LessonsChinese · Lesson 5
Chinese · Lesson 5 · From Hán-Việt readings to Mandarin
Hán-Việt (Sino-Vietnamese) readings are how Vietnamese pronounces Chinese characters — a pronunciation fixed around the Tang dynasty, over a thousand years ago. Mandarin kept changing after that, but it changed by rule. So the Vietnamese reading of a character predicts a good deal of its Mandarin sound: the tone, the ending and often the initial. If you know Vietnamese this lesson is a shortcut; if you do not, it shows why the two languages share so much.
🔊 The sound in these lessons comes from your device’s own Mandarin voice, so it differs from device to device. Recorded audio will follow.
The figures below were counted on 2,053 readings of 1,899 characters found in words the two languages share (国家 – quốc gia, 学生 – học sinh…). They are tendencies, not laws: the percentage says how often the pattern holds, and the number in brackets is how many characters were counted.
🎵 Tones: the most reliable pattern
| Hán-Việt tone | Mandarin tone | Holds | Examples (tap to hear) |
|---|---|---|---|
| ngang after m n nh ng l v d, or with no initial | tone 2 | 76% (209) | |
| ngang after any other initial | tone 1 | 94% (330) | |
| huyền | tone 2 | 91% (263) | |
| hỏi | tone 3 | 89% (212) | |
| ngã after m n nh ng l v d | tone 3 | 93% (76) | |
| ngã after any other initial | tone 4 | 64% (39) | |
| sắc | tone 4 | 86% (296) | |
| nặng | tone 4 | 94% (280) |
In short: huyền → tone 2, hỏi → tone 3, sắc and nặng → tone 4. Ngang and ngã split by the initial: after the sonorants m, n, nh, ng, l, v, d, ngang gives tone 2 and ngã gives tone 3; after other initials ngang gives tone 1 and ngã usually gives tone 4.
🔚 Endings: what Mandarin kept and what it dropped
| Hán-Việt ending | Mandarin | Holds | Examples |
|---|---|---|---|
| -m | -n | 98% (97) | |
| -nh | -ng | 97% (117) | |
| -n | -n | 99% (346) | |
| -ng | -ng | 99% (271) | |
| -p -t -c -ch | no final consonant | 100% (348) |
🔡 Initials: from near-certain to scattered
| Hán-Việt | Mandarin | Holds | Examples |
|---|---|---|---|
| m | m | 100% (77) | |
| l | l | 98% (100) | |
| n | n | 97% (31) | |
| ch | zh | 99% (68) | |
| gi | j | 89% (38) | |
| d | y | 83% (65) | |
| ph | f | 81% (95) | |
| đ | d · t | 99% (132) | |
| b | b · p | 97% (107) | |
| tr | zh · ch | 93% (98) | |
| v | w · y | 96% (51) | |
| h | h · x | 95% (187) | |
| kh | k · q | 88% (85) | |
| qu | g · j | 88% (49) | |
| c / k | j · g · q | 98% (164) | |
| nh | r · y | 84% (51) | |
| ng | y · w | 75% (59) | |
| th | sh · t | 69% (177) |
Where a row lists two Mandarin sounds, the vowel that follows usually decides: before i and ü, older g, k and h became j, q and x — so Vietnamese h gives h (好 hǎo) or x (学 xué), and kh gives k (可 kě) or q (气 qì). Vietnamese t is the most scattered of all: x, z, s and j each take about a fifth (心 xīn – tâm, 子 zǐ – tử, 三 sān – tam, 进 jìn – tiến).
🧪 Try it
Cover the pinyin and guess the tone from the Vietnamese reading, then tap to hear: đồng (huyền → 2), thuỷ (hỏi → 3), đại (nặng → 4), thiên (ngang after th → 1), nhân (ngang after nh → 2).
How it was counted: each character counts once per reading; only word pairs whose shared origin Wiktionary’s etymologies confirm were used. Hán-Việt readings come from Unihan and Wiktionary, pinyin from CC-CEDICT.