The Evolution of Meme Machines

International Congress on Ontopsychology and Memetics
Milan, May 18-21 2002
Also published in Meneghetti, A. et al. (2003) Ontopsychology and Memetics. Rome, Psicologica Editrice, pp 233-240

Susan Blackmore

The science of memetics faces a serious problem. The concept of the meme emerged from evolutionary biology and the theory of replicators, and within this context it is well understood, if highly controversial. But out on the web, and in popular discourse, the word ‘meme’ is horribly abused. It is confused with ‘idea’ or ‘concept’ or treated as something ethereal or non-material floating about quite separate from behaviours and artefacts.

No one can prevent people from using using the term ‘meme’ in these other ways, but the power of the idea may be lost, and much confusion may be generated. So I hope that it may be helpful for me to outline the origins of memetics and the basic concept of the meme as a replicator in an evolutionary process. I will then use this to explore the evolution of human machines and of the other newer meme machines that have evolved from us.

The story begins in 1976 with the publication of evolutionary biologist, Richard Dawkins’s, best-selling book The Selfish Gene. This book popularised the growing view in biology that natural selection proceeds not in the interest of the species nor of the group, nor even of the individual, but in the interest of the genes. Although selection takes place largely at the individual level, it is the information in the genes that is copied. They are the replicators and it is their competition that drives the evolution of biological design.

In explaining this, Dawkins wanted to emphasise the principle of Universal Darwinism. Darwin’s fundamental insight was brilliant; an idea of such stunning power and simplicity that it has been called “the best idea anybody ever had”. It is this: – if living things vary in ways that affect how well they can survive, and if they produce more offspring than can possibly survive, and if the few survivors pass on their characteristics to the next generation, then the characteristics that helped them survive will be more common in the next generation. That is, the members of the next generation will have evolved in some way compared with the previous one – they will be better adapted to the environment in which the selection took place. This, as Darwin saw, is an inevitable process that simply must occur if the conditions are fulfilled. Dennett has called this the evolutionary algorithm. If you have variation, heredity, and selection, then you must get evolution. You get “Design out of Chaos without the aid of Mind” (Dennett, 1995, p 50).

There are two important features to notice about this process. First, it simply musthappen if the three functions are in place. There is no magic here, nor any incomprehensible theories. Once you understand the effects of variation with selective copying the result is obvious – it is beautifully simple. Second, the process requires no designer and no plan. It is not heading inexorably towards anything in particular because all the changes are the product of chance and necessity. This is because selection is not carried out by anyone with a plan or scheme or project in mind but is shaped by wind and weather, lack of food or oxygen, and the appetites of predators. Biology needs no God. Evolution has no project. This is what Dennett calls “Darwin’s dangerous idea”. It is often said that nothing in biology makes sense except in the light of this theory of evolution.

We are all familiar with how this process works for genes, but to explore the notion of universal Darwinism, Dawkins wondered whether there any other replicators on this planet. Since then many other examples have been discovered, including the immune system and neural processing, but Dawkins argued that staring us in the face, though still drifting clumsily about in its primeval soup of culture, is another replicator – a unit of imitation. He gave it the name “meme” (to rhyme with “cream” or “seem”), from the Greek for ‘that which is imitated’. As examples he suggested “tunes, ideas, catch-phrases, clothes fashions, ways of making pots or of building arches.” (Dawkins, 1976, p 192).

By 1998 the term had entered the English language and first appeared in the Oxford English Dictionary, defined as follows;  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”. This means that whatever is copied from person to person is a meme. Everything you have learned by copying it from someone else is a meme; every word in your language, every catch-phrase or saying. Every story you have ever heard, and every song you know, is a meme. The fact that you drive on the left (or perhaps the right), that you drink lager, think sun-dried tomatoes are passé, and wear jeans and a T-shirt to work are memes. The style of your house and your bicycle, the design of the roads in your city and the colour of the buses – all these are memes.

There is nothing ethereal about these memes. They are the very behaviours and artefacts that fill our lives. They are whatever is copied.

We can see that much of culture consists of memes. However, it is easy to get carried away and think of all experiences as memes and this is not helpful. We need instead to stick to a clear definition. The whole point of the meme is that it is information copied from one person to another. Therefore a great deal of what goes on in the human mind is nothing to do with memes. First, perception and visual memory need not involve memes. You can look at a beautiful scene, or taste a delicious meal, and remember them in detail without any memes being involved (unless you put words to your experience).

Second, not all learning involves memes. What you learn by yourself through classical conditioning (association) or by operant conditioning (trial and error) need not be memetic. Many other creatures are capable of these processes, and of extensive learning, but they do not have memes because they cannot pass on what they learn to anyone else. There may be a limited capacity for imitation in song birds, dolphins, and possibly some primates. Chimpanzees and orangutans may be capable of very limited forms of imitation, but only humans are capable of the kind of widespread and general imitation that makes a second replicator possible, and so leads to memetic evolution.

We should remember that this new kind of evolution proceeds not in the interest of the genes, nor in the interest of the individual who carries the memes, but in the interest of the memes themselves. This is why both memes and genes are described as “selfish”. Replicators are not selfish in the sense of having desires and plans as we do – they couldn’t have – they are only bits of information, either coded on DNA or copied by imitation. They are selfish in the sense that they will get copied if they can. In the case of memes, they will use us to get themselves copied without caring about the effect on us, or on our genes, or on our planet.

We can now begin to take on the “the meme’s eye view” and from this perspective the important question is why some memes survive and get copied into many brains or artefacts, while others do not. The general principle might be stated like this: Some memes succeed in getting copied because they are good, useful, true, or beautiful, while others succeed even though they are false or useless. From the meme’s point of view all this is irrelevant. If a meme can survive and get replicated it will. Generally we humans do try to select true ideas over false ones, and good over bad; after all our biology has set us up to do just that, but we do it imperfectly, and we leave all kinds of opportunities for other memes to get copied – using us as their copying machinery.

We may consider some examples of selfish memes that survive well in spite of being useless, false, or even harmful. At the simplest end of the continuum are self-replicating viral sentences, or simple groups of memes. A group of memes that works together is called a ‘co-adapted meme-complex’ or “memeplex”. An example is the common sort of email virus that urges you to pass on an urgent warning to all your friends. These messages often warn of a non-existent threat, such as a virus that will destroy everything on your hard disk. If you believe them, and pass on the message, this little memeplex can go on to be copied many more times. In fact the message itself is the virus. Not only have such viruses clogged up whole systems, but when people realise their mistake they often send out new messages telling people not to believe it, and so clog up the system again. Some of these viruses have lasted for five years or more.

The basic structure of such viruses is an instruction to “copy me” backed up by threats and promises. This same structure can be seen in other, more important, memeplexes too. For example, Dawkins uses Catholicism as an example of a group of memes that have succeeded for centuries in spite of being false. At Holy mass, the wine is supposed to turn literally into the blood of Christ. Clearly this is nonsense, in the sense that the wine still smells and tastes as it did before and would not show up as Christ’s blood in a DNA test. Yet millions of people routinely believe the claim, as well as believing in heaven and hell, an invisible and all-powerful God, the virgin birth and the Holy Trinity.

Why? Part of the answer is that these memeplexes have the same structure as simple viral memes. But religions use other memetic tricks too. The idea of God appeals because of our desire to understand our origins and purpose here on earth, and to have a greater being who protects us. Of course if God could be seen you could discover that he did not exist, so invisibility is a good ploy. God can see all your sins and will punish you, but you have to wait for proof of that until you are dead. And in case you do show an inclination for checking up on things, you may be reminded that faith is good and questioning is bad (the opposite of how it is in science). In addition, the memes include exhortations to marry another Catholic and bring up lots of children in the faith, or to convert others. Giving your money away to the poor will raise your stakes in heaven, as will contributing to the building and maintenance of great churches, cathedrals, and monuments which will inspire further meme hosts. In all these ways money and effort is diverted into the spreading of memes. The memes make us work for their propagation.

Memes such as religions, cults, fads and ineffective therapies, have been described as viruses of the mind because they infect people and demand their resources in spite of being false. Some authors have emphasised these pernicious kinds of meme and even implied that all memes are viral. However, memes can vary across a wide spectrum. As a general principle we can say that some memes succeed because they are good, true, useful or beautiful, while others succeed even though they are none of these things. And some just pretend to be good or useful. Towards one end are the viruses, religions, cults and false beliefs. Towards the other are our most valuable tools for living (such as our languages, technology and scientific theories). Without memes we could not speak, write, enjoy stories and songs, or do most of the things we associate with being human. Memes are the tools with which we think and our minds are a mass of memes.

Note that successful memeplexes were not deliberately designed by anyone, but were created by the process of memetic selection. Presumably there have always been countless competing memes – whether religions, political theories, ways of curing cancer, clothes fashions, or musical styles – the point about memetic evolution is that the ones we see around us now are those that survived in the competition to be copied. They had what it takes to be a good replicator.

The theory of memetics provides a completely new way of looking at the world in general, and at human evolution in particular. For example, it provides new explanations for both the evolution of the enormous human brain, and for the evolution of language – both of which are difficult to explain on ordinary evolutionary hypotheses.

The size of the human brain is one mystery. Expensive to build and maintain, and dangerous to give birth to, it is about as big as the genes can safely make it – about three times bigger, relative to body weight, than the brains of other great apes. But why? Traditional theories look to genetic advantage, in improved hunting or foraging skills, or the ability to sustain larger groups with complex social skills. Memetics provides a completely different explanation.

The turning point was when the first hominids began to imitate, perhaps two and half million years ago, before the advent of stone tools and expanding brains. True imitation means copying a novel behaviour or skill from another animal. It is difficult to do and needs a lot of brain power, and is correspondingly rare in the animal kingdom. But once it arose, we may imagine our early ancestors imitating useful new skills in hunting, carrying and preparing food, or making fire or clothes.

As these early memes spread, it became increasingly important to be able to acquire them. So people who were better at imitation thrived, and the genes that gave them that ability, and the bigger brains it required, spread in the gene pool. Everyone got better at imitation, increasing the pressure to increase brain size still further.

Once everyone began imitating, the memes were let loose and could begin competing with each other to get copied. Alongside useful skills such as building fires, go less useful ones like fancy body decoration, and downright costly ones like energetic but futile rain dances. From the genes’ point of view people ought to be choosy about what they imitate, with genes for indiscriminate imitation being eliminated. But how can genes ensure that their carriers copy only useful memes when the memes keep changing? One useful strategy might be to copy the best imitators because they are most likely to have accurate versions of currently useful memes. This gives added status to the best imitators, improves their survival chances, and so helps to spread the genes that made them good imitators – genes for imitating rain dances as well as useful skills. If this memetic evolution goes too far the genes will respond with improvements in selective imitation, but their response will always lag behind the memetic competition. In this way the memes take hold of the leash. This is the process I have called memetic drive. Memes compete among themselves and evolve in one direction, genes then respond by improving selective imitation, and this means increasing brain power and size.

In a final twist, it would pay for people to mate with the best imitators too, because by and large they have the best survival skills. This means that sexual selection, guided by memes, could have played a role in creating our big brains. By choosing the best imitator for a mate, women help propagate the genes needed to copy religious rituals, colourful clothes, singing, dancing or painting, depending on the direction memetic evolution has taken. By this process, the legacy of past memetic evolution becomes embedded in the structures of our brains and we become musical, artistic and religious creatures. Our big brains are selective imitation devices built by and for the memes, as much as for the genes.

The origins of language can be explained by the same mechanism. Questions about the origins and function of language have been so contentious that as long ago as 1866 the Société de Linguistique de Paris banned any more speculation on the issue, and even now there is no generally agreed explanation. The most popular theories naturally appeal to genetic advantage. In contrast, the theory of memetic drive is based on advantage to the memes.

To understand how this works we must ask which kinds of memes would have survived best and spread in the emerging meme pool of our early ancestors. The general answer for any replicator is those with high fidelity, fecundity and longevity – in other words, ones that make many accurate and long-lived copies of themselves.

Sounds can be copied by more people at once than can gestures or other physical actions. Some sounds would be copied more accurately or more frequently than others depending on their value in communication or on the limitations of people’s ears, voices, and memories. The sounds themselves would compete for roles in signalling, and in this competition the best replicators would thrive. Streams of sound broken into separate words would be copied more accurately because digitisation makes for more effective replication. The use of different word orders in different circumstances would open up new niches for more memes. In this competition the highest quality replicable sounds would swamp out the poorer ones.

Now consider the effect on the genes. The best imitators acquire better survival skills, higher status and the best mates. Therefore genes for the ability to imitate the winning sounds increase in the gene pool. I suggest that by this process the successful sounds gradually drove the genes into creating a brain that was especially good at copying them. The result was the human capacity for language. It was designed by memetic competition and meme-gene coevolution.

The whole process of memetic driving is an example of replicators coevolving with their copying machinery. Just as DNA must have once evolved along with its cellular replication machinery, so memes have coevolved with the human brains that copy them. But human imitation can be inaccurate. Unlike imitation in other species, human imitation is clearly good enough to sustain memetic evolution, but there is plenty of room for improvement. So we might expect better copying machinery to have appeared – and it has. From pen and paper to the printing press, from telephones to the fax machine, and from computers to the Internet, copying machinery has been improving, and more memes are spreading further and faster.

We can take a simple example; the invention of the fax machine. When fax machines became available people realised that they could get information faster and so they bought a machine. This encouraged them to send more faxes and encouraged their friends and colleagues to buy machines too. The memes sent, and the machines that copied them, increased together – and because faxes get there quicker than letters the whole process of memetic exchange speeded up. The same process happened with the Internet. Once email was possible, more and more people wanted it, and they sent more and more messages. The infosphere expanded rapidly.

Looked at from our point of view, we might see the Internet as wonderful technology created by us for our own pleasure and to make our lives easier. Looked at from the meme’s eye view, we humans are the early meme machines that helped create better meme machines for the benefit of the memes themsleves. When you see an office full of people enslaved by the flod of memes they have to deal with – typing away all day – rushing to deal with yet more information – you might reasonably wonder who this is all for. According to memetics, it is all a vast evolutionary process happening for the sake of replicating the memes. Today’s information explosion is just what we should expect.

Finally, memetics has implications for the nature of our very selves. According to Dennett a person is “a particular sort of ape infested with memes”. We all pick up countless memes throughout our lives and these (along with our genes and the environment in which we live) make us the unique individuals we are. But isn’t there a real self inside who lives this life? Isn’t there a real ‘me’ who makes my decisions, and holds my beliefs. Isn’t there a real self who has consciousness and free will? I would say no. The self is just a word around which memes can gather. All sorts of memes benefit by us having the false idea of a self inside. So the self is just a complicated memeplex, created by and for the memes themselves for their own protection and replication.

How then do we live our lives if we are just memeplexes? Some philosophers have argued that the only result would be a helpless fatalism or deep depression. In fact it is possible to drop the idea of an inner self and simply live life as a memeplex. Oddly enough this does not seem to make people worse, or more miserable, but to be a kind of liberation. Dawkins ended The Selfish Gene with the words  “We, alone on earth, can rebel against the tyranny of the selfish replicators”. I would argue, instead, that we are meme machines, created by and for the selfish replicators. Our only true freedom comes not when we rebel against the tyranny of the selfish replicators but when we realise that there is no one to rebel.


Aunger, R.A. (Ed) (2000) Darwinizing Culture: The Status of Memetics as a Science, Oxford University Press

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

Dawkins,R. (1976) The Selfish Gene Oxford, Oxford University Press (new edition with additional material, 1989)

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

Note that this article is also available in French, Portuguese, Italian and Russian.

See also my Note on Ontopsychology