On the afternoon of a nondescript day in 1983 at the University of California San Francisco, a subject sat in a dimly lit laboratory with electrodes attached to their scalp, watching the sweep of a clock hand. Their instructions were simple: flex their wrist whenever they felt like it, and note the clock position at the moment they first felt the urge to move. Benjamin Libet, a neurophysiologist who had spent decades probing the timing of conscious experience, was waiting to see something he had been theorizing about for years. The EEG traces scrolled across his equipment, recording the electrical activity of the motor cortex in the half-second before each movement. What those traces revealed would generate more philosophical debate than almost any other finding in the history of neuroscience — and they would force a reckoning with one of the oldest questions in human thought.

The readiness potential, a slow negative electrical wave already known to precede voluntary movements, appeared in Libet's data approximately 550 milliseconds before each wrist flexion. This was expected. What was not expected — or rather, what was now measured with unprecedented precision — was the timing of the subjects' reports of their conscious intention. They marked the clock at a position corresponding to roughly 200 milliseconds before the movement. The arithmetic was stark: the brain had begun preparing for the action approximately 350 milliseconds before the person became aware of intending it.

This was Libet's 1983 experiment, published in Brain (Vol. 106, pp. 623-642), and it became the most cited, most argued-about, and most misunderstood result in the neuroscience of volition. Its implications — or at least its apparent implications — for the question of free will were immediate and explosive.

Key Definitions

Free will — the philosophical concept that agents have the capacity to choose and act in ways that are genuinely up to them; the precise meaning depends heavily on whether the definition requires that choices be undetermined by prior physical causes.

Determinism — the thesis that every event, including every human action and mental state, is the inevitable result of prior causes operating according to physical laws; given the state of the universe at any moment, only one future is possible.

Compatibilism — the philosophical position that free will and determinism are not mutually exclusive; free will on this view means acting according to one's own desires, values, and reasons without external compulsion, regardless of whether those desires are themselves causally determined.

Hard determinism — the view that determinism is true and that this fact eliminates free will in any meaningful sense; moral responsibility, deserved punishment, and genuine agency are therefore illusory.

Libertarian free will — not to be confused with political libertarianism — the view that humans possess genuine contra-causal agency, meaning choices are not fully determined by prior physical causes; requires either agent causation or some source of genuine indeterminacy in neural processes.

Readiness potential (Bereitschaftspotential) — a slow, surface-negative EEG wave that begins over the motor cortex and supplementary motor area several hundred milliseconds to over a second before a voluntary movement; first described by Hans Helmut Kornhuber and Luder Deecke in 1965.

Epiphenomenalism — the view that conscious mental states are caused by physical brain processes but do not themselves cause anything; they are effects without causal power, like the shadow of a moving car.

Veto hypothesis — Libet's own interpretation of his results; while unconscious processes initiate voluntary actions, conscious awareness can still intervene to cancel or inhibit the action before it is executed.

Moral luck — the problem, identified by Thomas Nagel and Bernard Williams, that people are held morally responsible for outcomes that depend significantly on factors beyond their control.

The Libet Experiment in Detail

The methodological design of Libet's 1983 study was deceptively simple. Subjects sat before a clock whose hand completed a revolution every 2.56 seconds. They were instructed to flex their right wrist at any moment of their choosing, and to note the position of the clock hand at the moment they first became aware of the wish or urge to move — their "W time." EEG electrodes recorded the readiness potential from the scalp overlying the supplementary motor area, while EMG electrodes on the forearm registered the actual moment of muscle contraction.

The results established three time points relative to the moment of movement:

  • -550 ms: Onset of the readiness potential (Type II — spontaneous movements without pre-planning)
  • -200 ms: Subjects' reported time of conscious intention (W time)
  • 0 ms: Actual muscle movement (EMG onset)

The gap between readiness potential onset and conscious intention was approximately 350 milliseconds. The gap between conscious intention and actual movement was approximately 200 milliseconds. Libet interpreted this as indicating that the brain's preparatory process — whatever neural computation underlies a voluntary action — begins well before the actor is conscious of intending to act.

Libet himself was careful not to draw the conclusion that free will is an illusion. His 1999 paper in the Journal of Consciousness Studies (Vol. 6, pp. 47-57) laid out what became known as the veto hypothesis: even if unconscious processes initiate the action, conscious awareness could still function as a "control" that can cancel the pending movement before execution. The window between W time (-200 ms) and movement onset (0 ms) provides time for conscious veto. Libet saw this not as a consolation prize but as a genuine form of top-down control — will operating as prohibition rather than initiation.

How the Readiness Potential Was Reinterpreted

For almost three decades, the dominant interpretation of Libet's results was that the readiness potential represents unconscious neural decision-making — that the brain "decides" before the mind knows. Then, in 2012, Aaron Schurger, Jacobo Sitaram, and Stanislas Dehaene published a paper in the Proceedings of the National Academy of Sciences (Vol. 109, No. 42, pp. E2904-E2913) that fundamentally reframed what the readiness potential actually measures.

Schurger and colleagues modeled the neural dynamics underlying spontaneous voluntary movements using a leaky stochastic accumulator — essentially a neural circuit that integrates noisy, fluctuating signals over time until a threshold is crossed, at which point movement is triggered. The key insight was that if subjects are instructed to move "at any time they feel like it," they will move precisely when random neural noise happens to push the accumulator over threshold. The slow negative drift of the readiness potential, on this account, is not a signal of unconscious pre-decision; it is the statistical signature of random neural fluctuations sampled precisely because they resulted in threshold crossing.

This reinterpretation is important because it removes the most philosophically damaging element of the original Libet story. If the readiness potential is noise crossing a threshold rather than a deliberate preparatory signal, then there is no "unconscious decision" that precedes conscious intention by 350 milliseconds. The brain is simply running continuously, and movements emerge from the moment neural activity happens to be sufficient to trigger them. Schurger's model made specific predictions that the standard Libet interpretation could not, and subsequent experimental tests have supported the accumulator model.

The fMRI Challenge: Predicting Choices Seconds in Advance

Schurger's reinterpretation did not close the challenge to traditional notions of free will. In 2008, John-Dylan Haynes and colleagues at the Max Planck Institute for Human Cognitive and Brain Sciences published a study in Nature Neuroscience (Vol. 11, pp. 543-545) that was, in many respects, a more serious challenge than Libet's.

Haynes used high-resolution fMRI to record brain activity while subjects freely decided whether to press a button with their left or right hand, pressing as soon as they felt the urge. Using multivariate pattern analysis of activity in prefrontal and parietal regions — particularly the frontopolar cortex — the researchers could predict which hand the subject would use up to 7 to 10 seconds before the subject reported making the decision.

The implications are more difficult to dismiss than Libet's. Libet's gap was 350 milliseconds — short enough to fit within the noise-threshold account. Haynes's gap was 7,000 to 10,000 milliseconds. At this scale, whatever neural patterns predict the choice are not plausibly random fluctuations; they represent sustained, structured information about upcoming behavior that exists entirely below the level of conscious deliberation. Haynes himself was careful to note that prediction accuracy was not perfect (roughly 60% for a 50% chance baseline) and that the relationship between these prefrontal patterns and the eventual decision is not one of simple determination — the patterns reflect probabilistic biases, not fixed outcomes.

Soon et al. (2008) extended these findings, showing that activity in the supplementary motor area and pre-SMA predicted the timing of spontaneous movements, and that frontopolar cortex activity predicted left-right choice, establishing that multiple distinct neural computations underlying voluntary action precede awareness in distinct brain regions.

Philosophical Positions on Free Will

Hard Determinism

Hard determinists hold that the causal closure of the physical world — the principle that every physical event has a sufficient physical cause — entails that human behavior is fully determined by prior physical states. Given the exact configuration of a person's brain and environment at any moment, their next action is the only action physically possible. This view has ancient roots in Stoic philosophy and was given rigorous modern expression by Baron d'Holbach in the 18th century and by contemporary neuroscientists including Robert Sapolsky.

The hard determinist position has significant implications for moral responsibility. If no one could ever have done otherwise, the retributive intuition that wrongdoers deserve punishment becomes difficult to justify. The perpetrator of any crime was shaped by genetics, development, and experience none of which they chose — and that chain of causes extends back before their birth.

Libertarian Free Will

Philosophical libertarians about free will maintain that genuine agency requires that agents could, in identical circumstances, have acted differently — that choices are not fully determined by prior physical causes. This is a metaphysically demanding position. It requires either that humans possess a form of agent causation irreducible to physical causation, or that the fundamental indeterminacy of quantum mechanics somehow propagates into neural computation in a way that generates genuine openness.

The quantum indeterminacy route is widely regarded as insufficient: random quantum events do not produce freedom; they produce randomness. An action that is the result of a quantum coin-flip is not more free than one that is determined — it is simply unpredictable. Agent causation accounts face the challenge of explaining how an agent can be a causal source without being identical to (or at least supervenient on) the physical states of their brain.

Libertarian free will remains the intuitive position for many non-philosophers — the sense that one is genuinely the originating cause of one's choices — but it commands relatively little support among professional philosophers who have examined the arguments in detail.

Compatibilism

Compatibilism is the dominant position in academic philosophy. According to PhilPapers' 2020 survey of approximately 1,700 professional philosophers, roughly 59% endorsed compatibilism, with additional percentage endorsing positions that were broadly compatibilist-leaning. Among those who specialize in philosophy of action and metaphysics, support is higher still.

The compatibilist strategy is to argue that the kind of free will worth wanting — the kind that grounds moral responsibility, self-authorship, and rational agency — does not require that actions be uncaused. What it requires is that actions flow from the agent's own desires, values, deliberations, and reasons, without external compulsion, manipulation, or coercion.

Daniel Dennett, in Elbow Room (1984) and Freedom Evolves (2003), argues that the deliberative processes in human cognition are real causal processes — not illusory — and that they produce real outcomes. When a person reasons about what to do, that reasoning causally influences what they do. The fact that the reasoning process is itself physical and causally embedded does not make it less real or less the person's own. Dennett's position is sometimes called "naturalistic compatibilism" because it grounds free will in the evolved capacities for self-representation, deliberation, and self-correction that distinguish human cognition from simpler systems.

Harry Frankfurt's influential 1971 paper "Freedom of the Will and the Concept of a Person" (Journal of Philosophy, Vol. 68, No. 1, pp. 5-20) proposed that what distinguishes free will is the capacity for second-order volitions — desires about one's own desires. A person acts freely when they act on a desire they identify with and endorse at a higher level of reflection, regardless of whether that desire was itself determined.

The Neuroscience of Deliberation

The prefrontal cortex — particularly the dorsolateral prefrontal cortex (dlPFC) and the anterior cingulate cortex (ACC) — plays a central role in what neuroscientists call executive function: the capacity to hold multiple options in working memory, evaluate them against goals, suppress prepotent responses, and select among competing courses of action. This is the neural substrate of deliberation.

Damage to the dlPFC impairs the capacity for goal-directed reasoning and impulse control. Studies of patients with frontal lobe damage show precisely the pattern that hard determinists cite: that what we call "character" and "rational agency" are functions of physical brain structures susceptible to injury, disease, and pharmacological manipulation. The same point applies to development — adolescents' impulsivity reflects the incomplete myelination of prefrontal circuits, not a moral deficit.

Robert Sapolsky, a Stanford neuroendocrinologist and primatologist, synthesized decades of this evidence in Determined: A Science of Life Without Free Will (2023, Penguin Press). Sapolsky's argument operates by expanding the time horizon of biological causation. A person's behavior at any moment is a function of their biology in that moment; that biology was shaped by their experiences in the preceding hours, days, and years; those experiences interacted with a developmental trajectory shaped by early childhood; childhood was influenced by prenatal environment; prenatal environment was shaped by maternal biology; and so on, through evolutionary history, none of which was chosen. Sapolsky concludes that the chain of causation leaves no room for genuine agency.

Sapolsky advocates replacing retributive punishment with consequentialist frameworks — not because wrongdoers are not dangerous or that their behavior should not be constrained, but because the concept of deserved suffering requires that they could have done otherwise, which he denies.

Position Core Claim Implication for Moral Responsibility
Hard Determinism All actions are causally necessitated; no one could have done otherwise Retributive punishment is incoherent; only forward-looking justifications (deterrence, rehabilitation) survive
Libertarian Free Will Genuine contra-causal agency exists; agents could have done otherwise Traditional moral responsibility intact; deserved praise and blame are justified
Compatibilism Free will = acting from one's own reasons without compulsion; compatible with determinism Moral responsibility survives; grounded in the quality of reasons-responsiveness rather than causal indeterminism
Hard Incompatibilism Neither determinism nor indeterminism leaves room for ultimate moral responsibility Responsibility practices require revision; basic desert is illusory (Derk Pereboom)

Moral Luck and Its Implications

Philosophers Thomas Nagel and Bernard Williams independently raised the problem of moral luck in 1976, and it remains one of the most uncomfortable challenges to ordinary moral thinking. The problem is this: we routinely hold people morally responsible for outcomes that depend substantially on factors beyond their control. Two drivers may behave identically recklessly, but only one happens to strike a pedestrian who steps off a curb. We judge them very differently. Two soldiers in an atrocity may be equally complicit, but one was not at the critical location when the order was given. Their moral fates diverge by chance.

More broadly, people's characters — their tendencies toward cruelty or compassion, impulsivity or deliberation, empathy or indifference — are substantially the products of genetics, upbringing, and circumstance that the person did not choose. If a person grew up in a household that normalized violence, with a genetics that predisposed them to low impulse control and high threat sensitivity, the moral luck problem asks: in what sense do they deserve condemnation in a way that someone raised with advantages does not?

The moral luck problem does not require that determinism is true to have force. Even under libertarian free will, the circumstances that shape choices are unchosen. This is why Nagel concluded that the problem is insoluble within our current moral framework — not that moral judgment should be abandoned, but that moral luck exposes deep tensions that ordinary thought tends to suppress.

Experimental Philosophy and Folk Intuitions

Joshua Knobe, a philosopher at Yale University who helped found the field of experimental philosophy, has conducted studies showing that ordinary people's intuitions about free will and responsibility are more nuanced and context-dependent than philosophers often assume. When people read about a deterministic universe in abstract terms, many report that it eliminates free will. But when they read about specific agents committing harmful actions in a deterministic universe, they tend to reinstate attributions of responsibility — what Knobe and colleagues call the "asymmetry" between abstract and concrete judgments.

A 2010 study by Nichols and Knobe published in Noûs (Vol. 44, No. 4, pp. 663-685) found that when people imagined a fully deterministic universe, they were more likely to deny free will in the abstract while simultaneously affirming it in concrete cases involving harmful behavior. This suggests that ordinary free will attributions are not primarily metaphysical judgments about causal structure — they are responses to the practical needs of social life, particularly the need to assign blame and coordinate behavior.

This finding has been interpreted as supporting a version of compatibilism: people are, at the level of their practical reasoning, already compatibilists, even if they endorse incompatibilism when the question is posed abstractly. This "folk compatibilism" suggests that compatibilist free will is not merely a philosopher's technical revision but corresponds to something in the ordinary practice of holding people responsible.

The Consciousness-Will Relationship

One of the deeper puzzles raised by the Libet results is the relationship between the subjective sense of agency — the experience of willing something — and the neural processes that actually produce behavior. Libet's W time was a report of conscious experience: subjects said "I felt the urge at this clock position." The readiness potential was a neural measurement. The mismatch between them does not necessarily mean that consciousness is epiphenomenal — that it has no causal role — but it raises the question acutely.

Patrick Haggard and Benjamin Libet (2001) proposed that the sense of voluntary action — what Haggard later called "the experience of agency" — is a reconstruction. Rather than reflecting a real-time signal, the W report is assembled retrospectively from multiple sources of neural information and projected backward in time, creating the phenomenal impression of a temporally earlier intention. The experience of will, on this account, is real as experience but does not map cleanly onto the causal sequence that produced the movement.

Daniel Wegner's The Illusion of Conscious Will (2002, MIT Press) pushed this argument further, proposing that the experience of willing is a post-hoc inference rather than a direct read of causal efficacy — that we experience ourselves as causing our actions because thought and action are correlated, not because the thought actually caused the action. Wegner's view has been widely criticized for overclaiming, but his evidence for the gap between experienced agency and actual causal structure is substantial.

The more moderate view — and the one most consistent with current neuroscientific evidence — is that conscious deliberation is causally efficacious, not epiphenomenal, but that its role is different from the naive picture. Consciousness may not initiate actions from nowhere; it may function as a monitoring and modulating system that adjusts ongoing processes, exercises the veto function Libet described, and shapes future dispositions through learning.

Why the Deliberative Process Matters Even Without Libertarian Free Will

Perhaps the most important insight from the compatibilist tradition is that the debate about libertarian free will, however philosophically interesting, does not determine the practical significance of deliberation. Even if every mental state is physically determined, the processes of reasoning, evaluation, and decision-making are real causal processes with real consequences.

When a person deliberates — weighing evidence, considering consequences, imagining alternatives, reflecting on values — those cognitive operations causally influence the neural states that produce behavior. Information matters. Arguments matter. Emotional regulation matters. All of this is true whether or not the deliberative process is itself determined by prior physical states.

This is the compatibilist rescue: not that we have some metaphysically special free will beyond the causal order, but that the causal order includes the kind of self-directed, reasons-responsive processes that constitute the only free will worth having. The neuroscience of decision-making tells us that these processes are physically realized, developmentally shaped, and susceptible to improvement through education, therapy, and experience. None of this requires the supernatural.

"The point is not that the will is free in some spooky metaphysical sense, but that the human capacity for deliberation is real, causally powerful, and the proper object of moral education and legal design." — Daniel Dennett, Freedom Evolves (2003)

Cross-References

References

Tags

free will neuroscience philosophy of mind determinism compatibilism consciousness psychology

Frequently Asked Questions

What did Benjamin Libet's 1983 experiment show about free will?

Libet found that EEG-detected neural activity called the readiness potential began approximately 550 milliseconds before a voluntary wrist movement, yet subjects only reported conscious awareness of the intention to move about 200 milliseconds before the movement. This gap of roughly 350 milliseconds suggested the brain initiates voluntary actions before conscious intention arises. Libet himself proposed a 'veto' hypothesis — that while unconscious processes begin action, conscious will can still cancel it.

What is compatibilism and why do most philosophers accept it?

Compatibilism holds that free will and determinism are not mutually exclusive. On the compatibilist view, free will means acting according to your own desires, values, and reasons without external compulsion or coercion — not that actions are uncaused. According to the PhilPapers survey of professional philosophers, approximately 59 to 75 percent endorse compatibilism, making it the dominant position in academic philosophy. Daniel Dennett is its most prominent contemporary defender.

How did Aaron Schurger's 2012 research change the interpretation of Libet's findings?

Schurger and colleagues proposed that the readiness potential does not represent a pre-decision or unconscious intention at all. Instead, it reflects random fluctuations in neural activity that gradually cross a threshold required to trigger movement. When subjects are told to move 'whenever they feel like it,' they tend to move precisely when this neural noise peaks — making the readiness potential a statistical artifact of the experimental design rather than evidence of unconscious decision-making.

Can brain scanners predict decisions before people are conscious of them?

Yes. John-Dylan Haynes and colleagues published a 2008 study in Nature Neuroscience showing that patterns of activity in prefrontal and parietal cortex could predict whether a subject would press a button with their left or right hand up to 7 to 10 seconds before the subject consciously reported deciding. This is a stronger challenge to conscious free will than Libet's results because it extends the pre-conscious lead time dramatically.

If determinism is true, does criminal punishment make sense?

This is one of the most consequential practical implications of the free will debate. Hard determinists like Robert Sapolsky argue that retributive punishment — the idea that people deserve to suffer for their choices — makes no sense if behavior is fully determined by biology and environment. However, forward-looking justifications for punishment (deterrence, incapacitation, rehabilitation) remain coherent even under determinism, since they treat punishment as a causal intervention rather than a deserved response.

What is libertarian free will in philosophy?

Philosophical libertarianism about free will — not to be confused with political libertarianism — holds that humans possess genuine contra-causal agency: the ability to have done otherwise in identical circumstances. This view requires that human choices are not fully determined by prior physical causes. Most libertarian free will positions involve either agent causation (a person as an irreducible causal source) or quantum indeterminacy as a source of genuine openness in neural processes.

What does Robert Sapolsky argue in 'Determined' (2023)?

In 'Determined: A Science of Life Without Free Will,' Sapolsky argues that every behavior is the product of a cascade of biological and environmental causes stretching back before birth — genes, hormones, prenatal environment, childhood experiences, culture — none of which the individual chose. He concludes that libertarian free will is biologically impossible and advocates for a legal and moral system based on consequentialist rather than retributive grounds.

Share this article