by Amartya Sen
214
r e a l i z a t i o n s , c o n s e q u e n c e s a n d a g e n c y
c u l m i n a t i o n a n d c o m p r e h e n s i v e o u t c o m e s
Since consequence-based arguments are often seen as being concerned with outcomes (and in some cases interpreted to be concerned only with outcomes), it would be useful, in understanding Arjuna’s arguments, to examine the notion of ‘the outcome’ more closely and critically than the way it is usually treated. The outcome is meant to be the state of affairs that results from whatever decision variable we are concerned with, such as action or rule or disposition. Even though the possibility of describing any state of affairs ‘in its entirety’ is not credible (we can always add some more detail, if necessary by using a magnifying glass aimed at events and actions), the basic idea of a state of affairs can be informationally rich, and take note of all the features that we see as important.
There is no particular reason to insist on an impoverished account of a state of affairs in evaluating it. In particular, the state of affairs, or the outcome in the context of the choice under examination, can incorporate processes of choice, and not merely the narrowly defined ultimate result.
The content of outcomes can also be seen as including all the agency information that may be relevant and all the personal and impersonal relations that may be seen as important in the decisional problem at hand.
In my earlier work on decision theory and rational choice, I have argued for the importance of paying particular attention to ‘comprehensive outcomes’ that include actions undertaken, agencies involved, processes used, etc. along with the simple outcomes seen in a way that is detached from processes, agencies and relations – what I have been calling ‘culmination outcomes’.* This distinction can be central to
* The distinction between culmination outcomes and comprehensive outcomes was discussed in the Introduction, and is quite important for the approach to justice in this work, in which comprehensive outcomes have a role that cannot be played by culmination outcomes. Indeed, part of the problem with what are taken to be
‘consequentialist’ theories of practical reason lies in the tendency to focus on culmination outcomes only. On the broad reach of the distinction, see my essays ‘Maximization and the Act of Choice’, Econometrica, 65 (1997); ‘Consequential Evaluation and Practical Reason’, Journal of Philosophy, 97 (2000); and my book Rationality and Freedom (Cambridge, MA: Harvard University Press, 2002).
215
t h e i d e a o f j u s t i c e certain problems in economics, politics, sociology, and in the general theory of rational decisions and games.* As it happens, the distinction is also crucial in assessing the reach of consequence-based reasoning, since a consequence is more than just the aftermath. The appraisal of comprehensive outcomes can be an integral part of the assessment of states of affairs, and thus a crucial building block in consequential evaluation.
How is this distinction relevant for understanding Arjuna’s arguments? In philosophical discussions on the content of the Gita, it is common to see, as was mentioned earlier, Krishna as the quintessential deontologist, focusing relentlessly on duty, and Arjuna as the typical consequentialist, basing assessment of actions entirely on the goodness (or badness) of the consequences that follow from those actions. In fact, both these interpretations are significantly misleading. There is nothing to prevent a general deontological approach from taking considerable note of consequences, even if the approach begins with the importance of independently identified duties; so it would be a mistake to see Krishna’s somewhat emaciated morality as archetypal deontology. We cannot, for example, understand Immanuel Kant’s deontology on the basis of Krishna’s extremism.† Krishna’s deontology is of a particularly purist form, which goes beyond seeing the importance of duty-based reasoning, and denies the relevance of any concern, particularly any consequential concern, in determining whether some action should be undertaken or not.
Similarly, Arjuna is not a typical no-nonsense consequentialist, insisting on ignoring everything other than the culmination outcomes, which is indeed the way the typically narrow version of consequentialism is defined. In contrast, Arjuna’s moral and political reasoning
* To illustrate one of the issues in a decisional context with a very simple example of the relevance of processes and agencies in the evaluation of a state of affairs, a person may quite like being assigned to a very comfortable chair at a long-lasting party, but not be particularly inclined to run to the most comfortable chair before others get to it. The structures of many decisions and games change when such process-based considerations are accommodated.
† The extent to which Kant is concerned with consequences in the exposition of his basic deontological position is, in fact, quite striking; see, for example, Critique of Practical Reason (1788; translated by L. W. Beck (New York: Bobbs-Merrill, 1956)).
It is hard to think of these arguments as not being part of his overall ethical position.
216
r e a l i z a t i o n s , c o n s e q u e n c e s a n d a g e n c y
is deeply concerned with outcomes in their comprehensive form. The idea of social realizations, as has been explained earlier, demands that outcomes be seen in these broader terms, taking note of actions, relations and agencies. It has already been discussed that Arjuna does make substantial room for his idea of duty, taking into account his responsibility for his own agency, and also acknowledging his special relationship with many of the potential victims of the war (in addition to his general grief at the prospect of massive human death and deliberate killing). That is certainly much broader than consequentialism based on culmination outcomes.
It is part of the approach of the work presented in this book that a comprehensive understanding of states of affairs can be integrated with an overall evaluation of social realizations. While consequences
– even culmination outcomes – are taken seriously among other concerns, there is no defence here of the standard version of consequentialism as it has emerged from two centuries of work led by the utilitarian school. It is, however, useful to ask in what sense, if any, Arjuna’s position is consequentialist, even if not an archetypal one.
c o n s e q u e n c e s a n d r e a l i z a t i o n s It is not easy to identify any definition of consequentialism that would satisfy all those who have invoked that idea, either in defence or in criticism. As it happens, the term ‘consequentialism’ was devised by enemies rather than by proponents of consequential evaluation, and it has been invoked mainly to be refuted, often with colourful counter-examples that have added a good deal of spice – and some intellectual fun – to moral philosophy. To admit to being a ‘consequentialist’ is almost like introducing oneself by saying, ‘I am a wog from London’
(or a ‘frog’ or a ‘limey’ from wherever). Indeed, the term ‘consequentialism’ is unattractive enough to be sensibly bequeathed to anyone who wants to take it away.*
* While I have no great interest in proposing any definition of what consequentialism really is, I should note here that Arjuna’s approach is certainly compatible with Philip Pettit’s definition of consequentialism, as presented in the introduction to the distinguished collection of essays on the subject edited by him: ‘Roughly speaking,’
217
t h e i d e a o f j u s t i c e It is, however, important to see that consequence-sensitive reasoning is necessary for an adequately broad understanding of the idea of responsibility. This has to be a part of the discipline of responsible choice, based on the chooser’s evaluation of states of affairs, including consideration of all the relevant consequences viewed in the light of the choices made and the comprehensive outcomes associated with what happens as a result.4 This substantive issue is not, of course, directly concerned with the use of the term ‘consequentialism’.
Whether the ideas of responsibility and social realizations, as explored here, should be placed in some wide enough basket called ‘consequentialism’ is not a qu
estion of much substantial interest (in the way that the ideas themselves are).*
It is true that the importance of personal responsibility has not always been adequately recognized within what has been called consequentialist ethics. The standard versions of utilitarian ethics have been especially short in this, in particular by ignoring all consequences other than utilities, even when they are part and parcel of the state of affairs (for example, the actions of particular agents that have actually occurred). This has followed from the utilitarian programme of combining consequentialism with additional demands, particularly ‘welfarism’, which insists that states of affairs must be judged exclusively Pettit says, ‘consequentialism is the theory that the way to tell whether a particular choice is the right choice for an agent to have made is to look at the relevant consequences of the decision; to look at the relevant effects of the decision on the world’
( Consequentialism (Aldershot: Dartmouth, 1993), p. xiii). Since there is no insistence here that the accounting of consequences be confined to culmination outcomes only, ignoring the relevance of agencies, processes or relations, capturable in the picture of a comprehensive outcome, there is no tension in seeing Arjuna as a consequentialist in Philip Pettit’s sense.
* There is, in fact, also a ‘signalling’ issue that makes consequentialism an oddly unsuitable name for an approach that begins with – and focuses on – the evaluation of states of affairs. To see the states of affairs as ‘consequences’ raises the immediate question: consequence of what? So even though philosophers who see themselves as consequentialists seem inclined to start with the evaluation of states of affairs (and then proceed to the evaluation of other things such as acts or rules), the term consequentialism points in the opposite direction – to the prior relevance of something else (an action or a rule or whatever) of which a state of affairs is a consequence. It is like, first, defining a country merely as a colony, and then striving hard to show not only that the colony is important independently of the metropolis, but also that the metropolis itself should be assessed entirely in the light of the colony.
218
r e a l i z a t i o n s , c o n s e q u e n c e s a n d a g e n c y
by the utility information (such as happiness or desire-fulfilment) related to them – no matter what the other features of the consequent states of affairs may be, such as the performance of particular acts, however nasty, or the violation of other people’s liberties, however personal.*
r e a l i z a t i o n s a n d a g e n c i e s Here endeth my discussion of consequentialism. But the substantial issues do, of course, remain, and there will be plenty of engagement with them in the rest of the book. But I do want to make a couple of further points before closing this chapter. I have emphasized the importance of recognizing that the perspective of social realizations is a great deal more inclusive than the narrow characterization of states of affairs seen as culmination outcomes. A person not only has good reason to note the consequences that would follow from a particular choice, but also to take an adequately broad view of the realizations that would result, including the nature of the agencies involved, the processes used and the relationships of people. Some of the deontological dilemmas that are presented, with evident relevance, to discredit narrowly consequentialist reasoning, need not arise, at least in those forms, in dealing with responsible choice based on assessment of social realizations that would follow from one choice or another.
Given the importance of states of affairs in social realizations, a question that would occur to many critics of consequential reasoning is this: if we want to take note of agencies, processes and personal relations, is there is any real hope of getting a consistent system of
* Indeed, utilitarian reasoning is an amalgam of three distinct axioms: (1) consequentialism, (2) welfarism, and (3) sum-ranking (the last stands for the requirement that utilities of different people must simply be added up to assess the state of affairs, paying no attention to, say, inequalities). On the factorization of utilitarianism, see my ‘Utilitarianism and Welfarism’, Journal of Philosophy, 76 (September 1979), pp. 463–89, and Amartya Sen and Bernard Williams (eds.) Utilitarianism and Beyond (Cambridge: Cambridge University Press, 1982); see particularly our joint Introduction.
219
t h e i d e a o f j u s t i c e evaluation of social realizations on which reasoned and responsible decisions can be based? Given the demands of consistency, how can two persons value the same state of affairs differently, depending on their respective actions and responsibilities? The perceived problem here clearly arises from the temptation to see the evaluation of social realizations in strictly impersonal terms. The insistence that you and I, if we follow the same system of ethics, must value a comprehensive outcome in exactly the same way does correspond to the demands of utilitarian ethics, which is of course a classic case of consequential reasoning, but informationally a highly restrictive one. To insist on the same requirement in the evaluation of comprehensive outcomes, even when we are concerned with agencies and relations and processes, would seem to be entirely arbitrary, and indeed motivationally contradictory.5
In fact, if the roles of different persons in the development of a state of affairs are totally different, it would be rather absurd to make the odd demand that the two must value that state of affairs in exactly the same way. This would make a nonsense of taking note of agencies that are integral parts of social realizations. When, for example, Othello explains to Lodovico that he has killed Desdemona by saying,
‘That’s he that was Othello; here I am’, it would indeed be ridiculous to insist that Othello must see what has happened in exactly the same way as Lodovico would. The understanding of the nature of the deed and of his own agency in it that makes Othello take his own life would also demand that he cannot view what has happened without considering his own role in the murder, which would make his perspective altogether different from that of others. Othello’s positionality is central to the evaluation – not a detail that can be lost in his own evaluation of the event.*
It is not surprising that Arjuna’s consequence-sensitive reasoning attached particular importance to the fact that he himself would have
* As was discussed in Chapter 7, ‘Position, Relevance and Illusion’, whether a positional connection is an important concern for a person’s assessment of a state of affairs, or merely a distorting influence that should be overcome, is a matter for reasoned evaluation. In this case it would be hard to argue that Othello’s role in Desdemona’s murder is a distracting detail that should be overlooked as Othello evaluates what exactly has happened.
220
r e a l i z a t i o n s , c o n s e q u e n c e s a n d a g e n c y
to do a good deal of the killing, and that some of the people killed would be those related to him for whom he has affection. Sensitivity to consequence does not demand insensitivity to agencies and relations in evaluating what is happening in the world. There can be good reasons to take note of both agent-relative and agent-independent concerns in the appraisal of what happens in the world, and thus in assessing justice in the sense of nyaya.* There is, however, no exemp-tion from personal scrutiny, or indeed public argument, in the evaluation of their respective relevance and importance. The demand for reason in the assessment of reasonableness applies to both.
* The idea of responsibility can have very different bearings depending on the context and the purpose of the investigation. For some important distinctions, which I have not addressed here, see Jonathan Glover, Responsibility (London: Routledge, 1970); Hilary Bok, Freedom and Responsibility (Princeton, NJ: Princeton University Press, 1998); and Ted Honderich, On Determinism and Freedom (Edinburgh: Edinburgh University Press, 2005), among a number of other relevant studies. See also Samuel Scheffler, ‘Responsibility, Reactive Attitudes, and Liberalism in Philosophy and Politics’, Philosophy and Public Affairs, 21 (Autumn 1992).
221
.
p a r t iii
The Mat
erials of Justice
.
11
Lives, Freedoms and Capabilities
Twenty-five hundred years ago, when young Gautama, later known as Buddha, left his princely home in the foothills of the Himalayas in search of enlightenment, he was moved specifically by the sight of mortality, morbidity and disability around him, and it agitated him greatly. He was also distressed by the ignorance he encountered. It is easy to understand the sources of Gautama Buddha’s agony, particularly the deprivations and insecurities of human life, even if we may have to ponder more about his subsequent analysis of the ultimate nature of the universe. It is not difficult to appreciate the centrality of human lives in reasoned assessments of the world in which we live.
That, as has already been discussed in the Introduction and later, is a central feature of the perspective of nyaya in contrast with the rule-bound niti, even though the idea of nyaya is not at all alone in pointing to the relevance of human lives for assessing how a society is doing.
Indeed, the nature of the lives people can lead has been the object of attention of social analysts over the ages. Even though the much-used economic criteria of advancement, reflected in a mass of readily produced statistics, have tended to focus specifically on the enhancement of inanimate objects of convenience (for example, in the gross national product (GNP), and the gross domestic product (GDP), which have been the focus of a myriad of economic studies of progress), that concentration could be ultimately justified – to the extent it could be
– only through what these objects do to the human lives they can directly or indirectly influence. The case for using instead direct indicators of the quality of life and of the well-being and freedoms that human lives can bring has been increasingly recognized.1