• Models

It may sound unlikely, but in 1969, Mr. Sum studied law at the University of Utrecht while making a living as a computer programmer. The social sciences had discovered the computer as a thing that brought statistics within reach. He had limited confidence in statistics, but he found programming fantastic.

Cogito (Mr. Sum’s first name) thought programming was a fascinating experience. The ability to tell a machine what to do appealed to him. Especially because assignments could be mathematical and literary. To learn the trick, he visited teachers of mathematics who had also become fascinated. A turning point was the lecture in which it was clarified that a computer has a memory containing assignments and data, while there are assignments that can change the content in memory anywhere. Although Mr van Eck, who gave the lecture, acknowledged on request that a program could change itself in this way, he did not recommend it, given the inconsistencies this would cause. Mister Sum thought differently, more in terms of the ability to learn.

At the time (and, surprisingly, even today), one technique was in the spotlight. Principal component analysis. The idea is to formulate a suspicion of what facts might affect something you want to know – say a series of measures for the spread of the COVID-19 virus. And that you then let the computer calculate the most important factors in the spread of the virus.

Obviously, in such a setup it should be measured whether there have been infected individuals near the subject of the investigation, and what the nature of their interaction was, if any. And furthermore that it must be measured during periods. Those periods and those distance measures make lattices. A lattice is understood here as a regular arrangement (of points, moments or things) in an area or a space. These lattices, like the language, provide glasses for looking through to what was measured. (Varimax rotation?)