What is recursive reliability?
Recursive reliability asks whether a system can assess, preserve and repair the trustworthiness of its own models and reliability estimates.
Recursive reliability is the problem of reliability turned back onto itself. A system does not only need a model of the world. It also needs some way of estimating whether that model can be trusted, when it is failing, and what kind of repair is required.
Why it matters
In ordinary settings, a bad model can be corrected by checking it against an external constraint. In harder settings, the failure may affect the system’s own self-model, memory, goals or reliability estimate. The system then faces a deeper question: can the mechanism that judges reliability itself remain reliable enough to guide repair?
Mirror Programme use
Within the Mirror Programme, recursive reliability is a bridge between observerhood, memory continuity and viable action. Volume I treats reliability not as a decorative confidence score, but as a condition for persistence under perturbation.