Loading field map
Free will is the capacity of an agent to act or choose among alternatives without total constraint by prior causes, external forces, or fate. The philosophical problem of free will arises because two seemingly obvious features of human life appear to conflict. First, we experience ourselves as genuine authors of our actions: we deliberate, weigh reasons, and feel that it is up to us which option we take. Second, we live in a world that science describes as governed by causal laws, and we increasingly understand human behavior as the product of genetics, environment, and neural processes. The central question of the subfield is whether these two pictures can both be true, and if not, which one must give way.
The stakes are not merely abstract. Most legal systems assume that people can be held responsible for their actions only if they could have done otherwise. Moral praise and blame, feelings of guilt and pride, and the practice of punishment all seem to presuppose that the agent is the true source of her conduct. If free will is an illusion, it is unclear what remains of moral responsibility, and some philosophers argue that our practices of punishment would need radical revision. Others argue that responsibility can survive even without free will, or that the very terms of the debate are confused.
The free will problem is best understood as a cluster of related questions. The first concerns the nature of the will itself: is the will a distinct faculty of choice, or is it simply a name for the way desires and beliefs produce action? The second concerns causation: if every event, including every human decision, has a cause, then how can any choice be free? The third concerns alternatives: does freedom require that the agent could have done otherwise in exactly the same circumstances, or is it enough that the agent acted in accordance with her own desires and character?
Three concepts recur throughout the literature. Determinism is the thesis that every event, including every human action and decision, is the necessary result of prior states of the world together with the laws of nature. Given the past and the laws, only one future is possible. Compatibilism is the view that free will and determinism can both be true. Incompatibilism is the view that they cannot. Incompatibilists divide further: hard determinists accept determinism and conclude that free will does not exist; libertarians accept that free will exists and therefore deny that determinism is true, holding that human choices involve a genuine openness not found elsewhere in nature.
A further distinction, introduced by the philosopher Harry Frankfurt, has reshaped the debate. Frankfurt distinguished between alternative possibilities—the ability to do otherwise—and sourcehood—the idea that the agent is the ultimate origin of her action. He argued that moral responsibility depends on sourcehood, not on alternative possibilities. This distinction matters because it suggests that even in a deterministic world, an agent might be responsible if her action flows from her own values and commitments, provided those values are genuinely hers.
The problem of free will is ancient, though the modern terms of the debate emerged only gradually. In ancient Greek philosophy, the issue was often framed as a question of fate and moral responsibility. The Stoics held that everything happens according to divine reason or fate, yet they still maintained that human actions are up to us, because our assent to impressions is within our power. Aristotle discussed voluntary action and responsibility without a fully developed concept of determinism, focusing instead on whether an action originates in the agent.
The Christian tradition sharpened the problem by adding divine foreknowledge and omnipotence. If God knows everything that will happen, and if God's will is the cause of all things, then how can humans sin freely? Augustine and later medieval thinkers struggled with this, developing distinctions between necessity and compulsion, and between what God foreknows and what God causes. Thomas Aquinas argued that God's foreknowledge does not destroy contingency in human actions, because God knows future events as present, not as necessary. The problem of divine foreknowledge remains a distinct sub-problem within the broader free will debate, though it is often treated separately from the problem of causal determinism.
The modern problem took shape in the seventeenth and eighteenth centuries, as the scientific picture of a law-governed universe replaced the older theological framework. Thomas Hobbes argued that freedom is simply the absence of external impediment, and that a free action is one that follows from the agent's own will, even if that will is itself caused. This is the first clear statement of compatibilism. David Hume developed the view further, arguing that the conflict between liberty and necessity is merely verbal: liberty is the power of acting according to the determinations of the will, and necessity is the constant conjunction of motives and actions. For Hume, the real threat to freedom is not causation but chance, since random events would undermine responsibility just as much as compulsion.
Immanuel Kant took a different path. He argued that the causal determinism of the phenomenal world—the world as it appears to us—is compatible with freedom because freedom belongs to the noumenal realm, the world as it is in itself, which is not subject to time and causation. This solution is widely regarded as obscure, but it influenced later German idealism and the distinction between the standpoint of explanation and the standpoint of agency.
In the twentieth century, the debate became more technical and more closely tied to developments in physics, psychology, and the philosophy of language. The rise of quantum mechanics raised the question of whether indeterminism at the microscopic level could provide the kind of openness that libertarians require. Most philosophers concluded that quantum randomness is not the same as free will, since random events are not under the agent's control. Meanwhile, the behavioral and cognitive sciences increasingly treated human decision-making as a natural phenomenon subject to causal explanation, intensifying the pressure on libertarian views.
Contemporary compatibilism has moved well beyond the simple claim that freedom is just the absence of external constraint. The most influential modern version is the hierarchical theory, developed by Harry Frankfurt and others. On this view, a person has free will when her first-order desires—desires to do particular things—are aligned with her second-order desires, her desires about what she wants to desire. A drug addict who wants to quit but cannot resist the craving is unfree because her first-order desire conflicts with her second-order desire. A satisfied addict, who embraces her addiction, is free in Frankfurt's sense, because her will is wholehearted. The key point is that freedom is a matter of the structure of the will, not of causal origins.
A second major compatibilist approach is the reasons-responsive view, associated with philosophers such as John Martin Fischer and Mark Ravizza. On this view, an agent is morally responsible if her mechanism of practical reasoning is responsive to reasons—that is, if she would have acted differently had there been sufficient reason to do so. This does not require that she could have done otherwise in the same circumstances; it requires only that her decision-making process is the kind that tracks reasons. A person who acts from a reasons-responsive mechanism is responsible even if the mechanism is itself determined.
A third approach, sometimes called mesh compatibilism, holds that freedom consists in the agent acting in accordance with her deep self—her values, character, and commitments—provided that the deep self was formed in a way that the agent can endorse. This view faces the challenge of explaining what makes the deep self genuinely the agent's own, rather than merely a product of conditioning. The problem of manipulation is a standard test case: if an evil neuroscientist implants desires in a person's brain, the person may act in accordance with her deep self, yet we are reluctant to call her free. Compatibilists must explain what distinguishes ordinary causal formation from manipulation.
Libertarians agree with hard determinists that free will requires indeterminism, but they deny that determinism is true. The central challenge for libertarianism is to explain how indeterminism can make an action free rather than merely random. If a decision is not fully caused by prior states, then in what sense is it the agent's decision? The classic worry is that indeterminism introduces luck: if two possible futures are open, and the agent chooses one, the choice seems arbitrary.
Contemporary libertarians have developed several responses. Event-causal libertarians hold that free actions are caused by the agent's reasons, but that the causal process is nondeterministic. The agent's beliefs and desires make the action probable but not necessary, and the agent's deliberation is a genuine causal factor in which outcome occurs. The problem of luck remains: if the same prior state could lead to either outcome, the difference between choosing and not choosing seems to be a matter of chance.
Agent-causal libertarians argue that the agent herself, as a substance, causes her actions without being caused to do so. On this view, the agent is a cause that is not itself an event, and her causation is not reducible to the causation of her desires or beliefs. This view, defended by philosophers such as Roderick Chisholm and Timothy O'Connor, preserves the intuitive idea that the agent is the true origin of her action. Critics object that agent causation is mysterious and that it posits a kind of causation that is nowhere else in nature.
A third libertarian strategy, associated with Robert Kane, is the ultimate responsibility view. Kane argues that free will requires that the agent be ultimately responsible for her character and motives, not just for her immediate actions. This requires that at some point in her life, the agent performed self-forming actions, in which she chose between competing sets of values under conditions of genuine uncertainty. These self-forming actions are undetermined, but they are not random because they are the agent's own efforts to shape herself. Critics question whether such actions are possible and whether they can ground responsibility for later, determined actions.
The view that free will does not exist has been defended on both deterministic and indeterministic grounds. Hard determinism is the classical version: if determinism is true, and if free will requires indeterminism, then free will is an illusion. Some hard determinists, such as the philosopher Galen Strawson, argue that free will is impossible regardless of whether determinism is true. Strawson's basic argument holds that to be truly responsible for an action, one must be responsible for one's character, but to be responsible for one's character, one must have chosen it, which would require a prior character, leading to an infinite regress. Therefore, ultimate responsibility is impossible.
Hard incompatibilism, defended by Derk Pereboom, extends this skepticism to both deterministic and indeterministic worlds. Pereboom argues that neither event-causal nor agent-causal libertarianism can provide the required control, and that compatibilist accounts fail because they cannot exclude manipulation cases. He concludes that free will, understood as the kind of control required for moral responsibility in the desert sense, does not exist. However, Pereboom and other skeptics do not conclude that we should abandon all moral practices. They argue for a revised conception of responsibility that is forward-looking rather than backward-looking: we may still hold people responsible in the sense of engaging with them, encouraging better behavior, and protecting society, but we should not blame them in the sense of treating them as deserving punishment.
A more recent development is revisionism, associated with philosophers such as Manuel Vargas. Revisionists argue that the traditional concept of free will is indeed incoherent or unsupported, but that we should not simply abandon the term. Instead, we should revise our concept of free will so that it serves the useful functions it has always served—such as supporting moral responsibility and social coordination—without relying on false metaphysical assumptions. Revisionism is thus a middle path between compatibilism and skepticism: it agrees with skeptics that the traditional concept is flawed, but it agrees with compatibilists that we can keep a meaningful notion of free will by changing what we mean by it.
The major approaches are not simply rival answers to a single question; they often disagree about what the question is. Compatibilists and libertarians disagree about whether free will requires the ability to do otherwise. Hard incompatibilists and revisionists disagree about whether the concept of free will can be salvaged. The debate is structured by a few key fault lines.
The first fault line is the consequence argument, formulated by Peter van Inwagen. The argument holds that if determinism is true, then our actions are the consequences of the past and the laws of nature, over which we have no control. Since we cannot change the past or the laws, we cannot change the consequences, and therefore we cannot do otherwise. Compatibilists respond in various ways: some deny that "cannot do otherwise" follows from determinism in the relevant sense; others argue that the relevant sense of "can" is not the one the argument targets; still others accept that we cannot do otherwise but deny that free will requires this.
The second fault line is the manipulation argument. If a person is manipulated into acting in a certain way, we do not hold her responsible. But if determinism is true, then all actions are the result of factors outside the agent's control, and the difference between manipulation and ordinary causation seems to be only a matter of degree. Compatibilists must explain why ordinary causal history is different from manipulation. Some argue that the difference lies in the agent's endorsement of her desires; others argue that it lies in the responsiveness of her reasoning to reasons. Critics respond that a skilled manipulator could create an agent who endorses her desires and is responsive to reasons, yet is still not free.
The third fault line concerns the nature of moral responsibility. Compatibilists typically hold that responsibility is a matter of the agent's relationship to her action: whether she is the appropriate target of reactive attitudes such as resentment and gratitude. Libertarians typically hold that responsibility requires that the agent be the ultimate source of her action. Skeptics argue that neither account can satisfy the demands of genuine desert. The debate thus depends on what we take responsibility to be, and this is itself a contested question.
The contemporary free will debate is characterized by several developments. First, the debate has become increasingly interdisciplinary. Neuroscientists such as Benjamin Libet have conducted experiments that appear to show that brain activity precedes conscious intention, raising questions about whether conscious will is causally effective. Philosophers have responded with careful analysis of what such experiments can and cannot show, noting that the timing of neural events does not settle the question of whether the agent is the author of her action. The relevance of neuroscience to free will remains a live and contested issue.
Second, the debate has expanded beyond the traditional Western canon. Philosophers working in Buddhist, Hindu, and Confucian traditions have contributed perspectives that do not fit neatly into the compatibilist-libertarian framework. For example, some Buddhist thinkers argue that the idea of a substantial self is an illusion, and that the question of free will must be reformulated in terms of the conditions of action and responsibility without a permanent agent. These traditions complicate the assumption that free will is a universal problem with a single set of terms.
Third, the debate has become more attentive to the social and political dimensions of responsibility. Feminist philosophers have examined how attributions of responsibility are shaped by power relations, and how the concept of free will has been used to hold oppressed groups responsible for conditions they did not choose. This line of inquiry does not simply ask whether free will exists, but who gets to be treated as a free agent and under what conditions.
Fourth, the debate has seen a revival of interest in moral responsibility without free will. Some philosophers argue that even if free will is an illusion, we can still make sense of responsibility as a set of social practices that regulate behavior. This view, sometimes called responsibility without desert, holds that we can hold people responsible in the sense of expecting them to conform to norms and responding to violations with sanctions, without claiming that they deserve those sanctions in a deep metaphysical sense.
The field remains deeply divided, and no consensus is in sight. What unites the various approaches is a shared commitment to taking seriously both the phenomenology of choice and the scientific picture of the world. The free will debate is not merely a technical exercise; it is an attempt to understand what it means to be an agent in a world that does not obviously make room for agency. The persistence of the debate suggests that the question touches something fundamental about human self-understanding, and that the tension between our experience of freedom and our knowledge of causation is not easily resolved.