ClearFactAll topics

Do we actually have free will?

12 peer-reviewed studies for, 4 against

Key takeaways

Stronger evidence: For

Verdict

Taken together, the empirical literature neither cleanly vindicates nor decisively refutes free will, but it does undercut some simple stories on both sides. Libet-style EEG work and later fMRI decoding studies show that unconscious neural processes reliably precede reported decisions and can weakly predict upcoming actions. On a naïve picture where free will requires a conscious “first cause” that initiates behavior ex nihilo, this is clearly bad news.

However, more recent modeling and review work shows that these findings are compatible with views of free will grounded in reasons-responsive control. Accumulator models reinterpret readiness potentials as stochastic threshold crossings rather than fixed prior decisions, and philosophical–empirical work indicates that many people understand free will as acting from one’s own motives even in a deterministic brain. Comprehensive reviews explicitly argue that current neuroscience does not demonstrate the nonexistence of free will, but instead invites more precise accounts of how conscious control is implemented within causal processes. On balance, the best-supported position is that some forms of robust free will are still on the table, though they must be compatible with a brain that begins preparing and biasing actions before conscious awareness catches up.

Strongest study on each side

Ranked by citation count.

Strongest study for

An accumulator model reinterprets the readiness potential as threshold-crossing of neural noise, not a preconscious decision (Proceedings of the National Academy of Sciences, 2012)

678 citations

Schurger and colleagues propose that the classical readiness potential preceding self-initiated movement reflects a stochastic accumulator approaching threshold rather than a specific, time-locked decision to act. Simulations and EEG data show that random fluctuations can generate the characteristic slow buildup when trials are aligned to movement onset. This substantially weakens the inference from readiness potentials to the claim that the brain has already decided before conscious intention arises, leaving room for conscious control in how and when thresholds are set.

Strongest study against

Unconscious brain activity can predict which option a person will choose up to 10 seconds before awareness (Nature Neuroscience, 2008)

2,700 citations

Using fMRI and multivariate pattern analysis, Soon and colleagues showed that patterns in frontopolar and parietal cortex predicted whether a participant would press a left or right button several seconds before they reported deciding. Although prediction accuracy was modest, the above-chance performance suggests that the neural decision process is underway well before conscious awareness of choosing, reinforcing worries that conscious will is more a readout of prior neural events than the true originator.

Open the interactive page →

The evidence

For (12)

An accumulator model reinterprets the readiness potential as threshold-crossing of neural noise, not a preconscious decision (Proceedings of the National Academy of Sciences, 2012)

678 citations

Schurger and colleagues propose that the classical readiness potential preceding self-initiated movement reflects a stochastic accumulator approaching threshold rather than a specific, time-locked decision to act. Simulations and EEG data show that random fluctuations can generate the characteristic slow buildup when trials are aligned to movement onset. This substantially weakens the inference from readiness potentials to the claim that the brain has already decided before conscious intention arises, leaving room for conscious control in how and when thresholds are set.

Folk intuitions often treat free will and moral responsibility as compatible with determinism (Philosophical Psychology, 2005)

Eddy Nahmias, Stephen Morris et al. · 358 citations

Survey experiments by Nahmias and colleagues present lay participants with deterministic scenarios in which agents have their mental states fully caused by prior events. Many respondents still ascribe free will and moral responsibility, suggesting that ordinary concepts of freedom do not require metaphysical indeterminism. This supports compatibilist theories on which free will is about acting from one’s reasons-responsive capacities, something that can exist even in a causally ordered brain.

Experimental work on folk intuitions supports compatibilist judgments under mechanistic descriptions (Midwest Studies in Philosophy, 2007)

181 citations

In further experiments, participants read scenarios where agents’ choices are produced by deterministic mechanisms. A substantial proportion still judge the agents to act freely and be responsible, especially in concrete, emotionally charged cases. These findings undermine the claim that commonsense treats determinism as automatically destroying free will, and instead suggest that many people understand free will in terms of internal control and responsiveness, which may be consistent with neuroscientific causal stories.

Neuroscientific data do not straightforwardly disprove free will and are compatible with control-based accounts (Neuroscience and Biobehavioral Reviews, 2019)

113 citations

This review argues that common interpretations of Libet-style readiness potentials and fMRI prediction studies overstate their threat to free will. The authors show that these signals can be understood as parts of a multi-stage decision process, that timing measures are noisy, and that many philosophical accounts of free will focus on rational control rather than momentary initiation. They conclude that existing neuroscientific findings do not demonstrate that free will is an illusion, but instead invite more precise models of how conscious control is implemented in the brain.

Readiness potential shape and timing support stochastic accumulator interpretations over fixed preplanning (eNeuro, 2018)

42 citations

Follow-up work links trial-by-trial variation in the readiness potential to properties of an underlying stochastic accumulator process. The study shows that both the distribution of waiting times and the average EEG waveform can be reproduced without positing a deterministic, unconsciously fixed decision far in advance. These results further erode the simple picture in which early neural activity straightforwardly proves that conscious decisions are epiphenomenal.

Making sense of agency: Belief in free will as a unique and important construct (Social and Personality Psychology Compass, 2017)

Gilad Feldman · 42 citations

Feldman's comprehensive review examines how belief in free will functions as a distinct construct affecting motivation, self-control, academic performance, and learning outcomes. The paper integrates evidence showing that free will beliefs predict higher achievement, job satisfaction, and capacity for personal agency—distinct from self-efficacy or self-control alone.

Does Belief in Free Will Increase Support for Economic Inequality? (Collabra: Psychology, 2020)

Brett Mercier, Dylan Wiwad et al.

Five studies examining whether belief in free will influences support for economic inequality. Study 1 finds belief in free will is associated with increased support for inequality at country and individual levels. Studies 3-4 show people more willing to support inequality in scenarios where free will exists versus not, suggesting belief in free will justifies inequality.

Against (4)

Unconscious brain activity can predict which option a person will choose up to 10 seconds before awareness (Nature Neuroscience, 2008)

2,700 citations

Using fMRI and multivariate pattern analysis, Soon and colleagues showed that patterns in frontopolar and parietal cortex predicted whether a participant would press a left or right button several seconds before they reported deciding. Although prediction accuracy was modest, the above-chance performance suggests that the neural decision process is underway well before conscious awareness of choosing, reinforcing worries that conscious will is more a readout of prior neural events than the true originator.

Readiness potentials precede reported conscious intention by hundreds of milliseconds, suggesting unconscious initiation of voluntary acts (Brain, 1983)

2,300 citations

Libet’s classic experiments compared the onset of the readiness potential with participants’ reported time of deciding to move. The EEG signal typically began about 500 ms before movement, whereas conscious intention was reported only around 200 ms before action. Libet interpreted this as showing that the brain initiates voluntary acts before conscious will, leaving consciousness only a narrow time window for veto. Many take this as evidence that conscious free will does not initiate actions in the way we normally think.

Abstract intentions can be decoded seconds in advance from frontopolar cortex activity (Proceedings of the National Academy of Sciences, 2013)

366 citations

In a related paradigm, Soon and colleagues asked subjects to freely choose which task to perform and found that brain activity in anterior frontopolar cortex predicted the content of these abstract intentions several seconds before participants reported forming them. The existence of such preconscious neural signatures for high-level intentions is often cited as further evidence that what feels like a spontaneous, undetermined choice is in fact determined by prior brain states.

Predictive brain activity can be explained by subpersonal evidence accumulation rather than conscious deliberation (Neuroscience and Biobehavioral Reviews, 2014)

150 citations

Bode and colleagues review work on predicting decisions from early brain signals and argue that many findings fit models of unconscious evidence accumulation. On these views, choices emerge from gradual, noisy integration of information that crosses a decision threshold before entering awareness. If most of the choice-forming work is done by such subpersonal processes, this challenges traditional pictures of a conscious agent freely originating actions at will.

The debate

Monica: Priya, I know you think free will is just an illusion, but the latest neuroscience actually supports our ability to make genuine choices. Yes, those famous Libet experiments showed brain activity before conscious decisions, but that doesn't kill free will—it just means we need a more sophisticated understanding of it.

Priya: Come on, Monica. If my brain is already deciding what I'll do before "I" even know it, how is that free? Those EEG studies and fMRI scans prove that what we call "choice" is just our consciousness catching up to decisions already made by unconscious neural processes. We're passengers in our own minds.

Monica: That's exactly the oversimplified view that recent research has debunked! You're assuming free will requires some magical "first cause" that creates decisions out of nothing. But the new accumulator models show those readiness potentials aren't fixed prior decisions—they're more like stochastic processes building up to a threshold. Your brain isn't deciding for you; it's preparing options that you can still influence.

Priya: But influence how? If it's all just neurons firing according to physical laws, where's the room for genuine agency? Every thought, every deliberation, every "choice" is just the inevitable result of prior brain states. You can dress it up in fancy terminology, but determinism is determinism.

Monica: Here's what you're missing—free will doesn't require escaping causation. It requires reasons-responsive control. When I choose chocolate over vanilla because I prefer the taste, that's free will in action, even if my preference emerges from a physical brain. The key insight from philosophy and empirical work is that most people intuitively understand free will as acting from their own motives and values, not as some impossible self-creation.

Priya: But those motives and values themselves come from genetics, upbringing, brain chemistry—factors completely outside your control. How can you call it "free" when every link in the chain traces back to things you never chose?

Monica: Because freedom isn't about ultimate self-origination—it's about the capacity for rational deliberation and control. Look, comprehensive reviews of the neuroscience explicitly state that current evidence doesn't demonstrate free will's nonexistence. Instead, it's helping us understand how conscious control works within causal processes. My brain preparing actions before I'm aware isn't a bug—it's a feature that allows for sophisticated, multi-level control.

Priya: That sounds like rationalization to me. You're redefining free will to match what the science allows, rather than accepting what it reveals about our lack of ultimate control.

Monica: No, I'm recognizing that the naive picture of free will was wrong from the start. The evidence shows we have robust forms of agency—we can respond to reasons, reflect on our choices, and act according to our values. That's real freedom, even in a deterministic universe. Free will is still very much on the table; we just understand it better now.

Frequently asked questions

Did the Libet experiment disprove free will?

Libet found that a brain signal called the readiness potential began about 500 ms before movement, while people reported deciding only about 200 ms before acting, which he took as unconscious initiation of action. Later modeling suggests this signal may reflect noisy neural activity drifting toward a threshold rather than a decision already made. Critics also argue that results from arbitrary choices like when to move cannot be extended to deliberate decisions.

Can scientists predict your decisions before you make them?

fMRI studies have predicted which button a person would press, and which task they would choose, several seconds before they reported deciding. The prediction accuracy was modest, though above chance. Some researchers see this as evidence that choices form before awareness, while others interpret it as one stage of a longer decision process.

Is free will compatible with determinism?

Compatibilist accounts hold that free will means acting from one's own reasons-responsive capacities, which can exist in a causally ordered brain. Survey experiments show many people still judge agents to act freely and be responsible even when their choices are fully caused by prior events. On this view, free will does not require a choice that is uncaused.

Does believing in free will affect behavior?

A meta-analysis of 145 experiments found that arguments against free will reduce people's belief in it, but found no clear evidence of downstream effects on behavior. Other research links belief in free will to motivation, achievement and personal agency, and some studies associate it with greater support for economic inequality.

Related questions