Japanese has pitch accent (高低アクセント, kōtei akusento — literally "high-low accent"). Every word carries a fixed melody: each mora is pronounced either high or low, and in most words there is a single point — the downstep — after which the pitch falls and never rises again within that word. This melody is lexically fixed and is part of the word, exactly like its vowels and consonants. Two words spelled and segmented identically can mean completely different things depending only on where the pitch drops.
Most beginner courses ignore pitch accent entirely, and you can absolutely be understood without it. But pitch is the difference between a clear, native-sounding rhythm and the classic "textbook foreigner" accent — perfect vowels, perfect consonants, and a melody that is subtly, persistently wrong. The encouraging news: pitch accent is learnable, it follows tidy rules, and paying attention to it early is one of the highest-leverage things you can do for your accent.
Pitch, not stress — this is the core mental shift
English does not have pitch accent. English has stress: a stressed syllable is louder, longer, and clearer, while unstressed syllables get reduced (the a in "banana" becomes a mushy "uh"). English speakers reach for that toolkit automatically — they make one syllable loud and long and call it "the accent."
Japanese accent is none of those things. It is pitch — the musical highness or lowness of your voice (fundamental frequency), nothing more. In a Japanese word:
- Every mora is roughly the same loudness.
- Every mora is roughly the same length (Japanese timing is governed by the mora, covered in Pronunciation Overview).
- Vowels are never reduced — the a in a "low" mora is just as full as the a in a "high" mora.
The only thing that moves is the melody: your voice steps up or down between beats, like two notes on a piano. A "high" mora and a "low" mora differ the way two piano keys differ — in pitch, not in force.
The downstep: one drop, and it stays down
The heart of a Japanese word's melody is the downstep (also called the accent kernel, 核 kaku): the specific mora after which the pitch falls. Standard Tokyo Japanese obeys two ironclad rules:
- The first two morae always differ in pitch — a word starts either low-then-high or high-then-low, never high-high or low-low across its opening beats.
- Once the pitch falls, it never rises again inside the same word. There is at most one downstep.
So a word's melody is fully described by a single question: is there a downstep, and if so, after which mora? If there is no downstep at all, the word simply rises on the second mora and stays high — this is the accentless pattern. (The full four-way typology is on The Four Pitch Patterns.)
We write the downstep with a small drop mark ꜜ placed right after the accented mora. So neꜜko means "high on ne, then drop": ne(High) ko(Low).
猫がテーブルの上で寝ている。
neꜜko ga tēburu no ue de nete iru
The cat is sleeping on the table. (猫 neko drops after the first mora: HL.)
Minimal pairs: when melody is the only difference
The reason pitch accent is worth taking seriously is that it is phonemic — it distinguishes words. The textbook examples are pairs that are identical in every segment and differ only in pitch.
雨 vs 飴 — rain vs candy
| Word | Reading | Pitch (alone) | Meaning |
|---|---|---|---|
| 雨 | aꜜme | HL (drop after mora 1) | rain |
| 飴 | ame | LH (no drop, stays high) | candy |
雨だから傘を持っていって。
aꜜme dakara kasa o motte itte
It's raining, so take an umbrella. (雨 'rain' = HL.)
子供に飴をあげた。
kodomo ni ame o ageta
I gave the kid some candy. (飴 'candy' = LH, stays high.)
箸 vs 橋 vs 端 — the classic three-way set
The most famous example in all of Japanese phonetics is hashi, which is three different words:
| Word | Reading (alone) | With が | Pitch pattern | Meaning |
|---|---|---|---|---|
| 箸 | haꜜshi | haꜜshi ga (H L L) | drop after mora 1 | chopsticks |
| 橋 | hashiꜜ | hashiꜜ ga (L H L) | drop after mora 2 | bridge |
| 端 | hashi | hashi ga (L H H) | no drop | edge, end |
箸でご飯を食べる。
haꜜshi de gohan o taberu
I eat rice with chopsticks. (箸 = HL.)
橋を渡って駅に行く。
hashiꜜ o watatte eki ni iku
I cross the bridge to get to the station. (橋 = LH, drop lands on the next beat.)
The particle is the diagnostic
Look carefully at 橋 hashiꜜ and 端 hashi in the table above. Said alone, they sound identical — both are low-then-high, ha(L) shi(H). The downstep on 橋 is on its last mora, and a drop can only be heard when there is a following mora to drop onto. Since 橋 has nothing after it in isolation, the drop has nowhere to land.
Attach a particle — が, は, を — and the difference appears instantly:
- 橋が = hashiꜜ ga → ha(L) shi(H) ga(L) — the が drops.
- 端が = hashi ga → ha(L) shi(H) ga(H) — the が stays high.
橋がとても古い。
hashiꜜ ga totemo furui
The bridge is very old. (が drops → this is 橋 'bridge.')
端が少し破れている。
hashi ga sukoshi yaburete iru
The edge is torn a little. (が stays high → this is 端 'edge.')
This is why, throughout your study of Japanese pitch, the particle is the test. The melody of a word's final mora often only becomes audible on the beat that follows it, and that beat is usually a particle. When you look up a word's accent, you are really learning where its drop is relative to a following particle.
Do you have to learn this?
Honestly: you can communicate without it, and context rescues most misfires — nobody hearing 雨 vs 飴 in a real sentence is confused for long. So pitch accent is not a gatekeeper the way vowels and geminates are. But it is the difference between "understandable" and "natural," and — unlike some aspects of accent — it responds very well to conscious study. Because Tokyo pitch is rule-governed and mostly predictable from patterns, a little early attention pays off for years. Ignoring it, by contrast, is exactly what freezes many learners at the "fluent but unmistakably foreign" stage. Learn the framework now, add per-word accents gradually, and let your ear do the rest.
Common mistakes
Mistake 1: Substituting English stress for pitch. Making the accented mora louder and longer instead of higher. This is the master error from which most others follow.
❌ HA-shi(loud, long first syllable)
HA-shi (stressed, not pitched)
Incorrect — English stress on mora 1 instead of a pitch step.
✅ 箸を持ってきて。
haꜜshi o motte kite
Bring the chopsticks. (higher pitch on ha, then a clean drop — same volume throughout.)
Mistake 2: Treating pitch as optional / flattening everything. Speaking in a monotone with no melody at all. Every Japanese word has a shape; a flat delivery erases the LH rise that even the simplest word needs at its start.
Mistake 3: Lengthening the "accented" syllable. Because English stress adds length, learners stretch the high mora — turning 猫 neko into "nee-ko." Japanese morae stay even; only the pitch changes.
Mistake 4: Assuming pitch drops can rise back up. Adding a second peak inside one word (a "low-high-low-high" bounce). Tokyo words fall at most once and then stay low. There is no second rise.
Mistake 5: Thinking it's like Chinese tone. Japanese pitch is per word, described by one downstep — not a contour carried on each syllable. You are not singing a tone on every mora; you are placing a single step somewhere in the word (or nowhere).
Key takeaways
- Japanese accent is pitch (melody), not English stress (loudness/length); every mora is equal in force and length.
- A word has at most one downstep (ꜜ); after it, pitch stays low. The first two morae always differ.
- Accent is phonemic: 雨/飴, 箸/橋/端 differ only in melody.
- The particle reveals the accent — learn new nouns with が attached.
- Pitch is optional for being understood but essential for sounding natural, and it is very learnable — start early.
Now practice Japanese
Reading grammar gets you part of the way. The exercises are where it sticks — free, no signup needed.
Start learning Japanese→Related Topics
- The Four Pitch PatternsN3 — Standard-Tokyo words fall into four accent types — heiban, atamadaka, nakadaka, odaka — and the particle が is what tells odaka and heiban apart.
- Pitch Accent Minimal PairsN3 — Words spelled and segmented identically that Tokyo Japanese keeps apart by pitch alone — 雨 vs 飴, the 箸/橋/端 triple, 花 vs 鼻, 神 vs 髪 — and why context is not the whole story.
- Japanese Sounds: An OverviewN5 — The big picture of Japanese pronunciation for English speakers — five pure vowels, a small consonant set, mora-timed rhythm, and pitch instead of stress.