FREE WILL: ARE WE REALLY IN CONTROL OF OUR DECISIONS?

You pause before replying to a difficult message. An impulse arrives, you consider the consequences, and you choose different words. It feels like control. Yet your decision also has a history: habits, emotions, reasons and a brain already at work before you describe what you intend. The question of free will begins in that tension. Neuroscience can investigate parts of the process. Whether those findings abolish freedom depends on what freedom is supposed to require, and that remains a philosophical dispute as well as a scientific investigation.

Which kind of freedom are we asking about?

Free will can refer to acting for reasons, having meaningful alternatives or being the appropriate source of an action. These ideas are related, but an experiment may challenge one without disposing of all the others. A person forced to hand over a wallet and a person choosing to give money can perform similar movements under very different conditions. Before asking when the brain decided, we need to ask what distinguishes a controlled action from coercion, compulsion or accident. The definition is part of the problem, not a detail to skip. Stanford Encyclopedia of Philosophy: free will

What Libet’s experiment found

In the influential 1983 study by Benjamin Libet and colleagues, participants made simple self-initiated movements and reported when they first became aware of an intention to move. An averaged electrical signal, the readiness potential, began before their reported awareness. In one reported comparison, the gap averaged about 350 milliseconds. The observation challenged the intuitive story of a fully conscious command appearing first and all relevant preparation following afterward. It did not directly measure the beginning of every kind of decision, or establish that a person’s reported moment was an exact timestamp for intention itself. Libet and colleagues: conscious intention and cerebral activity, 1983

The seductive headline is that the brain decides before you do. But the word you creates a hidden separation: your brain is part of the person making the decision. The important issue is what role conscious deliberation plays within that person, not whether decisions have neural activity. A clock report, an electrical average and a philosophical concept are different kinds of evidence. Moving from the first two to a verdict about all human freedom requires additional premises that the apparatus cannot supply on its own.

Why an average need not reveal a completed decision

Schurger, Sitt and Dehaene offered an alternative interpretation in 2012. Their accumulator model showed how fluctuating neural activity, approaching a movement threshold, could produce a readiness-potential-like average when trials were aligned to movement onset. This matters because a smooth buildup in an average need not mean that every individual trial contains a settled decision steadily unfolding from the same early point. The model does not prove free will. It challenges one inference about what a familiar signal represents, making the experimental question more precise. Schurger, Sitt and Dehaene: an accumulator model, 2012

Prediction is not the same as inevitability

Soon and colleagues’ 2008 brain-imaging study found information about an upcoming simple choice before participants reported deciding. This kind of decoding is striking, but its interpretation depends on predictive accuracy, the task and what the measured activity encodes. Above-chance prediction is not perfect foreknowledge, and a predictor may capture a developing bias rather than an irrevocable commitment. Even a successful prediction in a constrained laboratory task does not establish that complex decisions are fixed in the same way. Prediction, causation and metaphysical determinism should not be treated as synonyms. Soon and colleagues: unconscious determinants of decisions, 2008

Consequential choices can look different

A 2019 experiment by Uri Maoz and colleagues compared arbitrary choices with deliberate decisions involving donations to nonprofit organizations. The expected readiness potential appeared for arbitrary choices but was absent in their deliberate condition. The result challenges the generalization from purposeless button presses to decisions guided by reasons. It does not show that deliberate choice escapes the brain, or demonstrate an uncaused self. It shows why the type of decision matters. A study of when to move is not automatically a study of how someone weighs competing values. Maoz and colleagues: deliberate and arbitrary choice, 2019

An action can still be cancelled

In a 2016 brain-computer-interface experiment, Schultze-Kraft and colleagues used early movement-related activity to trigger stop signals. Participants could still cancel movements after the readiness potential had begun, provided the stop signal arrived early enough. In that task, cancellation became difficult near a point roughly 200 milliseconds before movement onset. This identifies a limit within a particular motor-control setup. It is not a universal deadline for changing one’s mind, nor proof of a nonphysical veto. It does show that an early preparatory signal need not make the action unstoppable. Schultze-Kraft and colleagues: cancelling self-initiated movements, 2016

Why philosophers disagree about determinism

Determinism, roughly, says that the past and the laws of nature fix a single future. Compatibilists argue that meaningful freedom can nevertheless involve acting through one’s own reasons and capacities without the wrong kind of constraint. Incompatibilists reject that compatibility; some defend a freedom requiring alternatives under the same past, while others doubt that such freedom exists. These are disagreements about what control requires, not three laboratory findings. The fact that an action has causes does not, without further argument, establish which account of responsibility is correct. Stanford Encyclopedia of Philosophy: compatibilism

Randomness does not automatically solve the problem. If a decision changed because of an uncontrolled chance event, it would not obviously become more yours. A libertarian account of agency therefore needs more than unpredictability, just as a compatibilist account needs more than saying that we do what we want. It must explain which wants, influences and capacities support control. Neuroscience can inform these arguments by describing mechanisms and limitations; it cannot select a philosophical definition simply by producing a scan. Stanford Encyclopedia of Philosophy: free will

What remains useful even without a final verdict

The everyday distinction between reconsidering a reason and being physically prevented from acting does not disappear because the ultimate debate remains open. But turning that distinction into a theory of moral desert requires further ethical argument. A careful reader can reject both sensational shortcuts: that one brain signal proves nobody controls anything, or that the subjective feeling of choice proves independence from every cause. Our article on AI consciousness examines a related mistake, confusing a persuasive report with a complete explanation of a mind.

The strongest conclusion is narrower and more interesting than a verdict that freedom is real or fake. The feeling of deciding is not a transparent view of every process producing an action. Preparation can precede awareness, tasks can recruit different mechanisms, and cancellation can remain possible after preparation begins. What kind of freedom survives those facts is still contested. Keeping the experiment, its interpretation and the philosophical claim separate is how the debate becomes more illuminating. For another boundary between explanation and ultimate questions, read why anything exists.

Related reading: Can artificial intelligence ever become conscious?.