Expensive Will

Why the most popular argument against free will might be wrong

Have you ever changed your mind?

Not been persuaded. Not received new information that updated a conclusion. Changed your mind. Been mid-decision — felt the momentum running toward one option, felt the weight of habit and emotion and self-interest pulling you in a direction you could already see was wrong — and shifted.

Not easily. Not because you wanted to. But because some part of you caught the process in motion and altered it from the inside.

For me this typically happens when I am about to get angry, realise I have a choice not to, and so I avoid that state.

If you have experienced this, then you get what free will is. But one of the most influential public intellectuals of the last decade says that free will is an illusion. You are apparently deluding yourself. I am not so sure.

Last week Australia Post sent me a notification. Failed delivery. Their driver claimed no one was home.

Both Sarah and I were working from home that day. The doorbell was not rung. We know this because we were both in the house, ten metres from the front door, all day. The driver did not attempt delivery. He marked it as failed and left the parcel at the local post office, which meant I now had to drive to collect something that was supposed to arrive at my door.

This is not unusual. Couriers do this because last-mile delivery is expensive. It is cheaper to dump parcels at a collection point than to walk up a driveway and ring a bell. Every other industry understands that a contract to do something means you have to fully complete a scope of work to a specification. Only this industry thinks that they can ignore job specification — that diverting to a different destination and calling it a failed delivery is acceptable service.

I could feel the anger building. Not as a thought. As a physical state. The tension that is gearing up for a conflict. The momentum running toward a very specific kind of phone call — the kind where you are technically correct about everything and manage to make the person on the other end defensive, which means nothing gets resolved, which makes you angrier, which confirms that the system is broken and nobody cares.

I have made that phone call too many times. About different things, to different organisations, with the same result. The anger is righteous. The anger is also useless. L1 produces the heat, L2 adds the wound — nobody respects the agreement, nobody delivers what they promise, you always have to chase — and L3 builds the case: you are right, they are wrong, let them have it.

(If L1, L2, and L3 are unfamiliar, see Too Clever by Half for the full framework. Short version: L1 is the body's survival system, L2 is the emotional and social landscape, L3 is the rationalising mind. They interact in ways that matter for what follows.)

Three conditions are required for anger to be appropriate. Something unfair or unjust or threatening has occurred. You cannot just ignore what happened. And you have to be the most focused, energised, pain-resistant version of yourself to deal with it. The superhuman version of yourself that can go too far and become psycho if you are not controlled.

Condition one was met. Condition two — maybe. I could have just driven to the post office and collected the parcel. It would have taken fifteen minutes. The world would not have ended. Condition three was not met — because the superhuman focused version of angry me is not actually more effective on that phone call than calm me. Calm me states facts. Calm me asks for what he wants. Calm me gets results. Angry me gets conflict.

And then I noticed.

Not the content of the anger. The momentum. I watched it the way you watch a wave forming — from slightly outside, with the part of me that has learned, at great expense, to observe my own process while it is happening. The anger was real. The grievance was legitimate. The driver genuinely did not ring the bell. And in the moment of noticing, the weights shifted. Not because I decided to be calm. Because observation changed the system. I called Australia Post, stated the facts without heat, got an apology and a note on the account. Fifteen minutes. No adrenaline hangover. No ruined afternoon.

That experience — banal, private, unremarkable to anyone watching — is what Sam Harris says cannot happen. Or at least, he says the part where I intervened in my own process is an illusion. The anger would have resolved however it resolved, and my sense of having redirected it is a story I told myself after the fact.

I think he is wrong. And I think he is wrong in a way that matters for everything Paragentism is trying to do.

The Best Argument You Will Ever Lose

Harris's case against free will is possibly the strongest version of that argument ever constructed. It deserves to be presented at full strength, because anyone who dismisses it without feeling its weight will never understand what it actually costs to disagree.

Sit down. Close your eyes. Watch. A thought appears. You did not decide to think it. It arrived, unbidden, from somewhere you cannot access. Watch another. Same thing. Sit for an hour and every single thought, emotion, impulse, and intention will appear in consciousness without your permission. This is not theory. This is direct first-person observation available to anyone who sits still for ten minutes. It is one of the most replicable experiments in human experience.

Now make the harder move. Look for the Witness that is watching these thoughts. Turn attention back on itself. Where is the self that supposedly decides? You will find awareness. Openness. But no entity behind awareness pulling levers. The homunculus — the little person inside your head who you imagine is running the show — cannot be located. Not because it is hidden. Because it does not appear to be there. But this is a bit like trying to touch the tip of your nose with the tip of your nose. Structurally impossible in one sense and silly in another sense because the tip of your nose is already in touch with itself.

This is not just meditation. The neuroscience converges. Libet's experiments showed the brain producing readiness potential before the conscious experience of deciding. Your brain has already committed to a motor action hundreds of milliseconds before "you" feel like you have chosen it. Haynes pushed this further with fMRI — predicting a person's decision up to ten seconds before they were consciously aware of making it. Ten seconds. That is not a timing discrepancy. That is the conscious self arriving late to a party the brain already threw.

Then the philosophical layer, and this is where it becomes genuinely difficult to refute. Every decision you make is the product of factors you did not choose. Your genetics — not chosen. Your parents — not chosen. Your childhood — not chosen. Your neurochemistry — not chosen. The specific configuration of fears and wounds that shape every reaction you have — not chosen. Even the moment of "choosing" — what caused you to notice at that particular instant? Prior states. Prior causes. An infinite regress of things you did not author, producing the next thing you call "my decision."

And here is where Harris is at his most seductive. If free will is an illusion, then hatred is incoherent. You cannot truly hate someone for actions they could not have done otherwise. The murderer was produced by causes — genetics, abuse, brain chemistry, circumstance. You restrain them the way you would contain a flood — with clear eyes and without vengeance. This produces more compassion, not less. You stop blaming people for being the output of processes they did not design. The world becomes kinder when you understand that everyone is doing the only thing they could have done given everything that preceded this moment.

That is the argument. It is elegant. It is internally consistent. It is empirically supported. It is experientially verifiable. It contradicts nothing in physics. It explains everything.

Sit with it. Feel the weight of it. Because what follows only matters if you have taken Harris seriously enough to understand what you are arguing against.

The Amp

Harris built his case from meditation. From sitting. From the mind at rest, receiving whatever arises, authoring nothing.

In that mode, everything he says is confirmed. Thoughts arrive. No author is found. The self dissolves under inspection. The Witness — his own term for the singular point of conscious awareness — just watches. The show runs itself.

But here is the question I do not think he asks often enough: is that the only mode?

An amplifier must amplify. That is what it does. Turn it on with nothing plugged in and it will pick up the 240-volt mains hum from the electrical wiring in the walls — because an active amplifier with no signal input cannot do nothing. It amplifies whatever is there, even if all that is there is noise.

The brain is similar. It has to process. A brain deprived of input finds this intolerable and generates its own content to work with — the way nature abhors a vacuum, or the way dividing by zero breaks mathematics. The thoughts that arise unbidden during meditation are not evidence that the brain only produces unbidden thoughts. They are evidence that a brain with no input will create some, because a processing system that must process cannot sit idle.

Harris observed the brain in input-deprived mode and documented what he found with extraordinary rigour and extraordinary honesty. And then — I believe — he declared the noise floor is the ceiling. The mind in receive-only mode is the mind in its only mode. Thoughts arise unbidden, therefore all mental events are unbidden. The amp hums, therefore the amp cannot play music.

My assertion is different. We observe the secretion of thoughts during meditation as a protection mechanism for a system that must be active. But when there is real stimulation — a live decision, a genuine stake, a problem that requires engagement — the brain operates in a different mode than when it is deprived of input.

And there is a subtlety here that matters. Just before a thought arises, there is a short space of no thought. You can only notice this after the thought has appeared — literally, there is silence between the thoughts. This silence is not nothing. It is the Witness in its resting state. And the practice of noticing that silence — catching the gap between thoughts — is how you cultivate the Witness to have the capacity to provide choice. The gap is where the finger meets the scale.

The Witness Breaks Its Own Argument

This is where I think Harris's framework runs into trouble using its own best concept.

The Witness. Harris insists it is real. It is the foundation of his entire meditation teaching, his app, his public philosophy. There exists a singular point of conscious awareness that can observe thoughts, emotions, sensations, and the self itself. The Witness is what remains when you strip away everything that can be observed. It is the clearing in which experience happens.

In passive mode — meditation, sitting, idle — the Witness watches. Thoughts arise, the Witness notes them, they pass. The Witness does not intervene. Harris is right about this. In passive mode, the Witness is a spectator and the show runs on autopilot.

But the Witness does not only exist on the cushion.

When you are engaged with a live decision — when options have surfaced and momentum is building toward one of them and the outcome matters — the Witness, in my experience, does something Harris does not adequately account for. It observes the momentum. And the observation changes the momentum.

This is not mysticism. It is not even controversial in any other domain. Observation changes systems. A team performs differently when watched. A market moves differently when monitored. A negotiation shifts when one party becomes aware of their own pattern. The act of attending to a process alters the process. Harris knows this about every system in the world — and he asserts that the Witness has a choice to mindfully witness. I am not convinced he has fully reckoned with what that means for his own case.

The Witness, in active deliberation, is not a spectator. It is a participant. Its engagement with the decision process introduces a new causal factor — awareness of the process itself — that changes the weights. The option that was winning on autopilot may not win under observation. Emotional charge gets contextualised: that is L1 fear, not L3 insight. New options become visible. The momentum shifts. Not because a ghost in the machine intervened from outside the causal chain. Because the Witness is part of the causal chain. And when it is active, the chain runs differently than when it is passive.

I think Harris may have found the mechanism of agency and mistaken it for proof of agency's absence. He found the guitar and concluded it was more mains hum.

Three Moments

Here is what actually happens when I make a decision. Not what the theory says. What I experience.

Moment One — Momentum. First, options surface. I do not claim to author them. They arise from memory, pattern recognition, associative processing, habit, imagination — all of it below conscious control. Harris can have this. I did not choose what occurred to me. The menu appeared without a chef. One option is winning. I can feel the pull. The pull has texture — excitement, anxiety, familiarity, avoidance. It is not neutral. It is not rational. It is the accumulated weight of every prior state Harris correctly identifies. Genetics, childhood, imprinting, neurochemistry. All of it is present. All of it is pushing.

Moment Two — The Witness Activates. Something in me steps back. Not always. Not reliably. Not on command. But sometimes — and more often now than ten years ago, which is itself evidence of development — the Witness catches the momentum in flight. It does not stop the momentum. It contextualises it. That pull is fear. That excitement is avoidance dressed up. That familiarity is the pattern that trapped you last time. The momentum does not disappear. But it is no longer the only force acting.

Moment Three — The Shift. The decision changes. Not dramatically. Not heroically. The weights adjust. The option that was winning on autopilot loses a few points. Another option — one that was present but suppressed — gains visibility. The outcome is different from what it would have been without the Witness's engagement.

Harris would say: the Witness's activation was itself caused. Prior meditation practice caused the neural pathway. Prior failures caused the pattern recognition. The "shift" was just another determined event in the causal chain, experienced from the inside as choice.

Fine. I accept that the Witness was developed by prior causes. I accept that its activation in any specific moment has preconditions. I accept that the entire system is causally embedded.

But here is what I do not accept: that "causally embedded" means "no agency." A thermostat is causally embedded. It still regulates temperature. Its regulation is real — it changes the system it is part of. The fact that the thermostat was designed, installed, and calibrated by prior causes does not make the regulation illusory. The room is actually warmer because the thermostat exists.

The Witness is a thermostat for decision-making. It was built by prior causes. It operates within the causal chain. And when it is active, decisions go differently than when it is not. That is not an illusion. That is a function.

The Finger on the Scale

I am not claiming libertarian free will — the uncaused cause, the ghost in the machine, the self that exists outside physics and intervenes from nowhere. Harris demolished that. It was always incoherent.

I am claiming something smaller and, I think, more defensible.

Agency is not the ability to choose from outside the causal chain. It is the ability to develop an internal function — the Witness — that, when active, changes the weights within the causal chain. You cannot choose your genetics. You cannot choose your childhood. You cannot choose your initial neurochemistry. But you can, over time, develop the capacity to observe your own process in real time — and that observation is not neutral. It changes outcomes.

This is not free will. It is expensive will. It costs years of practice, failure, and the willingness to sit with discomfort rather than act on autopilot. It is not available equally to everyone — some people have better starting conditions for developing it. It is not reliable — the Witness fails more often than it succeeds, especially under stress. It is not complete — even when it works, it shifts weights rather than overriding them.

But it is real. And it is developable. And the fact that it was developed by prior causes does not make it illusory, any more than the fact that a muscle was built by prior exercise makes the muscle illusory.

Harris's move — declaring that because the Witness was caused, its effects are not "real" agency — proves too much. By the same logic, no function in any system is real. The heart does not really pump blood because it was caused by genetics and development. The immune system does not really fight infection because it was shaped by evolution. These claims are absurd. Functions are real even when their origins are causal. The Witness is a function. Its effect on decision-making is real. Calling it an illusion because it has causes is a category error.

Libet — Revisited

Libet's experiments show the brain initiating motor actions before conscious awareness. Harris presents this as devastating evidence against free will. But let us look more carefully at what was actually measured.

Libet asked subjects to perform a simple action — flex a wrist — and report when they felt the urge to move. The readiness potential preceded the conscious report by about 350 milliseconds. The action is already underway before "you" decide.

But Libet himself noted something Harris tends to underemphasise. Subjects could veto the action after the readiness potential had fired. The brain initiated movement; the conscious self could still abort it. Libet called this "free won't" rather than free will. The readiness potential fires, but the Witness can catch it and say no.

This is exactly the mechanism I am describing. Not uncaused choice. Causal intervention. The Witness does not initiate from nothing. It monitors what the system has already initiated and adjusts. That is not the libertarian free will Harris demolished. It is the functional agency I am asserting.

Haynes's ten-second prediction window is more challenging. But even there, the prediction accuracy was around 60% — better than chance, far from certainty. Forty percent of decisions went against the brain's "commitment." That is not a clean kill of agency. That is a system where momentum is strong but not absolute — where something, sometimes, shifts the outcome against the predicted trajectory.

I am suggesting that the something is the Witness. And that Harris, having correctly identified the Witness as the deepest structure of consciousness, inexplicably refuses to grant it any causal power in the system it observes.

The Cost of Harris Being Right

If Harris is right — if agency is truly illusory — then Paragentism collapses. Not just mine. Every framework that asks people to take responsibility for their own development, their own choices, their own engagement with systems. If you cannot actually steer, then steering frameworks are just more moral theatre.

Harris would not flinch at this. He would say: yes, some people will read my work and behave differently. That behavioural change was caused — by reading the book, by the neural impact of the ideas, by everything that preceded the reading. The change is real. The sense that "I chose to change" is the illusion.

But here is the problem with that move. If the sense of choosing is an illusion but the behavioural change is real, then the illusion is functional. It produces different outcomes than its absence. A person who believes they can steer — who acts as though the Witness has causal power — makes different decisions than a person who does not. Even if the belief is "caused," its presence in the system changes the system.

Harris has to either accept that the belief in agency is a useful causal factor — in which case it is not merely an illusion but a functional component of the decision-making system — or insist that the belief makes no difference, which contradicts his own observation that meditation (a practice of cultivating the Witness) changes behaviour.

He cannot have it both ways. If developing the Witness changes outcomes, then the Witness has causal power. If the Witness has causal power, then calling it an illusion is mislabelling a real function to preserve a philosophical commitment.

So choose.

Growing Up

If Expensive Will is real — if the Witness can intervene and the intervention changes outcomes — then the question becomes: what develops this capacity? And what prevents it from developing?

This is not a stage model. There is no Hero's Journey. You do not pass through dependence to independence to interdependence like levels in a game. You grow up or you do not. In fits and starts. Sometimes gradually. Sometimes not at all for years. Sometimes backwards.

But there are descriptions of where people get stuck, and these descriptions are useful even without a map.

Dependence. The system should work. Should be fair. Should reward merit. Should apply rules equally. When it does not, you are outraged — not because you are naive, but because you are still operating from the assumption that someone is responsible for making things right. Outrage is the emotional signature of dependence. It assumes a parent. The Witness is asleep. Momentum runs unchecked. L1 and L2 produce the reaction and L3 calls it righteous anger.

Independence. You have separated from the system emotionally. You see through it. You comply strategically but no longer believe. This feels like maturity — "I see through the bullshit" is the most seductive waystation in human development because it pattern-matches to wisdom. But it is not wisdom. It is L2 pain — prestige and belonging not being met — recruiting L3 to build a fortress around the wound. The Witness is active but co-opted. It observes accurately and then L3 constructs an impregnable analysis of why the world is broken and action is pointless. The system is rigged is less painful than I am not being chosen. The cynic has built a prison out of correct observations and called it a watchtower. This is the Too Clever by Half failure mode. The intelligence that should produce agency instead produces sophisticated paralysis.

Interdependence. Re-engagement from structural strength. Not because the system deserves it. Not because you have had a developmental breakthrough. Because the desert is worse than the game. Outside of society is a bleak existence and a desert of purposelessness. Fuckwittery is stability — it is just shit stability if you are high-agency and not an elite.

The Witness, in this mode, does the expensive thing. It observes the L1 fear and does not obey it. It observes the L2 wound and does not build a fortress around it. It observes L3 constructing rationalisations and does not mistake them for truth. And then it acts. Not freely. Not easily. Not reliably. But differently than it would have acted on autopilot. The finger on the scale.

The capacity that makes this possible — the thing that distinguishes the person who grows from the person who stays stuck — has no good name. Every candidate has been colonised. Self-mastery implies domination. Mindfulness implies calm. Self-authoring has a price tag and a Jordan Peterson brand. Sovereignty implies isolation. Self-control implies suppression.

It is the capacity to have the L1/L2 response and not be the L1/L2 response. To relate to your internal states as data rather than as truth. Agency applied inward before it is applied outward.

I cannot name it without reducing it to something smaller than it is.

But I can tell you what it costs. It costs every comfortable story you ever told yourself about why you are stuck. It costs the fortress the cynic built. It costs the outrage the dependent lives on. It costs the peace Harris offers. It costs the illusion that someone else is responsible for what happens next.

It is not free will. Free will is a philosopher's abstraction. This is earned will. Developed will. Will that was built by decades of getting it wrong and occasionally, at great expense, getting it right.

Expensive Will.

I am still developing the capacity. Because I choose to.