Li Shizhen and the Bencao Gangmu

Twenty-seven years, eight hundred sources, nearly two thousand substances — and he corrected his predecessors constantly, which is the part that matters.

Who they were

Li Shizhen (1518–1593), a physician from Qizhou in Hubei province, and the compiler of the Bencao Gangmu — the Compendium of Materia Medica — which is to Chinese pharmacology what Dioscorides is to the European tradition, and is a great deal larger.

Where they came from

A medical family. His grandfather was an itinerant doctor and his father a physician who wanted better for him — medicine was low-status work — and pushed him toward the imperial examinations. Li sat them three times, failed, and gave up on an official career to practise medicine, which turned out to be the right decision for everyone.

He served for a period at the imperial medical academy in Beijing, which gave him access to libraries and to the imperial collections. What he found there was the problem that occupied the rest of his life: the existing materia medica literature was full of errors, duplications and contradictions that had been copied forward for centuries.

He spent twenty-seven years on the correction. He consulted over eight hundred earlier works. He travelled to collect specimens, questioned farmers, hunters, fishermen and woodcutters, and tested preparations — including on himself. The finished work covers 1,892 substances, of which several hundred were newly added, with over 11,000 prescriptions and more than a thousand illustrations.

It was finished around 1578. He could not find a publisher willing to take on something that size, and it was printed in 1596, three years after he died.

What they taught

A classification system, which is the intellectual achievement. Earlier materia medica were organised by the three-grade toxicity scheme of the Shennong Bencao Jing. Li reorganised everything into sixteen divisions and sixty categories by natural kind — waters, fires, earths, metals and minerals, then herbs, grains, vegetables, fruits, trees, then insects, scaled animals, shelled animals, birds, beasts, and human substances — with subdivision within each.

This is a hierarchical natural classification arrived at nearly two centuries before Linnaeus, and historians of science treat it as a genuine independent achievement rather than a curiosity. Charles Darwin cites the Bencao Gangmu in The Variation of Animals and Plants under Domestication, referring to it as the ancient Chinese encyclopaedia.

And a method. Each entry gives the name and its variants, an explanation of the name, a collected account of what earlier sources said, Li's own correction where he judged them wrong, the substance's properties, and the prescriptions using it. The correction sections are the heart of it: he names predecessors and says where they were mistaken, often on the basis of having gone and looked.

What we can and cannot say

We can say the empirical instinct was real and was unusual. A tradition that treats its classics as authoritative does not usually produce a man who spends nearly three decades systematically checking them. His willingness to go to a woodcutter rather than to a text, and to say plainly that an earlier authority was wrong, is the same instinct that Withering showed with foxglove two centuries later.

We can say the work is not a modern pharmacopoeia and should not be read as one. Alongside the careful botany sit substances that do nothing, substances that are toxic, and a section on human materia medica that Li himself was uneasy about and included partly to argue against. Human substances are in there — hair, bone, and worse — and he is explicit that some of the practices are repugnant.

We can say it contains the artemisinin reference indirectly, and this matters. Tu Youyou's source for the cold-water preparation of qinghao was Ge Hong's fourth-century handbook rather than Li — but it was the survival of that entire textual tradition, of which the Bencao Gangmu is the summit, that made the search possible at all. A pharmacological literature that was preserved, compiled and corrected for two thousand years yielded one Nobel Prize, and that is neither nothing nor a great deal.

And we can say what the work does not settle. That the Bencao Gangmu is a monumental, carefully made, historically important book is not evidence that the substances in it work. It is evidence that someone did a very good job of recording what people believed, checked a great deal of it against observation, and produced the best reference anyone could have produced without chemistry. The same is true of Dioscorides, and both traditions ended the same way — superseded by people who could measure things.

We cannot evaluate 1,892 substances. What can be said is that Li Shizhen was doing science under conditions that made science nearly impossible, that he corrected his own tradition more rigorously than most of its later defenders do, and that he died before anyone would print him.

Who they were

Where they came from

What they taught

What we can and cannot say

Further reading