Daniel Dennett

1. Free will worth wanting

Most arguments about free will begin with a definition: free will is the power to have done otherwise in exactly the same circumstances, or to be the uncaused cause of one's own actions. Then they ask whether anything in nature could have that power. Usually the answer is no, and free will is declared an illusion.

Dennett turned the question round. Before asking whether we have free will, he asked what we want it for. We want to be able to plan, to weigh reasons and act on the best of them, to avoid harm we can see coming, to be held responsible and to hold others responsible, and to be the authors of our lives rather than puppets. Each of these, he argued, is something a well-designed physical system can have, and some systems have more of it than others. A free will that would require a break in the laws of physics adds nothing to that list. It is a free will nobody has ever needed.

That is the sense of his subtitle, the varieties of free will worth wanting. The philosopher's ideal of an uncaused choice is like a perpetual motion machine: impressive on paper, impossible in practice, and not what anyone's life actually depends on.

2. Bugbears

Why does determinism seem so frightening? Elbow Room argues that the fear is fed by a gallery of bogeymen: vivid images that pump our intuitions in the wrong direction. There is the invisible jailer, who keeps us in a cell we cannot see. There is the nefarious neurosurgeon, who has wired our brains to his own ends. There is the puppeteer, and the bogeyman who knows our every move in advance. Each image takes a real threat to freedom, being controlled by another agent, and smuggles it into the bare idea that our actions have causes.

But causes are not controllers. The laws of physics do not want anything from us. A clockwork world does not contain a hidden agent who is playing us like a fish on a line. Once the bogeymen are dismissed, Dennett says, much of the anxiety goes with them.

The most memorable bugbear is an insect. The digger wasp Sphex drags a paralysed cricket to its burrow, leaves it at the threshold and goes inside to check the burrow. If an experimenter moves the cricket a few inches while she is inside, she brings it back to the threshold and goes in to check again. The story, as it was told to Dennett, is that she will do this forty times without ever thinking of simply dragging the cricket straight in. Her behaviour looks purposeful, but it is a loop. The fear is that we are all like Sphex: mechanisms whose apparent wisdom would fall apart if anyone tested it.

Dennett's answer is that the difference between Sphex and us is real, and it is a difference of design. We can notice that we are in a loop, step back and try something else. Sphexishness is a matter of degree, and human beings sit at the far, flexible end of the scale. (Later fieldwork suggests even the wasp was less rigid than the famous story says, which rather strengthens the point.) Being a mechanism is not the problem. Being a stupid mechanism is.

3. Determinism is not inevitability

In Freedom Evolves Dennett used John Conway's Game of Life, invented in 1970, to make the point vivid. The Life world is a grid of cells, each alive or dead. At every tick a single, completely deterministic rule decides the next state of each cell from its eight neighbours. Nothing random ever happens. Yet from that rule emerge stable shapes, travelling shapes called gliders, and "eaters" that consume gliders and repair themselves. At the level of cells, everything is fixed. At the level of gliders and eaters, there are things that avoid, survive and destroy.

The lesson is that determinism is a fact about the bottom level of description, while freedom is a fact about higher levels. "Determined" does not mean "inevitable". Something is inevitable if it will happen whatever you do. In a deterministic world, a great many things happen only because of what you do, and would not have happened if you had been designed differently. Determinism says that the future is fixed by the past. It does not say that your deliberation plays no part in fixing it.

4. Avoiding the baseball

Consider a simple act of avoidance. A ball is thrown at your head, you see it coming, and you duck. Did you avoid being hit? Of course you did. Was being hit ever going to happen? In a deterministic world, the answer seems to be no: given the past, you were always going to duck. So what did you avoid?

Dennett's reply is that avoidance does not need a future that was really going to happen. It needs a projected future, the trajectory your brain extrapolates from what it sees, and a system that acts on the projection. That is what eyes and nervous systems are for. A creature with no eyes cannot avoid anything; one with good eyes and quick reflexes avoids things all the time, deterministically. Freedom, in its humblest form, is just this: the capacity to see what is coming and do something about it.

The same deterministic laws run in both versions of the figure. What differs is the design of the creature. That, for Dennett, is where freedom lives: not in exemptions from the laws, but in the capacities the laws make possible.

5. Could have done otherwise

The traditional test of free will is that the agent could have done otherwise. If determinism is true, then in exactly the same circumstances, atom for atom, you would always do exactly the same thing. So, the argument goes, you never could have done otherwise, and you are never free.

Dennett borrowed an example from the Oxford philosopher J. L. Austin. Austin misses a short putt and says he could have holed it. What does he mean? Not, surely, that if every atom in the universe were exactly as it was, the ball would have gone in. That is not something anyone could ever test, and it is not what golfers care about. He means that he has the skill, and that in situations very like this one, he usually makes the shot. The miss was a failure of execution, not of ability.

The useful sense of "could have done otherwise" looks at a range of nearby situations, not one frozen instant. Knowing whether an agent would have done differently if things had been slightly different is the information we need to predict, train, praise and blame. Whether they would have done differently if nothing at all had been different is idle. Determinism removes only that idle possibility.

6. Here I stand

In 1984 Dennett published a paper with a provocative title: "I Could Not Have Done Otherwise – So What?" His example was Martin Luther, who is said to have told the Diet of Worms in 1521 that he could do no other than stand by his writings. Luther was not pleading an excuse. He was saying that his conscience had so formed him that recanting was not a live option. We do not think less of him for it. We think he was taking full responsibility, and that the impossibility of doing otherwise was the mark of his character, not a limit on his freedom.

Many of our most admirable acts are like this. A decent person cannot bring themselves to torture a child for money; the option is closed to them. If being free required that every act could have gone the other way, the people we admire most would be the least free. That is a sign the requirement is wrong. Whether someone could have done otherwise is a clue we use in judging responsibility, Dennett argued, not the thing itself. Harry Frankfurt had reached a similar conclusion by a different road; see his page.

7. Giving libertarians what they want

Libertarians about free will, in the philosophers' sense, hold that free will is real and requires that determinism be false. In a 1978 essay, "On Giving Libertarians What They Say They Want", Dennett tried to build them the best version of their view, to see whether it gained them anything.

Suppose, he said, that a decision has two stages. In the first, some process, which might be genuinely random, throws up considerations: options, memories, arguments, ideas. In the second, a thoroughly reliable process evaluates what has come up and selects among it. The randomness would give the libertarian a future that is truly open. The reliable selection keeps the result under the agent's rational control. He named this the Valerian model, after the French poet Paul Valéry, who remarked that it takes two to invent anything: one to make up combinations, the other to choose. William James had sketched something similar a century earlier.

Then came the twist. Swap the truly random generator for a pseudo-random one, the deterministic kind that computers use. Would anything we care about be lost? The agent would deliberate just as well, be just as responsive to reasons, and deserve exactly the same praise or blame. From the inside and the outside, nothing would change. So the indeterminism, which libertarians thought was essential, turns out to be doing no work. If it does no work, we need not insist on having it. Robert Kane later developed a more sophisticated libertarian view partly in response; see Kane.

8. Making yourself small

In 1983 the neuroscientist Benjamin Libet reported that a burst of brain activity, the readiness potential, began about half a second before a simple voluntary movement, and several hundred milliseconds before people said they had decided to move. Many concluded that the brain decides before "you" do, and that conscious will is a latecomer. You can try a version of the experiment in the lab.

Dennett's response was that the conclusion depends on a picture of the self he had spent decades attacking: a tiny inner observer who sits at a single point in the brain, receives reports and issues commands. If that is who you are, then of course anything happening elsewhere in the brain is happening to you, not by you. In a phrase from Elbow Room, "if you make yourself really small, you can externalize virtually everything".

But you are not a point. You are the whole organised system, including the processes that started the readiness potential. There is no reason why your decisions should be made in a single place at a single instant, or why they should be done by the part of you that later reports on them. Asking exactly when, to the millisecond, a decision became conscious, Dennett argued, assumes there is a finish line in the brain where consciousness happens. On his view of consciousness, there is no such line. See also Libet and Mele.

9. Freedom evolves

The title of Dennett's 2003 book states its thesis. Freedom is not a timeless metaphysical fact that we either have or lack. It is a biological and cultural achievement that came into the world gradually, and is still growing. Bacteria have a kind of freedom: they swim towards food and away from poison. Animals with eyes and brains can avoid far more. Animals that learn can avoid dangers they have never met before. And one species acquired language, which let it share reasons, ask for them, and argue about them.

Language changes everything, in Dennett's account, because it lets us represent our own reasons to ourselves and to others. A creature that can be asked "why did you do that?", and can answer, can also be held to its answer. It can make promises, take on commitments and become the sort of thing that can be persuaded. That is where moral responsibility comes in. Free will, on this view, is a social and cultural achievement built on biological foundations, and it comes in degrees. A small child has less than an adult, and education, health and good institutions can give people more.

10. The moral agents' club

Why hold people responsible at all? Dennett's answer looks forward and back at once. Being treated as a responsible agent is like being a member of a club. Membership brings great benefits: you are trusted, you can make binding agreements, you are left free to run your own life. In return you accept the rules, including that you will be blamed and sometimes punished if you break them. Most of us, he pointed out, want to be members, and would be insulted to be told we are not responsible for what we do.

To borrow a sporting comparison, a footballer who commits a foul in the penalty area concedes a penalty. Nobody thinks the rule requires the foul to have been uncaused, or the player to have been able to do otherwise in some deep metaphysical sense. The rule is justified because the game is better with it, and the player knew the rules. Punishment, Dennett argued, can be deserved in the same way: not because it balances some cosmic ledger, but because it is part of a fair system that free people have good reason to want.

This is the heart of his disagreement with Gregg Caruso in Just Deserts. Caruso denies that anyone ever deserves punishment in the "basic" sense; Dennett says that the only sense of desert worth wanting is the one his system supplies. We set the two side by side on Do we deserve blame?

11. Against the deniers

Dennett spent his last years arguing with a new wave of writers who declared free will an illusion: the neuroscientist Sam Harris, the philosopher Gregg Caruso, and, in the year before Dennett's death, the biologist Robert Sapolsky, whose Determined (2023) argued that there is no free will at all. Dennett respected their humane motives. Many of them want a kinder justice system, less hatred and less blame for people whose lives went wrong. He shared those aims.

What he objected to was the headline. Telling people that free will is an illusion, he argued, is like telling them that colour is an illusion because physics has no colours in it. It is true of something, a philosopher's fantasy of freedom, and false of the thing people actually mean. And it could do harm: experiments have suggested that people told they lack free will can behave a little worse, though how robust those findings are is disputed. If the deniers mean only that nobody is an uncaused cause, Dennett said, they should say so, and stop implying that our deliberations, plans and promises make no difference. See Harris, Caruso and Sapolsky.

12. Objections

Compatibilism is the oldest answer to the free will problem, and it has always had critics. Kant called it a wretched subterfuge, and said that a freedom which is merely the working of internal causes would be the freedom of a turnspit, which, once wound up, carries out its movements by itself. Dennett's reply was that a turnspit cannot deliberate, learn or respond to reasons, and we can. The difference between a turnspit and a person is not a gap in causation. It is a vast difference in design.

The sharpest modern objections come from the incompatibilists. Peter van Inwagen's consequence argument says that if determinism is true, our acts are consequences of the laws of nature and events in the remote past, neither of which is up to us. Galen Strawson argues that to be truly responsible for what you do you would have to be responsible for how you are, which would require creating yourself. Derk Pereboom describes an agent manipulated by neuroscientists, then shows that the manipulation can be made more and more like ordinary causal history, with no point at which responsibility returns.

Dennett's general reply is that each argument assumes the very notion of ultimate, from-nowhere responsibility that he rejects. Nobody makes themselves from scratch. But we do make ourselves, gradually, from what we were given, by the choices we take and the reasons we accept. Whether that is enough is the deepest question on this site, and the thinker pages set out both sides.

In brief

• Ask what free will is for. Every variety worth wanting is compatible with determinism.
• Causes are not controllers. The bogeymen of determinism are images of other agents, not of laws.
• Determinism is not inevitability: in a deterministic world, many things happen only because of what you do.
• "Could have done otherwise" matters in nearby situations, not in an identical replay of the universe.
• Indeterminism adds nothing to responsibility; a pseudo-random generator would serve as well.
• Freedom evolved, comes in degrees, and grows with language, reasons and culture.
• Responsibility is membership of a club most of us very much want to belong to.