Free will
A Socratic walk-through of free will — reasoned out one step at a time, not lectured.
The question we started with
THE QUESTION #If every choice has a cause behind it, in what sense is any choice yours?
You chose tea over coffee. Ask why, and there is always an answer: you slept badly, you had coffee yesterday, you were raised in a house where tea meant comfort. Keep asking and the chain runs back past your birth, and nothing in it was chosen by you, because there was no you yet to choose.
That is uncomfortable, and the reflex is to look for a gap in the chain — some point where the causes run out and you step in. But before hunting for the gap, it is worth asking what we would do with one if we found it.
Reasoning it through
REASONING #Set out the worry as a clean fork, because it is sharper than it first appears.
Suppose determinism is true: the state of the world at any moment, plus the laws of nature, fixes every later state. Then the state of the world before you were born, together with the laws, entails that you would reach for the tea. You have no power over the distant past and none over the laws — so, Peter van Inwagen argued, you have no power over what they entail. That is the consequence argument, and it depends on no picture of the brain. It only needs entailment.
Now suppose determinism is false, as quantum mechanics may suggest at some scales, and some events are genuinely undetermined. Does that help? Imagine the undetermined event sits right where your decision does, so that in identical circumstances you might have reached for coffee instead. Then which one you took was, at the crucial moment, a matter of chance — and chance is not control. A choice that flips on a subatomic coin is no more yours than one dictated by physics. This is the luck objection.
So the trilemma closes: either your choices are settled by things you did not choose, or partly random, and neither looks like authorship. Hard incompatibilists such as Derk Pereboom accept that and argue we should give up the free will that grounds deserved blame.
Here is where most readers expect the argument to end, and where academic philosophy mostly disagrees. Around three in five philosophers in the 2020 PhilPapers survey accept compatibilism — the view that free will and determinism are consistent, because the argument above quietly assumes a definition of freedom nobody needs.
What does the compatibilist substitute? Roughly: an action is free when it flows from you in the right way — without coercion, compulsion or manipulation — and when you are responsive to reasons, so that had there been good reason to choose coffee, you would have. On this view being caused is not the opposite of being free. Being caused by the wrong things is. A person acting on their own considered desires and a person acting under hypnosis or addiction are both fully caused; the difference between them is real, and it is what our practices of praise and blame have always tracked. Harry Frankfurt refined this by asking whether you endorse the desire you act on — the addict who wishes he were not one is unfree in a way the contented addict is not.
The libertarian holds out for something stronger: genuine alternative possibilities, secured by indeterminism placed carefully enough to escape the luck objection. Robert Kane's version puts it at moments of inner conflict, so that whichever way you resolve them, the resolution is your own effort. Whether that answers the luck objection or relocates it is exactly what is argued about.
Two words about the experiments everybody cites. Benjamin Libet found that a scalp-recorded readiness potential begins building several hundred milliseconds before people report becoming aware of an urge to flex a finger, and popular accounts concluded the brain decides before you do. That is over-read on several counts. The task was an arbitrary flick with no reasons on either side, which is not the sort of decision the debate is about. The reported moment of awareness is measured by having subjects note a spot on a fast clock, which is itself unreliable. And Aaron Schurger and colleagues have argued the readiness potential is not a decision signal at all but the visible part of ordinary fluctuating neural activity drifting until it crosses a threshold — on that reading, averaging trials backwards from the movement manufactures the ramp. Libet himself thought a conscious veto remained. Whatever settles this question, it will not be that experiment.
The analogy
THE ANALOGY #Consider a river reaching the sea through a delta. Which channel the water takes is fixed entirely by the terrain, the sediment and the flow — nothing chooses. Yet we still distinguish a river that has cut its own bed over centuries from one shunted into a concrete culvert. Both are fully determined; only one is doing what the river's own history and structure dispose it to do.
a river has no perspective on its channel, whereas the compatibilist's central move depends on your being able to reflect on your desires and be moved by reasons — so the analogy shows how "caused" and "unconstrained" can coexist, and shows nothing about whether that is enough for responsibility.
Clarifying the model
THE MODEL #The dispute is largely about what "free" should mean, not about the physics. Nobody in it denies that choices have causes. The incompatibilist says freedom requires being able to do otherwise given exactly the same past; the compatibilist says it requires being able to do otherwise if you had wanted to, and that this is the sense our practices care about.
Determinism is also not fatalism. Fatalism says the outcome arrives whatever you do; determinism says your deliberation is one of the causes that produces it. Deliberating is not theatre on a determined view — it is a link in the chain, and removing it changes the result.
And the practical stakes attach to responsibility rather than freedom in the abstract. Even philosophers who deny free will mostly keep forward-looking practices — protection, deterrence, reform — while questioning backward-looking desert. That distinction is where the live argument now sits.
A picture of it
THE PICTURE #How to readstart at the rounded terminal and take the first diamond — both branches lead somewhere unwelcome, which is the trilemma. The slanted boxes are the two objections, one per branch, and the bracketed box is the libertarian's attempted repair, which rejoins the main line rather than escaping it. Everything funnels into the second diamond, which is not a fact about the world but a choice of definition, and the two terminals below are the positions that choice delivers. The dashed back-edge is the live loop: denying desert sends you straight back to arguing about what freedom had to mean.
What became clearer
WHAT CLEARED #The problem is not a missing fact about the brain. Both horns of the trilemma are arguments, and the escape most professional philosophers take is to deny that freedom ever required an uncaused choice — while a serious minority insist that anything less is not the thing we wanted. Libet's experiments do not decide it, and the question remains genuinely open.
Where to go next
ONWARD #- Frankfurt cases, which try to show you can be responsible for an act even when you could not have done otherwise.
- What happens to punishment, apology and gratitude if backward-looking desert is abandoned but forward-looking practices are kept.
Key terms
TERMS #| Term | What it means |
|---|---|
| Determinism | the thesis that the state of the world plus the laws of nature fix every later state. |
| Consequence argument | van Inwagen's argument that since we control neither the past nor the laws, we control nothing they entail. |
| Compatibilism | the view that free will requires freedom from coercion and compulsion, not freedom from causation. |
| Readiness potential | a slow build-up of scalp-recorded brain activity preceding voluntary movement, central to Libet's experiments. |
Every term the collection defines is gathered in the glossary.