You win, I win

Philip Kitcher

  • Unto Others: The Evolution and Psychology of Unselfish Behaviour by Elliott Sober and David Sloan Wilson
    Harvard, 400 pp, £18.50, May 1998, ISBN 0 674 93046 0

Organisms that contribute to the reproductive success of their species by doing things that decrease the size of their own brood appear to be inevitable losers in the Darwinian struggle. Since the 19th century, biologists have regarded the evolutionary possibility of altruism as an important theoretical puzzle, and in past decades, it has become clear that it can’t be solved by vague appeals to the idea that co-operative behaviour is good for the flock, the herd or the species. There are alternatives, however. If altruists direct their helpful behaviour towards relatives, then the genes associated with altruism may spread, because they are present in the beneficiaries and transmitted to their offspring; or if today’s altruist is tomorrow’s recipient, the present loss may be made up with interest. Models of kin selection and reciprocal altruism are widely regarded as solving the puzzle.

These explorations involve a special notion of altruism: biological altruists are organisms who contribute to the reproductive success of others at reproductive cost to themselves. By contrast, our everyday concept of altruism emphasises intentions, and we think of altruists as people who want others to thrive. Megalomaniacal spermdonors aside, few of us place reproductive considerations at the top of our list of priorities, so the biological focus on reproductive costs and benefits is alien to our broader vision of well-being. Moreover, biological altruism cares nothing about motives: indeed, it covers organisms not noted for the richness of their psychological lives – we can speak of altruistic behaviour in insects and even plants. Despite what some of the louder sociobiologists have claimed, an ability to account for the evolution of biological altruism does not have direct implications for psychological altruism, the kind of other-directedness that really matters to us.

Just as the puzzle of biological altruism has sometimes prompted Darwinians to deny that altruism exists in nature, so, too, there has been widespread scepticism about psychological altruism. Influential ideas in philosophy, economics and political theory have combined to offer a simple picture of human behaviour as directed towards maximising the satisfaction of the agent. Confronted with apparent examples of people who act to promote the welfare of others, fans of Homo economicus maintain that such people’s deep motives involve, if not the expectation of future benefit, at least the inner glow that comes from ‘doing good’. So, it’s suggested, we are creatures dedicated to our own pleasure and (the illusion of) altruism is most prominent when there are social means of inducing us to take pleasure in the happiness of others.

Elliott Sober and David Sloan Wilson are clear that there are two notions of altruism, as well as two challenges to its possibility, stemming from quite different sources, but their wide-ranging book is intended to tackle both. They begin with biological altruism, offering their own perspective on how this puzzle should be resolved, and discussing the ways in which natural selection of social structures may have figured in the history of our species. In the second half of Unto Others, they turn to psychological altruism, arguing that debates between those who believe that human beings are sometimes other-directed and their sceptical opponents cannot be settled either by philosophical arguments or by psychological experiments (at least, not by the kinds of psychological experiment typically designed by participants in the debates). They conclude by maintaining that natural selection would be expected to favour desires for the welfare of others based neither on the expectation of reciprocation nor on some ‘warm feeling’. Thus both types of altruism are made intellectually respectable.

Sober and Wilson offer a distinctive approach to the problem of biological altruism, one that attempts to incorporate the accepted solutions within a unified theory. For two decades, Sober, an internationally prominent philosopher of biology, has provided welcome clarification of the concept of natural selection, while, for an even longer period, Wilson, a well-known theoretical biologist, has campaigned to rehabilitate one of the most vilified views about the nature of selection. Unto Others opens with a manifesto for group selection.

Faced with apparent examples of biological altruism, theorists in the first half of the 20th century floated the idea that these forms of behaviour endured because they were ‘good for the group’: birds gather on telephone wires (so the story goes), assess their numbers, and regulate their procreation so that population size remains within bounds. Rigorous formulations of the idea proved disconcerting, however. If we imagine a group of altruistic organisms who reduce their own reproduction (for example), then, in the simplest scenario, the group can be invaded by a selfish mutant that reproduces without restraint; birds who ignore the census and procreate freely will spread whatever genes are associated with their anti-social behaviour. In succeeding generations, selfish types will become increasingly prevalent until they eventually take over completely. Individual selection for selfishness will always defeat group selection for altruism. Or so it appears.

Sober and Wilson demonstrate that this dismal conclusion can be avoided if we complicate the scenario. Suppose we have a large population of organisms, some selfish and some altruistic, which divides up into groups. Within each group, the selfish types increase in frequency, but when the groups dissolve, giving rise to a new large population, those with more altruists contribute more members to the global population. When the algebra is done, it’s not hard to see that, overall, the altruists can become steadily more prevalent. On this account, the association of similar organisms is the key to the evolution of altruism: the altruists become more numerous because they keep company with one another, and kin selection works because association with relatives brings altruists together. Reciprocal altruism is effective because it pairs organisms who alternate as donor and beneficiary. At its most ambitious, the revival of group selection delivers a stern rebuke to the traditional supposition that kin selection and reciprocal altruism are rigorous replacements for the sloppiness of group selection. Rather, these fashionable models are disguised variants of the fundamental group selectionist idea.

The models that evolutionary theorists construct can be taken in two different ways: one is hard-line while the other is relaxed and pluralistic, recognising that many such models are possible and celebrating the pragmatic benefits of diversity. Sober and Wilson could be seen as providing us with a useful new model to supplement old theories of kin selection or reciprocal altruism. They obviously want more than this, however. At various points, they maintain that decisions about which model to use are not matters of convenience: only one way of conceptualising selection processes captures the causal facts about what occurs in nature, and, in instances of biological altruism, they see group selection as what is really going on.

In the Seventies, Richard Dawkins took a similarly hard line when claiming that natural selection is really a process in which ‘selfish genes’ struggle against one another. Later, he changed to a more relaxed view, suggesting that such processes could be fully represented in terms of intergenic competition, even though some of them could also be conceived in other ways. Whether the hard line proposed supports the omnipresent selection of genes or defends the claim that groups are sometimes the real targets of natural selection, it is equally unjustified. Emphasising the word ‘real’ in this context is little more than chest-thumping, the impulse to engage in which results from a failure to recognise that the appeal to natural selection is itself a metaphor.

When intelligent beings select it makes sense to inquire into the grounds on which they’re choosing. Did the voters pick Tony Blair because he is young and dynamic, because they approved of his policies or because they were tired of the Tories? Controversies about the real units of natural selection arise when analogies are illegitimately invoked in complex contexts in which organisms of different types survive and reproduce at different rates – contexts that Darwin taught us to see in terms of the metaphor of natural selection – and it’s unnecessary to probe causal intricacies in search of the entities on which selection is ‘acting’. Sober and Wilson formulate their account of group selection without a great deal of mathematical detail (making it easy to grasp the essentials), but one drawback of this is that they don’t show that their examples of group selection can’t be understood in other terms.

In fact, there’s a well-known technique for constructing models of different kinds that yield the same consequences. Where one model postulates a group selection process, another sees the same phenomena in terms of a contribution to the reproductive success of individuals that comes from their being in a special environment (one affected by the other organisms with whom they interact); some organisms do well at Darwin’s game because they interact frequently with friendly types. Yet another model supposes that genes at particular loci have chances of being transmitted to future generations that depend on the environments in which they find themselves – where the environments extend all the way from the other parts of the genome to the characteristic situations in which the bearers of the genes interact with other bearers of genes; some genes are replicated more frequently because of the genetic company they keep, and because they inhabit bodies likely to interact with other bodies carrying ‘unselfish’ genes (more exactly, genes that tend to make their bearers co-operative). Unto Others fails to answer the pluralist argument that the strategies for generating alternative representations of selection processes are general, and that it is consequently misguided to claim that natural selection is rightly understood from any one perspective.

There is, even so, much to be learnt by deploying the kinds of account of human social evolution Sober and Wilson favour. They draw a helpful distinction between two types of altruism, primary (for example, sharing the food obtained on a hunt) and secondary (enforcing the giving of a fair share to everyone in the group). Both types can be understood in terms of their model of group selection (and from the perspective of competition among individuals – or even genes). But secondary behaviour evolves more easily than primary, since the costs of engaging in enforcement are small. Of course, once a group has arrived at a rich repertoire of secondary behaviour, it becomes advantageous to each individual in it to abide by the rules, and forms of co-operative behaviour that might previously have been strongly opposed by individual selection may now become prevalent. Sober and Wilson outline a case for what they call the ‘amplification of altruism’, and offer a survey of randomly selected cultures to show how their theory might be applied.

Nothing that I’ve considered so far touches the issue of psychological altruism, however. Group selection might work wonders in nature, leading all kinds of organisms to behave in ways that raise the reproductive success of those around them, and co-operation might be especially prominent in our own species because of the amplification of altruism. Yet those organisms whose psychological lives are rich enough for them to act intentionally might invariably do so from selfish motives. Any attempt to rebut this depressing view must start from a clarification of what is at stake. Sober and Wilson begin by distinguishing two kinds of case. People may want something simply as a means to an end, as a lobbyist directs his actions towards pleasing a politician, not because he is concerned about the officeholder’s happiness but because he wants favourable treatment for himself. Alternatively, something can be wanted for its own sake. Scepticism, however, suggests that there are always plausible explanations of apparently altruistic wishes that see them as ultimately directed towards selfish ends. You think you genuinely want your friend, child or partner’s happiness for its own sake, but how do you know (nudge, nudge) that you aren’t really concerned with your own well-being, to which end the happiness of these others is merely a means?

The obvious way to settle the issue is to ask whether your wish that the other person benefit would remain even if it weren’t associated with a selfish end. If the politician’s contentment had no bearing on the likelihood of his support for the lobbyist’s causes, the lobbyist wouldn’t have the slightest desire to please. Sober and Wilson see that there are complications here. Sceptics propose that apparently altruistic behaviour is directed to a selfish end: the agent prefers the state in which the other person flourishes because this is one in which the selfish end is attained.

If this end is no longer correlated with the other’s welfare, there are several possibilities. Suppose you and I share a garden and divide up the chores of fertilising, watering and weeding. Your flowers and my vegetables will be entered for prizes at the village fête. We might both win, neither of us might win, or one of us might win but not the other. Our psychological profiles on the egoism-altruism axis are determined by how we rank the options. Maybe you are so altruistic that you rank highest, and are indifferent between, the cases in which I alone win and we both win; you are also indifferent between the case in which you alone win and the case in which we both lose; only my success matters to you. Perhaps I am a pure egoist, regarding the case in which I alone win as just as good as that in which we both win, and the case in which you alone win as no better than that in which neither wins. Or maybe I am a bit more altruistic, preferring that I win and you do not rather than that you win and I do not, but still thinking it better that you alone should win than that neither should win. I could even be more altruistic still, preferring that if one of us wins it should be you. So there are pure altruists, pure egoists and two intermediate types.

Psychological egoists find no place for pure altruism or for either of the intermediate types. Sober and Wilson, by contrast, recommend the pluralist view that human beings are not always egoistic. They show convincingly that philosophical analysis can rebut popular arguments for the inevitability of egoism. The frequently heard suggestion that we must be egoists because all our intentional actions are brought about by our own desires is vulnerable to the point – familiar since Bishop Butler – that some of those desires might be directed towards the well-being of others. Philosophical considerations can’t, however, take us further. Nor, when we turn to the most serious and well-designed attempts to resolve the issue by psychological experiments, do we find compelling evidence. Sober and Wilson argue that the data generated can be understood in terms either of pluralism or of egoism.

Where, then, can we look to discover whether we deceive ourselves in thinking that human beings are sometimes altruistic? To the likely operation of natural selection. In their last chapter, Sober and Wilson contrast two possible explanations for parents’ actions towards their children. In the altruistic explanation, parents perform an action when they think it will help their children; in the hedonistic one, parents act when they believe they will obtain pleasure by doing so. Sober and Wilson maintain that natural selection ought to favour the altruistic alternative. Their argument is that, even though the parents’ pleasure might be correlated with the thriving of their offspring, the pleasure-seekers will be less reliable in aiding their offsping than the altruists, and, in consequence, their children will fare less well; since genetic representation in future generations depends on how the children do, altruists will outcompete hedonists.

Interestingly, this argument does not draw on the multi-level picture of selection drawn by Sober and Wilson in the first part of their book. It invokes simple individual selection: parents who desire the well-being of their children are expected to have more descendants in future generations. More importantly, Sober and Wilson are vulnerable to the criticism often urged against sociobiology – despite the fact that their conclusion is so much more optimistic about the springs of human behaviour than sociobiologists usually are. Have they worked out a precise model that will consider the costs and benefits, in terms of reproductive success, of different motivations? Have they considered that the development of a mechanism for desiring the well-being of others might have behavioural consequences outside the contexts in which aid to children is given or withheld? Do they have reason to think that charity begins at home, or might the attitudes towards children be derived from broader psychological mechanisms that govern our behaviour towards friends or sexual partners?

Unto Others does not adequately address such questions. Sober and Wilson briefly float the possibility that their pluralism might play a role in social relations, that it ‘is a plausible design solution for the problem of getting group members to take care of one another’, but they do not confront the obvious egoistic retort that the problem might be better solved if the group members focused narrowly on their own happiness and aided others in cases in which they might gain from a return of the favour. Suggestive as it is, their argument is incomplete. Nevertheless, they have considerably clarified what is at stake in the debate about psychological altruism, and have demonstrated how an evolutionary perspective might bear on it. The challenge now is to explore whether their preliminary case can be made more extensive and more rigorous. Defenders of psychological altruism can no longer be automatically indicted on grounds of wishful thinking, but they still have much work to do before psychological egoism is exposed as the vulgar deception most of us would like it to turn out to be.