The phonology of Japanese features about 15 consonant phonemes, the cross-linguistically typical five-vowel system of /a, i, u, e, o/, and a relatively simple phonotactic distribution of phonemes allowing few consonant clusters. It is traditionally described as having a mora as the unit of timing, with each mora taking up about the same length of time, so that the disyllabic [ɲip.poɴ] ("Japan") may be analyzed as /niQpoN/ and dissected into four moras, /ni/, /Q/, /po/, and /N/.
Standard Japanese is a pitch-accent language, wherein the position or absence of a pitch drop may determine the meaning of a word: /haꜜsiɡa/ "chopsticks", /hasiꜜɡa/ "bridge", /hasiɡa/ "edge" (see Japanese pitch accent).
Unless otherwise noted, the following describes the standard variety of Japanese based on the Tokyo dialect.
|Stop||p b||t d||k ɡ|
|Affricate||(t͡s) (d͡z)||(t͡ɕ) (d͡ʑ)|
|Fricative||(ɸ)||s z||(ɕ) (ʑ)||(ç)||h|
|Special moras||/N/, /Q/|
See below for more in-detail descriptions of allophonic variation.
|/b/ > bilabial fricative [β]:||/abareru/ > [aβaɾeɾɯ] abareru 暴れる 'to behave violently'|
|/ɡ/ > velar fricative [ɣ]:||/haɡe/ > [haɣe] hage はげ 'baldness'|
However, /ɡ/ is further complicated by its variant realization as a velar nasal [ŋ]. Standard Japanese speakers can be categorized into 3 groups (A, B, C), which will be explained below. If a speaker pronounces a given word consistently with the allophone [ŋ] (i.e. a B-speaker), that speaker will never have [ɣ] as an allophone in that same word. If a speaker varies between [ŋ] and [ɡ] (i.e. an A-speaker) or is generally consistent in using [ɡ] (i.e. a C-speaker), then the velar fricative [ɣ] is always another possible allophone in fast speech.
/ɡ/ may be weakened to nasal [ŋ] when it occurs within words—this includes not only between vowels but also between a vowel and a consonant. There is a fair amount of variation between speakers, however. Vance (1987) suggests that the variation follows social class, while Akamatsu (1997) suggests that the variation follows age and geographic location. The generalized situation is as follows.
In the middle of compound words morpheme-initially:
So, for some speakers the following two words are a minimal pair while for others they are homophonous:
To summarize using the example of hage はげ 'baldness':
Some phonologists posit a distinct phoneme /ŋ/, citing pairs such as [oːɡaɾasɯ] 大硝子 'big sheet of glass' vs. [oːŋaɾasɯ] 大烏 'big raven'.
The palatals /i/ and /j/ palatalize the consonants preceding them:
|/m/ > palatalized [mʲ]:||/umi/ > [ɯmʲi] umi 海 'sea'|
|/ɡ/ > palatalized [ɡʲ]:||/ɡjoːza/ > [ɡʲoːza] gyōza ぎょうざ 'fried dumpling'|
|/n/ > Alveolo-palatal nasal [ɲ̟]:||/nihoN/ > [ɲihoɴ] nihon 日本 'Japan'|
|/s/ > alveolo-palatal fricative [ɕ]:||/sio/ > [ɕi.o] shio 塩 'salt'|
|/z/ > alveolo-palatal [d͡ʑ] or [ʑ]:||/zisiN/ > [d͡ʑiɕiɴ] jishin 地震 'earthquake';|
/ɡozjuː/ > [ɡod͡ʑɯː] ~ [ɡoʑɯː] gojū 五十 'fifty'
|/t/ > alveolo-palatal affricate [t͡ɕ]:||/tiziN/ > [t͡ɕid͡ʑiɴ] ~ [t͡ɕiʑiɴ] chijin 知人 'acquaintance'|
/i/ and /j/ also palatalize /h/ to a palatal fricative ([ç]): /hito/ > [çito] hito 人 ('person')
Of the allophones of /z/, the affricate [d͡z] is most common, especially at the beginning of utterances and after /N/, while fricative [z] may occur between vowels. Both sounds, however, are in free variation.
In the case of the /s/, /z/, and /t/, when followed by /j/, historically, the consonants were palatalized with /j/ merging into a single pronunciation. In modern Japanese, these are arguably separate phonemes, at least for the portion of the population that pronounces them distinctly in English borrowings.
|/sj/ > [ɕ] (romanized as sh):||/sjaboN/ > [ɕaboɴ] shabon シャボン 'soap'|
|/zj/ > [d͡ʑ ~ ʑ] (romanized as j):||/zjaɡaimo/ > [d͡ʑaɡaimo] jagaimo じゃがいも 'potato'|
|/tj/ > [t͡ɕ] (romanized as ch):||/tja/ > [t͡ɕa] cha 茶 'tea'|
|/hj/ > [ç] (romanized as hy):||/hjaku/ > [çakɯ] hyaku 百 'hundred'|
The vowel /u/ also affects consonants that it follows:
|/h/ > bilabial fricative [ɸ]:||/huta/ > [ɸɯta] futa ふた 'lid'|
|/t/ > dental affricate [t͡s]:||/tuɡi/ > [t͡sɯɡi] tsugi 次 'next'|
Although [ɸ] and [t͡s] occur before other vowels in loanwords (e.g. [ɸaito] faito ファイト 'fight'; [ɸjɯː(d)ʑoɴ] fyūjon フュージョン 'fusion'; [t͡saitoɡaisɯto] tsaitogaisuto ツァイトガイスト 'Zeitgeist'; [eɾit͡siɴ] eritsin エリツィン 'Yeltsin'), [ɸ] and [h] are distinguished before vowels except [ɯ] (e.g. English fork vs. hawk > fōku [ɸoːkɯ] フォーク vs. hōku [hoːkɯ] ホーク). *[hɯ] is still not distinguished from [ɸɯ] (e.g. English hood vs. food > [ɸɯːdo] fūdo フード). Similarly, *[si] and *[(d)zi] usually do not occur even in loanwords so that English cinema becomes [ɕinema] shinema シネマ; although they may be written スィ and ズィ respectively, they are rarely found even among the most innovative speakers and do not occur phonemically.
The contrast between /d/ and /z/ is neutralized before /i/ and /u/: [(d)ʑi, (d)zɯ]. By convention, it is often assumed to be /z/, though some analyze it as /d͡z/, the voiced counterpart to [t͡s]. The writing system preserves morphological distinctions, though spelling reform has eliminated historical distinctions except in cases where a mora is repeated once voiceless and once voiced, or where rendaku occurs in a compound word: つづく[続く] /tuduku/, いちづける[位置付ける] /itidukeru/ from |iti+tukeru|. Some dialects retain the distinctions between /zi/ and /di/ and between /zu/ and /du/, while others retain only /zu/ and /du/ but not /zi/ and /di/, or merge all four (see Yotsugana).
These assimilations occur beyond word boundaries.
While Japanese features consonant gemination, there are some limitations in what can be geminated. Most saliently, voiced geminates are prohibited in native Japanese words. This can be seen with suffixation that would otherwise feature voiced geminates. For example, Japanese has a suffix, |ri| that contains what Kawahara (2006) calls a "floating mora" that triggers gemination in certain cases (e.g. |tapu| +|ri| > [tappɯɾi] 'a lot of'). When this would otherwise lead to a geminated voiced obstruent, a moraic nasal appears instead as a sort of "partial gemination" (e.g. |zabu| + |ri| > [(d)zambɯɾi] 'splashing').
In the late 20th century, voiced geminates began to appear in loanwords, though they are marked and have a high tendency to devoicing. A frequent example is loanwords from English such as bed and dog that, though they end with voiced singletons in English, are geminated (with an epethentic vowel) when borrowed into Japanese. These geminates frequently undergo devoicing to become less marked, which gives rise to variability in voicing:
The distinction is not rigorous. For example, when voiced obstruent geminates appear with another voiced obstruent they can undergo optional devoicing (e.g. doreddo ~ doretto 'dreadlocks'). Kawahara (2006) attributes this to a less reliable distinction between voiced and voiceless geminates compared to the same distinction in non-geminated consonants, noting that speakers may have difficulty distinguishing them due to the partial devoicing of voiced geminates and their resistance to the weakening process mentioned above, both of which can make them sound like voiceless geminates.
There is some dispute about how gemination fits with Japanese phonotactics. One analysis, particularly popular among Japanese scholars, posits a special "mora phoneme" (モーラ 音素 Mōra onso) /Q/, which corresponds to the sokuon ⟨っ⟩. However, not all scholars agree that the use of this "moraic obstruent" is the best analysis. In those approaches that incorporate the moraic obstruent, it is said to completely assimilate to the following obstruent, resulting in a geminate (that is, double) consonant. The assimilated /Q/ remains unreleased and thus the geminates are phonetically long consonants. /Q/ does not occur before vowels or nasal consonants. This can be seen as an archiphoneme in that it has no underlying place or manner of articulation, and instead manifests as several phonetic realizations depending on context, for example:
|[p̚] before [p]:||/niQ.poN/ > [ɲip̚.poɴ] nippon 日本 'Japan'|
|[s] before [s]:||/kaQ.seN/ > [kas.seɴ] kassen 合戦 'battle'|
|[t̚] before [t͡ɕ]:||/saQ.ti/ > [sat̚.t͡ɕi] satchi 察知 'inference'|
Another analysis of Japanese dispenses with /Q/. In such an approach, the words above are phonemicized as shown below:
|[p̚] before [p]:||/nip.poN/ > [ɲip̚.poɴ] nippon 日本 'Japan'|
|[s] before [s]:||/kas.seN/ > [kas.seɴ] kassen 合戦 'battle'|
|[t̚] before [t͡ɕ]:||/sat.ti/ > [sat̚.t͡ɕi] satchi 察知 'inference'|
Gemination can of course also be transcribed with a length mark (e.g. [ɲipːoɴ]), but this notation obscures mora boundaries.
Various forms of sandhi exist; the Japanese term for sandhi generally is ren'on (連音), while sandhi in Japanese specifically is called renjō (連声). Most commonly, a terminal /N/ on one morpheme results in /n/ or /m/ being added to the start of the next morpheme, as in tennō (天皇, emperor), てん ＋ おう > てんのう (ten + ō = tennō). In some cases, such as this example, the sound change is used in writing as well, and is considered the usual pronunciation, though in other cases, such as abbreviating ...-no-uchi (〜の家, ...'s house) to 〜んち (-nchi) this is only done in speech, and considered informal. See 連声 (in Japanese) for further examples.
Except for /u/, the short vowels are similar to their Spanish counterparts.
Vowels have a phonemic length contrast (i.e. short vs. long). Compare contrasting pairs of words like ojisan /ozisaN/ 'uncle' vs. ojiisan /oziisaN/ 'grandfather', or tsuki /tuki/ 'moon' vs. tsūki /tuuki/ 'airflow'.
Some analyses make a distinction between a long vowel and a succession of two identical vowels, citing pairs such as 砂糖屋 satōya 'sugar shop' [satoːja] vs. 里親 satooya 'foster parent' [satooja]. They are usually identical in normal speech, but when enunciated a distinction may be made with a pause or a glottal stop inserted between two identical vowels.
Within words and phrases, Japanese allows long sequences of phonetic vowels without intervening consonants, pronounced with hiatus, although the pitch accent and slight rhythm breaks help track the timing when the vowels are identical. Sequences of two vowels within a single word are extremely common, occurring at the end of many i-type adjectives, for example, and having three or more vowels in sequence within a word also occurs, as in aoi 'blue/green'. In phrases, sequences with multiple o sounds are most common, due to the direct object particle を 'wo' (which comes after a word) being realized as o and the honorific prefix お〜 'o', which can occur in sequence, and may follow a word itself terminating in an o sound; these may be dropped in rapid speech. A fairly common construction exhibiting these is 「〜をお送りします」 ... (w)o o-okuri-shimasu 'humbly send ...'. More extreme examples follow:
|/hoː.oː.o.o.oː/ [hoː.oː.o.o.oː]||hōō o oō (
||'let's chase the fenghuang'|
|/toː.oː.o.oː.oː/ [toː.oː.o.oː.oː]||tōō o ōō (
||'let's cover Eastern Europe'|
|/kutu/ > [kɯ̥t͡sɯ]||kutsu 靴 'shoe'|
|/atu/ > [at͡sɯ̥]||atsu 圧 'pressure'|
|/hikaN/ > [çi̥kaɴ]||hikan 悲観 'pessimism'|
Generally, devoicing does not occur in a consecutive manner:
|/kisitu/ > [kʲi̥ɕit͡sɯ]||kishitsu 気質 'temperament'|
|/kusikumo/ > [kɯɕi̥kɯmo]||kushikumo 奇しくも 'strangely'|
This devoicing is not restricted to only fast speech, though consecutive voicing may occur in fast speech.
To a lesser extent, /o, a/ may be devoiced with the further requirement that there be two or more adjacent moras containing the same phoneme:
|/kokoro/ > [ko̥koɾo]||kokoro 心 'heart'|
|/haka/ > [hḁka]||haka 墓 'grave'|
Japanese speakers are usually not even aware of the difference of the voiced and devoiced pair. On the other hand, gender roles play a part in prolonging the terminal vowel: it is regarded as effeminate to prolong, particularly the terminal /u/ as in arimasu. Some nonstandard varieties of Japanese can be recognized by their hyper-devoicing, while in some Western dialects and some registers of formal speech, every vowel is voiced.
Japanese vowels are slightly nasalized when adjacent to nasals /m, n/. Before the moraic nasal /N/, vowels are heavily nasalized:
|/seesaN/ > [seːsãɴ]||seisan 生産 'production'|
At the beginning and end of utterances, Japanese vowels may be preceded and followed by a glottal stop [ʔ], respectively. This is demonstrated below with the following words (as pronounced in isolation):
|/eN/ > [eɴ] ~ [ʔeɴ]:||en 円 'yen'|
|/kisi/ > [kiɕiʔ]:||kishi 岸 'shore'|
|/u/ > [ɯʔ ~ ʔɯʔ]:||u 鵜 'cormorant'|
When an utterance-final word is uttered with emphasis, this glottal stop is plainly audible, and is often indicated in the writing system with a small letter tsu ⟨っ⟩ called a sokuon. This is also found in interjections like あっ and えっ. These words are likely to be romanized as ⟨a'⟩ and ⟨e'⟩, or rarely as ⟨at⟩ and ⟨et⟩.
|Marginal combinations mostly found in Western loans|
[ɴ, m, n, ɲ, ŋ, ɰ̃]
(geminates the following consonant)
Japanese words have traditionally been analysed as composed of moras; a distinct concept from that of syllables. Each mora occupies one rhythmic unit, i.e. it is perceived to have the same time value. A mora may be "regular" consisting of just a vowel (V) or a consonant and a vowel (CV), or may be one of two "special" moras, /N/ and /Q/. A glide /j/ may precede the vowel in "regular" moras (CjV). Some analyses posit a third "special" mora, /R/, the second part of a long vowel (a chroneme). In this table, the period represents a mora break, rather than the conventional syllable break.
|Mora type||Example||Japanese||Moras per word|
|V||/o/||o 尾 'tail'||1-mora word|
|jV||/jo/||yo 世 'world'||1-mora word|
|CV||/ko/||ko 子 'child'||1-mora word|
|CjV||/kjo/1||kyo 巨 'hugeness'||1-mora word|
|R||/R/ in /kjo.R/ or /kjo.o/||kyō 今日 'today'||2-mora word|
|N||/N/ in /ko.N/||kon 紺 'deep blue'||2-mora word|
|Q||/Q/ in /ko.Q.ko/ or /ko.k.ko/||kokko 国庫 'national treasury'||3-mora word|
/N/ is restricted from occurring word-initially, and /Q/ is found only word-medially. Vowels may be long, and the voiceless consonants /p, t, k, s, n/ may be geminate (doubled). In the analysis with archiphonemes, geminate consonants are the realization of the sequences /Nn/, /Nm/ and sequences of /Q/ followed by a voiceless obstruent, though some words are written with geminate voiced obstruents. In the analysis without archiphonemes, geminate clusters are simply two identical consonants, one after the other.
In English, stressed syllables in a word are pronounced louder, longer, and with higher pitch, while unstressed syllables are relatively shorter in duration. Japanese is often considered a mora-timed language, as each mora tends to be of the same length, though not strictly: geminate consonants and moras with devoiced vowels may be shorter than other moras. Factors such as pitch have negligible influence on mora length.
Standard Japanese has a distinctive pitch accent system: a word can have one of its moras bearing an accent or not. An accented mora is pronounced with a relatively high tone and is followed by a drop in pitch. The various Japanese dialects have different accent patterns, and some exhibit more complex tonic systems.
As an agglutinative language, Japanese has generally very regular pronunciation, with much simpler morphophonology than a fusional language would. Nevertheless, there are a number of prominent sound change phenomena, primarily in morpheme combination and in conjugation of verbs and adjectives. Phonemic changes are generally reflected in the spelling, while those that are not either indicate informal or dialectal speech which further simplify pronunciation.
In Japanese, sandhi is prominently exhibited in rendaku – consonant mutation of the initial consonant of a morpheme from unvoiced to voiced in some contexts when it occurs in the middle of a word. This phonetic difference is reflected in the spelling via the addition of dakuten, as in ka, ga (か／が). In cases where this combines with the yotsugana mergers, notably ji, dzi (じ／ぢ) and zu, dzu (ず／づ) in standard Japanese, the resulting spelling is morphophonemic rather than purely phonemic.
The other common sandhi in Japanese is conversion of つ or く (tsu, ku), and ち or き (chi, ki), and rarely ふ or ひ (fu, hi) as a trailing consonant to a geminate consonant when not word-final – orthographically, the sokuon っ, as this occurs most often with つ. So that
Some long vowels derive from an earlier combination of a vowel and fu (see onbin). The f often causes gemination when it is joined with another word:
Most words exhibiting this change are Sino-Japanese words deriving from Middle Chinese morphemes ending in /t̚/, /k̚/ or /p̚/, which were borrowed on their own into Japanese with a prop vowel after them (e.g. 日 MC */nit̚/ > Japanese /niti/ [ɲit͡ɕi]) but in compounds as assimilated to the following consonant (e.g. 日本 MC */nit̚.pu̯ən/ > Japanese /niQ.poN/ [ɲip̚.poɴ]).
Sandhi also occurs much less often in renjō (連声), where, most commonly, a terminal /N/ or /Q/ on one morpheme results in /n/ (or /m/ when derived from historical m) or /t̚/ respectively being added to the start of a following morpheme beginning with a vowel or semivowel, as in ten + ō → tennō (天皇: てん + おう → てんのう). Examples:
Another prominent feature is onbin (音便, euphonic sound change), particularly historical sound changes.
In cases where this has occurred within a morpheme, the morpheme itself is still distinct but with a different sound, as in hōki (箒 (ほうき), broom), which underwent two sound changes from earlier hahaki (ははき) → hauki (はうき) (onbin) → houki (ほうき) (historical vowel change) → hōki (ほうき) (long vowel, sound change not reflected in kana spelling).
However, certain forms are still recognizable as irregular morphology, particularly forms that occur in basic verb conjugation, as well as some compound words.
The polite adjective forms (used before the polite copula gozaru (ござる, be) and verb zonjiru (存じる, think, know)) exhibit a one-step or two-step sound change. Firstly, these use the continuative form, -ku (-く), which exhibits onbin, dropping the k as -ku (-く) → -u (-う). Secondly, the vowel may combine with the preceding vowel, according to historical sound changes; if the resulting new sound is palatalized, meaning yu, yo (ゆ、よ), this combines with the preceding consonant, yielding a palatalized syllable.
This is most prominent in certain everyday terms that derive from an i-adjective ending in -ai changing to -ō (-ou), which is because these terms are abbreviations of polite phrases ending in gozaimasu, sometimes with a polite o- prefix. The terms are also used in their full form, with notable examples being:
Other transforms of this type are found in polite speech, such as oishiku (美味しく) → oishū (美味しゅう) and ōkiku (大きく) → ōkyū (大きゅう).
The morpheme hito (人 (ひと), person) (with rendaku -bito (〜びと)) has changed to uto (うと) or udo (うど), respectively, in a number of compounds. This in turn often combined with a historical vowel change, resulting in a pronunciation rather different from that of the components, as in nakōdo (仲人 (なこうど), matchmaker) (see below). These include:
In some cases morphemes have effectively fused and will not be recognizable as being composed of two separate morphemes.