Mandarin tones, explained for people who think they're tone-deaf

If you've ever said "I'm tone-deaf, Mandarin isn't for me", here's some good news: you already use pitch all the time. You raise your voice at the end of "Really?", you drop it at the end of "Fine." when you're annoyed, and you can tell a question from a statement without thinking. Mandarin tones use the same equipment. The difference is that in Mandarin the pitch shape belongs to each syllable, and changing it changes the word.

This guide explains the tones in plain language, the few rules that change them, and the practice that actually helps.

The four tones, plus one

Every Mandarin syllable has a tone: a pitch shape. The classic example is the syllable ma, which becomes four different words. Tap each one and listen to the shape, not the sound.

A few things the classic tone diagrams don't tell you:

The first tone is high, not loud. Pick a comfortable high note and stay on it.
The second tone rises like a question, from your middle range up.
The third tone is mostly just low. Textbooks draw it dipping down and back up, but in normal speech, in the middle of a sentence, it usually stays low and creaky. The full dip-and-rise mostly appears when the word is on its own or at the end of a sentence.
The fourth tone falls sharply, from high to low, and it's short. Think of it as decisive.
The neutral tone has no shape of its own. It's light and quick, and its pitch depends on the tone before it. It's marked in zhuyin with a dot before the syllable, ˙ㄇㄚ, and in pinyin by leaving the tone mark off.

In zhuyin the tones are marked after the syllable: nothing for the first, ˊ for the second, ˇ for the third, ˋ for the fourth. In pinyin the marks sit over the vowel: mā, má, mǎ, mà.

The tone changes everyone trips on

A few common words change tone depending on what follows. You don't need to memorise a chart; three rules cover most of what you'll hear (AllSet Learning's pronunciation wiki has a fuller list).

Two third tones in a row: the first becomes a second tone. That's why 你好 is said ní hǎo, not nǐ hǎo.

不 (bù, "not") becomes bú before a fourth tone:

一 (yī, "one") stays yī when you count, but becomes yí before a fourth tone and yì before the other tones:

Dictionaries and most textbooks write 一 and 不 with their original tones and expect you to apply the change. Lin says them the way people actually do, which is the version worth copying.

Say it out loud with Lin

Bopomo is an iPhone app for speaking Mandarin with Lin, your AI tutor, on a trip around Taiwan. The first 500 on the waitlist get the founding price. Join the waitlist.

Why tones feel hard, and what actually helps

Tones feel hard for two reasons that have nothing to do with musical talent. First, in English, pitch carries mood, not meaning, so your brain has learned to ignore it when it's working out which word you heard. Second, when you're concentrating on new words, pitch is the first thing to go flat. Both get better with the right kind of practice.

Learn words with their tones from day one. If you learn 馬 as "ma, horse" and add the tone later, you'll be fixing it for months. Learn it as "mǎ, low".

Say things out loud, in full phrases. Tones in real speech are shaped by their neighbours. Practising 謝謝 or 不客氣 as whole phrases teaches the rhythm along with the tones.

Exaggerate, then relax. Overdo the shapes at first: very high, very rising, very low, very falling. Your natural speech will settle into something clear.

Use your hand. Drawing the tone in the air while you say it (flat, up, low, down) sounds silly and works, because it gives your voice something to follow.

Record yourself and compare. Play Lin's clip, record yourself saying the same thing, and listen to both. Hearing your own pitch from the outside is the fastest way to notice where it goes flat.

Pairs beat single syllables. Practise two-syllable words, which cover every combination of tones: 謝謝, 你好, 好吃, 早安. They're much closer to real speech than ma, má, mǎ, mà.

Don't panic about mistakes. Context does a lot of the work. If you say 我要一杯奶茶 at a drink stand with a slightly wrong tone, people will still hand you a milk tea.

What an app can and can't check

It's worth being honest about this, because a lot of apps aren't.

Most language apps, Bopomo included, use the phone's speech recognition to turn what you say into text and check whether it matches what you meant to say. That's genuinely useful: if your 謝謝 comes out as 謝謝, you were understood, which is what matters in a real conversation. But it isn't the same as grading each tone. Speech recognition is built to guess the right words even when the pronunciation is off, so it can forgive a wrong tone that a person would notice, especially in common phrases.

Here's how Bopomo handles it today. Lin plays every line in a Taiwanese voice, the shape of each syllable's tone is drawn above it so you can see the melody you're aiming for, and your attempt is checked for whether it was understood. Bopomo doesn't yet score each tone on its own: we've been testing that, and in our own tests it wasn't accurate enough to ship. When it is, it'll say so. Until then, recording yourself and comparing with Lin is still the best tone check there is.

A five-minute daily routine

Listen to five two-syllable words, one at a time, drawing the tones in the air.
Say each one three times out loud, exaggerating the shapes.
Record one short phrase and compare it with the original.
Use one phrase for real that day, even if it's just 謝謝 at the shop.

Do it most days and within a few weeks the tones stop being something you think about and start being part of the word.

Questions

How many tones does Mandarin have?

Four main tones plus a neutral tone: high and level, rising, low, falling, and a short, light neutral tone.

Can you learn Mandarin if you're tone-deaf?

Almost certainly. Most people who call themselves tone-deaf can hear the difference between a question and a statement, and that's the same skill. Regular listening and speaking practice is what makes tones easier.

Why does 你好 sound like "ní hǎo"?

When two third tones come together, the first one changes to a second tone. It's a regular rule, and everyone says it that way.

Do tones matter if people understand me from context?

Context helps a lot, especially with common phrases. But clear tones make you much easier to understand, especially with words you'd use less predictably.

Are tones different in Taiwan?

The four tones are the same. Taiwan Mandarin uses the neutral tone a little less than Beijing Mandarin, so some second syllables keep their full tone.