Game theory has made it possible to explore how and why various strategies might evolve. Trivers used a game called the Prisoner’s Dilemma in which two people are kept apart and told they are accused of a crime with a penalty of, say, ten years in prison. If both stay silent they can be convicted only on a lesser charge and both get a shorter sentence, say three years, but if one gives evidence against the other the defector gets off free. What should they do? Obviously the best outcome all round is for both to stay silent – but there is a strong temptation to defect – and what if the other one is tempted? – you might as well be tempted too. There are many other versions using points, money, or other resources. The important point is that a perfectly rational and selfish person will always gain by defecting. So how does cooperative behaviour ever come about?
The answer is that in a one–off game it never should, but life is not a one–off game. We meet people again, and form judgements about their trustworthiness. The answer to the Prisoner’s Dilemma lies in repetition. In iterated Prisoner’s Dilemmas people assess the other’s likely behaviour and then both can gain by cooperating. Players who have not met before often copy each other, cooperating with cooperators and not with defectors. Persistent defectors are shunned, and so lose their chance of exploiting others.
Games like this are also used by economists, mathematicians and computer modellers. In 1979, the American political scientist Robert
Axelrod set up a tournament and asked computer programmers to submit strategies for playing the game. The fourteen entries each played 200 times against all the others, themselves, and a random program. To many people’s surprise, the winning program ‘Tit–for-tat’ was both simple and ‘nice’. Tit–for-tat began by cooperating and then simply copied what the other player did. If the other player cooperated then both continued to cooperate and both did well; if the other player defected, Tit–for-tat retaliated and so did not lose out too badly against defectors. In a second tournament over sixty programs tried to beat Tit–for-tat but failed.
Subsequent research has used more complex situations, with many players, and has been used to simulate evolutionary processes. It turns out that unless Tit–for-tat begins against overwhelming numbers of defecting strategies, it will spread in a population and come to dominate it. It is what is known as an ‘evolutionarily stable strategy’. However, the real world is more complex, and Tit–for-tat does not do so well when mistakes are made, or when there are more players and more uncertainty. Nevertheless, this approach shows how group advantage can emerge out of purely individual strategies without the need to appeal to evolution for the ‘greater–good’.
Is this how cooperative behaviour actually evolved? If so it would need some kind of nice behaviour to get it started, and Trivers has suggested that kin selection might have provided the starting point. Animals already supplied with feelings of affection and caring towards kin could easily begin generalising and so give nice Tit–for-tat the start it needed.
Note that the Prisoner’s Dilemma is a non–zero-sum game. In a ‘zero–sum’ game what I gain you lose, and vice versa. This is not so for many real–life situations. Half a blood meal means life or death to a hungry young vampire bat but no more than an easy way to buy future favours for a well–fed more experienced hunter. This exposes the rather unpleasant concept of bargain hunting – giving deliberately to others who are in great need because their debt to you will be all the greater. This approach has also been used to show how moralising might evolve, since it pays to punish defectors and even to punish people who fail to punish defectors. In this kind of game, trustworthiness becomes a valuable currency. It pays you to be seen to be cooperative because you may reap the reward at some later date.
I have given only a few examples of how sociobiology has dealt with the problem of altruism (more extended treatments can be found in Cronin 1991; Matt Ridley 1996; and Wright 1994) but I hope these are enough to see just how successful it has been. In a sense this approach takes the
altruism out of altruism. Acts of kindness and cooperation can be explained because they ultimately help the survival of the selfish genes on which they depend. Is that the problem solved, then? Does all of human altruism ultimately come down to kin selection and reciprocal altruism?
The oddities of human altruism
In today’s world we frequently deal with people who are unrelated to us and whom we know we will never meet again. This suggests that society ought to be becoming less kind and cooperative, but this does not seem to be happening. Psychologists have long studied helping and cooperative behaviour. Experiments in the 1970s concentrated on bystander apathy–the depressing finding that people often do nothing to help a person injured in the street. They found that helping is greatly increased if the bystander is the only one who can help, and is decreased if other people can be seen not helping – so this is another situation in which people imitate each other. More recent studies, however, show that people will offer help in a wide range of situations. Experiments teasing out the effects suggest that people help because they feel empathy for the sufferer, and not because they are related to them, nor because they can expect any reward for helping (Batson 1995).
Try to think of the most altruistic of human acts you can. Dawkins gave the example of giving blood. In Britain every healthy adult is encouraged (or at least invited) to give blood twice a year, and donors are not paid – you get a cup of tea and a biscuit, and a little badge after ten donations. He suggested this was a case of ‘pure, disinterested altruism’ (Dawkins 1976, p. 230). Others have suggested giving a large tip in a restaurant you will never visit again, or going to Ethiopia to help starving orphans. We might add picking up valuables found in the street and handing them in to the police, clearing away someone else’s abandoned rubbish, recycling your waste, or setting up a standing order to a charity whose members you will never meet. Then there are dogs’ and cats’ homes, and many people who care for birds with broken wings or maltreated donkeys. All these may appear to be examples of ‘true’ altruism but, sociobiologists would argue, they are really the byproducts of kin selection and reciprocal altruism. We are most generous to our relatives (or those we think might be relatives) and we are nice to others so as to build up a reputation for being good and trustworthy. Is this explanation adequate?
Let us take a few examples in more detail. Imagine an Australian who sends money to the starving in Africa, or an American who sends money to Bangladesh. Many people do this and some make no fuss about it. They send off a cheque and never even tell anyone they have done so. This cannot be kin selection because the final recipients are probably about as unrelated to the average donor as they could possibly be. You might even argue that on a planet with limited resources, this kind of generosity is strongly against the genetic interests of the donor – over and above the cost of the gift. So is it reciprocal altruism? Clearly not in any straightforward sense because the donor never expects to see the recipients or to be thanked by them in any way. However, evolutionary psychologists argue that such generosity is a way of building up the donor’s reputation as a generous person (Matt Ridley 1996). In that case, though, we should expect people to brag about their donations, which often they do not. Even this can be explained as part of reciprocal altruism on the theory that the feeling of guilt is evolution’s way of making sure the system works, and so these hidden acts of generosity are just mistakes – the price we pay for having our uniquely human emotions.
The examples I have given so far are mostly isolated acts of generosity, but altruism is much more deeply embedded in our lives than that. Vast numbers of people choose to do jobs that are badly paid, poorly rewarded, have very long hours, and are highly stressful, because they want to be of service. Such jobs include social work, psychotherapy, working in old people’s homes, looking after delinquent children, and environmental protection. Why would anyone want to spend several years training to become a nurse and then spend their life working irregular hours, long shifts, dealing with difficult people, clearing up horrible messes, spending hours giving out pills and making beds in an environment of sickness and disease, all for an uncomfortably low salary? The answer cannot be for material gain or genetic advantage. Nurses may say it is because they want to help people, because it makes them feel fulfilled, because they believe that life is only worth living if you help others, because they are grateful to be healthy and want to help those who are not, because they recognise that money alone is not the way to happiness, and so on.
According to sociobiological theory these reasons must all be byproducts of reciprocal altruism, but to me that stretches the theory to breaking point. The problem is that natural selection is ruthless and the cost of this kind of generosity could be very high indeed. People who managed to avoid paying it in the past would have been at an advantage
and would have passed on their genes for avoiding it. Evolutionary psychologists might argue that our emotional system was designed for the hunter–gatherer lifestyle and must be expected to go wrong (and perhaps to produce excess generosity) in a rich technological world. Perhaps the knowledge that ‘I will never ever see this person again’ is no match for underlying emotions programmed by the genes in times past, but then we are back to explaining away our behaviour as just a mistake.
So is there an alternative?
Until now there have been only two major choices in accounting for altruism. The first is to say that all apparent altruism actually (even if remotely) comes back to advantage to the genes. On this view there is no ‘true’ altruism at all – or rather, what looks like true altruism is just the mistakes that natural selection has not managed to eradicate. That is the sociobiological explanation. The second has been to try to rescue ‘true’ altruism and propose some kind of extra something in human beings – a true morality, an independent moral conscience, a spiritual essence or a religious nature that somehow overcomes selfishness and the dictates of our genes; a view that finds little favour with most scientists who want to understand how human behaviour works without invoking magic. Neither choice appears satisfactory to me.
Memetics provides a third possibility. With a second replicator acting on human minds and brains the possibilities are expanded. We should expect to find behaviour that is in the interests of the memes, as well as behaviour serving the genes. Magic is no longer required to see why humans should differ from all other animals, nor why they should show far more cooperative and altruistic behaviour.
We can ask our meme–selection question again.
Imagine a world full of brains, and far more memes than can possibly find homes. Which memes are more likely to find a safe home and get passed on again?
I suggest that among the successful memes are altruistic, cooperative, and generous ways of behaving.
Altruism in the service of the memes
Imagine two people. Kevin is an altruist. He is kind, generous, and thoughtful. He gives good parties and buys people drinks in the bar. He often has friends round for meals and he sends out lots of birthday cards. If his friends are in need he takes the trouble to ring, to help them out, or to visit them in hospital. Gavin is mean and selfish. He resents buying other people drinks, and thinks birthday cards are a waste of money. He
never invites people round for a meal, and if his (few) friends are in trouble he always has something more important to worry about. Now the question is – who will spread more memes?
Other things being equal, Kevin will. He has more friends and spends more time talking to them; they like him and they listen to him. The memes he spreads might include the stories he tells, the music he likes, the clothes he wears, and the fashions he follows. They might be the scientific ideas he likes to discuss, the economic theories he espouses, and his political views. Most important, they will also include all those memes that make him the way he is – memes for giving good parties, for sending out lots of cards, for helping people in need and for buying them drinks. Psychological experiments confirm that people are more likely to be influenced and persuaded by people they like (Cialdini 1994; Eagly and Chaiken 1984). So his friends will imitate his popular behaviour and thus his altruism will spread. And the more friends he has, the more people can potentially pick up his ways of making himself popular. We could call Kevin a meme–fountain (Dennett 1998).
Meanwhile, Gavin has few friends. He makes few opportunities for talking to the ones he does have, and he rarely finds himself chatting over a drink or passing the time of day with a neighbour. His memes have few chances to replicate because the few people who could potentially imitate him rarely do so. Whatever he thinks about the state of the nation or the best way of making apple pie, his ideas are unlikely to spread far because people do not listen to him, and if they do they do not adopt his ideas because they do not like him. We might call Gavin a meme–sink.
This difference forms the basis of a memetic theory of altruism. The essential memetic point is this – if people are altruistic they become popular, because they are popular they are copied, and because they are copied their memes spread more widely than the memes of not–so-altruistic people,
including the altruism memes themselves.
This provides a mechanism for spreading altruistic behaviour.
Note that I am not the first to treat altruistic acts as memes. As we shall see, Allison (1992) proposes quite a different mechanism, and Du Preez (1996) considers selfish and altruistic discourses to be evolving memes, though without explaining exactly why altruism should spread in spite of its cost. There are many ways of being altruistic and I have lumped them all together here, but they include generosity, kindness, caring behaviour, and so on – anything that makes it more likely that others will want to spend time with those people, and emulate them, and so will pick up their memes. Note that for this kind of memetic altruism to work two things must be true. First, that people are capable of imitation, and second, that
they more often imitate altruists. If both these are true we should expect people just to find themselves being helpful and altruistic, without necessarily knowing why.