## Tofu
**tofu** (noun) — a soft white food made by coagulating soy milk and pressing the resulting curds into solid white blocks.
From Japanese *tōfu* (豆腐), itself a direct borrowing of Chinese *dòufu* (豆腐), a compound of *dòu* (豆, "bean") + *fǔ* (腐, "fermented" or "curdled"). The character *腐* carries a sense of controlled decomposition — the same transformation applied to bean milk that rennet applies to animal milk.
## The Chinese Invention
The manufacture of tofu is traditionally attributed to Prince Liu An (劉安) of the Han dynasty, around 164 BCE, though the documentary record is patchy and the attribution may be legendary in the way that many food origins are. What is clear is that by the Tang dynasty (618–907 CE), tofu was established as an ordinary food across China, with the compound *dòufu* firmly in use.
The word itself encodes the process. *Dòu* (豆) — originally depicting a tall ceremonial vessel, later applied to the legume served in such vessels — shifted entirely to mean "bean" in common use. *Fǔ* (腐) denotes a state of controlled transformation, the kind of change that produces something useful rather than merely rotted. The compound describes a technology: beans brought to a curdled state.
## Buddhism as Vector
If there is a single institutional force responsible for carrying *dòufu* across East Asia, it is Buddhism — specifically the Mahayana tradition's prohibition on meat. As Buddhist monasticism spread from China into Korea (roughly 4th century CE), Japan (6th century CE), and Southeast Asia, it carried with it both the dietary requirement and the culinary infrastructure to satisfy it. Tofu, protein-dense and meat-free, was indispensable to the monastic kitchen.
### Japan: Tōfu
Japanese monks returning from Tang China brought the food and its name. The Chinese compound *dòufu* became *tōfu* (豆腐) in Japanese, preserving the written characters while shifting the pronunciation through the regular phonological patterns of Sino-Japanese borrowing. The *dòu-* initial became *tō-* (the standard Sino-Japanese reading of 豆), and *fǔ* became *fu*. The word entered Japanese as a fully Chinese compound, not a native coinage — the same path taken by hundreds of Buddhist and technical terms.
### Korea: Dubu
Korean received the same characters but rendered them according to Korean phonology: *두부* (*dubu*). Here the initial *dòu* becomes *du*, and *fǔ* becomes *bu* — a slightly different phonological adaptation from the same Chinese source. The food arrived via the same Buddhist channel and the same monastic trade routes.
### Southeast Asia: Tahu, Đậu Hũ
Indonesian and Malay *tahu* reflects a different transmission path — via Hokkien-speaking Chinese traders and migrants rather than through monastic Buddhism. Hokkien (Min Nan) pronounces the compound as *tāu-hū*, which Indonesian adapted as *tahu*, losing the final syllable's vowel. Vietnamese *đậu hũ* is closer to the Mandarin form, *đậu* corresponding to *dòu* and *hũ* to *fǔ*, with Vietnamese tonal marking applied.
Four languages, four phonological adaptations — all from the same Tang Chinese source, distributed by two overlapping networks: Buddhist monasticism moving north and east, Sinophone merchant diaspora moving south.
## Why English Borrowed the Japanese Form
English *tofu* is not borrowed from Chinese. It comes from Japanese *tōfu*, and the reason is chronology. Portuguese traders reached Japan in 1543; the Dutch East India Company established a trading post at Dejima (Nagasaki) in 1641. These were the primary conduits through which East Asian vocabulary entered European languages in the early modern period. Words like *sake*, *kimono*, *zen*, and *tofu* all entered English through this Japanese-Dutch or Japanese-Portuguese pipeline.
China, during the same period, was less open to direct trade with European powers. The Canton System restricted foreign merchants to Guangzhou and limited cultural exchange. By the time English speakers had sustained contact with Chinese speakers sufficient to borrow vocabulary, many Chinese inventions were already known to them under their Japanese names.
## The Pattern: Chinese Inventions, Japanese Names
*Tofu* is not an isolated case. *Zen* is the Japanese pronunciation of Chinese *chán* (禪), itself a borrowing of Sanskrit *dhyāna* (meditation). *Ramen* is the Japanese adaptation of Chinese *lāmiàn* (拉麵, "pulled noodles"). The pattern is consistent: European traders encountered Japan before they had deep contact with the Chinese interior, and the vocabulary arrived accordingly.
This is Bopp's comparative method applied to trade history. The word *tofu* is a stratigraphic record — it shows the Chinese layer (the invention, the compound, the characters), the Buddhist transmission layer (Japan, Korea, the monastic network), the Hokkien merchant layer (Indonesia, Southeast Asia), and the European colonial contact layer (the reason English says *tofu* rather than *dòufu*).
## Summary of Forms
| Language | Form | Source reading | |----------|------|----------------| | Mandarin | dòufu (豆腐) | native | | Japanese | tōfu (豆腐) | Sino-Japanese | | Korean | dubu (두부) | Sino-Korean | | Vietnamese | đậu hũ | Sino-Vietnamese | | Hokkien | tāu-hū | Min Nan | | Indonesian | tahu | via Hokkien | | English | tofu | via Japanese |
Every form is descended from the same Tang Chinese compound. The divergence is a map of trade routes, religious missions, and colonial contact — six centuries of movement compressed into two syllables.