The meme’s eye view

In Darwinizing Culture: The Status of Memetics as a Science, Ed. R.A.Aunger, Oxford; Oxford University Press.

The new replicators

Robert Aunger, in his Introduction, has suggested that memeticists such as myself face a challenge: either to provide an existence proof for memes, or to come up with supported, unique predictions from meme theory. I suggest, however, that no existence proof is required and we would do better to concentrate on whether meme theory can be of any scientific value or not.

The reason no existence proof is required is the way ‘meme’ is defined. When Dawkins (1976) first coined the term he wanted an example of a replicator other than the gene. He based the name for his new cultural replicator on the Greek word mimeme — meaning that which is imitated. He intended imitation “in the broad sense” (a point to which I shall return) but was very clear that whatever is passed on when people imitate each other – that is the meme. This clarity is reflected in the new Oxford English Dictionary definition of “meme (mi:m), n. Biol. (shortened from mimeme … that which is imitated, after GENE n.) An element of a culture that may be considered to be passed on by non-genetic means, esp. imitation”. Although many authors use widely differing definitions I suggest we stick to this simple one. Doing so avoids many problems. It also becomes clear why no existence proof is required. As long as we accept that people do, in fact, imitate each other, and that information of some kind is passed on when they do then, by definition, memes exist.

We might, however, be a little more strict in our requirements and demand that memes must be shown to be replicators to count as existing. To be a replicator something must be capable of sustaining the evolutionary process of heredity, variation, and selection (Dawkins, 1976) or blind variation with selective retention (Campbell, 1960). It must, as Dennett puts it, undergo the evolutionary algorithm – that blind, mechanical procedure which creates “Design out of Chaos without the aid of Mind” (Dennett 1995, p 50).

Whichever scheme you prefer, memes fit. By definition they are inherited because they are passed on by imitation. They undergo selection in the sense that people are exposed to far more memes than they can possibly remember, let alone pass on again. And memes vary, whether by degradation (as occurs with errors of perception, memory, or reconstruction) or by creative recombination (as when different memes are put together to produce new combinations). The former is not helpful to memetic evolution in that memes are likely to lose any of the “good tricks” they have accumulated (Dennett, 1995). Recombination should be a more effective way of producing viable memes that will outperform those produced by degradative variation. In any case, memes clearly vary and therefore fit neatly into the evolutionary algorithm. In other words memes are replicators. The importance of this is that replicators are the ultimate beneficiaries of any evolutionary process. Dennett (1995) urges us always to ask Cui Bono? or who benefits? and the answer is the replicators. This means that if we have a new replicator – the meme – there is a new entity whose interests must be taken into account.

I suggest that no further proof of the existence of memes is required.

The interesting question then, is not whether memes exist, but whether taking the memes’ point of view can lead to any useful scientific work. In other words, is memetics a worthwhile endeavour? I believe that it is — not just because I am enjoying the startling new meme’s eye vision of the world — but because memetics provides new solutions to old problems, among them the origins of our large brain with its specialised language and unique intelligence.

Why the emphasis on imitation?

Before I discuss the advantages of the memetic perspective, I would like to consider one further issue associated with the definition of memes. I have chosen to stick to Dawkins’s original formulation of memes as information that is passed on by imitation. Others differ here. For example Cavalli-Sforza and Feldman (1981) base their model of cultural transmission on traits that can be passed on by imprinting, conditioning, observation, imitation or direct teaching. Durham’s (1991) coevolutionary model refers to both imitation and learning. Runciman (1998) refers to memes as instructions affecting phenotype passed on by both imitation and learning. Laland and Odling Smee (THIS VOLUME – PLEASE ADD REFERENCE BOB) argue that all forms of social learning are potentially capable of propagating memes.  Among meme-theorists Brodie (1996) includes all conditioning as memetic, and Gabora (1997) counts all mental representations as memes regardless of how they are acquired.

My reason for restricting meme acquisition to imitation (i.e., excluding other kinds of learning) is my suspicion that only imitation is capable of sustaining a true evolutionary process (Blackmore, in press). In individual learning (such as imprinting, classical conditioning and operant conditioning), nothing is copied from one individual to another, so there is no basis for a replicator to operate. In other forms of social learning, such as stimulus enhancement or local enhancement, the behaviour of two individuals is involved and that of the learner ends up similar to that of the original performer, but the behaviour is not copied from one individual to another. For example, cultural traditions such as tits learning to open milk bottles or chimpanzees using sticks to fish for termites, are thought to spread by stimulus enhancement. Each individual learns the skill anew, having had its attention drawn to the location, the available materials or the stimulus of a pecked bottle top. In such traditions, as Tomasello, Kruger and Ratner (1993) point out, there is no accumulation of modifications over generations – no cultural ratchet effect. Similarly, Boyd and Richerson (THIS VOLUME – PLEASE ADD REF HERE BOB) argue that only observational learning of novel behaviours allows cumulative cultural change.

Jablonka (1999) provides the useful distinction between reproduction and replication of behaviours. You could say that in other forms of social learning the same behaviour is apparently reproduced (such as washing sweet potatoes or pecking at milk bottle tops), but it is not replicated —  i.e. copied. This means there is no opportunity for variations on the copied behaviours to compete with each other, for truly novel behaviours to spread, or for cumulative change. In other words, without imitation there is no replicator and no new evolutionary process.

To some extent this difference could be seen as an issue of copying fidelity. You could argue that other forms of social learning can reproduce new behaviours with sufficiently high fidelity to count as replication and to sustain evolution. This is an empirical question worth researching if these issues are to be resolved (Blackmore, in press). The question would be which kinds of social learning can reproduce behaviours with sufficient fidelity to maintain them intact over several generations of copying, and to allow for selection between variants and for cumulative change. Such research may reveal that in fact other kinds of social learning can sustain such an evolutionary process, in which case they should be included as processes that replicate memes. However, working without such information and with the current uncertainties over definitions, I would argue that only imitation has the capacity to sustain an evolutionary process and this is a good reason for restricting the definition of memes to that which is imitated.

There is also the related question of whether you choose to apply the word “culture” to behaviours that are spread by other forms of social learning. If you do, then some monkeys, rats and birds have culture but, by my definition at least, they do not have memes. On the other hand, dolphins, some song birds, and possibly elephants and chimpanzees do have memes because they are capable (at least to some extent) of copying novel behaviours or sounds by imitation.

A different question arises at the other end of the scale when we think about memes passed on by complex human processes such as reading, writing and direct instruction. I presume that Dawkins meant to include these when he used the phrase “imitation in the broad sense”. We may not wish to count these as forms of imitation, but I would argue that they build on the ability to imitate and could not occur without it. Learning language requires the ability to imitate sounds, and instructed learning and collaborative learning emerge later in human development than does imitation and arguably build on it (Tomasello, Kruger & Ratner 1993). All these complex human skills clearly entail the copying of information from one person to another. Variation is introduced both by degradation due to failures of human memory and communication, and by the creative recombination of different memes. Selection occurs because of the limitations on available communication channels, time, memory and other sorts of storage space. Information passed on by these means therefore fits the evolutionary algorithm.

A final problem concerns creativity. Many people seem to think that imitation is a crude and blindly mechanistic process that is the antithesis of human creativity, which is conscious and purposive. This is indeed a very different view from my own, and entirely misses the point that evolutionary processes are creative — arguably the only creative processes on the planet. The alternative, first sketched out by Campbell (1960), is that just as biological creations come about through natural selection, so human artistic, literary and scientific creations come about through memetic selection. In both cases the creative force is the evolutionary algorithm. Human achievements are no less creative for that, but our own role has to be seen as that of the clever imitation machine taking part in this new evolutionary process, rather than a conscious entity who can stand outside of it and direct it.

The human brain

I suggest that memetics can provide a new explanation for the origins and evolution of the human brain. Since memes are, by definition, passed on by imitation, they must have first appeared when our ancestors became capable of imitation. This would have made an enormous difference in evolutionary terms because memes were a new replicator that started evolving in their own way and for their own replicative ends. Since that time human evolution has been driven by two replicators, not just one. I suggest that this is why humans are unique. It was the advent of the new replicator that changed the ground rules forever. Since then meme-gene coevolution has produced the enormous human brain which is designed not just for the benefit of genes, but for the propagation of memes.

The sheer size of the human brain requires some kind of evolutionary explanation. It is roughly three times larger than would be expected for an ape of our size and weight. It uses a prodigious amount of energy both to produce and to run, and is dangerous to give birth to. Not only is it unusually large but it has been restructured  in various ways and apparently specially adapted for the production and comprehension of language.

Early theories to explain the big brain focused on hunting and foraging skills, but their predictions have not generally held up, so more recent theories have emphasised the complex demands of the social environment (Barton & Dunbar 1997). Chimpanzees live in complex social groups and it seems likely that our common ancestors did too. Making and breaking alliances, remembering who is who to maintain reciprocal altruism, and outwitting others, all require complex, fast decision making and good memory. The “Machiavellian Hypothesis” emphasises the importance of deception and scheming in social life and suggests that much of human intelligence has social origins (Byrne & Whiten 1988; Whiten & Byrne 1997). Dunbar (1996) argues that gossip is the human equivalent of grooming in that it enables large social groups with complex relationships and reciprocal altruism to be maintained. This, he argues, explains the evolutionary advantage of language, and the need for language drove the increase in brain size.

Most of these theories entail gradual changes in abilities and brain size, but others include one or more transitions. For example, Donald (1991) proposes a three stage account of how human brains, culture, and cognition coevolved. His first step is a “revolution in motor skill” (Donald 1993 p739) that he calls mimetic skill. He argues that the anatomical changes necessary to support speech evolved in a mutually reinforcing manner with the lexical capacity. However, his term “mimetic” is quite unrelated to the term “memetic”. By mimesis he means “the ability to produce conscious, self-initiated, representational acts that are intentional but not linguistic” (Donald 1991 p 168). He specifically excludes “simple imitative acts” and concentrates on the importance of representation — whether externally to someone else or internally to oneself. This emphasis on symbolic representation makes Donald’s theory quite different from the one proposed here, which rests entirely on the premise that copying actions from one individual to another creates a new replicator. Whether those actions represent anything or are symbolic is quite irrelevant to their role as replicators. Also Donald’s theory, like most other theories of human evolution, ignores the possibility of a second replicator, and treats all adaptations as ultimately for the benefit of genes.

One possible exception is Deacon’s (1997) theory of the coevolution of language and the human brain. Deacon argues that once that simple languages appeared, they created selection pressure for bigger and better brains able to understand them. Although he does not use the term ‘meme’, he likens language to a parasitic organism with some of its features evolved for the purpose of passing the language on from host to host, even at the expense of the host’s adaptations. He refers to symbolic adaptation as a ‘mind virus’ that has turned us into the means for its own propagation (Deacon, 1997, p 436). However, his theory differs from the one I am proposing here in that the critical turning point was not the appearance of imitation but the point at which our ancestors crossed the “symbolic threshold”. For Deacon symbolic reference provided the only conceivable selection pressure for the evolution of hominid brains.

These various theories differ in many other ways but most share the conventional neo-Darwinian assumption: that the human brain was designed by evolution for the benefit of genes. In other words, their answer to Dennett’s Cui Bono? question is genes. I propose, instead, that the human brain was primarily designed for the benefit of memes.

Memetic drive

I propose that there was a critical turning point in human evolution: when our ancestors acquired the ability to imitate. From this point on, memes started driving genes to produce a brain that was especially good at replicating those memes.

Imitation can be a “good trick” from the genes’ point of view because it reduces the costs of learning. We might liken imitation to stealing learned behaviour from some one else without having to take the risks involved, or put in the time and effort needed, to acquire it by trial and error or other forms of individual learning. Mathematical modelling has shown that social learning, including imitation,  is worthwhile if the environment changes, but not too fast (Richerson & Boyd 1992). The point is that, although imitation might initially benefit the genes of the person who could do it, those genes had no foresight. They could not predict that they were letting loose a new replicator; one that need not “be subservient to the old” (Dawkins 1976, p 194).

Although speculating on the lives of our early ancestors is always dangerous, I would guess that the first memes were useful ones (that is, useful to the genes), such as new ways of hunting or preparing food, ways of making baskets or simple tools, or dealing with social relationships. However, once imitation was possible, memes could spread for many reasons other than their value to the genes that gave rise to them in the first place. So other not-so-useful memes would soon begin to exploit the new copying machinery and spread by imitation as well. These might include rituals, body decoration, burial rites or music. In even a simple culture like this we have the basic ingredients for what I have called “memetic driving”.

The mechanism works like this. The people who are best at imitation have an advantage over the rest because they can most easily acquire any useful new skills or artefacts, and most easily put together old memes to produce new ones — we may call these people “meme fountains”. As long as there is some genetic basis to what made them meme fountains in the first place, then genes for being good at imitation will tend to spread (on ordinary Darwinian principles). Assuming that imitation is a difficult skill that requires a bigger brain, we have a simple argument for the increase in human brain size — although this far my argument is the same as many previous theories.

The next step is that once memes are around, everyone has to start making decisions about whom and what to imitate. In general it will pay others to copy the meme fountains because they are more likely to have useful survival-related memes. This gives a further survival advantage to the meme fountains, and their genes, in terms of improved power and status. If there are genes for imitating the best imitators, these genes will also spread in the gene pool. However, this may mean copying a fancy head-dress, or a pleasing song or dance, as well as a new way of making stone tools or baskets. Memetic evolution now gets under way with various kinds of dances, head-dresses and songs competing with each other to be copied.

We now have two effects operating. First, everyone gets gradually better at imitating the successful memes, which means that more and more memes are created and culture expands. Second, genes for the ability to copy meme fountains and their popular memes have an advantage and more people come to behave this way. But this now creates selection pressure for the ability to discriminate between useful and useless memes (that is, useful or useless from the genes’ point of view), because copying a popular meme just might prove fatal. As memes evolve in one direction or another, according to the outcome of memetic selection and the kinds of memes the meme-fountains happen to be best at propagating, survival increasingly depends on the ability to choose which memes to copy and which to avoid.

This is essentially the basis of memetic drive. Memes compete with each other to be copied and the winners change the environment in which genes are selected. In this way, memes force genes to create a brain that is capable of copying the successful memes of the time.

One final step to the argument is that for similar reasons there may be an advantage to mating with the meme fountains. Sexual selection may therefore add to the pressures on genes to produce brains capable of imitating the currently successful memes.

This opens the way for an explanation of how the brain has been designed for language and other specialised abilities. The argument depends on the power of the successful memes, so which are they? The answer, according to general principles of evolution, should be memes of high fidelity, fecundity and longevity (Dawkins, 1976). Language is a good way of creating memes with high fecundity and fidelity. For example, sound carries better than visual stimuli to several people at once. Sounds digitised into words can be copied with higher fidelity than continuously varying sounds. Using different word orders in different circumstances opens up more niches for memes to occupy and so on. For this general reason we should expect language memes to succeed in memetic evolution, and then memetic driving to cause the spread of the genes that make that language possible. That is, in an environment in which simple language is spreading memetically, the meme fountains will have the best command of the new language because they are good at imitation, while the people who cannot pick it up will be at a disadvantage in a way they never would have been before language appeared. In addition, those who are best at picking up the new language may be preferentially chosen as mates. For these reasons, any genes involved in the ability to copy the language will tend to spread. As the evolving language changes through memetic competition, so genes are forced to follow. On this argument, the function of language is to spread memes. The genes had no choice but to follow where the memes led and produce a brain that was not only as big as the genes could carry, but was designed especially for propagating memes through language.

Is this theory testable? Some of the assumptions on which it depends could be tested. For example, it assumes that imitation is a difficult skill that requires a lot of processing capacity. Brain scan studies might refute this if it turned out that imitation does not use large areas of the brain, or that it does not involve the evolutionarily newer parts of the human brain. Computer simulations and mathematical models are already being used to test whether memetic drive really could produce an increase in brain size. For example, Higgs (under review) has developed a model in which memes can have both positive and negative effects on the fitness of a population of individuals. He not only found that genes for imitative ability are selectively favoured but that imitative ability increases slowly until a rapid transition occurs, after which memes spread like an epidemic and there is a dramatic increase in imitative ability as well as mean fitness. Kendal and Laland (under review ) have built on models from gene-culture coevolution theory and shown that the strategy of imitating enhanced imitators will spread under a wide variety of conditions, opening up new ways of testing these hypotheses.

Not just big but special

Since I first proposed this argument (Blackmore, 1999) many colleagues and critics have raised difficulties or questions about the proposed process of meme-gene coevolution. In particular, some critics of my book The Meme Machine have gained the impression that I believe the human brain is an all-purpose meme machine, designed to copy any old memes, and that its size is the only mystery to be explained. This is clearly not so, for our copying is highly selective. From soon after birth infants imitate facial expressions, hand movements and so on, but they do not imitate just anything they see; their imitation is selective (Brugger and Bushnell 1999). Adults imitate speech, and certain kinds of actions and behaviours, but not others. In view of this criticism, I would like to make clear the implications of the memetic driving hypothesis.

The theory is meant to be an argument of the following general form. Once memes arise they evolve by memetic competition (high quality memes spread at the expense of low quality memes) and they evolve faster than the genes that made them possible in the first place. Meme fountains (who have all the useful memes as well as all the ones that have spread for other reasons) survive better both because they have more useful memes and because other people copy them, giving them added power and status. So the memes that succeed in memetic competition change the environment in which genes are selected, giving an advantage to genes which help a person imitate the currently successful memes — whatever those memes happen to be. In addition, meme fountains may be selectively chosen as mates, though this is not essential to the argument.

I used this argument to provide an explanation for the “language instinct” or “language organ” but this may have been a poor choice not least because it is such a contentious issue. I will therefore try the argument out on something less contentious — the human enjoyment of music. Why do we humans, apparently alone among animals, invest large amounts of time and energy in producing and enjoying complex music? It is difficult, if not impossible, to provide an answer based on advantage to human genes (Pinker, 1997). Dennett (1999) imagines an early hominid, for no particular reason, just happening to bang on a log and enjoying the sound, and someone else coming along and copying it. For reasons to do with perceptual systems, memory, or features of the environment, some versions of the drumming, and then humming, are more infectious and spread at the expense of others. And so the process goes on, with the advantage being to the bangings, whistlings and hummings (i.e. the memes) — not necessarily to the hominids’ genes. Dennett then speculates that females might be more receptive to the winning hums.

Note that the idea that culture evolved by sexual selection is not new. Miller (1999) has argued that human culture in general, and music and art in particular, are mainly a set of adaptations for courtship. He cites evidence that musicians and artists are predominantly male and at their most productive during young adulthood. His theory does not, however, involve memes and is therefore slightly different from that proposed here. The difference is this. On Miller’s theory the songs (or other productions) are the cultural displays that guide females in their choice of mates – comparable to the peacock’s tail. Presumably the songs evolve only because of differential female choice. However, on Dennett’s theory and on the memetic driving hypothesis proposed here, the songs themselves compete to be copied. This memetic competition takes place in both males and females; the outcome being determined by features of the songs themselves (e.g. how easy they are to sing or remember) and of the perceptual systems and vocal tracts of the people trying to copy them. Meme-meme competition thus determines the direction taken by the evolution of music, dance, art and literature, as well as sexual selection.

Exactly the same argument can be applied to religions. This is also a contentious topic and relates to Dawkins’s (1993) suggestion that religions are viruses of the mind. He pointed out that some of the world’s great religions may have spread not because they are true, nor because they help anyone, but because they are successful memes — successful because they are essentially “copy-me” instructions backed up with threats, promises and ways of preventing their claims from being tested. Rather than discussing a vast memeplex like Roman Catholicism, we may take the simpler example of a ritual dance supposed to bring rain. The rain dance may, by chance, coincide with the advent of rain, and so be copied. If some meme fountain does a particularly flamboyant version this meme may be copied even more, outperforming other versions. Being powerful in this society (and hence acquiring a survival advantage) now becomes linked with being able to copy these winning dances. People not only copy these successful memes, but mate with the people who display them, so that any genes implicated in being good at these dances (or prayers, or fervent displays of belief in God, or singing hymns, or …. ) will tend to increase. We end up with brains specifically designed to pick up and copy religious memes. I suggest this is why religion, belief in God, and religious ritual still thrive in a modern scientific culture that rationally rejects them. Our brains are especially good at picking up these kinds of memes because of our long history of coevolution with them.

The argument takes the same form as before. Successful memes spread. They then change the environment in which genes are selected. The consequence is a brain that is better designed for spreading those particular memes.

The brain is good for genes as well as memes

In claiming that the human brain was built to copy memes, I have perhaps implied that it was of no benefit at all to genes. Yet our big brains have clearly provided all sorts of survival benefits enabling us to occupy widely varying niches all over the planet. My mistake was perhaps to over-emphasise the role of the most arbitrary, useless or even dangerous memes. But I did this mainly because it is on this issue that memetics diverges most strongly from traditional sociobiology or gene-culture coevolution theory.

In these disciplines genes provide the capacity for culture. Maladaptive (for genes) traits can arise and can even persist (Cavalli-Sforza and Feldman 1981, Feldman and Laland 1996) but competition between these traits is not taken into consideration, and benefit to the traits themselves is not the driving force. On the model I am proposing here, memes compete with memes and the outcome of that competition affects genes. Many memes survive precisely because they are useful to genes, but others survive for other reasons. They are not just maladaptive for genes; they are adaptive for themselves and for the memeplexes of which they are part. The whole package – the two-replicator creature – is an extremely efficient survival machine but we can only understand it by considering the effects of memetic competition. These effects are most obvious when they run counter to the interests of genes, so this is why I have tended to emphasise them.

I admit I had imagined the brain as having been driven to its huge size mostly by viral memes — in other words being somewhat analogous to a parasite that must be carried at some cost to genes — but this raises the interesting question of whether the brain is best seen as parasite, symbiont, commensal or something else (a question raised in discussions with Pinker, Dennett and some of their students). This has led to the following analogy.

We can imagine the brain as analogous to the immune system. Memetic driving forces genes to produce a bigger brain especially good at copying any successful memes that are around. The genes fight back by producing ways of selecting only memes that are useful for them. This requires a complex system for recognising which memes are useful and which not — something like the way the immune system has to recognise self from invader.

An example may help. Suppose that our hypothetical meme fountain is especially good at hunting with the latest tools, as well as dancing the latest flamboyant rain dance and flaunting his status by wearing the latest clothes. He has a survival advantage and so any genes he has that predispose him towards copying these memes are passed on. Other people copy him because he has the best memes, but there is another competition going on here. The people who selectively copy the useful tool skills from him while ignoring the dance will do better (biologically) than those who copy any and all of his memes. Although the simple heuristic — copy the meme fountain — works up to a point, the ability to select genetically beneficial memes from among those displayed by the meme fountain does even better. Meanwhile, memes are moving on in their purely memetic competition, outsmarting whatever selective tricks genes have come up with so far and putting on more pressure to be able to select among memes even more cleverly. The result is a brain that is very good at imitation, highly selective, and whose selective capabilities have been shaped by memetic competition.

Whether this comparison will prove useful I do not know. However, the central point here is this — on this theory the brain was designed to copy successful memes, and this means both memes that succeed for purely memetic reasons, and memes that actually help survival of genes. In other words, it is a compromise between the forces of memetic and genetic evolution. Human intelligence is, in this view, all about the selection of memes, and future research should focus on which memes we do and do not copy, and why. This is a new way of looking at the function of human intelligence. The human brain is a selective imitation device.

Can memes get off the leash?

Lumsden and Wilson (1981) famously declared that “the genes hold culture on a leash”. Most people who have modelled gene-culture coevolution have agreed. Even Durham (1991) who is one of the few to use the term ‘meme’, and who provides examples of maladaptive traits that spread successfully, argues that organic and cultural selection work on the same criterion — inclusive fitness. As far as I can see only Cloak (1975), and Boyd and Richerson (1985) truly treat their cultural trait as a replicator in its own right — an idea that is fundamental to memetics.

My arguments about meme-gene coevolution implied a complex interaction between the two replicators in which each affects the other — two dogs on the same leash you might say. But then the question arises whether the new dog can escape the leash altogether.

Among the factors that may be relevant are whether memes are transmitted vertically (parent to child) or horizontally (between unrelated people, possibly of the same age) (Cavalli-Sforza & Feldman, 1981). A related issue is the relative speeds of change of the two replicators. If all memes are transmitted vertically then memetic change tracks genetic change and there can be no meaningful coevolution (indeed no leash at all). I have assumed that during most of human evolution memes were transmitted largely vertically, and changed at speeds not very different from human genetic change, but that there was sufficient horizontal transmission to make memetic driving possible. Nowadays, however, memetic transmission is very fast and largely horizontal. Although most people still get their first language, their basic social rules, and their religion from their parents, most of the memes they acquire during their lifetime come from school, radio, television, books, magazines, the Internet, their peers and even their own children.

In such an environment, genes can hardly be expected to track memes. They may still be affected, for example by memes such as birth control, technological medicine, genetic engineering and so on, but memes are moving too fast for any detrimental effect they may have to exert any control. If the memes you come across are going to kill you or prevent you having children, the demise of your genes will come too late to exert any control over the spread of those memes. In other words, memes are getting off the leash.

Can this idea be formalised in any way? Bull has recently used an artificial life model to simulate the interactions between two replicators of different speeds (Bull, Holland and Blackmore 1999). When there is low dependence between the ‘genes’ and ‘memes,’ relative speed makes no difference to either replicator, but with slightly higher interdependence, increasing the rate of meme evolution provides rapid benefits to the memes, and gene evolution degrades to a random walk. This is only a simple and abstract model but suggests ways in which some implications of meme-gene coevolution might be tested.

Even if human genetic evolution is now no more than a random walk, it could still be argued that memes depend on genes for their propagation because they still build the brains which carry out the imitation — and it is these brains, with their endless penchant for food, sex, and violence, that determine the success of magazines, TV programmes and websites. In this sense, then, memes cannot be truly independent.

However, we may indulge in science fiction speculations and imagine the day — possibly not so far off — when humans are no longer required to maintain the hardware of the Internet because self-replicating computers have been designed. Even without that step, we can easily imagine information that is copied in the Internet without any human making the decisions. For example, there are already sites that generate a new academic paper, complete with references and footnotes, every time you visit. Imagine a program that chooses among these and distributes copies to other sites and you have memetic evolution without human intervention. Another possibility is that simple programs that currently pretend to be human users of chat rooms and discussion lists will evolve into much cleverer memeplexes, selectively copying behaviours from each other and from human users and so acting as autonomous, evolving, meme-selecting devices.

Such speculations are always dangerous but I mention this only to make a last, general point about the replicator power of memes. If memes are true replicators in their own right, as I have assumed, then we should expect them to coevolve along with the machinery for their own replication. Genes have done this — the exquisitely accurate machinery that copies DNA cannot have sprung into existence fully formed, but must have evolved gradually from simpler copying mechanisms (Maynard Smith & Szathmáry, 1999). Memes are now doing the same. The process of meme-gene coevolution I have described can be seen as one step in this process — that is, the memes and the brains that copied them coevolved. But later steps are now far more significant. These include the invention of writing; the building of roads, railways and ships; the development of printing and books; the invention of the telephone, fax and mobile phone; and most recently the Internet. Each step has improved the methods by which memes can be copied and stored, and made possible the creation of ever more memes. In our new memetic view of the universe, we should see these great steps in copying technology not as inventions consciously created for our own benefit, but as the inevitable consequences of memetic evolution. And cui bono? The memes. It is this process that may one day let memes right off the leash.


Memetics provides a new vision of human nature, in which memes succeed wherever and whenever they can. Memes spread not necessarily because they benefit either the genes that made their evolution possible, or the survival chances or happiness of the people who copy them, but because they benefit themselves.

In this vision, all the cultural entities around me are there because they are the current winners in a fearsome competition to be copied. My own body is a meme machine designed by a long history of meme-gene co-evolution. It is furnished with plenty of memes it has already copied, and surrounded by masses more potentially copyable memes from which it has to choose.

On the optimistic side, there are several mechanisms by which altruistic behaviours may get themselves copied even though they are costly both to the person who carries them out and to their genes. Most simply, if altruistic people attract more friends who copy them then their altruistic behaviours obtain an advantage. Religions and cults may survive because they use clever memetic tricks to get themselves passed on, and to persuade their carriers to work hard and invest time and money in their propagation. Alternative therapies that don’t work may thrive in modern environments, because of the powerful placebo effect combined with a fear of high-tech medicine. Even bizarre ideas like four-foot high aliens who come and abduct people from their beds at night can usefully be seen as memes that succeed in spite of being false.

Perhaps most challenging is the idea that my inner self, which seems to have consciousness and free will, is in fact a memeplex created by and for the replication of memes. “My” beliefs and opinions are survival tricks used by memes for their own perpetuation. “My” creativity is really design by memetic evolution. On this view, human nature is a product of memes and genes competing for replication in a complex environment, and there is no room for mysterious guiding principles or inner selves with free will.

In these and other ways memetics might utterly transform our view of ourselves. Indeed I suspect that taking the memes’ eye view will provide as dramatic a transformation in our understanding of human nature as taking the genes’ eye view has done in evolutionary biology.


Barton,R.A. and Dunbar,R.I.M. (1997) Evolution of the social brain. In Machiavellian Intelligence II: Extensions and Evaluations. (ed A.Whiten and R.W.Byrne), Cambridge, Cambridge University Press, 240-263

Blackmore,S.J. (1999) The Meme Machine, Oxford, Oxford University Press.

Blackmore,S.J. (in press) Evolution and memes: The human brain as a selective imitation device. Cybernetics and Systems

Boyd,R. and Richerson,P.J. (1985) Culture and the Evolutionary Process. Chicago, University of Chicago Press

Brodie,R. (1996) Virus of the Mind: The New Science of the Meme. Seattle, Integral Press

Brugger,A.E. and Bushnell,E.W. 1999 Imitative strategies employed by 15- and 21-month old infants for learning to work novel objects. Poster, Conference of the Society for Research in Child Development, Albuquerque, April 1999

Bull,L., Holland,O. and Blackmore,S. (1999) On meme-gene coevolution. Submitted for publication.

Byrne,R.W. and Whiten,A. (Eds) (1988)  Machiavellian Intelligence: Social Expertise and the Evolution of Intellect in Monkeys, Apes and Humans.  Oxford, Clarendon Press.

Campbell,D.T. (1960) Blind variation and selective retention in creative thought as in other knowledge processes. Psychological Review, 67, 380-400

Cavalli-Sforza,L.L. and Feldman,M.W. (1981) Cultural Transmission and Evolution: A quantitative approach. Princeton, NJ, Princeton University Press

Cloak,F.T. (1975) Is a cultural ethology possible? Human Ecology, 3, 161-182

Dawkins,R. (1976) The Selfish Gene. Oxford, Oxford University Press

Dawkins,R. (1993) Viruses of the mind. In Dennett and his Critics: Demystifying Mind (ed. B.Dahlbohm)  Oxford, Blackwell

Deacon,T. (1997) The Symbolic Species: The Co-evolution of Language and the Human Brain. London, Penguin

Dennett,D. (1995) Darwin’s Dangerous Idea, London, Penguin

Dennett,D. (1999) The evolution of culture. Charles Simonyi Lecture, Oxford, February 17.

Donald, M (1991) Origins of the Modern Mind: Three Stages in the Evolution of Culture and Cognition. Cambridge, Mass, Harvard University Press.

Donald, M (1993) Precis of Origins of the Modern Mind: Three Stages in the Evolution of Culture and Cognition. Behavioral and Brain Sciences, 16, 737-791

Dunbar,R. (1996) Grooming, Gossip and the Evolution of Language. London, Faber & Faber

Durham,W.H. (1991) Coevolution: Genes, Culture and Human Diversity. Stanford, Ca., Stanford University Press.

Gabora,L. (1997) The origin and evolution of culture and creativity. Journal of Memetics, 1,

Higgs,P.G. (under review) Selection for learning by imitation gives rise to a mimetic transition in primitive cultures. BOB. THIS HAS BEEN ACCEPTED SUBJECT TO REVISION. BY THE TIME OF PUBLICATION WE MAY BE ABLE TO SAY IN PRESS.

Jablonka,E. (1999) Between development and evolution: the generation and transmission of cultural variations. Paper presented at “The Evolution of Cultural Entities” The British Academy, April.

Kendal,J.R. and Laland,K.N. (under review) Mathematical models for memetics. BOB – THIS IS WITH A FAST ELECTRONIC JOURNAL AND MAY BE PUBLISHED IN TIME TO REFERENCE PROPERLY.

Lumsden,C.J. and Wilson,E.O. (1981) Genes, Mind and Culture. Cambridge, Mass., Harvard University Press.

Maynard-Smith,J. and Szathmáry,E (1999) The Origins of Life: From the birth of life to the origin of language, Oxford University Press.

Miller,G.F. (in press) Sexual selection for cultural displays. In The Evolution of Culture(ed. R.Dunbar, C.Knight and C.Power) Edinburgh, Edinburgh University Press

Pinker,S. (1997) How the Mind Works. Penguin

Richerson,P.J. and Boyd,R. (1992) Cultural inheritance and evolutionary ecology. In Evolutionary Ecology and Human Behaviour (ed. E.A.Smith and B.Winterhalder) 61-92.

Runciman,W.G. (1998) Greek hoplites, warrior culture, and indirect bias. Journal of the Royal Anthropological Institute, 4, 731-751

Tomasello,M., Kruger,A.C. and Ratner,H.H. (1993) Cultural Learning. Behavioral and Brain Sciences, 16, 495-552.

Whiten,A. and Byrne,R.W. (1997) Machiavellian Intelligence II: Extensions and Evaluations. Cambridge, Cambridge University Press.

Buy this book from   or