Rational change of belief

Chapter 6 examines diachronic norms governing belief change over time. Easwaran introduces conditional probability and conditionalization—the rule that upon learning evidence, one should update by conditioning on it. He demonstrates conditionalization's power through confirmation theory, showing how it vindicates induction against Humean skepticism, resolves Hempel's ravens paradox, and grounds Bayes' Theorem linking priors, likelihoods, and posteriors. However, the “problem of the priors” shows confirmation depends on subjective prior distributions even when likelihoods agree. Easwaran presents pragmatic justifications (Dutch book arguments, Good's theorem) and alethic justifications (Greaves-Wallace expected accuracy, Briggs-Pettigrew dominance arguments). He discusses alternatives including Jeffrey conditionalization for learning without certainty, and externalist challenges from cases like unmarked clocks. The chapter concludes with reflection principles—norms relating current credences to expectations about future credences—interpreting them as expressions of commitment rather than strict rational requirements.