Libertarian Free Will

The Philosophical Problem and Its Scope

Libertarian free will concerns the conditions under which an agent can be the determining source of a choice when the state of the world and its laws do not settle that choice in advance. In the terminology of the contemporary debate, libertarianism combines two claims: free will of the relevant kind is incompatible with determinism, and at least some free choices occur. The first is shared with hard incompatibilists; the second separates libertarianism from skeptical positions. It is important to state both claims. The denial of determinism alone establishes neither a free decision nor an agent whose action deserves a particular attribution of responsibility (Clarke, 2003; O’Connor & Franklin, 2022).

The fundamental question is whether a person can choose for reasons in a way that is genuinely up to that person even when the complete antecedent state of the world and its governing laws leave multiple futures open. This question contains several distinguishable requirements. A decision must count as an action; it must be attributable to the agent rather than simply happening in the agent; it must be guided or explained by reasons; and, on many accounts, the agent must have been able to settle a different alternative at the time. These requirements can support one another, but none is conceptually identical with the others. One may perform a voluntary movement without facing an open alternative, deliberate under uncertainty in a deterministic system, or experience competing reasons without obtaining the metaphysical power of choosing between them.

The object of inquiry is a family of philosophical theories, not the political doctrine also called libertarianism. Its central claims concern metaphysics, causation, agency, practical rationality, and the grounding of moral responsibility. The research separates the existence question—whether agents ever possess this freedom—from the conceptual question—what possession of it would require—and the explanatory question—how an instance of it could occur. Those questions also distinguish a defensible metaphysical possibility from an empirically established account of ordinary human decision making. Contemporary discussions increasingly recognize this separation when evaluating both philosophical models and neuroscientific findings (Mele, 2006; Franklin, 2018).

A stringent theory of libertarian freedom therefore requires an account of more than the negation of deterministic causation. It must specify the bearer of control, the relevant control relation, the location of indeterminism in action production, and the role of the reasons for which an action is done. It must also explain what would be lost if the same process were causally determined. The philosophical attraction of libertarianism is that it seeks a deep form of self-determination; its enduring challenge is that an undetermined transition can appear less controlled rather than more controlled.

Determinism, Indeterminism, and the Meaning of Freedom

Causal determinism is ordinarily formulated as the thesis that a complete description of the world at a time, together with its laws of nature, entails a unique future. This is a thesis about the structure of the world under laws, rather than a thesis about predictability, psychological compulsion, inevitability in ordinary speech, or the confidence of an observer. A deterministic process can be computationally unpredictable, while a simple probabilistic process may be predictable in the statistical sense. Neither feature by itself settles whether an agent exercises free will (van Inwagen, 1983; O’Connor & Franklin, 2022).

Indeterminism is the denial of determinism. It may take the form of genuinely open causal outcomes, stochastic transitions, or accounts on which an agent exercises a distinctive power to settle a choice. The word itself is therefore permissive rather than explanatory. An indeterministic event can be a random decay, a probabilistic neural transition, or an action whose production a libertarian theorist hopes to show is under the agent’s control. If the same metaphysical label is applied to each, the difference between chance and agency has yet to be given.

Incompatibilism maintains that the freedom under dispute cannot exist in a deterministic world. Libertarianism affirms that such freedom sometimes exists; hard determinism denies it because determinism is true; hard incompatibilism remains skeptical of the freedom even where determinism might be false. Compatibilism argues that free agency can coexist with determinism, often identifying sufficient control with suitable responsiveness to reasons, capacities for self-regulation, lack of bypassing coercion, or appropriately owned decision-making mechanisms. These rival positions do not always mean precisely the same thing by “free will.” A comparison must therefore keep the target capacity fixed rather than declare a result on the basis of a verbal switch.

A useful analytical distinction concerns the modality of an agent’s abilities. An ability in a broadly dispositional sense may be possessed even when its exercise is blocked in the actual circumstances. A categorical opportunity to exercise the ability otherwise in exactly the relevant circumstances may require that more than one outcome remain possible. Compatibilists and incompatibilists frequently diverge over which of these conditions free will demands. Franklin’s later work places special emphasis on opportunities to exercise reflective self-control in more than one way, rather than merely possessing a general capacity that might have manifested in altered circumstances (Franklin, 2018, ch. 3).

This distinction also prevents a common inference error. From “the agent could have chosen otherwise” in one modal sense, it does not follow that the relevant causal process was indeterministic; and from “the process was indeterministic,” it does not follow that the agent had an intelligible power to choose otherwise. A libertarian account must bridge both gaps. Its philosophical success depends on explaining how alternative outcomes are connected to the same agency rather than to merely different possible happenings.

Ancient Foundations: Choice, Fate, and Responsibility

The ancient material establishes a genealogy of the problem without supplying one uniform ancient “libertarian” doctrine. Aristotle’s Nicomachean Ethics, Book III, distinguishes voluntary and involuntary action, analyzes compulsion and ignorance, and treats prohairesis (προαίρεσις), deliberative choice, as especially revealing of character. In III.2–3 (1111b4–1113a14), choice is neither simply appetite nor opinion. It is connected with deliberation about things within our power. Aristotle’s investigation clarifies the agency and reasons dimensions of later debates, but it should not be identified wholesale with the modern thesis that deterministic laws are incompatible with responsibility (Aristotle, Nicomachean Ethics, III.1–5).

This qualification matters because the ancient questions are arranged differently. The agent’s internal origin of action, the epistemic conditions of deliberation, and the formation of character are central to Aristotle’s ethics. Whether the same antecedent microphysical conditions could issue in distinct actions is a later style of question. Even when Aristotle says that a principle of motion lies in the agent, the interpretive task is to determine whether he asserts a libertarian metaphysical alternative or the practical authorship appropriate to his account of virtue. A reconstruction that conflates these themes would erase the reasons for which his discussion remains relevant to both compatibilist and incompatibilist readers.

Epicurean sources provide a sharper confrontation between necessity, chance, and human agency. The Letter to Menoeceus, transmitted by Diogenes Laertius at X.133–134, distinguishes occurrences arising by necessity, by chance, and through ourselves, linking the last with praise and blame. This tripartite distinction anticipates the central libertarian pressure: freedom seems threatened both by all-encompassing necessity and by mere accident. The Epicurean atomic swerve, known through sources including Lucretius and Cicero, is an important ancient response to deterministic physical explanation. Yet the swerve also creates the familiar question whether adding physical indeterminacy genuinely establishes rational authorship.

Stoic discussions of fate, assent, and responsibility introduce a sophisticated contrasting tradition. Their attempts to explain responsible assent within a rational cosmic order complicate the idea that ancient ethical responsibility uniformly depended on indeterminism. Cicero’s De fato records disputes among competing schools; Alexander of Aphrodisias’ On Fate offers an influential Peripatetic criticism of Stoic deterministic tendencies. The surviving ancient archive is already a dispute about distinct relations among causal necessity, what is “up to us,” and the grounds of praise and blame (Cicero, De fato; Alexander, On Fate).

The philosophical result is not a claim of single historical priority. The ancient debate produced several conceptual resources later libertarian theories appropriate: action from an internal principle, rational deliberation, an open future, and the difference between a responsible act and a chance occurrence. Their integration into one contemporary metaphysical theory is a further philosophical achievement, requiring arguments the historical sources do not always supply.

Medieval Accounts of Will and Self-Determination

The Latin tradition develops liberum arbitrium in relation to rational judgment, voluntary choice, moral evil, grace, and divine foreknowledge. Augustine’s De libero arbitrio (On Free Choice of the Will) treats choice as a condition of imputability and connects the misuse of will with moral evil. His positions develop across different works and phases, especially as questions about grace and the human capacity for good become more pressing. An analysis limited to one early passage would misrepresent this intellectual development. The historian must distinguish the capacity to choose, freedom from disorder, and freedom in relation to grace (Augustine, De libero arbitrio, II–III; see the historical analysis in the Stanford Encyclopedia of Philosophy entry on Augustine).

Thomas Aquinas, in Summa theologiae I, q. 83, examines whether human beings possess free choice, whether it is a power, and how intellect and will cooperate. His answer grounds choice in rational judgment about alternatives and in the appetitive power that accepts a practical conclusion. He denies that freedom requires an agent to be an absolutely uncaused first cause. This point is essential: Aquinas argues for human free choice while preserving a metaphysical order of primary and secondary causation. Thus medieval affirmation of free choice cannot be made equivalent, without qualification, to the later claim that the complete natural past and laws must leave an undetermined future.

John Duns Scotus is important for a different reason. Scholarly reconstructions of his theory of will emphasize the contingency of volition and an account of a will capable of alternatives in relation to the same moment. This medieval formulation of synchronic contingency becomes relevant to contemporary accounts of alternative possibilities. Its exact theological and metaphysical setting, however, differs from modern natural-law determinism. It should be examined as an account of voluntariness, contingency, and rational power within scholastic metaphysics, rather than translated directly into twentieth-century philosophy of action.

The medieval dispute establishes a fundamental distinction between choosing through rational judgment and possessing a power whose exercise is not predetermined by earlier conditions. It also shows that “the will is free” may answer several different questions: whether human action is voluntary, whether choice can be contrary to present inclination, whether the agent is morally accountable, and whether some prior causal order fixes the action. Later libertarian theories sharpen the last question, but inherit the earlier attention to reason, orientation toward the good, and the structure of power.

Early Modern Liberty and the Causal Order

Early modern philosophy brings free will into contact with increasingly general accounts of causal necessity. Thomas Hobbes’ discussion of liberty in Leviathan, chapter XXI (1651), defines freedom principally with reference to the absence of external impediments to motion and action. In this approach a voluntary act can be caused and necessitated while the agent remains free to do what the agent wills. The analysis clarifies a leading compatibilist possibility: freedom of action may concern the successful execution of willing rather than the metaphysical origination of willing.

David Hume’s Enquiry Concerning Human Understanding, section VIII (1748), connects liberty and necessity through regularities of conduct and the dependence of action on motives, character, and circumstances. His account is significant because it makes causal intelligibility part of ordinary moral judgment. If actions were wholly disconnected from character and motives, praise and blame would lose their normal objects. The libertarian must therefore distinguish the rejection of deterministic necessitation from the rejection of intelligible causal and motivational explanation. Hume’s challenge persists whenever indeterminism is introduced without an account of why the decision is attributable to the agent.

Thomas Reid’s Essays on the Active Powers of Man (1788), especially Essay IV, offers a countertradition. Reid articulates a conception of moral liberty through the agent’s power over determinations of will, and refuses to collapse active power into necessitating relations among prior events. His distinction between motives and efficient causes supplies an important historical precursor to agent-causal libertarian theories. On this view, to cite a motive is to make action intelligible, while the bearer of active power is the person whose agency brings about the choice. The difficult issue is whether an appeal to this power explains the production of a particular choice or merely names the locus of responsibility.

Immanuel Kant complicates the usual categories. In the third section of the Groundwork of the Metaphysics of Morals (1785), he connects freedom, autonomy, and the intelligible standpoint of rational agency; his Critique of Practical Reason (1788) treats freedom in a distinctive relation to moral law. Kant does not offer a straightforward empirical hypothesis about stochastic causes in the brain. His transcendental account raises the question whether practical freedom and causal explanation belong to explanatory standpoints whose relation cannot be settled by importing the assumptions of physical determinism into the practical standpoint without argument. It influences later debates about autonomy while resisting simple assimilation to event-causal libertarianism.

The early modern inheritance is thus plural. Hobbes and Hume show how agency might be free despite necessity. Reid emphasizes originating power. Kant asks what freedom rational self-legislation requires. Contemporary libertarianism emerges in a field already containing these distinct definitions of freedom, and it must confront the best form of the compatibilist alternative rather than only the image of a passive person pushed mechanically by external forces.

The Modern Incompatibilist Argument

Peter van Inwagen’s An Essay on Free Will (1983) crystallizes the consequence argument. Let (P) represent the remote past, (L) the laws of nature, and (A) a present action. If determinism is true, then (P) together with (L) entails (A). If agents have no power over the remote past or the laws, and if the relevant inability transfers across the entailment, agents cannot determine a different outcome while those conditions remain fixed. The argument has force because it presents a control problem: the actual action would be a consequence of circumstances outside the agent’s control.

One schematic reconstruction is:

1. Under determinism, the complete past and the laws entail every subsequent action.
2. A present agent cannot now change the remote past or the laws.
3. If a consequence follows unavoidably from conditions entirely beyond the agent’s control, the agent lacks the categorical power to prevent that consequence.
4. Therefore, under determinism, no present agent has that categorical power concerning an action.
5. If the freedom at issue requires that power, determinism excludes it.

The third and fifth premises are philosophically contested. Compatibilists may reject the proposed transfer principle, challenge the understanding of an ability to do otherwise, or deny that categorical alternatives are necessary for responsible control. Some accounts interpret “could have done otherwise” counterfactually through what the agent would have done if reasons, intentions, or decisions had differed. Incompatibilists reply that altering such conditions may abandon the requirement to keep the actual past and laws fixed. The dispute therefore hinges on what ability and opportunity must mean for a free action (van Inwagen, 1983; Franklin, 2018).

The consequence argument is an argument against compatibility, not a demonstration that libertarian agency exists. Even granting all its premises leaves unanswered whether indeterminism obtains in the relevant decisions, whether the agent controls the outcome, and whether the resulting kind of control supports responsibility. The transition from a negative argument about determinism to a positive ontology of free choice is precisely where libertarian theories diverge.

Alternative Possibilities and Frankfurt Cases

A traditional principle of alternative possibilities says that an agent is morally responsible for an action only if the agent could have done otherwise. Harry Frankfurt’s 1969 challenge separates the availability of alternatives from responsibility in the actual sequence. In a standard Frankfurt-style case, a counterfactual intervener is prepared to ensure one decision if necessary, but the agent actually makes that decision without intervention. The agent appears responsible although the intervener would have prevented a different choice. The example presses on whether the moral significance of agency lies in alternative options or in how the actual decision came about (Frankfurt, 1969).

Libertarians can respond in several ways. They may distinguish freedom from moral responsibility and concede that some responsibility does not require alternatives while maintaining that freedom does. They may argue that genuinely successful Frankfurt cases presuppose deterministic indicators, thereby begging the question against incompatibilism. Or they may identify a remaining alternative—sometimes called a “flicker of freedom”—and ask whether its availability matters morally. Each response has costs. A residual flicker might be too slight to support significant responsibility, while the denial of all possible Frankfurt cases requires an independent argument about the structure of prediction and intervention.

Sourcehood-based libertarians shift attention to whether the agent originates the action even if no alternate outward act is available. Leeway libertarians retain the requirement that alternative exercises of decision-making power remain open. The two emphases can coexist within one theory, but they impose different explanatory burdens. A model proving mere alternative outcomes has not thereby proved appropriate sourcehood; a model explaining sourcehood has not automatically vindicated a robust opportunity to act otherwise.

The deeper philosophical point is that “could have done otherwise” contains several questions. Was another bodily movement possible? Was another intention possible? Could the agent exercise the same rational capacities differently? Was there an accessible alternative while the actual past and laws stayed fixed? Could a morally significant different response have been made? The answers may diverge. A comprehensive libertarian account must state which alternative matters and why that alternative has a constitutive role in freedom or moral responsibility.

Sourcehood, Origination, and Ultimate Responsibility

Sourcehood concerns the ownership of the process producing action. An agent can have multiple options without being their author: the outcome of a lottery is open but is not settled by a participant’s free choice. Conversely, an intentional, reason-responsive performance can be strongly attributable to an agent even where its history is causally determined. Libertarian sourcehood adds the requirement that the agent be an ultimate or nonderivatively relevant source, not merely a stage in a causal sequence fixed by conditions beyond the agent’s control.

Robert Kane argues that ultimate responsibility requires more than satisfying desires determined by past character. An agent must be responsible, at least in suitable formative episodes, for the sources from which later choices arise. Kane’s self-forming actions are meant to answer the otherwise threatening regress: if a person acts from character, and character was fixed by an unchosen past, the mere conformity of action to character may be insufficient for ultimate responsibility (Kane, 1996, ch. 5).

The difficulty can be developed explicitly. If responsibility for choice (C) requires responsibility for every prior condition influencing (C), the explanation appears to retreat without end. No finite agent seems to be responsible for the entire origin of its existence, temperament, social circumstances, and early capacities. A libertarian may answer that ultimate responsibility does not require responsibility for every cause; it requires the occurrence of pivotal actions in which the agent’s active control makes a genuine difference to the direction of character. The explanatory demand then moves from an impossible total self-creation to a more limited and testable account of self-forming choice.

Yet the more modest answer still needs a grounding condition. A formative decision must already be attributable to the agent, or it cannot confer responsibility on later character. If it is attributable simply because it arises from present reasons and capacities, a compatibilist can ask why those same features would cease to ground responsibility if causally determined. If it requires the exercise of an irreducible power, the libertarian must explain that power. Thus the debate about ultimate responsibility remains intertwined with the fundamental ontology of agency.

Noncausal Libertarian Theories

Noncausal theories maintain that a free action or a crucial act of deciding need not be caused by prior events or by a further substance-causal relation. Carl Ginet and Hugh McCann are prominent modern advocates of versions of noncausal freedom. Their motivation includes the thought that an act of willing must sometimes be an exercise of the person’s active capacity rather than the effect of antecedent conditions. Such views often place significant weight on the intrinsic act-character of deciding, intending, or trying (Ginet, 1990; McCann, 1998; Clarke, 2003).

This approach avoids one form of the deterministic worry: a free decision is not simply necessitated by earlier motives and states. It also resists the demand to introduce a special form of substance causation. Its main problem is to account for rational explanation and ownership. If nothing causes the decision, why does one decision occur rather than another? Merely saying that the action is intentional does not answer the question unless the theory clarifies how intending and acting are connected and how the decision is controlled by the agent.

A noncausal account may allow reasons to explain an act teleologically without causing it. Such explanation states what the agent was acting for. The critical question is whether having and recognizing a reason can explain why the agent acted for that reason on this occasion while remaining causally irrelevant to its production. The theory needs to distinguish a justified retrospective description from an explanatory relation that makes the action intelligible as the agent’s doing.

Noncausal libertarianism deserves treatment as a substantive alternative, not as an incoherent absence of explanation by definition. Its success turns on an account of agential exercise that makes reasons, intentions, and the production of choices mutually intelligible. The challenge is acute because a brute uncaused event and an uncaused act of agency cannot be distinguished merely by observing that neither had a prior sufficient cause.

Event-Causal Libertarian Theories

Event-causal libertarianism keeps the ordinary causal explanatory architecture of actions but rejects determination at a decisive point. Mental events such as apprehending reasons, desiring ends, and deliberating can causally contribute to a choice without necessitating one outcome. This theory has the advantage of allowing reasons to do explanatory work within a familiar model of action production. Its central problem is to specify exactly how an undetermined decision is nevertheless made by the agent rather than merely occurring as a chance outcome of competing causal propensities (Clarke, 2003; Mele, 2006).

There are several possible locations for indeterminism. It may arise in the formation or weighting of preferences, in the deliberative process that attends to reasons, in an effort of will, or immediately before a decisive intention. The location matters. If chance enters after a fully settled intention, it appears to interfere with the execution of a decision rather than deepen freedom. If chance enters before the agent deliberates, it may affect the materials of deliberation but leave the choice itself deterministic. Event-causal theories try to place openness where it can matter to an exercise of control.

One can express the minimum architecture without presupposing its success. The agent has relevant reasons (R_1) and (R_2) and capacities (K). Under the actual circumstances, these factors permit but do not necessitate either decision (D_1) or decision (D_2). Each decision is a possible outcome of a process governed by the agent’s practical reasoning. The positive libertarian claim adds that the agent’s exercise of capacities (K) genuinely governs which decision becomes actual. Without the last claim, the model describes a nondeterministic selection process without proving self-determination.

A sophisticated event-causal libertarian need not treat all random events as exercises of freedom. The proper comparison is between a random occurrence unrelated to a person’s deliberative organization and a nondeterministic decision issuing from a sustained, rationally structured capacity. Whether that structure provides greater control than a deterministic but equally reason-responsive process remains a separate question. This distinction defines much of the modern dispute over the value of libertarian freedom.

Kane and Self-Forming Actions

Kane’s account locates significant indeterminism in difficult choices involving competing motivations in which the agent makes efforts of will. Such actions are “self-forming” because the agent’s choice helps shape the commitments and character through which future decisions are made. Most ordinary free actions need not involve this formative struggle; their freedom can derive from a history in which some earlier self-forming choices supplied the agent’s own direction. The theory thus distributes freedom temporally rather than requiring every action to arise from a freshly undetermined event (Kane, 1996, chs. 4–5).

Consider a person torn between an immediate advantage and a costly but sincerely recognized moral obligation. Both alternatives may be supported by real motives; the person attempts to resolve the conflict and to act for the chosen reasons. In Kane’s account, the undetermined result is connected to the agent’s effort under tension. This is a stronger picture than a coin flip inserted at the end of deliberation: the agent has been working toward each relevant possible outcome, and either resolution could count as an intentional one rather than as an extraneous accident.

The account must still answer why this concrete resolution, rather than its alternate, was sufficiently under the agent’s control. An appeal to prior efforts can help if those efforts are intentional, rationally directed, and genuinely theirs. Yet if the efforts do not causally determine which action succeeds, the difference between the two possible outcomes remains unsettled. Calling both outcomes voluntary does not demonstrate that the particular actual outcome was controlled to the degree required for ultimate responsibility.

A further tension concerns retrospective endorsement. Agents often accept a decision after making it. That later acceptance can integrate an event into their developing practical identities, but it cannot by itself create antecedent control of the original decision. Similarly, the fact that a person later acts consistently with a self-forming choice demonstrates its influence on character, not automatically the freedom of the choice at the moment it occurred. These points form the strongest pressure on Kane’s attempt to connect formative indeterminism and historical ownership.

Kane’s enduring contribution is to show why a serious libertarian theory must address the process of choosing, the development of character, and the unequal moral significance of different decisions. Even where the account remains disputed, it prevents the debate from collapsing into the unrealistic requirement that every trivial movement contain an independent metaphysical miracle. Its unresolved problem is the exact relation between nondeterministic effort and sufficient agential settlement.

Agent-Causal Libertarian Theories

Agent-causal theories distinguish an agent as an enduring bearer of powers from the events and states associated with that agent. A free choice, or a constitutive component of the choice, is caused by the agent in a manner that is not exhausted by prior events causing later events. Reid’s account of active power supplies a historical resource; Roderick Chisholm’s twentieth-century papers develop a modern version; Timothy O’Connor provides a sustained contemporary metaphysical model (Reid, 1788, Essay IV; Chisholm, 1966; O’Connor, 2000).

The attraction is direct. If the problem is that an event caused indeterministically by prior events seems to be settled by chance, then the theory proposes a different source of settlement: the person. A free decision occurs because the agent exercises a power to cause an intention or determination, without that exercise itself being necessitated by the prior event history. This seems to supply exactly what the event-causal picture lacks—a metaphysical role for the agent as a genuine originator.

The agent-causal thesis is therefore stronger than saying that an action happened within a person’s body or followed from that person’s reasons. It asserts a distinctive explanatory relation. O’Connor’s account analyzes free deciding through the agent’s causing the coming-to-be of an intention, and attempts to connect this causation with the agent’s reasons and deliberative capacities. This is intended to avoid describing the agent as an arbitrary first mover disconnected from all practical reasons. Nonetheless, the theory must explain how reasons influence an exercise of active power without determining it (O’Connor, 2000, ch. 5).

Three difficulties are central. First, the causal relation itself requires intelligible metaphysical characterization: what does it mean for a persisting substance to cause a decision if the causation is neither an event-event relation nor a mere summary of ordinary psychological processes? Second, there is an integration problem: how does such causation fit with established explanations of bodily and neural processes? Third, there is a selection problem: given that an agent might cause either of two intentions for available reasons, what determines which agent-causal exercise occurs? Replacing “a chance event happens” with “the agent causes the event” may make agency explicit while leaving the decisive contrast unexplained.

Supporters can deny that every free choice must have a prior contrastive sufficient cause. On that reply, demanding a determining explanation would beg the question against libertarianism. The philosophical dispute then concerns the kind of explanation a rationally guided free action requires. A theory may allow genuine causation without a sufficient explanation of why this outcome obtained. Whether that is an adequate account of control cannot be decided simply by the vocabulary of causes; it depends on how well the proposed power connects action, reasons, and authorship.

Reasons, Deliberation, and Rational Explanation

Every leading libertarian strategy faces a reasons problem. Free action in the morally significant sense is ordinarily done for reasons the agent can recognize, evaluate, and endorse or resist. If reasons necessitate every decision, then a straightforward incompatibilist loses the required openness. If reasons do not bear on the production of the decision, the decision risks becoming unintelligible as action. Libertarian explanation must occupy a more precise middle position: reasons guide action without fully fixing which action will occur (Hume, 1748, §VIII; Clarke, 2003).

The distinction between a motivating reason and a normative reason is useful here. A normative reason counts in favor of an act; a motivating reason explains why this agent acted. The two may differ. An agent may act from a mistaken belief or a morally bad desire, while still acting intelligibly for a reason. A theory of freedom must therefore explain the causal or practical role of the reason actually motivating the action, not simply list considerations objectively favoring one outcome.

Contrastive explanation sharpens the problem. Suppose an agent chooses (A) over (B) when each has serious reasons in its favor. A noncontrastive account can explain the choice of (A) by identifying the agent’s recognition of reasons for (A). A contrastive account asks why (A) rather than (B) occurred in a situation where reasons for both were operative. If the decisive difference is an uncaused or probabilistically fluctuating variable, the account may be explanatorily adequate at one level while still struggling to vindicate robust control.

An adequate theory need not demand the impossible proposition that every intentional choice has a complete deterministic explanation. Explanation has several forms: a reason can make an action intelligible, a causal process can explain its production, a capacity can establish how the agent could act, and a normative analysis can establish why an action merits assessment. The task is to identify which combination justifies the claim that the undetermined choice was the agent’s settlement. The distinction prevents the luck objection from silently importing determinism as the sole standard of explanation, while preserving a legitimate demand that action be more than a fortuitous event.

The Luck, Rollback, and Control Objections

The luck objection is the most persistent internal challenge to libertarianism. Imagine two possible worlds with identical pasts and laws through the moment of choice. In one, the agent decides (A); in the other, (B). If the agent’s character, reasons, capacities, and preceding efforts are exactly alike, what accounts for the difference? If nothing within the agent’s control explains it, the realized decision seems lucky. Repeating the world from the same predecision conditions—the rollback image—makes the challenge vivid: an outcome that varies across repetitions might appear less attributable to the agent than one reliably produced by a controlled mechanism (van Inwagen, 1983, ch. 4; Mele, 2006).

This argument should not be flattened into the claim that every probability is necessarily luck. A practiced player can intentionally attempt a difficult shot, succeed in some repetitions, and fail in others. The outcomes vary, but the performance is genuinely the player’s exercise of skill. Libertarians can borrow this distinction: probabilistic success may coexist with control. The skeptical response is that the relevant issue is the control of the choice itself, not merely an outcome following a completed intention. The analogy must establish that the agent’s exercise settles a decision, not only that a decision exposes the agent to risk of success or failure.

The objection also has a comparative component. Suppose a deterministic and an indeterministic decision process possess identical reasons-responsiveness, self-understanding, and resistance to coercion. What additional positive power does the indeterministic process confer? If the answer is simply that either decision could occur, this may secure openness but not enhanced authorship. Franklin’s minimal libertarianism responds by combining a developed control structure with indeterminism at specified junctures, rather than insisting that openness by itself creates control (Franklin, 2018).

Agent-causal theories offer a different reply: the person, rather than chance, settles which intention is caused. Yet if the person has identical powers and reasons in either possible world, a contrastive question remains. The absence of a determining prior event is essential to the theory, so the theory must explain how an undetermined exercise of agent causation remains an intentional application of power. Merely inserting a substantial agent at the final stage does not fully answer this challenge.

The strongest formulation of the luck objection distinguishes an ontological question from an epistemic one. It may be metaphysically true that an agent exercises control even when no deterministic antecedent fixes an outcome. Our difficulty in predicting or explaining why that particular outcome occurred does not logically entail the absence of control. Still, an account that cannot identify the character of the power or its conditions of successful exercise does not yet give a sufficient positive theory of liberty. The objection is a demand for explanatory articulation, not a proof that indeterminism is incoherent.

Manipulation, Character, and the History of Agency

Manipulation arguments ask whether a person can be responsible when desires, commitments, or decision mechanisms have been deliberately shaped by another agent. The case is particularly troubling when the manipulated person deliberates coherently and acts according to internal reasons. If manipulation can undermine responsibility despite apparent reasons-responsiveness, then merely identifying internal psychological causes may not establish adequate sourcehood. Incompatibilists often deploy this point against compatibilist theories; hard incompatibilists extend it to challenge the difference between manipulated and naturally determined agents (Pereboom, 2001).

Libertarianism may answer that a causally open formative choice supplies the person with a history of genuine self-determination. But this reply has conditions. The agent must possess adequate understanding, practical competence, and opportunity at the formative moment, and the openness must be exercised rather than merely encountered. A history containing one undetermined event will not suffice if that event was forced, manipulated, or disconnected from rational control. An appeal to earlier free choices therefore has explanatory force only when the earlier choices independently satisfy the theory’s freedom requirements.

The relation between character and choice is also reciprocal. Character influences choices; repeated choices shape character; institutions and social environments affect both. The existence of influence is not itself the existence of coercion. A workable libertarian account cannot require independence from all inherited dispositions, language, education, and interpersonal formation. Such independence would remove much of the rational material through which agents exercise capacities. The more plausible target is the agent’s ability to use, assess, and sometimes transform those influences in an appropriately self-governed manner.

This observation exposes a broader vulnerability in radical self-creation requirements. If responsibility demanded that the agent create the entirety of its earlier self, no finite agent could meet the standard. The question is whether the libertarian can define a limited but sufficient form of originating control. It must be more than mere historical causal continuity, yet less than impossible authorship of every condition of existence.

Libertarianism and Moral Responsibility

Libertarian theories are often motivated by moral responsibility, but the relationship is complex. Responsibility can mean causal attribution, answerability for reasons, accountability within interpersonal practices, the appropriateness of praise and blame, or the basic desert of suffering or benefit independently of beneficial consequences. These are not equivalent. A theory may plausibly ground the first two while disputing that the agent deserves punishment or suffering in a strong retributive sense.

Frankfurt’s challenge to alternative possibilities and P. F. Strawson’s analysis of reactive attitudes show why the moral significance of free will cannot be reduced to a simple modal test. Resentment, gratitude, forgiveness, demands for justification, and recognition of an agent’s quality of will arise in relationships structured by norms and expectations. The libertarian can regard deeper sourcehood as a condition for some forms of desert while agreeing that ordinary moral practices also depend on rational competence, mutual regard, and social roles (Frankfurt, 1969; P. F. Strawson, 1962).

Hard incompatibilists such as Derk Pereboom argue that neither compatibilist nor libertarian models establish the freedom required for basic-desert responsibility. Pereboom also contends that rejecting such desert need not abolish ethical concern or practical regulation of behavior. This is a significant rival position: the libertarian must show why its distinctive form of control is present, rather than infer that it must exist because ethics would otherwise collapse. Moral relationships may remain possible under substantially different metaphysical accounts (Pereboom, 2001, 2009).

A mature libertarian account should thus separate at least three claims. First, agents sometimes have causally open and reason-guided powers. Second, exercising these powers can make actions attributable to them in a special originating sense. Third, that special attribution justifies particular moral attitudes or institutional consequences. Each transition requires an argument. Even a successful metaphysics of free choice would not automatically determine the just design of punishment, forgiveness, or repair. That normative conclusion needs its own account of proportionality, standing, harm, and opportunity.

The analysis leaves an important possibility open: different forms of moral responsibility may require different degrees or forms of freedom. The philosophical task is to avoid smuggling the strongest conception of desert into every ordinary judgment that a person acted, can explain an action, or should respond to a request for reasons.

The Scientific and Neuroscientific Evidence

Neuroscience investigates mechanisms, temporal patterns, and correlates of decision-making; it does not directly read off a philosophical definition of ultimate sourcehood from a brain measurement. Benjamin Libet and colleagues’ 1983 experiments compared the timing of readiness potentials before simple voluntary movements with participants’ reported times of conscious intention. Brain activity preceding a reported intention has stimulated claims about the unconscious preparation of action. But the design involved simple, low-stakes movements and depends on contested assumptions about the timing and interpretation of intention (Libet et al., 1983).

Schurger, Sitt, and Dehaene (2012) proposed an accumulator interpretation of the readiness potential in which time-locking spontaneous neural fluctuations to movement onset yields an apparent buildup in the average signal. This model changes how the earlier results should be interpreted: an early averaged signal need not identify a fully formed unconscious decision already made prior to awareness. Neither the original observation nor the accumulator account establishes metaphysical libertarianism; together they show that a single neural precursor cannot bear the entire philosophical weight of the free-will debate.

The more general scientific problem has two sides. Evidence that neural activity causally influences decisions is compatible with many libertarian theories, especially event-causal approaches, because libertarianism does not require action to be detached from physical processes. Conversely, evidence for stochasticity does not establish agent control: randomness in a brain process is still randomness unless an adequate theory explains its relation to intentional agency. Experimental findings concern mechanisms, timescales, and behavior; the inference to categorical alternative possibilities, agent causation, or basic moral desert needs added premises.

The scientific agenda can be more specific. Research may ask where deliberative flexibility occurs, how competing reasons influence decisions, how self-regulation develops, how cognitive constraints alter choice, or how neural interventions change agency. These are relevant constraints on a theory of free will. They do not replace the task of specifying what freedom means. A productive relationship between philosophy and science proceeds by distinguishing explanatory levels and then identifying what observational findings would support or undermine the competing models.

Current Research and the Limits of the Major Models

Recent work has deepened rather than dissolved the divisions among libertarian theories. Clarke’s systematic comparison evaluates noncausal, event-causal, and agent-causal approaches against problems of randomness, explanation, and control. His assessment is important because it separates local replies to objections from the global adequacy of a theory. A model may withstand a particular luck objection and nevertheless lack an account of sourcehood sufficient for the freedom it claims to explain (Clarke, 2003).

Mele investigates freedom, ability, decision, and luck without assuming that every philosophical difficulty is exclusive to indeterminism. This approach avoids granting deterministic theories an automatic victory over agency merely because the rival theories confront luck. Philosophical control comes in degrees and kinds; explaining which capacities are relevant to a specified freedom claim remains necessary on either side of the compatibility debate (Mele, 2006).

Franklin’s A Minimal Libertarianism (2018) is especially instructive. It proposes that an event-causal libertarian may begin with an otherwise plausible compatibilist architecture and require indeterminism at appropriate moments of self-control. The result is conceptually economical. It also makes the dispute more exact: the challenge is whether a relatively modest change to the causal structure secures a kind of opportunity or control unavailable under determinism. Franklin explicitly identifies a further concern about the reduction of agency to mental states and events. This concern applies beyond libertarianism, because compatibilist event-causal theories also depend on understanding the relation between an agent and the events that constitute acting.

The Stanford Encyclopedia’s discussions of free will and nondeterministic theories document ongoing disagreement over the location of indeterminism, the role of alternatives, and the grounding of sourcehood. That continuing disagreement is intellectually productive. It prevents the simple classification of libertarianism as one theory with one solution and allows different accounts to be judged against precisely specified requirements. The current state of research supports a comparative judgment: no single mechanism of indeterminism has been established as a generally accepted explanation of rational self-determination (O’Connor & Franklin, 2022; Clarke, 2021).

The broader significance extends to philosophy of action, metaphysics, moral psychology, and philosophy of mind. The correct interpretation of a reasons-responsive system may depend on its functional organization, its causal history, and the ontology of the bearer of agency. These intersections motivate future research into the distinction between practical competence and ultimate sourcehood. They do not independently settle whether humans, or other possible agents, possess libertarian freedom.

A Comparative Assessment of Libertarian Strategies

The three major libertarian families can be compared by asking the same questions of each. What does the agent do? What causes the free choice? Where is indeterminism located? How do reasons guide it? What distinguishes it from chance? What type of responsibility is supposed to follow?

Noncausal theories emphasize the act-character of deciding and resist defining agency through antecedent efficient causation. Their achievement is to preserve the idea of a direct exercise of choice. Their outstanding task is to explain how rationally informed agency differs from an uncaused occurrence in the absence of causal production. An account may appeal to intrinsic intentionality, but the relation between that intentionality and the choice of one alternative over another needs independent analysis.

Event-causal theories integrate freedom with ordinary causal explanation and can specify experimentally and psychologically plausible processes of deliberation. Their achievement is continuity with the framework in which reasons, intentions, desires, and capacities already explain action. Their outstanding task is to establish that adding indeterminism at a chosen point contributes appropriate control rather than a merely additional route for chance to influence the outcome.

Agent-causal theories assign an irreducible role to the agent as a bearer of active power. Their achievement is a direct answer to the demand for origination. Their outstanding task is a satisfactory metaphysics of agent causation and an account of why the agent’s undetermined exercise of power is a reason-governed settlement. The fact that a cause is called an agent does not by itself supply the contrastive explanation the theory needs.

All three face the historical-ownership problem. Even if a choice is genuinely open and rationally guided, the agent’s reasons and powers arise within a history not wholly of its making. The theorist must explain how limited self-formation is sufficient for responsibility. All three also face a justificatory question: why should the precise metaphysical freedom each supplies be regarded as necessary for a particular moral response? A solution to the metaphysics does not substitute for a normative argument.

One can therefore identify a shared explanatory demand without erasing substantive disagreement. Libertarianism must vindicate an exercise of reason-guided, agent-attributable control over a causally open choice. The forms of causation differ across theories. The underlying demand for intelligible agency remains constant. This is the point at which comparison becomes a philosophically constructive method rather than a list of rival labels.

Open Questions and Philosophical Synthesis

Several questions remain open in the contemporary debate. First, is there a coherent metaphysical account of an agent’s power that is neither a disguised deterministic mechanism nor a bare stochastic selection? Second, what kind of explanation of this choice rather than its alternative must a free agent provide? Third, can reasons be genuinely action-guiding without either necessitating the action or remaining causally ineffective? Fourth, how much formative control must agents possess to ground responsibility for later character-shaped choices? Fifth, is the freedom required for basic desert stronger than the freedom required for rational accountability?

A further question concerns the evidential relation between theory and life. People often experience deliberation as open, but the phenomenology of choosing does not directly establish the metaphysical structure of possible worlds. Scientific evidence may reveal the neural organization and limitations of decision-making, yet an account of readiness potentials does not settle the transfer principle of the consequence argument or the ontology of agent causation. The existence question therefore requires a coordinated philosophical and empirical program with explicit bridging premises.

The synthesis of the historical and analytical material yields five conclusions. First, libertarian freedom is a positive theory of control and cannot be defined adequately by indeterminism alone. Second, its historical antecedents supply multiple dimensions of agency—voluntariness, deliberation, active power, and contingency—that later theories organize in distinct ways. Third, the modern incompatibilist argument concerns the dependence of choice on a fixed past and laws; it does not independently establish a positive account of freedom. Fourth, the luck problem becomes philosophically decisive wherever a theory fails to distinguish openness of outcome from the agent’s rational settlement of outcome. Fifth, responsibility has several normative senses, so the ethical value of libertarian control must be justified with respect to the particular responsibility claim being made.

The most promising direction of inquiry is consequently a precise theory of the exercise of agential powers: an account specifying how deliberation, attention, intention, reasons, and active capacities jointly constitute the agent’s control in a situation where the outcome remains genuinely open. Its success would depend on explaining the particular role of indeterminism rather than placing indeterminism at the center of a picture whose account of agency remains unchanged. Where such an account cannot be supplied, a libertarian may still establish that deterministic antecedents seem inconsistent with a desired freedom, but will not yet establish that a causally open choice realizes it.

This result does not license a premature verdict that every form of libertarianism is false. It identifies the explanatory conditions under which a theory becomes philosophically adequate. By distinguishing necessity, possibility, sourcehood, and responsibility, the debate can progress beyond the recurring choice between mechanical compulsion and pure accident.

Conclusion

Libertarian free will is a coherent research program organized around the claim that genuine self-determination requires an open causal future and that some agents sometimes exercise the relevant freedom. Its ancient and medieval precursors developed the vocabulary of choice, fate, rational power, and responsibility; early modern philosophy intensified the conflict between causal necessity and originating agency; contemporary theories made the nature and location of indeterminism a precise issue in the metaphysics of action.

The arguments surveyed show why libertarianism is attractive and why its central explanatory problem remains difficult. The consequence argument supplies reasons to doubt categorical freedom under determinism. Noncausal, event-causal, and agent-causal theories offer different accounts of how undetermined decisions might nevertheless be the agent’s own. Kane’s formative account and Franklin’s minimal model refine the relation of openness to control; O’Connor’s agent-causal metaphysics emphasizes the originating bearer of power. The luck, contrastive explanation, and responsibility objections demand that each model explain the actual exercise of freedom rather than merely postulate possible alternatives.

The fundamental result is that a libertarian account must establish a controlled transition from reasons and capacities to a particular undetermined decision. Causal openness supplies the space for such a transition; its agential character requires further philosophical explanation. Moral appraisal then requires a distinct argument about the kind of responsibility that transition supports. These conclusions define the durable problem at the center of libertarian free will and provide an explicit research standard for evaluating future metaphysical and empirical proposals.