AIM Framework glossary
Self-Model Transience Principle
Canonical definition
The Self-Model Transience Principle holds that the Bayesian Self-Model is not itself located in any Decoherence Regime, because it is not a single Belief but a transient aggregate output, the current posterior over the Hypothesis Space, continually rewritten by Belief-Register Updates. Each component Belief composing the self-model occupies one of the five Decoherence Regimes; the self-model is their weighted superposition and therefore inherits no single regime. Its accuracy is accordingly a transient degree equal to the regime-weighted composition of its components, bounded strictly between 0 and 1: it can never be wholly accurate, because some components are at any moment Inaccessible or Abandoned, and never wholly inaccurate, because some are Truth-Tracked. A component's regime can be ascribed only by deliberate examination, which is rationed and cannot be performed for every Belief at once; consequently most components run un-ascribed. Warranted confidence in any Belief tracks its ascribed regime, and the act of ascription is what earns that confidence; a Belief whose regime has not been ascribed warrants Epistemic Suspension, since it is not yet known whether reality will collect on it.
Explanation
Your sense of who you are is not one Belief with one truth-value; it is a shifting bundle of many Beliefs, each of a different kind, and it is being rewritten all the time. That is why you can never call your self-image simply true or simply false, it is always partly solid, partly confabulated, partly about things you can't yet even see about yourself, and the mix keeps moving. You can only work out which kind any one Belief is by stopping and examining it against these five categories, and you can't do that for every Belief you hold, so most run unchecked. The honest rule that follows: trust a Belief as much as its kind warrants once you've checked which kind it is, and until you've checked, hold your confidence in suspense, because you don't yet know whether reality is going to come collect.
Epistemic status
Framework-Specific Principle.