I have been turning this over for a while, and recently spent enough time on it to try to write the position down in one place. What follows is a walk through the major arguments in the debate, with notes on which ones I find compelling, followed by the position those arguments leave me with.

A heads up. Most of the moves in the position are not new — the dilemma argument, the basic argument, the special-physics-for-brains objection, the use-theory of meaning, and the deflationary stance toward the compatibilist / incompatibilist debate all have a long pedigree. What I'd claim for myself is mostly the way the pieces fit together. The label that comes closest is something like a non-reductive (or "wide") computational functionalism with a process-based account of awareness — but the easiest way to see what that means in practice is to walk through the arguments first and the synthesis after.

The latter half of the post drifts somewhat naturally from free will into consciousness. The positive account I want to give of what will is ends up making claims about awareness and qualia, since the two pull on each other in ways that aren't easy to separate cleanly.

Common senses of "free will"

It's worth flagging up front that "free will" is used in at least two notably different ways in the philosophical literature, and more variably in ordinary speech. I'll give each a label and reference them in the arguments below.

Sense [C] — compatibilist. On the usage favoured by Hume, Frankfurt, Dennett, and most contemporary academic philosophers — and by many ordinary speakers in everyday contexts — "free will" means something like: acting on one's own desires and reasons without external compulsion or coercion. A person who deliberates, chooses, and isn't being forced is "free" in this sense. By design, this is consistent with determinism: it doesn't require the choice to be uncaused, only that it issue from the agent's own deliberation.

Sense [L] — libertarian / incompatibilist. On the usage common in metaphysical free-will discussions — and, I suspect, common when ordinary speakers press on what "really" makes a choice theirs — "free will" requires more: the genuine ability to have done otherwise, holding the prior state of the world fixed. The choice can't be the inevitable output of prior causes; alternative possibilities have to be really open at the moment of choice. This conception is incompatible with determinism by construction.

Sense [O] — other ordinary uses. Outside the philosophical disputes, "free will" is often used more loosely — to mean "absence of coercion," "control over my future," "responsibility-aptness," or just a vague gesture at agency without a clear metaphysical commitment. Whether any of these maps cleanly onto [C] or [L] is itself an open question.

I'm not endorsing any of these as the "right" meaning here — that would itself be a substantive move, and one I take a position on later. Most of the arguments below attach specifically to sense [L] (either defending it or arguing against it). The compatibilism cluster (Section 4) specifically negotiates between [C] and [L]. I'll mark each argument with the sense it bears on.

Part I: The arguments

There are more arguments in the literature than the ones I'll cover; I'm picking the ones I find most pivotal. For each, I include the gist of the argument, and a note on my position.

1. The phenomenology argument

Used in defence of [L].

The argument. Nearly everyone reports the felt experience of choosing between options, deliberating, and committing to one. That experience is itself a datum any adequate theory has to explain. If determinism eliminates choice, why does the experience of choosing not also disappear?

My position. I take the experience seriously. I just don't think the existence of an experience tells us what produced it. Visual experience presents us with images and motion; it doesn't present us with photoreceptors, optic nerve signals, or V1 cortical processing. Everything we know about how vision works we had to discover by means external to the experience of seeing. The same goes here. The felt sense of deliberation is real. It does not, as part of itself, deliver an account of how the deliberation produces the choice. So it isn't, by itself, evidence about whether the underlying mechanism is determined, random, or something else.

A further observation, based on my own experience in discussions about this, is that the primary motivation for most people who assert libertarian free will is exactly this kind of introspection-without-perceived-cause. They observe their own thoughts and deliberations — and they have privileged access to those in a way no one else does — and from inside, they don't see anything that looks to them like a "cause" for what's happening. So they infer there isn't one. The inference is the same one I'm rejecting above, but it's worth flagging because, in my experience, this implicit move shows up across both compatibilist and incompatibilist circles, and is much closer to the actual root of the conviction than any of the sophisticated philosophical arguments.

That said, this is genuinely an empirical claim about what other people are doing, and my warrant for it is only my own experience in such conversations. Settling whether the pattern actually holds beyond that sample would require real investigation — surveys, structured interviews, experimental-philosophy work. Treat it as a hypothesis I'm flagging, not an established finding.

A related defence shows up in many of the same conversations: the claim "I act as if I have free will." It's in the same phenomenology-defending family as the introspection appeal above — both treat some feature of the speaker's experience or conduct as evidence for free will — but where the introspection move appeals to an inner state, this one appeals to outward behaviour. I'd push back on the phrasing itself, though I want to be careful about claiming to know what people mean by it. I suspect at least two readings are common.

One reading: acting as if they have biases, beliefs, and desires, and as if those have an impact on the world. If that's the intended meaning, all of those things hold under determinism — the phrase isn't doing the work that people seem to expect.

Another possibility: the phrase implicitly conflates determinism with fatalism — taking determinism to mean deliberations don't matter, or have no impact on the future. That presupposition is wrong. Under determinism, your deliberations genuinely affect what happens next; they are part of the causal chain.

The two contexts where I usually hear "I act as if I have free will" are worth specifying, because they apply across either reading. (a) An interlocutor asserts it during a debate or deliberation, treating the very conduct of debating as evidence for libertarian agency. (b) A determinist themselves says something like "sometimes I forget I'm determined and act as if I have free will." Both usages — whichever reading is operative — presuppose that, on a determinist view, you wouldn't or shouldn't be deliberating, debating, or weighing biases. A determinist who uses the phrase themselves risks reinforcing the confusion — implicitly granting that determinism somehow weakens or eliminates agency, when it doesn't.

Either way, "acting as if I have free will" doesn't seem to pick out a distinctive behaviour. I can't think of any observable conduct that would distinguish a libertarian agent from a deterministic one acting on the same biases and beliefs.

2. The dilemma: determined or indeterminate?

Argues against [L].

The argument. Classical physics is deterministic; quantum physics is indeterministic. Neither, on its own, looks like "free choice." If an act is fully accounted for by prior causes, it isn't free. If it has some uncaused or stochastic component, it's just random. Neither matches what people seem to mean by "freely chose."

My position. I find this argument, and a couple of sharper restatements of it, decisive against libertarian free will.

The first sharpening — what I'd call causal incoherence — makes the problem about the concept of cause: causation seems non-sensical anywhere it is neither determined nor random. If a choice was caused, it was caused by prior states the chooser did not author. If it was uncaused, the word "cause" has lost its meaning, and so has the phrase "the agent caused the choice." There is no coherent third regime where causation does the work but isn't one of these two. Asserting one without specification just relabels the mystery.

A note on naming. This argument is most commonly referred to in the literature as the luck objection (or luck argument, luck problem); Al Mele's Free Will and Luck (2006) is the canonical book-length treatment. I prefer the "causal incoherence" framing for a substantive reason. The "luck" framing treats outcomes as the unit of analysis and implicitly suggests that every indeterminate output is, in some sense, lucky. But agents are systems, and systems can be organised to produce specific outputs in coordinated ways toward an end. We wouldn't say a computer is "lucky" to execute its program correctly — it's an organised system of matter producing a structured output. Calling the outputs of an organised system "lucky" misses what's happening. The "causal incoherence" framing avoids that by targeting the concept of cause directly: it asks the libertarian to specify what "cause" refers to in the proposed third regime — neither determined nor random — rather than letting them retreat to "whatever-it-is-it-isn't-luck."

A common dialectical move worth flagging: when pressed on causal incoherence, libertarians often retreat back to the phenomenology cluster — typically in the specific form of "I don't need any other evidence than my experience being the evidence." That closes the circle but doesn't escape it. The phenomenology appeal was already addressed upstream (experiences don't deliver their own causes), and reusing it as a response to causal incoherence asks the same observation to do work it can't do.

The second sharpening — denial of the prior self — runs the same point through personal identity, and makes more specific what the libertarian's "I could have done otherwise" claim actually requires. It isn't simply asserting a different outcome. It's asserting that I — at the moment of decision — could have had different motivations, different biases, different beliefs, different attention than I actually had. That is, it requires specific facts about me at that moment to have been other than they actually were. To assert "I could have chosen otherwise" given my actual state is therefore to deny that the actual me, with that actual state, existed as that specific person at that moment. The escape — that the same me, with the same state, could have produced a different output — concedes the dilemma: the difference would now be uncaused or stochastic, not authored by my control. And the libertarian would be just as blind to whatever was actually causing that difference as they are to the biological photoreceptors detecting different wavelengths of light during ordinary vision. The phenomenology can't deliver the mechanism in either case.

The same inconsistency surfaces in how libertarians usually describe choices. Someone will say a person "had a bias" or "had a preference" for one option, and in the same breath assert they "freely chose" another. But choosing against the asserted bias would require, at that moment, a different bias profile than the one the libertarian just attributed to the person. The framing holds together only by implicitly contradicting itself — granting the biases that would produce one outcome, then attributing the opposite outcome to a self that, by hypothesis, had different biases than the libertarian had just granted.

These are two angles on one underlying point. The libertarian needs to hold both I could have done otherwise and I am the same self that did it, and these two claims are in tension with each other.

3. The argument from reason

Used in defence of [L] (specifically against determinism).

The argument. A subtler family of arguments suggests that thoroughgoing determinism is self-undermining. If every belief is just the deterministic output of physical history, the "belief" that determinism is true is also just an output, with no rational-evaluative weight. Versions appear in Kant, C. S. Lewis, and (more carefully) some contemporary philosophy of mind.

My position. I don't find this argument persuasive. It relies on an unstated assumption — that deliberation could be something other than a serial, local, time-consuming causal process. Weighing competing beliefs, retrieving stored information, balancing biases, and building toward a threshold where an action becomes the output is exactly what physical systems do. The existence of a process called "reasoning" is not incompatible with determinism, and it isn't evidence for or against determinism either. A deterministic system can deliberate, and so can a stochastic one. The argument from reason gets its force only by tacitly comparing deliberation as a causal process with some non-physical alternative that nobody specifies. Once you stop comparing against that imagined alternative, the worry stops biting.

A useful analogy: how a computer generates output. At any moment, producing an output requires accessing relevant memory, applying stored rules and biases, integrating new inputs, and progressing through some sequence of operations until a result emerges. This isn't a quirky feature of computers — it's roughly what "producing an output" has to look like in any complex system that makes use of internal state. The system needs some pipeline that gets it from "has these inputs and state" to "has now produced this output."

The interesting question, the one that's reasonable to ask, isn't why such a process should exist — that's already addressed above; some pipeline of this shape is what generating an output requires. The reasonable question is why we are able to perceive this particular part of the process at all. Why does deliberation surface in awareness the way it does, when so much of the rest of the underlying machinery doesn't? That's a substantive question any account of consciousness has to address.

But once we grant that we do perceive it, whether the underlying process runs deterministically or with some stochastic component is a separate question entirely. A deterministic computer running through this kind of pipeline produces outputs; a stochastic one does too. The shape of the pipeline — retrieval, weighing, threshold-building, output — is consistent with both. So being able to perceive this process doesn't tell us anything about whether the underlying causation is deterministic or indeterministic. At best, all we can say is that we perceive it. From perception alone, we can't read off what the process actually is, or what's actually causing it — those are separate questions, not settled by being able to notice the process is happening.

4. Compatibilism and its hard-incompatibilist replies

Negotiates between [C] and [L]: which sense should the term carry?

The arguments. A different move on the metaphysical question: even granting determinism, maybe what we mean by "free will" (and "responsibility") is consistent with it. The hard-incompatibilist replies — there are several influential ones — argue that this concedes the substantive point.

• Compatibilism (Hume, Frankfurt, Dennett): free will, properly understood, is acting on one's own desires without external compulsion — not contra-causal magic — and that is fully realised even in a deterministic world. Frankfurt's cases are influential here: they argue moral responsibility doesn't even require alternative possibilities — even if you couldn't have done otherwise, you can still be responsible if the actual cause of your action was you.
• Moral responsibility presupposes free will: We routinely hold people responsible. That practice is either rationally grounded — in which case some kind of free will is real — or it is a mass error and our entire legal and moral system rests on a confusion. The latter is, at minimum, a heavy bullet to bite. (This is the everyday motivation for taking the compatibilist side rather than letting the metaphysics dissolve responsibility.)
• The consequence argument (van Inwagen, 1983): if determinism is true, every act is a logical consequence of the laws of nature plus the state of the universe in the distant past. We do not control the laws and we did not control the distant past, so we do not control the acts.
• The basic argument (Galen Strawson): to be morally responsible for an act, you would have to be responsible for the way you are at the time you act. To be responsible for the way you are, you would have to have caused yourself to be that way. Nothing is causa sui. Therefore no one is morally responsible.
• Manipulation arguments (Pereboom): imagine an agent who is (1) directly programmed by neuroscientists, (2) programmed at birth, (3) shaped by an unusually deterministic upbringing, (4) shaped by an ordinary upbringing in a deterministic universe. We'd typically not call (1) free; nothing morally relevant distinguishes adjacent cases; therefore (4) is also not free.

My position. I think this whole cluster, taken at face value, is largely a verbal dispute. Each of the hard-incompatibilist replies is making a structurally similar move: "whatever you call this thing, it isn't what we usually mean by free will or responsibility." Each compatibilist reply is the inverse: "no, this is what we usually mean." Both sides usually agree about the physics; they disagree about which phenomenon should wear the words.

It seems perfectly reasonable to me to adopt either sense of the term in specific contexts where it's pragmatic to do so. I don't think the choice of usage points to some metaphysical truth that one definition gets right and the other gets wrong; the terms are simply functional to different ends, and I'll switch which one I'm using based on what conversation I'm in.

The corollary I'd add: the belief that the term is truth-tracking — that one definition is correct and others mistaken — is itself a practical obstacle to productive conversation. Parties who insist on forcing one usage onto every interlocutor tend to bog the dialogue down arguing about word choice rather than the substantive issue underneath, which is rarely what anyone needed to resolve in the first place.

5. Reactive attitudes and the practice of holding responsible

Bears on whether [L]-style responsibility is needed for moral practice or [C] suffices.

The arguments.

• Reactive attitudes (P. F. Strawson, "Freedom and Resentment", 1962): Whatever physics says, we cannot in practice stop holding each other responsible. Gratitude, resentment, indignation, love, and guilt are too central to human life to be optional. P. F. Strawson argued this fact has its own normative weight independent of the metaphysics. (A different P. F. Strawson from Galen Strawson, who appears in Section 4.)
• Revisability (Pereboom and others): The practice of basic-desert moral responsibility can be revised without abandoning all the reactive attitudes that matter. We can keep gratitude, love, even forms of accountability, while giving up the idea that anyone fundamentally deserves blame or punishment in the metaphysically loaded sense.

My position. Both sides of this one are doing the same thing the dispute in Section 4 does — arguing about the metaphysics of a framework whose force comes from how it functions in practice. Reactive attitudes work because of the role they play in actual human life, not because they refer to some underlying metaphysical fact about desert. So the question "does determinism erode reactive attitudes?" is the wrong question — the language was never tracking that kind of metaphysics in the first place.

6. The argument from belief effects

Pragmatic, not metaphysical. Bears on belief in [L] (or sometimes the looser [O]) and its practical consequences.

The argument. Even if libertarian free will doesn't exist, undermining people's belief in it might have bad consequences. Vohs and Schooler (2008, Psychological Science, "The Value of Believing in Free Will: Encouraging a Belief in Determinism Increases Cheating") found that participants exposed to deterministic primes cheated more on a math task and overpaid themselves on a cognitive task, mediated by a decreased belief in free will. The proposed conclusion: believing in free will (or at least not actively undermining that belief in others) has prosocial effects, and this is a reason to be cautious about how the metaphysical case is communicated.

My position. Worth taking seriously, but two things take most of the force out of it.

First, the empirical findings haven't held up well. Notable replication attempts include Nadelhoffer et al. (2020), "Reconsidering the Value of Believing in Free Will" (Cognition), and the Many Labs 5 registered replication of Vohs & Schooler's Experiment 1 (Buttrick et al., 2020). Both used much larger samples than the original studies, with several preregistered designs, and largely failed to reproduce the original effects. The proposed link between manipulated free-will belief and immoral behavior turns out to be much less robust than the early studies suggested.

Second, even granting the original findings at face value, I think the practical worry rests on a confusion between determinism and fatalism — the same confusion that animates the "I act as if I have free will" pushback in Section 1. The bad-behavior effect, to whatever extent it's real, seems to come from people interpreting "determinism" as "my actions don't matter" or "nothing I do has consequences." That isn't determinism; it's fatalism. Under determinism, your deliberations, restraints, and consequences are all part of the causal chain — they all matter for outcomes, just deterministically. The right response to the practical worry isn't to preserve a metaphysical belief most people don't actually need; it's to communicate the determinist position in a way that doesn't invite the fatalistic reading, and to clarify what the position actually implies for ordinary practice.

So the argument doesn't push me to soft-pedal the metaphysics. It does push me to be careful about framing — not to talk about determinism in ways that invite fatalistic readings, and to take the time to spell out what the position actually does and doesn't entail.

7. The empirical case

Argues against [L], with caveats on each sub-argument.

This isn't a single argument but a cluster of empirical pressures. None is logically decisive on its own — they don't strictly rule out libertarian metaphysics — but each makes the gap libertarian free will would have to live in keep shrinking.

Neuroscience timing data (Libet, Soon, Haynes)

The argument. Readiness potentials in motor cortex precede the conscious experience of deciding to move by several hundred milliseconds. Patrick Haggard, John-Dylan Haynes, and others have replicated and extended the finding for more cognitive choices. The empirical claim is contested in detail — does the early signal truly fix the later choice, or just bias it? — but the broad pattern is robust: conscious choice is not the first thing in the causal chain.

My position. Interesting but not dispositive of determinism vs. indeterminism. Conscious choice not being first in the chain is consistent with both pictures; the temporal data leaves the underlying metaphysics undetermined.

The empirical-determinism case (Sapolsky)

The argument. Robert Sapolsky's Determined (2023) assembles the empirical case from neurobiology: behaviour at any moment is, on his account, exhaustively determined by the joint action of genes, prenatal hormones, childhood environment, recent stress, sleep, glucose level, peer effects, and a long list of other factors. None of these is up to you.

My position. Also interesting but not dispositive. The case is evidential, not a logical-necessity argument — a committed libertarian can in principle hold out against it. But it does drain the libertarian position of empirical support: every behavioural variation we can study turns out to have causes that aren't authored by the agent.

No special physics for brains

The argument. Many free-will defences depend, implicitly or explicitly, on brains being subject to physical laws different from those that govern rocks, computers, and weather. There is no independent reason to expect such an exemption, and lots of empirical reason to deny it. Whenever physicists carefully measure forces — orbits, beam deflections, spinning tops — they find no unexplained pushes. Outside of brains and bodies, there is no signal of an extra force. So the libertarian picture has to commit to physics working differently inside human skulls than it does anywhere else in the world, and to that special physics being undetectable by direct measurement everywhere else.

My position. I find this objection structurally decisive on its own. It's the argument that does the most work in shaping my view. The libertarian commitment is much stronger than the everyday version of the picture usually makes explicit: not just "we lack the data to predict the new trajectory" but "the additional force is not characterisable even in principle." That's a strong claim, and the rest of physics gives no independent reason to expect it.

8. The quantum gambit

A defence of [L] (the most-discussed contemporary version) and the structural critique against it.

The argument. Quantum mechanics is genuinely indeterministic at the level of single events. That seems, at least at first, to give the libertarian a venue: if some outcomes are not predetermined, perhaps the will lives in those genuinely undetermined gaps without violating any law. Penrose and Hameroff put this in microtubule quantum coherence; Eccles put it in synaptic quantum tunneling.

My position. Against, but not chiefly for the usual decoherence-based reason. The decoherence argument (Tegmark 2000) says brains are too warm and noisy to sustain the quantum coherence these proposals need. I don't have a basis for asserting whether that's right. Given my positions on the other arguments above, what I'd need to take a libertarian quantum proposal seriously is actual empirical evidence that something other than ordinary QM is happening — evidence that brain-scale events are not simply behaving with the predictable stochastic "footprint" we'd expect from standard quantum mechanics. Hypothesis isn't enough. Without that evidence, the libertarian quantum proposal isn't yet a reasonable position to accept; it's a placeholder waiting on data.

The argument I find more decisive is structural. Quantum field theory says more than "single events are random" — it predicts an exact probability distribution over them, and physicists have measured those distributions trillions of times. Any non-physical will that biased the per-event probabilities would accumulate, in aggregate, as a measurable statistical anomaly. Any will somehow clever enough to never perturb the aggregate would still need a continuous physical information channel to track the system — which is itself a measurable interaction. Either way, the fingerprint is detectable in principle, or the proposal collapses into pure dualism with no quantum content doing any work.

The unfortunate practical caveat is that actually performing the measurement on brain-scale processes — at the precision needed to detect the kind of bias the libertarian position would produce — may not be feasible with current or near-future instrumentation. I think that's a large part of why the position survives as an argument: the structural critique is sound on paper, but the empirical falsification remains out of reach in practice, so the libertarian can keep occupying the gap. The argument doesn't fail because of this; the position is still without positive evidence, and the burden of proof is still on whoever wants to assert there is one. But it does mean the dispute may not be settled by lab data any time soon.

Part II: My position

If you accept the arguments I find compelling — the phenomenological one as evidence-of-experience-without-evidence-of-mechanism, the dilemma (sharpened via causal incoherence and denial of the prior self) as ruling out libertarian metaphysics, the no-special-physics-for-brains objection as ruling out the libertarian implicit picture, and the structural pressure on the quantum gambit (along with what I take to be a suspicious motivation behind it) as deflating the most popular escape route — you arrive somewhere fairly definite about what is not true. You're left to figure out what is.

Here's my position.

The metaphysics: effectively determined

My best guess about the actual physical world is that something like determinism is most likely true at the level that matters. On the dominant interpretation, quantum mechanics is indeterministic at the level of individual interactions — some interpretations (Bohmian mechanics, certain readings of many-worlds) dispute this, but it's the standard view. At the scale of brains, decoherence, thermal averaging, and the law of large numbers seem to smooth out most of that indeterminacy long before it could plausibly steer a human-scale decision.

Even granting that the universe contains real randomness in places, I don't think it shows up where the free-will debate cares about it. The brain is, for the purposes of this kind of argument, an effectively classical system whose state at a given moment was produced by its prior state and its inputs.

The compatibilist / incompatibilist dispute is largely semantic

Once you grant the metaphysics, what remains in the standard philosophical debate is mostly an argument about which phenomenon should wear the words "free will." I don't think disputes about the "real" meaning of an ordinary-language term — taken in isolation from any specific context of usage — are productive. I'm also not convinced there is some single unified usage of "free will" to appeal to in the first place, and there is some empirical support for that skepticism. Experimental-philosophy work on folk free-will intuitions (Nahmias et al. 2006, Nichols & Knobe 2007, Murray & Nahmias 2014, Sarkissian et al. 2010) consistently shows significant variation in lay responses depending on how scenarios are framed; closely related work on folk metaethics by Michael B. Gill (Indeterminacy and Variability in Meta-Ethics, Philosophical Studies, 2009) makes the structurally identical argument that ordinary moral discourse is much less uniform than philosophers tend to assume. My guess, then, is that people use "free will" to accomplish particular ends, those ends vary across speakers and settings, and there isn't any meaning beyond what is functional in a given context. Words don't have essential meanings sitting in some Platonic space waiting to be uncovered.

The lineage here is Wittgensteinian. In Philosophical Investigations (1953) §43, Wittgenstein writes that "for a large class of cases … the meaning of a word is its use in the language." That formulation — the use-theory of meaning — has been refined and contested across philosophy of language since (Quine, Davidson, Brandom, ordinary-language philosophy more broadly), but the basic move remains widely held: a word's meaning is anchored in how it functions in particular linguistic contexts, not in a fixed external referent. Applied to free-will language, this is what underwrites my view that the compatibilist / incompatibilist dispute is largely about which usage to adopt, not about which one is right.

The corollary is that I will switch which definition of free will I'm operating with based on the conversation I'm in. If I'm talking with a committed incompatibilist, it's usually most useful to adopt the incompatibilist sense of the term — that's the conversation we're having. The same goes the other way around. Demanding a verbal correction first just delays the substantive question.

I don't deny that "will" exists

Hard-determinist positions are sometimes read as eliminativist about agency — as denying that anything of the everyday phenomena of deliberation, intention, and choice survives the metaphysics. That isn't my view.

Humans deliberate. They retrieve information, weigh beliefs and biases against each other, build toward an action threshold, and produce intentions that issue in behaviour. The phenomenon is real, and the word "will" picks it out fine.

What I reject is the additional metaphysical move libertarians make on top of the phenomenon — the claim that any real will has to be non-deterministic, and specifically has to escape physical characterisation in principle, not merely be perceived as such. A deterministic deliberation system that retrieves, weighs, and produces an output is still a will. You can praise it, blame it, argue with it, change its mind. The libertarian commitment is much stronger than what the everyday concept actually requires.

What I think will is

A complex processing system runs in human bodies. Sensory inputs and internal states trigger cascades of further activity; some of that activity surfaces in awareness, and a great deal of it doesn't. What I would call a "will" is one particular thing that processing produces — a desire or intention that shows up in awareness carrying some kind of motivational weight, then re-enters processing as another input. Will isn't the whole system. It's a specific product the system generates, surfaced in awareness, looped back as further input.

This view is broadly functionalist: mental states are defined by their causal/computational role, not by any specific physical realisation. The underlying processing might be substrate-independent in principle. But I'd put a couple of refinements on the standard functionalist picture.

First, on multiple realizability: I'm conceptually fine with it. The view that you could in principle implement a mind in something other than wet biology doesn't strike me as obviously wrong. What I'd flag is that "in principle" buys less in practice than it tends to sound. A neuron's contribution to any specific mental state isn't determined by the neuron in isolation; it's determined by the rest of the system it's wired into. Reproducing a hundred neurons firing in a particular sequence wouldn't reproduce a specific thought without also reproducing the connectivity and the surrounding system that gives those firings their functional role. So practical realisability of a specific thought, in a different system, ends up requiring something close to recreating the original system — and at that point the question of whose self the thought belongs to gets thorny.

There's also a related concern about analog vs. digital substrates. Biological neural systems are a combination of electrical, chemical, and mechanical processes, which we can consider analog more or less generally — and analog systems aren't always cleanly reproducible the way digital ones are. The specific kind of experience humans actually have may turn out to require the analog dynamics of that combination in a way that digital substrates can't straightforwardly mimic. That isn't an in-principle barrier to other forms of experience — some other form of experience, in a different substrate, may well be possible. It just may not be the same kind of experience.

Second, awareness isn't a momentary state. Even a single word during deliberation takes time to unfold — neurons firing in sequence, alongside other processes — so a thought is better described as a series of events stretched out over an interval, not a state at an instant. A given concept doesn't even need to be realised by exactly the same neurons every time; different firing patterns in the same brain can plausibly do the same conceptual work.

If I were forced to pick a label, non-reductive or wide computational functionalism with a process-based account of consciousness is the closest fit — but the easier description is the one above: will and awareness are both features of how a complex processing system runs in time, neither is the whole system, and neither needs an immaterial or non-deterministic ingredient on top.

I'm not arguing for illusionism either

The strongest form of illusionism (Frankish, some readings of Dennett) denies things like qualia — claims that introspection systematically misrepresents the existence of phenomenal experience. I wouldn't go that far. I don't deny qualia. What I'd say instead is that qualia and experience aren't, or don't need to be, static atomic things. They could plausibly be processes unfolding over time within a complex system, and that's how I'd expect them to turn out on closer inspection. That is a structural reframing, not a denial. (Unveiling the Incommunicable makes a related move on a related topic — explaining why awareness has the limits it does without invoking either illusion or new physics.)

On moral practice

A common worry is that determinism erodes moral practice — that if everything was always going to happen, we have no business holding each other accountable, praising, blaming, or treating each other as agents.

I don't think that worry survives the deflationary view applied consistently. It presupposes that moral language was supposed to be tracking a metaphysical fact in the first place. On the view I'm defending, it isn't. Moral terms work the way "free will" works: pragmatically, in context, in the service of communication and coordination. The metaphysics doesn't set the practice. The practice is the practice.

So the deflation runs all the way through. Whether you call physics "deterministic," whether you call your last decision "free," whether you call your neighbour "responsible" for what they did — these are all instances of language being used in context, with purposes, doing work. None of them are reading off facts about the cosmos's metaphysical fine structure. The question of how to use them is settled by what we're trying to do with them, and that question is real and answerable, just not in the place the standard debate keeps looking for it.

Closing

If this sounds like fence-sitting, I'd argue it isn't. The standard ways of being un-fence-sitting in this debate (committed libertarianism, full-strength compatibilism, full-strength illusionism, full-strength eliminativism) all turn out, on inspection, to be making metaphysical commitments stronger than the evidence justifies. The view I'm defending is decidedly less committal — but it's committal in specific ways:

• on the physics: effectively deterministic at the level that matters;
• on quantum-mediated free will: I'm not prepared to say the structural argument from statistical footprints closes the question experimentally — at least in principle, future evidence could go either way. What I'm skeptical of is the motivation for the position. It looks much more like a search for explanation in a place where something currently can't be tested, applied to a phenomenon (deliberation, choice) that doesn't need further explanation. That "quantum" consciousness and "quantum" free will tend to come bundled together is evidence of this: ordinary deterministic processes already give a perfectly good account of what we experience, and the bundling looks like proponents looking for unexplained mystery to explain things that aren't mysterious;
• on the compatibilist / incompatibilist debate: largely a verbal dispute about which phenomenon should wear the words;
• on agency: real, in the form of a particular output of a complex processing system, not as anything immaterial or non-deterministic;
• on awareness and qualia: real, but better understood as processes unfolded over time, not as static atomic things;
• on moral practice: it survives, because it was never tracking the metaphysics determinism would or wouldn't erode.

That's the position. Whether or not you find it convincing, it has the virtue of being specific about where it's committed and where it isn't.