Principlev1
Require schemas to make specific, falsifiable predictions
Require schemas to make specific, falsifiable predictions about future observations rather than merely explaining past events post-hoc.
Why This Is a Principle
Derives from Automatic Narrative Generation Precedes Conscious Evaluation (narrative generation precedes evaluation), Hindsight Bias and Calibration Necessity (memory reconstructs toward outcomes), and Learning occurs when outcomes differ from predictions, (learning requires prediction error). The principle prescribes demanding forward-looking predictions to avoid post-hoc rationalization.