The Idea of Justice

by Amartya Sen

  And here we do come to a parting of ways. In contrast with such institutional approaches, there are theories of justice and of social choice that take extensive note of the social states that actually emerge in order to assess how things are going and whether the arrangements can be seen as just. Utilitarianism takes such a view (even though its assessment of social states is confined to the limited perspective of utilities generated, ignoring everything else), but much more generally, so does social choice theory as an approach to evaluation and justice, as explored in a framework established by Kenneth Arrow, broadly in line with the normative approaches explored by Condorcet and Adam Smith, among others. There is no necessity here to rely only on utilities for the assessment of states of affairs, or, for that matter, only on ‘end states’ (as Robert Nozick calls them), ignoring the huge significance of the processes used. Rather, the comprehensive states of affairs that actually emerge are seen to be critically important in assessing whether we are doing the right thing, or could do better.

  In the inclusive perspective of nyaya, we can never simply hand over the task of justice to some niti of social institutions and social rules that we see as exactly right, and then rest there, and be free from further social assessment (not to mention anything like ‘freedom from morality’, to use David Gauthier’s colourful phrase). To ask how things are going and whether they can be improved is a constant and inescapable part of the pursuit of justice.

  86

  4

  Voice and Social Choice

  As Alexander the Great roamed around north-west India in 325 bc, he engaged in a series of battles against the local kings in and around Punjab and won them all. But he was not able to generate enthusiasm among his soldiers to take on the powerful Nanda imperial family that ruled over the bulk of India from their capital city Pataliputra in eastern India (now called Patna). Alexander was not, however, ready to return quietly to Greece, and as a good student of Aristotle spent some considerable time holding relaxed conversations with Indian philosophers and theorists – religious as well as social.*

  In one of the more vigorous debates, the world conquerer asked a group of Jain philosophers why they were neglecting to pay any attention to him. To this question, he received the following broadly democratic reply:

  King Alexander, every man can possess only so much of the earth’s surface as this we are standing on. You are but human like the rest of us, save that you are always busy and up to no good, travelling so many miles from your home, a nuisance to yourself and to others! . . . You will soon be dead, and then you will own just as much of the earth as will suffice to bury you.1

  * India was full of intellectual heterodoxy at that time, roughly the period when the great epics, the Ramayana (in particular the Valmiki Ramayana) and the Mahabharata, were composed, which are dated between the seventh and the fifth centuries bc. The huge heterodoxy of beliefs and reasonings within the epics is discussed in my Foreword to the new Clay Sanskrit Library edition of the Valmiki Ramayana, edited by Richard Gombrich and Sheldon Pollock (to be published by New York University Press). This was also the time when the rebellious teachings of Gautama Buddha and Mahavira Jain, from the sixth century bc, offered a huge challenge to the dominant religious orthodoxy.

  87

  t h e i d e a o f j u s t i c e We learn from his biographer, Arrian, that Alexander responded to this stern egalitarian reproach with the same kind of intense admiration that he had shown in his encounter with Diogenes, expressing huge respect for the interlocutor and conceding the argument made against him. But his own personal conduct, Arrian also noted, remained altogether unchanged: ‘the exact opposite of what he then professed to admire’.2

  Clearly, debates and discussions are not always effective. But they can be. Indeed, even in the case of Alexander, it is possible that these apparently idle chats – with Diogenes, with the Jains, and with many others – did have some effect on the expanding reach and liberality of his thinking and on his firm rejection of intellectual parochialism.

  But no matter what happened to Alexander himself, the channels of communication that his visit to India established had profound effects, over the centuries, on Indian literature, drama, mathematics, astronomy, sculpture and many other pursuits, deeply influencing the face of India in many radical ways.*

  Understanding the demands of justice is no more of a solitarist exercise than any other human discipline. When we try to assess how we should behave, and what kind of societies should be understood to be patently unjust, we have reason to listen and pay some attention to the views and suggestions of others, which might or might not lead us to revise some of our own conclusions. We also attempt, frequently enough, to make others pay some attention to our priorities and our ways of thought, and in this advocacy we sometimes succeed, while at other times we fail altogether. Not only are dialogue and communi-

  * As will be discussed later (in Chapter 15, ‘Democracy as Public Reason’), it is also under Greek influence that the Indians would start their own experiments with democratic governance in municipal administration. On the other side, the Greeks also became much engaged in Indian ideas and philosophy, often in a somewhat romanticized form. On the similarities between Greek and Indian philosophies of that period, see the excellent study by Thomas McEvilley, The Shape of Ancient Thought: Comparative Studies in Greek and Indian Philosophies (New York: Allworth Press, 2002). Some of the similarities may have been independently generated, but there are also huge areas of influence and interaction as well. An important study, regrettably unpublished, is John Mitchener, ‘India, Greece and Rome: East–West Contacts in Classical Times’, mimeographed (Office of the UK Deputy High Commissioner, Kolk-ata, India, 2003).

  88

  v o i c e a n d s o c i a l c h o i c e cation part of the subject matter of the theory of justice (we have good reason to be sceptical of the possibility of ‘discussionless justice’), it is also the case that the nature, robustness and reach of the theories proposed themselves depend on contributions from discussion and discourse.

  A theory of justice that rules out the possibility that our best efforts could still leave us locked into some mistake or other, however hidden it might be, makes a pretension that would be hard to vindicate.

  Indeed, it is not defeatist for an approach to allow incompleteness of judgements, and also to accept the absence of once-and-for-all finality.

  It is particularly important for a theory of practical reason to accommodate a framework for reasoning within the body of a capacious theory – that, at any rate, is the approach to the theory of justice that this work pursues.

  Theories of justice are not, however, taken by most mainstream practitioners to be anything like as general and underspecified as a framework of reasoning. Rather, these specialists seem determined to take us straightaway to some fairly detailed formula for social justice and to firm identification, with no indeterminacy, of the nature of just social institutions. Rawls’s theory of justice illustrates this very well.

  As we have just seen, there is a lot of critical reasoning, involving respectively the pre-eminence of fairness, the conception of the original position, the nature of representation that is involved in the exercise and the type of unanimity that is expected in the choice of institutional principles in the original position. All such general reasoning takes us, we are assured, to quite clear-cut rules to follow as unambiguous principles of justice, with singular institutional implications. In the case of Rawlsian justice, these principles primarily include (as discussed in Chapter 2) the priority of liberty (the first principle), some requirements of procedural equality (first part of the second principle) and some demands of equity, combined with efficiency, in the form of giving precedence to promoting the interests and advantages of the worst-off group (the second part of the second principle). With all this particularized delineation in Rawlsian theory, there need be no great fear of being accused of indecisiveness.

  But is there too much decisivenes
s here? If the reasoning presented so far is correct, then this degree of specification requires us to close 89

  t h e i d e a o f j u s t i c e our eyes to a number of relevant, indeed vitally important, considerations. The nature and content of the Rawlsian ‘principles of justice’

  and the process through which they are derived may have the effect of leading to some seriously problematic exclusions, including: (1) ignoring the discipline of answering comparative questions about justice, by concentrating only on the identification of the demands of a perfectly just society;

  (2) formulating the demands of justice in terms of principles of justice that are exclusively concerned with ‘just institutions’, ignoring the broader perspective of social realizations;

  (3) ignoring the possibly adverse effects on people beyond the borders of each country from the actions and choices in this country, without any institutional necessity to hear the voices of the affected people elsewhere;

  (4) failure to have any systematic procedure for correcting the influence of parochial values to which any society may be vulnerable when detached from the rest of the world;

  (5) not allowing the possibility that even in the original position different persons could continue to take, even after much public discussion, some very different principles as appropriate for justice, because of the plurality of their reasoned political norms and values (rather than because of their differences in vested interests); and (6) giving no room to the possibility that some people may not always behave ‘reasonably’ despite the hypothetical social contract, and this could affect the appropriateness of all social arrangements (including, of course, the choice of institutions), made drastically simpler through forceful use of the sweeping assumption of compliance with a specific kind of ‘reasonable’ behaviour by all.*

  * Some of these limitations have already been discussed, and others will be taken up in the chapters to follow. The last item in this list of omissions and commissions has received some attention in the standard literature, in a somewhat stylized form, through the recognition of the need for theories that deal with ‘non-ideal’ conditions.

  The other items, however, are not helpfully understood in terms of the distinction between ‘ideal’ and ‘non-ideal’ theories, and must not be brushed under the same carpet. The reach and limits of ‘ideal theory’ are instigated in an illuminating sym-posium on ‘Social Justice: Ideal Theory, Non-Ideal Circumstances’ in Social Theory and Practice, 34 (July 2008), led by Ingrid Robeyns and Adam Swift.

  90

  v o i c e a n d s o c i a l c h o i c e If these invitations to close our eyes to significant issues related to justice are to be resisted, then the identification and pursuit of the demands of justice may have to take a much broader and more contingent form. The importance of a framework for public reasoning

  – much emphasized by John Rawls himself – is particularly important in that larger exercise.

  Perhaps the nature of the task can be clarified a little with the help of social choice theory, and I turn now to that line of inquiry.

  s o c i a l c h o i c e t h e o r y a s

  a n

  a p p r o a c h

  Discussions about ethics and politics are not new. Aristotle wrote on these subjects in the fourth century bc with great reach and clarity, particularly in Nicomachean Ethics and Politics; his contemporary Kautilya in India wrote on them with a rather more rigidly institutional approach in his famous treatise on political economy, Arthasastra (as was discussed in the last chapter). But the exploration of the formal procedures of public decisions and their underlying – often hidden – normative presumptions began much later. One of the ways of going into those issues can be found in social choice theory, which, as a systematic discipline, first came into its own at the time of the French Revolution.

  That subject was pioneered by French mathematicians working mostly in Paris in the late eighteenth century, such as Jean-Charles de Borda and the Marquis de Condorcet, who addressed the problem of arriving at aggregate assessments based on individual priorities in rather mathematical terms. They initiated the formal discipline of social choice theory through their investigation of the discipline of aggregation over individual judgements of a group of different persons.3 The intellectual climate of the period was much influenced by the European Enlightenment, and in particular by the French Enlightenment (as well as the French Revolution), with its interest in reasoned construction of social order. Indeed, some of the early social choice theorists, most notably Condorcet, were also among the intellectual leaders of the French Revolution.

  91

  t h e i d e a o f j u s t i c e The motivation that moved the early social choice theorists included the avoidance of both arbitrariness and instability in procedures of social choice. Their work focused on the development of a framework for rational and democratic decisions for a group, paying attention to the preferences and interests of all its members. However, their theoretical investigations typically yielded rather pessimistic results.

  Condorcet showed, for example, that majority rule can be thoroughly inconsistent, with A defeating B by a majority, B defeating C also by a majority, and C in turn defeating A, by a majority as well (a demonstration that is sometimes called the ‘Condorcet Paradox’). On the nature of these difficulties, a good deal of exploratory work (often, again, with further pessimistic results) continued in Europe through the nineteenth century. Indeed, some very creative people worked in this area and wrestled with the difficulties of social choice, for example Lewis Carroll, the author of Alice in Wonderland, who wrote on social choice under his real name, C. L. Dodgson.4

  When the subject of social choice theory was revived in its modern form by Kenneth Arrow around 1950 (Arrow also gave the subject its name), he too was very concerned with the difficulties of group decisions and the inconsistencies to which they may lead. Arrow put the discipline of social choice in a structured and analytical form, with explicitly stated and examined axioms, demanding that social decisions satisfy certain minimal conditions of reasonableness, from which the appropriate social rankings and choices of social states would emerge.5 This led to the birth of the modern discipline of social choice theory, replacing the somewhat haphazard approach of Condorcet, Borda and others with a recognition of the need to state explicitly which conditions must be satisfied by any social decision procedure in order to be acceptable, and allowing other contributors to vary Arrow’s own axioms and demands, after reasoned critique.

  That was the positive and constructive avenue that Arrow’s pioneering work opened up. However, so far as his own axioms were concerned, Arrow dramatically deepened the pre-existing gloom by establishing an astonishing – and hugely pessimistic – result of apparently ubiquitous reach, which is now known as ‘Arrow’s impossibility theorem’ (Arrow himself gave it the more cheerful name of ‘General Possibility Theorem’).6 This is a mathematical result of remarkable 92

  v o i c e a n d s o c i a l c h o i c e elegance and power, which shows that even some very mild conditions of reasonable sensitivity of social decisions to what the members of a society want cannot be simultaneously satisfied by any social choice procedure that can be described as rational and democratic (as Arrow characterized these requirements, with some plausibility). Two centuries after the flowering of the ambitions of social rationality in Enlightenment thinking and in the writings of the theorists of the French Revolution, the subject of rational democratic decisions seemed to be inescapably doomed, just at a time when a peaceful world, full of new democratic commitment, was emerging from the gore of the Second World War.7

  Arrow’s pessimistic theorem, and a cluster of new mathematical results that followed his pioneering lead, together with the wide-ranging general discussions that were generated by this largely technical literature, eventually had a major constructive impact on the discipline of social choice.* It forced the theorists of group decisions to look deeply into what caused the apparently reasonable requireme
nts of sensitive democratic practice to yield these impossibility results. It also emerged that while impossibilities and impasses of this kind can arise with considerable frequency and amazing reach, they can also be, in most cases, largely resolved by making the social decision procedures more informationally sensitive.8 Information on interpersonal comparisons of well-being and relative advantages turns out to be particularly crucial in this resolution.9

  Most of the mechanical procedures of political choice (like voting and elections) or economic assessment (like the evaluation of national income) can accommodate rather little information, except in the discussions that may accompany these exercises. A voting result, in itself, reveals nothing much except that one candidate got more votes

  * The motivational as well as analytical connections between the impossibility theorems and the constructive departures that emerged are discussed in my Nobel Lecture ‘The Possibility of Social Choice’, American Economic Review, 89 (1999), and in Le Prix Nobel 1998 (Stockholm: The Nobel Foundation, 1999). The mathematical relations involved are scrutinized in my Choice, Welfare and Measurement (Oxford: Blackwell; Cambridge, MA: Harvard University Press, 1997), and ‘Social Choice Theory’, in K. J. Arrow and M. Intriligator (eds), Handbook of Mathematical Economics, vol. 3 (Amsterdam: North-Holland, 1986).

  93

  t h e i d e a o f j u s t i c e than another. Similarly, the economic procedure of national income aggregation draws only on information about what was bought and sold at what prices, and nothing else. And so on. When all the information that we can put into the system of evaluation or decision-making takes such an emaciated form, then we have to be reconciled to those pessimistic results. But for an adequate understanding of the demands of justice, the needs of social organization and institutions, and the satisfactory making of public policies, we have to seek much more information and scrutinized evidence.