Reflective Self-Models
What would make a model genuinely about the system that maintains it?
The destination question
B-Theory began with the “I” in “I think.” The existing mathematics gives tools for states, distinctions, refinement, and structure. A self-model would need a further account of how a system represents its own boundary, current state, actions, or update process—and how that representation can be tested.
First formal target
Define a small agent with an internal state, observations, actions, and a model that makes predictions about some of its own future states. Then compare a model of the external world with one that includes the agent’s own changing state. Specify a task on which the difference can be measured.
Fixed points or self-reference may be useful mathematical objects, but a fixed point by itself is not evidence of self-recognition. A useful first result would show what extra conditions are needed for a reflective representation to predict or control something that a non-reflective one cannot.
Current status
This is a proposed research stub. It connects the research map to the larger question of The MIND without claiming that the present papers prove a theory of consciousness.