Rawls — justice as fairness

John Rawls's A Theory of Justice (1971) is the most important work of political philosophy of the twentieth century, and the point of departure for almost every debate you will meet in this paper. Rawls revived normative political philosophy from a long period in which questions of justice had been thought either matters of taste or disguised emotion. His ambition was Kantian in scale: to identify principles of justice that free and equal citizens could accept as the fair terms of their cooperation, and to show that these principles are superior to the utilitarianism that had dominated Anglophone moral thought since Bentham and Mill.

The conception Rawls defends is called justice as fairness. Its central device is a thought experiment — the original position — in which rational agents choose principles to govern their society from behind a veil of ignorance that hides their class, talents, sex, race, and conception of the good. Rawls argues that agents so situated would choose two principles: one guaranteeing equal basic liberties, the other — the difference principle — permitting social and economic inequalities only where they work to the benefit of the least advantaged.

The subject of these principles is not individual conduct but what Rawls calls the basic structure of society — the constitution, the economy, the legal system, and the family, taken together as the institutions that distribute rights, duties, and life-chances. This focus on the basic structure, and the priority Rawls gives to liberty, distinguish his liberalism sharply from both the utilitarian and the libertarian alternatives you will study.

This note reconstructs the argument from the original position to the two principles, examines Rawls's later turn to political liberalism, and surveys the major lines of criticism — Nozick's libertarian challenge, G.A. Cohen's egalitarian critique, Sen's capability objection, and the communitarian attack on the Rawlsian self. The debate connects directly to your jurisprudence reading on rights and Raz and on punishment and theories of justice, and it frames the discussion of liberty, authority, and equality in the weeks that follow.

To understand what Rawls was doing, you must understand what he was reacting against. By the middle of the twentieth century the dominant view in moral and political philosophy was a form of utilitarianism: the right action or institution is the one that maximises the aggregate satisfaction of desires, or welfare, across society. Utilitarianism had real strengths — it was systematic, secular, and egalitarian in its insistence that each person counts for one and none for more than one.

But Rawls thought it contained a fatal flaw. By summing satisfactions across persons and seeking the largest total, utilitarianism treats society as if it were a single individual maximising its good over time, and is therefore prepared in principle to impose severe losses on some for the sake of greater gains to others. In his most quoted charge, utilitarianism 'does not take seriously the distinction between persons'. The separateness of persons — the fact that there is no social super-person who experiences a net balance of satisfaction, only distinct individuals each living one life — is the moral bedrock of Rawls's project.

Rawls's alternative drew on the social contract tradition of Locke, Rousseau, and above all Kant, but raised it, in his words, 'to a higher level of abstraction'. The contract is not a historical bargain, nor an agreement to enter society, but a hypothetical agreement on the principles of justice themselves. From Kant he took the idea that just principles are those that free and rational agents would legislate for themselves; the original position is, in effect, a procedural interpretation of Kantian autonomy, stripped of its transcendental apparatus.

It is worth stressing how unfashionable this ambition was in 1971. Anglophone philosophy had largely retreated to meta-ethics and the analysis of moral language, leaving substantive normative questions to one side. Rawls's achievement was to show that principles of justice could be argued for with rigour, and in doing so he almost single-handedly revived political philosophy as a living discipline.

The original position and the veil of ignorance

The heart of justice as fairness is a thought experiment. Rawls asks us to imagine the principles of justice being chosen by rational agents in an original position of equality, tasked with selecting, once and for all, the principles that will govern the basic structure of their society. The fairness of the outcome is guaranteed by the fairness of the conditions of choice — this is the meaning of 'justice as fairness': not that justice and fairness are the same idea, but that the principles of justice are the product of a fair procedure (what Rawls calls pure procedural justice, where there is no independent standard of the right result, only a fair process whose outcome is therefore just).

What makes the procedure fair is the veil of ignorance. The parties choose without knowing any particular facts about themselves: not their class or social status, their natural talents and abilities, their race or sex, their conception of the good or plan of life, nor even their particular psychology — their optimism, pessimism, or attitude to risk. They do know general facts: the laws of economics and psychology, and the 'circumstances of justice' (moderate scarcity and limited mutual concern). The veil's purpose is to nullify the influence of morally arbitrary contingencies: no one should be advantaged or disadvantaged in the choice of principles by 'the outcome of the natural lottery' or by their social starting point, since these are facts for which they can claim no credit.

The parties are assumed to be rational and mutually disinterested: each seeks to advance their own conception of the good, whatever it proves to be, and they are neither envious nor altruistic. Because they do not know their conception of the good, they reason in terms of primary goods — the all-purpose means any rational person is presumed to want whatever else they want: basic rights and liberties, freedom of movement and free choice of occupation, powers and prerogatives of office, income and wealth, and the social bases of self-respect.

Maximin and the choice of principles

How would such parties reason? Rawls argues they would follow a maximin rule — maximise the minimum — selecting the option whose worst possible outcome is better than the worst outcome of every alternative. Behind the veil there is no reliable basis for assigning probabilities to ending up in any position; the stakes are one's entire life-prospects; the guaranteed minimum under his principles is acceptable, while the worst outcomes under, say, average utilitarianism (one might be the persecuted minority whose misery raises the average) are intolerable. Under these specific conditions, Rawls says, maximin is the rational rule — though he does not claim, and you should not say he claims, that maximin is the right way to make decisions under risk in general.

The two principles of justice

The parties, Rawls concludes, would choose two principles, stated in the Restatement (2001) as follows:

First Principle. Each person has the same indefeasible claim to a fully adequate scheme of equal basic liberties, which scheme is compatible with the same scheme of liberties for all.

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

These principles stand in lexical (serial) priority. The first is prior to the second: the basic liberties may not be sacrificed for, or traded against, economic and social advantages. Within the second, fair equality of opportunity is prior to the difference principle. The effect is to place liberty beyond the reach of utilitarian aggregation and to forbid buying greater prosperity at the price of basic freedoms.

The difference principle is the most distinctive and most contested element. It does not require equality: inequalities of income and wealth are permitted, but only insofar as they raise the position of the worst-off group (identified by their index of primary goods) above what it would be under strict equality. A society in which talented entrepreneurs grow rich is just under the difference principle if, and only if, the incentives that enrich them also lift the least advantaged. Inequality is thus licensed as a device for improving everyone's position, never as an end in itself or as a reward for desert — for Rawls insists that even our talents, and the character that lets us cultivate them, are 'arbitrary from a moral point of view'.

Political philosophy has no statutes, but it has canonical texts, and an examiner rewards precise knowledge of which argument belongs to which work. Rawls's thought falls into two phases.

A Theory of Justice (1971; revised edition 1999)

The foundational work. It sets out justice as fairness, the original position, the two principles, and the case against utilitarianism, across three parts: the principles and their derivation; the institutions that would embody them; and a long account of stability, moral psychology, and the 'congruence' of justice and goodness. Note that the revised edition of 1999 corrects the statement of the principles and several arguments; where the wording matters, cite the revised edition. Rawls came to think the Part III argument for stability was flawed, and that dissatisfaction drove the second phase.

Political Liberalism (1993)

Not a repudiation but a reframing. Rawls accepts that A Theory of Justice had presented justice as fairness as part of a comprehensive moral doctrine (a broadly Kantian liberalism), and argues that no such doctrine could be the shared basis of a stable, just society, because free institutions inevitably generate a plurality of reasonable but incompatible 'comprehensive doctrines' — the fact of reasonable pluralism. The book recasts justice as fairness as a political conception: freestanding, drawn from ideas latent in the public political culture, and capable of being the object of an overlapping consensus. It introduces the ideas of public reason and 'the reasonable'.

Justice as Fairness: A Restatement (2001)

Rawls's own late, authoritative summary, edited by Erin Kelly. It gives the cleanest statement of the two principles and clarifies many points — notably that the principles are best realised not by welfare-state capitalism but by a property-owning democracy or a liberal (democratic) socialism: regimes that disperse the ownership of capital and human capital rather than merely redistributing income after the fact. This is the text to cite for the mature view, and the property-owning-democracy point is a reliable way to show an examiner you have read beyond A Theory of Justice.

Two shorter works are worth knowing: The Law of Peoples (1999), which extends the theory to international justice, and the essay 'The Idea of Public Reason Revisited' (1997), his last substantial word on public reason.

Where a law note sets out landmark cases, a philosophy note sets out the key arguments and the devices that carry them. For Rawls, four are essential.

1. The argument from the original position

The original position is meant to be an argument, not merely an intuition pump. Its structure is: (i) a fair procedure for choosing principles is one that excludes morally arbitrary advantages; (ii) the veil of ignorance models exactly such a procedure; (iii) rational parties so situated would choose the two principles, via maximin, over the principle of average utility; (iv) therefore the two principles are the principles of justice. The weight of the theory rests on steps (iii) and (iv), and most criticism targets them — either the veil is said to be too thick (stripping the parties of information they need to choose) or too thin, or the maximin reasoning is said to be irrational, or the conclusion is said to be smuggled into the design of the premises.

2. The veil of ignorance and moral arbitrariness

The deep moral idea beneath the veil is that the distribution of natural talents and social starting-points is arbitrary from a moral point of view: no one deserves the genes or the family they are born into, so no one has a prior moral claim to the rewards those contingencies bring. The difference principle is the institutional expression of this thought: it treats the distribution of natural talents as a common asset, allowing the better-endowed to benefit from their gifts only on terms that also improve the position of the worse-endowed. This severing of distributive shares from desert is among Rawls's most radical and most disputed moves.

3. Reflective equilibrium

Rawls's method matters as much as his conclusions. We test proposed principles against our considered judgments — the moral convictions we hold with confidence and without distorting pressure (that religious persecution and chattel slavery are unjust, for instance). Where principle and judgment conflict, we may revise either, moving back and forth until they cohere. That state is reflective equilibrium; in its 'wide' form it also requires us to weigh the leading alternative conceptions and the arguments for them. Justification in ethics, for Rawls, is holistic and coherentist — not deduction from self-evident axioms, but mutual adjustment toward a considered whole.

4. The Kantian interpretation and the stability argument

Rawls offers a Kantian interpretation of the original position: to act from the principles chosen there is to express our nature as free and equal rational beings, so justice as fairness realises autonomy in something like Kant's sense without his metaphysics. Finally, A Theory of Justice closes with an argument for stability: a society governed by the two principles would tend to generate its own support, because citizens raised under just institutions acquire an effective sense of justice. Rawls's later judgment that this argument failed — because it tacitly assumed that everyone shares one comprehensive doctrine — is precisely what produced Political Liberalism.

The problem of stability and the fact of reasonable pluralism

The shift from A Theory of Justice to Political Liberalism is the development you are most likely to be examined on, and the most often misunderstood. Its driver is the problem of stability. In a free society, Rawls observes, human reason working under free institutions will always produce a plurality of comprehensive doctrines — religious, philosophical, and moral worldviews — that are mutually incompatible yet each reasonable. This is not a regrettable fact to be overcome but 'the normal result of the exercise of human reason within the framework of the free institutions of a constitutional democratic regime'. Call it the fact of reasonable pluralism.

If that is so, a conception of justice grounded in any single comprehensive doctrine — even Rawls's own Kantian liberalism — cannot serve as the shared basis of a just and stable society, because reasonable citizens will reasonably reject that doctrine. A Theory of Justice had quietly assumed a well-ordered society united on one comprehensive view; Rawls came to see this as inconsistent with the very liberty his theory protects.

The political conception and overlapping consensus

The solution is to make justice as fairness a political rather than a comprehensive conception. A political conception is (i) freestanding — presented on its own, not derived from any particular religious or philosophical doctrine; (ii) limited in scope to the basic structure; and (iii) worked up from fundamental ideas implicit in the public political culture of a democracy, above all the idea of citizens as free and equal and of society as a fair system of cooperation.

Stability is then secured by an overlapping consensus: the various reasonable comprehensive doctrines each endorse the political conception from within their own premises — the Kantian for reasons of autonomy, the utilitarian for reasons of welfare, the religious believer as an expression of their faith's respect for persons. Because each affirms it for its own reasons, the consensus is stable 'for the right reasons', and is not a mere modus vivendi — a temporary balance of forces that would dissolve the instant the balance shifted.

Public reason

The companion idea is public reason. When citizens, and especially officials, debate constitutional essentials and matters of basic justice, they should appeal to reasons and values that other reasonable citizens could reasonably be expected to accept, rather than to the whole truth as their own comprehensive doctrine sees it. This 'duty of civility' is moral, not legal, and Rawls's later 'wide view' permits comprehensive reasons to be introduced provided that properly public reasons are given in due course. Critics abound — that it muzzles religious citizens, or is self-undermining — and public reason is now a large literature in its own right. Flag it, but keep it subordinate to the political/comprehensive distinction in an exam.

Rawls's theory is the hub around which post-1971 political philosophy turns; learn the critics as systematically as the theory.

The libertarian critique: Nozick

In Anarchy, State, and Utopia (1974), Robert Nozick attacks all 'patterned' and 'end-state' principles of justice, the difference principle among them. His entitlement theory holds that a distribution is just if it arose through just original acquisition and just voluntary transfer, whatever pattern results; justice is historical, not a matter of fit to any formula. The Wilt Chamberlain argument is the engine: begin from any distribution you regard as just (say, equality), let many people each freely pay a basketball star a small sum to watch him play, and you reach a new, unequal distribution that no one can complain of — yet it breaches the original pattern. 'Liberty upsets patterns', so maintaining a pattern like the difference principle requires 'continuous interference with people's lives'. Nozick adds that taxation of earnings from labour is 'on a par with forced labour', grounded in a thesis of self-ownership. You take up Nozick in his own right next week.

The egalitarian critique: G.A. Cohen

G.A. Cohen's critique (in Rescuing Justice and Equality, 2008, building on earlier essays) comes from the left and is internal. The difference principle permits inequalities that benefit the worst-off — typically, incentives to the talented. But, Cohen argues, the talented only need incentives because they are unwilling to work hard without extra pay; in a society whose members genuinely shared an egalitarian ethos, the talented would not make such demands, and the inequalities the difference principle licenses would be unnecessary. Cohen therefore rejects Rawls's basic-structure objection — the claim that principles of justice govern institutions only, not personal choices — and insists that justice reaches into the daily choices of individuals. His provocation, and the title of his 2000 Gifford Lectures: 'If you're an egalitarian, how come you're so rich?'

The capability critique: Sen

Amartya Sen ('Equality of What?', 1979; Inequality Reexamined, 1992) attacks Rawls's metric. Primary goods are means, but people differ in their ability to convert means into actual functioning: a person with a disability may need far more income to achieve the same mobility. Justice, Sen argues, should attend to capabilities — the real freedoms people have to achieve valuable 'functionings' — not to the bundle of primary goods they hold. This launches the capability approach later developed by Martha Nussbaum, and feeds straight into next week's question of 'equality of what?'.

The communitarian and feminist critiques

Michael Sandel (Liberalism and the Limits of Justice, 1982) argues that the self behind the veil is an implausibly 'unencumbered' self, abstracted from the attachments and ends that in fact constitute identity; the original position smuggles in a contested liberal metaphysics of the person. Michael Walzer (Spheres of Justice, 1983) urges that justice is plural and culturally situated — different 'social goods' belong to different distributive spheres — and cannot be read off a single universal procedure. Susan Moller Okin (Justice, Gender, and the Family, 1989) presses a feminist objection: the family is part of the basic structure, yet Rawls barely subjects it to his principles; the parties in the original position must not know their sex, and a consistent Rawlsian must confront injustice within the household.

Finally, H.L.A. Hart's essay 'Rawls on Liberty and its Priority' (1973) presses an internal point you already meet in jurisprudence: Rawls's argument for the lexical priority of liberty is underspecified and gives no clear way to adjudicate conflicts among the basic liberties. Rawls accepted much of this and reworked the argument in the relevant lecture of Political Liberalism.

It helps to locate justice as fairness on the wider map of distributive theories.

Against utilitarianism, Rawls's decisive charge is the separateness of persons: by maximising an aggregate, utilitarianism can in principle justify grave burdens on a minority, whereas the lexical priority of liberty and the difference principle erect protections that aggregation cannot override. The price Rawls pays is that the difference principle attends only to the worst-off group, and is in that sense insensitive to large gains spread across everyone else.

Against libertarianism (Nozick), Rawls treats the distribution of talents as morally arbitrary and hence as a common asset, while Nozick treats it as the property of the individual who bears it. The dispute turns on self-ownership and on whether justice is patterned or purely historical.

Against the capability approach (Sen, Nussbaum), Rawls measures advantage in primary goods — objective, public, and readily indexed — whereas capability theorists measure it in real freedoms to function, capturing human diversity at the cost of a more contestable and harder-to-administer metric.

Finally, the luck egalitarians (Dworkin, Arneson, Cohen) can be read as radicalising Rawls's own insight about moral arbitrariness: if it is unfair for life-chances to track unchosen luck, justice should neutralise the effects of brute luck while holding people responsible for their genuine choices. Whether this is a friendly extension of Rawls or a rival to him is exactly the issue next week takes up. Seeing justice as fairness as one option among these four sharpens both its appeal and its costs.

Question: 'Why does Rawls think the parties in the original position would choose his two principles rather than the principle of average utility? Is he right?'

Introduction

The original position is engineered so that its output carries justificatory force; everything therefore turns on what rational parties behind the veil would actually choose. Rawls claims they would choose his two principles over the principle of average utility — the principle directing them to select the basic structure that maximises average well-being per head. The difficulty, pressed most sharply by Harsanyi, is that the very same device can be made to yield utilitarianism. This essay reconstructs Rawls's argument — the maximin rule and the three conditions he says make it rational — sets against it the expected-utility rival, and argues that although the narrow decision-theoretic case for maximin does not succeed, the broader argument from the separateness of persons, self-respect, and the strains of commitment secures most of what Rawls wants.

The choice and the two candidates

Behind the veil the parties choose principles to govern the basic structure, ignorant of their identity and reasoning in terms of primary goods. Two candidate conceptions confront them: Rawls's two principles, with the priority of liberty and the difference principle; and the principle of average utility, which ranks basic structures by the average level of welfare they produce. The question is which a rational, mutually disinterested party would prefer.

Rawls's maximin argument

Rawls argues that the parties would adopt a maximin rule: rank the alternatives by their worst possible outcomes and choose the one whose worst outcome is least bad. Under his principles the worst case is to be among the least advantaged, whose position the difference principle maximises; under average utility the worst case is to be the minority whose suffering raises the mean. Maximin is normally a very cautious rule, so Rawls specifies three conditions under which it is nonetheless the rational one: (i) the parties have no sound basis for assigning probabilities to occupying any particular position, and so cannot responsibly compute expected utility; (ii) the minimum guaranteed by the two principles is satisfactory, so the parties care little about possible gains above it; and (iii) the rejected alternatives carry outcomes the parties could not accept. Where these hold, staking one's entire life-prospects on an unsupported probability would be irrational.

The Harsanyi rival

John Harsanyi turns the device against Rawls. A rational chooser behind the veil, he argues, should assign equal probability to being any member of society (the principle of insufficient reason) and then maximise expected utility — which yields average utilitarianism. So the contractarian apparatus, far from refuting utilitarianism, can ground it. The dispute reduces to a question of rational choice under uncertainty: is it rational to use equiprobability and maximise expected utility, or to refuse probabilities and apply maximin?

Assessing the dispute

Three points decide it. First, on probabilities, Rawls denies the parties may invoke insufficient reason: the choice is unique and unrepeatable — not a bet one can make many times so that the averages come good, but one's single life — and the assumed probabilities would be groundless. This is plausible, but contested: a one-shot decision can still rationally maximise expected utility where the probabilities are warranted, and insufficient reason is a respectable principle.

Second, and most damagingly, the risk-aversion objection. Maximin presupposes a markedly conservative, risk-averse attitude to one's prospects. But Rawls's own veil excludes knowledge of one's attitude to risk. He therefore cannot help himself to the assumption that the parties are risk-averse without contradicting his own construction: either he builds risk-aversion in — in which case the choice of the two principles is rigged — or he does not, in which case nothing favours maximin over expected-utility maximisation. This objection, pressed by Hare and others, is widely thought to sink the decision-theoretic reading of the argument.

Third, Rawls's better reply is that the case need not rest on psychological timidity at all, but on structural features of the choice: it fixes one's life-prospects irrevocably, and there is a qualitative difference between securing an acceptable minimum and gambling it away for marginal gains above it.

The deeper, non-decision-theoretic argument

Rawls's most durable objections to average utility are not about decision theory. The first is the separateness of persons: by aggregating, average utility treats the losses of some as fully compensable by the gains of others, which parties each concerned for their own single life have decisive reason to refuse. The second is the strains of commitment: the parties must choose principles they can actually honour once the veil lifts and they learn their place, and principles that might consign them to misery for the sake of the average impose strains no one could in good faith undertake. The third is self-respect and publicity: the public knowledge that inequalities are arranged to benefit the worst-off supports the social bases of self-respect in a way utilitarianism cannot guarantee. None of these needs maximin; each explains why parties who take seriously that they are separate persons, each living one life, would refuse to wager that life on maximising an average.

Judgment

The narrow claim — that maximin is the uniquely rational rule of choice — fails, because Rawls cannot consistently attribute risk-aversion to parties he has stripped of knowledge of their risk attitudes, and Harsanyi's model is at least as defensible as a piece of decision theory; tellingly, Rawls himself leaned less on maximin in later work. But the conclusion — that the parties would reject average utility — survives, carried by the separateness of persons, the strains of commitment, and the concern for self-respect, which are moral rather than decision-theoretic considerations and which aggregation genuinely cannot accommodate. Rawls is therefore right that his parties would not choose average utility, though for reasons better than the maximin argument with which he first made the case. The wider lesson — a recurring one in this paper — is that the original position does not generate substantive conclusions mechanically; it dramatises moral convictions, here the irreducible separateness of persons, that must in the end be defended on their own merits.

1. Confusing the veil of ignorance with the original position. The original position is the whole hypothetical choice situation; the veil of ignorance is one feature of it (the information constraint). Examiners notice the slip immediately.

2. Describing the difference principle as 'equality'. It is not. The difference principle permits inequality, provided the inequality raises the position of the worst-off above what strict equality would give them. Nor is it utilitarian: it attends to the least advantaged, not to the aggregate or the average.

3. Forgetting the lexical priority. The first principle (equal basic liberties) is prior to the second, and within the second, fair equality of opportunity is prior to the difference principle. Treating the principles as a single balancing exercise misrepresents the theory; liberty cannot be traded for economic gain (in favourable conditions).

4. Saying Rawls endorses the welfare state. In the Restatement he is explicit that welfare-state capitalism does not realise the two principles; he favours a property-owning democracy or a liberal socialism that disperses ownership of capital, not mere after-the-fact redistribution. Citing this marks out a strong script.

5. Overstating maximin. Rawls does not claim maximin is the rational rule for decisions under risk in general — only under the three special conditions of the original position. Attributing to him a general theory of conservative choice invites the easy (and unfair) objection that maximin is usually irrational.

6. Collapsing early and late Rawls. A Theory of Justice presents justice as fairness as part of a comprehensive doctrine; Political Liberalism recasts it as a freestanding political conception in response to the fact of reasonable pluralism. And an overlapping consensus is not a modus vivendi. Blurring the two phases is the single most common error on Rawls.

7. Talking about desert. Rawls denies pre-institutional desert: natural talents are 'arbitrary from a moral point of view', so the difference principle does not reward the deserving. Saying the better-off 'deserve' their rewards under Rawls gets him backwards.

The practice questions for this topic — with model answers and mark schemes across foundation, standard, and challenge tiers — are set out below. Attempt each under timed conditions before revealing the model answer, then mark your work against the scheme.

The primary texts and the best secondary literature are listed below. For a first pass, read Rawls's Restatement (the cleanest statement of the mature view) alongside Freeman's Rawls as a guide, then turn to Nozick, Cohen, and Sen for the major critiques.