CONCEPT
The Ladder of Causation
Judea Pearl's three-rung hierarchy of intelligence—seeing, doing, imagining—and the claim that no amount of data carries you from one rung to the next.
The Ladder of Causation is the central image of
Judea Pearl's account of intelligence, and it is more than a metaphor: it is a hierarchy with a mathematical spine. It has three rungs—
association (seeing),
intervention (doing), and
counterfactual (imagining)—and each corresponds to a kind of question, a kind of cognitive operation, and a kind of formal machinery required to answer it. The rungs are nested, and the gaps between them are differences of kind, not degree: you cannot climb from one to the next by collecting more data, only by adding a new kind of knowledge, namely a
model of how the world works. This is the lens
[YOU] on AI invites us to look through without flinching—the question is never how impressive a machine's performance is, but which rung it is performing on. By Pearl's reckoning, all of contemporary machine learning, including the
large language models that now write and reason in fluent prose, lives entirely on the first rung, doing nothing more