Pinyin Is Not Pronunciation: q, x, zh, c and ü
Pinyin's letters were assigned to Mandarin sounds, not borrowed from English, and about half a dozen of them mislead an English speaker every time. q is not k. x is not ks. c is ts. zh, ch and sh are made with the tongue curled back. r is not the r in red. ü has no English equivalent at all. Pleco will play you each one; this is what your mouth is meant to be doing.
Pinyin was not written for you
The single most useful thing to know about pinyin is who it was built for. It was designed in China in the 1950s, by a Chinese committee, for Chinese people learning the standard pronunciation of their own language, and to fit twenty-six Latin letters onto a sound system with more distinctions than those letters naturally carry. English intuition was never a design constraint. It could not have been: the same letters had to serve readers whose first language was Cantonese or Hokkien or Sichuanese.
So the committee did what any engineer does with a fixed alphabet and too many phonemes: spent the obvious letters first, then assigned the leftovers - q, x, c and the digraph zh - to whatever had nothing left. Those assignments are perfectly consistent. Once you know what q is, it is that every time, with none of the exceptions English spelling is infested with. They are simply not the sounds those letters make in English, and no amount of reading a chart will make them so.
What follows is what your mouth should do for the six that go wrong most reliably. Where a comparison to English exists it is given; where none does, that is said plainly rather than papered over with an approximation that would teach you a wrong habit.
b and p differ by breath, not by voice
Start here, because it reframes half the consonant chart. In English, bin and pin differ mainly by voicing: your vocal folds buzz through the b and not through the p. In Mandarin, neither b nor p is voiced. Both are voiceless, and what separates them is the puff of air that follows - aspiration.
Mandarin b is the p in spin: a clean unaspirated stop with no puff. Mandarin p is the p in pin: same consonant, released with an audible breath. The same pairing runs right through the system, and once you see it the odd-looking letters start making sense:
| Unaspirated | Aspirated | Both are |
|---|---|---|
| b | p | a p, with and without the puff |
| d | t | a t, with and without the puff |
| g | k | a k, with and without the puff |
| z | c | a ts, with and without the puff |
| zh | ch | a retroflex ch, with and without the puff |
| j | q | a palatal ch, with and without the puff |
There is a physical test that settles it in ten seconds. Hold a strip of thin paper an inch in front of your lips and say the two syllables. On 怕 pà, afraid, the paper should jump. On 爸 bà, dad, it should barely move. English speakers get this one comparatively cheaply, because English aspirates its initial p, t and k anyway - the habit is already in your mouth, and only needs uncoupling from voicing.
q is not k, and x is not ks
These two cause more damage than anything else in the system, because English has trained you hard on both. In English, q is followed by u and pronounced kw, without exception. So 请 qǐng, please, gets read as king, and it is not remotely the same word.
j, q and x are one series, made in a place English does not use. Put the tip of your tongue down behind your lower front teeth and leave it there. Now arch the blade of the tongue up toward the hard palate, and spread your lips into a slight smile - no rounding. From that position:
- x is the fricative. It is near the sh in sheep, but sh has the tongue tip up and the lips pushed forward, and x has the tip down and the lips spread. That difference is the whole sound. 西 xī, west.
- j adds a stop closure to it: near the j in jeep, again with the tip down and the lips spread, and with no puff of air. 鸡 jī, chicken.
- q is j with the puff. Near the ch in cheap, tip down, lips spread, breath released. 去 qù, to go.
A shortcut worth having: j, q and x only ever appear before i or ü sounds, never before a plain u, a, o or e. That is not a coincidence - the sound needs the tongue already up front and high. If you have written qa, or xu meaning a real u, you have written something that does not exist.
zh, ch, sh and r are made with the tongue curled back
The h in zh, ch and sh is not a breathy h. It is a diacritic in disguise, marking the series as retroflex: curl the tip of your tongue up and back so it points at the roof of your mouth just behind the ridge, and make the sound from there. The result is darker and hollower than the English equivalent.
- sh - the sh in shoe, but with the tongue curled back. 书 shū, book.
- zh - the j in jug, curled back, unvoiced, no puff. 中 zhōng, middle.
- ch - the same with the puff. 吃 chī, to eat.
- r - the same tongue position as sh, but voiced. It sits somewhere between the s in measure and an American r, and it is emphatically not the r in red, which is made with the lips rounded and the tongue bunched. 人 rén, person. 日 rì, day.
This is also where native speech varies most. Many speakers in the south merge the retroflex series with the plain one, so shi comes out closer to si and zhi closer to zi. Nothing is wrong with those speakers or with your ears - it is a distinction not made everywhere, and you will meet both.
c is ts, and it was the last letter left
菜 cài, a dish or a vegetable, is tsài. Not kài, not sài. c in pinyin is the ts at the end of cats, moved to the front of the syllable and released with a puff of air. Its unaspirated partner z is the ds at the end of kids, likewise moved to the front and with the voicing taken out.
Starting an English word with ts feels impossible for about a day. The trick is to steal it from the middle of a word you already say: take cats, and say only the second half, -ts. Now put a vowel after it. That is c.
The choice of letter looks perverse only from English, where c is redundant - either a k or an s, never needed, and so going spare. Polish, Czech and Hungarian already used c for exactly this ts sound. Pinyin gave the letter nobody wanted a job it already held elsewhere: defensible, and completely opaque unless somebody tells you.
ü is a vowel English does not have
Every other item on this list is a sound you can reach from an English one. ü is not. It is a front rounded vowel: the tongue in the position for ee, the lips rounded as though for oo, both at once. German has it, French has it, English never did.
The reliable way in is to build it rather than imitate it. Say the vowel in see and hold it. Keeping your tongue exactly where it is - this is the part people cheat on - round your lips slowly into a whistle. Do not let the tongue slide back toward oo. The vowel that comes out is ü: 雨 yǔ, rain. 绿 lǜ, green. 女 nǚ, woman.
The written trap is that the dots are dropped after j, q, x and y, so jū, qù, xū and yǔ all contain a ü that is not shown. After n and l the dots stay, and they are doing real work: lù is road and lǜ is green, nǔ is to exert and nǚ is woman. The full set of rules is in the post on tone marks and where they go. To type any of it you use the letter v, which pinyin never otherwise uses - one of several places where the notation you write and the notation you read part company, which is the subject of tone numbers or tone marks.
Reading a description of a sound is not hearing it
Everything above is worth knowing, and none of it is sufficient. You can hold a complete and accurate model of where the tongue sits for x and still fail to tell 去 qù from 出 chū at conversational speed, because identifying a sound in half a second of unexpected speech is a different skill from describing it. The description lives in the part of your mind that reasons; the identification has to become a reflex.
That is the same gap that makes tones hard, and it is worth being blunt about which half of it Tingyin addresses. Tingyin trains tone identification and nothing else. It plays a clip, you pick which tone you heard, it tells you at once. It does not drill initials, it will not teach you to hear q against ch, it teaches no characters and no grammar, and it cannot hear you speak - so it can never tell you whether your own x has the tongue in the right place.
For that half, a person is the only real answer, and italki or a tutor is where you find one for the price of a coffee an hour. HelloChinese runs speech recognition over your pronunciation, which is a rougher instrument but immediate and cheap. Pleco is what you want for hearing a specific syllable on demand, one word at a time.
A dedicated drill earns its place in the tone half, because that is the part every general course covers once and then leaves alone. Tingyin's corpus is 637 clips. Level 1 is 80 single-syllable minimal pairs - the same syllable in each of its tones, so pitch is the only variable - and is free with no account at all; level 2 adds 200 more and is free once you sign in. Levels 3 and 4 hold the remaining 357 and are paid, with progress sync and offline downloads alongside, at $2.49 weekly, $5.99 monthly or $15.99 once on the web. Every clip that ships is a human recording under CC BY-SA 4.0 or CC0, carrying the name of the person who spoke it.
And pinyin misleads on tone too, more systematically than on any consonant. A word written with two third tones is not said with two third tones: 你好 is written nǐ hǎo and said ní hǎo. Tingyin labels every clip by what is spoken rather than what is written, which is why not one of the 637 is tagged third-then-third. The sandhi post has the rest of those rules, and why tones are hard to hear rather than hard to say makes the case for why identification needs its own practice. If you want to try it, the drill starts without an account.
Frequently asked questions
How is q pronounced in Chinese pinyin?
Roughly like the ch in cheap, but with the tip of the tongue held down behind the lower front teeth and the lips spread rather than rounded. It is never a k sound, so 请 qǐng is not king. Its unaspirated partner is j, which is the same sound without the puff of air. Pleco will play you a recording of any syllable you look up, which is worth doing alongside a description like this one.
Why does pinyin use letters so differently from English?
Because it was not derived from English. It was designed in China in the 1950s to map twenty-six Latin letters onto Mandarin's sound system for Chinese readers, and Mandarin has three series of sounds where English has one. The obvious letters were spent first and q, x, c and zh were assigned to what remained. The system is entirely consistent - it is just consistent with itself rather than with English spelling.
What is the difference between zh, ch, sh and z, c, s?
Tongue position. zh, ch and sh are retroflex - the tongue tip curls up and back, giving a darker, hollower sound - while z, c and s are made with the tongue flat behind the upper teeth, so c is the ts in cats and z is that same ts without the puff. Many speakers in southern China merge the two series, so you will hear both in the wild. That is variation, not error.
How do I pronounce ü in Mandarin?
Say the vowel in see, hold it, and round your lips into a whistle without letting the tongue move backwards. It is the German ü and the French u, and English has no equivalent, so build it rather than substituting something close. Watch the spelling: after j, q, x and y the dots are dropped, so qù and yǔ both contain a ü. After n and l they stay, because lù is road and lǜ is green.
Will a pronunciation guide fix my accent?
It will fix your model, which is the necessary first step and not the last one. Nothing written can tell you whether the sound leaving your mouth matches, and nothing here trains your ear to catch the distinction at speed. For feedback on your own production you need a person - italki, a tutor, a patient friend. Tingyin cannot help with any of that: it trains tone identification only, and it cannot hear you speak.