A model that explains everything explains nothing. This is not a modern epistemological complaint — it's the precise charge the Navya-Nyāya logician Gaṅgeśa leveled against rival philosophical systems in his 14th-century Tattvacintāmaṇi: a framework with no exclusionary power, one that can accommodate any outcome after the fact, has ceased to be a tool for thinking and become a tool for not-thinking. Gaṅgeśa called this vyāpti-failure — the collapse of the inferential link between reason and conclusion when the link stretches to cover too much. The parallel in how we actually use mental models is uncomfortable: most of us don't discard a model when it fails to predict; we add a caveat, a sub-category, an exception. The model survives. Our clarity doesn't. The discipline Gaṅgeśa demanded was not cleverness but surgical restriction — a good inference must be capable of being falsified by a specific counter-instance, not merely refined around it. For writers and thinkers alike, this means your most useful model is not your most comprehensive one; it's the one that would be broken by a clear example you can name in advance.

Pick one mental model you use regularly. What specific outcome, if it occurred, would force you to abandon it — and have you ever actually seen that outcome happen?

Drawing from Navya-Nyāya (New Logic school of Indian philosophy) — Gaṅgeśa Upādhyāya (14th-century Navya-Nyāya philosopher, Tattvacintāmaṇi)