Recursive self-models and minimal phenomenal experience
George Deane, Daphne Demekas preprint
A state of pure awareness stripped of conceptual content, known as minimal phenomenal experience (MPE), can arise from a computational architecture where a policy model generating behavior is recursively coupled to a program model that synthesizes structured explanations of that behavior. These programs function as hypothetical self-models that condition future behavior, creating attractor dynamics that stabilize into a coherent identity. MPE emerges when programs remain simple and interoceptively focused, the self-hypothesis distribution stays broad, and the agent sustains awareness without recruiting extended conceptual structures.