Talk:Japanese pitch accent

comparison with Navajo
just a comment:

Apache lanuages and Navajo are usually described as a tonal languages. Japanese is usually described as an accent language (although a paper has been written claiming otherwise). so the statement may be misleading to not-so-careful readers. - Ish ishwar 00:38, 2005 Jan 29 (UTC)

a nice paper
The Tokyo & Osaka pitch accent varieties are nicely summarized in


 * Haraguchi, Shosuke. (1999). Accent. In N. Tsujimura (Ed.), The handbook of Japanese linguistics (Chap. 1, p. 1-30). Malden, MA: Blackwell Publishers. ISBN 0-631-20504-7.

An online verison of this is available here:

http://www.blackwellpublishing.com/content/BPL_Images/Content_store/Sample_chapter/0631234942%5C001.pdf

Cheers! - Ish ishwar 22:08, 2005 Feb 7 (UTC)

perception by native Japanese speakers
I thought it important to point out that in my experience, most Japanese people are quite aware of the existence of pitch accent, and do not attribute it simply to differing kanji, as this article states.
 * LeeWilson 10:25, 8 Apr 2005 (UTC)


 * Same here. I've had people coach me on the proper placement of the accent in their dialect (and in Tokyo people seemed well aware of this too). People could rattle off minimal pairs, usually with the idea that the accent pattern was just the opposite in Tokyo. And of course they understand you when you get the accent wrong - in Kansai and Shikoku they hear it that way on TV all the time, and in Tokyo no one had the slightest problem with my Kansai accent.
 * One thing that's not clear about the article: is it saying that the syllable after the accent has a falling pitch? (That's the only way it can work with the examples given.) As far as I know, Kansai allows falling pitch on final morae only, whereas Tokyo doesn't allow any at all. kwami 06:35, 2005 May 22 (UTC)

Here is "nonsense" that Kwamikagami has just deleted:

''Accent is sometimes taught to non-Japanese learners of Japanese, but it is never taught to the Japanese themselves in grade school, so most of them are not explicitly aware of its existence. Most Japanese people who are not linguists will deny that there is any variation in pitch between Japanese syllables, since Japanese is not a tonal language like Chinese. Unfortunately, this does not necessarily mean they'll understand you if you say KA-ki ("oyster") when you should say ka-KI ("persimmon"). They realize that they do pronounce these two words differently, but curiously enough, attribute this to a difference not of pitch but of kanji.''

It seems to me that, while poor, this has some content that might be improved and preserved. Is accent taught to the Japanese at grade school? I hadn't thought so. Of course it's daft to attribute different pronunciations to different orthography, but this is what I have been told, in all seriousness -- of course by non-linguists. (I guess and hope that this is a minority opinion, even among non-linguists.) And, to comment on the objection raised above, of course people in Tokyo understand Kansai Japanese perfectly well and vice versa -- but that's not the point that was raised (rather clumsily) in the mostly-deleted passage, which is (unclearly) either (a) whether Japanese people understand the results of idiosyncratic L2 Japanese pitch accent, or (b) whether they understand the results of sprinkling one L1 dialect with samples of a different L1 dialect.

But I'd agree that a description of misunderstandings of pitch accent, if worth bothering with at all, is much less important than a clear presentation of a well-informed analysis of the facts. -- Hoary 02:53, 2005 Jun 7 (UTC)


 * I read this before it was deleted, and I can't help agreeing that it sounds very reasonble, though it would of course be good to have it referenced. I feel that kwami has of lately been way too eager to remove literally everything that is not to his personal agreement. As far as I know the comment about orthography sounds plausible considering how hard it seems for most people to let go of their native orthorgraphy when describing the phonetics of their own language. I mean, just have a look at our own language articles. The majority of the editors that don't have access to phonetic literature seem to revert to lists of how graphemes correspond to sounds by default when wanting to describe phonology, no matter if the orthography is phonemic i nature or not.
 * Japanese pitch accent is fairly similar to how tonal word accent works in Swedish, yet this is never explicitly taught in school to native speakers. This is simply because it's nothing that needs to be actively taught. Most Swedes have mastered the Swedish word accents before the age of 3 and are usually happily unaware that it is realized differently in other dialects than their own or might not even exist at all (like in Finland-Swedish).
 * Peter Isotalo 12:47, Jun 8, 2005 (UTC)


 * A comment that Japanese aren't taught standard (or even local) pitch accent in school is correct, as far as I know. For instance, many children across the country now use Tokyo pitch accent, and this is commonly attributed to the influence of television in spreading the standard language, not of school. I only took out the 'not learned in school' comment because it was a minor point in a misleading paragraph. The paragraph had two errors: getting the accent wrong does not make you unintelligibe as the paragraph suggested, the way that getting Chinese tone or English stress accent wrong can make you unintelligible; rather, it's more like speaking with a Southern US or Irish accent in English. Obvious, but not incomprehensible. (Rereading the paragraph, I see that the example was two foods, so that could be confusing if you just asked a shopkeeper if he sells "kaki" without any context to disambiguate.) And although I've met lots of Japanese who think the accent system is fun, and have taught me all sorts of minimal sets that they found amusing, I've yet to meet anyone who didn't recognize it or who attributed it to differences in kanji. I'm not saying that such people don't exist, but a blanket statement is at best misleading. I think some supporting evidence or testimonials (people who know from personal experience) would be in order. It sounds like Hoary does have such experience, so I'd be happy for him to restore the paragraph. It would be interesting to know whether people only say this about Sino-Japanese compounds, as in technical vocabulary, which people often disambiguate by drawing the kanji on their hand or in the air with their finger; or if he knows anyone who actually believes that the native Japanese words for oyster and persimmon are only differentiated by their kanji. (That would be odd.) And the examples given in the article are of everyday native vocabulary; technical terms of foreign origin, even if well integrated, create their own issues in many languages. (Think of Greek terms in English.) We should have the qualification, though, that lots of Japanese are quite aware of the pitch accent system. It's just as much in people's consciousness as desert and dessert are in English. (Who knows, maybe some English speakers believe the only difference between those words is their orthography!) As one concrete example, many (most?) monolingual Japanese dictionaries give the pitch patterns for all entries, and these are explicitly described in terms of pitch, so any Japanese who uses a dictionary is likely to know that Japanese has lexical pitch variations.


 * As for my removing anything I don't personally agree with, I only removed this after LeeWilson made the same comment, and no one contradicted him for six weeks. Previously I hadn't touched it. It may have been rash, but at least now we're correcting the passage, which no one had bothered to do before. (Funny how saying "I'm deleting this nonsense" gets a discussion going when "I think this is wrong" gets no response at all.) kwami 19:09, 2005 Jun 8 (UTC)

Accent vs. Emphasis
I was wondering if anyone else thinks that it would be worthwhile to point out the difference between the Japanese way of emphasizing a syllable and the English way. I have heard many Japanese speakers mention that English speakers who learn Japanese often retain the habit of lengthening the emphasized syllable in addition to the pitch change, whereas Japanese speakers retain the same rythm no matter whether the syllable is emphasized or not. Mattopia 06:41, 15 October 2005 (UTC)

Soundclips
I wonder if anyone can obtain or create some soundclips as examples. I think they'd be a great suppliment to this article.

80% of all Japanese words without accent?
The text claims that 80% of all Japanese words would be accentless -- of course without giving citation. This must be another mistake. How are words defined here? If you look for example at Shin Meikai Nihon-go akusento jiten, you get the impression, that about 50% have accent and 50% don't.


 * You seem to be about right, at least according to the /Accent/ paper above by Haraguchi, Shosuke. (1999). who says "The largest number of words (approximately 55 percent) are unaccented. Among the accented words, approximately 75 percent have accent on the antepenultimate mora" (7). However, the paragraph preceeding says "Let us begin with a discussion of nouns. ..." and the section heading is "1.3.1 The tonal melody of nouns", so I'm unsure whether the author refers to nouns or words. --209.151.140.79 (talk) 03:25, 15 October 2011 (UTC)

Common Ways to Represent Accent in Japanese
I've noticed a few textbooks (published in Japan) and the very few dictionaries (for English users) that bother to give the accent patterns employ a sort of bracket to show how to accent words. At least two examples of this method are Merriam-Webster's Japanese-English Learner's Dictionary and Kodansha's Basic English-Japanese Dictionary, but I've seen it used elsewhere. Would it make sense to try to incorporate this system into the article along with or rather than typing capital letters or the ↓ symbol for the sake of conforming to what some learners are accustomed to? Dambrosio 06:09, 15 November 2007 (UTC)


 * Normally we try sticking to the IPA, so that people who don't know the language can still follow the article. However, it wouldn't hurt to add an ad hoc transcription in addition to the IPA. kwami 12:05, 15 November 2007 (UTC)
 * But for pitch accent, the argument is futile. The arrows to represent pitch accent are not known to common people who know IPA. --188.99.142.18 (talk) 09:22, 6 September 2017 (UTC)

"Pitch accent" in Japanese
What are the common or standard ways to say "pitch accent" in Japanese? I found 高さアクセント and 高低アクセント but not sure if these are unambiguous or the most common or standard. Anatoli (talk) 22:35, 19 March 2009 (UTC)


 * 高低アクセント or ピッチ・アクセント. Ja Wiki uses 高低アクセント. Oda Mari (talk) 18:48, 20 March 2009 (UTC)


 * Thanks, Oda Mari. Anatoli (talk) 20:45, 20 March 2009 (UTC)

Intonation of "hashi"
Regarding this phrase:

háshiga 'chopsticks', hashíga 'edge', or hashigá 'bridge'.

It seems to indicate that the intonation is low-high-low for 端が (edge) and low-high-high for 橋が (bridge), but this is the opposite of what appears in the Goo dictionary.

http://dictionary.goo.ne.jp/srch/jn/%E3%81%AF%E3%81%97/m0u/

I'm confident that the current entry is wrong, but I don't want to step on any toes. Shall we change it? —Preceding unsigned comment added by 208.120.180.50 (talk) 17:11, 3 September 2009 (UTC)

Edit: I decided to make the change instead of waiting for peer review because of the Goo reference. —Preceding unsigned comment added by 208.120.180.50 (talk) 17:15, 3 September 2009 (UTC)


 * Yes, your correction brings that into line with the table, where I got it right. kwami (talk) 19:30, 3 September 2009 (UTC)

dialects
A lot of gaps filled in under 'dialects' need to be verified; different sources may not mean the same thing by 'eastern' & 'western' type accents. I don't have a good country-wide description available. — kwami (talk) 08:55, 17 June 2010 (UTC)

Category:Pitch (music)
I don't understand while the inclusion of this category has been reverted twice. The main article of this category says: "Pitch represents the perceived fundamental frequency of a sound" which is AFAIK the feature this article is about. Moreover, Japanese pitch accent sends to Pitch accent, which in turn sends to Pitch (music), i.e. to the main article of the category I tried to add. Perhaps Category:Pitch (music) or Pitch (music) should be renamed or merged, but I think Category:Pitch (music) is the most relevant available category for the moment (Category:Japanese phonology is not precise enough). Apokrif (talk) 20:58, 28 July 2010 (UTC)
 * I am a native ja speaker but I don't understand your claim at all. As far as I know, linguistic pitch has nothing to do with musical pitch. If you think they are relevant, please show me RS first. Thank you. Oda Mari (talk) 05:26, 29 July 2010 (UTC)
 * Do you mean that the Japanese pitch accent, or musical pitch, or both, are not "the perceived fundamental frequency of a sound" ? Apokrif (talk) 09:51, 29 July 2010 (UTC)
 * I mean the Japanese pitch accent. Oda Mari (talk) 14:30, 29 July 2010 (UTC)

Regarding the "normative Japanese pitch accents" example
The audio example gives examples of words in isolation. For example, 端 is pronounced differently from 橋 in the audio example, whereas the "downstep" analysis of spoken Japanese, which the rest of the article uses, suggests that there is no difference between them at all except when there is a following word such as a particle. Although the audio was provided by a native speaker, I do wonder if maybe she was hypercorrecting and in fact would not pronounce 端 as completely flat in normal conversation. (It's not unknown for native speakers of languages to do this. For example, in Spanish it is said that 'b' and 'v' are pronounced identically. But many will nevertheless pronounce them differently when, and only when, speaking extra "clearly" for this sort of lesson.) - furrykef (Talk at me) 00:41, 1 August 2010 (UTC)
 * Whether that's hypercorrectness, I don't know. I do know that I mess up a lot when I'm speaking quickly; all people do. But I don't consider my stumbles normative, correct or beneficial to communication. — Preceding unsigned comment added by 82.139.87.238 (talk) 00:13, 12 September 2011 (UTC)
 * The scalar pitch section and some discussion in another section below describe how these two could be different without it being hyper correction. There is a third, middle phonetic tone M that appears after the initial mora until the accented mora, which is H, or until the end in heiban words. However since only the downstep is phonemic, LM and LH would not be phonemically distinguishable in isolation to native speakers except extralinguistically. The difference would however be audible to speakers of other languages.165.91.13.85 (talk) 08:01, 13 February 2019 (UTC)

Few questions
1) How big is the accent? Would it register on a musical scale?

2) Is there an on-line dictionary to look up the pitch accent of words?

3) Are there any reliable scientific articles on the Downstep section? Some things in it seem suspect.

4) How important is pitch accent really? Given the differences between dialects, in how much is the meaning of a word derived from the accent, rather than from context?

Also, someone with no knowledge of Japanese will be looking for the case where the accent falls on the last mora, and not find it. It is heiban, but that isn't clear from the text at all. Wouldn't it be better to just say that if the accent falls on the last syllable and the word is spoken in isolation, it cannot be heard? And is it not the case, that if another word were to follow it, the pitch would drop? — Preceding unsigned comment added by 82.139.87.238 (talk) 00:13, 12 September 2011 (UTC)


 * (1) yes
 * (2) don't know. Most print dictionaries only indicate Tokyo dialect.
 * (3) several. What do you find suspect?
 * (4) not very. There are much greater barriers to communication between dialects than pitch accent, though it doesn't help.
 * — kwami (talk) 23:19, 31 December 2011 (UTC)

contour tones in Osaka?
Are [hî] (日) and [kǐ] (木) correct? I thought only Kagoshima had contour tones in one-mora words. Shouldn't they be [hí] (日) and [kì] (木)? Or is this a phonemic distinction only in Kagoshima? — kwami (talk) 22:33, 31 December 2011 (UTC)
 * In Kansai dialect one-mora words are pronounced with long vowels. So some books describe them as [híì] and [kìí]. 日 always has falling tone, and 木 always has rising tone. --180.10.81.176 (talk) 00:59, 1 January 2012 (UTC)
 * Ah yes. Forgot. Thanks. — kwami (talk) 03:20, 1 January 2012 (UTC)

Some problems
I am a Japanese and my English may be incorrect but there are some problems in this article so I want to discuss and revise it.

Why does this article say that the pitch rise in Tokyo dialect is gradual and "from mora to mora"? What reliable sources say so? In Tokyo dialect, pitch rise occurs between the first mora and the second mora, but the following moras are flat or even descend slightly (physiological natural falling) if you record the pitch with a machine.

It is inappropriate to say that some accent patters of Tokyo, Keihan, and Kagoshima type are similar, such as "Keihan (Kyoto–Osaka)-type dialects … have nouns with both patterns", or "Low-tone accented words …are like accented words in Kagoshima", because these three types of accent have independent systems of each other. While Kagoshima accent has only two patterns: accented (A type (A型), which has downstep) or unaccented (B type (B型), which has no downstep), Tokyo type can put an accent(or downstep) on every mora, so Tokyo type has N+1 patterns, and Keihan type has high/low-initial tone and accent(downstep), so Keihan type has 2N+1 patterns (accented words have high/low-initial tone differences and the accent(downstep) can be put every mora). You should not explain that high-(initial-)tone words in Keihan type are like Tokyo accent. If Keihan dialect speakers hear words of Tokyo dialect, they describe them as low-initial tone. If the first mora of a word has low pitch, the word is heard to have low-initial tone for Keihan type speakers. --180.10.81.176 (talk) 02:50, 1 January 2012 (UTC)


 * Had a source for gradual rise last year. Is it not in the article? I remember clear pitch traces, rising gradually and then dropping. In most languages, rising tones are gradual and falling tones abrupt. However, if only a two-way distinction is phonemic, then people will filter out the transitions as meaningless and will hear only high and low.
 * The other wording I'm sure can be improved. The problem is what "like" means. I meant this as a structural analogy, for the reader trying to come to grips with the possible patterns, not a phonetic resemblance. I'll try to clarify. — kwami (talk) 03:26, 1 January 2012 (UTC)
 * In Tokyo N+1, N is the number of syllables, not moras, correct? But in Osaka 2N+1, N is the number of moras? How high of an N do these patterns hold for? — kwami (talk) 03:49, 1 January 2012 (UTC)

Thanks for fixing low tone as marked. That's a big help. — kwami (talk) 07:02, 1 January 2012 (UTC)


 * (I am the same person to 180.10.81.176.)
 * Which source said? I can't find it. The graph of this page shows in 端が the pitch reachs its peak at the second mora, does not show gradual rising.
 * I also meant the differences as structural ones.
 * Yes. In Tokyo N is the number of syllables, but in Osaka N is the number of maras. --219.99.112.248 (talk) 09:31, 1 January 2012 (UTC)


 * You appear to be correct, but with a complication. The word in the graph you linked to is too short to distinguish the two cases. We need a word with a 3- or 4-syllable transition. Li (2006) The Handbook of East Asian Psycholinguistics: Japanese shows a rising tone on the first mora of an unaccented phrase, continuing a bit on the second, and normal prosodic declination after that. The actual pitch for that particular trace (for /uerumono/) is . All the accented examples I can see are accented on the second mora, though, so they don't really address the issue. /ueꜜrumono/, for example, is . One thing that differs from our account, however, is that there is no high pitch in the unaccented phrase: it's vs, not (LHHHH vs LH*LLL).


 * Haven't had a chance to more than skim it, but Kitahara (2001) Category Structure and Function of Pitch Accent in Tokyo Japanese has numerous examples of LMH*L. He calls the M's "phrasal high"s (the traditional term), which are fixed on the second mora. He notes that this phrasal "high" tone is not as high as an accentual high tone, and that this difference in pitch needs to be described in an adequate account. Indeed, in the pitch traces in Li, they're half as high. That is, the "high" tone on non-tonic syllables is phonetically mid tone, and only H*L accents are at the upper end of the pitch range. I always saw that second mora as having a merely transitional pitch, but it would now seem it's more than that: σσσσσ is LMMMM↘, whereas σσσꜜσσ is LMH*LL, a H*L superposed on the LMMMM↘. Kitahara cites previous research with akai seꜜetawa, but does not give a pitch trace to show if the pitch rises gradually over the akai to the se, or if it levels out on the kai and then jumps up on the se. Also, the pattern is phrasal intonation, across all the words in the phrase, while the H*L (with the fall on either mora, perhaps depending on phonetic environment?) is lexical and superposed on the phrasal intonation, rather like English primary and "secondary" stress.
 * Don't know what that means for what we call "high" in Osaka: if it also jumps up before the drop, then the contours of high- and low-tone words have similar shapes. It would be nice to see pitch traces. — kwami (talk) 11:38, 1 January 2012 (UTC)

(I am the same person to 180.10.81.176 and 219.99.112.248)

You may be correct in that the accented words in Tokyo dialect(or standard Japanese) is LMMMMHLL rather than LHHHHHLL, though I've never seen any descriptions as LMMMMHLL in books written in Japanese. The pitch may jump up before the drop, but I'm not sure which pattern Tokyo dialect has. I can find many words accented on other than the first and the second mora, such as /otoRto↓/(弟), /mizuu↓mi/(湖), and /hirosima↓keN/(広島県).

You said that there is no high pitch in the unaccented phrase, but I believe there is no meaning in considering whether /ueru/ is LMM or LHH. We should pay attention only to where the pitch rises or falls. LMM and LHH are the same meaning in that pitch rises between the first mora and the second mora. Unaccented words have pitch rising only between the first and the second mora, and accented words may have pitch rising before the accented mora as well as the initial rising, so the accented mora would have higher pitch than the unaccented words, though. But whether H or M has no relation to whether a word is accented. Japanese speakers differentiate whether an word is accented by hearing downstep. The pitch falling after the accented mora is abrupt. While /uerumono/ would be [u˩˨e˨˧ru˧mo˧˨no˨˩], normal declination on unaccented syllables are not perceived to be declination to Japanese speakers.

No. In high-initial-tone words of Keihan type, the pitch does not jump up before the downstep. All the moras preceding the downstep are high and flat. On the other hand, in low-initial-tone words, the pitch rises before the accented mora in Osaka.--219.99.112.4 (talk) 02:55, 2 January 2012 (UTC)

Oh, N would be the number of moras in Tokyo, too. "Mora or syllable" is difficult question as to Tokyo dialect. For one-syllable but two-moras words, /CVN/ or /CVR/, two patterns are permitted (so I said N is syllable), but /CV↓N/　is not the same accent to /CV↓/, so it would be better to consider the moras the units.--219.99.114.76 (talk) 08:44, 2 January 2012 (UTC)


 * But since */CVNꜜ/ is not possible in Tokyo, adding a more does not increase the number of patterns. Only adding a full syllable would do that, right? — kwami (talk) 13:02, 2 January 2012 (UTC)
 * Yes, you are right. But one-syllable words have three patterns: 1.unaccented, 2./(C)V↓/, and 3./(C)V↓N/ or /(C)V↓R/, though one-mora words have two.--219.99.112.200 (talk) 13:44, 2 January 2012 (UTC)
 * Can you think of a three-way contrast, say s.t. like our nihon example? — kwami (talk) 14:04, 2 January 2012 (UTC)

Ambiguous / Poorly Chosen Example in the 'Syllabic and Moraic' Section.
The second half of this section, which talks about the relevance of syllables and nasals in the Tokyo dialect, refers to "kai" as a single syllable. I have seen a few sorces that claim "ai" to be a diphthong, but most, including most relevant pages on this website, would consider "kai" to be two syllables in hiatus.

I recommend replacing "kai" with a word such as 精, "sei", /séè/, (polish / refinement), because it demonstrates the chroneme / double vowel that is much more often considered to be a moraic coda than a seperate vowel. — Preceding unsigned comment added by 107.77.218.106 (talk) 18:03, 12 July 2017 (UTC)

Arrows in IPA
I don't know Japanese. Curiosity led me to this article when I saw a Japanese IPA pronunciation that ends with an arrow in the final position. My naive logic makes me think this could not really affect the sound of the word itself, and would only become clear when the accent of the following sound was modified. Is my guess true, or a mistake? TooManyFingers (talk) 16:01, 30 May 2023 (UTC)
 * More or less. See § Scalar pitch, #3. Nardog (talk) 16:34, 30 May 2023 (UTC)