CharcoalMosquito Dear EB,
First a point of clarification: I think Dennett's project is only about explaining why we have the experience of acting in accordance with our will; he is not aiming to explaining how we can, in fact, act in accordance with our will.
I think you've now introduced a straw man Dennett to back up your straw man determinism argument that determinism precludes the actual free will to choose which jacket you wear when you get up in the morning. Compatibilism, AFAIU, is compatible with both microphysical determinism and macroscopic human free will precisely because the initial conditions allow for the evolution of thinking organisms, my cybernetic servomechanism (what I also call Everett's Relative State Observer), for whom macroscopic variables such as moral values, beliefs, social and legal norms, intentions and so on, actually can 'make a difference' and have causal efficacy in our macroscopic human world.
I’ve always understood Dennett's compatibilism as a deflation of both libertarian notions of free will and “greedy” reductive physicalist notions of determinism: from his preferred concept of determinism via Everett’s relative state interpretation of QM, the initial boundary conditions (Big Bang or Bounce), together with the dynamical laws of physics fixes the unitary evolution of the universal wave function \Psi_{\mathrm U} and thus the unfolding, possibly infinite-dimensional, correlation structure of the universe. Environmental decoherence then yields branch trajectories with stable, decohered, quasiclassical macroscopic patterns that the RSO can track and exploit as “real patterns”. Our CNS has evolved to cognitively navigate these real predictive control patterns (and not the microphysics) via reason, morals etc., in a 'reasons-responsive' way. While Dennett also rejects libertarianism, I think his compatibilist free will is, in practice, compatible with the libertarian free will of an evolved information-processing organism within the physical constraints of its universe. For me, the principal difference is how the two ‘wills’ define determinism – as in scale‑relative vs reductively microphysical.
On this picture, you don’t need to either deny determinism or insist on indeterminism: compatibilist agency lives at our human macroscopic scale of stable 'intervention–outcome' relations and is compatible with deterministic Everettian, Bohmian or stochastic microphysics. The historically repetitive impasse comes about when libertarian free will proponents, like yourself, and hard physicalist determinists both assume a reductive picture of 'determinism' as if it erased the causal structure we see in our macroscopic human world – with both agreeing, on the basis of that vaguely stated 'picture', that free agency is therefore incompatible with a deterministic world.
This phenomenological distinction enables him to make effective use of an argument from ignorance: assuming that our access to reality is limited, and that hence we cannot experience all the factors determining our behavior, we can have the experience of freedom even if determinism were true. But notice that there could also be another reason why we have the experience of freedom, namely for exactly the reasons as they appear to us: our future is not yet completely pre-specified by factors outside of our control.
WRT your understanding of compatibilism, Dennett isn’t offering an argument from ignorance based on a phenomenology of human finitude, in fact he argues that the kind of control that matters for human responsibility – information-sensitive, reasons-responsive and foresight-enabled predictive control – is actually a real macro-scale capacity of our evolved cognitive systems rather than some sort of metaphysical power to violate deterministic laws. That’s a positive explanatory account of genuine agency at the human scale rather than "at the level of the phenomenology of appearances" (your straw man compatibilism).
Aside from your reductive (and I think plainly wrong) notion of compatibilism, your definition of 'determinism' follows the naively reductive physicalist argument that sets it over against 'free will' such that:
Essentially, determinism means that everything that appears at any point in time is merely a projection of an initial state through some set of entailing laws. There is no room for intervention; intervention would just be another mere appearance contained in a larger deterministic system. In practice, this means that under determinism none of the macroscopic regularities can have any causal powers of their own. These regularities may be real patterns, in the sense that they pick out a subset of actual states, but again those states are only ephemeral transients horizontally projected across time via entailing laws from an initial state.
Real patterns aren't ephemera, they are stable, decohering, macroscopic realities – what we also call 'worlds', with people and things, history, physics, science, a cup of coffee, your loved ones and so on. You are reducing real macroscopic phenomena to microphysical states, then conflating those 'initial conditions' with a claim that precludes multi-scale causal relations at later stages. Deterministic microphysics is fully compatible with macro-scale causal relations, unless you reduce all 'real' causal relations to QFT ... which would be ridiculous don't you think?
But your straw man (and I think naive) definition of determinism – in terms of the microphysical actually exhausting all possible causal relations – sets up 'real patterns' as having no causal power, thus your (reductive physicalist) determinism allows only the "illusion of free will".
Accordingly, if determinism were true, we could not rationally rely on these real patterns to make predictions, given that they hold no intrinsic causal power. The patterns could vanish into thin air at any moment without violating any causal mechanism.
I find this argument rather incoherent, as many macroscopic regularities in our sense perception are remarkably stable, as you admit, such as those reflected in classical mechanics for example, or in decoherence selecting pointer bases that are redundantly propagated via quantum Darwinism, or thermodynamics, or in the motion of the sun, moon, planets and stars in the sky, and etc. The macroscopic reality doesn't just go 'poof' and vanish, with or without humans to perceive its real predictive macroscopic patterns. Recognising real predictive patterns, however, allows an evolved biological servomechanism to process that information and effect real changes downstream based on its decisions. I think your 'determinism' collapses into precisely that reductive picture you presuppose.
However, we do not experience the world to be a transient, fragile, ephemeral shape-shifting kind of place. This puts pressure on determinism, which ultimately becomes a kind of superdeterminism in which all of macroscopic appearance must be explained in terms of an unbelievably improbable super-fine-tuned initial condition of the universe. Rather than supporting natural science, determinism ultimately undermines scientific explanation.
And here you set up yet another straw man – basically equating determinism with superdeterminism – and concluding from that gross assumption that determinism is unscientific! So from your argument, IFF we:
- Ignore the fact that determinism does not entail superdeterminism – the latter being specifically related to Bell inequalities where deterministic Everettian interpretations have no need to invoke superdeterministic correlations. And:
- Reductively define determinism as 'there is only microphysical causality' and no macroscopic causality. And so:
- Claim that Dennett's concept of 'real patterns' and their causal relations is reducible to the microphysics (itself a reductive physicalist argument!) Then:
On the basis of assuming this reductive physicalist argument – that Dennett himself rejected – you misrepresent compatibilist and interventionist interpretations as 'arguments from ignorance of total conditions' even though they argue precisely that macro-scale causal relations are real, 'explanatorily autonomous' and irreducible to any complete microphysical description. In fact, compatibilism is entirely compatible with, and actually allows, your same conclusion that:
we are justified in believing that macroscopic patterns do make a difference - they do precisely what they appear to do, namely to constrain an open-ended space of possibilities, such that robust and stable patterns arise and persist as a moment-to-moment achievement.
And so we arrive at our compatibilist reality of a physically deterministic universal wave function within which macroscopic, evolved cybernetic servomechanisms really can freely decide what garment they might wear when they wake up in the morning – so long as they're not expecting miraculous open-ended garment possibilities!
