20.02.01 · philosophy / ethics

Theories of justice: Rawls, Nozick, and fairness

shipped3 tiersLean: none

Anchor (Master): primary sources: Rawls 1971, Nozick 1974, Sen 1992, Cohen 2008

Intuition Beginner

What does it mean for a society to be just? The question is old -- Plato asked it, Hobbes and Locke and Rousseau gave competing answers, and every revolution in modern history has claimed justice as its motive. But the twentieth century produced two answers so precise, so rigorously argued, and so opposed to each other that they reshaped how philosophers, economists, and policymakers think about fairness itself.

The first answer comes from John Rawls. In A Theory of Justice (1971), Rawls asked you to imagine designing the rules of society from behind a veil of ignorance -- a thought experiment where you do not know your own talents, your social class, your race, your gender, or even your conception of the good life. From this "original position," Rawls argued, any rational person would choose two principles.

The first principle is that each person should have the most extensive set of basic liberties compatible with a similar set for others. The second principle is that social and economic inequalities are permissible only if they benefit the least advantaged members of society. This second principle is the difference principle: inequalities are just only when they work to the advantage of the worst-off.

The second answer comes from Robert Nozick. In Anarchy, State, and Utopia (1974), Nozick argued that justice is not about the pattern of distribution -- who has how much -- but about the history of how people got what they have. If you acquired your holdings justly (through original acquisition or voluntary transfer), then any distribution that results is just, no matter how unequal. Taxation for redistributive purposes, on Nozick's view, is morally equivalent to forced labour. The core of his position is the entitlement theory: justice in holdings consists of justice in acquisition, justice in transfer, and rectification of past injustice.

These two positions define the main axis of debate in contemporary political philosophy. Rawls gives you a theory that judges outcomes by how they affect the worst-off. Nozick gives you a theory that judges processes by whether they respect individual rights. They disagree about what justice is for -- whether it exists to produce fair patterns or to protect individual entitlements -- and they disagree about what the state may permissibly do.

Why does this matter? Three reasons.

First, the Rawls-Nozick debate is the framework within which most real-world policy disagreements are ultimately argued. Arguments about progressive taxation, universal healthcare, welfare programmes, and property rights are, at bottom, arguments about whether justice is patterned or historical.

Second, the debate forces you to confront a question that is as much philosophical as political: what kind of equality matters? Rawls thinks equality of opportunity and equality in the distribution of primary goods matter. Nozick thinks equality before the law -- the equal right to acquire, hold, and exchange -- is the only equality justice requires. Sen and Nussbaum, later in the debate, argue that neither goods nor rights are the right metric: what matters is equality of capability -- what people are actually able to do and be.

Third, the disagreement is not merely about policy. It is about the foundations of moral reasoning itself. Rawls builds his theory on a hypothetical contract -- what free and equal persons would agree to under fair conditions. Nozick builds his on natural rights that individuals possess prior to any contract. The dispute between contractarian and libertarian foundations is a dispute about what kind of thing a moral theory can be: whether it derives from agreement, from rights, from consequences, or from something else entirely.

Visual Beginner

Imagine two societies. In Society A, the government collects taxes and redistributes income so that the poorest citizens have access to healthcare, education, and a social minimum. Inequalities are allowed, but only when the system that produces them also makes the poorest better off than they would be under any alternative arrangement. This is a Rawlsian society: the distribution has a pattern, and the pattern is judged by its worst-off member.

In Society B, the government enforces contracts and protects property rights but does not redistribute. People keep what they earn through voluntary exchange. If someone becomes fabulously wealthy through talent and voluntary transactions, the state has no business taking a portion to give to others. This is a Nozickian society: the distribution has a history, and the history is judged by the justice of each step.

The diagram captures the fundamental disagreement. Rawls looks at the end state and asks: is this fair to the least advantaged? Nozick looks at the process and asks: was each step free and voluntary? Both claim to be talking about justice. They are talking about different things.

Worked example Beginner

Consider the Wilt Chamberlain argument, Nozick's most famous thought-experiment (from Anarchy, State, and Utopia, pp. 160-164). Imagine you have achieved your favoured pattern of distributive justice -- call it D1. Whatever your theory says a just distribution looks like, D1 is it. Now imagine Wilt Chamberlain, a famous basketball player, signs a contract: from every ticket sold, twenty-five cents goes directly to him. A million people willingly pay to watch him play. At the end of the season, Chamberlain has $250,000 and the distribution has changed to D2, which is much more unequal.

Is D2 just? Nozick's argument is that any theory of justice that endorses D1 but condemns D2 must either (a) prohibit the voluntary exchanges that turned D1 into D2, which means prohibiting free people from spending their money as they choose, or (b) continuously interfere in people's lives to maintain the pattern, redistributing away whatever inequalities voluntary exchange produces. Nozick calls this "the tyranny of the pattern": if justice is a pattern, then maintaining it requires constant, liberty-violating interference.

Rawlsians respond by denying that D1 was chosen in a vacuum. The original position does not endorse a snapshot distribution; it endorses a structure of institutions -- including taxation -- that continuously maintains background justice. The difference principle does not demand that we freeze a pattern but that the basic structure of society be arranged so that inequalities track the advantage of the worst-off over time.

Chamberlain's earnings are just, on the Rawlsian view, if the institutional framework that allows them also maximises the position of the least advantaged -- for instance, through taxation that funds public goods from which everyone benefits. Nozick anticipated this reply and argued that it concedes the point: the Rawlsian framework must constantly reach in and rearrange people's holdings, which is precisely the interference individual rights are supposed to prevent.

The exchange between Rawls and Nozick on this point is the spine of the distributive-justice debate. Every subsequent contribution -- Dworkin on equality of resources, Sen on capabilities, Cohen's critique of the difference principle -- takes a position relative to this axis.

Check your understanding Beginner

Exercise (easy, multiple choice).

The "veil of ignorance" in Rawls's original position is best described as:

A. A literal physical barrier that prevents people from seeing each other. B. A hypothetical condition in which parties choosing principles of justice do not know their own social position, natural talents, or conception of the good. C. A legal doctrine that prevents courts from considering the identity of defendants. D. A psychological theory about how people form moral judgements.

Hint

The veil of ignorance is a thought experiment, not a real institution. Its purpose is to ensure that the choice of principles is not biased by knowledge of one's own interests.

Answer

Option B. The veil of ignorance strips the parties in the original position of knowledge about their particular characteristics -- their class, talents, race, gender, and comprehensive moral views. The purpose is to ensure that the principles chosen are fair to everyone, since no one can tailor principles to their own advantage.

Exercise (easy, short answer).

In your own words (2-3 sentences), state the difference principle and explain why Rawls thinks rational parties behind the veil of ignorance would choose it rather than simple equality.

Hint

The difference principle allows inequalities, but only under a specific condition. Why might allowing some inequalities be better for everyone than strict equality?

Answer

The difference principle states that social and economic inequalities are just only if they are attached to positions open to all under conditions of fair equality of opportunity, and only if they are to the greatest benefit of the least advantaged members of society. Rational parties behind the veil of ignorance would choose it over strict equality because some inequalities -- for instance, paying doctors more to attract talent to medicine -- can make everyone better off, including the worst-off. The difference principle captures this: it permits inequality only when the worst-off do better than they would under any alternative arrangement.

Exercise (easy, multiple choice).

Nozick's entitlement theory of justice holds that a distribution of holdings is just if and only if:

A. It maximises total social welfare. B. It arises from a process in which each holding was acquired and transferred justly, starting from just initial acquisitions. C. It satisfies the difference principle. D. It is approved by a democratic majority.

Hint

Nozick explicitly rejects both welfare maximisation and the difference principle. His theory is historical and process-based, not end-state-based.

Answer

Option B. The entitlement theory has three parts: justice in acquisition (how unowned things come to be owned), justice in transfer (how holdings change hands voluntarily), and rectification of injustice (how to correct past violations). If every step in the history of a holding satisfies the first two, the resulting distribution is just -- regardless of its pattern.

Exercise (medium, short answer).

Explain the Wilt Chamberlain argument in your own words. What is Nozick trying to show with it?

Hint

The argument starts from a distribution that any theory of justice would endorse and shows that voluntary exchange leads to inequality. What conclusion does Nozick draw about patterned principles of justice?

Answer

Nozick asks you to start with D1, your favoured just distribution. People voluntarily transfer some of their holdings to Wilt Chamberlain in exchange for watching him play basketball. The result is D2, a much more unequal distribution. Nozick argues that if D1 was just and each transfer was voluntary, then D2 is also just -- and any theory that condemns D2 while endorsing D1 must either prohibit voluntary exchange or continuously redistribute to maintain the pattern. The argument is meant to show that patterned principles of justice are incompatible with individual liberty: you cannot have both a fixed pattern of distribution and freedom to exchange.

Formal definition Intermediate+

The debate between Rawls and Nozick can be made precise. Let us define the main concepts formally and then examine the argument structure of each position.

Definition (Distributive justice). A theory of distributive justice specifies conditions under which a distribution of social goods $D = {g_{1}, g_{2}, \dots, g_{n}}$ among $n$ members of a society $S$ is just, where $g_{i}$ represents the bundle of primary goods, resources, or capabilities assigned to person $i$ .

Definition (Patterned principle). A principle of justice is patterned if it specifies that a distribution is just when it matches some pre-defined structural criterion -- for example, equality, need, merit, or the maximin condition. Formally, a patterned principle specifies a set $P$ of permissible distributions such that $D \in P$ iff $D$ satisfies the pattern.

Definition (Historical principle). A principle of justice is historical if it specifies that a distribution is just when it arises from a sequence of just acquisitions and voluntary transfers from a just starting point. Formally, $D$ is just iff there exists an initial just distribution $D_{0}$ and a sequence of transfers $(T_{1}, T_{2}, \dots, T_{k})$ , each voluntary or a just acquisition, such that $D_{0} T_{1} D_{1} T_{2} \dots T_{k} D$ .

Rawls's theory is patterned (the difference principle defines $P$ ). Nozick's theory is historical (justice is a function of the sequence of transfers, not of the final distribution).

Rawls's two principles of justice

Rawls states his two principles as follows, with lexical priority of the first over the second:

First Principle. Each person has an equal right to a fully adequate scheme of equal basic liberties which is compatible with a similar scheme of liberties for all.

Second Principle. Social and economic inequalities are to satisfy two conditions: (a) they are to be attached to offices and positions open to all under conditions of fair equality of opportunity; and (b) they are to be to the greatest benefit of the least advantaged members of society (the difference principle).

The lexical ordering means: basic liberties may not be traded for economic gains. No amount of additional wealth for the worst-off can justify restricting freedom of speech, conscience, or political participation.

Formal statement of the difference principle. Let $u_{i}$ denote the index of primary goods available to person $i$ . A distribution $D$ satisfies the difference principle if and only if:

$i min u_{i} (D) \geq i min u_{i} (D^{'}) for all feasible distributions D^{'} .$

That is, the worst-off position under $D$ is at least as good as the worst-off position under any alternative feasible distribution. The difference principle is a maximin criterion applied to the index of primary goods.

Nozick's entitlement theory

Nozick's theory consists of three principles:

Justice in acquisition. A person who acquires a holding that was previously unowned, in accordance with a just rule of acquisition (e.g., Locke's labour-mixing with the Lockean proviso that "enough and as good" is left for others), is entitled to that holding.
Justice in transfer. A person who acquires a holding through voluntary exchange, gift, or bequest from someone entitled to it, is entitled to that holding.
Rectification of injustice. If a holding was acquired or transferred unjustly, rectification is required to restore justice.

Formal statement. A distribution $D$ is just if and only if every holding in $D$ traces back, through a chain of just transfers, to a just initial acquisition. Symbolically:

$D is just ⟺ \forall i : holding (i) traces to just acquisition via just transfers .$

Key theorem -- the impossibility of liberty-upholding patterned justice

Nozick's Wilt Chamberlain argument can be reconstructed as a formal impossibility theorem:

Premise 1 (Liberty of transfer). Individuals have the right to transfer holdings to which they are entitled, to willing recipients, on mutually agreed terms.

Premise 2 (Pattern sensitivity). Let $D$ be any distribution satisfying pattern $P$ . There exists a finite sequence of voluntary transfers, each permitted by Premise 1, that transforms $D$ into $D^{'}$ such that $D^{'} \in / P$ .

Conclusion. No patterned principle of justice $P$ is compatible with liberty of transfer: maintaining $P$ requires interfering with voluntary exchanges.

The argument is valid. Its soundness depends on Premise 2, which is an empirical claim about the consequences of voluntary exchange under any substantive pattern. Nozick supports it with the Chamberlain example and argues that it generalises: any substantive pattern will be disrupted by voluntary transfers, because people have different preferences, talents, and conceptions of the good, and they will exchange in ways that diverge from any pre-specified pattern.

Rawlsians reject the argument's force by denying that the original position licenses a snapshot pattern. The two principles govern the basic structure of institutions over time, not a momentary distribution. Taxation and redistribution are part of the institutional background within which voluntary exchange takes place -- not an interference with liberty but a condition for maintaining background justice over time. The disagreement, then, is not about the formal validity of Nozick's argument but about the underlying conception of liberty: Nozick treats liberty as freedom from interference with voluntary exchange; Rawls treats liberty as a scheme of equal basic rights that includes the institutional conditions for fair cooperation over a complete life.

Counterexamples to common slips

"Rawls advocates equality of outcome." Rawls explicitly does not. The difference principle permits inequalities when they benefit the worst-off. A society where doctors earn more than janitors can satisfy the difference principle if the incentive structure attracts enough doctors to make healthcare -- and thus the position of the worst-off -- better than it would be under strict equality.
"Nozick is against all taxation." Nozick is against redistributive taxation -- taxation that takes from some to give to others for the purpose of altering the distribution of holdings. He permits taxation for the minimal state's core functions: protection against force, theft, and fraud, and the enforcement of contracts.
"The original position is meant to be realistic." Rawls is explicit that the original position is a device of representation, not an actual historical event. It models what it means for terms of cooperation to be fair: they are terms that free and equal persons, properly motivated and not biased by their particular circumstances, could agree to.
"Nozick's entitlement theory assumes a just starting point that never existed." This is correct and is the basis of the most serious objection to historical theories: in the real world, almost all holdings trace back to injustice -- conquest, theft, colonial expropriation. Nozick acknowledges this and concedes that rectification is required, but he never provides a detailed theory of rectification. This gap is the Achilles' heel of the entitlement theory.

Exercises Intermediate+

Exercise 1 (easy, argument reconstruction).

Reconstruct the argument from the original position to the difference principle. Be explicit about each step and the premises it requires.

Hint

The argument has three key moves: (1) the veil of ignorance removes knowledge of one's position; (2) the parties are rational and mutually disinterested; (3) given these conditions, the maximin strategy is rational.

Answer

Premise 1 (Veil of ignorance). The parties in the original position do not know their social position, natural talents, race, gender, or conception of the good.

Premise 2 (Rationality). The parties are rational and mutually disinterested -- each seeks to advance their own interests (whatever those turn out to be) without envy or altruism.

Premise 3 (Maximin rationality). Under conditions of radical uncertainty about one's own position, where the stakes involve the fundamental terms of social cooperation for a complete life, and where there is no prior agreement about the probability of occupying any particular social position, the rational strategy is maximin: choose the arrangement that makes the worst possible outcome as good as possible.

Conclusion. The parties would choose the difference principle, because it maximises the position of the worst-off -- the position any party might end up occupying once the veil is lifted.

The most contested premise is Premise 3. Critics (Harsanyi, among others) argue that under uncertainty, rational agents should maximise expected utility using an objective probability distribution (e.g., equal probability of occupying any position), which leads to average utilitarianism rather than maximin. Rawls responds that the original position is not a case of risk but of uncertainty in the Knightian sense: there is no basis for assigning probabilities, and the stakes are too high to gamble.

Exercise 2 (medium, critique).

State what you take to be the strongest objection to Rawls's difference principle and explain why it is strong. Then give the best Rawlsian reply you can construct.

Hint

Consider: Is maximin too risk-averse? Does the difference principle permit too much inequality? Does it ignore desert? Does it ignore the claims of the non-worst-off?

Answer

A representative objection (from Harsanyi and others): maximin is irrational as a decision rule because it is excessively conservative. Suppose there are two possible social arrangements. Under arrangement A, the worst-off have 100 units and the next-worst-off have 101. Under arrangement B, the worst-off have 99 units and everyone else has 500. The difference principle chooses A, even though B makes almost everyone dramatically better off and the loss to the worst-off is marginal. This seems irrational: a small gain for the worst-off should not outweigh massive gains for everyone else.

Rawlsian reply: the example misrepresents the choice situation. The difference principle applies within a framework where the first principle (equal basic liberties) is already satisfied and where fair equality of opportunity obtains. Within that framework, the "worst-off" is not a tiny minority but a significant social group -- those with the least share of primary goods over a complete life. Moreover, the choice is not between A and B in the abstract but between institutional structures evaluated over generations. A structure that marginally advantages the worst-off while severely constraining everyone else would not be stable -- and stability is a constraint Rawls builds into the argument. The difference principle is not pure maximin; it is maximin subject to the constraints of the first principle, fair equality of opportunity, and the stability of the system over time.

Exercise 3 (medium, application).

A government proposes the following policy: a flat tax of 30% on all income, with all revenue used to fund public healthcare and education. Under this policy, the worst-off are better off than under the current system (which has lower taxes but no universal healthcare). However, the policy is not the arrangement that maximises the position of the worst-off -- a more progressive tax system would do even better for them. Does this policy satisfy the difference principle? Explain.

Hint

The difference principle says inequalities must be "to the greatest benefit of the least advantaged." It does not say "to the benefit of the least advantaged."

Answer

No, the policy does not satisfy the difference principle on a strict reading. The difference principle requires that the arrangement maximise the position of the worst-off, not merely improve it. Since a more progressive tax system would make the worst-off even better off, the flat-tax policy falls short of the difference principle's demand.

However, Rawls allows that the difference principle applies to the basic structure of institutions, not to individual policies in isolation. If the more progressive system has hidden costs -- reduced incentives, capital flight, administrative complexity -- that make it infeasible or unstable, then the flat-tax system may be the best feasible approximation. The difference principle is an ideal to which real institutions should approximate, not a literal requirement that every policy choice be proven globally optimal for the worst-off. This pragmatic dimension of Rawls's theory is often underemphasised.

Exercise 4 (medium, comparison).

Compare how Rawls and Nozick would evaluate the justice of a system in which children inherit large fortunes from their parents. Each philosopher should get 2-3 sentences.

Hint

Rawls cares about the effects on the worst-off and fair equality of opportunity. Nozick cares about whether the transfer was voluntary.

Answer

Rawls: Large inheritances create inequalities in starting positions that violate fair equality of opportunity -- the second condition of the second principle. If inherited wealth allows some children access to dramatically better education, healthcare, and social networks, then the system does not satisfy the difference principle unless it is structured so that even those without inheritance are made as well-off as possible. Rawls endorsed some form of inheritance taxation to maintain background justice.

Nozick: If the parent acquired the fortune justly and the bequest is a voluntary transfer, the child's inheritance is just, regardless of the resulting inequality. The child is entitled to the holding because it came through a just transfer from someone entitled to it. Taxation of inheritance for redistributive purposes violates the entitlement of the parent to dispose of their holdings as they wish.

Utilitarian approaches to justice Master

Before examining the capabilities approach and the contemporary debate, it is essential to locate the Rawls-Nozick axis within the broader landscape of theories of justice. The most important alternative framework, and the one Rawls explicitly defined himself against, is utilitarianism.

Classical utilitarianism and justice

Classical utilitarianism holds that the just arrangement is the one that maximises total or average utility across all persons. Justice, on this view, is a subspecies of efficiency: a distribution is just when it produces the greatest aggregate well-being.

The formal statement is straightforward. Let $u_{i}$ denote the utility of person $i$ . Classical utilitarianism selects the distribution $D^{*}$ that maximises:

$D^{*} = ar g D max i = 1 \sum n u_{i} (D) .$

Average utilitarianism divides by $n$ ; the difference matters only when population size varies.

Rawls's central objection to utilitarianism is that it does not take seriously the separateness of persons. If the goal is to maximise the sum, then sacrificing the well-being of a few for the benefit of many is not only permitted but required whenever the gains to the many outweigh the losses to the few. But this, Rawls argues, conflates the choice that a single person would make about distributing resources across the stages of their own life (where it is rational to accept losses now for gains later) with the choice society makes about distributing resources among distinct persons (where no such compensation is guaranteed). The separateness of persons is the fundamental moral fact that utilitarianism fails to accommodate.

Nozick shares this objection but draws a different conclusion. For Nozick, the separateness of persons grounds individual rights that function as side-constraints on action: you may not violate a person's rights even to produce better overall outcomes. For Rawls, the separateness of persons grounds a fair procedure for choosing principles: the original position ensures that the terms of cooperation are acceptable to each person as a free and equal moral agent.

Rule utilitarianism and distributive justice

Rule utilitarianism -- the view that the just arrangement is the one governed by rules that, if generally followed, would maximise utility -- can accommodate some distributive concerns. A rule utilitarian might endorse a social minimum, progressive taxation, or even something resembling the difference principle, on the grounds that societies with these institutions produce more aggregate well-being over time than societies without them (because extreme inequality causes social instability, health problems, and wasted human potential).

The Rawlsian reply is that this gets the moral order backwards. Justice is not justified because it produces good consequences; it is justified because it is what free and equal persons would agree to under fair conditions. The right is prior to the good. A rule utilitarian who endorses the difference principle does so for the wrong reason -- because it maximises utility, not because it is fair -- and could in principle withdraw the endorsement if the utility calculus changed. Rawlsian justice is meant to be stable for the right reasons, not merely instrumentally effective.

Harsanyi's utilitarian original position

John Harsanyi (1953, 1975) offered a direct challenge to Rawls from within the contractarian framework. Harsanyi accepted the device of the original position and the veil of ignorance but argued that rational agents behind the veil should be modelled as maximising expected utility rather than applying maximin. If each person has an equal probability of occupying any position in society, then the rational choice maximises:

$i = 1 \sum n \frac{1}{n} u_{i} (D),$

which is average utilitarianism. Harsanyi's argument is that maximin is not a rational decision rule under uncertainty: it is appropriate only when one has reasons to believe the worst case is especially likely or especially catastrophic, and neither condition holds behind the veil of ignorance.

Rawls's response distinguishes between risk (where probabilities are known) and uncertainty (where they are not). Behind the veil of ignorance, the parties do not know the probability of occupying any particular position, and the choice involves the fundamental terms of social cooperation for a complete life. Under these conditions, Rawls argues, maximin is the rational strategy: the stakes are too high to gamble on expected utility calculations that depend on unknown probability distributions.

The Rawls-Harsanyi dispute remains unresolved in the literature. It turns on a foundational question in decision theory: what is the rational strategy under radical uncertainty? This is not merely a technical question; it is a philosophical question about what rationality demands when the consequences of a decision structure an entire life.

The capabilities approach: Sen and Nussbaum Master

The most important development in the theory of justice since the Rawls-Nozick debate is the capabilities approach, developed independently by Amartya Sen and Martha Nussbaum. Their shared insight is that neither Rawls's focus on primary goods nor Nozick's focus on rights and entitlements captures what matters most for justice: what people are actually able to do and be.

Sen's critique of Rawls

Sen's Inequality Reexamined (1992) argues that Rawls's theory is built on the wrong "space" -- the wrong metric for evaluating equality and inequality. Rawls measures justice by the distribution of primary goods (rights, liberties, opportunities, income, wealth, and the social bases of self-respect). But two people with the same bundle of primary goods may have vastly different capabilities to convert those goods into valuable functionings. A person with a disability may need more resources to achieve the same level of mobility as an able-bodied person. A person living in a harsh climate may need more fuel to achieve the same level of warmth. If we distribute primary goods equally, we have not achieved equality in any meaningful sense.

Sen's argument is that the appropriate space for evaluating justice is capability -- the set of valuable "functionings" (beings and doings) that a person is able to achieve. Functionings include being well-nourished, being healthy, being educated, participating in community life, and appearing in public without shame. Capability is the freedom to achieve these functionings: not just whether a person actually achieves them, but whether they have the real option to do so.

Formally, Sen defines a person's capability set $C_{i}$ as the set of functioning vectors $f$ that the person can achieve given their resources, personal characteristics, and social environment:

$C_{i} = {f : f is achievable by person i} .$

Justice, on Sen's view, requires expanding and equalising capability sets, not distributing goods. The metric of evaluation is what people can do, not what they have.

Sen deliberately does not specify a fixed list of capabilities. He argues that the selection of relevant capabilities is a matter for public deliberation and democratic discussion, not philosophical fiat. This open-endedness is both a strength (the approach is context-sensitive and democratic) and a weakness (it is unclear how to make the approach operational without specifying what counts).

Nussbaum's list

Nussbaum takes a different route. In Women and Human Development (2000) and subsequent work, she specifies a definite list of ten central capabilities:

Life
Bodily health
Bodily integrity
Senses, imagination, and thought
Emotions
Practical reason
Affiliation
Other species (living with concern for the natural world)
Play
Control over one's environment (political and material)

A society is just, on Nussbaum's view, if it provides all citizens with a threshold level of each capability. Below this threshold, a person is not living a life worthy of human dignity. The threshold is not a ceiling: people may exceed it. But justice requires that no one fall below it.

Nussbaum's list is controversial. Critics argue that it is culturally biased, that it presupposes a liberal-individualist conception of the good, and that the selection of capabilities is philosophically arbitrary without a deeper account of why these ten rather than others. Nussbaum responds that the capabilities are grounded in a cross-cultural account of human dignity and that the list is open to revision through democratic deliberation. But the tension between philosophical grounding and democratic legitimacy is a structural feature of the approach, not a bug to be patched.

The capabilities approach vs. Rawls vs. Nozick

The capabilities approach challenges both Rawls and Nozick from a different direction than each challenges the other.

Against Rawls: primary goods are the wrong metric. Two societies with identical distributions of primary goods can differ dramatically in the real freedoms people enjoy, because of differences in health, disability, geography, and social norms. A just society must attend to the conversion of goods into functionings, not just to the distribution of goods.

Against Nozick: entitlements to holdings are insufficient for justice. A society that respects all property rights and voluntary exchanges can still be deeply unjust if large portions of the population lack the capability to live with dignity. Nozick's theory is formally complete -- if every holding traces to a just acquisition, the distribution is just -- but substantively empty: it cannot recognise injustice that arises from natural disadvantage, social exclusion, or structural deprivation, so long as no one's rights have been violated.

The capabilities approach thus occupies a distinctive position: it shares Rawls's concern for the worst-off and for the material conditions of a decent life, but it shifts the metric from goods to freedoms. It shares Nozick's emphasis on freedom (capability is a kind of freedom), but it insists that formal freedom is insufficient without the material and social conditions that make freedom effective.

Distributive vs. retributive justice Master

The theories discussed so far concern distributive justice -- the just allocation of benefits and burdens in society. But the concept of justice has another major dimension: retributive justice -- the just response to wrongdoing. The two dimensions interact, and a complete theory of justice must address both.

Retributivism and its rivals

Retributivism holds that punishment is justified because the offender deserves it. The severity of punishment should be proportional to the seriousness of the offence. The classic formulation is Kant's: judicial punishment "can never be inflicted merely as a means to promote some other good for the criminal himself or for civil society. It must always be inflicted upon him only because he has committed a crime" (Metaphysics of Morals, 1797).

The rivals of retributivism are consequentialist theories of punishment: deterrence (punishment is justified if it prevents future crimes), rehabilitation (punishment is justified if it reforms the offender), and incapacitation (punishment is justified if it removes dangerous individuals from society). Each consequentialist theory faces the objection that it can justify punishing the innocent if doing so produces good consequences -- a retributivist insists that punishing the innocent is never just, regardless of consequences.

The retributivist position connects to both Rawls and Nozick. Rawls's theory is not directly a theory of punishment, but his framework of fairness -- in which just institutions are those that free and equal persons would agree to -- can support a retributivist conclusion: parties in the original position would agree to a system of proportional punishment because they would want protections against wrongdoing that do not depend on consequentialist calculations that could be turned against them. Nozick's rights-based framework is more straightforwardly retributivist: if individuals have rights that others must not violate, then punishment is the restoration of the moral balance that rights-violations disturb.

The Rawlsian perspective on punishment

Rawls himself addressed punishment in his 1955 paper "Two Concepts of Rules." He distinguished between justifying a practice and justifying a particular action under the practice. The institution of punishment is justified by its social benefits -- deterrence, protection, and the stabilisation of social cooperation. But a particular punishment is justified by the rules of the practice: the person being punished broke a just rule, and the punishment is proportional to the offence. This "rule-utilitarian" account allows Rawls to incorporate consequentialist considerations at the institutional level while maintaining retributivist constraints at the level of individual cases.

The move is characteristic of Rawls's philosophical style: he defuses oppositions by showing that the competing positions are answers to different questions. Retributivism answers the question "When is a particular punishment justified?" Consequentialism answers the question "Why have a system of punishment at all?" The two answers are compatible because they operate at different levels.

Structural injustice and the limits of retributivism

A more recent challenge to retributivism comes from the concept of structural injustice -- injustice that arises not from individual wrongdoing but from the normal operation of social institutions. If a society's economic system systematically deprives a group of education, healthcare, and political voice, the resulting harms are not traceable to any individual's rights-violation. Nozick's framework, which evaluates justice by tracking individual acquisitions and transfers, cannot detect structural injustice. Rawls's framework, which evaluates the basic structure as a whole, can -- but Rawls's focus on the difference principle may not capture the specific harms of structural exclusion that the capabilities approach identifies.

This is one of the major motivations for the capabilities approach: it provides a metric that can detect injustice at the structural level without requiring identification of individual wrongdoers. A society in which women systematically lack the capability for bodily integrity (because of domestic violence, lack of legal protection, or social norms that restrict movement) is unjust on Nussbaum's view, even if no individual rights-violation is being committed. The injustice is in the structure, not in any particular transfer or acquisition.

Cohen's critique of the difference principle Master

G. A. Cohen's Rescuing Justice and Equality (2008) is the most important internal critique of Rawls from within the egalitarian tradition. Cohen argues that Rawls's difference principle, properly understood, does not permit the inequalities Rawls allows.

The incentive argument and its discontents

Rawls allows that talented individuals may be paid incentive salaries -- more than others receive -- if the incentives are necessary to draw out their productive contribution in ways that benefit the worst-off. Surgeons may earn more than teachers if the extra pay attracts enough surgeons to make healthcare available to all.

Cohen's objection: the difference principle, as stated, applies to the basic structure of institutions. But Rawls also argues that citizens of a just society internalise the principles of justice and act on them in their daily lives. If citizens genuinely affirm the difference principle, they would not demand incentives to work in ways that benefit the worst-off. They would contribute their talents without demanding extra pay, because they accept that inequalities are justified only when they benefit the worst-off -- and an inequality that exists only because talented people insist on extra pay does not benefit the worst-off in the right way.

The point is not that talented people should be forced to work for less. It is that a society in which the difference principle is genuinely internalised would not need large incentives, because people would choose to contribute without them. The incentive argument, Cohen contends, exploits a tension in Rawls's theory between the institutional level (where incentives are permitted) and the individual level (where citizens are supposed to be motivated by justice).

The fact of injustice

Cohen's deeper argument is that the difference principle conflates two different things: the rules that should govern the basic structure of society, and the choices that individuals should make within that structure. A just society requires not only just institutions but also a just * ethos* -- a shared understanding that inequality requires moral justification and that those who can contribute more should do so without demanding extra reward.

This argument has been widely criticised. Rawlsians respond that Cohen's position demands too much: a theory of justice should not require saintly motivation, and the difference principle is designed for a society of ordinary people with diverse motivations, not for a society of moral saints. Cohen replies that this response concedes that the difference principle is not, after all, a principle of justice -- it is a principle of social policy that tolerates injustice for the sake of stability. If so, Rawls has not shown that inequality is just when it benefits the worst-off; he has shown only that it is prudent to tolerate it.

The Cohen-Rawls debate is a dispute about the relationship between justice and motivation. Rawls thinks justice must be realistic: it must be achievable by people as they are, not by people as they might ideally be. Cohen thinks justice must be pure: it must specify what is genuinely fair, not what is expedient given human selfishness. The disagreement is not resolvable by argument alone; it turns on a prior commitment about what a theory of justice is for.

Sandel's communitarian critique Master

Michael Sandel's Liberalism and the Limits of Justice (1982) attacks Rawls from a different direction. Sandel argues that the Rawlsian conception of the self -- a self prior to its ends, choosing its commitments from behind the veil of ignorance -- is incoherent. We are not disembodied choosers who happen to have attachments; we are constituted by our attachments, our communities, and our commitments. To ask me to abstract from my conception of the good, my relationships, and my formative experiences in order to choose principles of justice is to ask me to abstract from what makes me who I am.

Sandel's argument has two parts. First, the constitutive claim: our identities are partly constituted by our memberships and commitments -- family, community, nation, religion. These are not attributes we can set aside; they are conditions of our having any identity at all. Second, the normative claim: because our identities are constituted in this way, a theory of justice that abstracts from constitutive attachments cannot be a theory of justice for actual human beings. Justice must be situated within a conception of the good life, not prior to it.

Rawls responded by refining his position. In Political Liberalism (1993), he abandoned the claim that justice as fairness rests on a particular conception of the person (as autonomous and prior to her ends) and argued instead that justice as fairness is a political conception -- one that can be endorsed by citizens with diverse comprehensive doctrines (religious, philosophical, moral) without requiring them to abandon those doctrines. The original position is not a metaphysical claim about the self; it is a device for modelling the fair terms of cooperation among free and equal citizens.

The Sandel-Rawls exchange is foundational for the liberalism-communitarianism debate of the 1980s and 1990s. It raises the question of whether a theory of justice can be neutral among competing conceptions of the good life, or whether any theory of justice necessarily presupposes some conception of the good. Rawls argues for neutrality; Sandel argues that neutrality is impossible and that the attempt to achieve it disguises a particular (individualist, liberal) conception of the good as universal.

Connections Master

Social contract theory [20.01.NN] (pending). Rawls's original position is a modern reconstruction of the social contract tradition (Hobbes, Locke, Rousseau, Kant). The device of the veil of ignorance is Rawls's answer to the question: under what conditions would agreement on principles of justice be genuinely free and fair? The connection runs both ways: understanding Rawls requires understanding the social contract tradition, and understanding the social contract tradition in its contemporary form requires Rawls.
Utilitarianism and consequentialism [20.02.NN] (pending). Rawls defined his theory in explicit opposition to utilitarianism. The Rawls-utilitarian debate on the separateness of persons and the maximin vs. expected-utility question connects directly to the foundations of normative ethics.
Political legitimacy and authority [20.02.NN] (pending). Nozick's minimal state argument -- that a state can arise from anarchy through invisible-hand processes without violating anyone's rights -- connects to broader questions about the legitimacy of political authority and the obligations of citizens.
Feminist philosophy and justice [20.02.NN] (pending). The capabilities approach, especially Nussbaum's version, was developed partly in response to feminist concerns about the failures of traditional justice theories to address gender-based inequality, unpaid domestic labour, and the specific deprivations women face in many societies.
Philosophy of law: rights and obligations [20.02.NN] (pending). Nozick's side-constraint view of rights (rights as constraints on how others may treat you, not as goals to be balanced) connects to debates in legal philosophy about the nature and function of rights.
Economics: welfare economics and social choice [12.NN.NN]. Sen's capabilities approach has been influential in development economics and the design of the Human Development Index. Arrow's impossibility theorem and Sen's liberal paradox connect formal social-choice theory to the philosophical debate about justice.

Cross-domain to philosophy of social science: the measurement problem in justice theory -- how to measure primary goods, capabilities, or utility across persons -- is a live methodological question that connects to debates about commensurability, interpersonal comparison, and the role of indices in social science.

Historical and philosophical context Master

The modern debate about theories of justice begins with the publication of Rawls's A Theory of Justice in 1971, but its antecedents stretch back to Plato's Republic, Aristotle's theory of distributive justice (treat equals equally, unequals unequally in proportion to their merit), and the social contract tradition of Hobbes (Leviathan, 1651), Locke (Second Treatise of Government, 1689), Rousseau (The Social Contract, 1762), and Kant (Groundwork, 1785; Metaphysics of Morals, 1797).

Rawls's innovation was to combine the social contract tradition with a rigorous analytical method. The original position is a Kantian device -- it models what it would mean for the terms of social cooperation to be chosen by free and equal rational beings -- but it is given a specific institutional form through the veil of ignorance and the two principles. Rawls's method of "reflective equilibrium" -- the back-and-forth adjustment between considered moral judgements and theoretical principles until they cohere -- was itself a methodological contribution that changed how analytic philosophy approaches moral theory.

The immediate philosophical context of Rawls's work was the dominance of utilitarianism in mid-twentieth-century Anglo-American moral philosophy. The two major utilitarian systems -- Sidgwick's The Methods of Ethics (1874) and the rule-utilitarianism developed by Brandt and others -- were the default positions against which Rawls defined his alternative. Rawls's claim that the separateness of persons undermines the utilitarian aggregation of utility across individuals was the most philosophically distinctive move in A Theory of Justice and the one that has been most widely accepted, even by philosophers who reject the difference principle.

Nozick's Anarchy, State, and Utopia (1974) was written partly as a response to Rawls and partly as an articulation of the libertarian philosophical tradition that runs from Locke through Bastiat, Spencer, and Rothbard. Nozick's distinctive contribution was to combine Locke's natural-rights framework with a rigorous analytical method and a willingness to follow arguments to their conclusions, however counterintuitive. The entitlement theory, the Wilt Chamberlain argument, and the invisible-hand justification of the minimal state are all exercises in what Nozick called "philosophical explanation": starting from premises that seem obvious or innocuous and showing that they have far-reaching and surprising consequences.

The publication of Nozick's book, just three years after Rawls's, established the Rawls-Nozick axis as the dominant framework for Anglo-American political philosophy. The subsequent literature can be mapped by position relative to this axis: Dworkin's equality of resources (1981, 2000) is a refinement of the Rawlsian side; Sen's capabilities approach (1992) challenges both from a different metric; Cohen's egalitarian critique (2008) pushes Rawls leftward; and the libertarian tradition has been developed by Rothbard, Friedman, and others beyond Nozick's moderate position.

The communitarian critique of the 1980s -- Sandel, MacIntyre (After Virtue, 1981), Taylor (Sources of the Self, 1989), Walzer (Spheres of Justice, 1983) -- challenged the liberal individualism shared by both Rawls and Nozick. The communitarians argued that justice cannot be abstracted from the community, the tradition, and the shared understanding of the good within which moral reasoning takes place. Rawls's response, Political Liberalism (1993), repositioned justice as fairness as a political conception rather than a comprehensive moral doctrine -- one that could be endorsed from within multiple reasonable comprehensive views.

The capabilities approach emerged partly from the failure of both Rawlsian and Nozickian frameworks to address global poverty, gender inequality, and the specific deprivations faced by people in the developing world. Sen's Development as Freedom (1999) argued that development should be understood not as economic growth but as the expansion of human capabilities -- a reconception that has been influential in development policy (the UN Human Development Index is explicitly capability-inspired) and in the philosophical literature on global justice (Pogge, Nussbaum).

The contemporary state of the debate, as of the mid-2020s, involves several active research programmes: (i) luck egalitarianism (Dworkin, Arneson, Cohen), which holds that inequalities are unjust when they arise from brute luck rather than choice; (ii) democratic equality (Anderson), which argues that justice requires relational equality -- equal standing in a democratic society -- rather than distributive equality; (iii) sufficientarianism (Frankfurt, Huseby), which holds that justice requires ensuring that everyone has enough rather than ensuring equality; (iv) global justice extensions of the Rawlsian and capabilities frameworks (Pogge, Nussbaum, Beitz); and (v) the intersection of justice theory with race (Mills, The Racial Contract, 1997), gender (Okin, Justice, Gender, and the Family, 1989), and disability (Nussbaum, Kittay). Each of these programmes takes a position relative to the Rawls-Nozick axis while challenging one or both of its assumptions.

Bibliography Master

Foundational and historical:

Plato -- Republic (~375 BCE).
Aristotle -- Nicomachean Ethics, Book V (~340 BCE).
Hobbes, T. -- Leviathan (1651).
Locke, J. -- Second Treatise of Government (1689).
Rousseau, J.-J. -- The Social Contract (1762).
Kant, I. -- Groundwork of the Metaphysics of Morals (1785); Metaphysics of Morals (1797).
Sidgwick, H. -- The Methods of Ethics (1874; 7th ed. 1907).

Primary contemporary sources:

Rawls, J. -- A Theory of Justice (Harvard University Press, 1971; revised ed. 1999).
Rawls, J. -- Political Liberalism (Columbia University Press, 1993; expanded ed. 2005).
Nozick, R. -- Anarchy, State, and Utopia (Basic Books, 1974).
Harsanyi, J. C. -- "Cardinal utility in welfare economics and in the theory of risk-taking", Journal of Political Economy 61, 434-435 (1953).
Sen, A. -- Inequality Reexamined (Harvard University Press, 1992).
Sen, A. -- Development as Freedom (Oxford University Press, 1999).
Nussbaum, M. -- Women and Human Development: The Capabilities Approach (Cambridge University Press, 2000).
Cohen, G. A. -- Rescuing Justice and Equality (Harvard University Press, 2008).
Dworkin, R. -- "What is equality? Part 1: Equality of welfare" and "Part 2: Equality of resources", Philosophy and Public Affairs 10 (1981).
Dworkin, R. -- Sovereign Virtue: The Theory and Practice of Equality (Harvard University Press, 2000).
Sandel, M. -- Liberalism and the Limits of Justice (Cambridge University Press, 1982; 2nd ed. 1998).
Scanlon, T. M. -- What We Owe to Each Other (Harvard University Press, 1998).

Communitarian and critical responses:

MacIntyre, A. -- After Virtue (University of Notre Dame Press, 1981; 3rd ed. 2007).
Walzer, M. -- Spheres of Justice: A Defense of Pluralism and Equality (Basic Books, 1983).
Taylor, C. -- Sources of the Self: The Making of the Modern Identity (Harvard University Press, 1989).
Okin, S. M. -- Justice, Gender, and the Family (Basic Books, 1989).
Mills, C. W. -- The Racial Contract (Cornell University Press, 1997).

Contemporary developments:

Anderson, E. -- "What is the point of equality?", Ethics 109, 287-337 (1999).
Pogge, T. -- World Poverty and Human Rights (Polity Press, 2002; 2nd ed. 2008).
Frankfurt, H. -- "Equality as a moral ideal", Ethics 98, 21-43 (1987).
Arneson, R. -- "Equality and equal opportunity for welfare", Philosophical Studies 56, 77-93 (1989).

Prerequisites

none — this is a leaf unit

Used in

20.02.02
20.02.03

Tier anchors

beginner: Any intro ethics/political philosophy textbook; Sandel, Justice
intermediate: Rawls, A Theory of Justice (1971); Nozick, Anarchy, State, and Utopia (1974)
master: primary sources: Rawls 1971, Nozick 1974, Sen 1992, Cohen 2008

References

Rawls, J. — A Theory of Justice (Harvard University Press, 1971) · Ch. 1-3, 5
Nozick 1974, Anarchy State and Utopia — to be sourced · Ch. 1-3, 7 · source being verified
Sen 1992, Inequality Reexamined — to be sourced · Ch. 1-5, 8-11 · source being verified
Nussbaum 2000, Women and Human Development — to be sourced · Ch. 1-2, 4 · source being verified
Cohen 2008, Rescuing Justice and Equality — to be sourced · Ch. 1-3, 6 · source being verified
Sandel 1982, Liberalism and the Limits of Justice — to be sourced · Ch. 1-3 · source being verified
Dworkin 1977, Taking Rights Seriously — to be sourced · Ch. 6-7 · source being verified
Scanlon 1998, What We Owe to Each Other — to be sourced · Ch. 5 · source being verified

Estimated time

beginner: 20m
intermediate: 40m
master: 60m