Chinese Tones Explained: All 5 Tones + Change Rules
Mandarin has four tones plus a neutral tone, and the tone is part of the word itself: mā (妈 "mother") and mà (骂 "to scold") share nothing but a syllable. The four tones are high-flat, rising, low-dipping, and falling. Below you'll find each one with its pitch contour, minimal pairs to train your ear on — and the part most tone guides skim: the tone change rules that make real Mandarin sound different from dictionary pinyin.
- Four tones + neutral: high-flat (mā), rising (má), low-dipping (mǎ), falling (mà), plus short unstressed (ma).
- The third tone is usually just LOW in real speech — the full dip-and-rise is the exception, not the rule.
- Change rules matter: 你好 is said ní hǎo (3rd+3rd → 2nd+3rd), 不是 is bú shì, 一个 is yí ge.
- Drill tones in two-syllable words and sentences, not isolated syllables.
- Taiwan Mandarin uses the same four tones but far fewer neutral tones, and a few words carry different official tones.
What are the four Mandarin tones?
Each tone is a pitch shape, not a fixed note — linguists describe them with Chao numbers on a 1 (low) to 5 (high) scale. Here are all four plus the neutral tone on the classic syllable "ma," which happens to mean five different things:
| Tone | Contour | Pitch | How it sounds | Example |
|---|---|---|---|---|
| 1st — flat | 55 (high, level) | Sung, held high and steady — like a doctor's "say aah" | mā 妈 (媽) mother | |
| 2nd — rising | 35 (mid → high) | Rises like an English question: "what?" | má 麻 hemp; numb | |
| 3rd — low / dipping | 214 (low dip, rise at the end) | Drops to the bottom of your range; the final rise only appears in isolation | mǎ 马 (馬) horse | |
| 4th — falling | 51 (high → low) | Sharp and final, like an emphatic "no!" | mà 骂 (罵) to scold | |
| Neutral | short, unstressed | Quick and light; its pitch rides on the syllable before it | ma 吗 (嗎) question particle |
In pinyin the tone is the mark over the vowel (ā á ǎ à, nothing for neutral); in zhuyin, first tone is unmarked and ˊ ˇ ˋ ˙ mark the rest. If either system is new to you, start with our guides to how to read pinyin and how to read zhuyin — everything below uses pinyin marks.
Minimal pairs: hearing the difference
The fastest way to convince your brain that tones carry meaning is minimal pairs — same syllable, different tone, unrelated words:
| 1st | 2nd | 3rd | 4th |
|---|---|---|---|
| mā 妈 mother | má 麻 numb | mǎ 马 horse | mà 骂 scold |
| tāng 汤 (湯) soup | táng 糖 sugar | tǎng 躺 lie down | tàng 烫 (燙) scalding |
| — | shí 十 ten | mǎi 买 (買) buy | shì 是 to be · mài 卖 (賣) sell |
买 mǎi "buy" versus 卖 mài "sell" is the pair that keeps learners honest: get the tone wrong in a shop and you've reversed the transaction. Don't just read these — find audio and listen until the difference stops feeling subtle.
The neutral tone: what it actually does
The neutral tone (轻声 qīngshēng, 輕聲) isn't a fifth pitch shape — it's the absence of one. The syllable is short, light, and unstressed, and its actual pitch is decided by the tone before it: high-ish after a 3rd tone, low after a 4th. You'll meet it in three places: grammar particles (的, 了, 吗, 呢), the second syllable of many everyday words (妈妈 māma, 桌子 zhuōzi, 我们 wǒmen, 朋友 péngyou), and reduplicated verbs (看看 kànkan). Treat it as a real feature to imitate, not a syllable you're allowed to mumble — a learner who stresses every syllable equally sounds robotic in Mainland Mandarin. (Though, as we'll see below, Taiwan does exactly this on purpose.)
Tone change rules (sandhi): how real speech differs from the dictionary
Dictionary tones aren't always the tones people say. A handful of automatic rules — linguists call them tone sandhi — rewrite tones in connected speech. Pinyin is conventionally written with the original tones, so you have to know these rules to match what your ears hear. There are only four that matter:
Rule 1: 3rd + 3rd → 2nd + 3rd
When two third tones meet, the first becomes a second tone. The most famous word in Chinese class is the classic example:
In longer runs of third tones (我很好 "I'm fine"), speakers group the words and apply the rule inside each group — typically wó hén hǎo.
Rule 2: the half-third tone
Before any tone other than a third, a third tone loses its final rise and becomes a simple low tone: 老师 lǎoshī sounds like a low "lao" then a high flat "shi"; same in 很高 hěn gāo and 好吃 hǎochī. Since third tones are followed by something most of the time, the low half-third is the normal third tone — the full 214 dip-and-rise mostly surfaces when the syllable stands alone or ends a phrase. Learning "third tone = low," not "third tone = dip," is the single highest-value correction most self-taught learners can make.
Rule 3: 不 bù → bú before a 4th tone
不 also goes neutral in the middle of verb-not-verb questions: 是不是 shì bu shì.
Rule 4: 一 yī → yí before a 4th tone, yì before the rest
These four rules are standard and well documented — the Chinese Pronunciation Wiki has a thorough reference if you want every edge case. One practical note: because pinyin convention writes citation tones, most apps and books show you nǐ hǎo while every recording says ní hǎo. Literate Chinese has a "spoken tone marks" setting that rewrites the pinyin (and zhuyin) to show exactly these sandhi rules applied — or both forms at once — so what you read finally matches what you hear.
How do you actually learn tones? (Not one syllable at a time)
Drilling isolated syllables past the first week is the classic dead end: real Mandarin never presents tones one at a time, and half the rules above only exist between syllables. What works instead:
- Drill tone pairs. Two-syllable words are the natural unit of Mandarin, and there are only 20 tone-pair combinations (4×4 plus neutral endings). Collect a few example words for each pair — 今天 (1-1), 中国 (1-2), 很好 (3-3 → 2-3) — and drill the melodies, not the individual tones.
- Listen at sentence speed early. Your goal is recognizing tone shapes inside the stream of speech, where they're compressed and fast.
- Shadow. Play a sentence, say it back imitating the melody exactly — pitch, rhythm, everything. Ugly but effective.
- Read along with audio. Reading while hearing every word pronounced welds tones to vocabulary instead of storing them as separate trivia. This is exactly what tap-to-hear graded stories are for — try reading your first story with the audio on.
What mistakes do English speakers make with tones?
- Performing the full third-tone dip everywhere. It's usually just low (Rule 2). The exaggerated dip is why beginners sound slow-motion.
- Letting sentence intonation override tones. English raises pitch at the end of questions; doing that in Mandarin rewrites your final word's tone. Emotion in Mandarin lives in rhythm, volume, and particles — the tones stay.
- Storing tones as numbers, not sounds. If you have to recall "3rd tone" and then compute the pitch, you're too slow for speech. Store the sound of the whole word.
- Dropping tones on unstressed syllables. Second syllables of full-tone words keep their tone; only genuine neutral-tone syllables reduce.
- Betting that context will cover for you. It partly does in person — and then fails exactly when stakes are highest: names, numbers, 买/卖, and anything said over the phone.
Are tones different in Taiwan?
The four-tone system is identical, but two real differences are worth knowing. First, the neutral tone is far rarer in Taiwan: words that Beijing reduces — 先生 xiānsheng, 东西 dōngxi (東西) — are typically pronounced with full tones in Taipei (xiānshēng, dōngxī). Second, a modest set of words carries a different official tone in Taiwan's Ministry of Education dictionary than in the Mainland standard: 星期 is xīngqī on the Mainland but xīngqí in Taiwan, 企业/企業 is qǐyè versus Taiwan's qìyè, and 垃圾 is lājī on the Mainland but lèsè in Taiwan (a different reading entirely, not just a tone). Neither variety is "wrong" — they're two standards, and our dictionary deliberately encodes the Taiwan readings in zhuyin so both are represented accurately. For vocabulary-level differences, see Taiwan vs Mainland Mandarin; for the writing systems themselves, zhuyin vs pinyin.
Frequently asked questions
Can you speak Chinese without tones and still be understood?
In face-to-face conversation, context rescues a lot — and then fails on names, numbers, and short answers, where nothing else disambiguates. More importantly, toneless speech caps your listening: if you didn't store the tones, similar words blur together forever. Learn them from day one; they're cheaper early.
Which tone is hardest for English speakers?
The third, almost universally — mostly because it's taught as a full dip-and-rise when in connected speech it's usually just low. Relearn it as "the low tone" and both your speaking and listening improve immediately.
Do native speakers think about tone sandhi?
No — the rules run completely automatically, the way English speakers say "a" versus "an" without thinking. With enough listening and shadowing, they become automatic for learners too; the rules exist so you know what you're hearing in the meantime.
Are Cantonese tones the same as Mandarin tones?
No. Cantonese has six tones (by the usual count) with different contours, and knowledge doesn't transfer directly. This guide is Mandarin-only — though the learning method, tone pairs and shadowing, works for any tonal language.
Why does my textbook write nǐ hǎo if everyone says ní hǎo?
Pinyin convention writes the original dictionary tones and trusts you to apply sandhi. That's why a "spoken tone marks" display (which Literate Chinese offers as a setting) is useful — it shows the tones as actually pronounced.
