What Justice Requires

A hospital has one place left on a training programme and four applicants.

The first has the highest examination results by a clear margin. The second has worked for six years in difficult conditions to get here while the first was supported throughout. The third comes from the region the programme was set up to serve, where there are almost no doctors. The fourth was wrongly rejected last year through an administrative error that has since been admitted.

Each has a claim, each claim is of a different kind, and no amount of care about the facts will make three of them go away. The previous article said that every serious dispute about equality is a dispute about which differences between people are relevant. This one sets out the candidates.

Three bases for a share

Desert looks backward at the person and asks what they have made of themselves. It is the basis behind the thought that people who work harder should get more, that a prize should go to the best performance, and that a wrongdoer should suffer for the wrong. The intuition is powerful and nearly universal, and it is also the most philosophically embattled of the three, for reasons the aside below sets out.

Merit is regularly conflated with desert and the two come apart sharply. Merit is forward-looking and role-relative. It asks who will do the job well, not who has earned the chance, and the answer can be someone who did nothing to acquire the relevant capacity. A person with an unusual natural aptitude has merit for a role without deserving anything, and a person who worked twice as hard to reach a lower standard may have more desert and less merit.

The distinction explains a familiar argument that otherwise looks like people talking past each other. Appointment on merit is defended as fair because the best-qualified candidate gets the post, and attacked as unfair because qualifications track advantages that the candidate did nothing to earn. Both claims can be true. Merit is the right criterion for allocating a role whose performance matters to third parties, and it was never a claim about what anyone had earned.

Need looks at the person's condition rather than their record. It is the basis behind medical triage, disaster relief, and the provision of a minimum income, and it is the one basis on which the strongest claim can be held by the person who has contributed nothing.

The best-known statement of a needs principle is the formula popularised by Marx in the Critique of the Gotha Programme, written in 1875 and published by Engels in 1891: "from each according to his ability, to each according to his needs". Marx did not coin the phrase, which was already circulating in socialist writing, and he presented it as describing a condition that would become possible only after scarcity had been overcome rather than as a rule for allocating shares under present conditions.

Two difficulties attach to need wherever it is used. Needs must be distinguished from wants, and the line is genuinely hard to draw once anything beyond bare survival is included; a threshold set at subsistence is defensible and almost nobody thinks justice is satisfied by it. And a principle that responds only to need severs the connection between contribution and reward entirely, which the first three applicants above would have something to say about.

Fair procedures and fair outcomes

Everything so far concerns what the distribution should look like. A different family of answers concerns how it came about.

Rawls drew a three-way distinction in A Theory of Justice in 1971 that has become standard, and it is more useful than it first appears.

Perfect procedural justice applies when there is an independent standard for the right outcome and a procedure guaranteed to produce it. Dividing a cake equally is the illustration: let the person who cuts choose last, and the equal division follows.

Imperfect procedural justice applies when there is an independent standard for the right outcome but no procedure guaranteed to reach it. A criminal trial is the case. The right outcome is that the guilty are convicted and the innocent acquitted, this is fixed independently of what the trial produces, and no arrangement of rules and evidence reaches it every time. The best available procedure is followed and its verdicts are sometimes wrong.

Pure procedural justice applies when there is no independent standard at all, and the outcome is just because the procedure was. A lottery is the case. No particular winner was the right winner, and what makes the result fair is that the draw was fair.

The classification matters because a great many political arguments turn on which category a case falls into. If the distribution of income is a matter of pure procedural justice, then whatever emerges from fair transactions is just and there is nothing further to assess. If it is imperfect procedural justice, then there is a standard the outcome can fail to meet, and market results are open to criticism no matter how the transactions were conducted. That disagreement is not about economics; it is about which kind of procedural justice is in play, and it is one of the deepest divisions in the subject.

Aristotle separated this from distribution in Book V of the Nicomachean Ethics and the separation has held. Corrective justice is bilateral, holding between a wrongdoer and the person they wronged, and it takes no interest in how well off either of them is. A rich person who defrauds a poor one owes restitution, and a rich person who defrauds another rich person owes exactly as much. The measure is the wrong, not the resulting distribution.

The fourth applicant in the opening case is making a corrective claim and it should not be assessed alongside the others. Their argument is not that they need the place, or deserve it, or would do the job best. It is that a wrong was done to them and has not been undone, and the remedy is owed by whoever did it regardless of what the distributive picture looks like.

Whether corrective justice is genuinely independent of distributive justice, or whether restoring someone to a prior position presupposes that the prior position was itself just, is a serious question and the advanced tier takes it up.

Formal and substantive

The previous article set out formal equality and showed that it is nearly empty on its own. Substantive equality is what the demand becomes when it is given content: not merely that the rule is the same for everyone, but that people end up standing in something like equal condition.

The gap between them is where most practical argument lives. A rule requiring every applicant to sit the same examination is formally equal and produces results that track years of unequal preparation. A rule setting the same minimum wage everywhere is formally equal and means something quite different in places where the cost of living differs by a factor of three. Formal equality is a constraint on arbitrariness, satisfied by any consistently applied rule, and consistency is compatible with almost any outcome.

Substantive equality is the more demanding notion and it inherits every difficulty from the "equality of what" question in the previous article. It also collides directly with desert. A distribution that responds to how hard people worked will not be substantively equal, since people do not work equally hard, and any principle that insists on equal outcomes has to say what happened to the desert claim.

Next

Three of the four applicants were pressing claims about what they should receive. The fourth was pressing something different: a claim that something was owed to them and withheld. That form of claim has a vocabulary of its own, and it has become the dominant one in political argument. The next article introduces it.