Principlev1PR-732
Design your worldview's protective belt to generate novel…
Design your worldview's protective belt to generate novel predictions rather than only rationalize failures after the fact, using predictive success as a test of whether the worldview is progressing or degenerating.
Why This Is a Principle
Derives from world model accuracy bounding performance (The performance of an agent is bounded by the accuracy…), learning from prediction errors (Learning occurs when outcomes differ from predictions…), and belief perseverance (Belief Perseverance Against Contradictory Evidence). Lakatos's research programmes framework distinguishes progressive from degenerating programs by whether they generate novel predictions. The principle prescribes a specific test for worldview health. Clearly principle-level: actionable criteria for self-evaluation.