- The Idea of Justice by Amartya Sen
Allen Lane, 468 pp, £25.00, July 2009, ISBN 978 1 84614 147 8
At some time in the past the idea took hold that social justice was all about the state’s hoovering up resources and then blowing them at needy or deserving recipients. Some of these resources, money for example, were material, and others, like opportunities, were virtual. But there were various problems. One was that there had to be just one hoover, the state, sucking in goods, although other providers, being plural, threatened to disrupt the favoured distribution by shunting them around for purposes of their own. Some, and not only libertarians, wondered whether the state’s cornering the market in resources might not frustrate the point of social justice itself – or even directly negate it, if for example the point was to make people autonomous. Then the resources had to be treated as homogeneous to satisfy the principles that were meant to justify redistributing them.
This puréeing of resources also made it hard to keep sight of the ends that gave social justice its point. For example, if the purée is cash, and the aim is to give each person as good a life as possible, this end may not be well served by handing a person whatever income fills the gap between her actual earnings and, say, half the national average, or some other benchmark. As Amartya Sen has shown elsewhere, it’s not what you have that counts, so much as what you can do with it. So the currency of justice should be capabilities. But it is far from obvious how to distribute capabilities justly.
But why should one have bothered about social justice anyway? Although the most obvious reason to worry about who got what was its effect on welfare – how well people’s lives went – quite a lot of things that affected welfare couldn’t be parcelled out. For instance, charm, good looks, artistic talent, a winning personality and so forth at best allow only of very limited redistribution. Still, someone may say, that doesn’t mean that what can be redistributed shouldn’t be; and even those who suffer non-remediable conditions that make their lives worse could be tossed the odd bone by way of indemnity. Here, though, the usual questions about fungibility arise. How much is a physical handicap worth, for example? And the very notion of welfare as a generic good to be shunted around according to some notion of justice is problematic. Is each person the best judge of his or her welfare? Or should your welfare be gauged by what someone else, such as a rational and benevolent spectator, might want for you?
In answering these questions, The Idea of Justice revives several debates to which its author has made significant contributions. The book, indeed, has a slight air of Amartya Sen’s Greatest Hits about it. For instance, he reprises his 1970 proof of the impossibility of a Paretian liberal. According to the principles of Pareto efficiency, given two distributions, Y and Z, if nobody does worse in Y than in Z, and at least one person does better, Y is preferred to Z. This seems plausible. Sen showed, however, that it runs afoul of the minimum demands of liberty, taken as the claim that each person should be decisive over at least one matter, such as deciding what books one can look at. In Sen’s celebrated example, Lewd wants to read a pornographic book and Prude doesn’t. But if only one person is going to read it, Lewd prefers that it be Prude (he wants Prude to be agonised), while Prude thinks that it should be he, Prude (he is worried that Lewd will chuckle over the book). Prude’s ranking of preferences is neither>Prude>Lewd>both, whereas Lewd’s is both>Prude>Lewd>neither. It then follows by Pareto that Prude’s reading the book is preferred to Lewd’s reading it, since each prefers this, and nobody prefers the reverse – even though Prude would rather not read it. Hence liberalism clashes with Pareto efficiency.
But this is rather odd. Certainly, there is nothing to stop Lewd from preferring that Prude be forced to read the book, to reading it himself. But in that case, Lewd has got what he wants: the satisfaction of his other-regarding preference that Prude, willy-nilly, should read it. If liberalism means satisfying the preferences that people actually have, then it should put other-regarding and self-regarding preferences on the same footing; conversely, if liberalism rules other-regarding preferences out, then it is those preferences, rather than Pareto, which conflict with liberalism. Anyway, it’s hard to see why Sen revisits the debate at all: he says that it is ‘a contribution to public discussion’.
Sen has indeed contributed strenuously to public discussion. He is a heroic figure who has published major theoretical works in welfare economics, social choice theory and political philosophy, but has also been politically engaged. He remains prominent in public life in the UK and elsewhere, as a trustee of the Nalanda University project, a member of the UK National Security Forum and sometime co-chair of the Commission on Human Security. Sen also devised a widely used index for measuring poverty and has written about Rabindranath Tagore, a family friend. Accordingly, the material in this new book is exhilaratingly broad. He discusses human rights, the Mahabharata, bounded rationality, development aid, the Bengal famine of 1943, cognitive psychology and risk aversion, discursive democracy, equality, liberty, Adam Smith’s Theory of Moral Sentiments, medieval Arabian astronomy and a lot more. He brings this off with erudition, not a little wit, and the odd guffaw: ‘To admit to being a “consequentialist” is almost like introducing oneself by saying, “I am a wog.”’
As a result, perhaps, the book’s narrative voice is peripatetic and long on anecdote. Indeed, the caliginous aura of the high table, somewhere between the departure of the widgeon and the arrival of the Sauternes, hangs over much of the book. We learn what W.V. Quine wrote about the word ‘solstice’ in a personal letter to Sen, what John Sparrow said about the Good Samaritan parable over dinner at All Souls, and why Piero Sraffa used to rub his chin when chatting to Wittgenstein at Trinity. No doubt each of these tales forms a filament, however frail, in the pendant damask of our human story. But after a while the reader may wonder whether, amid the reminiscence and the engaging but heteroclite reflections, any distinctive theory is on offer.
The theory is buried in much other matter, but it is there. Political philosophy errs in formulating ideal principles of justice, which afford little help in resolving competing claims in our non-ideal world. In this world, the claims of justice are plural, in that people can often make competing claims on scarce resources with some show of plausibility. So there is little prospect that ideal principles will help resolve the dispute. Instead, insofar as a resolution is possible at all, it has to rely on public reason, which in modern political theory serves as a dialectical footbath, purifying the reasons that are put into the public realm and sterilising, in particular, the verruca of self-interest. Sen acknowledges that these reasons are plural, in that even if attention is confined to justice, it may remain uncertain what should be done. He gives the example of flautism. Suppose one asks who should get a flute: the person who made it, someone else who can already play it, or a third person who would like it, having no toys of his own? In the abstract, a question like this has no generally valid answer. But Sen argues that in situ one can reach an answer justified by the public reasons on offer, even if one cannot produce a complete ranking of all the options.
Sen is addressing this question: how are principles of justice related to the concrete circumstances in which just acts take place? On one view, the relation is like – indeed, is an instance of – the relation that theory bears to practice. Those who like to make the distinction sharp, including not a few academics, see theory as ‘prior’ to practice. Anybody could come up with principles of justice, and then imagine acting on them, perhaps in some utopia. Plato, whose thoughts about justice have often been read in this way, imagines Dike¯ , the personified figure of Justice in the Laws, tagging along behind the Almighty and meting out condign punishment to malefactors. In similar vein, the late G.A. Cohen argued that ‘principles’ were ‘prior’ to facts: any given fact will have practical implications only if there is some principle which tells us why we should pay heed to that fact. For instance, someone who thinks it makes a difference, when parcelling out slices of cake, whether or not the recipients are fat, must rely on some principle which tells cake-cutters, say, that in situations like this, the fatties should snag the bigger (or smaller) slices.
By contrast, one may affirm instead the primacy of practice over principle, as Sen does. This approach has a long ancestry. In his discussion of justice in the Nicomachean Ethics, Aristotle mentions something he calls the ‘Lesbian rule’. Students can get excited when this PowerPoint slide comes up but, crushingly, it turns out that Aristotle is thinking not of a body politic subject to correction by Dike¯ , but of stonemasons on Lesbos, who apparently used a bendy ruler made of lead for measuring curved window mouldings and the like. Bendy rulers adapt themselves to the shape of the material, and that, presumably, is part of Aristotle’s point. The absolute size and ratio of just distributive shares vary with context. More radically, the very metric by which the shares get parcelled out varies with context.
For Sen, the hunt for ideal principles of justice misses the point, because justice has to be implemented in non-ideal circumstances. He gets to this conclusion via a very short argument, of the following form. Suppose that Anna Karenina is the ideal novel; this knowledge doesn’t put anyone in a position to appraise other, non-ideal books. Knowing that the Tolstoy is the ideal doesn’t help in deciding whether a Joanna Trollope surpasses a Philip K. Dick in literary merit. And so, Sen concludes, with justice: we don’t need a grasp of ideal justice to know how to act justly in the non-ideal world.
Sen frames the prime contrast as between niti and nyaya, two Sanskrit terms for justice, which he renders, less pithily, as ‘transcendental institutionalism’ and ‘realisation-focused comparison’ respectively. The distinction comes out in the question: should theory concern only the design of ideal institutions, procedures and principles, or should it take account of how the world actually is? But this runs together two distinctions: between ideal and non-ideal theories of justice; and between theories which are institution-based and those which aren’t. The two distinctions are distinct. Someone could elect to design ideal institutions from scratch, but someone else might heed real-world imperfections, say by institutionalising complaints procedures. If niti and nyaya are respectively ideal and non-ideal approaches to justice, institutions are incidental.
As a result, the intended contrast proves elusive, as does what is at stake (Sen complicates matters further by drawing a parallel between this distinction and the debate between Krishna and Arjuna in the Bhagavad Gita). For clarity, let’s stick with the Sanskrit. Niti is encapsulated for Sen by the idea that justice should be done regardless of the consequences. He cites the Holy Roman Emperor Ferdinand I’s dictum, ‘fiat iustitia, et pereat mundus.’ When this tag surfaces in moral philosophy, as in debates over Kant’s view that one shouldn’t lie to a would-be murderer regarding the whereabouts of his intended victim, the opposing idea is usually consequentialism. But nyaya seems not to be simply a morality of consequences, which anyway may be as idealised as one could wish; while Kantians – anti-consequentialists par excellence – could be called ‘realisation-focused’, in not lying to the axe murderer. So the niti/nyaya contrast is not between a morality of right and one of consequences.
If we take the analogy with novels seriously, the idea of nyaya seems instead to be that one can, at least grosso modo, identify actual injustices such as slavery without a notion of an ideally just society. One objection to this would be that it mistakes complements for alternatives. It may be that one can identify slavery as wrong without recourse to ideal principles, but that doesn’t mean there’s anything wrong with ideal theory. Is it so wrong to focus on the ideally just world? One could answer that it depends what you are trying to do. After all, there are purely aesthetic versions of utopianism, such as News from Nowhere. But the point Sen wants to press is that if one is interested in applicability, knowing the ideal may not help much.
Someone may still object that one can’t rank sub-ideal specimens without knowing the ideal. In the novels example, it may be said, one needs some principle of appraisal in judging that the Trollope beats the Dick. According to this principle, the ideal novel maxes out on some property – psychological depth for instance – and so a novel with a given amount of this must beat any novel with less of it. Whoever is in a position to say that a novel is ideal must do so on the basis of criteria that will, at least in principle, enable him to rank those that fall short of the ideal.
Suppose, similarly, that justice is top dog among political values, the first virtue of societies etc, and further, that only one property is relevant to justice. Someone might think, say, that justice is all about equality, and infer that the more equal some set of arrangements is, the more just it is: one should ignore every other consideration, such as desert, merit, absolute measures of welfare and so on. So, in comparing two sub-ideal distributions, the one with less inequality must be the more just one. And for that, the objection goes, one needs a grasp of ideal equality.
But is it the case that more equality must be more just than less? Prioritarians favour a society in which the worst-off do better than they do anywhere else, sufficientarians one in which more people have enough: on some ways of measuring overall equality, each may prefer the slightly more unequal one. It will immediately be said that here the ground has shifted from justice qua equality, to welfare. But as debates about priority and sufficiency suggest, it’s when the pure principle is at issue that the grounds for endorsing it become significant. What makes equality just? The rationale for justice qua equality will rest either on nothing – which looks like dogma – or on a concern with something else, like people’s moral equality as expressed by their life prospects. The justice that demands equality will require other inequalities to achieve it.
The real world will be reflected imperfectly, if at all, by the ideal. Sen’s guiding thought is that one can, here and now, make comparative judgments without ranking all conceivable options. Quite often, political questions come down, basically, to ‘yes’ or ‘no’. The decisive judgment may be that some practice, like slavery, exemplifies injustice, rather than that such practices deviate from an already well-defined template of justice.
Suppose that more than one property goes to deciding if an arrangement is just. As Sen says, one arrangement may do better on one count, and another on another, so one may not know which is more just, all things considered. The idea that we cannot fully rank different arrangements relies either on scepticism about relative weightings, or on the claim that the different goods are incommensurable, that is to say, cannot both be measured on a single scale of value. Sen is a bit rude about incommensurability: he dismisses as ‘feeble’ the idea, often voiced by value-pluralists, that one might confront incommensurability in having to choose, say, between a Hazel Blears speech and a poke in the eye. So presumably Sen favours scepticism. But scepticism seems not to do the job: then one can’t say that, in deciding justice here and now, one needs a knowledge of ideal justice.
Sen’s implicit argument might be put this way. Justice is a practical matter. Feasibility is salient in practical matters: in thinking about what to do, the realisation that some action is practically impossible closes off that branch of the deliberative tree. So any mooted principles of justice that fly in the face of feasibility are not principles of justice at all; they are rather, in the pejorative sense, utopian, since they wish away practical constraints which will be with us not simply in the world as it is, but in any world one could sanely hope to reach. To take the most obvious example, distributive justice assumes, as Hume noted, relative scarcity: if everybody had as much as anyone could desire, the question of justice wouldn’t arise. If so, the problem of justice can be solved by imagining scarcity away. Naturally, that looks like a cheat. But imagining away non-compliance, lack of co-ordination, sub-prime motivation – and disagreement about justice itself – also seems like a cheat. Sen’s overriding point is that these worries can be put off by asking what it is best to do in the here and now, given what can be done.
Sen, then, calls for nyaya over niti. Unsurprisingly, his work on capabilities, developed with his sometime collaborator Martha Nussbaum, resurfaces in this connection. The original rationale, plausibly enough, lay in concern that the resource-based approach to distributive justice largely ignored the net impact which getting a particular handout has on its recipient’s life. One should look at capabilities net of resource inputs, rather than worrying only about who gets what. Take, for example, a disabled person on a high salary: using earnings to measure wellbeing overlooks the functional limitations that the disability imposes on him. This case seems straightforward. But generalising it to yield a basis for distributive justice is far from simple.
What is a capability? Sen’s earlier work on this bore the firm imprint, via Nussbaum, of Aristotle, who takes certain achieved capabilities as the mark of human flourishing. The Metaphysics distinguishes a dunamis, or capacity, from an achieved capability, or entelechia. Aristotle says the relation of entelechia to dunamis is like that between someone now using her capacity to see, and a normally sighted person whose eyes are shut. Sen sees capabilities as dunameis. For him, what matters is not just what someone actually does, but what they could do. But ‘could’ comes in different strengths: is a dunamis more like an ability which, though currently dormant, could be exercised presently, like my ability to drive; or more like a potential, which could be realised only with a lot of pump-priming, like my tragically undeveloped potential to speak Finnish? As far as resource allocation goes, it presumably makes sense to create abilities rather than unrealised potentialities, as the latter can only benefit their possessor if they are converted into abilities that can be exercised securely, as Jo Wolff and Avner de-Shalit underline in their book Disadvantage. They use Sen’s example of wild honey gatherers in West Bengal, who run a significant risk of being eaten by tigers. Those who gather do so from economic necessity, and the fact that they are capable of gathering the honey may not offer much consolation. They are disadvantaged by having to seek their income at risk of death.
So it seems that it is secure abilities which the capabilities approach should treat as the currency of justice. This still leaves some big questions. While education sometimes inculcates generic abilities like numeracy and literacy, often it has more specific capacities in view, such as bagpipe tuition. Then there are capacities that the state probably won’t fund at all, like that for flea-training. Not everyone will get help from the state to develop her preferred capacities. The state has to take a view about what kinds of life are better, in violation of modern liberals’ belief in neutrality, according to which the state should take no position about the nature of the good life.
Then there is the fact that for some people, lack of function is a desideratum. They include not just mystic fasters and voluntary amputees, but also the millions of normals who run to fat or submit themselves to a blitz of drugs, drink and junk TV. On each of these counts, Sen’s best bet is the basically Aristotelian position that there are generic capacities which the good human life displays, and the state can try to instil these through education, even though it can’t guarantee that people won’t succumb to akrasia and cretinise themselves.
Sen could be accused of echoing conservatives who treat theory with impatience, and assert the primacy of practice. Occasionally, indeed, conservatives adopt a pragmatist bone-headedness – stolidity and want of imagination – as a pose. But this pose need not be all that pragmatism amounts to. Insofar as conservatism is not simply the flag of convenience in which an otherwise naked status quo wraps itself, it makes extant practices the launch pad for future action, as conservative thinkers like Michael Oakeshott argue. That will mean, indeed require, asking how far those practices need shaking up. But doing so leaves open all the options that are open. If it says otherwise, conservatism isn’t just pretending to be bone-headed.
The Idea of Justice valuably adjusts justice theory’s preoccupation with principles and institutional design. Orthodox theorists will not like the focus on practice-dependency. It leaves space for the charge – which might be worn as a badge of honour – that the account of justice is imperfectly theorised, or that talk of justice makes sense only where the practices already exist. It also invites the objection that practice has eclipsed principle entirely, and that there is nothing but a glut of localised norms – what Martin Hollis used to call ‘liberalism for liberals, cannibalism for cannibals’. How far one feels bothered by this is no doubt a personal matter. The slope at whose bottom the pit of moral relativism froths has never wanted for lubrication by op-ed journalists and academics. But, as often, relativism is beside the point. Even if beliefs about justice were globally uniform, there would still be a question about whether practice should take its bearings from an ideal set of principles.
What price justice? It is worth remembering that, like injustice, it does have a price. As Sen implies, the price is not usefully paid in the specie of eternity, an idea behind some psychopathically meliorist projects of modern times. There is always a political question about whether and why to take an interest in something – even the justitia whose fiat has been joy-ridden by philosophers on behalf of an absconded deity.