Roads breed traffic. The M25 motorway round London eased congestion at first, and so tempted more drivers into more journeys. A belief that a good road is empty soon fills it up.
Game Theorists are not at all surprised. Roads are a splendid case of an Invisible Hand producing effects which no one intended or wanted. What is a rational choice for each of us need not sum to a result good for all or even any of us. Sometimes it does, as free marketeers claim for the free market. Sometimes an Invisible Hand makes mischief, especially when it suits each of us to be a free-rider and thus together destroy a public good which we each hoped to get for free. As Adam Ferguson remarked long ago, ‘history is the result of human action, not of human design.’
This theme has lately inspired a vast enthusiasm for applying the economists’ theory of Rational Choice to social life at large. Define a ‘rational agent’ as one who always chooses the most efficient means to his ends. This is homo economicus, master of the cost-benefit calculation, bargain-hunter, maximiser of his expected utility. Then define a ‘game’ as an encounter between two or more such agents, where the pay-off which each expects from his choice of action depends partly on what the others choose. (Thus the benefit of the M25 to each driver depends on how many drivers use it.) Game Theory is the abstract general analysis of such games, potentially as revealing about, for instance, marriage, domestic politics and international relations as about the workings of markets. Where it is stuck for a determinate answer, the theory of rational bargaining takes over. The implications for our view of organisations, public policies and ethics are fascinating.
With Logic and Society (1978) and Ulysses and the Sirens (1979), Jon Elster sprang to the fore as an exciting exponent. His gifts for simplifying technicalities, for novel applications and for scholarly connections have helped to spread the word across many disciplines. His rampant individualism – the conviction that history is only the result of individual human actions – has caught a current mood. At the same time he has always had a fine eye for the quirks and paradoxes of Rational Choice theory. Sour Grapes (1983) is subtitled ‘Studies in the Subversion of Rationality’ and Solomonic Judgments (1989) ‘Studies in the Limitations of Rationality’. This mixture of firm belief that Rational Choice holds the key, coupled with electrifying curiosity about its quirks, has opened many minds.
The line is stoutly opposed to all structural or functional explanations of social institutions and historical change, as he made beautifully clear in Explaining technical change (1983). Structures exist, function and change only through individual desires, beliefs and actions, which provide all the explanation we can have or need. When Marxist hackles rose, he pursued the theme in Making sense of Marx (1985). For instance, Marx had claimed that capitalism would be undermined by a falling rate of profit but failed to show why the dialectic of history would bring this about. There is no mystery, however, if it is rational for each capitalist to undercut the rest. The sum of these rational choices is intended by no one, disastrous for all and readily explained by the logic of this typical free-rider result. Unexpectedly Elster has thus been a pioneer of ‘Analytical Marxism’, an individualist reworking of Marx which startles the old guard. Meanwhile a swathe of essays on topics ranging from the philosophical puzzle of weakness of the will to the economic history of the American railroad has made Elster a trig point for a fertile area of the academic map.
After this preamble his three latest books come as a shock. The Cement of Society asks: ‘What is it that glues societies together and prevents them from disintegrating into chaos and war?’ Solomonic Judgments is a fresh set of studies in ‘rationality: its scope, limits and failures’, premised on ‘the normative privileges of rationality in the study of human behaviour’ and focused interestingly on court decisions in child custody cases. Nuts and Bolts is intended ‘to introduce the reader to causal mechanisms that serve as the basic units of the social sciences’. Readers will think they know what to expect, especially when Nuts and Bolts quotes Adam Ferguson on human history with approval and declares: ‘The elementary unit of social life is the individual human action. To explain social institutions and social change is to show how they arise as the result of action by individuals.’
But Elster has lost his bearings, or at least his faith. Each of the three books, he says, ‘reflects an increasing disillusion with the power of reason’. The tone has become hesitant, the analyses fragmentary. Coupled with a Hayekian sense that social planning can only fail is a deep pessimism about the prospects of social science as a source of general enlightenment. ‘At the present time the social sciences cannot aspire to be more than social chemistry ... The time for social physics is not yet here and may never come.’ Even the social chemistry is bitty, confined to ‘specifying small and medium-sized mechanisms for human action and interaction’. He would not be surprised, he remarks engagingly, if readers likened his bits and pieces to rabbits running around on the floor. I found the rabbits delightful but the simile altogether apt.
The source of trouble is the part played by social norms in the explanation of what glues societies together and accounts for social institutions and social change. Rational Choice theory presents us with instrumentally rational, self-interested agents. It sounds as if such individuals, knowing the price of everything and the value of nothing, will never contribute to the common weal unless they gain more than they pay. Indeed, since each calculates in terms of his marginal costs and his resulting marginal benefits, it sounds as if Rational Choice theory will be baffled to account for public goods of any sort.
A public good is anything which benefits a group of people but which cannot be confined to those members of the group willing to pay. For instance, clean air costs money but is breathed equally by those who do not pay to clean it. If picking one’s litter up is a bore, then each litter-bug can reason: ‘my own litter spoils my pleasure in the view less than the cost of my effort. That is true, whether others pick up theirs or not. So it is rational for me not to bother.’ If the reasoning is sound, rational agents will never Keep Britain Tidy.
Generalise the idea and the provision of public goods becomes a ‘game’ where it is never rational to co-operate and the Invisible Hand makes mischief every time. Since ‘social cement’ is among the public goods, there is a serious puzzle about how to account for the norms which hold us together, if individual actions are the elementary units of social life. The crucial norms are those like honesty, which serves to keep us truthful even when it would pay to lie and we could get away with it. Witness the common anxieties about public spirit and morals in an enterprise culture. Rational Choice theory threatens not only to explain how individual enterprise undermines an essential culture of norms but also to encourage the undermining by putting centrifugal thoughts into people’s heads.
This may be to pose the problem too starkly. First, not all norms invite free-riding. British drivers have adopted the norm of driving on the left and, once the norm is in operation, it suits no one to drive on the right. Some of our social cement is durable because it helps coordinate activities to the greater benefit of each of us. Secondly those who value their neighbours’ approval will abide by some norms which they would otherwise buck, especially against a background of law and education to deter free-riders. But, as Elster points out, this cannot be the whole story, since no society would stay glued, if its vulnerable norms were protected only by threats. Rational agents must be persuaded to obey willingly.
Thirdly, then, we need to be careful about dubbing rational agents ‘self-interested’. The term does not mean ‘selfish’. It means only that each is motivated by his own preferences, which can include love for friends, good will towards neighbours and altruism towards people in need. Provided that enough of us want to do what norms encourage everyone to do, Rational Choice theory might he able to account for norms in the abstract, and to show that any trouble with social cement in practice arises because not everyone is rational.
Elster takes some steps in this direction. For instance he thinks that some of us are ‘Kantians’, who do what is morally right unswervingly, and that others are moved by a norm of ‘fairness’, which bids us do our bit if enough people do theirs, and that others have other public-spirited elements in their preferences. But he remains convinced that no complete account of norms can result from such piecemeal observations. ‘I cannot offer a positive explanation of norms,’ he says on page 125 of The Cement of Society. ‘I do not know why human beings have a propensity to construct and follow norms, nor how specific norms come into being and change.’
Before we despair, however, it is worth reflecting that Elster conceives the problem of social cement as one of how rational agents come to have irrational motives. He thinks of ‘rational’ always as meaning instrumentally rational and hence hypersensitive to consequences. He thinks of social norms as a source of irrational motives, because obedience to them is (usually) ‘blind, compulsive, mechanical or even unconscious’ and hence wholly insensitive to consequences. So norms, being thus beyond the scope of Rational Choice theory, are like geographical features of the social landscape, mysterious in their origin and workings. Viewed in this light norms not only defy analysis but also leave Elster stuck with just the sort of collective or structural social facts whose existence he has always denied.
Is Rational Choice theory stuck with this view of norms and how they work? Well, for a start, not even unswervingly moral or public-spirited people are blind slaves of duty. There are always questions of which norms are relevant and what exactly they imply. For instance, when are loyal civil servants to blow the whistle and when are they to be economical with the truth? How should we choose in areas of our lives where social norms and personal commitments overlap and conflict? What if thou canst not love thy neighbour and love thine enemies? To plead that one is forced into blind, compulsive, mechanical obedience is self-deception. Social actors are morally active and accountable players of the social game.
Besides, norms are usually not static but in process of negotiation. Think, for example, of the current rows about the conduct of press and television. It is almost as hard to know what the norms governing privacy actually are as to decide what they should be. Neither philosophers nor sociologists are likely to accept Elster’s static, sharp-edged picture. Norms are no less effective for being fluid and no less real for being negotiable. Both ideally and actually the stuff which binds societies is more like mastic than cement.
The crux is what Rational Choice theory can make of the roles which institutions provide and people play. Given Elster’s approach, roles are parameters within which individuals pursue their individual ends. Alternatively, however, we can think of our roles (or some of them) as a way of expressing who we are and where we belong. Many norms and institutions, even if fixed for each of us, are malleable when we act together. Churches, political parties, business firms and glee clubs evolve in the process of playing the roles which they provide. Here norms are as much enabling as constraining. They enable not only by making collective action easier but also by creating forms of activity. The Catholic Church, for example, is not a vehicle for religious individuals but an expression of one form of Christianity. It defines what is important for the serious Catholics, who together define the evolving Church.
Each of us takes some roles more seriously than others. Some are means to our ends, as when doing a dull job for the money. Others matter deeply for who we are. But we make choices in the playing of both kinds and one wonders if Rational Choice theory needs to be confined to the former. That is a hard question, because its powerful analysis of games and decisions has all been worked out for instrumental choices by calculative individuals. A theory whose rational agents were more social in character and less trapped by their existing preferences would be very different from the familiar one. Yet economists have lately been exploring the thought and Jon Elster’s previous work is a regular source of inspiration for these attempts. So it is all the more disappointing that he seems to have lost hope for the moment.
It may be, of course, that social norms are truly beyond the scope of Rational Choice theory. I am not trying to pronounce on that. I mean only to point out why Elster is stuck. He is right to be disillusioned with the power of reason, if it indeed instructs him to drive a wedge between ‘rational’ motives, like personal gain, and ‘irrational’ motives, like social norms. Having posed the question of norms like this, he can offer us only rabbits running around on the floor. They are, as said very lively rabbits and I commend them warmly to anyone taken by the quirks and paradoxes of Rational Choice. But I hope equally that Elster will soon recover his faith and build them a new hutch.