A Tang-Period Borrowing Layer
When Tang dynasty officials governed Annam (northern Vietnam) between the 7th and 10th centuries CE, Vietnamese absorbed a large body of Chinese vocabulary. The historical value of this borrowing layer lies in the phonological distinctions it preserves. Modern linguists call this stratum Hán-Việt (漢越, Sino-Vietnamese).
Estimates place Hán-Việt words at roughly 60% of Vietnamese vocabulary, similar to the proportion of Latin/French-derived words in English. But unlike English, which drastically remodelled its borrowed phonology, Vietnamese retained many distinctions associated with the Tang-period source varieties.
Preservation of Final Stops
Middle Chinese (中古漢語, 7th–10th c.) had four categories of tone, and crucially, the checked tones (入聲, entering tones) ended in unreleased stop consonants: -p, -t, -k. Modern Mandarin lost all three — 國 guó once ended in -k, 學 xué in -k, 食 shí in -k.
Sino-Vietnamese readings preserve many of these finals:
- 學 → học (final -k intact)
- 國 → quốc (final -k intact)
- 食 → thực (final -k intact)
- 十 → thập (final -p intact)
- 八 → bát (final -t intact)
Eight Categories from Four
Middle Chinese had four tones: 平 (level), 上 (rising), 去 (departing), and 入 (entering). Each one split in two according to whether the initial consonant was voiced — a process called tonogenesis — giving the eight tone categories (四聲八調) that Sino-Vietnamese readings still sort into. Vietnamese writes them with just six tones, because the two 入 categories fell together with the two 去 categories: 平 gives ngang and huyền, 上 gives hỏi and ngã, 去 gives sắc and nặng, and the old 入 syllables, the ones ending in -p, -t, -c, -ch, take sắc and nặng as well. So the six marks of quốc ngữ map neatly onto the reconstructed Middle Chinese grid. Modern Mandarin collapsed that same grid into four tones, losing a vast amount of tonal information.
Why Beijing Lost What Hanoi Kept
The difference reflects geographic and political history. Mandarin (普通話) derives mainly from northern dialects that had already undergone centuries of internal restructuring — loss of final stops, merger of tones — long before the standard was codified. Vietnam, separated from China after 938 CE, continued using its Tang-era loan layer without the pressures that reshaped northern Chinese. The Hán-Việt stratum became effectively fossilised.
Sino-Vietnamese readings are therefore one source of comparative evidence for reconstructing Middle Chinese. Alongside Cantonese, Hokkien and Sino-Korean readings, they preserve distinctions that Mandarin alone does not show.
文化 văn hóa (vi) — 文化 wénhuà (zh): same characters, and here the two agree — Vietnamese v- and Mandarin w- are both regular reflexes of the Middle Chinese 微 initial (Baxter mjun), whose original *m- survives instead in Cantonese man4 and Hokkien bûn.
Common words such as quốc, học and văn therefore preserve a borrowing layer established many centuries before modern Standard Chinese.