Japanese pitch accent

Japanese pitch-accent types

  Keihan type (pitch drop plus tone)

  Tokyo type (variable pitch drop)

  N-kei (1-3 pattern) type (fixed pitch drop)

  No accent

  intermediate (Tokyo–Keihan)

  intermediate (Tokyo–none)

Japanese pitch accent is a feature of the Japanese language that distinguishes words by a distinct tone that focuses on a specific mora ('accent') in most Japanese dialects. The nature and location of the accented syllable for a given word may vary between dialects. For instance, the word for "river" is /kawá] in the Tokyo dialect, with the accent on the second mora, but in the Kansai dialect it is /káwa/.

Reading of the first two paragraphs of the chapter 1 of Botchan in the Tokyo accent

Reading of the same part of Botchan in the Kansai accent

Tokyo Japanese accent

This section deals principally with the accent of modern standard Japanese, the Yamanote variety of the Tokyo dialect.

Falling accent and the nucleus

In contrast to stress accent (強弱アクセント・強さアクセント, kyōjaku akusento/tsuyosa akusento; lit. 'strong-and-weak accent/strength accent') in English, Japanese is said to have pitch accent (高低アクセント・高さアクセント, kōtei akusento/takasa akusento; lit. 'high-and-low accent/height accent'). This means that Japanese speakers make primary use of pitch to distinguish words, while English speakers also make use of length, articulatory strength and vowel quality.

Japanese pitch accent is characterized as "a steep drop from a relatively high pitch to a relatively low pitch," henceforth an "accentual fall" or "drop." In Tokyo Japanese, it is the presence or absence, as well as the exact location, of the accentual fall within an utterance that are lexically contrastive, not the exact pitch contour. This sets Japanese apart as a "pitch-accent" language, as opposed to a "tone" language such as Mandarin, where it is the pitch contour of a word that is contrastive.

The mora at the start of an accentual fall can be called "accented," or an "accent nucleus" or "accent kernel" (アクセント核, akusento-kaku). It carries a high tone and is followed by a mora with a low tone. "Falling nuclei" (下げ核, sagekaku) are characteristic of Tokyo Japanese, but it is by no means the only way for accent to be realized; for example, the Hirosaki dialect features "rising nuclei" (昇り核, noborikaku), which involve an accentual rise from a low pitch to a high one.

Types of accent patterns

Tokyo Japanese belongs to a group of dialects with n+1-type (n+1型) accent. This means that a Tokyo noun phrase with n morae has, theoretically, n possible locations for the nucleus, as well as not having a nucleus at all, hence n+1 possible patterns in total; however, quinquemoraic or longer nouns do not have all n+1 possibilities, but at most four. Other dialects may only have 2- (such as in the Kagoshima dialect) or 3-type (such as in the Oki Islands dialects) accent.

The n+1 possibilities of accent patterns can be classified as follows (in the diagrams to the right, the nuclei are marked in red):

• Unaccented/Atonic (平板式・平板型・無核型, heibanshiki/heibangata/mukakugata; lit. 'flat-style/flat-type/unnucleated'): The phrase in question has no nucleus. An unaccented phrase can be of any length.
• Accented/Tonic (起伏式・起伏型・有核型, kifukushiki/kifukugata/yūkakugata; lit. 'up-and-down-style/up-and-down-type/nucleated'): The phrase has exactly one nucleus which is the first mora, the nth mora, or any mora between them.
  • Initial-accented (頭高型, atamadakagata; lit. 'high-head-type'): The nucleus is the first mora. An initial-accented phrase can be of any length. An initial-accented unimoraic phrase is accentually indistinguishable from an unaccented unimoraic phrase, and the accent in this particular case is only realized before a particle or auxiliary.
  • Medial-/Internal-accented (中高型, nakadakagata; lit. 'high-middle-type'): The nucleus is neither the first nor last mora. Thus, a medial-accented phrase must necessarily be trimoraic or longer.
  • Final-accented (尾高型, odakagata; lit. 'high-tail-type'): The nucleus is the last mora, but not simultaneously the first mora in a unimoraic phrase which, if accented, would be initial-accented instead. Thus, a final-accented phrase must necessarily be bimoraic or longer; it is accentually indistinguishable from an unaccented phrase, and the accent is only realized before a particle or auxiliary.

The actual pitch pattern of a final-accented phrase is generally considered indistinguishable from that of an unaccented phrase. The "final" accent of a word only surfaces as "medial" accent when it is followed by a particle or auxiliary, which is also the only time when a final-accented word is indeed distinct from an unaccented one. However, there are indeed some speakers who do make distinctions between no accent and final accent, for example by raising the overall pitch contour on the final-accented hana (花; /haná/ 'a flower') higher than on the unaccented hana (鼻; /hana/ 'a nose'). In one experiment, one college-educated female Tokyo speaker was able to not only produce a consistent, better-than-chance distinction between no accent and final accent, but to also perceive it consistently when she was listening to her own recordings, something none of the other subjects could do.

The phonetic accent contour

The actual pitch pattern is not simply a fluctuation between some "high" and "low" tones, but is a phonetically complex contour. In an accented phrase, the basic contour:

1. Starts from a relatively low pitch on the first mora;
2. Rises sharply to a relatively high pitch on the second mora;
3. Continues to rise higher to the highest pitch on the nuclear mora; the "high" on the nucleus is actually higher than the "high" on the second mora;
4. Falls sharply to a very low pitch on the postnuclear mora;
5. Declines further to the last mora.

In contrast, in an unaccented phrase, the basic contour:

1. Starts from a relatively low pitch on the first mora;
2. Rises sharply to a relatively high pitch on the second mora;
3. Declines gradually towards a relatively low pitch on the last mora.

Even in an unaccented phrase, the contour declines towards the end of the phrase; and in an accented phrase, the "low" pitch on the postnuclear mora continues to decline as well. Both these cases exemplify a phonetic feature known as "declination."

Initial lowering

Henceforth, the complex pitch contour can be simplified into a dichotomy of "highs" and "lows" as follows (the corner indicates an accentual fall):

• Accented: μμμμμμ
• Unaccented: μμμμμμ

In both cases, that the first mora has a low tone and the second mora a high tone is referred to as "initial lowering" or "initial dissimilation." This is a phonetic feature whose possible function is to demarcate the beginning of a phrase in Tokyo Japanese. It is a predictable property of a phrase, not of a word. Consider the following:

• pɯɾiNtaR ɡa ← purintā ga (プリンターが; 'printer')
• ano pɯɾiNtaR ɡa ← ano purintā ga (あのプリンターが; 'that printer')
• tomodacɕi̥ kaɾa kaɾita ano pɯɾiNtaR ɡa ← tomodachi kara karita ano purintā ga (友達から借りたあのプリンターが; 'that printer which I borrowed from my friend')
• kinoR tomodacɕi̥ kaɾa kaɾita ano pɯɾiNtaR ɡa ← kinō tomodachi kara karita ano purintā ga (昨日友達から借りたあのプリンターが; 'that printer which I borrowed from my friend yesterday')

In all these cases, initial lowering is never applied to any particular word, but to the entire phrase. Other dialects may have "final lowering" (such as in the Hirosaki dialect) or "initial raising" (such as in the Bessho dialect in Yurihama, Tottori).

The two aforementioned basic contours occur during the spans of several morae, which necessitates a minimum mora count, but there are not always enough morae for the contours to fully realize. A unimoraic phrase only carries a high pitch of a would-be second mora, whether that phrase is supposed to be accented or not, making the initial-accented ki (木; /kí/; 'a tree') and the unaccented ki (気; /ki/; 'spirits') sound identical without a following particle or auxiliary. Similarly, a bimoraic phrase whose first mora is unaccented has room for initial lowering, but the second mora has the same high pitch whether it is accented or not, making the final-accented hana (花; /haná/; 'a flower') and the unaccented hana (鼻; /hana/; 'a nose') sound identical without a following particle or auxiliary. An initial-accented bimoraic or longer phrase does not have room for initial lowering, and the pitch must immediately peak on the first mora, which still begins with a brief, albeit heavily compressed, initial rise from low to high. On the other hand, a second-accented trimoraic or longer phrase does not have room for the slight incline from the high on the second mora to the highest high on the nucleus, and the pitch must peak immediately on the second mora. Thus:

• Unimoraic, accented or not: μ
• Bimoraic or longer, initial-accented: μμμμμμ
• Bimoraic, accented or not: μμ
• Trimoraic or longer, second-accented: μμμμμμ

Conventionally, accent dictionaries still make clear distinctions in notation between ki (木; /kí/), ki (気; /ki/), hana (花; /haná/) and hana (鼻; /hana/):

• Accented: キ, ハナ
• Unaccented: キ, ハナ

Initial lowering may be optionally suppressed, which means that the unaccented first mora simply carries a high tone and flatly transitions into the second mora, if the second mora is /R/ (the lengthening mora) or /N/ (the moraic nasal); in these cases, initial lowering has been said to occur only in "slow, careful, unnatural" enunciation. If the second mora is /Q/ (the moraic obstruent), the initial rise only happens from the second to third mora. Thus:

• The second mora is lengthening: μRμμμμ ~ μRμμμμ
• The second mora is nasal: μNμμμμ ~ μNμμμμ
• The second mora is obstruent: μQμμμμ

For the sake of consistency, textbooks and pronunciation guides ignore all these variations and mark all notations with initial lowering:

• The second mora is lengthening: μRμμμμ
• The second mora is nasal: μNμμμμ
• The second mora is obstruent: μQμμμμ

Minor and major phrases

Japanese accent is not realized on a word level, but on a phrase level. Even words when pronounced in isolation form their own phrases.

A phrase on which one of the aforementioned accent contours is realized is called an "accent(ual) phrase," "minor (phonological) phrase," or "accent unit" (アクセント単位, akusento tan'i). A minor phrase can have maximally one nucleus, and begins with initial lowering. In this article, minor phrases are demarcated with a single vertical bar /|/ (the IPA symbol for a minor break): μμμμμμ | μμμμμμ | μμμμμμ | μμμμμμ.

Multiple minor phrases form together an "intermediate phrase," "intonation phrase," or "major (phonological) phrase." In this article, major phrases are demarcated with a double vertical bar /‖/ (the IPA symbol for a major break): μμμμμμ | μμμμμμ ‖ μμμμμμ | μμμμμμ.

Downstep

When an accented minor phrase follows another accented minor phrase, the pitch contour of the second is significantly lower. This is a phonetic feature known as "downstep," "downdrift" or "catathesis." Downstep is said to produce a more natural-sounding contour, making downstepped highs successively lower after each nucleus, rather than keeping all highs equally high as one would when deliberately and slowly enunciating accented minor phrases.

The domain of downstep is the major phrase, in other words, across minor phrases rather than within them.

Combining minor phrases

Shorter minor phrases can combine to make longer minor or major phrases. This is typical, and separating minor phrases without special emphasis would make one sound artificial like a first-grade schoolchild reciting in class. Accented and unaccented minor phrases combine differently; when two minor phrases combine:

If the first minor phrase is unaccented, it can merge with the second minor phrase to form an even longer minor phrase. The second minor phrase no longer exists separately, and is thus "dephrased":

• Unaccented + unaccented: μμμμμ | μμμμμ → μμμμμ μμμμμ
• Unaccented + accented: μμμμμ | μμμμμ → μμμμμ μμμμμ

If the first minor phrase is accented, the two minor phrases can remain separate, with the start of the second minor phrase marked with initial lowering, although it is not impossible for the second phrase to be dephrased. If the second minor phrase is also accented, it undergoes downstep:

• Accented + unaccented: μμμμμ | μμμμμ ~ μμμμμ μμμμμ
• Accented + accented: μμμμμ | μμμμμ ~ μμμμμ μμμμμ

In the above cases, if the first minor phrase is final-accented and the second minor phrase is an initial-accented particle or particle sequence, the latter is highly likely dephrased:

• Final-accented + initial-accented particle(s): μμμμμ | μμμ → μμμμμ μμμ

If a minor phrase is "focused" to convey particularly important information, its highest pitch is raised even higher. The focused unaccented first minor phrase stays separate instead of merging with the second minor phrase:

• Unaccented (focused) + unaccented: μμμμμ | μμμμμ
• Unaccented (focused) + accented: μμμμμ | μμμμμ

Focusing on the second minor phrase resets the pitch range and starts an entirely new major phrase, which means downstep does not apply:

• Unaccented + unaccented (focused): μμμμμ ‖ μμμμμ
• Unaccented + accented (focused): μμμμμ ‖ μμμμμ
• Accented + unaccented (focused): μμμμμ ‖ μμμμμ
• Accented + accented (focused): μμμμμ ‖ μμμμμ

Different major phrasings yield different meanings, even if all the words are exactly the same:

• umai sobaya ga aru (うまいそば屋がある)
  • ɯmai ‖ sobaja ɡa | aɾɯ "there's a good SOBA diner [not a curry diner]"
  • ɯmai | sobaja ɡa ‖ aɾɯ "there IS a good soba diner"
• kinō purintā ga kowareta (昨日プリンターが壊れた)
  • kinoR pɯɾiNtaR ɡa kowaɾeta "the printer broke yesterday"
  • kinoR ‖ pɯɾiNtaR ɡa kowaɾeta "what happened yesterday was the PRINTER broke [so I couldn't print my report]"
• ototoi karita purintā ga kowareta (一昨日借りたプリンターが壊れた)
  • ototoi kaɾita pɯɾiNtaR ɡa ‖ kowaɾeta "the printer, which I borrowed the day before yesterday, broke"
  • ototoi ‖ kaɾita pɯɾiNtaR ɡa kowaɾeta "the printer, which I borrowed, broke the day before yesterday"
• kesa katta bakari no kasa o nakushita (けさ買ったばかりの傘をなくした)
  • kesa ‖ kaQta bakari no | kasa o nakɯɕi̥ta "this morning, I lost the umbrella I'd just bought [I lost it this morning]"
  • kesa | kaQta bakari no | kasa o nakɯɕi̥ta "I lost the umbrella I'd just bought this morning [I bought it this morning]"
• Natsume Sōseki no Botchan ni deaimashita (夏目漱石のぼっちゃんに出会いました)
  • natsɯme soRseki no ‖ boQcɕaN ni | deaimaɕi̥ta "I happened on Botchan by Natsume Sōseki [Botchan means his novel]"
  • natsɯme soRseki no | boQcɕaN ni | deaimaɕi̥ta "I happened on Natsume Sōseki's son [botchan means his son]"

Special and deficient morae

Most morae are known as "autonomous/independent morae" (自立拍・自立モーラ, jiritsuhaku/jiritsu mōra) and they are capable of bearing an accent and being a nucleus. The morae that are not are known as "dependent morae" (非自立モーラ, hijiritsu mōra), "special morae" (特殊拍・特殊モーラ, tokushuhaku/tokushu mōra), "special phonemes" (特殊音素, tokushu onso), "special segments," "mora(ic) phonemes" (モーラ音素, mōra onso) or "moraic segments." There are four such morae:

• The mora(ic) nasal (撥音/撥ねる音, hatsuon/haneru on), spelt ん or ン in kana and notated phonologically /N/.
• The mora(ic) obstruent (促音/詰まる音, sokuon/tsumaru on), spelt っ or ッ in kana and notated phonologically /Q/.
• The second half of a long vowel, henceforth called the lengthening mora (引（き）音/長音, hikion~in'on/chōon), spelt あ・ア・い・イ・う・ウ・え・エ・お・オ・ー in kana and notated phonologically /R/ or /H/.
• The second half of a "diphthong," henceforth called the offglide mora, spelt い・イ・う・ウ and notated phonologically, individually /i, ɯ/, or collectively /J/. Note that if /i, ɯ/ do not directly follow a vowel, or are the second half in /ii → iR, ɯɯ → ɯR/, they are not offglide morae. The "diphthongs" of Japanese include principally /ei, ai, oi, ɯi/ and less frequently /aɯ/ (/iɯ, eɯ, oɯ/ are probably not diphthongs). /ei/ is alternatively pronounced /eR/ with a lengthening mora, except in ei (鱏; /éi/ 'a ray'), rei (レイ; /ɾéi/ 'a lei'), etc; and purportedly except across morpheme boundaries, perhaps in careful pronunciation, as in me-isha (眼医者; /mé-iɕa/ 'an oculist'; compare mei-sha (名車; /mei-ɕa ~ meR-ɕa/ 'a famous car')), yake-ishi (焼け石; /jake-iɕi/ 'a hot stone'), etc.

Proponents of the Japanese "syllable" consider it a sequence of a autonomous mora optionally followed by a special mora. Examples of Japanese "syllables" would be /a, aN, aQ, aR, ai, aɯ/. If a Tokyo syllable is to be accented, only its first mora is the actual nucleus, hence /a, aN, aQ, aR, ai, aɯ/. Henceforth, the nucleus is marked with an acute accent for simplicity, as in /á, áN, áQ, áR, ái, áɯ/. The syllable is accentually more salient in other dialects, such as Kagoshima Japanese, and if a Kagoshima syllable is to be accented, both its morae carry high tones.

The class of special morae are not the only ones that cannot be nuclei. In certain contexts, even "autonomous" morae can behave similarly in failing to be nuclei as well. These morae can form a larger class known as "deficient morae," which include morae with devoiced and/or epenthetic vowels.

Accent shifts

Since deficient morae cannot bear accent, if an accent rule would put an accent on them, an "accent shift" is necessitated. Labrune (2012) counts all instances where an accent is not in the exact location generalized by a rule as "accent shifts;" Vance (2008) only counts instances involving devoiced and/or epenthetic vowels, and treat those involving special morae as "accented syllables" rather than "accented morae."

Consider the suffix ‑shi (市; 'a city'). The accent rule for it is that when it follows a city's name, it places an accent on that name's last mora:

• Ōsakashi (大阪県; 'Osaka City'): /oRsaká-ɕi/

If that last mora is special, a leftward shift is necessitated; or, the above rule can be restated as "it places an accent on that name's last syllable":

• Leftward shift away from a moraic nasal: Muroranshi (室蘭市; 'Muroran City'): /*mɯɾoɾaŃ-ɕi → mɯɾoɾáN-ɕi/
• Leftward shift away from a lengthening mora: Kitakyūshūshi (北九州市; 'Kitakyushu City'): /*kitakjɯRɕɯŔ-ɕi → kitakjɯRɕɯ́R-ɕi/
• Leftward shift away from an offglide mora: Sendaishi (仙台市; 'Sendai City'): /*seNdaí-ɕi → seNdái-ɕi/

Exceptions to these leftward shifts involve two consecutive special morae, in which case the accent usually stays on the first special mora and does not shift any further leftward. Compare the suffixes ‑bashi (橋), ‑jin (人) and ‑kko (っ子), the last of which is capable of causing two consecutive special morae:

• Rondonbashi (ロンドン橋; 'London Bridge'): /*ɾoNdoŃ-baɕi → ɾoNdóN-baɕi/
• gendaijin (現代人; 'a modern person'): /*ɡeNdaíʑiN → ɡeNdái-ʑiN/
• Rondonkko (ロンドンっ子; 'a Londoner'): /ɾoNdoŃ-Qko ^?(→ ɾoNdóN-Qko)/
• gendaikko (現代っ子; 'one of the kids these days'): /ɡeNdaí-Qko ^?(→ ɡeNdái-Qko)/

Loanwords may contain consecutive special morae to begin with, which blocks further shifts, or allows options with no shift (here not marked with asterisks):

• Rinkānhai (リンカーン杯; 'Lincoln Cup'): /ɾiNkaRŃ-hai → ɾiNkaŔN-hai ^?(→ ɾiNkáRN-hai)/

The seemingly offglide mora /i/ in the sequences /iN/ and /iQ/ in loanwords can be a nucleus, compared to the actual offglide mora /i/ in the above examples. In the following examples, an accent is triggered by a suffix, or placed by default on the antepenultimate mora (the "antepenultimate rule"):

• Raingawa (ライン川; 'the Rhine'): /*ɾaiŃ-ɡawa → ɾaíN-ɡawa *(→ ɾáiN-ɡawa)/
• Supeinkaze (スペイン風邪; 'Spanish flu'): /sɯ̥peiŃ-kaze → sɯ̥peíN-kaze *(→ sɯ̥péiN-kaze)/
• koinshō (コイン商; 'a coin dealer'): /koiŃ-ɕoR → koíN-ɕoR *(→ kóiN-ɕoR)/
• hoissuru (ホイッスル; 'a whistle'): /*hoiQ́sɯɾɯ → hoíQsɯɾɯ *(→ hóiQsɯɾɯ)/
• sandoitchi (サンドイッチ; 'a sandwich'): /saNdoíQcɕi *(→ saNdóiQcɕi)/
• kuikku (クイック; 'quick'): /kɯíQkɯ *(→ kɯ́iQkɯ)/

Accent dictionaries also claim a leftward shift is possible if the last mora is devoiced, but this has been disputed as incorrect in the current paradigm.

• Leftward shift away from a devoiced vowel: Nagasakishi (長崎市; 'Nagasaki City'): /naɡasakí̥-ɕi → ^?naɡasáki̥-ɕi/

A leftward shift away from a devoiced vowel, if indeed possible, is said to only happen if there are three or more morae before the devoiced vowel. And unlike shifts away from special morae which are obligatory, shifts away from devoiced vowels are optional at best and increasingly infrequent. The current tendency is to keep the accent where it is supposed to be on the devoiced vowel.

If there are less than three morae before an accented devoiced vowel, a rightward shift may occur instead.

• Rightward shift away from a devoiced vowel:
  • kite (来て; 'comes and'): /kí̥te → ki̥té/
  • futte (降って; 'descends and'): /ɸɯ̥́Qte → ɸɯ̥Qté/
  • Ashikaga (足利): /*aɕí̥kaɡa → aɕi̥káɡa/
  • Katsushika Hokusai (葛飾北斎): /katsɯ́ɕi̥ka | *hokɯ̥́sai → katsɯ́ɕi̥ka | hokɯ̥sái/

Amusuterudamu (アムステルダム; 'Amsterdam') is an example of a shift away from an epenthetic vowel. In rendering the Dutch name, pronounced /ˌɑmstərˈdɑm/, the Japanese vowel /ɯ/ as in /ɾɯ/ is inserted to transcribe the Dutch consonant /r/, hence epenthesis. The Japanese name ought to follow the antepenultimate rule, and yet there seems to be a leftward shift.

• Leftward shift away from an epenthetic vowel: Amusuterudamu (アムステルダム): /*amɯsɯ̥teɾɯ́damɯ → amɯsɯ̥téɾɯdamɯ/

Consecutive devoiceable vowels and accent

When there are two consecutive devoiceable vowels, the regularly accented one has been said to stay voiced, although it is common nowadays for both of them to be devoiced. Examples with contradictory claims would be adjectival gerunds ending in ‑shikute; compare:

• /i/ is accented and voiced, /ɯ/ is devoiced: /kanaɕíkɯ̥te/
• /i/ devoiced, /ɯ/ is voiced, the accent shifts leftward: /kanáɕi̥kɯte/
• Both /i/ and /ɯ/ are devoiced, /i/ is accented: /kanaɕí̥kɯ̥te/
• Both /i/ and /ɯ/ are devoiced, the accent shifts leftward: /kanáɕi̥kɯ̥te/

Pitch raising

The defining characteristic of accent in Japanese is a steep fall from a high tone to a low tone. This means that an accented mora is necessarily voiced for it to have any tone at all, much less specifically a high one. Yet, it is frequently the case that devoiced vowels, particularly the high vowels /i, ɯ/, are accented. The transition from an illusory, voiceless "high" to the following low is significantly different from a true high, in that the pitch of the following low is higher in the former case. This phenomenon has been termed pitch raising. Experiments have shown that an accented yet voiceless mora raises the pitch of the following low tone significantly. The degree of raising depends on how close the voiceless "high" is to the low, which is raised the most when it comes right after the "high", for example in the test nonsense phrase /ɸɯ̥ɕi̥tsɯ̥́ to/. In this case, the low tone of /to/ is raised because of the voiceless "high" of /tsɯ̥́/. If the "high" and the low are separated by voiceless "lows", there is less pitch raising as the accented "high" becomes farther away from the low, as in /ɸɯ̥ɕí̥tsɯ̥ to/ and /ɸɯ̥́ɕi̥tsɯ̥ to/. Pitch raising is suppressed if there is a voiced high before the devoiced region, as in /ɸɯ̥ɕítsɯ̥ to/. Pitch raising has been attributed to an overlap of activity between the cricothyroid and sternohyoid muscles, the former of which is said to regulate pitch but to still be active during devoicing, while the latter is said to cause pitch falls, in words where the accented mora is voiceless. Such overlap is not present in words where the accented mora is voiced.

Accent rules

Main article: Japanese pitch accent rules

Examples of words that differ only in pitch

In standard Japanese, about 47% of words are unaccented and around 26% are accented on the ante-penultimate mora. However, this distribution is highly variable between word categories. For example, 70% of native nouns are unaccented, while only 50% of kango and only 7% of loanwords are unaccented. In general, most 1–2 mora words are accented on the first mora, 3–4 mora words are unaccented, and words of greater length are almost always accented on one of the last five morae.

The following chart gives some examples of minimal pairs of Japanese words whose only differentiating feature is pitch accent. An accentual fall is indicated with the IPA symbol for "downstep," [ꜜ], a notation used by Vance (2008), but it is not to be confused with actual downstep or catathesis, which is not a fall from a high to low tone, but the lowering of high tones only.

Romanization	Accent on first mora			Accent on second mora			Accentless
hashi はし	/haꜜsi/ [háꜜɕì] háshì	箸	chopsticks	/hasiꜜ/ [hàɕíꜜ] hàshí	橋	bridge	/hasi/ [hàɕí] hàshí	端	edge
hashi-ni はしに	/haꜜsini/ [háꜜɕìɲì] háshì-nì	箸に	at the chopsticks	/hasiꜜni/ [hàɕíꜜɲì] hàshí-nì	橋に	at the bridge	/hasini/ [hàɕīɲī] hàshi-ni	端に	at the edge
ima いま	/iꜜma/ [íꜜmà] ímà	今	now	/imaꜜ/ [ìmáꜜ] ìmá	居間	living room
kaki かき	/kaꜜki/ [káꜜkì] kákì	牡蠣	oyster	/kakiꜜ/ [kàkíꜜ] kàkí	垣	fence	/kaki/ [kàkí] kàkí	柿	persimmon
kaki-ni かきに	/kaꜜkini/ [káꜜkìɲì] kákì-nì	牡蠣に	at the oyster	/kakiꜜni/ [kàkíꜜɲì] kàkí-nì	垣に	at the fence	/kakini/ [kàkīɲī] kàki-ni	柿に	at the persimmon
sake さけ	/saꜜke/ [sáꜜkè] sákè	鮭	salmon				/sake/ [sàké] sàké	酒	alcohol, sake
nihon にほん	/niꜜhoɴ/ [ɲíꜜhòɴ̀] níhòn	二本	two sticks of	/nihoꜜɴ/ [ɲìhóɴ̀] nìhón	日本	Japan

In isolation, the words hashi はし /hasiꜜ/ hàshí "bridge" and hashi /hasi/ hàshí "edge" are pronounced identically, starting low and rising to a high pitch. However, the difference becomes clear in context. With the simple addition of the particle ni "at", for example, /hasiꜜni/ hàshí-nì "at the bridge" acquires a marked drop in pitch, while /hasini/ hàshi-ni "at the edge" does not. However, because the pitch drop occurs after the first mora of the accented syllable, a word with a final long accented syllable would contrast all three patterns even in isolation: an accentless word nihon, for example, would be pronounced [ɲìhōɴ̄], differently from either of the words above. In 2014, a study recording the electrical activity of the brain showed that native Japanese speakers mainly use context, rather than pitch accent information, to contrast between words that differ only in pitch.

This property of the Japanese language allows for a certain type of pun, called dajare (駄洒落, だじゃれ), combining two words with the same or similar sounds but different pitch accents and thus meanings. For example, kaeru-ga kaeru /kaeruɡa kaꜜeru/ (蛙が帰る; lit. the frog will go home). These are considered quite corny, and are associated with oyaji gags (親父ギャグ, oyaji gyagu; dad joke).

Since any syllable, or none, may be accented, Tokyo-type dialects have N+1 possibilities, where N is the number of syllables (not morae) in a word, though this pattern only holds for a relatively small N.

The accent system of Tokyo dialect

accented syllable	one-syllable word	two-syllable word	three-syllable word
0 (no accent)	/ki/ (気, mind)	/kaze/ (風, wind)	/tomeru/ (止める, to stop)
1	/kiꜜ/ (木, tree)	/haꜜru/ (春, spring)	/iꜜnoti/ (命, life)
2	—	/kawaꜜ/ (川, river)	/tamaꜜɡo/ (卵, egg)
3	—		/kotobaꜜ/ (言葉, word)

Other dialects

Pitch-accent systems of Japanese. Blues: Tokyo type. Yellow-orange: Kyoto–Osaka (Keihan) type. Pink: Two-pattern accent. White: No accent. Speckled areas are ambiguous.

Accent and tone are the most variable aspect of Japanese dialects. Some have no accent at all; of those that do, it may occur in addition to a high or low word tone.

The dialects that have a Tokyo-type accent, like the standard Tokyo dialect described above, are distributed over Hokkaido, northern Tohoku, most of Kanto, most of Chūbu, Chūgoku and northeastern Kyushu. Most of these dialects have a more-or-less high tone in unaccented words (though first mora has low tone, and following morae have high tone); an accent takes the form of a pitch drop, after which the tone stays low. But some dialects, for example, dialects of northern Tohoku and eastern Tottori, typically have a more-or-less low tone in unaccented words; accented syllables have a high tone, with low tone on either side, rather like English stress accent. In any case, the pitch drop has phonological meaning and the syllable followed by pitch drop is said to be "accented".

Keihan (Kyoto–Osaka)-type dialects of Kansai and Shikoku have nouns with both patterns: That is, they have tone differences in unaccented as well as accented words, and both accentual falls in some high-tone words and a high-tone accent in some low-tone words. In the neighboring areas of Tokyo-type and Keihan-type such as parts of Kyushu, northeastern Kanto, southern Tohoku, around Fukui, around Ōzu in Ehime and elsewhere, nouns are not accented at all.

Kyushu (two-pattern type)

In western and southern Kyushu dialects (pink area on the map on the right), a high tone falls on a predictable syllable, depending only on whether the noun has an accent. This is termed a two-pattern (nikei) system, as there are two possibilities, accented and not accented. For instance, in the Kagoshima dialect unaccented nouns have a low tone until the final syllable, at which point the pitch rises. In accented nouns, however, the penultimate syllable of a phonological word has a high tone, which drops on the final syllable. (Kagoshima phonology is based on syllables, not on morae.) For example, irogami 'colored paper' is unaccented in Kagoshima, while kagaribi 'bonfire' is accented. The ultimate or penultimate high tone will shift when any unaccented grammatical particle is added, such as nominative -ga or ablative -kara:

[iɾoɡamí], [iɾoɡamiɡá], [iɾoɡamikaɾá]

[kaɡaɾíbi], [kaɡaɾibíɡa], [kaɡaɾibikáɾa]

In the Shuri dialect of the Okinawan language, unaccented words are high tone; accent takes the form of a pitch drop after the second syllable, or after the first syllable of a disyllabic noun. However, the accents patterns of the Ryukyuan languages are varied, and do not all fit the Japanese patterns.

Nikei accents are also found in parts of Fukui and Kaga in Hokuriku region (green area on map).

No accent versus one-pattern type

In Miyakonojō, Miyazaki Prefecture (the small black area on map), there is a single accent: all phonological words have a low tone until the final syllable, at which point the pitch rises. That is, every word has the pitch pattern of Kagoshima irogami. This is called an ikkei (one-pattern) accent. Phonologically, it is the same as the absence of an accent (white areas on map), and is sometimes counted as such, as there can be no contrast between words based on accent. However, speakers of ikkei-type dialects feel that they are accenting a particular syllable, whereas speakers of unaccented dialects have no such intuition.

Keihan (Kyoto–Osaka type)

See also: Kansai dialect § Pitch accent

Near the old capital of Kyoto, in Kansai, Shikoku, and parts of Hokuriku (the easternmost Western Japanese dialects), there is a more innovative system, structurally similar to a combination of these patterns. There are both high and low initial tone as well as the possibility of an accented mora. That is, unaccented nouns may have either a high or a low tone, and accented words have pitch accent in addition to this word tone. This system will be illustrated with the Kansai dialect of Osaka.

The tone-accent system of Osaka dialect

	accented mora	one mora	two-mora word	three-mora word	gloss
high tone	(no accent)	/ki/ [kíí]	/kiɡa/ [kíɡá]	/kikara/ [kíkáɾá]	'mind' (気)
		—	/kaze/ [kázé]	/kazeɡa/ [kázéɡá]	'wind' (風)
		—		/jameru/ [jáméɾɯ́]	'stop' (止める)
	1	/hiꜜ/ [çíì]	/hiꜜɡa/ [çíɡà]	/hiꜜkara/ [çíkàɾà]	'day' (日)
		—	/kaꜜwa/ [káwà]	/kaꜜwaɡa/ [káwàɡà]	'river' (川)
		—		/siꜜroi/ [ɕíɾòì]	'be white' (白い)
	2	—	(none)	/ataꜜma/ [átámà]	'head' (頭)
	3	—		(few words, if any)
low tone	(no accent)	/˩ki/ [kìí]	/˩kiɡa/ [kìɡá]	/˩kikara/ [kìkàɾá]	'tree' (木)
		—	/˩ito/ [ìtó]	/˩itoɡa/ [ìtòɡá]	'thread' (糸)
		—		/˩okiru/ [òkìɾɯ́]	'to get up' (起きる)
	2	—	/˩haruꜜ/ [hàɾɯ́ ~ hàɾɯ̂]	/˩haruꜜɡa/ [hàrɯ́ɡà]	'spring' (春)
		—		/˩kusuꜜri/ [kɯ̀sɯ́ɾì]	'medicine' (薬)
	3	—		/˩maQtiꜜ/ [màttɕí ~ màttɕî]	'match' (マッチ)
Low tone is considered to be marked (transcribed /˩/). Not all patterns are found: In high-tone words, accent rarely falls on the last mora, and in low-tone words it cannot fall on the first. One-mora words are pronounced with long vowels.

Accented high-tone words in Osaka, like atama 'head', are structurally similar to accented words in Tokyo, except that the pitch is uniformly high prior to the pitch drop, rather than rising as in Tokyo. As in Tokyo, the subsequent morae have low pitch. Unaccented high-tone words, such as sakura 'cherry tree', are pronounced with a high tone on every syllable, and in following unaccented particles:

High tone /ataꜜma/, accent on ta: [átámà], [átámàɡà], [átámàkàɾà]

High tone /sakura/, no accent: [sákɯ́ɾá], [sákɯ́ɾáɡá], [sákɯ́ɾákáɾá]

Low-tone accented words are pronounced with a low pitch on every mora but the accented one. They are like accented words in Kagoshima, except that again there are many exceptions to the default placement of the accent. For example, tokage is accented on the ka in both Osaka and Kagoshima, but omonaga 'oval face' is accented on mo in Osaka and na in Kagoshima (the default position for both dialects); also, in Osaka the accented is fixed on the mo, whereas in Kagoshima it shifts when particles are added. Unaccented low-tone words such as usagi 'rabbit' have high pitch only in the final mora, just as in Kagoshima:

Low tone /˩omoꜜnaɡa/, accent on mo: [òmónàɡà], [òmónàɡàɡà], [òmónàɡàkàɾà]

Low tone /˩usaɡi/, no accent: [ɯ̀sàɡí], [ɯ̀sàɡìɡá], [ɯ̀sàɡìkàɾá]

Hokuriku dialect in Suzu is similar, but unaccented low-tone words are purely low, without the rise at the end:

/˩usaɡi/: [ɯ̀sàŋì], [ɯ̀sàŋìŋà], [ɯ̀sàŋìkàɾà];

sakura has the same pattern as in Osaka.

In Kōchi, low-tone words have low pitch only on the first mora, and subsequent morae are high:

/˩usaɡi/: [ɯ̀sáɡí], [ɯ̀sáɡíɡá], [ɯ̀sáɡíkáɾá].

The Keihan system is sometimes described as having 2n+1 possibilities, where n is the number of morae (up to a relatively small number), though not all of these actually occur. From the above table, there are three accent patterns for one-mora words, four (out of a theoretical 2n+1 = 5) for two-mora words, and six (out of a theoretical 2n+1 = 7) for three-mora words.

Correspondences between dialects

There are regular correspondences between Tokyo-type and Keihan-type accents. The pitch drop on high-tone words in conservative Keihan accents generally occurs one syllable earlier than in the older Tokyo-type accent. For example, kokoro 'heart' is /kokoꜜro/ in Tokyo but /koꜜkoro/ in Osaka; kotoba 'word' is /kotobaꜜ/ in Tokyo but /kotoꜜba/ in Osaka; kawa 'river' is /kawaꜜ/ in Tokyo but /kaꜜwa/ in Osaka. If a word is unaccented and high-tone in Keihan dialects, it is also unaccented in Tokyo-type dialects. If a two-mora word has a low tone in Keihan dialects, it has a pitch drop on the first mora in Tokyo-type dialects.

In Tokyo, all nakadaka verbs have the pitch drop after the second to last mora, except in cases of deficient mora. This is the last mora of the verb stem, as in /shiroꜜi/ 'be white' and /okiꜜru/ 'get up'. In Kansai, verbs have high- and low-tone paradigms as nouns do. High-tone verbs are either unaccented or are accented on the penultimate mora. Low-tone verbs are either unaccented or accented on the final syllable, triggering a low tone on unaccented suffixes. In Kyoto, verbal tone varies irregularly with inflection, a situation not found in more conservative dialects, even more conservative Kansai-type dialects such as that of Kōchi in Shikoku.

Syllabic and moraic

Japanese pitch accent also varies in how it interacts with syllables and morae. Kagoshima is a purely syllabic dialect, while Osaka is moraic. For example, the low-tone unaccented noun shinbun 'newspaper' is [ɕìm̀bɯ́ɴ́] in Kagoshima, with the high tone spread across the entire final syllable bun, but in Osaka it is [ɕìm̀bɯ̀ɴ́], with the high tone restricted to the final mora n. In Tokyo, accent placement is constrained by the syllable, though the pitch drop occurs between the morae of that syllable. That is, a stressed syllable in Tokyo dialect, as in 貝 kai 'shell' or 算 san 'divining rod', will always have the pattern /kaꜜi/ [káì], /saꜜɴ/ [sáɴ̀], never */kaiꜜ/, */saɴꜜ/. In Osaka, however, either pattern may occur: tonbi 'black kite' is [tóm̀bì] in Tokyo but [tòḿbì] in Osaka.

Bibliography

• Akamatsu, Tsutomu (1997). Japanese Phonetics: Theory and Practice. Munich: Lincom Europa. ISBN 3-89586-095-6.
• Akinaga, Kazue (20 December 1998) [25 April 1998]. 共通語のアクセント. In NHK Broadcasting Culture Research Institute (ed.). NHK日本語発音アクセント辞典 (in Japanese) (5th ed.). Shibuya, Tokyo: NHK Publishing. pp. 174–221. ISBN 4-14-011112-7.
• Bloch, Bernard (1950). "Studies in Colloquial Japanese IV: Phonemics". Language. 26 (1). Linguistic Society of America: 86–125. doi:10.2307/410409. JSTOR 410409.
• Haraguchi, Shosuke (September 1975). The Tone Pattern of Japanese: An Autosegmental Theory of Tonology (doctoral thesis). Cambridge, Massachusetts: Massachusetts Institute of Technology.
• Haraguchi, Shosuke (2002) [1999]. "1. Accent". In Tsujimura, Natsuko (ed.). The Handbook of Japanese Linguistics. Malden, Massachusetts; Oxford: Blackwell Publishers. pp. 1–30. ISBN 0-631-23494-2.
• Hirayama, Teruo, ed. (25 December 1961) [30 June 1960]. 全国アクセント辞典 (in Japanese) (3rd ed.). Chiyoda, Tokyo: Tokyodo.
• Kindaichi, Haruhiko; Akinaga, Kazue, eds. (1 April 2025) [10 April 2014]. 新明解日本語アクセント辞典 (in Japanese) (2nd ed.). Chiyoda, Tokyo: Sanseido. ISBN 978-4-385-13468-0.
• Kōri, Shirō (1 September 2024) [1 May 2020]. 日(に)本(ほん)語(ご)のイントネーション――しくみと音(おん)読(どく)・朗(ろう)読(どく)への応(おう)用(よう) (in Japanese) (2nd ed.). Bunkyō, Tokyo: Taishukan. ISBN 978-4-469-21379-9.
• Kubozono, Haruo (2002) [1999]. "2. Mora and Syllable". In Tsujimura, Natsuko (ed.). The Handbook of Japanese Linguistics. Malden, Massachusetts; Oxford: Blackwell Publishers. pp. 31–61. ISBN 0-631-23494-2.
• Kubozono, Haruo (2015). "Japanese Dialects and General Linguistics". Gengo Kenkyu 言語研究. 148. The Linguistic Society of Japan: 1–31. doi:10.11435/gengo.148.0_1. ISSN 2185-6710.
• Kubozono, Haruo (2023). Lahiri, Aditi (ed.). Word and Sentence Prosody: The Endangered Dialect of Koshikijima Japanese (PDF). Phonology and Phonetics. Vol. 31. Berlin; Boston: De Gruyter Mouton. ISBN 978-3-11-073505-5.
• Labrune, Laurence (2012). The Phonology of Japanese. Oxford; New York: Oxford University Press. ISBN 978-0-19-9545834.
• Martin, Samuel Elmo (2004) [1975]. A Reference Grammar of Japanese (revised ed.). Honolulu: University of Hawai'i Press. ISBN 0-8248-2818-6.
• Matsumori, Akiko; Nitta, Tetsuo; Kibe, Nobuko; Nakai, Yukihiko (1 May 2026) [20 September 2012]. An Introduction to Japanese Accent 日本語アクセント入門 (in Japanese) (3rd ed.). Chiyoda, Tokyo: Sanseido. ISBN 978-4-385-36531-2.
• McCawley, James David (1968). The Phonological Component of a Grammar of Japanese. The Hague: Mouton. OCLC 438994.
• NHK Broadcasting Culture Research Institute, ed. (20 December 1998). NHK日本語発音アクセント辞典 (in Japanese) (5th ed.). Shibuya, Tokyo: NHK Publishing. ISBN 4-14-011112-7.
• NHK Broadcasting Culture Research Institute, ed. (20 June 2025) [25 May 2016]. NHK日本語発音アクセント新辞典 (in Japanese) (15th ed.). Shibuya, Tokyo: NHK Publishing. ISBN 978-4-14-011345-5.
• Pierrehumbert, Janet; Beckman, Mary Esther (1988). Japanese Tone Structure. Cambridge, Massachusetts; London, England: The MIT Press. ISBN 0-262-16109-5.
• Poser, William John (1984). The Phonetics and Phonology of Tone and Intonation in Japanese (doctoral thesis). Cambridge, Massachusetts: Massachusetts Institute of Technology.
• Shibatani, Masayoshi (1990). The Languages of Japan. Cambridge: Cambridge University Press. ISBN 0 521 36070 6.
• Vance, Timothy John (1987). An Introduction to Japanese Phonology. Albany, New York: State University of New York Press. ISBN 0-88706-360-8.
• Vance, Timothy John (1995). "Final accent vs. no accent: utterance-final neutralization in Tokyo Japanese" (PDF). Journal of Phonetics. 23. Elsevier: 487–499.
• Vance, Timothy John (October 2008). The Sounds of Japanese with Audio CD. Cambridge; New York: Cambridge University Press. ISBN 978-0-521-61754-3.

Further reading

• Kōri, Shirō (30 May 2015). 助詞・助動詞のアクセントについての覚え書き—直前形式との複合形態の観点からの分類— (PDF). 言語文化共同研究プロジェクト (in Japanese). Graduate School of Language and Culture, Osaka University: 63–74. doi:10.18910/53334.
• Kōri, Shirō (31 July 2020). 日本語の助詞・助動詞類のアクセント—一覧と使い分け，変化の方向性— (PDF). 言語文化共同研究プロジェクト (in Japanese). Graduate School of Language and Culture, Osaka University: 13–24. doi:10.18910/77063.

External links

External videos
Linguistics Lecture "Diversity in the Accents of Dialects" (YouTube)

• Japanese word accent speech analysis
• Formulation of Japanese pitch accent
• Online Japanese Accent Dictionary, from the University of Tokyo