Chinese pronunciation: 64 minimal pairs to listen to

English uses pitch for the sentence; Mandarin uses it for the word. That is why four of the 13 groups below are about tones, and why the pairs English speakers confuse most — the second tone against the third, the first against the fourth — each have a group of their own. The other groups are sounds: zh against j, z against zh, the vowel of zi and shi, ü, n against ng, the r, and b against p.

What happens here: English speakers hear pitch as attitude, not as part of the word, and so they do two things with Mandarin tones. They ignore them when listening, and when speaking they lay the English sentence melody over the top: every last word of a question rises, every last word of a statement falls, whatever its own tone. Mandarin keeps the contour of each syllable whatever the sentence is doing. The first tone sits high and does not move, like a note held on one pitch. The second climbs from the middle of the voice to the top, the way English asks “what?” in surprise. The third starts low and sinks lower; said on its own or at the end of a sentence it then comes back up a little, but inside a sentence it mostly stays at the bottom, and that low stretch is the part to aim for. The fourth starts at the top and drops hard, like a firm “no!”. Practise each word on its own first, then in the sentence, without letting the sentence bend it. Two third tones in a row do not both dip: the first one rises, so 你好(nǐhǎo) is said níhǎo, although pinyin keeps writing both as third tones.

What happens here: Learners are often taught the third tone as a dip and a rise, and that is where the confusion starts: a third tone that comes back up sounds very like a second tone with a slow start, and an English ear, which listens for the end of a pitch movement, hears both as “rising”. The difference lies at the start and in the middle. The second tone begins in the middle of the voice and climbs at once, steadily, to the top. The third tone begins low and goes lower, with a slight creak in many voices, and in most positions never climbs at all — only at the end of a phrase, or when the word stands alone, does it rise again, and then only part of the way. So aim the second tone upwards from its first moment, and keep the third tone down: think of the third as the low tone rather than the dipping one. A second test helps when listening: if the voice went down before it went up, it was a third tone.

What happens here: Both tones start at the top of the voice, so the whole difference is what happens next, and English gets in the way twice. First, English has no habit of holding a pitch level: a word said firmly on its own tends to fall at the end, the way any English statement does, and a first tone that sags is heard as a fourth. Secondly, English speakers, told that the first tone is “high”, often pitch it in the middle of their voice, where it sounds flat rather than high, and then the fourth tone has no room to fall from. Pitch the first tone higher than feels natural, as if singing one long note, and keep it there to the end of the syllable, even when it is the last word of the sentence. Make the fourth tone short and decisive, falling all the way from the top to the bottom; a fourth tone that only falls a little is the English statement melody, and that is exactly what makes it unclear.

What happens here: Many syllables in Mandarin lose their tone in certain words. Pinyin writes them without a tone mark, and they are said short, light and quickly, with a pitch set by the syllable before: after a first, second or fourth tone they sit low, after a third tone a little higher. English speakers miss this in two ways. Having learned that every syllable has a tone, they give the neutral syllable a full one and say 妈妈(māma) as two equal syllables; and they hear the shortness as English stress and lengthen the first syllable instead of shortening the second. In a handful of words the neutral tone is the only difference between two meanings: 东西(dōngxī) is east and west, 东西(dōngxi) a thing; 地道(dìdào) is a tunnel, 地道(dìdao) authentic. Those pairs are written with the same characters and cannot be put to a voice that reads characters, so the pairs below differ in writing as well. The commonest neutral syllable is the noun ending 子(zi), which has a full third tone, zǐ, in other words — and that is exactly the difference in most of them.

What happens here: English has one sh, one ch and one j, made with the tongue in between the two Mandarin positions, and an English speaker uses that one sound for both sets. Mandarin has two. For zh, ch and sh the tip of the tongue is raised and drawn back towards the hard ridge behind the upper gum, without curling right over, and the lips are not rounded: the sound is darker and hollower than English sh. For j, q and x the tip of the tongue rests behind the lower front teeth and the middle of the tongue is pressed up flat against the hard palate, with the lips spread as for a smile: the sound is brighter and thinner than English sh. The spelling helps once this is known, because the two sets never stand before the same vowel: j, q and x are always followed by i or ü (written u after them), and zh, ch and sh never are. So the danger is not that one consonant is heard as the other, but that the vowel after it is heard wrongly and the whole syllable moves.

What happens here: Pinyin z is not the z of “zoo”: it is the ts at the end of “cats”, without voice, and c is the same with a strong puff of breath — the ts of “it’s hot” — not k or s. Many English speakers say c as ch or as k and z as an English z, and both move the word. The harder part is that z, c and s are made with the tongue tip forward, just behind the upper front teeth, while zh, ch and sh are made with the tip raised and drawn back. In a fast English-shaped pronunciation the two sets drift together, and then 四(sì) (four) and 是(shì) (to be) are the same word. Keep the s sharp and forward, a thin hiss at the teeth, and let sh be a broader, darker sound further back. Speakers from some southern regions merge the two sets in their own speech, and a learner will hear that in everyday life; the standard keeps them apart, and so should a learner.

What happens here: Pinyin uses the letter i for three different vowels, and English speakers say ee for all of them. After j, q, x and most other consonants it is ee. After z, c and s it is a buzz with the tongue held where the consonant was made, close to the teeth: say a long z as in “buzz”, then keep the tongue where it is and let the voice run on, and that is the vowel of 字(zì). After zh, ch, sh and r it is the same idea further back, with the tongue tip raised: something like the vowel of American “sir”, with the lips spread rather than rounded. The syllables zi, ci, si, zhi, chi, shi and ri therefore never sound like “zee” or “she”. Said with ee, they collide with the syllables that really have ee — 字(zì) becomes something close to 记(jì) — and said with an English “uh”, shi and chi move towards she and che, which are other words again.

What happens here: The letters mislead before anything else does. English reads q as k or kw and x as ks or z, so “qi” is said like “key” and “xi” like “zee”, and neither is anywhere near. Pinyin q is a ch made with the tongue tip down behind the lower teeth and the front of the tongue flat against the palate, released with a clear puff of air; x is the hiss made in the same place, like a sh pronounced with a wide smile. Once the letters are set aside, the two are told apart the way English tells “cheap” from “sheep”: one starts with a closure and a burst, the other is a steady hiss from the beginning. A useful exercise is to say “cheese” and “she” with the lips spread and the tongue tip pressed down, and listen to how both brighten. The ch sound must not be rounded, either: English speakers push the lips forward for ch and sh, and Mandarin q and x keep them back.

What happens here: English has no vowel that combines the tongue of “see” with the lips of “too”, and English speakers fill the gap with one of the two: ü said as oo turns 绿(lǜ) (green) into 路(lù) (road), and ü said as “you” adds a y-glide that makes the syllable longer than it is. Say a long ee, keep the tongue exactly where it is, and slowly round the lips into a small tight circle, as if to whistle: the vowel that comes out is ü. If the sound slides into oo, the tongue has moved back; check with a finger on the jaw that nothing moves but the lips. The spelling hides ü in four places: after j, q, x and y pinyin writes plain u, because u never follows those letters, so ju, qu, xu and yu all have ü. The two dots appear only where both vowels are possible, after n and l — and that is where the pairs are.

What happens here: English has both endings, in “sin” and “sing”, so the sounds themselves are not new. What is new is that Mandarin syllables can only end in one of these two consonants, or in a vowel, and that the choice changes the vowel as well. Before -n the vowel is made further forward and higher; before -ng it moves back. English speakers tend to swallow the ending and let the vowel decide, or nasalise the vowel and close with nothing at all, which leaves a listener guessing. Finish every -n syllable with the tip of the tongue firmly against the ridge behind the upper teeth, as at the end of “fun”, and every -ng syllable with the tip down and the back of the tongue up, as at the end of “song”. The vowel follows: fēn has a vowel like the u of “fun”, fēng one further back, close to the o of “song”. Many speakers from the south merge the two endings, but the standard keeps them apart.

What happens here: Mandarin r at the start of a syllable is made in the same place as sh — tongue tip raised towards the ridge behind the gum — but with voice, so that it sits somewhere between the r of “red” and the s of “measure”. An English r is close enough to be understood, but it carries the habit of rounding the lips, and English speakers often add a w in front of it, which Mandarin does not have. The confusions that change words come from elsewhere. Some speakers, and some regions, let r slide towards l or towards n, and an English speaker who does not yet place the tongue firmly does the same: 入口(rùkǒu) (entrance) becomes 路口(lùkǒu) (crossing). For l the tongue tip touches the gum ridge and the air goes round the sides; for n it touches and the air goes through the nose; for r it does not touch at all, and the voice buzzes through a narrow gap above the raised tongue.

What happens here: English tells b from p mainly by the voice: the vocal cords vibrate during b and not during p. Mandarin does not use voice here at all. Both its b and its p are voiceless; what separates them is breath. Mandarin p, t and k are followed by a strong puff of air before the vowel starts, stronger than in English “pin”, and b, d and g have no puff at all, like the p of English “spin” or the t of “stop”. An English speaker who voices b, d and g is still understood. The real trouble is in listening: to an English ear a sound without voice is a p, so a Mandarin b, d or g is easily heard as p, t or k. Listen for the puff, not for the voice. Hold a sheet of paper in front of the mouth: it should flutter for 跑(pǎo) and stay still for 饱(bǎo). The same contrast runs through the other pairs of consonants: j against q, z against c and zh against ch.

What happens here: Pinyin e alone, as in 歌(gē) or 喝(hē), is not the e of “bed”. It is made with the back of the tongue raised, roughly as for o, but with the lips spread and relaxed, and many speakers let it open slightly as it goes on, so that it sounds like the “uh” of “but” with a touch of the vowel of “her”. English speakers say it as e, which no listener expects, or round it into o, and then 歌(gē) (song) drifts towards 锅(guō) (pot). The letter e plays a different part in combinations: in ie, üe and ei it is the e of “yes”. The second trap is in ian. The a of ian is raised by the i before it and the n after it, so the syllable sounds close to “yen”, not “yan”; iang keeps a full, open a, as in “father”, and ends in ng. English speakers who read the letters say both with the same a and let the ending decide, and since the ending is easily lost, 脸(liǎn) (face) and 两(liǎng) (two) are hard to tell apart.