One might claim that a metaphysical commitment to strong determinism is only porous to quantum indeterminacy or atomic indeterminacy (decay behavior for instance). Those two can be lumped together and simply called subatomic indeterminacy or something. Everything else is conceptually derivative of state evolution and therefore deterministic. So does that mean that my model for R fails unless I can invoke these two candidates? My suggestion of amplifying thermodynamic noise doesn’t really cut the mustard (an amusing semantic drift from pass muster, perhaps) because it only appears random and solely characterizable by these macroscopic variables like pressure and temperature, not because it actually is random in the molecule swirl.
But I can substitute an atomic decay counter for my thermodynamic amplifier, or use a quantum random number generator based on laser measurements of vacuum fluctuations. There, I’ve righted the ship, though I’ve jettisoned my previous claim that randomness is not necessary for R’s otherwises. Now it is, but it is not sufficient because of the need for a device like the generative subsystem that uses randomness in a non-arbitrary way to revise decisions. We do encounter a difficulty in porting subatomic indeterminacy into a human analog, of course, though some have given it a try.
But there is some new mathematics for causal emergence that fits well with my model. In causal emergence, ideas like necessity and sufficiency for causal explanations can be shown to have properties in macroscale explanations that are not present at microscales. The model used is a simple Markov chain that flips between two states and information theory is applied to examine a range of conceptual structures for causation running from David Hume’s train of repeating objects (when one damn thing comes after another and then again and again, we may have a cause), up through David Lewis’s notion of counterfactuals in alternative probabilistic universes (could it have happened that way in all possible worlds?), and a few others as well.
So is R’s emergent decision-making system subject to causal emergence and is this a win for free will? R’s emergent system is essentially a tracking model for others and the environment. It’s building decision trees with compression through scoring and rescoring against increasingly fine-grained observations of causality. Sufficiency and necessity are intrinsic to R’s goal-seeking behavior in that without them (assuming they are just a consequence of evolutionary survival or my fiendish robot-building direction) R can’t sustain functioning. Bringing back in computational irreducibility, a system that is trying to model an irreducible system is a macroscale model that compresses the microscale into this causal emergent predictive framework.
So we end with two candidates: (1) libertarian free will with some “metaphysically real” subatomic indeterminacy being tapped to assist with the reorganization paths. The generative machinery therefore disconnects from this strong determinism and gets a bit of real non-determinism to shape the evolution of the decision-making apparatus. R proceeds accordingly to make otherwise decisions. (2) Causal emergence that has greater explanatory power than noisy micro states results in something that is super-compatibilist while maintaining a metaphysical commitment to strong determinism. Let’s call that “causal emergence deterministic super-compatibilism.” It’s “super” because it is has some more meat to it than just saying something like that our minds seem to have freedom to choose various otherwises and that’s compatibilist enough.
It’s fun to imagine what metaphysical alternatives there are to these ideas. Perhaps the most obvious one is some kind of divine spark or a divine design that logically expands into something like an idealist’s commitment to mind alone. For the former we have to wrestle with the obvious problem of physical brains being tangled up with this spark in a dualism. For the latter, however, there is no physicality at all—it’s a mind byproduct. Alas, there is not much to be gained from these models beyond a kind of scholastic attempt to reconcile guesses about the properties of souls or some kind of detailed phenomenology.
There is a pragmatic argument that treating physicality and materialism as true leads to the entire edifice of science and treats human suffering as addressable rather than as a theological problem, has served us well, and fits within this motif of emergence that appears to be pervasive. Dan Dennett’s universal acid burns through the muck again.