Peirce's pragmatist colleagues remembered him as brilliant and maddening in equal measure — partly because he kept insisting that the meaning of any concept is nothing more than its practical consequences. Not what it claims to be. What it *does*. This maps onto a persistent failure mode in mental models: we keep models alive not because they produce useful predictions, but because they feel coherent. The model explains things cleanly after the fact, which our minds register as evidence it works — when really we've just built a comfortable story. The philosopher of science Imre Lakatos called this a 'degenerative research program' — a theoretical framework that keeps generating ad hoc patches to explain away its failures instead of ever being genuinely tested. The leadership version looks like this: you diagnose a team problem using your usual model (incentives are misaligned, communication broke down, leadership was unclear), and each time it doesn't resolve, you adjust the explanation rather than the model. The model stays. The patches accumulate. Lakatos's practical test was blunt — ask whether your framework is making novel predictions that turn out to be correct, or only explaining things that already happened. Apply that to your current mental models and the discomfort is immediate.

Pick a mental model you've used at least three times this month — what did it predict in advance that you could have been wrong about, versus what did it only explain after the fact?

Drawing from Philosophy of Science — Imre Lakatos