Shanghainese pronunciation: the sounds, one by one

Shanghainese has sounds Mandarin does not have, and the most important of them is one English does not distinguish either. This is the whole inventory, with a word for each sound that you can listen to.

Updated 18 September 2026 · 6 min read

Tap 🔊 to hear an example. Shanghainese examples play the recordings from Verbavia’s course, made with a Shanghainese voice; Mandarin examples play a Mandarin voice.

The one thing to know first

Mandarin has two kinds of stop consonant: unaspirated, as in 布 bù ‘cloth’, and aspirated, as in 怕 pà ‘afraid’. Shanghainese has three. Alongside those two it keeps a voiced series that Mandarin lost: 步 bu ‘a step’ is neither of the other two, and no amount of listening to Mandarin will have trained your ear for it.

This matters far beyond the consonant itself, because in Shanghainese the initial consonant tells you which of two tone registers the syllable is in. Get the voicing wrong and the tone follows it wrong. Tone is dealt with on its own page — tones and tone sandhi — and this page is about the consonants and vowels that carry it.

The initials

Twenty-six consonants can start a syllable, plus a syllable that starts with no consonant at all. The middle column is how Verbavia’s course spells the sound, following the Wu Association romanisation; the IPA is from the reference descriptions listed at the foot of this page. Tap a word to hear the sound in use.

Four rows there deserve a sentence of their own.

• /ɲ/, written ny, is the sound in the middle of English canyon, used as a consonant in its own right. It is why 人 nyin ‘person’ sounds nothing like Mandarin’s rén.
• /ɦ/, written gh, is a breathy h made with the vocal folds loosely vibrating. Before i and iu the course writes it y, and before u it writes it w — same sound, three spellings, chosen so the words look the way an English reader would expect.
• /z/ covers two sounds. Before the high front vowels it is really [ʑ], the voiced partner of sh; elsewhere it is a plain z. They never contrast, so the romanisation does not separate them.
• A syllable can begin with nothing. 衣 i ‘clothing’ has no initial consonant at all, and the difference between that and 移 yi ‘to move’ is that the second is in the voiced register — which is exactly how the writing tells you the tone.

The finals

A final is everything after the initial consonant. Shanghainese has a larger vowel inventory than Mandarin — Chen and Gussenhoven count around twenty vowel qualities, twelve of them contrastive. These are the plain finals; each can also take a i-, u- or iu- glide in front of it, which is where the rest of the inventory comes from.

Two rows at the bottom are not vowels at all. 五 ng ‘five’ is a syllable made entirely of an ng, and 呒 m ‘to not have’ is a syllable made entirely of an m — which is how 呒没 m meh ‘there isn’t any’, one of the commonest words in the language, begins.

Two others are worth pointing at. The y of 知 tsy ‘to know’ is not the English y: it is the buzzing vowel that a Mandarin speaker knows from zi, ci, si. And the nasal finals — an, aon, en — do not end in a consonant you can hear separately. The vowel itself is nasalised, as in French blanc, rather than closed off with an n the way Mandarin closes lán.

The glottal stop, and why syllables are short

Three finals in the table end in -h: ah, oh, eh. That h is not breath. It is a glottal stop — the catch in the middle of uh-oh — and it cuts the syllable short.

These are the checked syllables, and they are the remains of the Middle Chinese syllables that ended in -p, -t or -k. Cantonese still has all three of those endings; Shanghainese has collapsed them into one stop, and Mandarin has lost them altogether. It is why so much of the everyday vocabulary is abrupt where the Mandarin equivalent is not: 勿 veh ‘not’, 一 ih ‘one’, 吃 chih ‘to eat’, 脚 ciah ‘foot’.

Checked syllables also carry their own tones, which is dealt with on the tones page. In the middle of a phrase the stop is often not fully made and you hear only a shortened vowel.

Putting a syllable together

The shape is always the same: an optional initial consonant, an optional glide, then a rime. 光 kuaon ‘light’ is k + u + aon. 脚 ciah ‘foot’ is c + i + ah. 我 ngu ‘I, me’ is ng + u, with no glide. And 五 ng ‘five’ is a rime on its own.

Two habits will carry you a long way. Do not add a vowel after a final -h — the syllable simply stops. And do not close the nasal finals with a hard n or ng; let the vowel go through the nose and end there.

Why you will see two spellings for the same word

There is no official romanisation of Shanghainese, so the same word is spelt differently depending on which system you are reading. The two you will meet most are the Wu Association system, which this course uses, and Wugniu, which is the commonest online. They agree on the consonants and disagree on a few vowels, and they disagree on the glottal stop, which Wu Association writes -h and Wugniu writes -q.

One more difference is worth knowing if you are comparing sources. Wu Association keeps two spellings, e and ae, where the reference inventory of Xu and Tang (1988) lists a single final — so 来 le ‘to come’ and 兰 lae ‘orchid’ are written apart here and together there. The full history of the rival systems is on the writing page.

What to practise first

In order of how much they will buy you:

1. The three-way stop contrast. Say 布 pu ‘cloth’, 怕 pho ‘afraid’ and 步 bu ‘a step’ one after another until the third is not the first. Everything else in the sound system rests on this.
2. The glottal stop. Cut 一 ih ‘one’ and 勿 veh ‘not’ short. A learner who lets them run on is audibly reading Mandarin.
3. The nasal vowels. 冷 lan ‘cold’ and 忙 maon ‘busy’ end in the nose, not on the tongue.
4. /ɲ/ and /ɦ/. 人 nyin ‘person’ and 鞋 gha ‘shoe’ are the two sounds an English speaker most often replaces with something else.

Do not drill tones in isolation. Shanghainese tone is decided at the level of the word rather than the syllable, so a word learned whole comes out right; that is the argument of the tones page, and it is the single biggest way this language is easier than Mandarin.

Sources

1. Chen, Y. and Gussenhoven, C. (2015). Shanghai Chinese. Journal of the International Phonetic Association 45(3).
2. Wikipedia: Shanghainese, initials and finals (citing Xu Baohua and Tang Zhenzhu 1988, 《上海市区方言志》, and Qian Nairong 2007)
3. Ladefoged, P. and Maddieson, I. (1996). The Sounds of the World’s Languages, pp. 64–66, on slack voice.
4. Norman, J. (1988). Chinese, p. 216, on the retention of Middle Chinese voiced initials in Wu — cited in Wikipedia: Yue Chinese.
5. Wikipedia: Romanization of Wu Chinese
6. Rime input method: Wugniu schemes