Why does Shanghainese sound like Japanese?

Shanghainese sounds like Japanese to a lot of first-time listeners, and the impression has real causes. It keeps the voiced consonants b, d, g, v and z that Mandarin lost; its tones spread across whole words the way a Japanese pitch accent does; and both languages hold on to an old layer of Chinese pronunciation that Mandarin let go. None of that makes the two languages related. Here is what each reason amounts to.

Updated 16 October 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 short answer

• Voiced consonants. Japanese and Shanghainese both have a real b, d, g and z. Mandarin has none.
• Pitch over the whole word. In a Shanghainese word the first syllable’s tone sets the melody for the rest, which is close to how Japanese pitch accent works. Mandarin gives every syllable its own tone.
• Plain vowels and clipped syllables. Shanghainese flattened many of the gliding vowels Mandarin keeps, and it cuts some syllables short with a catch in the throat.
• A shared old layer. Words like person and day start with n in everyday Shanghainese and in Japanese’s older readings of the same characters, where Mandarin has an r.

It is not that Japanese came from Shanghainese, or the other way round. Japanese is not a Chinese language at all. What the two share is a set of features that Mandarin, the language most people compare them with, happens to have lost.

1. The voiced consonants Mandarin lost

Mandarin’s b, d and g are not what an English speaker means by those letters. Pinyin uses them for unaspirated p, t and k, and Mandarin has no voiced stops at all. Shanghainese has a whole voiced series, which Wikipedia notes is rare outside the Wu and Xiang languages:

Japanese has the same set, the sounds its kana mark with two small strokes: ga, za, da, ba. A run of Shanghainese with v, z and g in it therefore has a texture that a Mandarin sentence never has, and that texture is one a Japanese sentence does have.

One detail is worth knowing if you listen for them. At the very start of a word the Shanghainese voiced stops are not fully voiced: Wikipedia describes them as voiceless there, with a slack, breathy quality on the vowel after them, and truly voiced only between vowels. So they come through most clearly in the middle of a word or phrase, as in 生日 san nyih ‘birthday’ or 上海 zaon he ‘Shanghai’ said inside a sentence.

2. Tones that behave like a pitch accent

Mandarin is the textbook tone language: four tones, and each syllable keeps its own. Shanghainese has five tones on a syllable said alone, but in a word of two or more syllables the first syllable’s tone spreads across the rest and the others lose theirs. A whole word comes out as one pitch shape. Wikipedia’s article on the language says it plainly: Shanghainese has a system of tone sandhi similar to Japanese pitch accent.

That is the single biggest reason for the impression. Japanese is also a language in which pitch belongs to the word rather than to each syllable. Listen to a Mandarin sentence and you hear the pitch jump on every syllable; listen to a Shanghainese one and you hear smoother contours that rise or fall once per word, which is much closer to how Japanese moves. How the spreading works, word length by word length, is set out on the tones and tone sandhi page.

3. Plain vowels and clipped syllables

Where Mandarin has a gliding vowel, Shanghainese often has a single steady one:

Japanese has only five vowels and holds each one steady, so a language that turned ai into e and a moves towards it. Shanghainese also ends many syllables with a glottal stop, written -h here, as in 吃 chih ‘to eat’ and 一 ih ‘one’. Wikipedia notes that inside a sentence these are often just short vowels. Japanese, too, is built from short, even beats, and a run of clipped Shanghainese syllables adds to the likeness.

4. The old readings both languages kept

Japanese reads its Chinese characters in more than one way, and the two main layers of Chinese-derived readings are called go-on and kan-on. Go-on is the older. Its name, 呉音, means Wu sound, and Wu is the name of the region around Shanghai and of the language family Shanghainese belongs to. Kan-on came later, in the 7th to 9th centuries, from the pronunciation of the Tang capital, Chang’an.

It is tempting to stop there and say Japanese borrowed its oldest readings from Shanghai. That goes further than the evidence. Wikipedia’s account is that the route go-on took is still debated: Japanese reference works call it a southern layer, but discuss transmission through Korea, by way of the kingdom of Baekje, as well as from southern China, and the dictionary Nihon Kokugo Daijiten warns that the label Wu itself encouraged the theory of a southern origin. Shanghai, for its part, did not yet exist: go-on was in Japan centuries before the name Shanghai appears in any record, as the page on the city’s name explains.

What can be said is that the two languages kept the same old features. Wikipedia notes that go-on preserves the old distinction between voiced and voiceless consonants more often than kan-on does, and that some old nasal consonants stay nasal in go-on but become b, d or z in kan-on. Shanghainese shows the same split between its own two layers of readings: an everyday layer that keeps the nasal, and a bookish layer that does not.

Read across a row and the pattern is plain. The everyday Shanghainese word and the older Japanese reading both start with a nasal, n, ny or m. The bookish Shanghainese reading and the later Japanese one have z, j, v or b instead. Mandarin has moved on from both, to r and w. Even Japan shows it: Japanese calls itself 日本 Nihon, with the old nasal, while Shanghainese says 日本 zeh pen, with its bookish reading.

So when a Shanghainese speaker says 人 nyin ‘person’ where a Mandarin speaker says rén, a Japanese listener hears something close to their own nin. That is not borrowing in either direction. It is two languages that both kept an old sound. Why Shanghainese has two readings for so many characters is the subject of the literary and colloquial readings page.

What it does not mean

• The two are not related. Japanese belongs to the Japonic family. Shanghainese is a Chinese language, in the Wu branch; its family is on the Wu Chinese page.
• A Japanese speaker cannot understand Shanghainese, and the shared readings help only with single words that came from Chinese.
• The resemblance is in the sound, not the grammar. Japanese puts the verb last; Shanghainese, like Mandarin, usually does not.

The quickest way to hear all four effects at once is to listen to Shanghainese words rather than read about them. Every example on this page has the course’s Shanghainese reading, and the free Shanghainese course plays each word in a Shanghainese voice.

Sources

1. Wikipedia: Shanghainese — voiced obstruent initials, rare outside Wu and Xiang; tone sandhi “similar to Japanese pitch accent”; voiced stops voiceless with slack voice word-initially and truly voiced between vowels; checked rimes shortened inside a sentence.
2. Wikipedia: Go-on — the older layer of kanji readings; its route debated (southern China, or Korea and Baekje); the Nihon Kokugo Daijiten on the label; voicing and nasal correspondences with kan-on.
3. Wikipedia: Kan-on — readings from the Chang’an pronunciation of the 7th to 9th centuries.
4. Chen, Y. and Gussenhoven, C. (2015). Shanghai Chinese. Journal of the International Phonetic Association 45(3).
5. Wiktionary: 人, 日 and 物 — the go-on readings nin, nichi and motsu and the kan-on readings jin, jitsu and butsu.