker. Would you like me to check it with some native speakers? I can easily ask people. Anyway remove it if it's upsetting you, but I thought it was rather a good example of the process described. --DannyWilde 08:38, 13 October 2005 (UTC)

well, it is a possible utterance, but not something that anyone would expect to hear. I put it there because it was fun for me. The example is from Akamatsu (1997). There is a similar example in Bloch (1950) "let's cover the tail" (ō wo ōo). Bloch says "Such an oddity is of course not intelligible to Japanese speakers at first hearing; but when it has been explained to them, they can both understand and repeat it." I think Akamatsu was just trying top Bloch with 7 vowels (over Bloch's 6). If you ask a native speaker, they will probably be amused. It's not serious: you can take it out. – ishwar  (speak) 16:29, 13 October 2005 (UTC)

It wasn't me who edited it, sorry I don't remember the user name, but someone edited it and put a comment "this is unnatural". As for being un-understandable, if you went up to your average English native speaker and said "Let's cover the tail" or "Cover Eastern Europe", you might get some funny looks - taking anything out of context no one will understand it. --DannyWilde 22:31, 13 October 2005 (UTC)

There's an important difference though: if you say to someone "Cover eastern Europe" they might give you odd looks, but primarily because they do understand what you're saying. Whereas the Japanese version is unintelligble. The fact that people "get it" when explained doesn't change that - the fact that it has to be explained means that it's unintelligible. Shinobu 11:01, 3 August 2006 (UTC)

[edit] what palatal fricative?

Doesn't sound like it... hito/ → [çit̥o] hito 人 'person' sounds like h to me -Iopq 03:54, 8 December 2005 (UTC)

Just as hu in japanese is pronounced as [ʍɯ] or even (rarely) [fɯ], so the h in hi is altered by the vowel which follows it to [ç]. Tomer^talk 04:48, 8 December 2005 (UTC)

I thought it was [ɸɯ] for /hu/ -Iopq 01:44, 10 December 2005 (UTC)

Iopq's right. I have never seen [fɯ], much less [ʍɯ], in the literature, nor have I ever heard them.

However, please remember that what you hear is not always what someone said, and this is especially true across languages. (I remember the first time I heard a Fijian say [β], and I thought for all the world it was [v].) Godfrey Daniel 21:11, 22 March 2006 (UTC)

I listened to some people pronounce it carefully and I can definitely say that some pronounce it [h] and some pronounce it [ç]. It was in an anime and the title of a corporation is "Asahi". Some people said [asaçi] and others said [asahi]. It seems the same person would say it the same way so it's not just free variation in terms of strength of articulation. I am going to go to a library and research this issue. -Iopq 20:17, 2 May 2006 (UTC)

[edit] Phonetically long vowels

The article has written:

Phonetically long vowels, then, are treated as a sequence of two identical vowels, i.e. ojiisan is /ojiisaɴ/ not /ojiːsaɴ/

and then lists /ɡjooza/ → [ɡʲoːza] and /seesaɴ/ → [seːsãɴ]. Is there some reason that these two are treated differently than /ii/?  freshgavinΓΛĿЌ  14:05, 22 February 2006 (UTC)

[edit] Nasal vowel

I heard ichi man en as icɕimãẽɴ how does this conform to the moraic nasal section?

That sounds like a simplification (or over-simplification) of very fast speech to me. Even when it's not fully formed the nasal qualities of the mora still exist to a small degree.  freshgavinΓΛĿЌ  12:06, 23 February 2006 (UTC)

That's true, but I don't know how to write morae in IPA. Plus, it seems that a lot of the time in very brisk speech ɴ doesn't count as a mora. You can tell this from songs because when the song requires two syllables a word could be sung as "shunkan" if you need three syllables it could be sung as "shunkaN" and when you need four it could be sung as "shuNkaN." I don't know if that's relevant because lots of crap in songs has nothing to with the language. -Iopq 13:04, 24 February 2006 (UTC)

That's true : ). Modern Japanese music (especially rap) tends to English-ize a lot of the sounds, one of the most interesting being sounds like /naku/, pronounced instead as something like [næk]. I'd still argue that in general, /n/ gets it's own mora in most Japanese music, though. Even though you can interpret "shunkan" as 2 syllables, that doesn't mean there're only 2 morae; it's not the same thing. In fact, you could say the whole idea of "syllables" don't exist in Japanese at all.  freshgavinΓΛĿЌ  03:37, 1 March 2006 (UTC)

According to the research I've done and observations I've made, mora nasals never get lost, but non-native speakers may miss them. In fact, I believe a more accurate transcription of relaxed-speech 一万円 (in isolation) would be something like [itʲɕimãə̃ẽə̃] or even [itʲɕimããẽẽ], where there's one more nasalized vowel between the a of 万 and the e of 円. (BTW, I don't believe that the affricate is fully palatal, but that's perhaps a matter of taste.)

ん always gets counted as a mora. However, as Freshgavin pointed out, syllables and moras aren't the same. However, I believe both are necessary for an accurate description of Japanese. For example, the location of (pitch) accent in a verb can best be described by incorporating the syllable, whereas without it, you have to explain the seemingly-abberant accent in káeru in an unprincipled manner. Godfrey Daniel 21:06, 22 March 2006 (UTC)

[edit] L and R

I noticed that Japanese doesn't have an "l" or "r". Is that why some native Japanese speakers mix l and r when they speak English? I'm not trying to offend anyone here.Cameron Nedland 00:18, 6 March 2006 (UTC)

Though it's not the greatest article, check out Engrish, which explains a little bit. You basically answered your own question. In Japanese, there is no sound equivalent to the English "l" or "r" sound. Because of this—though it may sound suprising to a native English speaker—many Japanese lack the ability to reliably differenciate between the two sounds (the same applies for "v" and "b", and the many sounds symbolized by the letter "a").

The fact that many/most Japanese people don't generally think in roman letters, but instead in katakana (which, of course, lacks individual spellings for "r" and "l"), makes it difficult for them to remember which words are spelled with "r" and which with "l", even if they are among the minority who can pronounce them. I've copied this question to the Wikipedia:Reference desk/Language, which would be a good place to ask a general question like this in the future!  freshgavinΓΛĿЌ  00:28, 6 March 2006 (UTC)

OK sorry.Cameron Nedland 16:12, 12 March 2006 (UTC)

[edit] Interesting idea I got from a book

It seems you can actually treat Japanese and Sino-Japanese phonology in a different way.

Vowels = a u e o

Y + Vowel = ya yu ye(i) yo

this way you can say that ye is pronounced i and combines with consonants the same way ya yu and yo do

s + ya = sha s + yu = shu s + yo = sho s + ye = shi (because ye is pronounced i)

This stems from the fact that i always palatalizes, while e never does so there is no overlap between the two

so the Japanese syllable table could be written with four columns and the y row will be all filled

the writing system does not accomodate this though -Iopq 07:08, 5 May 2006 (UTC)

[edit] Question

The underlined part was recently added to this article:

The name of the pop group in the movie, Dezaato (i.e., Dessert), is purposely romanized differently throughout the movie: "Dessart", "Dessret" and "Desert". But when pronounced, it sounds close to the words "Death Threat."

I know that z is a popular replacer for th in transliteration, but when I try to pronounce the zaa part, it doesn't sound anywhere near "death threat". My question is that is z also pronounced something like th, so that dezaato sounds like "death threat"? I don't think so (since my first language is pronounced very similarly to Japanese and I'm somewhat familiar with Japanese pronunciation), but can anyone verify the claims of that sentence? --TonyM ^{ｷﾀ━( °∀° )━ｯ!!} 19:57, 4 July 2006 (UTC)

It seems like they're trying to find something where there isn't. I could imagine some logic if I assume that the person who thought of the tidbit was thinking along these lines:

• Dessert is traditionally transliterated to the Japanese phonetic system as "dezaato"
• Attempts by some Japanese to figure out the English spelling of "dezaato" are often haphazard
• One of these haphazard trials of spelling appears in the movie as "dessret"
• The English pronunciation of "dessret" once again fitted to the Japanese phonetic system would be "dessu retto"
• "Dessu retto" sounds very similar to the Japanese interpretation of the words "death threat", transliterated as "dessu suretto"

As you can see, this is a ridiculously complex sequence of events to assume without any source and thus should be quickly deleted from Wikipedian and forgotten forever.  freshofftheufoΓΛĿЌ  05:08, 5 July 2006 (UTC)

Thank you! --TonyM ^{ｷﾀ━( °∀° )━ｯ!!} 12:16, 5 July 2006 (UTC)

[edit] Desu

"desu is pronounced [desɯ̥]."

It might be just me, but I hear [des] very often. But then again, I'm getting used to the idea that voice actresses and pop stars speak a completely different language than textbook Japanese :-) Shinobu 11:05, 3 August 2006 (UTC)

It's probably better to not that the [ɯ̥] vowel is in general pronounced with a lighter stress than the other four, and it often shortened or omitted, especially at the end of a phrase/sentence. That would at least account for the half-pronunciation, and the way the mouth forms the sound even if it isn't pronounced.  freshofftheufoΓΛĿЌ  21:13, 7 August 2006 (UTC)

It's just an unvoiced vowel. That's pretty much all there is to it. -Iopq 00:00, 8 August 2006 (UTC)

So I was right after all. I know this kind of devoicing is common, but the statement above doesn't take it into account at all, leading the reader to believe devoicing for some reason shouldn't happen in desu. Given the rest of the article it's hard to see why the statement there at all, actually. Shinobu 08:20, 8 August 2006 (UTC)

Of course you're not wrong, but if you were write that desu is pronounced [des] you wouldn't be entirely right either. I think it is necessary to say first what the basic, original, or "pure" form (whatever that means) is, and then explain how it is actually communicated in common speach.  freshofftheufoΓΛĿЌ  15:47, 10 August 2006 (UTC)