Commonly Mispronounced Sounds

Why these ten contrasts, and how to work through them

Learners do not mispronounce Chinese at random. Errors cluster around a small number of contrasts where the articulatory difference is subtle, where the learner's first language has no equivalent distinction, or both. The ten pairs collected here cover the overwhelming majority of what a listener will actually notice. They are not equally urgent, and they should not be drilled simultaneously: working on two or three contrasts in a focused block produces far better results than spending ten minutes on each of ten pairs in a single week.

Each entry below follows the same shape. First the articulatory difference — what the tongue, lips and soft palate are actually doing, in enough detail that you can feel the change. Second, diagnostic word pairs that differ only in the contrast, so that getting them wrong changes the word. Third, a drill, which is deliberately short. A drill is not practice for its own sake; it is a way of making the difference perceptible to your own ear. Once you can hear the contrast reliably, producing it usually follows.

For the full inventory of sounds these contrasts sit inside, see the pinyin chart. For a contrast that is about tone rather than articulation — the third-tone and 一/不 changes — see tone change rules.

z / zh, c / ch, s / sh: flat tongue versus curled tongue

This is the single most valuable group to fix, because it affects six sounds at once, and because the underlying adjustment is the same for all three pairs.

The articulatory difference

For z, c, s the tongue tip stays low and forward, just behind the upper front teeth, touching or nearly touching the ridge there. The tongue is flat and relaxed. For zh, ch, sh the tongue tip lifts and curls back towards the hard palate, so the contact point moves noticeably rearward and the tongue surface becomes hollow. The lips are slightly rounded for the retracted set in many speakers. Everything else — the amount of air, whether there is a stop followed by friction or just friction — is identical between the two columns.

Words that reveal the contrast

• zìjǐ (自己, oneself) versus zhīchí (支持, to support) — compare the first syllables.
• cài (菜, a dish) versus chái (柴, firewood).
• sān (三, three) versus shān (山, mountain).
• zǎo (早, early) versus zhǎo (找, to look for).
• cā (擦, to wipe) versus chā (插, to insert).
• sù (素, vegetarian) versus shù (树, tree).

A drill

Say the three pairs slowly, alternating: zǎo — zhǎo, zǎo — zhǎo. The point is not speed but the physical sensation of the tongue moving forward and back. Hold your hand in front of your mouth: the amount of air should feel identical in both members of each pair, so if aspiration is changing, you are compensating for the wrong thing. Then work through cā — chā and sù — shù the same way. Practise in two-minute blocks, twice a day, for a week. Most learners find the s/sh pair easiest, because the fricative is continuous and can be held; use it as your anchor when the affricates are not yet clear.

an / ang, en / eng, in / ing: the nasal final split

This is the contrast that most often survives into otherwise fluent speech, and it is the one most worth revisiting at intermediate level.

The articulatory difference

Nasal finals end in one of two places. With -n the tongue tip moves forward and closes against the ridge behind the upper front teeth — exactly the position of the initial d. The mouth aperture stays fairly small and the sound resonates in the front of the mouth. With -ng the back of the tongue rises to the soft palate — exactly the position of the initial g — and the mouth stays more open, with resonance at the back. The transition is not about how long you hold the nasal; it is about where the closure happens.

Words that reveal the contrast

• bān (搬, to move) versus bāng (帮, to help).
• chén (陈, a surname) versus chéng (城, a city).
• jīn (金, gold) versus jīng (京, the capital; 经, to pass through).
• mén (门, a door) versus mēng (蒙, to deceive).
• pán (盘, a plate) versus páng (旁, beside).
• xīn (新, new) versus xīng (星, a star).

A drill

The most effective technique for this pair is to extend the nasal. Say bān and hold the final -n for two seconds, feeling the tongue tip pressed behind the teeth; then say bāng and hold the -ng for two seconds, feeling the back of the tongue against the soft palate. Exaggerating the ending makes the place of articulation obvious, which is what you need. Once you can feel the difference, shorten the hold back to normal and work through the whole list. A useful secondary check is to put two fingers lightly on the bridge of your nose: front nasals produce more vibration there, back nasals more at the base of the nose and throat — a rough guide, but a helpful one when your ear is still unreliable.

f / h: a substitution that travels in one direction

The articulatory difference

For f the upper teeth touch the lower lip, and the air escapes between them. For h the mouth is open and the constriction is further back, between the back of the tongue and the soft palate. The two sounds have nothing in common except that both are voiceless fricatives. Crucially, Mandarin h is pronounced further back and with more friction than the English h; it is close to the Scottish "ch" in "loch" or the German "ch" in "Bach". English speakers tend to substitute a weak, glottal h, which is too soft, and in some southern Chinese varieties of the learner's own language the two sounds are merged.

Words that reveal the contrast

• fēi (飞, to fly) versus hēi (黑, black).
• fàn (饭, cooked rice, a meal) versus hàn (汉, a Han person).
• fāng (方, a square, a direction) versus hāng — no common word, which itself is a useful signal: hāng is rare, so if you hear a word that seems to be fāng, it almost certainly is.
• hūn (婚, marriage) versus fēn (分, to divide).

A drill

Alternate fēi — hēi slowly, and pay attention to whether your lips are moving. If you produce hēi with lip movement, you are making an f. Then say hǎo (好, good) ten times in a row, deliberately producing the friction far back in the throat, almost as if clearing your throat gently. The goal is a rougher, deeper h than English uses. Reversing the substitution — saying h where f belongs — is uncommon among English speakers but regular among speakers of some other languages, so check yourself both ways if the contrast is not native to you.

n / l, and r / l: two different problems with one shared tongue position

The articulatory difference

All three sounds are made with the tongue tip at the ridge behind the upper front teeth. n is nasal: the tongue tip closes completely against the ridge and the soft palate lowers so air escapes through the nose. l is lateral: the tongue tip also touches the ridge, but the soft palate is raised, so the air travels out around the sides of the tongue. r in Mandarin is neither: the tongue tip curls back towards the hard palate without closing, and air passes through a narrow channel, producing friction. So n and l differ in the airstream, not the tongue position, while r and l differ in the tongue shape.

Words that reveal the contrast

• nà (那, that) versus là (辣, spicy).
• nán (南, south) versus lán (蓝, blue).
• niú (牛, a cow) versus liú (流, to flow).
• ròu (肉, meat) versus lòu (漏, to leak).
• rén (人, person) versus léng (冷, cold).
• nǎo (脑, the brain) versus lǎo (老, old) — same tone, so the only difference is the initial.

A drill

Press your thumb lightly against one nostril and say nà. You should feel vibration through the nose. Now say là in the same position; the vibration should largely disappear. That physical test is definitive, and it works better than listening to yourself, which is unreliable because you hear your voice through bone conduction. For r, start from shī (湿, wet) — which uses the same curled tongue position — and simply add voicing; the friction sound of sh becomes the friction sound of r. Then contrast ròu and lòu, checking that the tongue curls for the first and flattens for the second.

j q x with ü: a reading problem more often than a speaking one

The articulatory difference

The initials j, q, x are made with the tongue flat and high, the blade against the hard palate, and the lips spread — not rounded. The vowel ü is a front rounded vowel: make the tongue position for "ee" and then round your lips without moving the tongue forward or back. In pinyin, after j, q, x and y, the vowel ü is written as plain u. So jú (局, a bureau), qù (去, to go), xū (需, to need) and yǔ (雨, rain) all contain ü, not u.

Words that reveal the contrast

• júzi (橘子, a tangerine) — the vowel is ü.
• qùnián (去年, last year) — ü.
• xūyào (需要, to need) — ü.
• yǔyán (语言, language) — ü.
• Against nǚ (女, a woman) and lǜ (绿, green), where the umlaut is written because n and l combine with both vowels.

A drill

The problem here is usually not the vowel itself but the letter u on the page, which drags the reader towards the English "oo". Practise by writing the syllables out with an explicit umlaut — jü, qü, xü, yü — and reading them aloud from that version until the habit is broken. Compare with pairs where the distinction is visible: lù (路, road) versus lǜ (绿, green), nù (怒, anger) versus nǚ (女, woman). Because the ambiguity exists only in the spelling, and never in the sound, this contrast is almost always resolved by attention rather than by physical training.

Which errors come from your first language, and how to prioritise

The most efficient way to decide what to work on is to ask which contrasts your first language does not mark, because those are the ones your ear will not report. A contrast that does not exist in your native language is not merely harder to produce; it is harder to hear, which is why learners are frequently convinced they are saying two different sounds when a native listener hears one. Below is where the common confusions usually come from. These are tendencies, not rules — individual learners vary, and multilingual learners may face several at once.

• Speakers of English and most European languages: the aspiration contrasts (b/p, d/t, g/k, j/q, zh/ch, z/c) are the systematic problem, because English distinguishes voicing instead. The z/zh, c/ch, s/sh split also needs explicit work, since English has no retroflex series. Front versus back nasals are usually either fine or fixed quickly.
• Speakers of languages with no retroflex consonants (Japanese, Korean, many Romance languages): the whole zh/ch/sh/r column collapses into the flat series, giving zīchí for zhīchí. This needs the most work of anything on this page.
• Speakers of languages without a front rounded vowel (English, Spanish, most Slavic languages except those with ü): ü is produced as plain u, so nǚ and nù merge. Since English has no ü at all, this needs deliberate practice — but note it is the vowel, not the j/q/x spelling, that has to be learned.
• Speakers whose own language merges n and l (some southern Chinese varieties, some Caribbean and West African languages): n/l is a genuine perceptual merger and needs the nostril test described above.
• Speakers whose own language lacks /f/ (Korean, some Japanese and Indic contexts): f and h are substituted for each other in both directions.
• Speakers of tonal languages with a different tone system (Thai, Vietnamese, Cantonese): the tones themselves transfer imperfectly, and the sandhi patterns differ, so the tone change rules are the priority rather than the consonants.

Prioritise in this order. First, whatever contrast causes a native listener to misunderstand you — usually the retroflex series or the nasal finals, since those create real homophones. Second, whatever contrast your own ear cannot reliably detect, because that error will persist indefinitely without attention. Third, accent features that do not impede comprehension: a slightly too-soft h, an over-long third tone, a rounded j. Fixing these makes you sound more native, but they cost time and should come after communication is secure.

Checking your own pronunciation against the standard

Self-assessment is the weak point of pronunciation practice, for a simple reason: you cannot hear your own errors reliably, because you perceive your intended sound rather than the sound you produced. Two workarounds make a real difference. The first is the mirror-and-hand method used throughout this page — checking whether the lips move, whether the nostrils vibrate, how much air hits your palm — because these give you information that bypasses listening. The second is to compare your production against a written standard rather than against your memory, which is where an annotated text is genuinely useful.

Paste a short passage into the online pinyin tool and it returns the text with the correct pinyin and tone marks above every character. Read it aloud against the annotation, paying particular attention to the contrasts on this page: if the annotation says zhīchí and you hear yourself say zīcí, that is exactly the kind of error a mirror will not catch but a written reference will make visible. The tool also flags characters with more than one reading, so you are not silently practising a wrong reading — see common polyphonic characters for why that matters, and 500 essential characters if you want a frequency-ordered list to build drill material from. Everything is free and runs in the browser; nothing you paste is uploaded.

The series continues in the Tutorials, which also covers making printable annotated reading material if you are preparing practice texts for your own study or for other learners. And to apply the method from this page directly, open the online pinyin tool, paste a short passage, and read the annotated result aloud while watching for the specific contrasts you have chosen to work on: the written reading above each character is the standard the drills are training you towards, and seeing it while you speak is the fastest way to catch the errors your own ear will not report.