I'm currently working through SICP, so I appreciate your concept. For about a decade, I've been considering a tree structure to represent the Universe, but data/code, perhaps my view isn't so different.
Regarding free will however, I think you are making assumptions that are not definitely established. Why must what we perceive as free will rely on callbacks? It seems just as likely that there is hidden state within the closure that determines the output of that function.
While the complexity of decisions, choices, and actions seem to be free will, I've seen no evidence to suggest that it isn't actually predetermined. Given the shear number of possibilities, why can't free will just represent a traversal of some massive tree?
Quantum Mechanics imposes an interesting twist on this view point. If the wave function doesn't collapse until the lazy evaluation of the quantum state, it too could be a contributing factor to determining free will. I'm unaware of any finding that suggests there is a definite relationship between quantum state and the macroscopic, but I haven't given up on that possibility.
Even if there was concrete evidence to support that theory, it could just be that the model we use is just a very good approximation of the system and it doesn't reveal the implementation -- after all, George implemented it and we have abstracted that implementation away.
