How do you record what no book contains?

That time, that place

In the persecution of 1801 Jeong Yak-jeon was exiled to Heuksando; his younger brother Yak-yong went to Gangjin. Neither came back.

Heuksando is far off the mainland, reached by boat only now and then. He lived there fifteen years and died there.

What he did in exile was look at the sea.

Why this question

He decided to record the fish. The trouble came quickly.

The fish in the Chinese literature and the fish in front of him did not match. Many had no name at all. Where names existed, the islanders' names were not the literature's names.

The scholarly method was to ground everything in texts. By that method nothing could be written.

So he changed the method. Where the literature had nothing, he wrote that it had nothing. He marked which names were the island's and which were his own. And he asked the people who actually knew fish.

What was found

The man he asked was a fisherman named Jang Chang-dae, island-born and sea-wise. Jeong went to his house and observed fish alongside him.

In the preface he set down this man's name — naming a collaborator, which was not a common thing to do in the records of the time.

The book holds 226 species: shape and size, where they are caught and when they are plentiful, how they taste, how to cook them, what they are good for medicinally.

There is no classification system. Nothing here resembles the framework Linnaeus was building in Europe around the same time. What there is instead is a separation of what he saw from what he was told, and a willingness to write that something is unknown.

And half the knowledge the book records belonged to a man who could not write.