Up to Us

Everything you choose is done by atoms that follow the laws of physics. Can the choice still be yours? What physicists, mathematicians, neuroscientists, biologists and philosophers really disagree about.

Raise your hand, or don’t. Either way the choice felt like yours: you considered it, and something you would call you settled it. But the hand is made of cells, the cells of molecules, and the molecules of particles that, as far as anyone has been able to measure, do exactly what the equations of physics say they will. Somewhere between the equations and the hand there seems to be no gap through which you could reach in. Einstein, compressing an argument of Schopenhauer’s, put the puzzle in one line: a person can do what he wants, but cannot want what he wants.[1]

The questions this essay tries to answer are the ones people ask first. Do the laws of physics really fix everything in advance, and are they deterministic at all? If consciousness emerges from the brain the way temperature emerges from moving molecules, could it have powers its parts lack, free will among them? How could a decision that seems to come from outside the physical world move a single atom inside it? And the order of time itself, the fact that entropy grows and the past differs from the future, is a property that emerges from simple interactions: does that show something about choices? Finally, are these real problems, or signs that we do not yet understand something, or that we are asking the wrong question? Physicists, mathematicians, neuroscientists, biologists and philosophers have all answered, often against each other. What follows sets out their arguments as fairly as I can, says where each side is strongest, and separates what is known from what is not.

59%of professional philosophers think free will is compatible with determinism

0.55 show early a brain signal rises before a spontaneous movement, in Libet’s experiment

60%how often brain scans predicted a left-or-right choice seconds ahead, where chance is 50%

10⁻¹³ sthe longest a quantum superposition would survive in a warm brain, by one estimate

What is being asked

Three things people mean by free will

Much of the confusion comes from the fact that “free will” names several different things, and people who seem to disagree are often defending different ones.

The first is being able to do otherwise: at the moment of choice, more than one future was really open, and which one happened depended on you. The second is being the source: the choice came from you in a deep sense, not from causes that ran through you from before you were born. The third is control: you acted for your own reasons, could have responded to other reasons, and were not coerced, compelled or deceived. The first two are what most people mean when they say that a world fixed by physics leaves no room for freedom. The third is what courts, teachers and parents mostly check for when they decide whether someone is to blame.

The philosophical positions follow from which of these is thought necessary. Incompatibilists say that free will, in the sense that matters, cannot exist if the past and the laws fix every future. Some of them, the libertarians (no relation to the political word), conclude that the world is not fixed in that way and that we are free; others, the hard determinists and hard incompatibilists, conclude that we are not. Compatibilists say that free will is the third thing, control, and that it is perfectly at home in a world governed by laws, deterministic or not. Behind all of them stands the reason anyone cares: whether people can deserve praise and blame, and punishment, for what they do.

History

An older question

The problem is much older than physics, and for more than a thousand years it was argued about God rather than about atoms. If God knows today what you will do tomorrow, can you do otherwise? Augustine wrote his dialogue On Free Choice of the Will in the late fourth century to reconcile divine foreknowledge with human responsibility. Boethius, awaiting execution in the sixth, answered that God does not foresee but simply sees, because eternity is “the possession of endless life whole and perfect at a single moment”.[2] In 1524 Erasmus defended free choice against Luther, who replied the next year in The Bondage of the Will that the human will “is, as it were, a beast between the two. If God sit thereon, it wills and goes where God will… If Satan sit thereon, it wills and goes as Satan will.”[3] One of the most ingenious reconciliations was worked out in Portugal: the Spanish Jesuit Luis de Molina, who taught at Coimbra and Évora, published his Concordia in Lisbon in 1588, arguing that God knows what each person would freely do in every possible circumstance.[4] Laplace’s intelligence, in the next section, is the same worry with God taken out; the shape of the argument hardly changed.

Physics

Does physics decide?

The classic statement of determinism is Laplace’s, from 1814: an intelligence that knew, at one moment, all the forces of nature and the positions of everything in it, and could analyse the data, would find that “nothing would be uncertain and the future, as the past, would be present to its eyes.”[5] Newton’s mechanics seemed to make this literally true. Two things have complicated it since, and a third has not.

The first is chaos. In 1963 the meteorologist Edward Lorenz found that a simple set of deterministic equations for convection was so sensitive to its starting values that tiny differences, smaller than any measurement could detect, grew until the predictions were worthless.[6] A chaotic system is completely determined and still unpredictable beyond a horizon. That separates two ideas that are often run together: determined is a fact about the world, predictable is a fact about what someone inside it can know. Even Newton’s equations are not quite as tidy as Laplace thought. The philosopher of physics John Norton showed in 2008 that a ball resting on the top of a specially shaped dome can, consistently with Newton’s laws, stay there forever or start rolling at any moment in any direction, with nothing to decide which: determinism can fail even in classical physics, if only in idealised cases.[7]

The second is quantum mechanics. Its equation for how a system’s state changes, Schrödinger’s, is deterministic. But what it predicts for measurements are probabilities, and in the textbook version an individual result, which slit the electron goes through, when the atom decays, is not fixed by anything at all. Whether that randomness is part of nature depends on an interpretation that physicists still dispute. In the collapse theories the randomness is real. In Everett’s many-worlds view the whole evolution is deterministic and every outcome happens, in branches that cannot see each other. In Bohm’s pilot-wave theory particles have definite positions guided by the wave, and the whole thing is deterministic again, though unpredictable in practice. All three reproduce the same experimental results so far. Physics today does not tell us whether the world is deterministic.

What experiments have settled is narrower and stranger. Since 2015, tests free of the main loopholes have confirmed that measurements on entangled particles[a] are correlated in ways no theory of local hidden causes can reproduce, a result recognised by the Nobel Prize in Physics in 2022.[8] Those tests assume that the experimenters’ choice of what to measure is independent of the particles being measured. The mathematicians John Conway and Simon Kochen turned the assumption around in their “free will theorem”: if the experimenters’ choices are not determined by the past, then neither are the particles’ responses, and so, in their words, “if indeed we humans have free will, then elementary particles already have their own small share of this valuable commodity”.[9] Their “free will” means only “not a function of the past”, a much thinner thing than the philosophers’. A few physicists, among them Sabine Hossenfelder and Tim Palmer, take the opposite branch, superdeterminism: the choices and the particles are correlated from the start, so there is no independence to assume.[10] Most physicists reject it as making science impossible to test; a “cosmic Bell test” in 2017 chose the measurements using the colour of light from stars about 600 light years away, so that any such correlation would have had to be set up at least six centuries ago; a later one, using quasars, pushed the date back by billions of years.[11]

The luck problem

Chance is not choice

It would seem that quantum randomness rescues free will: if the future is not fixed, there is room for us. Most philosophers on all sides think it does not, at least not by itself. If, at the moment you decide, nothing about you settles whether you raise your hand, then whether you raise it is a matter of chance, and chance is no more yours than a law. David Hume made the point in 1748: “liberty, when opposed to necessity, not to constraint, is the same thing with chance”.[12] This is the luck objection, and it is why the question “is physics deterministic?” matters less to the debate than it seems to.

Libertarians have answers. Robert Kane argues that the indeterminism matters only in rare “self-forming” decisions, when a person is torn between, say, duty and ambition, and is really trying to do both; the chaotic brain amplifies microscopic indeterminacy, and whichever effort wins, the result was something the person was trying to do, and so is hers.[13] Critics reply that the outcome is still, in the end, a matter of luck. Others, the agent-causal libertarians, say the agent herself, not any event in her brain, is the cause of the choice. That avoids luck at the price of a kind of causation physics does not describe, which is the problem of the next section.

Even if randomness were wanted, it is not clear the brain could use quantum randomness. Roger Penrose and Stuart Hameroff proposed that consciousness involves quantum states in microtubules, protein tubes inside neurons.[14] In 2000 the physicist Max Tegmark estimated how long such superpositions would survive in the warm, wet brain before interactions with their surroundings destroyed them, a process called decoherence:[b] between about 10⁻²⁰ and 10⁻¹³ of a second, against at least a thousandth of a second for anything a neuron does.[15] Hameroff and colleagues answered with longer estimates, and the dispute is not closed, but most neuroscientists treat the brain as a classical machine with ordinary noise: ions crossing channels, molecules released or not at synapses. That noise is real and measurable, whatever its ultimate source.

Fig. 1 — Timescales, seconds, on a logarithmic scale where each gridline is ten thousand times the one before. Decoherence estimates are Tegmark’s for the brain (2000); the dashed marker is the longer estimate of Hagan, Hameroff and Tuszyński (2002) for microtubules. Neural timescales: a nerve impulse lasts about a millisecond; perception and decisions take from tens of milliseconds to about a second.

The computer scientist Scott Aaronson has proposed a subtler role for quantum physics. If some of the brain’s decisions depend sensitively on microscopic details whose quantum states nobody, even in principle, could have measured, which he calls “freebits”, then no machine could ever predict a particular person perfectly, not because the person is random but because the information needed does not exist anywhere to be copied. What he defends is “a certain kind of in-principle physical unpredictability that goes beyond probabilistic unpredictability”; he presents the idea as speculative, and as a question about prediction rather than a proof of freedom.[16]

Mind and body

A push from outside

The picture of a will that reaches into the world from outside it is old, and its difficulty was stated almost as soon as the picture was. In May 1643 Princess Elisabeth of Bohemia wrote to Descartes, who held that the mind is an unextended thinking thing, to ask: “Given that the soul of a human being is only a thinking substance, how can it affect the bodily spirits, in order to bring about voluntary actions?” Every motion seemed to need a push, and a push needs contact and extension.[17] Descartes’s replies did not satisfy her, and three and a half centuries have not produced a better one for that picture. In 1994 the Portuguese neuroscientist António Damásio called a book Descartes’ Error for a related mistake: the idea that reason could work at all apart from the body and its feelings.[18]

Physics has since made the difficulty sharper. Energy is conserved; every physical event that has a cause seems to have a sufficient physical cause; and the forces that act at the energies of a brain are, physicists such as Sean Carroll argue, fully known; the title of a 2010 essay of his says it plainly: “The laws underlying the physics of everyday life are completely understood”.[19] A non-physical will that moved neurons would show up as atoms behaving in ways those laws forbid, and no experiment has seen that. The philosopher Jaegwon Kim turned this into a dilemma: if each brain event already has a complete physical cause, then a mental cause distinct from it either does nothing or doubles up on work already done.[20] Either the mind is physical, or it seems to have nothing left to do.

This is the point where the question “how can something physical influence the physical world from outside?” gets its most common answer: it doesn’t, and it doesn’t need to. Very few researchers now defend a will outside physics. On most current views, including most libertarian ones, a decision is a process in the brain, and it moves the hand the way a hurricane uproots a tree: not by pushing on the molecules from outside, but by being a particular pattern of them. The remaining argument is whether the pattern, the higher level, is a real cause at all, or a convenient redescription of what the particles do anyway.

Emergence

More is different

The physicist Philip Anderson gave the higher levels their charter in 1972: “The ability to reduce everything to simple fundamental laws does not imply the ability to start from those laws and reconstruct the universe.” At each level of complexity, he argued, “entirely new properties appear”, and studying them is as fundamental as studying the parts.[21] Superconductivity, the rigidity of a crystal and the wetness of water are not in any single atom. Nor is the behaviour of a worm in its wiring diagram, as this blog argued in The Complete Parts List.[c]

Philosophers distinguish two kinds of emergence, and the distinction decides almost everything here.[22] Weak emergence is novelty that is surprising and often impossible to calculate in practice, but that follows in principle from the parts and their laws: a traffic jam, a hurricane, the temperature of a gas. Strong emergence would be a higher-level property that does not follow from the lower level even in principle, and that can act back on it with powers of its own. If consciousness or will were strongly emergent, the “push from outside” would come from inside after all, from a new level of nature. The cost is that the particles in a brain would then, at some point, behave differently from the same particles anywhere else. That is a testable claim, and so far there is no sign of it. David Chalmers, who takes consciousness more seriously than almost anyone, thinks consciousness may be the one case of strong emergence, and accepts that this would be a radical thing.

Weak emergence still leaves room for higher-level causes, its defenders argue. The cosmologist George Ellis points out that the program running on a computer decides which way its electrons flow, without violating any law of electromagnetism: the higher level sets the conditions under which the lower laws act.[23] Erik Hoel, Larissa Albantakis and Giulio Tononi measured how much a description at each level constrains what comes next, and found systems in which the coarse description is the more informative one: “macro can beat micro”.[24] The philosopher Christian List goes further: free will needs intentional agency, alternative possibilities and causal control, and all three are real at the level of psychology, as an economy is real at the level of markets, even if at the level of physics there might be no alternatives at all.[25] Critics answer that all of these concern our descriptions, not the world: a better description is not an extra cause.

So the honest reply to the question whether consciousness, as an emergent property, could have free will has two parts. If it is weakly emergent, it can be a real cause in the same sense that a program or a hurricane is, which is enough for the compatibilist kind of freedom, but it is not exempt from the laws of its parts. If it is strongly emergent, it could be exempt, but that would be new physics, and nobody has found any.

The arrow of time

A direction that is not in the rules

The best-understood example of something that emerges from simple laws, and is not in them, is the direction of time. The fundamental laws, Newton’s and quantum mechanics’ alike, run equally well backwards (a tiny asymmetry in some decays of particles plays no role here). Yet eggs break and do not unbreak, heat flows from hot to cold, and we remember yesterday and not tomorrow. The second law of thermodynamics, which says that entropy, a measure of disorder, tends to increase, gives time its arrow, and it emerges from the interactions of particles that have no arrow at all.

It does not emerge from the interactions alone, though, and that is the lesson worth taking from it. In 1876 Josef Loschmidt objected to Boltzmann’s explanation of the second law that if you took any system whose entropy had grown and reversed every velocity, the same laws would carry it back, entropy falling, to where it began.[26] The objection is correct. The way out is that disordered states vastly outnumber ordered ones, so almost any starting point leads to more disorder; and that the universe began in a state of extraordinarily low entropy, so there has been a long way down ever since. The philosopher David Albert calls that assumption the Past Hypothesis.[27] The arrow is in the beginning, not in the rules.

A small computation shows it. The figure follows 4,000 particles in a box, moving without friction and bouncing off the walls, with whole-number positions and speeds so that every step is exact and exactly reversible. They start crowded into the left tenth of the box. The entropy of their spread rises within a few hundred steps to almost its maximum, and stays there. Then every velocity is reversed. Nothing in the rules has changed, and the gas runs back, entropy falling, until at step 800 every particle is exactly where it started.

Fig. 2 — A deterministic, reversible toy gas: 4,000 particles with whole-number speeds from 1 to 12 in a box of 2,000 positions, starting in the left tenth. Entropy is that of the particles’ distribution over 20 equal cells, as a share of its maximum. At step 400 every velocity is reversed. Computed by scripts/free_will.py.

What does this say about choices? Three things, and they point in different directions. First, emergent quantities can be completely real, measurable and law-like while being fixed by the dynamics of the parts: entropy is not an illusion, and it is not exempt. Compatibilists read this as the model for the will: real at its level, nothing spooky underneath. Second, the very asymmetry of choosing may be thermodynamic. We deliberate about the future and never about the past; we have records of the past and none of the future. Philosophers of physics such as Albert and Barry Loewer, and physicists such as Carroll and Carlo Rovelli, argue that this asymmetry of knowledge and influence is itself a consequence of the entropy gradient:[28] the openness of the future that we feel when we choose would be, in part, the same arrow that breaks the egg. Rovelli’s view of time was the subject of World without time on this blog in 2018. Third, the analogy has limits. Entropy is defined relative to a coarse description, the 20 cells of the figure, chosen by us. Incompatibilists reply that showing a higher-level quantity is useful does not show that a person could have done otherwise; the gas, after all, had no choice about reversing.

Mathematics

Determined, and still unknown to itself

Mathematics adds a result that sounds like a technicality and may be more than that: some processes cannot be predicted faster than they run. Stephen Wolfram called this computational irreducibility: for many simple programs there is no shortcut, and the only way to find out what the system will do is to let it do it. “And it is this, I believe,” he wrote, “that is the ultimate origin of the apparent freedom of human will.”[29] The physicist Seth Lloyd applied the idea to decision-makers. Using the same reasoning that shows no program can decide in general whether other programs will halt, he argued that a system capable of deciding cannot, in general, predict its own decisions before making them; to know what it will choose, it must go through the choosing. “Even when our decisions arise from a completely deterministic decision-making process,” he wrote, “the outcomes of that process can be intrinsically unpredictable, even to – especially to – ourselves.”[30] Earlier, Donald MacKay had argued that even a perfect prediction of your choice is not one you could correctly believe, because believing it would change the brain it predicts.

Supporters see this as the core of what freedom feels like from the inside: the future is open to the chooser not because physics is undecided but because there is no way to read the answer without doing the work. Critics see a confusion of ignorance with freedom: the weather is computationally irreducible too, and nobody thinks it chooses.

Neuroscience

Half a second early

The experiment that brought neuroscience into the debate was Benjamin Libet’s, in 1983. Volunteers flexed a wrist whenever they felt like it, and reported, by watching a fast clock, when they first felt the urge. Electrodes on the scalp recorded a slow build-up of activity, the readiness potential, which began on average about 350 milliseconds before the moment people reported first feeling the urge, and about 800 milliseconds before it when they had planned somewhat ahead.[31] In Libet’s later summary, the build-up started about 550 milliseconds before a spontaneous movement, and the urge was felt about 200 milliseconds before it. The brain, it seemed, had started before its owner knew. Libet himself did not conclude that free will was an illusion. People could, he found, “veto” a movement in the last 100 to 200 milliseconds, and he proposed that the role of conscious will “would be not to initiate a specific voluntary act but rather to select and control volitional outcome”; that evidence, though, came from acts planned for a set time and then cancelled, not from spontaneous ones.[32]

Later work went further. In 2008 John-Dylan Haynes’s group used brain scans to predict whether people would press a button with the left or the right hand from activity in the frontal and parietal cortex seven seconds before they were aware of choosing, and, allowing for the slowness of the blood-flow signal the scanner measures, perhaps up to ten. The prediction was right only about 60 per cent of the time, where guessing gives 50, as read from the paper’s figures.[33] In 2011 recordings from 256 neurons in the brains of epilepsy patients predicted a movement 700 milliseconds before the reported decision, with an accuracy above 80 per cent.[34]

Fig. 3 — What the brain shows before a simple movement, seconds before it, on a logarithmic scale. The shaded bar is the range over which Soon and colleagues decoded the choice from brain scans. Sources: Libet et al. (1983), Fried et al. (2011), Soon et al. (2008).

Each of these findings has been reinterpreted. In 2012 Aaron Schurger, Jacobo Sitt and Stanislas Dehaene showed that the readiness potential may largely be what you get when spontaneous fluctuations in brain activity are averaged backwards from the moments when they happened to cross a threshold. In their words, the moment of movement “is largely determined by spontaneous subthreshold fluctuations in neuronal activity”, and averaging makes these look like “a gradual exponential-looking increase”. The slow ramp would then reflect not a decision already made, but the noise that, on that occasion, tipped the person over.[35] In 2016 Haynes’s own group let a computer read the readiness potential in real time and signal people to stop: they could cancel a movement until about 200 milliseconds before the muscles moved, well after the build-up had started, though in that experiment the go signal came from a light.[36] And in 2019 Uri Maoz and colleagues compared arbitrary choices with decisions that mattered, which of two charities would receive a real $1,000 donation; they found the expected readiness potentials before the arbitrary choices, but they were “strikingly absent for deliberate ones”.[37] The experiments measure the timing of whims, and whims may be the one kind of decision where nothing much is decided.

Damásio’s own work shows the unconscious from the other side: not a brain that decides before you, but one that helps you decide well. With Antoine Bechara and colleagues he gave people four decks of cards to draw from, two that paid well at first but lost money over time and two that paid less and won. Healthy players drifted towards the good decks before they could say why, and their skin began to sweat slightly, a sign of alarm, just before they reached for a bad one. “Normals began to choose advantageously before they realized which strategy worked best, whereas prefrontal patients continued to choose disadvantageously even after they knew the correct strategy”; the patients, with damage to the front of the brain above the eyes, never produced the warning signal.[38] Damásio’s patient Elliot, whose frontal lobes had been compressed by a tumour, kept an IQ “in the superior range” but could spend a whole afternoon deciding how to sort his papers; another patient with similar damage weighed two possible dates for his next appointment, listing reasons for and against each, “for the better part of a half-hour”.[18] Damásio’s explanation, the somatic marker hypothesis, is that signals from the body mark options as good or bad, “an automated alarm signal which says: Beware of danger ahead”, and that without them reasoning goes on without end. Tiago Maia and James McClelland replied that, asked more carefully, players knew more than had been assumed, so that “there is no need to appeal to nonconscious somatic markers”; the two sides answered each other in print, and the dispute is not closed.[39]

Damásio also draws the line that Maoz’s experiment drew. “Most of the notions that we associate with deliberation and decisions that are important for one’s life are not taken the same way that we move this finger or we pick up the glass,” he said in 2010. “Sometimes you deliberate for minutes or hours or weeks or months and you do it not in the moment of execution of the action. You do it offline.”[40] Much of that deliberation happens below awareness, like the unconscious work that comes before a mathematician’s sudden insight, discussed on this blog in The Bus at Coutances: “Nobody has watched the unconscious work. Its output has been counted.” Whether that makes the unconscious a part of you or one more cause you did not choose is, again, where the two sides divide.

Other findings are harder to set aside, because they concern the reasons people give rather than the timing. In patients whose brain hemispheres had been surgically separated, Michael Gazzaniga found that the speaking left hemisphere would readily invent explanations for actions the right hemisphere had performed for reasons it could not know.[41] The psychologist Daniel Wegner argued that the feeling of willing an act is the mind’s inference about its own behaviour, not a direct view of its cause: experiences of conscious will “arise from processes whereby the mind interprets itself – not from processes whereby mind creates action”.[42] And in 2003 neurologists reported a man who developed paedophilic urges, which vanished when a tumour in his right orbitofrontal cortex was removed; according to later accounts, they returned when it grew back, and went again after a second operation.[43] No one proposes that he chose the tumour. Whether such cases show that nobody is responsible, or only that some people sometimes are not, is exactly where the neuroscientists divide.

Biology

Agents before minds

The sharpest current argument among scientists is between two biologists of the brain. Robert Sapolsky, a professor of biology and neurology at Stanford, is a neuroendocrinologist and a primatologist: for some two decades he returned each year to a troop of wild baboons in Kenya to study how rank and stress shape their bodies and their behaviour, a life he described in A Primate’s Memoir.[44] In Determined (2023) he argues that every behaviour is the product of what happened a second before in the neurons, hours before in the hormones, years before in childhood and the womb, and millennia before in evolution and culture, and that at no point in that chain is there room for a freedom that is not itself caused: “We are nothing more or less than the cumulative biological and environmental luck, over which we had no control, that has brought us to any moment.” He adds that it will suffice if readers conclude that there is “so much less free will than you thought”.[45]

His most distinctive argument is aimed at the place where most people keep their freedom when they grant everything else: effort. You may not choose your desires, the thought goes, but resisting them is up to you. Sapolsky’s reply is that the capacity to resist is itself a working of the brain, above all of the prefrontal cortex, the region that makes us do the harder thing when it is the right thing, and that how well it works depends on genes, on hormones before birth, on childhood, and on how tired, hungry or stressed a person is at that moment, none of which the person chose. Grit, on this view, is luck too: nobody can will himself to have more willpower.[45] From there he draws the conclusion that blame and retribution make no sense, though people who are dangerous can still be restrained to protect others.

The neurogeneticist Kevin Mitchell, in Free Agents (2023), argues the opposite from the same biology. Living things, from bacteria up, are not billiard balls pushed by causes: they hold themselves together against the physics that would dissolve them, act on information about the world, and do so for reasons, staying alive, that belong to them. Over evolution these control systems grew layers of prediction and deliberation, and in humans they can reflect on their own reasons. Free will, on this account, is a real capacity of real organisms, which evolved gradually and comes in degrees, and the noise in nervous systems is a resource it uses to avoid being predictable, not a threat to it.[46] The neurobiologist Björn Brembs has made a similar case from flies: tethered fruit flies in a featureless arena still vary their turns in structured, non-random ways, which he sees as the evolutionary root of spontaneous choice.[47]

Critics of Sapolsky, compatibilists above all, reply that he shows that every choice has causes, which they never denied, but not that people lack control, which is a different claim. Even a reviewer who largely agreed with him wrote that there still needs to be some way to tell a brain that can moderate its impulses from one that cannot, and concluded: “perhaps the concept of compatibilism isn’t entirely redundant after all.”[48] On Sapolsky’s account, such differences are real but are themselves luck: they can justify treating people differently, but not praising or blaming them.

The two biologists agree on the facts more than the titles suggest. Neither thinks anything violates physics. What they disagree about is whether the long causal history behind a choice takes the choice away from the person, or is simply what being a person consists of. Daniel Dennett, who died in 2024, spent fifty years arguing for the second view: evolution built creatures that can anticipate, weigh and revise, and that is “the kind of free will worth wanting”.[49]

Psychiatry and the law

Degrees of control

Psychiatry meets every day the cases that philosophers imagine. In compulsions and addictions people act against what they themselves want to want; Harry Frankfurt’s “unwilling addict”, who takes the drug while wishing that desire did not move him, is his example of a person whose will is not free even though he does what he wants.[50] How much control addiction takes away is disputed. Nora Volkow, director of the US National Institute on Drug Abuse, and her colleagues conclude that “neuroscience continues to support the brain disease model of addiction”.[51] The psychologist Gene Heyman points out that in every national survey in the United States “most of those who met the criteria for dependence on an illicit drug no longer did so by age 30”, usually without professional help, and takes it as evidence that addiction remains partly a matter of choice; dependence on alcohol, he notes, lasts much longer.[52] The neuroscientist Marc Lewis argues that addiction is a form of learning rather than a disease.[53] What the three agree on is that control comes in degrees, and can be lost and regained.

The law has had to draw the line somewhere. The English rule of 1843 excused a defendant only if he was “labouring under such a defect of reason, from disease of the mind, as not to know the nature and quality of the act he was doing; or, if he did know it, that he did not know he was doing what was wrong”.[54] The Portuguese Penal Code asks about control as well as knowledge: a person is not criminally responsible who, because of a mental anomaly, is unable at the time of the act “de avaliar a ilicitude deste ou de se determinar de acordo com essa avaliação”, to judge that it is wrong or to act according to that judgement, and it provides for capacity that is markedly diminished.[55] Since July 2001 Portugal has also treated the possession of drugs for personal use as an administrative offence, handled by commissions for the dissuasion of drug addiction rather than by the criminal courts,[56] a policy that treats people with an addiction as having impaired control without treating them as having none. Neither rule asks whether the universe is deterministic.

Consciousness

Is consciousness the problem?

It is natural to think that free will and consciousness are the same mystery. They overlap, but they are separable, and most of the free-will debate does not depend on solving the hard one. David Chalmers called it the hard problem: why any physical process is accompanied by experience at all, why there is something it is like to be you.[57] Nobody has an accepted answer. The leading scientific theories, such as Dehaene’s global workspace, explain what consciousness does, making information available to the whole brain at once for flexible use, rather than why it feels like anything.[58] On views like Anil Seth’s, discussed on this blog in Your brain hallucinates your conscious reality, experience is the brain’s running prediction of the world and of the body, and the feeling of being in control may be one such prediction.

Where consciousness does matter is in two places. One is the neuroscience above: if conscious intentions always came too late to cause anything, the self that feels it chooses would be a spectator. Reviewing the evidence in 2011, Roy Baumeister and colleagues concluded that conscious thought does shape behaviour, though mostly indirectly, by simulating options, planning and integrating information over time, rather than by triggering each movement.[59] The other is the deeper worry that experience might be epiphenomenal: produced by the brain but causing nothing, like the whistle of a steam engine. If consciousness is a weakly emergent feature of brain processes, it causes what those processes cause; if it is something extra, Kim’s dilemma returns in full.

So, to the question whether consciousness is an emergent property that could somehow have free will: on most current views it is the wrong unit. It is not consciousness that would be free or unfree, but the person, the whole organism acting for reasons; consciousness is one of the ways a person keeps track of those reasons. Solving the hard problem would tell us what experience is. It would not, by itself, tell us whether the person who has it could have done otherwise.

Philosophy

Where the philosophers stand

The most influential argument for incompatibilism is Peter van Inwagen’s Consequence Argument: “If determinism is true, then our acts are the consequence of laws of nature and events in the remote past. But it’s not up to us what went on before we were born, and neither is it up to us what the laws of nature are. Therefore, the consequences of these things (including our present acts) are not up to us.”[60] Galen Strawson’s Basic Argument goes further and claims that free will is impossible whether or not determinism is true: you do what you do because of the way you are; to be truly responsible for that, you would have to have made yourself that way, which requires an earlier self that made it, and so on without end.[61]

Compatibilists have two main replies. Harry Frankfurt imagined a person who decides to do something on his own, while a neuroscientist stands by, ready to force the same decision if he wavers; he never wavers, so the device is never used. He could not have done otherwise, and yet he seems fully responsible, because the act came from him.[50] If that is right, the ability to do otherwise is not what responsibility needs. And P. F. Strawson, Galen’s father, argued in 1962 that resentment, gratitude and indignation are not theories that physics could refute, but part of what it is to live with other people.[62] Hard incompatibilists such as Derk Pereboom answer that we can keep almost everything worth keeping, love, gratitude, the protection of society, while giving up only the idea that anyone deserves to suffer for what they did.[63]

In the largest survey of professional philosophers, run in 2020, the compatibilists are a clear majority.

Fig. 4 — “Free will: compatibilism, libertarianism or no free will?”, answers of 1,758 philosophers in faculty posts at leading departments, per cent who accept or lean towards each view; respondents could choose more than one, and the remaining tenth or so gave other answers (agnostic, the question unclear, none of these). Source: PhilPapers Survey 2020 (Bourget and Chalmers).

A majority is not proof, and philosophers of religion, who are more often libertarians, and philosophers of mind and cognitive scientists, among whom the “no free will” view is commoner, pull in different directions.[64] People who are not philosophers are split in revealing ways. In one study, told about a fully deterministic universe in the abstract, 86 per cent said that people in it cannot be fully morally responsible. Told about a particular man in that universe who kills his wife and three children to be with his secretary, 72 per cent said that he is.[65] The pattern is not only American: in a study in the United States, Hong Kong, India and Colombia, “in all four cultural groups, the majority of participants said that (a) our universe is indeterministic and (b) moral responsibility is not compatible with determinism”.[66] Our intuitions seem to be incompatibilist about the cosmos and compatibilist about our neighbours.

Consequences

Does believing it matter?

In 2008 a widely cited experiment found that students who read a passage arguing that free will is an illusion cheated more on a test afterwards.[67] The result has not held up: the large replication project of 2015 found no effect on cheating, nor did a larger replication in 2020, and a review of 145 experiments in 2022 found that such passages do weaken belief in free will a little, but “could not find evidence for downstream consequences” for behaviour.[68] The fear that telling people the truth about determinism would make them behave badly has, so far, little evidence behind it.

The law is a more serious case. Criminal law in most countries asks about control, whether the accused understood what he was doing and could have conformed to the law, which is the compatibilist question; it does not ask about determinism. The psychologists Joshua Greene and Jonathan Cohen argued in 2004 that this is why neuroscience changes nothing for the law in principle, and everything in practice: as people see more of the mechanism behind crimes, their intuitions about retribution will weaken, and punishment will be justified more by its effects than by desert.[69] Whether that would make justice kinder or more invasive, since prevention can justify more intervention than desert can, is one of the questions Dennett and the philosopher Gregg Caruso debated in Just Deserts (2021).[70]

The open questions

Ignorance, confusion, or both?

Are these problems only a lack of understanding? Some of them are, and could be settled by discovery. Whether the fundamental laws are deterministic depends on which interpretation of quantum mechanics is right, and that may yet be decided by experiment. Whether anything in nature is strongly emergent is an empirical question too: a brain whose particles disobeyed the laws would be detectable. How decisions are put together in the brain is being worked out, slowly. And what consciousness is, and whether it causes anything, is the largest piece of ignorance in the whole subject; some philosophers, such as Colin McGinn, suspect it may be beyond what human minds can understand.[71]

Other parts would survive any discovery, because they are about what the words should mean and what matters. Hume thought that “the whole controversy has hitherto turned merely upon words”: once you see that liberty means acting according to your will, and not the absence of causes, the dispute dissolves.[12] Compatibilists have said much the same since. Incompatibilists reply that it is not a dispute about words whether a person deserves to suffer for a choice that was fixed before she was born; that is a question about justice. No measurement will settle whether control is enough for desert.

And some of the difficulty may come from the question itself. It asks whether you or your brain decided, as if they were two things competing for the same job. On almost every scientific view they are the same thing described at different levels, the way the temperature of a gas and the motion of its molecules are. Whether the description at the level of persons, with its reasons, choices and responsibility, is as true as the description at the level of particles is what the debate is really about.

Coda

What to take from it

Some things are close to settled. There is no evidence that a mind outside physics pushes on the brain, and most researchers on all sides no longer think free will requires one. Randomness alone does not give freedom, so the fate of free will does not hang on the interpretation of quantum mechanics. The early-brain-signal experiments show that simple movements are prepared before we notice them; they do not show that considered decisions are made without us. And a determined world need not be a predictable one, least of all by the agent inside it.

The rest is open, and each side has a strong case. The incompatibilist’s is that if everything you do follows from causes you did not choose, it is hard to see how you could ultimately deserve credit or blame, and that the long chain of genes, upbringing and circumstance behind every crime and every achievement is not an excuse but a fact. The compatibilist’s is that the chain is what a person is made of, not what imprisons her; that the capacity to weigh reasons and act on them is real, evolved and measurable, and comes in degrees; and that a level of description can be true without being fundamental, like temperature or the arrow of time. The libertarian’s is that our experience of open alternatives is the most direct evidence we have about ourselves, and that physics, which has not decided whether the world is deterministic, cannot overrule it. Which of these is right depends on what freedom has to be, and that is the question that physics, mathematics, neuroscience and biology can inform, but not answer on our behalf.

On method and tools

This piece was written collaboratively with Claude Opus 5.5 (Anthropic): human specification and editorial direction; machine research, analysis and drafting. Research agents gathered and checked the sources; quotations were read in the original, except where noted in the source list. The toy gas and the four figures are computed by scripts/free_will.py, which writes docs/free-will-results.json; the timings, timescales and survey shares in the figures are typed in from the cited papers. The essay takes no side: where the argument is open, it gives each side its strongest form as briefly as it can.

Cover: “the Flammarion engraving”, an anonymous wood engraving first published in Camille Flammarion, L’atmosphère: météorologie populaire (Paris, 1888), public domain, via Wikimedia Commons, placed on a wider background.

Authored by: Luis Matos Ferreira — Physicist, Developer, Writer

Updates and corrections

Last updated 21 October 2026.

1. 21 October 2026. Expanded the passage on Robert Sapolsky: his work as a primatologist, his argument that willpower is itself luck, and the compatibilist reply to it, with two new sources. Later source numbers moved up by one.
2. 21 October 2026. Added a short history of the problem (Augustine, Boethius, Luther and Erasmus, Molina), António Damásio’s work on emotion and unconscious decision, a section on addiction and the law, including Portugal’s, and a cross-cultural study of intuitions, with fourteen new sources.

Go deeper

Ideas this essay uses without stopping to explain them, and where this blog does. The letters in the text point here; the links go straight to the section.

[a] Entanglement

What it means for two particles to share one state, and how to test it. The Interference Machine, “Entanglement, and how to check for it”.

[b] Decoherence

Why quantum states are destroyed by contact with their surroundings. The Interference Machine, “The noise problem”.

[c] The connectome

Why a complete wiring diagram of a worm’s brain does not predict what the worm does. The Complete Parts List, “The connectome does not contain the animal”.

Related essays on this blog

1. Many Ways to Be Clever — how different brains and different AI models solve problems.
2. Grown, Not Built — systems whose behaviour nobody wrote, and who is responsible for them.
3. The Complete Parts List — why biology can list its parts and still not predict the whole.
4. The Interference Machine — what quantum computers do, and the physics underneath.
5. World without time — Rovelli and the order of time (in Portuguese, 2018).

Sources

1. A. Einstein, “What I believe”, Forum and Century 84, 1930, reprinted as “The world as I see it”; English wordings of the line vary. A. Schopenhauer, Prize Essay on the Freedom of the Will (1839), trans. E. F. J. Payne, Cambridge University Press, 1999, chapter II.
2. Augustine, De libero arbitrio (c. 388–395); Boethius, The Consolation of Philosophy (c. 524), trans. H. R. James, book V, prose 6.
3. Erasmus, De libero arbitrio diatribe, Basel, 1524; M. Luther, The Bondage of the Will (1525), trans. H. Cole, section XXV.
4. L. de Molina, Concordia liberi arbitrii cum gratiae donis, Lisbon, 1588; see T. Flint, Divine Providence: The Molinist Account, Cornell University Press, 1998.
5. P.-S. Laplace, A Philosophical Essay on Probabilities (1814), trans. F. W. Truscott and F. L. Emory, Wiley, 1902, chapter II.
6. E. N. Lorenz, “Deterministic nonperiodic flow”, Journal of the Atmospheric Sciences 20, 1963.
7. J. D. Norton, “The dome: an unexpectedly simple failure of determinism”, Philosophy of Science 75, 2008.
8. B. Hensen et al., “Loophole-free Bell inequality violation using electron spins separated by 1.3 kilometres”, Nature 526, 2015; Nobel Prize in Physics 2022 to A. Aspect, J. F. Clauser and A. Zeilinger.
9. J. Conway and S. Kochen, “The free will theorem”, Foundations of Physics 36, 2006; “The strong free will theorem”, Notices of the AMS 56, 2009 (the quotation).
10. S. Hossenfelder and T. Palmer, “Rethinking superdeterminism”, Frontiers in Physics 8, 2020.
11. J. Handsteiner et al., “Cosmic Bell test: measurement settings from Milky Way stars”, Physical Review Letters 118, 2017; D. Rauch et al., Physical Review Letters 121, 080403, 2018.
12. D. Hume, An Enquiry Concerning Human Understanding (1748; titled Philosophical Essays until 1758), section VIII, “Of liberty and necessity”.
13. R. Kane, The Significance of Free Will, Oxford University Press, 1996.
14. S. Hameroff and R. Penrose, “Consciousness in the universe: a review of the ‘Orch OR’ theory”, Physics of Life Reviews 11, 2014.
15. M. Tegmark, “Importance of quantum decoherence in brain processes”, Physical Review E 61, 2000; S. Hagan, S. Hameroff and J. Tuszyński, “Quantum computation in brain microtubules: decoherence and biological feasibility”, Physical Review E 65, 2002 (10⁻⁵ to 10⁻⁴ s, from the preprint’s abstract).
16. S. Aaronson, “The ghost in the quantum Turing machine”, arXiv 1306.0159, 2013.
17. Elisabeth of Bohemia to Descartes, 6/16 May 1643, trans. J. Bennett, Early Modern Texts; see also L. Shapiro (ed. and trans.), The Correspondence between Princess Elisabeth of Bohemia and René Descartes, University of Chicago Press, 2007.
18. A. R. Damásio, Descartes’ Error: Emotion, Reason, and the Human Brain, Putnam, 1994; quotations from the Avon edition, 1995, chapters 3 and 8.
19. S. Carroll, “The laws underlying the physics of everyday life are completely understood”, 23 September 2010; The Big Picture, Dutton, 2016.
20. J. Kim, Mind in a Physical World, MIT Press, 1998.
21. P. W. Anderson, “More is different”, Science 177, 1972.
22. D. J. Chalmers, “Strong and weak emergence”, in The Re-Emergence of Emergence, Oxford University Press, 2006.
23. G. Ellis, How Can Physics Underlie the Mind? Top-Down Causation in the Human Context, Springer, 2016.
24. E. P. Hoel, L. Albantakis and G. Tononi, “Quantifying causal emergence shows that macro can beat micro”, PNAS 110, 2013.
25. C. List, Why Free Will Is Real, Harvard University Press, 2019.
26. J. Loschmidt, “Über den Zustand des Wärmegleichgewichtes eines Systems von Körpern mit Rücksicht auf die Schwerkraft”, Sitzungsberichte der Akademie der Wissenschaften Wien 73, 1876; see J. Uffink, “Boltzmann’s work in statistical physics”, Stanford Encyclopedia of Philosophy.
27. D. Z. Albert, Time and Chance, Harvard University Press, 2000.
28. D. Z. Albert, Time and Chance, chapter 6, and After Physics, Harvard University Press, 2015; B. Loewer, “Counterfactuals and the second law”, in H. Price and R. Corry (eds.), Causation, Physics, and the Constitution of Reality, Oxford University Press, 2007; S. Carroll, From Eternity to Here, Dutton, 2010; C. Rovelli, The Order of Time, 2018.
29. S. Wolfram, A New Kind of Science, Wolfram Media, 2002, chapter 12, “The phenomenon of free will”, p. 750.
30. S. Lloyd, “A Turing test for free will”, Philosophical Transactions of the Royal Society A 370, 2012.
31. B. Libet, C. A. Gleason, E. W. Wright and D. K. Pearl, “Time of conscious intention to act in relation to onset of cerebral activity (readiness-potential)”, Brain 106, 1983.
32. B. Libet, “Unconscious cerebral initiative and the role of conscious will in voluntary action”, Behavioral and Brain Sciences 8, 1985.
33. C. S. Soon, M. Brass, H.-J. Heinze and J.-D. Haynes, “Unconscious determinants of free decisions in the human brain”, Nature Neuroscience 11, 2008.
34. I. Fried, R. Mukamel and G. Kreiman, “Internally generated preactivation of single neurons in human medial frontal cortex predicts volition”, Neuron 69, 2011.
35. A. Schurger, J. D. Sitt and S. Dehaene, “An accumulator model for spontaneous neural activity prior to self-initiated movement”, PNAS 109, 2012.
36. M. Schultze-Kraft et al., “The point of no return in vetoing self-initiated movements”, PNAS 113, 2016; the 200 ms figure as given in summaries of the abstract.
37. U. Maoz, G. Yaffe, C. Koch and L. Mudrik, “Neural precursors of decisions that matter: an ERP study of deliberate and arbitrary choice”, eLife 8, 2019.
38. A. Bechara, H. Damásio, D. Tranel and A. R. Damásio, “Deciding advantageously before knowing the advantageous strategy”, Science 275, 1997.
39. T. V. Maia and J. L. McClelland, “A reexamination of the evidence for the somatic marker hypothesis: what participants really know in the Iowa gambling task”, PNAS 101, 2004; A. Bechara et al., “The Iowa Gambling Task and the somatic marker hypothesis: some questions and answers”, and T. V. Maia and J. L. McClelland, “The somatic marker hypothesis: still many questions but no answers”, Trends in Cognitive Sciences 9, 2005.
40. A. Damásio and S. Hustvedt, “A neurological basis for free will”, Big Think, recorded 2 July 2010 (archived copy).
41. M. S. Gazzaniga, “Cerebral specialization and interhemispheric communication: does the corpus callosum enable the human condition?”, Brain 123, 2000.
42. D. M. Wegner, The Illusion of Conscious Will, MIT Press, 2002; quotation from his “Précis of The Illusion of Conscious Will”, Behavioral and Brain Sciences 27, 2004.
43. J. M. Burns and R. H. Swerdlow, “Right orbitofrontal tumor with pedophilia symptom and constructional apraxia sign”, Archives of Neurology 60, 2003 (abstract); the recurrence as reported in later accounts, not checked in the original.
44. R. M. Sapolsky, A Primate’s Memoir: A Neuroscientist’s Unconventional Life Among the Baboons, Scribner, 2001.
45. R. M. Sapolsky, Determined: A Science of Life Without Free Will, Penguin Press, 2023; quotations from the excerpt in the Harvard Gazette, 17 January 2024; the argument about willpower (chapter 4, “Willing Willpower: The Myth of Grit”) as summarised in reviews, not checked in the book.
46. K. J. Mitchell, Free Agents: How Evolution Gave Us Free Will, Princeton University Press, 2023.
47. B. Brembs, “Towards a scientific concept of free will as a biological trait: spontaneous actions and decision-making in invertebrates”, Proceedings of the Royal Society B 278, 2011.
48. P. Badger, “Determined by Robert Sapolsky”, Philosophy Now 168.
49. D. C. Dennett, Elbow Room: The Varieties of Free Will Worth Wanting, MIT Press, 1984; Freedom Evolves, Viking, 2003.
50. H. G. Frankfurt, “Alternate possibilities and moral responsibility”, Journal of Philosophy 66, 1969; “Freedom of the will and the concept of a person”, Journal of Philosophy 68, 1971.
51. N. D. Volkow, G. F. Koob and A. T. McLellan, “Neurobiologic advances from the brain disease model of addiction”, New England Journal of Medicine 374, 2016.
52. G. M. Heyman, Addiction: A Disorder of Choice, Harvard University Press, 2009; quotation from “Addiction and choice: theory and new data”, Frontiers in Psychiatry 4, 2013.
53. M. Lewis, The Biology of Desire: Why Addiction Is Not a Disease, PublicAffairs, 2015.
54. M’Naghten’s Case (1843) 10 Cl & F 200, answers of the judges to the House of Lords, as quoted in Davis v. United States, 160 U.S. 469 (1895).
55. Código Penal, artigo 20.º, “Inimputabilidade em razão de anomalia psíquica”.
56. Lei n.º 30/2000, de 29 de novembro, Diário da República, in force 1 July 2001; amended by Lei n.º 55/2023.
57. D. J. Chalmers, “Facing up to the problem of consciousness”, Journal of Consciousness Studies 2, 1995.
58. S. Dehaene and L. Naccache, “Towards a cognitive neuroscience of consciousness: basic evidence and a workspace framework”, Cognition 79, 2001.
59. R. F. Baumeister, E. J. Masicampo and K. D. Vohs, “Do conscious thoughts cause behavior?”, Annual Review of Psychology 62, 2011.
60. P. van Inwagen, An Essay on Free Will, Oxford University Press, 1983; wording as quoted in the Stanford Encyclopedia of Philosophy, “Arguments for incompatibilism”.
61. G. Strawson, “The impossibility of moral responsibility”, Philosophical Studies 75, 1994.
62. P. F. Strawson, “Freedom and resentment”, Proceedings of the British Academy 48, 1962.
63. D. Pereboom, Living Without Free Will, Cambridge University Press, 2001.
64. D. Bourget and D. J. Chalmers, “Philosophers on philosophy: the 2020 PhilPapers Survey”, Philosophers’ Imprint 23, 2023; free will results.
65. S. Nichols and J. Knobe, “Moral responsibility and determinism: the cognitive science of folk intuitions”, Noûs 41, 2007; E. Nahmias et al., “Surveying freedom”, Philosophical Psychology 18, 2005.
66. H. Sarkissian, A. Chatterjee, F. De Brigard, J. Knobe, S. Nichols and S. Sirker, “Is belief in free will a cultural universal?”, Mind & Language 25, 2010.
67. K. D. Vohs and J. W. Schooler, “The value of believing in free will”, Psychological Science 19, 2008.
68. Open Science Collaboration, “Estimating the reproducibility of psychological science”, Science 349, 2015 (replication by Embley, Johnson and Giner-Sorolla); N. Buttrick et al., “Many Labs 5”, AMPPS 3, 2020; O. Genschow et al., “Manipulating belief in free will and its downstream consequences: a meta-analysis”, Personality and Social Psychology Review 27, 2023.
69. J. Greene and J. Cohen, “For the law, neuroscience changes nothing and everything”, Philosophical Transactions of the Royal Society B 359, 2004.
70. D. C. Dennett and G. D. Caruso, Just Deserts: Debating Free Will, Polity, 2021.
71. C. McGinn, “Can we solve the mind–body problem?”, Mind 98, 1989.