English Pronunciation for Chinese Speakers

The English words that trip up Mandarin speakers are often not the long ones. They are short, everyday words: three, very, street, bag. A sentence can have flawless grammar and still stall on one of them, because the listener hears free, "wery" or back and has to ask again.

Most of these slips have one root: the shape of a Mandarin syllable. It starts with at most one consonant, and it ends in a vowel, /n/ or /ŋ/ (leaving aside the -r suffix). English stacks consonants at both ends of a word and can end a word on almost any of them. Add a few sounds Mandarin does not have, a different way of telling /b/ from /p/, and a rhythm that squeezes some syllables and stretches others, and you have the whole list.

This guide takes those areas one at a time, with example words, what tends to go wrong, and a fix you can practise today. It is the mirror image of our guide to Mandarin tones for English speakers: there, English speakers meet a feature their language does not use; here, the direction is reversed. The learner patterns described below are the ones set out in the chapter on Chinese speakers in Swan and Smith's Learner English (full reference at the end). Where a point applies only to speakers of some Chinese varieties, the section says so.

About the transcriptions. Each table gives a US and a UK transcription. They come from a pronunciation lexicon IPAtics maintains, checked by hand, and we print them exactly as it gives them. Each column differs from most printed dictionaries in one way: the US entries mark stress but not vowel length (street is /ˈstɹit/), and the UK entries mark length but leave stress out. Both write English r as /ɹ/, the precise IPA symbol for it.

Final consonants and clusters

Dropped or padded endings

A Mandarin syllable can end in /n/ or /ŋ/ and in no other consonant. English words end in /t/, /d/, /k/, /s/ and many more, all the time. Learner English describes the two ways Chinese speakers tend to handle this: they drop the final consonant, or they release it with a small extra vowel, so bed comes out close to "bed-uh".

Both habits cost meaning. Dropping the ending merges words that differ only there:

Take the endings off tide, tight and time and all three become tie. English also keeps its grammar at the end of the word, so a dropped final sound can delete the past tense or the plural: worked turns into work, cats into cat.

The fix: close for the last consonant and stop there. English speakers often do not release a final /t/, /k/ or /p/ at all. The tongue or lips close, the sound ends, and there is no puff of air and no vowel after it. Say tight, freeze with the tip of your tongue on the ridge behind your top teeth, and count to two before you let go. If a vowel keeps sneaking in, whisper the final consonant for a few days, or join it to a following vowel: seek out is easier than seek on its own.

Consonant clusters

Mandarin has no consonant clusters: a syllable starts with one consonant at most, followed by a vowel or a glide such as the u in guó. English allows three consonants at the start of a word and up to four at the end. According to Swan and Smith, Chinese learners repair clusters in two ways: they insert a vowel between the consonants, so street sounds like "suh-treet", or they leave one out, so asked becomes ask.

The last four rows matter more than they look. The -ed and -s endings build clusters in almost every English sentence, and dropping one changes the tense or the number.

The fix: build the cluster from the vowel outwards. For street, say eat, then treat, then hold a long /s/ and slide into it: "sss-treat". Keep the tip of your tongue near the ridge behind your teeth for the whole of s-t-r; a vowel can only appear if the tongue drops away. For clusters at the end, work the same way from the other side: ask, "ask-t", asked. A following vowel helps here too, because asked it lets the /t/ start the next syllable, the position where Mandarin allows any of its consonants.

The th sounds /θ/ and /ð/

Standard Mandarin has neither of the English th sounds. Swan and Smith note that Chinese learners usually replace voiceless /θ/ with /s/ or /f/, and voiced /ð/ with /z/. Think turns into sink or fink, and breathe into breeze.

Of the two, /ð/ comes up far more often, because it starts some of the most frequent words in English: the, this, that, they, there. Fixing it in those few words changes how every sentence sounds.

The fix: put the tip of your tongue lightly against the back of your top teeth, or just between them, and blow. Nothing presses hard; the air slips out around the tongue. For /ð/, do the same with your voice on, and rest a finger on your throat to feel the buzz. A mirror helps in the first week: if you cannot see the tip of your tongue, you are most likely saying /s/. The phoneme pages for the voiceless th /θ/ and the voiced th /ð/ have more example words you can listen to.

The /v/ and /w/ mix-up

Mandarin has /w/ but no /v/. In Beijing speech a word-initial /w/ is often made with the lower lip close to the top teeth, a sound between English /v/ and /w/ (see the Wikipedia article on Standard Chinese phonology). So the two English sounds are close to a single sound for many Mandarin speakers, and Learner English lists them as a pair that gets confused: very comes out as "wery", and wet can drift toward vet.

The fix: use the teeth test. For /v/, your top teeth rest on your lower lip and you can hold the sound for as long as you like: "vvvv-est". For /w/, the teeth touch nothing. The lips round as if for oo and glide straight into the next vowel, and you cannot hold it. Watch in a mirror: on /v/ you should see your top teeth, on /w/ only rounded lips.

/n/, /l/ and the English /r/

/n/ and /l/, if your home variety merges them

Standard Mandarin keeps /n/ and /l/ apart, so this point applies only to some speakers. Most Southwestern Mandarin dialects, spoken in Sichuan, Chongqing and neighbouring provinces, have lost the distinction, and many Cantonese speakers merge the two as well. Swan and Smith note that speakers from regions like these tend to carry the merger into English, where nine and line are different words.

The fix: pinch your nose. Say a long "nnnn" and the sound stops, because /n/ sends all its air through the nose. Say a long "llll" and it carries on, because /l/ lets the air out over the sides of the tongue. The tip of the tongue sits in the same place for both, so the nose is the only thing you need to control.

The English /r/

English /r/, written /ɹ/ in the tables, is an approximant: the tongue curls or bunches toward the roof of the mouth but never touches it, and the lips round slightly. The Mandarin sound written r in pinyin, as in rén, has audible friction; Wikipedia's article on Standard Chinese phonology compares it to the s in measure said with the tongue tip curled back. If your English r hisses or buzzes, you are using the pinyin sound. Open the gap between tongue and palate until the noise disappears.

The heart row shows a choice you have to make. American English pronounces r after a vowel; standard British English does not, and lengthens the vowel instead. Both are correct. What sounds odd is an r that comes and goes, so pick one accent and use its column.

Vowels: ship and sheep

English has more vowel contrasts than Mandarin. Swan and Smith single out long and short pairs such as ship and sheep, which Chinese learners tend to pronounce the same.

Compare the two columns and the table shows something that "long ee, short i" hides. The UK entries mark length: /ʃɪp/ against /ʃiːp/. The US entries do not mark length at all, and still write two different vowels: /ˈʃɪp/ against /ˈʃip/. What both columns agree on is the vowel itself. /i/ is high, tense and pushed forward, with the lips spread; /ɪ/ is lower, looser and a little nearer the centre of the mouth. Stretching or clipping the same vowel is not enough. The tongue has to move.

The back pairs work the same way. For /u/ in fool, round the lips and push them forward; for /ʊ/ in full, relax them and let the tongue drop a little.

The fix: for the tense vowels, smile slightly for /i/, push the lips forward for /u/, and tighten; for the lax ones, let the jaw and tongue go slack, as if you were tired. Drill the words as pairs, never alone, because hearing both is what trains your ear. Our list of minimal pairs English learners confuse has more pairs for this contrast and for the others in this guide.

Voicing: why bag sounds like back

Mandarin separates its stops by aspiration, not by voicing. Pinyin b and p are both voiceless; p has a puff of air after it and b does not. English separates /b/ from /p/ mainly by voicing, the buzz of the vocal cords, and by the side effects of voicing.

At the start of a word the Mandarin habit mostly works, because English /p/, /t/ and /k/ are aspirated there too, so pig and big usually come out distinct. The trouble is at the end of a word, where there is no puff to lean on. Swan and Smith list voiced and voiceless consonants among the contrasts Chinese learners find hard, and at the end of a word the difference is easy to lose: bag sounds like back, bed like bet.

The fix: lengthen the vowel. In English, a vowel before a voiced consonant is noticeably longer than the same vowel before a voiceless one, and listeners rely on that length (Roach calls the shortening before voiceless consonants "pre-fortis clipping"). The /æ/ in bag is longer than the /æ/ in back. Stretch the vowel in bag, cab and bed, keep your voice going into the closure, and stop without adding a vowel. That alone keeps the pairs apart, even before the final consonant itself is perfect.

Stress, weak syllables and linking

Every full syllable in Mandarin carries its own tone and takes roughly the same time; only neutral-tone syllables, like the second syllable of māma, are short and light. English works the other way round. One syllable in each word is stressed, made longer, louder and higher, and many of the others shrink to the schwa /ə/. Learner English describes the usual result for Chinese speakers: every syllable gets equal weight and each word is said separately, which makes the English sound choppy even when every single sound is right.

The UK entries in this dictionary do not mark stress, so read it from the US column, where the mark sits before the stressed syllable. Look at photograph and photography: when the stress moves, the vowels move with it. The second syllable is a schwa in photograph and a full /ɑ/ in photography, and the first syllable, stressed in photograph, shrinks to a schwa in photography. Banana has one full vowel and two schwas.

The fix: find the stressed syllable before you say a new word, then make the other syllables smaller rather than making the stressed one louder. The neutral tone is a good model: the unstressed syllables of banana are as short and light as the second syllable of māma. Our guide to the English schwa shows where /ə/ turns up and how to hear it.

Between words, English joins a final consonant to a following vowel. Pick it up sounds like "pi-ki-tup", and turn it off like "tur-ni-toff". Practising these links also helps with the final consonants from the first section, because the consonant now starts the next syllable, where Mandarin habits work for you instead of against you.

A practice plan

1. One area a week. Start with the section that matches the words people most often ask you to repeat.
2. Pairs before single words. Drill the table words as pairs: listen first, then speak.
3. Record yourself. Read a table aloud, play it back, and mark the rows where both words sounded the same.
4. Convert your own text. Paste a sentence you need to say into the English to IPA converter and look for the endings, clusters and stress marks this guide covers.

Trying it yourself

These patterns are easiest to catch in the English you already read every day. IPAtics shows the IPA for any text you select, in any app, so a final /d/, a cluster or a stress mark is one shortcut away.

Sources

• Chang, J. (2001). "Chinese speakers." In M. Swan & B. Smith (eds.), Learner English: A Teacher's Guide to Interference and Other Problems, 2nd ed. Cambridge University Press. The source for every learner pattern described above.
• Han, F. (2013). "Pronunciation Problems of Chinese Learners of English." ORTESOL Journal 30, 26–30. Describes the same patterns for vowel length, /θ/ and /ð/, /v/ and /w/, added final vowels and rhythm.
• Roach, P. (2009). English Phonetics and Phonology: A Practical Course, 4th ed. Cambridge University Press. Pre-fortis clipping and unreleased final plosives.
• Wikipedia, Standard Chinese phonology: syllable-final consonants, the aspiration contrast, the Beijing pronunciation of /w/ and the pinyin r.
• Wikipedia, Southwestern Mandarin and Proper Cantonese pronunciation: the merger of /n/ and /l/.