Japanese Phonology

Japanese Lexical Strata

• Yamato (和語): inherited native vocabulary
• Mimetic (擬態語): Sound-symbolic words (e.g., がらがら)
• Sino-Japanese (漢語): Chinese-origin words (comprise ~60% of dictionary entries, ~20% of spoken Japanese)
• Foreign (外来語): Modern loanwords (mostly English)

Consonants

Lenition

The voiced plosives /b, d, ɡ/ may weaken to voiced fricatives [β, ð, ɣ] after a vowel word-internally or across word boundaries without a pause:

• /b/ -> [β]: あぶる/aburu/ -> [aβɯɾɯ]
• /d/ -> [ð]: 友達/tomodati/ -> [tomoðat͡ɕi]
• /ɡ/ -> [ɣ]: 描く/egaku/ -> [eɣakɯ]

Moraic Consonants

Certain consonant sounds count for a mora (a unit of timing or prosodic length).

Moraic Nasal (/N/)

The moraic nasal (ん, ン) acts as a syllable-final nasal. Its pronunciation assimilates to the place of articulation of the following sound:

• Bilabial [m] before /p, b, m/: 三枚[sammai]
• Velar [ŋ] before /k, ɡ/: 三回[saŋkai]
• Dorso-palatal [ŋʲ] before [kʲ, ɡʲ]: 元気[ɡeŋʲkʲi]
• Lamino-alveolar [n] before [t, d, ts, dz, n]: 三年[sanneɴ]
• Alveolo-palatal [ɲ̟] before [tɕ, dʑ, ɲ̟]: 三兆[saɲ̟tɕoː]
• Alveolar or Postalveolar [n̺], [n̠] before /r/: 三塁[san̺ɾɯi], 真理[ɕin̠d̠ɹ̝̆i]

Moraic Obstruent (/Q/)

The moraic obstruent (っ, ッ) acts as a syllable-final obstruent. It creates a long (geminate) consonant by assimilating to the place and manner of articulation of the following voiceless obstruent:

• Unreleased bilabial plosive [p̚] before /p/: 日本[ɲippoɴ]
• Unreleased alveolar plosive [t̚] before /t/: 切手[kitte]
• Unreleased alveolo-palatal affricate [tɕ̚] before /tɕ/: マッチ[mattɕi]
• Unreleased velar plosive [k̚] before /k/: 學校[gakkoː]
• Prolonged alveolar or alveolo-palatal fricative [s], [ɕ] before /s, ɕ/: 雑誌[zaɕɕi]

Coronal Affricates and Palatalization (/t/, /d/)

In core native vocabulary, the voiceless alveolar plosive /t/ palatalizes to [tɕ] before /i/ and affricates to [ts] before /u/. Similarly, its voiced counterpart /d/ becomes [dʑ] before /i/ and [dz] before /u/.

• /t/ + /i/ -> [tɕi] (ち/チ)
• /t/ + /u/ -> [tsɯ] (つ/ツ)
• /d/ + /i/ -> [dʑi] (じ/ジ)
• /d/ + /u/ -> [dzɯ] (ず/ズ)

Loanword Adaptation

• Older Loanwords: Foreign sequences were adapted to existing native Japanese allophones (/ti/ -> [tɕi], /di/ -> [dʑi], /tu/ -> [tsɯ]). (e.g., Ticket (チケット)/tɕiketto/, Radio (ラジオ)/radʑio/, Tour (ツアー)/tsɯaː/, Tool (ツール)/tsɯːrɯ/)
• Recent Loanwords: Use novel kana spellings (ティ, ディ, トゥ) to preserve non-palatalized sounds ([ti], [di], [tɯ]). (e.g., Party (パーティー)/paːtiː/, Disney (ディズニー)/dizuniː/, Tattoo (タトゥー)/tatɯː/)

Voiceless Bilabial Fricative ([ɸ])

In core native vocabulary, the voiceless glottal fricative /h/ is realized as the voiceless bilabial fricative [ɸ] exclusively before /u/. In this context, [ɸ] can be analyzed as an allophone of /h/.

ふね/hune/ -> [ɸɯne], ふた/huta/ -> [ɸɯ̥ta]

Loanword Adaptation

• Older Loanwords: Foreign [f] before vowels other than /u/ was traditionally adapted by replacing it with /h/ or by inserting [ɸɯ]. (e.g., Coffee (コーヒー)[koːhiː], Platform (プラットホーム)[purattohoːmɯ], Film (フイルム)[ɸɯirɯmɯ])
• Modern Loanwords: [ɸ] contrasts with [h] and [ç] before other vowels/glides (e.g., ファ, フィ, フェ, フォ, フュ). (e.g., Fin (フィン)[ɸiɴ], Ferry (フェリー)[ɸeɾiː], Fan (ファン)[ɸaɴ], Form (フォーム)[ɸoːmɯ], Fusion(フュージョン)[ɸjɯː(d)ʑoɴ]) - Minimal Pair: Fork[ɸoːkɯ] (フォーク) vs. Hawk[hoːkɯ] (ホーク)

Voiced Labiodental Fricative ([v])

Spellings with the kana ヴ were introduced in an attempt to render the foreign voiced labiodental fricative [v], a sound that does not natively exist in Japanese.

Loanword Adaptation

• General Realization: Most Japanese speakers do not phonetically distinguish foreign [v] from native /b/. There is no meaningful phonological difference in pronunciation between Violin (ヴァイオリン / バイオリン)[baioriɴ].
• Phonetic Realization: Innovative speakers attempting to approximate foreign [v] realize it as a voiced bilabial fricative [β], which is already an allophone of /b/. (e.g., Venice (ヴェネツィア)[βenetsia])

Velar Nasal Onset (bidakuon 鼻濁音)

For some Japanese speakers, the voiced velar plosive /ɡ/ is realized as a velar nasal [ŋ] in word-internal onset position (after a vowel or the moraic nasal /N/).

禿/hage/ -> [haŋe], 音楽/oNgaku/ -> [oŋŋakɯ]

Loanword Adaptation

In native Yamato vocabulary and Sino-Japanese words, speakers who consistently use bidakuon always pronounce word-internal /ɡ/ as [ŋ]. In modern foreign loanwords, usage varies and standard [ɡ] is often retained, but medial [ŋ] frequently appears in older loanwords or words that originally contained [ŋ] in the source language:

órgão(オルガン)/orugaN/ -> [orɯŋaɴ], inglês(イギリス)/igirisu/ -> [iŋirisɯ]

Vowels

Long vowels and vowel sequences

Long vowels are pronounced with around 2.5 or 3 times the phonetic duration of short vowels, but are considered to be two moras long at the phonological level.

Devoicing

Japanese vowels are sometimes phonetically voiceless. There is no phonemic contrast between voiced and voiceless versions of a vowel, but the use of voiceless vowels is an obligatory feature of standard Tokyo Japanese.

Between Voiceless Consonants Devoicing

A high vowel (/i/ or /u/) between two voiceless consonants is very likely to be devoiced if the second consonant is a stop or affricate, or if the first is a stop and the second is a voiceless fricative other than /h/.

鹿/sika/ -> [ɕi̥ka], 靴/kutu/ -> [kɯ̥tsɯ], 草/kusa/ -> [kɯ̥sa]

Phrase-Final Devoicing

high vowels are devoiced at the end of an utterance following a voiceless consonant, regardless of what follows (or even if nothing follows).

です/desu/ -> [desɯ̥], ます/masu/ -> [masɯ̥]

Atypical Devoicing

A high vowel may occasionally be devoiced after a voiceless consonant even when the following sound is voiced.

心kokoro -> ko̥koɾo], 誇りhokori -> [ho̥koɾi], 墓/haka/ -> [hḁka], 化粧/kesyoː/ -> [ke̥ɕoː]

Prosody

Moras

Japanese words have traditionally been analyzed as composed of moras, a distinct concept from that of syllables. Each mora is perceived to have the same time value. Regular Moras: vowel (V), consonant and a vowel (CV) Special Moras: /N/, /Q/, glide/j/ in regular moras (CjV), the second part of a long vowel (/R/).

Mora type	Example	Japanese	Moras per word
V	/o/	尾, o, ‘tail’	1-mora word
jV	/jo/	世, yo, ‘world’	1-mora word
CV	/ko/	子, ko, ‘child’	1-mora word
CjV	/kjo/	巨, kyo, ‘hugeness’	1-mora word
R	/R/ in /kjo.R/	今日, kyō, ‘today’	2-mora word
N	/N/ in /ko.N/	紺, kon, ‘deep blue’	2-mora word
Q	/Q/ in /ko.Q.ko/	国庫, kokko, ‘national treasury’	3-mora word

Pitch Accent

Standard Japanese is a pitch-accent language: each word has a lexically specified pattern of high and low pitch across its moras, rather than fixed stress.

Types of Accent Patterns

The key notion is the accented mora after which the pitch drops from high to low. If the first mora is accented, that mora is high and every following mora is low. Otherwise, the first mora is low, the pitch rises to high on the second mora and stays high through the accented mora, then drops to low right after it.

• Unaccented (heiban): has no accent mora. - 桜[sakɯɺa], 日本語[nihoŋɡo], ゴキブリ[ɡokibɯɺi], 虫[mɯɕi], 花[hana]
• Initial-accented (atamadaka): the first mora is accented. - 猫[neꜜko], 緑[miꜜdoɾi], 蜜柑[miꜜkaɴ], 神経[ɕiꜜŋkei], 目[meꜜ]
• Medial-/internal-accented (nakadaka): accented mora is neither the first nor last. - お菓子[okaꜜɕi], 平仮名[hiɾaɡaꜜna], 飲み物[nomimoꜜno], 心[kokoꜜɺo], 伊勢海老[iseeꜜbi]
• Final-accented (odaka): the last mora is accented. - 妹が[imoːtoꜜɡa], 男が[otokoꜜɡa], 父が[tɕitɕiꜜɡa], 娘が[mɯsɯmeꜜɡa], 頭が[atamaꜜɡa]

Syllables

Normally, each syllable contains at least one vowel and has a length of either one mora (light syllable) or two moras (heavy syllable) Thus the typical Japanese syllable structure is (C)(j)V(V/N/Q), where C is an onset consonant, V is a vowel, N is a moraic nasal, Q represents a moraic obstruent, components in parentheses are optional, and components separated by a slash are mutually exclusive.

Light Syllables

The majority of syllables in Japanese are light with the form (C)(j)V.

Heavy syllables

Heavy syllables (two moras long) may potentially take any of the following forms:

• (C)(j)VN: ending in a short vowel + /N/
• (C)(j)VQ: ending in a short vowel + /Q/
• (C)(j)VR: ending in a long vowel (R).
• (C)(j)V₁V₂, where V₁ is different from V₂.

Morphophonology

Rendaku (連濁)

Rendaku is a pronunciation change seen in the middle of some compound words. It replaces a voiceless consonant /k s t h/ with a voiced consonant /ɡ z d b/. It modifies the consonant at the start of the second (or later) part of the compound. Rendaku may have originated from the fusion of consonants with preceding nasal sounds derived from reduction of either の or に.

• /k/ -> /ɡ/:
  • te + kami -> /tegami/ (手紙)
• /s/ -> /z/:
  • [s] -> [z]: kuro + sato -> /kurozatoː/ (黒砂糖)
  • [ɕ] -> [dʑ]: tate + shima -> /tatezima/ (縦縞)
• /t/ -> /d/ or /z/:
  • [t] -> [d]: hi + tane -> /hidane/ (火種)
  • [ts] -> [dz]: ari + tsuka -> /ariduka/ (蟻塚)
  • [tɕ] -> [dʑ]: hana + chi -> /hanadi/ (鼻血)
• /h/ -> /b/:
  • [h] -> [b]: se + hone -> /sebone/ (背骨)
  • [ç] -> [b]: hana + hi -> /hanabi/ (花火)
  • [ɸ] -> [b]: kawa + fune -> /kawabune/ (川船)

Lyman’s Law

Rendaku is blocked when the second element already contains a voiced obstruent /ɡ z d b/.

• /ɡ/: ko + kage -> /kokage/ (木陰)
• /z/: umi + kaze -> /umikaze/ (海風)
• /d/: machi + kado -> /machikado/ (街角)
• /b/: hitori + tabi -> /hitoritabi/ (一人旅)

Sino-Japanese Gemination

When the second mora of a Sino-Japanese morpheme is つ, く, ち, or き and it is followed by a voiceless consonant, this mora is sometimes replaced by the sokuon っ, forming a geminate consonant.

• つ (tsu) -> っ (Q) / _ [voiceless obstruent]: Systematically becomes /Q/ before any voiceless obstruent.
  • itsu + sho -> /issjo/ (一緒)
• く (ku) -> っ (Q) / _ /k/: Systematically becomes /Q/ before /k/.
  • gaku + kō -> /gakkoː/ (学校)
• ち (chi) -> っ (Q) / _ [voiceless obstruent]: Optionally becomes /Q/ before voiceless obstruents in certain morphemes.
  • nichi + kō -> /nikkoː/ (日光)
• き (ki) -> っ (Q) / _ /k/: Optionally becomes /Q/ before /k/ in certain morphemes.
  • seki + kaku -> /sekkaku/ (折角)

Vowel Fusion & Colloquial Contraction

Vowel fusion refers to historical and modern phonological processes where two adjacent vowels merge into a single long vowel or diphthong.

Historical Rules

• /au/, /ou/ -> /oː/
• /eu/ -> /yoː/
• /iu/ -> /juː/

Modern Colloquial Fusions

Sequences like /ai, oi, ui/ fuse into long [eː] or [iː] in casual and masculine speech.

• /ai/ -> [eː]: /nai/ -> [neː] (ない -> ねえ)
• /oi/ -> [eː]: /sugoi/ -> [sugeː] (凄い -> 凄え)
• /ui/ -> [iː]: /warui/ -> [wariː] (悪い -> わりい)

Onbin (音便)

Historical sound reductions, lenitions, or coalescences that mainly occur in verb and adjective conjugations.

i-onbin (イ音便)

Occurs when the consonant is dropped and replaced by /i/:

• Verb Conjugation (ki, gi + te):
  • kaki + te -> kakite -> /kaite/ (書いて)
  • oyogi + te -> oyogite -> /oyoide/ (泳いで)
• Adjective Origin (k deletion → modern -i adjectives):
  • atsuki -> /atsui/ (熱い)
  • utsukushiki -> /utsukushii/ (美しい)

u-onbin (ウ音便)

Occurs when the consonant is dropped and replaced by /u/, fusing with preceding vowels:

• Kansai Dialect Verb Conjugation (/pi, bi, mi/ + /te/):
  • omopi + te -> omopite -> omoute -> /omoːte/ (思うて)
  • yobi + te -> yobite -> youde -> /yoːde/ (呼うで)
• Adjective Conjugation (Polite Expressions):
  • yoku -> you -> /yoː/ (良う)
  • arigataku -> arigatau -> /arigatoː/ (有り難う)
  • hayaku -> hayau -> /hayoː/ (早う)

Sokuonbin (促音便)

Occurs when the mora assimilates into the moraic obstruent っ:

• Verb Conjugation (/ti, ri, pi/ + /te/):
  • kati + te -> katite -> /katte/ (勝って)
  • tori + te -> torite -> /totte/ (取って)
  • kapi + te -> kapite -> /katte/ (買って)
• Verb 行く (iku):
  • iki + te -> ikite -> /itte/ (行って)
• Emphatic Adverbs & Nouns:
  • opi + to -> opito -> /otto/ (夫)
  • masiro -> /massiro/ (真っ白)

Hatsuonbin (撥音便)

Occurs when the mora assimilates into the moraic nasal ん:

• Verb Conjugation (/mi, bi, ni/ + /te/):
  • yomi + te -> yomite -> /yonde/ (読んで)
  • tobi + te -> tobite -> /tonde/ (飛んで)
  • shini + te -> shinite -> /sinde/ (死んで)
• Noun Assimilation:
  • manaka -> /mannaka/ (真ん中)

Polite Adjective Forms

The polite adjective forms (used before ござる and 存じる) exhibit a one-step or two-step sound change. The continuative ending -く undergoes u-onbin to become -う, which then fuses with the preceding vowel:

arigataku -> arigatau -> /arigatoː/ (有り難う / ありがとう) omedetaku -> omedetau -> /omedetoː/ (お目出度う / おめでとう)