The four pitch patterns, the mistakes most learners make, and how to actually train your ear and voice to fix them.
Feb 26, 202610 min read
Link Copied to Clipboard!
The Four Pitch Accent Patterns at a Glance
Every word in standard Japanese fits one of these four shapes
Odaka looks identical to heiban in isolation — the drop only becomes audible when a particle follows the word.
Introduction
You've been studying Japanese for months. Grammar is clicking. Your vocabulary has hit a few hundred words. You can read hiragana without thinking about it. Then you actually open your mouth in front of a native speaker, and they tilt their head slightly — not because they didn't hear you, but because something sounds off and they can't quite place why.
The vocabulary was right. The grammar was right. The problem was the pitch.
Japanese pitch accent is the part of pronunciation that most textbooks skip entirely, which means most learners skip it too, and then spend years wondering why they sound like they're reading out loud rather than speaking. This guide covers what pitch accent actually is, the four patterns you need to know, and how to start training for it without making it a second full-time job.
What Pitch Accent Actually Is
Japanese pitch accent — kotei akusento (高低アクセント) — is a system where the relative highness or lowness of your voice on a given syllable carries information. Not in the sweeping tonal way of Mandarin, where pitch determines word meaning entirely. Japanese is more subtle. But subtle doesn't mean unimportant.
In English, stress does the heavy lifting. You say CONtract (noun) versus conTRACT (verb) by putting more force on a syllable. Japanese doesn't work that way. Syllables aren't louder or softer — they're higher or lower. The pattern of highs and lows across a word is what pitch accent describes.
The hashi problem
The classic example: hashi (はし). Depending on your pitch, you're saying three completely different things.
HA-shiHigh-Low → 橋 (bridge)
ha-SHILow-High → 箸 (chopsticks)
ha-shiFlat → 端 (edge)
In context, native speakers almost always know which you mean. Out of context — or at the start of a conversation before they've calibrated — getting this wrong causes a small but real beat of confusion.
True minimal pair confusion is relatively rare in practice. Context covers most of it. But the cumulative effect of consistently wrong pitch is what makes a learner sound robotic or foreign even when the words are correct. That's the problem worth fixing.
Key Terms
These terms come up in dictionaries, learner forums, and resources like the OJAD database. Here's what they actually mean.
Mora (拍)
The basic unit of Japanese rhythm — roughly one kana character. Pitch accent is measured in mora, not syllables. To-kyo is three mora (to-o-kyo) because the long vowel counts separately.
Pitch Drop
The point where pitch falls from high to low. The location of this drop — or whether one exists at all — defines each pattern. Dictionaries mark it with a number in brackets: [0] means no drop; [1] drops after the first mora.
Standard Accent (共通語)
The Tokyo-based accent used in NHK broadcasts and formal speech — what most learners study. Kansai, Kyushu, and other regional dialects have their own distinct systems.
Shadowing
Repeating what you hear immediately after a native speaker, overlapping with their speech. One of the more effective techniques for pitch training because it forces your mouth to match your ear without analytical interference.
Minimal Pair
Two words identical except for pitch pattern — like hashi (bridge) and hashi (chopsticks). Useful for training the ear to hear pitch differences that disappear when you're focused on meaning.
OJAD
Online Japanese Accent Dictionary — a free, university-maintained database with pitch patterns and audio for tens of thousands of words, including verb conjugations and compound forms.
The Four Patterns
Standard Japanese organizes every word into one of four patterns. That's the good news — four patterns, not forty. The patterns below use H (high) and L (low) across mora.
1
Heiban (平板) — Flat
L H H H H...
Starts low on the first mora, rises on the second, stays high for the rest of the word — and for any following particle. No drop anywhere. The most common pattern in Japanese. If you're not sure what pattern a word uses, heiban is statistically the safest guess.
Examples
sakana 魚 — fish: sa-KA-NA
tamago 卵 — egg: ta-MA-GO
Dictionary notation: [0]
2
Atamadaka (頭高) — Head-High
H L L L...
Starts high and drops immediately on the second mora. The pitch peak is at the very beginning. Everything after the first mora stays low. Common in short two-mora words — hit the first mora high and stay down.
Examples
ame 雨 — rain: A-me
hashi 橋 — bridge: HA-shi
Dictionary notation: [1]
3
Nakadaka (中高) — Middle-High
L H...H L L
Rises in the middle of the word, then falls before the end. The drop happens inside the word itself — before any particle. More common in three- and four-mora words. The exact drop position varies, which is why the dictionary number matters.
Examples
atama 頭 — head: a-TA-ma
ikebana 生け花: i-KE-ba-na
Dictionary notation: [2], [3], etc.
4
Odaka (尾高) — Tail-High
L H H (particle: L)
Starts low, rises, and stays high — but drops on any following particle. In isolation it sounds almost identical to heiban. The distinction only becomes audible in a sentence: otoko-ga has an audible drop; heiban words don't.
Examples
otoko 男 — man: o-TO-KO(-ga↓)
hashi 箸 — chopsticks: ha-SHI(-ga↓)
Dictionary notation: [n] where n = mora count
Reading dictionary notation: [0] = heiban (no drop). [1] = atamadaka (drops after the first mora). Higher numbers indicate where in the word the pitch falls. This system is consistent across Jisho, OJAD, and most modern learner dictionaries.
Tokyo vs. Kansai: Pick One and Stay There
If you've spent time listening to Japanese from different regions, something sounds off between a Tokyo speaker and an Osaka speaker even when they're using identical words. That's regional pitch accent — and Kansai (Osaka, Kyoto, Kobe) is the most distinct variant you'll encounter.
It's not just a different accent — it's a fully separate pitch system. Words that are atamadaka in Tokyo can be heiban in Kansai and vice versa. Kirei (beautiful) is Low-High in Tokyo; in Kansai it may be High-Low or flat depending on the city and speaker.
Tokyo (Standard)
More defined high/low distinction. Used in NHK broadcasts, national media, and formal speech. Understood everywhere in Japan. What most learners study and what the four-pattern system describes.
Kansai (Osaka/Kyoto)
A fully separate pitch system with strong regional identity. Patterns often inverted from Tokyo. Mixing the two is immediately noticeable and sounds natural in neither direction.
For most learners, Tokyo/standard pitch accent is the right starting point — neutral, formal, and understood everywhere. Once you have it, you can pick up regional variation through exposure if you spend real time in a specific area. The main thing to avoid early on is mixing the two systems.
What Actually Goes Wrong
Most pitch accent mistakes fall into a small number of categories. Knowing them makes them easier to catch in your own speech.
①
Flat intonation
The most common issue for English speakers. When you haven't been trained to think about pitch, you produce everything at roughly the same level — no rise, no drop, just monotone. The single biggest contributor to sounding foreign, and also the hardest to self-diagnose because it sounds completely normal to you while you're doing it.
②
English stress patterns bleeding in
English builds emphasis by making syllables louder and longer. Japanese doesn't do that. When an English speaker says tabemasu and puts extra energy on the "be," they're applying English stress logic to a language that doesn't use it. The result sounds like something between English and Japanese — which is its own kind of hard to understand.
③
Over-correcting after learning the rules
This happens a lot after learners discover pitch accent. They know it exists, they understand the patterns, and then they perform them — deliberately, audibly, slightly theatrically. The goal is for pitch to become automatic, not to layer a pitch performance on top of normal speech. Overcorrection produces its own kind of unnatural sound.
④
Memorizing patterns without hearing them
You can learn every rule for all four patterns from written descriptions. You can pass a quiz about them. And then go listen to a native speaker and not recognize any of it, because what you memorized was an abstraction and what you're hearing is real speech. Reading about pitch accent and listening to pitch accent are not the same activity.
Compound Words: Where It Gets Complicated
Individual word pitch accent is learnable. Compound words are where the system starts to feel genuinely hard.
When two words combine in Japanese, their pitch patterns interact and produce a new pattern that isn't always predictable from the components. Tegami (手紙 — letter) combines te (hand) and kami (paper), each with its own standalone pitch. In the compound, a different set of rules takes over.
The practical approach
The general principle is that the first element in a compound tends to dominate the overall pitch shape, but there are enough exceptions that you can't rely on it. Look up compound words directly in OJAD rather than trying to derive pitch from components. Let the pattern emerge through repeated exposure. This is one of the reasons pitch accent mastery takes years rather than months — the individual patterns are learnable in weeks, but compound word behavior builds slowly through continuous listening and use.
How to Actually Train It
Understanding pitch accent and producing it consistently are different problems. Here's what actually works, in the order it works best.
Train your ear before your mouth
Deliberate Listening
Before you practice producing pitch, practice hearing it. Take a word you know, look up its pitch pattern in OJAD, find audio, and listen to the pattern rather than the meaning. Do this until the shape of the sound feels distinct from other patterns.
Minimal Pair Drills
Five to ten minutes of focused minimal pair listening is more useful than an hour of passive audio. The Pitch Accent Lab has quizzes organized by pattern type for exactly this purpose.
Shadow before you produce
Shadowing — repeating what you hear immediately after a native speaker, overlapping with their speech — bypasses the analytical layer and forces your mouth to match what your ear just processed. Shadow at the sentence level, not the word level. Pick a short clip of one or two sentences and repeat until the rhythm feels less forced. Then move on.
Record yourself
This is the step most people skip. Record yourself saying a word or sentence, then compare to native audio. The gap will be obvious. Without this feedback loop, you can practice for months without knowing whether you're making progress or just reinforcing the same mistakes more confidently. A voice memo next to a native audio clip is sufficient.
Start with high-frequency words
Don't start with obscure vocabulary. Start with what you'll use every day: arigatou, sumimasen, kudasai, wakarimasen. Pitch patterns on uncommon vocabulary matter much less — context covers most of the ambiguity in practice.
A realistic 15-minute daily block
Min 1–5
Listening
Take 5–8 words from a dictionary with audio. Look up their pitch notation, listen to each word twice while tracking the H/L pattern. Don't try to repeat yet.
Min 6–10
Shadowing
Pick two or three sentences from a clip you've been working with. Repeat them overlapping with the audio until the rhythm feels less forced.
Min 11–15
Recording
Say three words or one sentence, record it, compare to native audio. Note what's off. You don't need to fix it today — just observe it. That's your data for tomorrow.
Keeping It in Perspective
Pitch accent has a tendency to become an obsession for intermediate learners — the thing they worry about too much while other skills atrophy. You will meet people who have spent two years perfecting their atamadaka and still freeze when someone speaks to them at normal speed. You'll also meet people with noticeably imperfect pitch who communicate fluently. The latter group is more useful to emulate.
What matters more
Politeness levels, vocabulary range, and basic grammatical fluency affect real interactions more than pitch accuracy — especially in the first two years.
What native speakers notice
Native speakers are not auditing your pitch. They're trying to understand your meaning — and often surprised (positively) that you're speaking Japanese at all.
The realistic goal
Develop enough pitch awareness that you're not producing uniformly flat intonation. That moves you from sounding foreign to sounding like a learner. Beyond that, improvement is gradual.
You'll still sound like a learner for a while. Pitch accent doesn't change that. It just makes the learner sound slightly more natural, which is worth pursuing without being worth obsessing over. Perfect pitch with halting grammar is not a winning combination.
FAQ
Do I actually need to study pitch accent?
It depends on what you're trying to do. For survival Japanese on a short trip, you can skip it. For sustained conversation that sounds natural and doesn't create constant comprehension friction, yes. But it doesn't need to be your top priority until your grammar and vocabulary are reasonably solid. Get those right first.
How long does it take to get good at this?
Most learners develop a working intuition for pitch within a year of consistent, mindful listening. Producing it reliably in spontaneous speech takes longer — usually somewhere in the second or third year, if you're actively working on it. The four patterns themselves are learnable in a matter of weeks. Applying them automatically in real conversation is the slow part.
Will Japanese people judge me for getting it wrong?
No, and that's not just an encouraging platitude — it's how most interactions actually go. Japanese speakers are not running pitch audits on your sentences. Don't let pitch anxiety become a reason to avoid speaking. Avoidance is a much bigger problem than imperfect intonation.
Should I mark pitch patterns on my flashcards?
If you're already making vocabulary flashcards, adding the pitch notation (the bracketed number from the dictionary) takes about three seconds per card and builds a useful habit over time. Don't restructure your entire study process around it, but if the infrastructure is already there, it's worth doing.
What's the best resource for looking up pitch patterns?
The OJAD (Online Japanese Accent Dictionary) is the most useful reference — free, university-maintained, covers verb conjugations and compound words. For minimal pair practice and pattern recognition quizzes, the Pitch Accent Lab on this site is a good starting point.
Conclusion
The four patterns — heiban, atamadaka, nakadaka, odaka — are the core of Japanese pitch accent, and they're learnable. The dictionary notation system gives you a consistent way to look up any word. The training approach (ear before mouth, shadowing before production, recording for feedback) gives you a method that actually transfers to real speech rather than just abstract knowledge.
Flat intonation is the main thing to fix. Not perfect atamadaka on every word — just enough pitch variation that you don't sound like you're reading from a screen. That threshold is reachable within months of consistent work. Everything beyond it is a long, gradual process that happens mostly through listening and use.
Final Thought
The learners who make the most progress with pitch accent tend to treat it the way they treat kanji: not as a separate subject to master before moving on, but as one layer of the language that gets a little better every week through continuous exposure. Reviews in the morning, then real Japanese for the rest of the time. That balance is sustainable. Treating pitch accent as a prerequisite to speaking is not.
Tools Worth Using Alongside This Guide
For drilling pitch patterns, practicing kana, and getting speaking feedback — the parts you have to do, not just read about.
Practice Pitch
Pitch Accent Lab
Interactive quizzes for minimal pairs, pattern recognition, and verb conjugation pitch. Start here to find out what you can actually hear.