Choice, Self-Interest, and Commitment

In a forceful critical analysis, Sen challenges the standard picture of rational choice in which this is narrowly viewed as (roughly) a matter of an individual ranking attainable options in a complete and transitive preference ordering and selecting one to which no other is preferred. It is a neat, perhaps again a ‘pure’, picture but one which in his view sells reason short.

Sen’s critique has many branches, compellingly drawn together in the Introduction to Rationality and Freedom (Sen 2002). Here I focus on what I take to be the central limb, brought to prominence through the outspoken thrust of a famous early paper, ‘Rational Fools: A Critique of the Behavioural Foundations of Economic Theory’ (Sen 1977b). This branch of Sen’s case has attracted the hatchets or at least the secateurs of dissenters in a still ongoing tussle. Pettit, a relatively restrained opponent, explains what he takes to be at stake, writing: ‘One of Amartya Sen’s most distinctive claims, and perhaps also one of his most controversial, is that there is an altruistic attitude toward others that does not make sense within the terms of rational choice theory, however broadly.. .construed’ (Pettit 2005: 15). Pettit views Sen as ‘mistaken’, but on a ‘charitable’ reading guilty only of ‘terminological infelicity’, exhibiting ‘idiosyncrasy’ rather than ‘intellectual oversight’ (ibid.: 31-32).

But whose terms are infelicitous? Sen contends in ‘Rational Fools’ that the term ‘preference’ is ambiguous: in preferring A to B, you may simply be choosing it or you may be judging that A will make you better off (if in terms of pleasure, by giving you pleasure directly, or else indirectly through help to someone with whom you are in sympathy).^[1] The risk then comes from eliding the two possible meanings, so that whatever is chosen is simply assumed to make the chooser better off, when in fact choices may be made for reasons that are not to do with one’s own welfare—out of what Sen calls ‘commitment’ (Sen 1977b: 326).^[2] Sen holds that ‘much of traditional economic theory’ makes this elision, thus neglecting the distinction between the preference concepts and assuming ‘the invariable pursuit of self-interest in each act’, whereupon (provocative words) ‘thepurely economic man is...close to being a social moron.a rational fool decked in the glory of his one all-purpose preference ordering’ (ibid.: 329, 343, 336; italics in original).

Sen’s point was developed in a significant series of further contributions. For instance, an initial example (ibid.: 328-329) of Boy 1 offering Boy 2 the chance to choose a larger apple from a pair, but miffed when Boy 2 does so, successfully showed Boy 1 had not made the offer from sympathy but left a question hanging about his commitment—did his discomfiture come from affront to the etiquette he had felt one should be committed to or from frustration of a sneaky strategy to get the apple for himself without seeming impolite? In his important 1997 paper contrasting maximisation with optimisation,^[3] Sen explored game-playing possibilities of this sort in a familiar case of choosing fruit from a dinner party bowl, under the heading ‘Strategic nobility’. Go to that paper also for illustration of a range of possible reasons, some corresponding with commitment, for not rushing to take the single comfortable chair available at a garden party (though perhaps—in what Sen calls ‘chooser dependence’—content to occupy it if specially ushered to it by the host).^[4]

Some of the examples on which attention has focused—the ‘hanging crimes’ of college life—may seem on the trivial side, but to mount objection on that score would be to miss the nature of philosophical thought experiments, in which keeping distraction surrounding more harrowing matters out of examples designed to pinpoint issues of principle can be very much the name of the game. Many weightier applications are offered by Sen, both in relation to remarkable individuals—‘We do not have to be a Gandhi, or a Martin Luther King Jr, or a Nelson Mandela, or a Desmond Tutu, to recognize that we can have aims or priorities that differ from the single-minded pursuit of our own well-being only’ (Sen 2009: 18)—and in relation to major economic concerns, such as work motivation (comparing Japanese experience with that in other developed countries), and environmental responsibility towards other species and future generations (Sen 2005b: 9-12).

The ‘rational fools’ assertion met from the first a prickly reception. Careful criticism did not make the mistake of supposing Sen was insulting the actual par?ticipants in economic markets by dubbing them fools: the target of the epithet was the folly of an economic tradition that in lacking sufficient structure tended to prompt much too narrowly conceived a model of man. But there was, and still is, widespread objection to the claim that the conception was indeed narrow.

Hahn, an early challenger, did not disagree about variety of motivation: ‘Clearly Sen is correct at a level of analysis which concerns states of mind’, continuing however, ‘But I think he has confused others...when it comes to action’. In the reaching of a decision, will not any commitment considerations enter, along with other influences, into the formation of a person’s preference ranking, whereby the merits of available options are weighed against each other? Then the standard model, encompassing this process, would not presume shallow thinking: it is a matter of drawing the thinking together—as Hahn put it, of having ‘an integrated personality’ (Hahn 1991: 7-8).^[5] Recently, Hausman has vehemently defended the standard model of choice against Sen on very similar grounds, agreeing with him on motivations, disagreeing on how to model them, and insisting that ‘economists should employ a single notion of preference’ as ‘total subjective comparative evaluation’—preference, all things considered (Hausman 2012: 21, 64).^[6] But Sen, who is not unsympathetic to viewing maximisation, generously understood, as a necessary (though not sufficient)^[7] condition of rationality, is already well aware of the theoretical scope for capturing factors of concern in a maximisation exercise by broadening the maximand or introducing fresh constraints (Sen 2002: 41-42). This does not eliminate the potential pitfalls he finds in not following the strategy of giving commitment^[8] a place in the foreground (Sen 1991, 2002, 2005b, 2009): leaving it tucked away, however neatly, in the shadows increases the risk of final decisions being interpreted as self-interested after all (albeit sometimes self-interested in terms of the extended self-interest onto which Becker and—extending further—Jolls, Sunstein, and Thaler have shone a spotlight) (Sen 2002: 37-42).

Inspired by the range of ‘moral sentiments’ adduced by Adam Smith (whom he much admires), Sen urges recognition of further richness within the commitment concept itself. Commitment can take the form of choosing non-self-interested goals, such as the pursuit of social justice, and this is one aspect of rationality (‘It is the power of reason that allows us to consider our obligations and ideals as well as our interests and advantages’ (Sen 1999a:

272)). But commitment can also take the form of ‘following rules of.. .behaviour that restrain our inclination to be guided exclusively by the promotion of our own goals’, whether self-interested goals or not, thus making room to ‘let others be’ (‘Being considerate of the desires and pursuits of others need not be seen as a violation of rationality’ (Sen 2009: 193)).

Several critics have called into question the coherence of this second, norm- based, variety of commitment. One contention is that in acting so as to allow for the goals of others, you must in some sense be adopting them yourself. In rebuttal, Sen presents the imagined case of a plane passenger who is enjoying the sunshine through an adjacent window but agrees, when asked, to lower the shade/blind so that a neighbour can better pursue his goal of playing a computer game—a game which the chooser regards as silly and a poor use of his neighbour’s time: it would seem odd then to say that the chooser has adopted the playing of this game by his neighbour as a new goal of his own. Another widely voiced protest is that Sen’s second category of commitment clashes with our understanding of action, for ‘how could one ever fail to act on one’s own goals?’ (Pettit 2005: 15). If one’s original goals are ‘restrained’ (or ‘displaced’ in Pettit’s terms), must not a new goal—that of ‘being considerate’, say—have come into play? But for Sen, there is nothing out of joint about exercising restraint over one’s goals and no necessity to represent this—with, it might be held, a loss of informational richness—as itself goal-driven. He is unshaken, then, in holding both forms of commitment to be significant aspects of our exercise of reason that are all too liable to be overlooked if forced into the standard confining mould.

It could be contended that some defenders of the standard approach against Sen’s challenge have stretched the meaning of its central concepts beyond normal parlance. If you hold back from what you aimed to do, for them you are still doing what you aim. No loss there then. Again, self-interest can appear ubiquitous: as Sen put it in Identity and Violence, a frequent argument drawn from supposed demands of rationality—‘an allegedly knockout argument’— takes the form of asking, ‘If it is not in your interest, why would you have chosen to do what you did?’ (Sen 2006a: 21). But if every action, however seemingly unselfish, is in this sense driven by self-interest, the term surely loses any bite—worse, it risks engendering serious misunderstanding if still taken to have teeth. Terminology that better marks the range of different behaviours— as through Sen’s concept of commitment—may be felicitous after all.^[9]

[1] Hausman (2012: 21-22) appears to be dazzled by the first meaning when he alleges Sen is ‘mistaken’ inseeing the second.
[2] Sen allows for committed behaviour of negative, as well as positive, type.
[3] Buridans ass had no way of optimising but if content to maximise, Sen argues, need not have starved.
[4] Here too is discussion of ‘menu dependence’ (as Sen illustrates this, you would choose to visit youracquaintance to take tea until you realise that cocaine is also on the menu (see also Sen 1993: 502)).
[5] Though perhaps revealingly he amplifies this with ‘a man knows what he wants’.
[6] Hausman notes that ‘people sometimes use the term.. .“preference” in other ways’, but says that, amongother reasons, ‘the justification for retaining the word “preference” for.. .total subjective rankings.. .is thatit matches economic practice’ (Hausman 2012: 64).
[7] He points to the need for reasoned scrutiny of goals—of what to maximise.
[8] Also, when relevant, menu dependence, incomplete preferences, and process sensitivity.
[9] Admittedly, it can take practice to retain the names of some of Sen’s importantly distinct categories. ButCambridge M.Phil. students, set the challenge of trying to devise more readily memorable names for hisilluminating but complexly interwoven concepts of ‘self-centred welfare’, ‘self-welfare goal’, and ‘self-goalchoice’, found Sen’s terms very difficult to replace.