Japanese Mora Timing: The Rhythm That Makes You Sound Natural

A mora is a beat, not a syllable

English counts syllables, and English syllables come in wildly different sizes: strengths is one syllable and so is a. Japanese counts morae, and each mora takes roughly the same amount of time. The kana script makes this visible. Each full-size kana is one mora, so さくら sakura is three beats, and the only adjustment is that a small ゃ, ゅ, or ょ glues onto the kana before it, so きょ kyo is a single beat.

This is why Tōkyō is not a two-beat word. Written とうきょう, it has four morae, to-o-kyo-o, and a Japanese speaker gives all four their time. The English pronunciation with two stressed chunks, TO-kyo, sounds foreign not because of the vowels but because two beats have gone missing. Once you start hearing words as counts, the rhythm of the language becomes something you can practise deliberately.

Why English stress timing squashes Japanese

English is stress-timed. One syllable in each word gets louder, longer, and higher, and the syllables around it shrink to fit, often reducing their vowels to a vague uh sound. Say banana and notice how the first and last vowels almost vanish. That habit is automatic, and English speakers carry it straight into Japanese, where it does real damage.

Apply English stress to sakura and it becomes SAK-ra, a two-beat word with a lost vowel. Apply it to arigatō and the middle collapses while the final ō gets stretched into a stressed syllable. The vowels themselves may be fine; the timing has broken the word. The fix is not to speak slowly for ever but to distribute time evenly across the beats and remove the stress bump altogether.

The four things that take a full beat

Most beats are easy: a consonant plus a vowel, said once. The trouble comes from three sounds that English does not count as separate units and that Japanese does. Learn to feel these as full beats and the majority of your timing errors disappear at once.

• A kana: each consonant-vowel pair such as ka, shi, or tsu is one beat, and a kana with a small ya, yu, or yo attached (kya, shu, cho) is still one beat, not two.
• The second half of a long vowel: in ō, ū, and the ei of sensei the vowel is held for a second full beat, so obāsan has one beat more than obasan.
• The small tsu: in kitte or gakkō the mouth moves into position for the next consonant and holds it silently for a whole beat, a pause you can count even though nothing is voiced.
• The moraic n: ん is a beat of its own, so shinbun is four beats and hon is two; do not fold it into the vowel before it the way English folds the n of pin.

Each pair above is a genuinely different word, not a stylistic variant. Say them back to back, clapping, until the extra beat feels like a physical event: a held vowel, a silent hold, or a hummed n. The silent beat of the small tsu is the one most learners skip, because nothing is heard, but a Japanese listener hears the gap as clearly as any sound.

Counting morae in common words

Count the beats of a new word once, out loud, and you will rarely lose them again. The table lists everyday words with their beat count, the breakdown, and the place where learners most often steal a beat. Notice how often the culprit is a long vowel or a final n rather than a difficult consonant.

The clap drill

The clap drill trains the body to keep time so that your mouth does not have to think about it. Set a slow, steady pulse, around one clap per second at first, and say one mora per clap. Silent beats still get a clap: for kitte you clap ki, clap in silence while your tongue holds the t, then clap te. Long vowels get two claps with the same vowel sustained across both.

• Start with single words from the table and clap through each one three times without speeding up.
• Add a particle and keep clapping, because ga, o, and wa are beats too and often get swallowed.
• Move to the sentences below, counting the total beats first so you know what the sentence should feel like.
• Shadow a native recording of the same sentence while clapping; if your claps and the speaker's beats drift apart, you have found the beat you lose.
• Only when every clap lands do you raise the tempo, and never by more than a little at a time.

Why timing decides whether you are understood

Japanese listeners segment speech by beats before they process anything else. A word with the wrong number of beats does not sound like a slightly odd version of the intended word; it sounds like a different word, or like no word at all, and the listener has to stop and reconstruct what you meant. This is why a learner with imperfect vowels but steady timing is often understood more easily than one with beautiful vowels and a stress habit.

Devoiced vowels are a special case worth knowing. In desu, shimasu, and suki the u or i is whispered rather than voiced, and learners take this as permission to delete the beat. It is not. Desu is still two beats and suki is still two; the vowel is quiet but the time it occupies stays. Keep the beat and let the vowel fade, and the word sounds right.

How mora timing and pitch accent fit together

Pitch accent is assigned beat by beat, which is the simplest reason to fix timing first. A long vowel is two morae, so it can be high on the first and low on the second: jū is JU-u, with the drop inside the vowel. The moraic n is a mora, so it can carry the high pitch, as in se-N-SE-e for sensei, where the final held e is low. If you shorten the vowel or swallow the n, the pitch drop has nowhere to land and the word comes out wrong twice over.

Pitch patterns themselves are covered in the pitch accent guides on this blog; here the point is that the beat count is the grid the melody is drawn on. Practise the count until it is automatic, then add the drop, and you will find that many pitch mistakes were timing mistakes in disguise.

Speeding up without collapsing beats

Natural speed is not achieved by dropping beats. Native speakers compress every mora evenly, so a fast sentence keeps the same proportions as a slow one: a long vowel is still twice a short one and the small tsu is still a full hold. Learners speed up the way English does, by keeping the stressed beats and crushing the rest, and the special morae are always the first to go.

The remedy is to treat speed as a tempo setting rather than a shortcut. Take a sentence you can clap perfectly at one beat per second, then raise the tempo in small steps, checking after each step that the long vowels, silent holds, and n sounds are still there. If one disappears, drop back a step and rebuild. Three tempos, slow, medium, and natural, are enough for any sentence.

• Record the slow version first, because you will compare every faster version against it.
• Speed up whole sentences, not single words, so the particles and endings get compressed together with everything else.
• Listen to a native model at natural speed and count its beats; the count never changes, only the tempo does.
• When a long sentence falls apart, shorten it to the first four or five beats, fix those, and add the rest back a chunk at a time.

Check yourself: record, count, and compare

Your own ear forgives missing beats because it knows which word you meant. A recording does not. Record the clap-drill sentences without clapping, play them back, and tap a finger for every beat you hear. Where your taps come up short against the count you wrote down, you have found the beat that goes missing when you stop concentrating, and that is the one to drill for the rest of the week.

Then compare with a native model of the same sentence: a long vowel or a silent hold that is visibly shorter in your recording is the fix. Finally, speak the lines in a live conversation with your AI Sensei in Unihongo's immersive 3D classroom, where the pronunciation coaching points out the words whose length drifted and the session report gives you the corrected versions to clap through again.

Read each check line at the three tempos and count the beats every time. If the count holds at natural speed, the rhythm is yours; if it slips, you know exactly which beat to take back to the clap drill.