Japanese is characterised largely by its small number of vowels and consonants (five and fourteen, respectively). Pronunciation of each syllable is highly regular with the written system and there are only a few exceptions such as vowel devoicing. This is in stark contrast to English where the written and spoken language can differ a great deal (e.g. the vowel digraph "ou" in "noun" and "cough" and the consonant "g" in "goat" and "giraffe").

Apart from a single isolated consonant (the moraic nasal, "n") and double consonants (e.g. "itte" and "kekkon") all consonants must be followed by a vowel to form syllables. Double consonants are always a pair of the same consonant, though vowel devoicing sometimes makes different consonants sound one after the other (e.g. "suki" and "suteki").

Japanese has a great deal of homophones that make correct pronunciation quite important. While language learners may have difficulty hearing the difference between nuances like long and short vowels, native speakers are used to these and might not understand incorrectly pronounced words.

The syllabary

There are five vowels in Japanese, normally transcribed into the English alphabet as: "a", "i", "u", "e" and "o".

Vowel	a	i	u	e	o
Approximate sound	father	meaty	food	egg	old

*This sound has no approximation in English. See: http://en.wikipedia.org/wiki/Close_back_rounded_vowel#Close_back_compressed_vowel

Spanish and Italian speakers may note that Japanese vowels produce the same sounds as their Spanish and Italian equivalents.

Japanese vowels always represent distinct phonemes and don't form digraphs — i.e. they don't blend together or sound differently when joined. When one vowel follows another they are pronounced separately. Examples are the names Sae (sa.e) and Aoi (a.o.i)

The rest of the syllabary is formed by combining the above vowels with a consonant.

Clear							Voiced							Plosive							Clear medial y					Voiced medial y					Plosive medial y
	a	i	u	e	o			a	i	u	e	o			a	i	u	e	o			ya	yu	yo			ya	yu	yo			ya	yu	yo
k	ka	ki	ku	ke	ko		g	ga	gi	gu	ge	go									ki	kya	kyu	kyo		gi	gya	gyu	gyo
s	sa	shi	su	se	so		z	za	ji	zu	ze	zo									shi	sha	shu	sho		ji	ja	ju	jo
t	ta	chi	tsu	te	to		d	da	ji	zu	de	do									chi	cha	chu	cho		ji	ja	ju	jo
n	na	ni	nu	ne	no																ni	nya	nyu	nyo
h	ha	hi	fu	he	ho		b	ba	bi	bu	be	bo		p	pa	pi	pu	pe	po		hi	hya	hyu	hyo		bi	bya	byu	byo		pi	pya	pyu	pyo
m	ma	mi	mu	me	mo																mi	mya	myu	myo
y	ya		yu		yo
r	ra	ri	ru	re	ro																ri	rya	ryu	ryo
w	wa				o

Note that the sound which is written with a "y" is not considered a vowel, but a consonant. This will come as little surprise to German speakers where the same sound is written with a "j".

The -i line (ki, gi, shi, ji, chi, ni, hi, bi, pi, mi, ri) can be combined with the y- line (ya, yu, yo) to create the medial y combinations. These are just like regular consonant + vowel syllables, in that they should be pronounced as one mora (syllabic sound).

Mora

Japanese is quite regular in the timing and stress of its syllables. The basic timing unit is called mora. Each mora is pronounced with equal stress and should take about the same amount of time. Two morae should sound twice as long as a single one.

The following take up one mora:

• a short vowel
• a consonant followed by a short vowel
• a medial y
• a moraic nasal

Whereas these take up two morae:

• a long vowel
• a double consonant

Long vowels

A long vowel takes two morae. In rōmaji it's written with a macron: ā, ī, ū, ē and ō.

In hiragana, it's written with an extra "あ" (a), "い" (i) or "う" (u) depending on the vowel. In katakana, it's marked by appending a dash-like symbol "ー".

Word	Japanese	Meaning	Soundbyte
Ōsaka	大阪	Osaka city	Ja-Osaka.ogg
Tōkyō	東京	Tokyo city	Ja-Tokyo.ogg
dēta	データ	data	Ja-deeta-data.ogg
gyūnyū	牛乳	milk	Ja-gyuunyuu-milk.ogg
hō	頬	cheek	Ja-hoo-cheek.ogg

Devoicing

In standard Japanese the vowels i and u are not usually voiced when they occur between voiceless consonants (k, s, sh, t, ch, h, f). The phenomenon seems to have developed to facilitate the falling pitch intonation in the Kanto dialect. The mouth forms shape of the vowel and lasts for one mora, but the sound is not voiced. For final [su] in 'desu' and '-masu', all vestiges of the vowel have disappeared in standard Japanese, leaving a naked sibilant. Devoicing is not otherwise standard for word terminal i or u. Consecutive devoicing is rare, although exceptions exists (e.g. futsuka, 2nd day of the month, pronounced f-ts-ka). Devoicing can depend on context. E.g. 'Suzuki' has no devoicing; 'Suzuki-san' has a devoiced i: Suzuk-san. Some dialects do not demonstrate devoicing, notably Kansai.

Some examples:

Consonant variation

There are a couple of consonants that are pronounced differently from English:

Except for the doubled consonants and the n (which we will cover later), consonants can never end a syllable. They can only begin it.

Moraic nasal

Normally, Japanese consonants must be followed by a vowel except where they double. There is an exception to this: the moraic nasal which is transliterated as n. It is usually found at the end of words, but can be found in the middle of composite words.

The difference between the moraic nasal and the syllables "na", "ni", "nu", "ne" and "no" can be difficult for language learners to spot, while native speakers may have difficulty understanding incorrect pronunciation.

• kin'en (ki-n-e-n) no smoking vs. kinen (ki-ne-n) commemoration.
• hon'ya (ho-n-ya-) bookstore (not ho-nya)

The pronunciation of the moraic nasal changes depending on what sound follows it. This is not so much an irregularity as a shortcut to bridge the sounds between the two morae. When followed by the bilabial plosives, "b" and "p", the moraic nasal is pronounced like an "m". An example:

• "shinbun" is read as: shimbun

Listen to the audio (OggVorbis, 151 KB)

1. At the end of a word:
   • dan 段 "level"
   • kin 金 "gold"
   • fun 糞 "dung"
   • zen 善 "goodness"
   • hon 本 "book"
2. Directly before a consonant:
   • banzai 万歳 "hurrah", "long live (the Emperor)"
   • kingyo 金魚 "goldfish" (pronounced like "ng")
   • kunrei 訓令 "directive"
   • zenchi 全知 "omniscience" (pronounced like "n")
   • honten 本店 "main office" (pronounced like "n")
3. Before m, b, p
   • genmai 玄米 "unmilled rice"
   • honbu 本部 "headquarters"
   • tenpura 天ぷら (battered and fried vegetables or fish)
4. Before a, i, e, y
   • zen'aku 善悪 "good and evil"
   • ken'i 権威 "authority"
   • han'ei 反映 "reflection"
   • sen'you 専用 "exclusive use"
5. Note that before a, i, e, and y, moraic n is written n' (with an apostrophe). This is to distinguish it from the regular consonant n, which is pronounced differently and can produce different words. Some examples of cases where this becomes important are:
   • kani 蟹 "crab" vs. kan'i 簡易 "simplicity"
   • kinyuu 記入 "fill in" vs. kin'yuu 金融 "finances"
   • konyakku コニャック "cognac" vs. kon'yaku 婚約 "engagement (to be married)"

Consonant doubling (gemination)

There are four consonants that can become geminates (get doubled) in native Japanese words: /p/, /t/, /k/, and /s/. The geminate (represented linguistically as "Q") takes up an extra mora, with the general effect being to insert a pause that sounds as long as a regular syllable with a short vowel. The geminate is /t/ before ch and ts, /s/ before sh.

Word	Meaning	Soundbyte
takkyū	table tennis	Ja-takkyuu-table_tennis.ogg
Hokkaido	Hokkaido prefecture	Ja-hokkaido.ogg
makka	bright red	Ja-makka-bright_red.ogg
gakkō	a school
dotchi	which (informal)	Ja-docchi-which.ogg
kuttsuku	to stick	Ja-kuttsuku-to stick.ogg }
settei	setting	Ja-settei-setting.ogg
chotto	a little
kissaten	a tea house
hissori	quiet(ly)	Ja-hissori-quiet(ly).ogg
juppun	ten minutes
Sapporo	Sapporo city	Ja-Sapporo.ogg

In the Japanese pronunciation of foreign loan words, the voiced consonants /b/, /d/, /g/, and /z/ can also be doubled.

Practice