Pitch Accent Minimal Pairs

English tells record (the noun) from record (the verb) by stress: RE-cord versus re-CORD. Japanese has no stress in that sense, but it does the same job with pitch — a controlled drop from a high mora to a low one. In a handful of famous cases, two words share every single sound and every kana, and the only thing separating them is where (or whether) the pitch falls. These are pitch-accent minimal pairs, and they are the cleanest proof that pitch is a real, meaning-bearing part of the word — not decoration you can skip.

This page collects the classic Tokyo-standard pairs and works each one out mora by mora. If you have not yet met the four accent shapes (平板 heiban, 頭高 atamadaka, 中高 nakadaka, 尾高 odaka), read Pitch Accent Patterns first — this page assumes you know them.

Notation used here

We mark each mora H (high) or L (low), and write at the exact point the pitch drops (the "downstep"). Two rules of Tokyo pitch do most of the work:

  1. The first two morae always differ in pitch — if mora 1 is low, mora 2 is high, and vice versa.
  2. Once the pitch falls, it stays down for the rest of the word and any particles attached to it.

Because the drop only "shows" when there is something after it, we test each word with the particle が attached. That particle is the litmus strip that reveals the hidden downstep.

雨 (rain) vs 飴 (candy) — a first-mora drop vs no drop

This is the pair every learner meets first. Both are read ame.

WordReadingPatternAlone
雨 (rain)あꜜめ頭高 atamadaka [1]H·Lあꜜめが H·L·L
飴 (candy)あめ平板 heiban [0]L·Hあめが L·H·H

雨 starts high and drops immediately ( is the peak, and everything after is low). 飴 starts low, climbs on , and stays up — the particle rides high too.

今日は朝から雨が降っている。

kyō wa asa kara ame ga futte iru

It's been raining since this morning.

子供の頃、よく飴をなめていた。

kodomo no koro, yoku ame o namete ita

When I was a kid, I used to suck on candy a lot.

💡
The particle test is your best friend. Say the word, add が, and listen to the が. If が sags to a low tone, the word had a downstep (雨が). If が stays bright and high, the word was heiban (飴が). You cannot hear the difference on a two-mora word said in isolation nearly as well as you can with a particle stuck on the end.

箸 / 橋 / 端 — the perfect triple

All three are read hashi, and together they show every possible position of the downstep in a two-mora word: on the first mora, on the second mora, and nowhere.

WordReadingPatternAlone
箸 (chopsticks)はꜜし頭高 [1] — drop after mora 1H·Lはꜜしが H·L·L
橋 (bridge)はし尾高 odaka [2] — drop after mora 2L·Hはしꜜが L·H·L
端 (edge)はし平板 [0] — no dropL·Hはしが L·H·H

Look at the "Alone" column and notice the trap: 橋 (bridge) and 端 (edge) are identical in isolation — both are simply L·H. They diverge only when a particle follows. 橋 hides a downstep after the word, so the particle drops (はしꜜが); 端 has no downstep at all, so the particle stays high (はしが). 箸, meanwhile, drops early and is distinct even on its own (H·L). This is exactly why native speakers rely on the particle, not the bare word, to hear the difference.

このお箸、ちょっと短くない?

kono o-hashi, chotto mijikaku nai?

Aren't these chopsticks a little short?

あの橋を渡ったら、すぐ駅だよ。

ano hashi o watattara, sugu eki da yo

Once you cross that bridge, the station is right there.

危ないから、道の端を歩いてね。

abunai kara, michi no hashi o aruite ne

It's dangerous, so walk along the edge of the road, okay?

花 (flower) vs 鼻 (nose) — the odaka/heiban pair

Both are hana, and like 橋/端 they sound the same alone (L·H) and split only on the particle. This pair is the textbook demonstration of 尾高 vs 平板.

WordReadingPattern
花 (flower)はな尾高 [2]はなꜜが L·H·L
鼻 (nose)はな平板 [0]はなが L·H·H

机の上にきれいな花が飾ってあった。

tsukue no ue ni kirei na hana ga kazatte atta

There was a pretty flower arranged on the desk.

花粉症で鼻がずっとムズムズする。

kafunshō de hana ga zutto muzumuzu suru

Because of hay fever, my nose keeps tickling.

神 (god) vs 髪 (hair) — and a word pitch can't save

Both are kami. Here the difference is real and audible:

WordReadingPattern
神 (god)かꜜみ頭高 [1]かꜜみが H·L·L
髪 (hair)かみ尾高 [2]かみꜜが L·H·L
紙 (paper)かみ尾高 [2]かみꜜが L·H·L

神 falls early (H·L); 髪 rises and drops only on the particle. But notice the third row: 紙 (paper) has exactly the same accent as 髪 (hair) — both are 尾高 [2]. Pitch distinguishes 神 from the other two, but it does nothing to separate 髪 from 紙. Those two really are homophones, and only context tells them apart.

困ったときだけ神に祈るなんて、勝手だよね。

komatta toki dake kami ni inoru nante, katte da yo ne

Praying to god only when you're in trouble — that's pretty self-serving, huh.

髪、切った?なんか雰囲気変わったね。

kami, kitta? nanka fun'iki kawatta ne

Did you cut your hair? Something about your vibe changed.

💡
Pitch is not a magic homophone-eliminator. Japanese still has thousands of true homophones (橋 and 箸 differ, but 髪 and 紙 don't). What pitch buys you is the removal of a large class of near-collisions, plus the split-second of clarity where context has not caught up yet — the moment before the listener knows whether you meant rain or candy.

Does context make pitch optional? No.

English speakers reason: "If someone says ame ga futte iru ('_ is falling'), obviously it's rain, not candy — so who cares about the pitch?" Context does usually rescue the meaning. But three things make pitch worth learning anyway:

  • The disambiguating half-second. The listener parses your sentence in real time. Wrong pitch on 雨/飴 forces a momentary garden-path — they briefly hear the wrong word and have to back up. Do that on every ambiguous word and your speech becomes tiring to follow.
  • It is the "correct" cue, not a fallback. To a native ear, pitch is simply part of the word, the way vowels are. Saying 箸 with 橋's melody is not "an accent" — it sounds like you reached for the wrong word, the way ship for sheep does in English.
  • Flat, accentless speech marks you as foreign even when every word is technically right. Pitch is a big part of what learners are hearing when they say someone "sounds native."

A note before you drill these

Everything above is Tokyo / standard (標準語) accent — what NHK, dictionaries, and this guide use. In Kansai many of these pairs are reversed: the melody Tokyo puts on 雨 is roughly the one Osaka puts on 飴. Do not try to blend the two systems; pick one (Tokyo, if you are learning from media) and be consistent. See Regional Pitch Accent for how, and why, the systems mirror each other.

Common Mistakes

Because the "error" in a minimal pair is a melody, not a spelling, each pair below shows the sentence said with the wrong pitch and then the fix.

飴が降ってきた。

ame ga futte kita

❌ Said with 飴's flat rise (L·H·H) but meaning 'rain' — you've literally said 'candy started falling.' Rain (雨) must drop early: あꜜめ, H·L.

飴をなめる。

ame o nameru

✅ Reverse it: for candy (飴) keep it heiban — あめ, L·H, particle high. The early drop belongs only to rain.

橋でご飯を食べる。

hashi de gohan o taberu

❌ Meaning 'eat with chopsticks' but said as odaka 橋 (drop on the particle) — that's 'bridge.' Chopsticks (箸) drop after mora 1: はꜜし, H·L.

花がかゆい。

hana ga kayui

❌ 'flower is itchy'? You wanted 鼻 (nose), which is heiban — はなが, L·H·H. 花 (flower) drops on the particle: はなꜜが, L·H·L.

髪を信じる。

kami o shinjiru

❌ 'believe in hair.' 神 (god) is atamadaka — かꜜみ, H·L. Only 髪/紙 are the low-start, particle-drop kind.

The pattern behind all five: decide the accent when you learn the word, not when you need it. File 雨 as "high-drop," 飴 as "flat," 橋 as "particle-drop," and you will never have to reconstruct the melody mid-sentence.

Key Takeaways

  • Minimal pairs prove pitch carries meaning: 雨/飴, 箸/橋/端, 花/鼻, 神/髪 differ by melody alone.
  • Test with が: a sagging particle means the word had a downstep; a bright particle means heiban.
  • 橋 and 端 (and 花/鼻) are identical alone and split only on the particle — 尾高 vs 平板.
  • 箸 vs 橋 vs 端 shows all three downstep positions (first mora, second mora, none) on the same sounds.
  • Pitch removes many collisions but not all — 髪 and 紙 stay homophones. Context still matters; pitch just gets you there faster and sounds native.
  • These are Tokyo accents; Kansai often reverses them.

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 PatternsN3Standard-Tokyo words fall into four accent types — heiban, atamadaka, nakadaka, odaka — and the particle が is what tells odaka and heiban apart.
  • Pitch Accent: What It IsN4Japanese words carry a fixed pattern of high and low beats with one possible 'drop' — it's melody, not English loudness, and the particle after a word reveals it.
  • Regional Pitch Accent (Tokyo vs Kansai)N2Pitch accent varies by region: the Kyoto–Osaka (Keihan) system adds a register dimension Tokyo lacks and often mirror-images Tokyo melodies, while some areas have no lexical accent at all — yet everyone understands the Tokyo standard this guide teaches.