Linnaeus: the founder of databases

A couple of years ago I was in Exeter, and was chatting to Staffan Müller-Wille, who is an expert in the history of biology specialising in Linnaean taxonomy. He mentioned to me that Linnaeus had invented the index card in order to keep track of the increasingly large data set of species names he was receiving from his correspondents. He has now made a public presentation of this.

Apparently, Linnaeus initially numbered his manuscripts absolutely, which meant that each new edition had to have a new numbering system. Eventually he introduced index cards so he could, as it were, sort the data dynamically. This meant that he needed a unique record identifier, which it transpired, he already had: the binomial that all species have, of their genus and species epithet.

Linnaeus had to manage a conflict between the need to bring information into a fixed order for purposes of later retrieval, and the need to permanently integrate new information into that order, says Mueller-Wille. “His solution to this dilemma was to keep information on particular subjects on separate sheets, which could be complemented and reshuffled,” he says.

Towards the end of his career, in the mid-1760s, Linnaeus took this further, inventing a paper tool that has since become very common: index cards. While stored in some fixed, conventional order, often alphabetically, index cards could be retrieved and shuffled around at will to update and compare information at any time.

“Although a seemingly mundane and simple innovation, Linnaeus’ use of index cards marks a major shift in how eighteenth-century naturalists thought about the order of nature,” says Mueller-Wille. The natural world was no longer ordered on a fixed, linear scale, but came to be seen as a map-like natural system of multiple affinities.

I don’t know the context of that last comment. “Affinity” was a term that roughly played the role now played by homology, and meant the overall agreement of traits between groups. But Linnaeus used, in the scheme named after him, a single key for each division or group. He did attempt a “natural” classification late in life, probably about the time he started this indexing system, so maybe the report is right. I’d love to see the paper. Staffan rarely gets anything wrong (even on Linnaeus’ species “concept”).

But think what this means: Linnaeus has a system for recording, retrieving, and indexing data. That’s a database! Linnaeus invented the database. And although he doesn’t use a computer system (Wikipedia is just wrong to so restrict it to computers), he used what we might call data normalisation – each species was listed once, in one group. The tree structure of classification in Linnaean taxonomy roughly answers to a B-tree. He didn’t use the absence of a key as a defining trait. This is not an argument by analogy, either – the development of modern logic, on which databases are founded, was directly influenced by Linnaean taxonomy in the early and mid-19thC. Venn explicitly appeals to biological taxonomy, for example, and Jevons also.