Bayesian Updating

Canonical definition

Bayesian Updating is the rule under which probability mass is redistributed across the Hypothesis Space $\mathcal{H}$ in response to incoming evidence and the current Interference Term, such that posterior probabilities are computed from priors and likelihoods, and each posterior becomes the prior for subsequent updating cycles. The rule is run by Ground-Resolution, the evaluative operation of Conscious Cognition. Bayesian Updating is content-agnostic: it operates identically regardless of the accuracy, quality, or crisis-potential of the Hypotheses being processed, which is why it can produce either truth-discovery (when H is present in $\mathcal{H}$ with non-zero amplitude and doubt is sustained) or Confabulation Cascades (when H is absent from or dominated within $\mathcal{H}$ and doubt has decohered). It is stated canonically in Axiom 8 (Bayesian Belief Dynamics) and is the machinery presupposed by the Confabulation Rationality Lemma and the Bayesian Architecture of Mimetic Crisis.