Darwinian Soup

W.G. Runciman

  • The Meme Machine by Susan Blackmore
    Oxford, 264 pp, £18.99, March 1999, ISBN 0 19 850365 2

The word ‘meme’, popularised by Richard Dawkins in The Selfish Gene, has recently gained entry into the OED as ‘an element of a culture that may be considered to be passed on by non-genetic means, esp. imitation’. But the idea that culture is transmitted by imitation and learning in a manner analogous but not reducible to natural selection has been around for a long time, and many other terms have been used in describing it. Darwin himself, when he came to consider ‘the causes which lead to the advance of morality’, concluded that ‘natural selection effects but little’ and looked instead to such things as ‘the approbation of our fellow-men’ and ‘example and imitation’. Much later, when the mechanism of natural selection had come to be understood in greater depth and detail than had been possible for Darwin himself, the American psychologist Donald Campbell and others began to develop the notion that cultural evolution (including the emergence of such things as theories of cultural evolution) is also driven by a competitive process of variation and selection. Campbell called his Presidential Address of 1975 to the American Psychological Association, ‘On the conflicts between biological and social evolution and between psychology and moral tradition’. But how does cultural evolution actually work?

Any reader sufficiently interested in that question to read this review is likely to be aware that current discussion of it is permeated by controversy. The controversies are of two kinds. For some opponents of ‘memetics’ in any guise, the mere idea of it is totally misconceived, whether because (as the Reductionists argue) cultural evolution is really all about the workings of our genes, or because (as the Creationists argue) it’s really all about the souls implanted in us by God, or because (as the Post-Modernists argue) no so-called theory of cultural evolution will ever explain it any better than any other so-called theory. There are also not a few anthropologists and sociologists for whom the mere association of ‘memes’ with the name of Dawkins is enough to rule them out of mention in correct academic society. But to those, including myself, for whom the ideas of heritable variation and competitive selection are much the most serious game in town when it comes to explaining, and not merely narrating, the course of human sociocultural evolution, the issue is not whether to incorporate culture as such within neo-Darwinian theory, but how.

The disanalogies with natural selection are obvious, and Susan Blackmore is well aware of them. There aren’t, literally speaking, lineages of memes as there are of genes, and memes, unlike genes, can be discarded by the people who carry them and replaced by others. Moreover, the relation of the information encoded in memes to the rituals and artefacts of a given culture is not at all the same as that of the information encoded in genes to the organisms which carry them. But it doesn’t follow that memes can’t therefore be replicated and diffused by what is still a process of heritable variation and competitive selection. Some neo-Darwinian theorists, although they agree that in the human species genetic is supplemented by cultural inheritance, insist that the function of memes is still to serve the interests of genes: cultural mutations will continue to be replicated only if they somehow enhance, on average, the inclusive reproductive fitness of their carriers. But although that may be true in the long run, it demonstrably isn’t true in the short – and the short run covers some of the widest and most significant divergences between the cultures of the different human populations in the ethnographic and historical record. Blackmore firmly, and in my view rightly, sides with those who maintain that ‘memetic’ mutations can be culturally adaptive – that is, can be such that their continued replication will be favoured by the relationship of their carriers to their environment – even when they are at the same time genetically maladaptive – that is, such as to reduce the probability of replication of the genes of the people who carry them.

But what is – or isn’t – a ‘meme’? Blackmore sensibly eschews any attempt to find precise ‘memetic’ analogies for genetic concepts like alleles, meiosis and so forth. But having reviewed some of the many conflicting opinions in the current academic literature, she concludes that the best thing to do is to ‘use the term “meme” indiscriminately to refer to memetic information in any of its many forms; including ideas, the brain structures that instantiate those ideas, the behaviours these brain structures produce, and their versions in books, recipes, maps and written music’. This is not only confusing, but defeatist. It is true that there is no single model which can capture everything involved in the transmission of culture from adjacent or successive populations to one another. But if ‘memetics’ is, broadly speaking, about the way instructions affecting phenotype are transmitted from mind to mind by imitation and learning, it is presumably worth drawing an explicit distinction between the instructions transmitted, whether visually, orally or in written form, by parents, teachers, peer-group members or role-models, and the behaviour which the instructions dictate. Nobody can deny that such transmission is happening every day. Why, then, should even those whom Daniel Dennett calls ‘Darwin-dreaders’ object to the adoption of a single word for whatever specific bundles of instructions affecting phenotype – whether encoded in gestures, speech, writing or any other medium – do in fact guide the behaviour of the members of different cultures or sub-cultures?

The carriers (or as Blackmore prefers it, ‘vehicles’) of ‘memes’ are then the minds in which the instructions are held, and the behaviours which result from following the instructions – cooking soups, building churches, singing songs, wearing hats, performing plays, giving birthday parties – are the memes’ phenotypic effects. The instructions do not, however, have to be strictly deontic in logical form, or strictly prescriptive about every detail of behaviour. ‘Memes’ can include discrete units of memorable information (whatever they are) within the totality of values and beliefs which guide our cultural behaviour, as well as mental representations (however they function) of recipes, drawings, musical scores, prayer-books, folk maxims, manuals of etiquette and so forth. The different individual ‘cultures’ of the human populations in the ethnographic and historical record then consist of sets of related ‘memes’ (Blackmore calls them ‘memeplexes’) which have been handed down sufficiently widely and accurately over a sufficient number of successive generations to amount to a definable cultural tradition.

Not even the blessing of the OED, however, has converted to ‘memetics’ more than a handful of the increasing number of behavioural scientists, from archaeologists to linguists to economists, whose studies of one or another aspect of sociocultural evolution are conducted within a recognisable neo-Darwinian paradigm. This may in part be due to an innate suspicion of neologisms. But it is also due to a widespread feeling that ‘memes’ sound too much like ‘genes’ for the significance of the disanalogies between them to be allowed for, that the use of the word provokes more definitional quarrels than it resolves evidential disagreements, and that until we know much more than we do about where and how instructions affecting phenotype are stored in the human brain it is premature, if not seriously misleading, to talk about their replication and diffusion as if we did. Blackmore’s answer is to remind us that the theory of natural selection made enormous advances long before the underlying chemistry was understood and to claim that for all our ignorance about how memes are stored and transmitted ‘we certainly know enough to get started.’ But even those of her readers who are willing to assent to that proposition may be startled by her headlong gallop into problems as complex and intractable as language, altruism, religion and human self-consciousness itself.

The puzzling question about language, as Darwin observed, is why linguistic ability hasn’t evolved in other apes as well as humans. In both humans and other primates, there is a correlation between increasing group size and increasing size of the neocortex. But what were the ecological conditions which allowed the benefits of progressively bigger brains to outweigh the heavy energetic costs? And what exactly were those benefits? As Blackmore acknowledges, we have almost no idea when our remote forbears started using language, let alone what language was first used to say. Her claim that their brain size was ‘created by memes’ – ‘memes’ here standing for the newly evolved capacity for genuine imitation, as opposed to mere stimulus enhancement or operant conditioning – is purely speculative. It isn’t even as if genuine imitation hasn’t been observed in other species. Primatologists are divided about what sort of evidence is conclusive for this purpose, but Darwin had no doubt that ‘apes are much given to imitation.’ Might not language have helped the replication of memes more than memes helped the replication of genes responsible for the ‘language instinct’? Without a more detailed model of the feedback between the greatly enhanced capacity for imitation which Blackmore assumes to go with language and the physiological and anatomical changes which language required, it is difficult to see how the strength of her claim can be assessed. Her fellow-psychologists would probably agree that there is strong evidence for a link of some kind between the acquisition of language and the acquisition of a theory of mind which enabled our remote forebears to conceptualise other people as such. But nobody (yet) knows what it is, and invocation of ‘memes’ isn’t enough to furnish the answer.

Similarly, it is far from clear in Blackmore’s discussion of altruism how the strength of her claim that ‘meme-driven’ altruism spreads because people ‘infected with altruistic memes . . . are copied more than other people and so these memes spread more widely’ is to be tested. There is no lack of evidence that human beings will, under some conditions, behave generously towards others who are not their genetic relatives, and will sometimes do so even where they have no reason to expect that the recipients of their generosity will at some later stage do them a favour in return. But equally, there is no lack of evidence that under other conditions, human beings will cheat and exploit others who are their genetic relatives, and that once some members of a group or community start to renege on social contracts it is the behaviour of those who do so which the others will increasingly imitate. Particularly in large groups whose members are not genetically related, it is not at all easy to see how altruism can become the evolutionarily stable strategy uninvadable by defectors who will be imitated in preference to the altruists. So under what conditions can Blackmore confidently predict that altruistic strategies will be sufficiently widely imitated that they will come to dominate selfish ones (both being, presumably, ‘meme-driven’)? There is by now an extensive literature on this, much of it grounded in evolutionary game theory, including not merely formal models but field-studies and social-psychological experiments. It would be unreasonable to expect Blackmore to cover it in a single chapter of a short book which has in any case been written more as a manifesto than an encyclopedia. But the implication that it can all be absorbed in, or has been overtaken by, a ‘memetic theory of altruism’ whose ‘main assumption is that people preferentially copy the people they like’ may well leave the authors of that literature somewhat underwhelmed.

About religion, where sociobiology has the hardest time linking ‘meme-driven’ behaviour to inclusive reproductive fitness, Blackmore is on firmer ground. It is altogether more plausible to suppose that religious memes are replicated and diffused because of their selective function at the level of the group than because their carriers are individually more likely to have more children (although there are, admittedly, some well-documented religious communities where this does hold good). Instructions affecting phenotype which are underwritten by beliefs in supernatural sanctions, are encoded in sacred books, and enjoin the performance of elaborate ceremonies may well give the group a competitive advantage over other groups and thus enhance the probability of replication of memes carried by its members relative to theirs. Moreover, it may well do so by at the same time reinforcing altruism within the group through a mixture of punishments and rewards and permitting or even encouraging deceit, manipulation and aggression in relationships with the members of other groups.

Blackmore seems inclined to regard all religions as ‘memetically selected’ bundles of ‘clever tricks’ which fool people into believing things which are either untestable or false (as, let us remind ourselves, science often does, too). But this is to ignore the distinction between genuine conversion and mere ‘adhesion’, as Arthur Darby Nock called it. No doubt religions have often been helped in their quest for converts by apparently successful prophecies, answers to prayer and miraculous interventions in the natural course of events. But for many of its (mere) adherents, religion is not a matter of intellectual commitment so much as emotional involvement in the mourning of the dead, the celebration of victory, the re-enactment of myth, the collective expression of hope or fear, and the symbolising through rites of passage of transition from one stage of life to the next. If ‘memetics’ is to contribute anything significant to what is already well known to anthropologists, sociologists and historians of religion, it will have to be by identifying specific components of different religious ‘memeplexes’ whose probability of replication can be shown to have been enhanced by specific features of their carriers’ environment – a rewarding task, but a very difficult one.

When it comes to self-consciousness, Blackmore appears to assume that her readers will all be unreconstructed Cartesian dualists still clinging to what Gilbert Ryle famously called the ‘ghost in the machine’. But how does ‘memetics’ resolve the disjunction between the acknowledged findings of psychology and brain science on the one hand and the persistent phenomenological sense of self on the other? For Blackmore, the sense of self is another ‘clever trick’, which works because the various memes by which we are ‘bombarded’ can, once they get into our heads, ‘gain an advantage by becoming associated with a person’s self concept’. But if our sense of self, illusory as it may be, is intrinsic to any and all our beliefs, values and preferences, how does it influence the relative probability of replication of some memes as against others? On a ‘memetic’ view, competition between memes is what drives the various cultures between which the human species is distributed down one rather than another evolutionary pathway. But what actually happens in a culture where anti-Cartesian memes propagated by psychologists and philosophers start to invade minds in which what Blackmore calls the ‘selfplex’ has previously been undisturbed? If memes ‘associated with a person’s self concept’ have, on average, an advantage, are anti-Cartesian memes a mutation doomed to extinction? Or will they somehow combine with the memes constitutive of our phenomenologically unreconstructed beliefs, values and preferences in such a way as to be replicated along with them? And if they do, what difference will it make to which of the bundles of instructions by which our cultural behaviour is directed? Will we cook soups, build churches, wear hats, perform plays or give birthday parties any differently as a result?

Despite these reservations, nothing in this review is meant to imply that Blackmore’s emphasis on the significance of imitation in cultural evolution is necessarily misplaced. The relation of imitation to individual learning is itself a matter of some dispute. But the acquisition of an entire repertory of cultural behaviour by individual trial-and-error would be prohibitively costly, and the human capacity to absorb and retain rules of conduct observed in the behaviour of other people goes far beyond what could ever be achieved by operant conditioning and stimulus enhancement alone. On the other hand ‘socialisation’, so called, is not just a matter of children doing what they see their elders doing. It is also a matter of their doing what their elders tell them to do even if the elders don’t even pretend to do it themselves: ‘don’t do what I do, do what I say!’ And what about cultural behaviour which is the result not of imitation or learning but of enforced obedience to instructions from lawgivers and other people in positions not merely of influence but of power?

It may be that I am here disclosing a professional bias, since I am not a psychologist or anthropologist but a historical and comparative sociologist more at home in topics like class conflict, the formation of states and the displacement of slavery by wage-labour. But there is more to cultural evolution than just the relative psychological attraction of different memes. Blackmore sweepingly asserts that ‘social scientists study the way that people’s lives and selves are constructed by their roles, and by the texts in which they are embedded. But they have no evolutionary theory within which to understand the processes going on.’ That, however, is simply not so. I am not the only sociologist in the world who sees social roles as part of the same continuous evolutionary process out of which first genes and then memes have emerged and which stretches all the way from the chemical evolution of the nucleic acids to ‘intelligent’ machines. But social roles involve a further mode of information transfer affecting phenotype. As well as instructions transmitted genetically by inheritance and instructions transmitted memetically by imitation and learning, there are instructions encoded in formal rules underwritten by institutional inducements and sanctions which define the practices that make our roles what they are; about these Blackmore has little or nothing to say.

How successful, then, is The Meme Machine likely to be as, under Blackmore’s definition, a meme designed to further the replication and diffusion of the ‘meme’ meme? The answer, I suspect, is that it is unlikely to do for public understanding of the meme-centred approach to cultural selection what Dawkins’s The Selfish Gene has done for the public understanding of the gene-centred approach to natural selection. This isn’t just because Blackmore doesn’t have Dawkins’s altogether exceptional gift for accurate synthesis and lucid exposition. It is because she attempts too much too soon. No author of a book like this one can yet draw on the large volume of well-tested hypotheses and closely-documented examples available to writers on biological theory. Only detailed and wide-ranging re-examination of the relevant ethnographic, historical and archaeological evidence for the heritable variation and competitive selection of clearly identifiable ‘memes’ whose spread can accurately be traced and functions convincingly specified will turn ‘memetics’ from a project into a science.