Long vs short vowels in Japanese: why おばさん is not おばあさん

Japanese counts beats, not syllables

Japanese runs on the mora: a unit of time, roughly one kana long, and every mora gets the same amount of it. Say the word like a metronome and you are already most of the way to sounding right.

A long vowel is simply the same vowel occupying two of those beats. とうきょう is not two syllables, To-kyo. It is four beats: to-o-kyo-o. The second beat of each pair is the vowel continuing, not a new sound.

This is a timing feature, not a musical one. You are holding a vowel longer, not raising or lowering your voice. Japanese does have a separate pitch layer, but it is a different system from vowel length and the two should not be confused while you practice.

How length is written

In hiragana, a long vowel is written by adding the vowel kana: ああ, いい, うう, ええ. For the o sound, the standard spelling is おう rather than おお in most words — こうこう, とうきょう, ありがとう — even though the pronunciation is a held o, not an o gliding into a u. A handful of words keep おお: とおい, おおきい.

For え, the usual long spelling is えい, which most speakers pronounce as a held e: せんせい comes out closer to sensee than sensei.

In katakana, length is written with a bar: コーヒー, ケーキ, ビール. The bar is easier to read than the hiragana rules, and it makes borrowed words a good place to start drilling.

Romanization is the least reliable guide. Some systems write ō, some ou, some nothing at all — and "Tokyo" written the English way hides two long vowels.

How to hold a vowel

Set a beat. Tap your finger at a steady, slow tempo and say one kana per tap. Then say a long vowel across two taps without changing anything about your mouth: same tongue position, same lip shape, same volume, just twice as long.

The most common error is not shortening the vowel but changing it. English speakers turn a long o into a glide — "oh-oo" — because English long vowels are diphthongs. Japanese long vowels are pure. Freeze your mouth and let the sound run.

The second most common error is stressing the long vowel to make it feel longer. Do not push. Length in Japanese is duration alone; adding loudness makes it sound like emphasis instead of a different word.

Start at half speed. At half speed you can feel two beats. Then bring it up gradually, and check that the long vowel stays twice as long when the word gets fast.

Words to practice

Start with 東京 and 高校 — both are four even beats, both are common, and both are pronounced as two long vowels rather than two syllables. Say them against a tap: to-o-kyo-o, ko-o-ko-o.

Then ありがとう, five beats with the long vowel at the end, where learners usually cut it short.

Then the pairs where length carries the meaning. おばさん / おばあさん is the classic. Also ここ / こうこう, ゆき / ゆうき, おじさん / おじいさん. Say each pair back to back and make sure you can hear the difference yourself.

And take one katakana word seriously: コーヒー is ko-o-hi-i, four even beats with two long vowels and a Japanese ひ. It looks like a freebie and it is a trap.

What to listen for

Listen for evenness first. If any beat in the word is noticeably shorter than the others, your English stress habit is squeezing it. Japanese does not squeeze.

Listen for purity in the held vowel. If you hear the vowel move while you hold it, you have made an English diphthong and it will sound foreign no matter how long you held it.

Record a long-vowel word and a short-vowel word back to back and play them without looking. If you cannot tell which is which, neither can anyone else — and this is one of the few pronunciation problems where a recording gives you a completely reliable verdict.

TonePerfect scores every word you record mora by mora, so a rushed long vowel shows up in that word's score, and your sounds come back ranked weakest-first. That ranking is what tells you what to drill. It is not an accuracy percentage and not a JLPT level, and it does not reflect pitch accent — the app makes no pitch-accent claims at all.

Questions