Freedom and liberty: negative, positive, and free will

Intuition [Beginner]

Freedom is one of those words that everyone uses and almost nobody defines the same way twice. When someone says "I should be free to do what I want," they might mean that nobody should stop them. Or they might mean that they should have the real capacity — the resources, the education, the genuine opportunity — to pursue a worthwhile life. These are not the same claim, and the difference between them has shaped political philosophy for four hundred years.

In 1958, Isaiah Berlin gave an inaugural lecture at Oxford called "Two Concepts of Liberty" ^{[Berlin 1958]}. His aim was not to settle what freedom is. He did not believe there was a single correct answer. His aim was to show that two fundamentally different families of meaning were at work in modern political thought, and that confusing them had real consequences for how societies organise themselves.

The first family Berlin called negative liberty. This is freedom from interference. You are negatively free to the extent that no other person or institution prevents you from doing what you are otherwise able to do. If the government does not censor your speech, you have negative liberty of expression. If the police cannot detain you without cause, you have negative liberty of movement. The emphasis is on absence: negative liberty is the space within which a person can act unobstructed by others. Thomas Hobbes gave the classic statement in 1651: "Liberty is the absence of all the impediments to action that are not contained in the nature and intrinsical quality of the agent" ^{[Hobbes 1651]}.

The second family Berlin called positive liberty. This is freedom to realise yourself, to be your own master, to live in accordance with your authentic desires or your rational nature rather than under the control of external forces or internal compulsions. Positive liberty is not about the absence of obstacles but about the presence of self-direction. A person trapped in addiction might have enormous negative liberty — nobody is physically stopping them from doing anything — while having very little positive liberty, because they are not genuinely in control of their own choices.

The distinction matters because different political arrangements optimise for different kinds of freedom. A minimalist state that does not interfere with its citizens maximises negative liberty but may do nothing to help people who lack the education, health, or resources to act on their nominal freedoms. A state that provides those resources may increase positive liberty while restricting negative liberty through taxation, regulation, and paternalistic laws. The debate between these visions is not a technical dispute about definitions. It is a substantive disagreement about what a good society looks like.

There is a separate but related question that runs alongside the political one: do human beings have free will at all? If every event in the physical world is determined by prior causes, and human brains are physical objects, then every decision you make was already fixed by the state of the universe before you were born. This is the problem of determinism, and it threatens both negative and positive liberty at the root. If your choices are not genuinely yours, what does it mean to be free?

The free-will debate has three main positions. Hard determinists accept determinism and conclude that free will is an illusion. Libertarians (in the metaphysical, not political, sense) reject determinism and hold that agents have a genuine capacity to choose among alternatives. Compatibilists argue that free will and determinism are compatible — that freedom does not require the absence of causal determination, only the absence of coercion and compulsion.

These three positions do not map neatly onto the negative/positive liberty distinction, but they interact with it. If you are a compatibilist, you can be a negative-liberty theorist without worrying about determinism: freedom just means non-interference, and determinism does not imply interference. If you are a libertarian about free will, you might be drawn to positive-liberty theories that emphasise genuine self-authorship. The philosophical terrain is layered, and each layer constrains the others.

Visual [Beginner]

Picture two scenes. On the left, a person stands in a vast open field with no fences, no guards, no barriers in any direction. They can walk wherever they want. Nobody will stop them. This is negative liberty: the absence of external impediments.

On the right, the same person stands at the base of a mountain they have chosen to climb. They have trained for months, acquired the right equipment, studied the route. They are climbing because they decided to, not because anyone pushed them. The climb is difficult — there are obstacles everywhere — but each obstacle is one they chose to face. This is positive liberty: the presence of self-direction, even amid difficulty.

The gap between the two pictures is where most political disagreement about freedom lives. Expanding the open field does nothing to help someone who lacks the capacity to walk across it. Giving someone a mountain to climb means nothing if they are fenced in at the base. Both pictures capture something real about freedom, and neither is complete without the other.

Worked example [Beginner]

Consider a concrete case. A government passes a law requiring all citizens to vote in national elections. Failure to vote results in a fine.

From the perspective of negative liberty, this law reduces freedom. Before the law, you could choose to vote or not vote. After the law, you are compelled to vote under threat of penalty. The state has introduced an interference where none existed before. A negative-liberty theorist like Hobbes or the early Berlin would count this as a loss of liberty, full stop.

From the perspective of positive liberty, the analysis is more complicated. Compulsory voting might increase the effective political agency of citizens who would otherwise be ignored by the political system. If the reason you do not vote is that you are poorly educated, disenfranchised, or cynical about a system that has never represented you, then your nominal freedom not to vote is not exercising genuine self-direction — it is reflecting a failure of the system to include you. A compulsory-voting law, combined with civic education and accessible polling, might increase your positive liberty by making you a genuine participant in collective self-governance.

Neither analysis is wrong. They are tracking different things. The negative-liberty analysis tracks the presence or absence of legal compulsion. The positive-liberty analysis tracks whether the person has genuine agency over the conditions of their life. The compulsory-voting case is interesting precisely because the two analyses disagree about whether freedom has increased or decreased.

Berlin's warning was this: once you accept positive liberty as a political ideal, you open the door to the claim that the state can force people to be free. If genuine freedom is self-realisation, and the state knows what genuine self-realisation looks like better than the individual does, then coercion can be rebranded as liberation. Berlin was not arguing against positive liberty as a personal ideal. He was arguing that the historical record of states that claimed to enforce positive liberty — through forced re-education, through the suppression of "false consciousness," through authoritarian paternalism — was sufficiently grim that the negative-liberty tradition deserved priority as a political safeguard.

Check your understanding [Beginner]

Formal definition [Intermediate+]

Berlin's distinction can be made precise by giving each concept a formal structure. Gerald MacCallum proposed in 1967 that all uses of "freedom" share a single triadic form ^{[MacCallum 1967]}: a statement that an agent is free from certain obstacles to do or become certain things. The disagreement between negative and positive liberty theorists is then not about the logical form of freedom-talk but about which slot in the triad deserves emphasis.

Negative liberty. An agent $A$ is negatively free with respect to action $x$ if and only if no other agent intentionally interferes with $A$'s doing $x$. Formally, let $I(B,A,x)$ denote "agent $B$ interferes with agent $A$'s doing $x$." Then:

$$\text{NegFree}(A,x)⟺\neg \exists B(B\ne A\wedge I(B,A,x)).$$

Negative liberty is a relation between an agent, an action, and the set of other agents who might interfere. It is silent about whether $A$ has the capacity to do $x$; it tracks only whether anyone else is blocking $A$ from doing it. Hobbes's definition is the prototype: "a free man is he that in those things which by his strength and wit he is able to do is not hindered to do what he has the will to do" ^{[Hobbes 1651]}.

Positive liberty. An agent $A$ is positively free with respect to goal $g$ if and only if $A$ has the capacity, resources, and genuine self-direction to pursue and achieve $g$. The formalisation is less crisp because "genuine self-direction" is itself contested, but a first approximation:

$$\text{PosFree}(A,g)⟺\text{Capable}(A,g)\wedge \text{SelfDirected}(A,g).$$

The predicate $\text{Capable}(A,g)$ covers the material and psychological resources needed to pursue $g$. The predicate $\text{SelfDirected}(A,g)$ requires that $A$'s pursuit of $g$ reflects $A$'s authentic or rational desires rather than compulsion, manipulation, or adaptive preference formation. Both predicates are philosophically loaded: what counts as "authentic" desire and what level of capacity is required for genuine freedom are exactly the points where positive-liberty theorists disagree among themselves.

Mill's harm principle. John Stuart Mill proposed a specific limit on legitimate interference with negative liberty: "the only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others" ^{[Mill 1859]}. Formally, state coercion of agent $A$'s action $x$ is legitimate only if:

$$\text{LegitimateCoercion}(A,x)⟺\exists B(B\ne A\wedge \text{Harm}(x,B)).$$

The harm principle does not resolve the negative/positive debate; it presupposes a negative-liberty framework and then asks what justifies exceptions. The predicate $\text{Harm}(x,B)$ is itself contested — what counts as harm, whether offense counts, whether structural harm counts — and the boundaries of the principle have been the subject of sustained philosophical argument since 1859.

Compatibilism and incompatibilism. These are positions on the relationship between free will and determinism.

• Determinism ($D$): every event, including every human decision, is the unique causal consequence of prior events together with the laws of nature.
• Free will ($FW$): agents sometimes have the genuine capacity to choose among alternative futures.
• Incompatibilism: $D$ and $FW$ cannot both be true. Incompatibilism splits into:
  • Hard determinism: $D$ is true, therefore $FW$ is false.
  • Libertarianism: $FW$ is true, therefore $D$ is false.
• Compatibilism: $D$ and $FW$ can both be true. Freedom requires not the absence of causal determination but the absence of coercion, manipulation, or compulsion. A free action is one that flows from the agent's own desires and rational deliberation, even if those desires were themselves caused.

The central claim of compatibilism is that the predicate "free" in the free-will debate tracks a different distinction than the predicate "uncaused." An action can be fully caused and yet genuinely free, provided the causal chain runs through the agent's own cognitive architecture in the right way.

Key theorem with proof — the Consequence Argument [Intermediate+]

The most influential formal argument for incompatibilism is Peter van Inwagen's Consequence Argument ^{[van Inwagen 1983]}. It aims to show that if determinism is true, no one has any choice about anything — including their own actions. The argument is reconstructed here with explicit premises and a formal inference rule.

Definitions.

• $\square p$ ("necessarily $p$"): $p$ is true and no agent has any choice about whether $p$ is true.
• $N(p)$: no agent has any choice about the fact that $p$.

The operator $N$ is the key notion. $N(p)$ is true when $p$ is fixed — by the past, by the laws of nature, or by any combination of factors beyond the agent's control.

Premises.

P1 (Rule $\alpha$ — transfer of powerlessness). If $N(p)$ and $N(p\to q)$, then $N(q)$. That is: if no one has any choice about $p$, and no one has any choice about the fact that $p$ implies $q$, then no one has any choice about $q$.

P2 (Fixity of the past). If $p$ is a proposition about the state of the world at some past time $t_0$, then $N(p)$. No one now has any choice about what has already happened.

P3 (Fixity of the laws). If $L$ is a proposition expressing the laws of nature, then $N(L)$. No one has any choice about the laws of nature.

P4 (Determinism). The conjunction of the past state of the world at $t_0$ and the laws of nature logically entails every true proposition about the future. Formally, $P_0\wedge L\to F$, where $P_0$ is the past state and $F$ is any future-tense truth.

Argument.

1. $N(P_0)$ — by P2, since $P_0$ is a proposition about the past.

2. $N(L)$ — by P3.

3. $N(P_0\wedge L)$ — from 1 and 2: if no one has a choice about $P_0$ and no one has a choice about $L$, then no one has a choice about their conjunction. (This follows from a natural extension of rule $\alpha$.)

4. $P_0\wedge L\to F$ — by P4 (determinism).

5. $N(P_0\wedge L\to F)$ — from 4: if $P_0\wedge L$ logically entails $F$, then the conditional $P_0\wedge L\to F$ is a logical truth, and no one has any choice about logical truths.

6. $N(F)$ — from 3 and 5 by rule $\alpha$ (P1): no one has a choice about $P_0\wedge L$, no one has a choice about the fact that $P_0\wedge L$ implies $F$, so no one has a choice about $F$.

Conclusion. If determinism is true, then for every true proposition $F$ about the future — including every proposition of the form "agent $A$ performs action $x$" — $N(F)$. No one has any choice about anything that happens.

Why this matters. The consequence argument does not prove that determinism is false or that free will is an illusion. What it proves is that if you accept four premises — transfer of powerlessness, fixity of the past, fixity of the laws, and determinism — then free will is excluded. Each premise can be challenged:

• Rule $\alpha$ (transfer) is the most contested premise. Compatibilists including David Lewis have argued that the transfer rule is invalid in certain contexts, particularly when the relevant conditionals involve counterfactuals about what an agent would do. If the transfer rule fails, the argument does not go through.

• Fixity of the past might be challenged by theories that allow backwards causation, though these are not mainstream in physics.

• Fixity of the laws can be challenged by views on which the laws of nature are not "governing" but merely descriptive regularities (the Humean view of laws).

• Determinism itself can be rejected. Quantum mechanics is often cited as a source of indeterminism, though whether quantum indeterminacy is relevant to the kind of free will that matters is itself disputed.

The consequence argument structures the entire contemporary debate. Compatibilists must deny at least one premise. Incompatibilists accept all four and draw the conclusion. The substantive work is in deciding which premise to give up and what the cost of doing so is.

Exercises [Intermediate+]

Negative and positive liberty: the full argument [Master]

Berlin's 1958 essay does not merely define two concepts and move on. It constructs an argument about the political dangers of privileging positive liberty over negative liberty, and this argument has been both influential and contested for nearly seventy years.

The argument proceeds in three stages. First, Berlin distinguishes the two concepts. Second, he traces the historical tendency of positive-liberty theories to slide from a claim about individual self-realisation into a claim about collective self-realisation — from "I should be free to realise my true nature" to "the state, as the embodiment of collective rationality, has the authority to determine what your true nature is and force you to realise it." Third, he argues that this slide is not accidental but structural: the logic of positive liberty, once adopted as a political principle, contains an inherent tendency toward authoritarianism.

The second stage is the load-bearing one. Berlin's evidence is historical. He cites the trajectory from Rousseau's general will — "whoever refuses to obey the general will shall be compelled to do so by the whole body. This means nothing less than that he will be forced to be free" — through Hegel's identification of freedom with rational necessity, to the 20th-century totalitarianisms that claimed to realise true human freedom through state direction. The rhetorical pattern is consistent: freedom is redefined as conformity to the agent's "true" or "rational" self, and the state claims privileged access to what that true self wants.

Critics have responded on several fronts. MacCallum's 1967 paper ^{[MacCallum 1967]} argues that Berlin's dichotomy is false: all freedom claims are triadic (agent, obstacle, goal), and the negative/positive distinction collapses into a disagreement about which obstacles and which goals matter. Charles Taylor argues that Berlin's negative liberty is not as neutral as it pretends — any practical application requires distinguishing "important" from "minor" freedoms, which already invokes a substantive theory of the good life, which is what positive liberty is about.

Philip Pettit's republican freedom offers a third alternative ^{[Pettit 1997]}. Pettit argues that freedom is neither non-interference (negative liberty) nor self-realisation (positive liberty) but non-domination: the absence of arbitrary power over you. A slave with a kind master who never interferes has negative liberty but is not free in the republican sense, because the master could interfere at will. The slave lives in subjection, even in the absence of actual interference. Republican freedom tracks a structural condition — the absence of vulnerability to arbitrary power — that neither negative nor positive liberty captures cleanly.

Amartya Sen's capabilities approach ^{[Sen 1999]} reframes positive liberty in terms of functionings and capabilities. A functioning is a "being or doing" — being well-nourished, being literate, participating in community life. A capability is the real opportunity to achieve that functioning. Freedom, on Sen's account, is the expansion of capabilities: the range of genuine life-options available to a person. This framework has been widely adopted in development economics (the Human Development Index is built on it) and offers a precise way to talk about positive liberty without Berlin's feared slide into authoritarianism, because capabilities are defined by the range of options, not by the state dictating which option the agent must choose.

The contemporary landscape thus contains at least four distinct conceptions of political freedom — negative liberty, positive liberty, republican non-domination, and capabilities — each with independent motivation, each vulnerable to specific objections, and none reducible to the others. Berlin's warning about the dangers of positive liberty remains relevant, but it no longer defines the boundaries of the debate.

Compatibilism, incompatibilism, and the structure of free will [Master]

The free-will debate is not a single argument but a layered structure of arguments, counterarguments, and meta-arguments about what the question even means.

Incompatibilism: the consequence argument. Van Inwagen's consequence argument, reconstructed in the Intermediate section, is the central formal argument for incompatibilism. Its force is that it makes explicit the assumptions needed to get from determinism to the conclusion that no one has a choice. If those assumptions are acceptable, the compatibilist must either reject determinism or accept that free will is an illusion.

The most common compatibilist response attacks rule $\alpha$, the transfer principle. The principle says: if you have no choice about $p$, and no choice about $p\to q$, then you have no choice about $q$. This looks valid in the abstract. But David Lewis (1981) and others have argued that it fails in cases where $p\to q$ is a counterfactual conditional whose truth depends on what the agent would do. If the conditional "if I were to try to raise my hand, my hand would go up" is true in part because I would try, then the transfer principle lets a fact about what I would do feed back into a conclusion about what I have no choice about, which is circular. The debate over the validity of rule $\alpha$ is a technical dispute in modal logic with direct philosophical stakes.

Compatibilism: the conditional analysis. The classical compatibilist position, associated with Hobbes, Hume, and Mill, is that freedom is compatible with determinism because freedom requires not the absence of causation but the absence of specific kinds of causes — coercion, compulsion, mental illness. The conditional analysis of "could have done otherwise" captures this: "agent $A$ could have done otherwise" means "if $A$ had chosen to do otherwise, $A$ would have done otherwise." On this reading, the ability to do otherwise does not require the absence of determining causes; it requires only that the agent's choices are effective — that the causal chain runs through the agent's deliberation rather than bypassing it.

The conditional analysis faces a well-known objection from Chisholm (1964). Suppose a deterministic world in which an agent is coerced into performing an action. In this world, the agent could not have chosen otherwise (determinism), so the conditional "if the agent had chosen otherwise, the agent would have done otherwise" has a false antecedent and is vacuously true. The conditional analysis thus classifies the coerced action as free — which is wrong. Compatibilists have responded by refining the analysis: Harry Frankfurt's hierarchical model, P. F. Strawson's reactive attitudes framework, and John Martin Fischer's reasons-responsiveness account all attempt to provide a compatibilist condition for freedom that does not fall to the Chisholm objection.

Frankfurt cases and alternate possibilities. Frankfurt's 1969 paper ^{[Frankfurt 1969]} introduces a class of counterexamples to the principle of alternate possibilities (PAP): an agent is morally responsible for an action only if they could have done otherwise. In a Frankfurt case, the agent acts on their own, for their own reasons, but a standby mechanism ensures that they could not have done otherwise. The agent is intuitively responsible, yet PAP says they are not. If Frankfurt cases succeed, PAP is false, and the link between determinism (which removes alternate possibilities) and the loss of moral responsibility is broken. This does not prove compatibilism, but it removes one of the strongest intuitions driving incompatibilism.

Critics of Frankfurt cases — including Kane (1996) ^{[Kane 1996]} and Widerker (1995) — argue that the cases either fail to eliminate all alternate possibilities (the agent could still "flicker" between options before the mechanism activates) or depend on a specific account of moral responsibility that begs the question against PAP. The "flicker of freedom" response holds that as long as the agent has some alternative available — even a minor one, like hesitating before acting — PAP is satisfied. Frankfurtians respond that the relevant alternate possibilities are not robust enough to ground responsibility; the minor alternatives are not genuine choices. This exchange has generated a large technical literature on what counts as a "robust" alternative.

Libertarianism on free will. Metaphysical libertarians hold that free will exists and requires indeterminism. The challenge for libertarianism is explanatory: if an agent's choice is not determined by prior causes, what makes it the agent's choice rather than a random event? Indeterminism alone does not produce freedom; it produces chance. Robert Kane's event-causal libertarianism ^{[Kane 1996]} attempts to meet this challenge by locating indeterminism at the point of effort-of-will in difficult moral decisions. The agent makes an effort to choose the morally right course, and whether that effort succeeds is undetermined. If it succeeds, the agent is responsible because the effort was the agent's own. Critics object that this makes responsibility depend on luck — whether the indeterministic effort succeeds or fails is not under the agent's control.

Mill, harm, and the limits of liberty [Master]

Mill's On Liberty (1859) is the single most influential liberal text in the negative-liberty tradition ^{[Mill 1859]}. Its central principle — that the only legitimate ground for state coercion is harm to others — has been absorbed into the constitutional and legal frameworks of most liberal democracies. But the principle is more contested than its canonical status suggests.

The harm principle draws a line between the self-regarding and the other-regarding. Actions that affect only the agent are beyond the reach of legitimate coercion, no matter how foolish, self-destructive, or offensive they may be. Actions that cause harm to others can be regulated, provided the regulation actually prevents that harm. Mill's defence of this line relies on several supporting arguments: the argument from utility (social progress is maximised when individuals are free to experiment with ways of living), the argument from fallibility (censors may be wrong about what is harmful), and the argument from dead dogma (beliefs that are never challenged become empty formalities).

Each supporting argument has been attacked. The utility argument assumes that the social value of free experimentation outweighs the social cost of harmful actions — but this is an empirical claim, not a conceptual one, and it may be false in specific cases (e.g., the spread of dangerous misinformation). The fallibility argument cuts both ways: if censors are fallible, so are individuals deciding what is good for them. The dead-dogma argument assumes that challenged beliefs remain meaningful, but does not show that they would not remain meaningful under other conditions.

The deepest issue is the scope of "harm." Mill defines harm narrowly: tangible, direct damage to the interests of others. Offense, disgust, and moral disapproval do not count as harm on Mill's account. But this narrow definition is difficult to maintain in practice. If a neighbour plays loud music at 3 a.m., the harm is real but intangible (sleep disruption, annoyance). If a factory pollutes a river, the harm may be diffuse and statistical (increased cancer rates downstream) rather than direct. If a speaker advocates the violent overthrow of a democracy, the harm may be future and uncertain rather than present and tangible.

Joel Feinberg's four-volume The Moral Limits of the Criminal Law (1984–1988) is the most sustained attempt to work out the boundaries of the harm principle in legal theory. Feinberg introduces a distinction between harm (setbacks to important interests) and offense (unpleasant psychological states caused by the perceptions of others), and argues that the law may legitimately regulate the former but only under strict conditions regulate the latter. The Feinberg framework extends Mill's principle while preserving its core liberal commitment.

The harm principle also interacts with the positive-liberty tradition in an important way. If positive liberty requires certain material conditions — education, healthcare, a basic income — then the state's failure to provide those conditions constitutes a form of harm-by-omission. The negative-liberty framework, which focuses on what the state should not do, struggles to account for the claim that the state's inaction can be a form of coercion. Sen's capabilities approach ^{[Sen 1999]} resolves this by treating the deprivation of capabilities as itself a freedom deficit, not merely a welfare deficit. On this reading, the harm principle does not conflict with positive liberty; it requires positive liberty as a precondition for its own meaningful application.

Existentialist freedom: Sartre [Master]

The existentialist tradition introduces a conception of freedom that is neither negative nor positive in Berlin's sense, and neither compatibilist nor incompatibilist in the free-will sense. Jean-Paul Sartre's account, presented most accessibly in the 1946 lecture Existentialism is a Humanism ^{[Sartre 1946]} and developed at length in Being and Nothingness (1943), holds that human beings are "condemned to be free" — that freedom is not a contingent feature of our situation but a constitutive feature of human existence.

Sartre's argument has two stages. The first is phenomenological: when we examine our experience of choosing, we find that we are always already oriented toward possibilities, and that this orientation is not imposed on us from outside but constitutes our own mode of being. Consciousness, for Sartre, is nothingness — it is not a thing with fixed properties but a perpetual transcendence of whatever is given toward what is not yet. This structure is what Sartre calls freedom: the inability of consciousness to be fully determined by any set of facts about itself.

The second stage is normative: because we are free, we are responsible for everything we do and, in a more radical sense, for what we are. Sartre rejects the idea that there are excuses — circumstances, upbringing, nature, God's will — that relieve us of responsibility. To claim that you "had no choice" is, on Sartre's account, an exercise in bad faith (mauvaise foi): the deliberate self-deception of pretending to be a thing determined by its properties rather than a free consciousness transcending them.

Bad faith is Sartre's diagnosis of the human tendency to flee from the anxiety of radical freedom. The waiter who performs his role with excessive precision, identifying entirely with the social function "waiter," is in bad faith because he is pretending to be a fixed object — a waiter-essence — rather than a free consciousness that has chosen, and continues to choose, to work as a waiter. The person who says "I had no choice but to betray my friend" is in bad faith because they are treating themselves as a causal mechanism ("the situation made me do it") rather than as an agent who chose, in that situation, to betray.

The existentialist conception of freedom diverges from both negative and positive liberty in important ways. Negative liberty is a political condition (non-interference). Positive liberty is a condition of self-realisation. Sartrean freedom is an ontological condition — it is not something you can gain or lose but something you are, whether you acknowledge it or not. This makes existentialist freedom both more radical and less politically useful than Berlin's categories. It is more radical because it denies that any circumstance — including imprisonment, coercion, or physical constraint — can take away your freedom. Sartre insists that even the prisoner on death row is free: free to accept, resist, or reinterpret their situation. It is less politically useful because if freedom is inescapable, the political project of expanding or protecting freedom loses its urgency.

Sartre was aware of this tension and spent much of his later career (especially the Critique of Dialectical Reason, 1960) trying to reconcile existentialist freedom with Marxist social analysis. The reconciliation is uneasy. If material conditions determine consciousness (as Marxism holds), then radical freedom is an illusion of the bourgeoisie. If consciousness is radically free (as existentialism holds), then material conditions cannot fully determine it. Sartre's later work attempts to hold both claims simultaneously, with mixed success.

The existentialist contribution to the freedom-liberty debate is not a competitor to Berlin or Mill but a challenge to the framing. Berlin asks: under what conditions is a person free? Sartre asks: what kind of being is a person, such that the question of freedom can arise at all? The shift from political conditions to ontological structure changes what counts as a satisfactory answer. It also introduces a register of analysis — phenomenological, first-personal, attuned to the experience of choosing — that is absent from the analytical tradition's formal reconstructions.

Connections [Master]

• Determinism and the measurement problem 20.03.01 pending. The consequence argument's premise that the laws of nature determine all future events is directly connected to the interpretation of physical law. The measurement problem in quantum mechanics — whether the wave function evolves deterministically or collapses indeterministically — bears on whether premise P4 of the consequence argument is physically true. If QM introduces genuine indeterminacy, the determinism premise may be false. The phil-of-physics treatment at 20.03.01 pending provides the physical context for this premise.

• Consciousness and agency 20.06.01 pending. Frankfurt's account of moral responsibility, and Sartre's account of radical freedom, both presuppose substantive claims about what consciousness is and how it relates to agency. Frankfurt's hierarchical model requires that an agent can identify with or reject their own first-order desires — a capacity whose explanation lies in the philosophy of mind. The pending unit at 20.06.01 pending is the natural home for the agency-side analysis.

• Scientific realism and determinism 20.07.01 pending. The fixity-of-the-laws premise in the consequence argument depends on what laws of nature are. If laws are mere descriptive regularities (Humeanism), they may not "fix" anything, and the premise is weakened. If laws are governing relations (non-Humeanism), the premise is stronger. The general phil-of-science treatment of laws and realism at 20.07.01 pending provides the context.

• Moral responsibility 20.02.01 pending. The prerequisite unit on moral responsibility provides the foundational framework for understanding why free will matters: without it, the attribution of praise and blame lacks its standard grounding. The consequence argument and Frankfurt cases are responses to this prior question.

• Political philosophy and justice [20.02.NN]. Mill's harm principle and Berlin's negative/positive distinction operate within a broader framework of political theory that includes theories of justice, rights, and the legitimacy of state power. A treatment of distributive justice (Rawls, Nozick) would provide the broader institutional context in which the harm principle operates.

Historical & philosophical context [Master]

The distinction between two kinds of freedom has ancient roots. Isaiah Berlin traced the positive-liberty tradition to Plato and Rousseau, and the negative-liberty tradition to Hobbes, Locke, and Bentham. The naming convention is Berlin's, but the underlying tension is older.

Hobbes's Leviathan (1651) ^{[Hobbes 1651]} gives the paradigmatic negative-liberty definition: "Liberty, or freedom, signifieth properly the absence of opposition; by opposition, I mean external impediments of motion." Hobbes explicitly rejects the idea that liberty requires a capacity or a rational will — a man who is pushed through a door is "free" to pass through it, in Hobbes's sense, because nothing in the door frame impedes his motion, regardless of whether he chose to enter. This austere definition was contested almost immediately.

Locke's Second Treatise of Government (1689) ^{[Locke 1689]} introduces a richer concept: natural liberty as the freedom to act within the bounds of the law of nature, without being subject to the arbitrary will of another person. Locke's liberty is already more than Hobbesian non-interference — it contains an element of non-domination that anticipates Pettit's republican freedom by three centuries.

Rousseau's The Social Contract (1762) moves decisively into the positive-liberty register. Freedom, for Rousseau, is obedience to a law one has prescribed to oneself — autonomy in the literal sense. The general will is the expression of the collective self-legislation that makes each citizen simultaneously sovereign and subject. Berlin identifies Rousseau as the origin point of the "forced to be free" logic that he regards as the structural danger of positive liberty.

Mill's On Liberty (1859) ^{[Mill 1859]} is the canonical text of the negative-liberty tradition in its most influential form. Mill combines the harm principle with arguments from utility, fallibility, and the value of individuality. The essay was written in the context of Victorian England's moral and religious conformism, and Mill's target was not state tyranny but the "tyranny of the majority" — social pressure to conform that operates without legal coercion. This makes On Liberty an early treatment of what would now be called informal or structural constraints on freedom.

Berlin's "Two Concepts of Liberty" (1958) ^{[Berlin 1958]} was delivered as Berlin's inaugural lecture at Oxford. Written during the Cold War, it reflects the confrontation between Western liberal democracy and Soviet communism. Berlin's warning about positive liberty was shaped by his reading of the totalitarian claim that the state could force people to be free by directing them toward their "true" interests. The lecture has been reprinted in Four Essays on Liberty (1969) and Liberty (2002, ed. Hardy).

Frankfurt's "Alternate Possibilities and Moral Responsibility" (1969) ^{[Frankfurt 1969]} introduced the counterexample strategy that bears his name. The paper is short — fifteen pages — and its argument is deceptively simple: construct a case where an agent acts freely but could not have done otherwise, thereby severing the link between alternate possibilities and moral responsibility. The paper generated a literature of hundreds of articles and defined one of the central debates in contemporary action theory.

Van Inwagen's An Essay on Free Will (1983) ^{[van Inwagen 1983]} systematised the consequence argument and placed the formal structure of the free-will debate on the table for analytic philosophy. Van Inwagen presented the argument in multiple forms (the Consequence Argument, the Mind Argument, the Direct Argument) and argued that compatibilism is untenable given the validity of the transfer principle. The book is the standard reference point for contemporary incompatibilism.

Sartre's L'existentialisme est un humanisme (1946) ^{[Sartre 1946]} was delivered as a public lecture and published as a short book aimed at a general audience. It is the most accessible statement of existentialist freedom, though Sartre himself later distanced himself from some of its claims. The fuller treatment is in L'etre et le neant (Being and Nothingness, 1943), where Sartre develops the ontological argument for radical freedom through the phenomenology of consciousness, the body, and the gaze of the Other.

Bibliography [Master]

Historical sources:

• Hobbes, T. — Leviathan (1651), Ch. 14, 21. [Need to source.]
• Locke, J. — Second Treatise of Government (1689), Ch. 2–4. [Need to source.]
• Rousseau, J.-J. — The Social Contract (1762), Book I, Ch. 6–7. [Need to source.]
• Mill, J. S. — On Liberty (1859). [Need to source.]

Twentieth-century canonical:

• Berlin, I. — "Two Concepts of Liberty," in Four Essays on Liberty (Oxford University Press, 1969). Originally delivered 1958. [Need to source.]
• Frankfurt, H. — "Alternate Possibilities and Moral Responsibility," Journal of Philosophy 66, 5–20 (1969). [Need to source.]
• MacCallum, G. — "Negative and Positive Freedom," Philosophical Review 76, 312–334 (1967). [Need to source.]
• van Inwagen, P. — An Essay on Free Will (Oxford University Press, 1983). [Need to source.]
• Sartre, J.-P. — Existentialism is a Humanism (1946); English trans. (Yale University Press, 2007). [Need to source.]

Contemporary:

• Pettit, P. — Republicanism: A Theory of Freedom and Government (Oxford University Press, 1997). [Need to source.]
• Sen, A. — Development as Freedom (Oxford University Press, 1999). [Need to source.]
• Kane, R. — The Significance of Free Will (Oxford University Press, 1996). [Need to source.]
• Fischer, J. M. — The Metaphysics of Free Will (Blackwell, 1994). [Need to source.]
• Feinberg, J. — The Moral Limits of the Criminal Law, 4 vols. (Oxford University Press, 1984–1988). [Need to source.]

---

First ethics unit in the Codex, produced 2026-05-21. The 02-ethics chapter extends the section allocation beyond what PHILOSOPHY_PLAN v1 defined for 20.01–07. This unit declares hooks_out to 20.03.01 pending, 20.06.01 pending, 20.07.01 pending, and 20.02.01 pending via the Connections section; all are proposed pending target-domain reviewer attestation. Status: shipped pending external ethics/political-philosophy reviewer.