Sounds and rhythm

Japanese has a small inventory of sounds and a strict sense of timing, which is good news and bad news at once. The good news is that there is very little to learn: five vowels, a modest set of consonants, and nothing like the vowel zoo of English. The bad news is that the timing is not optional. Getting a beat wrong in Japanese does not produce an accent, it produces a different word, and this lesson is about the handful of places where that happens.

Five vowels, held steady

Every vowel in Japanese is short, pure and unchanging from the moment it starts to the moment it stops. あ sits near the a of "father" but clipped short. い is the ee of "machine", again short. う is made with relaxed, unrounded lips, not the pursed oo of "boot". え is the e of "bed". お is the o of "for", cut off before it can glide.

That last point is the one English speakers keep breaking. English vowels slide: "no" drifts from o towards u, "day" from e towards i. A Japanese vowel does not move. Say お and freeze the mouth in position until the sound ends. The same discipline applies inside a word: こえ is a clean お followed by a clean え, two beats, not a diphthong.

Three consonants also need care. The ら row is a single tap of the tongue against the ridge behind the teeth, closer to the middle of English "butter" than to either l or r, and English speakers who substitute an r sound noticeably foreign. ふ is blown between the two lips with no teeth involved, halfway between f and h. つ is a t and an s fused, the sound at the end of "cats", and it is hard only at the start of a word, where English never puts it.

The mora, not the syllable

The unit of Japanese rhythm is the mora, a beat of fixed length, and the writing system marks it for you: one kana is one mora, with small ゃ ゅ ょ folding into the character before them. Every mora takes the same time, whether it is a bare vowel, a consonant plus vowel, ん, or a small っ.

This is genuinely different from English, which stretches stressed syllables and squashes everything else. In Japanese, たまご is three beats of equal weight, not "ta-MA-go". Tap a steady pulse on the table and fit one kana to each tap. That mechanical, even delivery sounds far more natural than an English rhythm laid over Japanese words.

にほんご
nihongo
the Japanese language (four beats)

ともだち
tomodachi
friend (four beats)

しゃしん
shashin
photograph (three beats, because しゃ is one)

A long vowel is two beats

Doubling a vowel does not emphasise it, it holds it for two beats, and that length carries meaning. In hiragana the second beat is written with another vowel character; in katakana it is the bar ー.

おばさん
obasan
aunt

おばあさん
obāsan
grandmother

ゆき
yuki
snow

ゆうき
yūki
courage

Long o is usually written おう rather than おお, which is a spelling convention rather than a sound: おとうさん is otōsan, three long beats and a final ん, and the う is simply how the length is written. Long e behaves the same way, usually written えい, so せんせい is sensei in romanisation but sounds like a held e. The macrons over ō and ū in the romanised readings in this course always mean exactly this: hold the vowel for two beats.

The small っ is silence with a beat

A small っ before a consonant stops the airflow and holds it for one full beat before releasing. Romanised, it shows as a doubled consonant. Skip it and you say a different word.

さか
saka
a slope

さっか
sakka
a writer

おと
oto
sound

おっと
otto
husband

きって
kitte
postage stamp (three beats: き, っ, て)

The physical instruction is to close the mouth on the following consonant early and wait, as in the middle of the English phrase "hot tea" said carefully. The gap is real, measurable and audible, and native listeners rely on it.

ん is a full beat too

ん is never a mere tail on the vowel before it. It is its own beat, and it takes a whole one. しんぶん is four beats, not three, and it sounds wrong at three.

Its pronunciation shifts to suit what follows, without any change in spelling or timing. Before m, b or p it comes out as m, so しんぶん sounds like shimbun. Before k or g it is the ng of "singer". At the end of a word it is a nasal made at the back of the mouth, closer to French bon than to English "bun". None of this needs learning consciously, since the mouth does it by itself, but knowing it explains why the same character sounds different in different words.

Pitch, not stress

English marks a syllable by hitting it harder. Japanese marks it by putting it at a higher pitch, and the pattern is a property of the word. Every mora is either high or low, and once the pitch falls in a word it does not rise again.

Some pairs are separated by nothing else at all. Said with a high first beat and a low second, はし is chopsticks; with a low first and a high second it is a bridge. あめ is rain when it falls from high to low, and a sweet when it rises. かき is an oyster falling, a persimmon rising.

Do not try to memorise the pattern for every word. Do listen for it, copy what you hear, and keep your voice flat rather than adding English stress, which is a bigger giveaway than a wrong pitch. Context resolves nearly all of the ambiguity in practice: nobody eating dinner is confused about the chopsticks.

The vowels that vanish

Between two voiceless consonants, or at the end of a word after one, い and う often lose their voicing and are barely audible. This is regular and it is what makes recorded Japanese sound faster than the kana suggest.

すき
suki
liking, said very nearly as "ski"

した
shita
under, said very nearly as "shta"

です
desu
said as "des", with the u almost gone

The beat is still there, even when the vowel is not. Native speakers keep the timing and drop only the voicing, so copy that: shorten the vowel, but do not swallow the beat.

Vocabulary