How to sound more native in Mandarin Chinese

Most learners can say mā má mǎ mà correctly and still sound foreign in a sentence, because tones change shape when they meet. Native speakers learn words as tone patterns: 电话 diànhuà is "fall + fall", 中国 Zhōngguó is "high + rise". Practise the 20 two-syllable combinations (four tones times four tones, plus a neutral second syllable) until each pattern is automatic, and new words will slot into a pattern you already have.

Three rule-based tone changes apply in every sentence. Two third tones in a row become rising + third: 你好 nǐ hǎo is said ní hǎo. 不 bù becomes bú before a fourth tone (不是 bú shì). 一 yī becomes yí before a fourth tone and yì before the other tones, and stays yī when counting. A third tone before any other tone keeps only its low part (很忙 hěn máng is low, then rising): doing the full dip-and-rise everywhere is one of the clearest learner accents.

Keep the range wide. Learners often compress all four tones into a narrow band, so the first tone is not high enough and the fourth tone does not fall far enough. Native speech is not exaggerated, but the contrast between a high level tone and a low tone is large.

Mandarin has three sets of hissing and hushing sounds where English has roughly one, and mixing them is the most noticeable consonant accent. zh ch sh r are retroflex: the tip of the tongue curls up and back towards the roof of the mouth, and the sound is hollow and dark. j q x are alveolo-palatal: the tip rests behind the lower front teeth, the middle of the tongue rises to the hard palate, and the lips spread, so the sound is bright and thin. z c s keep the tip behind the upper teeth, like the ds in "kids" and the ts in "cats".

The aspiration pairs matter as much as the place: zh, j and z are unaspirated (no puff of air), ch, q and c are aspirated (a strong puff). Hold a strip of paper in front of your lips: it should move on ch, q and c only. j q x are only ever followed by i or ü (spelled u after them), so ju, qu, xu are jü, qü, xü.

r is not the English r. It is a light, buzzy sound made with the tongue in roughly the zh position and the lips relaxed, not rounded. In 日 rì and 热 rè, aim for something between a soft "zh" and an English r without the lip rounding.

Native speakers do not all sound the same here. In much of southern China and in Taiwan, many people pronounce zh ch sh close to z c s in everyday speech. Standard Mandarin (Putonghua, and the speech of news readers) keeps them apart, and learning the distinction is worth it: you will understand both kinds of speaker, and you can always relax it later.

ü: say "ee" and, without moving your tongue, round your lips as for "oo". If your tongue slides back you get "oo" and 绿 lǜ (green) turns into 路 lù (road). After j q x and y the two dots are dropped in pinyin, so 去 qù, 学 xué and 鱼 yú all contain ü.

-n and -ng: -n ends with the tongue tip on the ridge behind the upper teeth, -ng with the back of the tongue raised, as at the end of "sing". The vowel changes too: an is front and bright, ang is further back and darker. Contrast 山 shān (mountain) with 上 shàng (up), 心 xīn (heart) with 星 xīng (star), 门 mén (door) with 盟 méng (alliance). Speakers of some regional varieties merge some of these pairs as well, but standard Mandarin keeps them distinct.

-e: the e in 饿 è (hungry), 喝 hē (to drink) and 色 sè is not the e of "bed". It is a back vowel with the lips unrounded, a little like the "uh" in "duh" said further back in the throat. Inside ie, üe and ei, e is closer to the e of "yes". The -i after zh ch sh r z c s is not "ee" either: it is a buzz that continues the consonant, so 是 shì does not rhyme with "she".

Spelling traps worth knowing: iu is short for iou (六 liù), ui for uei (对 duì) and un for uen (论 lùn), and ian sounds close to "yen" (天 tiān), not "yan".

English squeezes unstressed syllables and stretches stressed ones; Mandarin gives most syllables their full shape. A learner who swallows syllables, English-style, sounds unclear. A learner who gives every syllable equal weight sounds robotic, because Mandarin has its own light syllables: the neutral tone.

Neutral-tone syllables are short, light and take their pitch from the syllable before: a little higher after a third tone, lower after a fourth. They include grammar words (的 de, 了 le, 吗 ma, 呢 ne), the second half of many common words (谢谢 xièxie, 东西 dōngxi "thing", 朋友 péngyou, 时候 shíhou) and repeated kinship terms (妈妈 māma). Saying 东西 as dōngxī changes it to "east and west". Learn the neutral tone with the word, from audio, because dictionaries and regions differ on some of them.

Speak in phrases. Group words the way they belong together (我想 / 去北京 / 学中文) and keep the pitch moving across the phrase instead of resetting after each syllable. Shadowing native audio phrase by phrase is the fastest way to absorb this rhythm.

Erhua is the -r added to the end of some syllables: 哪儿 nǎr (where), 一点儿 yìdiǎnr (a little), 玩儿 wánr (to play). It is strongest in Beijing and parts of northern China; many speakers in the south and in Taiwan rarely use it, and say 哪里 nǎlǐ and 一点 yìdiǎn instead.

You do not need erhua to sound native. Use it in the handful of common words where your teacher or your audio uses it, and do not add it to words that do not take it: a scattering of extra -r sounds more foreign than none at all.

Ten to fifteen minutes a day beats an hour on the weekend. Pick one contrast from this page for the week (say, sh against x), find ten words that contain it, and run each word through the same loop: listen to a native recording, say it, record it, play both back, then get it scored so you know which syllable and which part of it (tone, initial or final) slipped. When a word passes three days in a row, move it into a phrase.

For speaking practice beyond single words, the HSK oral exercises give you picture descriptions and questions to answer aloud, and a tutor or language partner can tell you what no automatic check measures: whether your sentence intonation and word choice sound natural.