Categorization is not neutral. In the 1930s, the linguist Alfred Korzybski coined a deceptively simple warning: the map is not the territory — but his deeper point, buried in Science and Sanity, was about what happens inside institutions when the map gets old and nobody notices. Every taxonomy you impose on data — every field name, every status code, every schema decision — quietly teaches the people who use it to perceive the world in that shape. The categories become the perception. In finance IT, this shows up as something you've probably seen but may not have named: teams that can only ask the questions their database schema permits. The fields that exist get queried; the fields that don't exist don't get missed. Korzybski called this 'semantic reaction' — the reflexive, pre-conscious fusion of symbol and reality — and his prescription was radical for 1933: build in deliberate structural reminders that the word is not the thing. For a schema, that looks less like documentation and more like purposefully keeping at least one field whose meaning is genuinely contested — a forcing function for the conversation the taxonomy is otherwise designed to make unnecessary.

Which field or category in a system you maintain has gone unquestioned longest — and what type of problem would it structurally prevent you from seeing?

Drawing from General Semantics — Alfred Korzybski — Science and Sanity, 1933