Geminate Consonants (っ)

Japanese has a sound that English speakers systematically underestimate: the geminate consonant, or sokuon（促音）, written with a small っ. It looks like a doubled letter — きて kite versus きって kitte — but it is not a spelling convention. It is a full mora of held silence, a genuine beat of time that you must produce with your mouth and your clock, not just your pen. Getting it wrong is one of the single most audible markers of a non-native accent, and getting it right is one of the fastest ways to sound more Japanese.

This page covers what the geminate actually is physically, where you hear it most clearly, and why the number-one beginner mistake — not holding it long enough — collapses the meaning of the word entirely.

What the small っ actually is

Standard Japanese is built out of morae — equal-length rhythmic beats. か is one mora, かん is two, かんこく is three. The small っ is one of those beats too, but it is a beat with no sound of its own. Instead, it is a timed pause in which your mouth gets into position for the following consonant and holds there.

Concretely, when you see っ before a stop consonant (k, t, p), your vocal tract closes for the coming consonant and simply stays shut for one full mora before releasing. In きって kitte, your tongue moves to the [t] position after ki, stays sealed for a beat of silence, and only then releases into te. The word is three morae long — き・っ・て — and that middle silent beat is as long as either spoken beat around it.

Before a fricative (s, sh), there is no silence — instead the [s] or [ʃ] hiss is prolonged for the extra beat. ざっし zasshi (雑誌, "magazine") is za followed by a long, sustained sh and then i: za・っ・し. You can hear the hiss stretch. Before affricates (ts, ch), you get a brief hold and then the burst.

Where you hear it most: k, t, p, s

The geminate is clearest — and matters most for comprehension — before the voiceless consonants k, t, p, s (and their kana partners ch, ts, sh). In native Japanese words the geminate essentially only appears before these voiceless sounds, which is why the silent-hold strategy works so reliably.

That short, everyday sentence has the geminate twice — chotto and matte — and a native speaker holds both. Rushing through them is exactly what makes a beginner sound rushed and flat.

Minimal pairs: the hold changes the word

Because the held beat is phonemic, adding or dropping it turns one word into a completely different one. These are true minimal pairs — identical except for the geminate:

That sentence deliberately holds kite (no hold) and kitte (one-beat hold) back to back, plus hatte (貼って, "stick on") — the length difference is the only thing keeping "come" and "stamp" apart.

It is a full mora — count it

The geminate does not just add "a little extra length." It adds an entire mora, which reshapes the rhythm of the whole word. がっこう gakkō（学校, "school"）is four morae: が・っ・こ・う (and う here lengthens the o — see Long Vowels). If you want the underlying rhythmic unit explained in full, see Mora vs. Syllable; for how the small っ is written and typed, see The Small っ (Sokuon).

Not like English "bookkeeper"

Here is the crucial contrast for English speakers. In English, doubled letters are almost pure orthography: the kk in bookkeeper, the tt in butter, the pp in happy are all pronounced as a single consonant. You do not hold them. English has no phonemic length distinction in consonants at all — hoping and hopping differ in the vowel, not in how long the p lasts.

Japanese is the opposite. The doubled consonant is a real, timeable event: your articulators genuinely stay in place for an extra beat. When an English speaker imports the "doubled letters are silent spelling" habit, きって kitte comes out as きて kite, and you have just asked the post office clerk to "come" instead of asking for "stamps."

Common mistakes

Mistake 1: Under-holding — collapsing っ into a single consonant. This is the big one.

Mistake 2: Slipping a vowel into the gap. English speakers uncomfortable with silence tend to fill it — turning kitte into "kit-o-te" or matte into "mat-uh-te." The geminate is silence, not a hidden vowel. Keep the tract sealed with no voice.

Mistake 3: Not prolonging fricative geminates. Before s/sh there is no silence — you must stretch the hiss. Learners often drop the hold entirely here because there is no "gap" to cue them.

Mistake 4: Inventing a geminate that isn't there. Over-correction runs the other way: once you notice geminates, you may start doubling consonants that should be single, e.g. saying itte for いて ite ("be/exist," te-form of いる). Only hold where a small っ is actually written.

Mistake 5: Forgetting the geminate counts as a beat when reckoning rhythm. In songs, poetry, and natural speech timing, いっぱい ippai is four morae (い・っ・ぱ・い), not three. Shortchanging the っ throws off the whole phrase's rhythm.

Key takeaways

• The small っ is one full mora — a real beat, not silent spelling.
• Before k, t, p: a beat of silent closure. Before s, sh: a prolonged hiss. Before ch, ts: a brief hold then a burst.
• The distinction is phonemic: きて/きって, 坂/作家, 音/夫 differ only in the hold.
• English gives you no practice with consonant length, so your reflex is to under-hold — deliberately over-hold until it feels natural.