Scientific ideas also evolve and there have been many theories that attempt to explain how. The influential philosopher Karl Popper, in one of his best known contributions to the philosophy of science, suggested that scientific knowledge is gained by the falsification of hypotheses, not by accumulating proof or evidence
for
theories. Science can then be seen as a competitive struggle between rival hypotheses in which only some survive.
Popper also applied Darwinian thinking in his three ‘cosmic evolutionary stages’: World 1 is the world of physical objects such as trees, tables and human bodies; World 2 is the world of subjective experiences including feelings, emotions and consciousness; and World 3 is the world of ideas; of language and stories, works of art and technology, mathematics and science. World 3 is largely autonomous, even though created by us (Popper 1972), and its contents have effects on the other worlds by a kind of downward causation. So, for example, scientific theories may appear as World 1 objects (the scientist, the journal papers, the
experimental apparatus, and so on), but they are more than just physical objects. The ideas
themselves
influence those objects. The problems, hypotheses, theories and intellectual struggles work through World 2 and into World 1. Scientific ideas really do change the world: ‘once theories exist, they begin to have a life of their own’ (Popper and Eccles 1977, p. 40).
How can an idea change the physical world? Popper was struggling here with a difficult and important problem, related to the value of reductionism in science and the viability of materialism as a world view. I do not think he solved it. His three worlds contain very different kinds of material and he has to propose a tricky kind of interactionism to link them. Interestingly, he touches on the role of imitation but without realising how it might help. For example, in explaining how artistic ideas can have real effects, he says ‘a sculptor may, by producing a new work, encourage other sculptors to copy it, or to produce similar sculptures’ (Popper and Eccles 1977, p. 39). In his terms, the ideas in the sculptor’s mind (World 3) affect the experiences of others (World 2) and thus lead to new sculptures (World 1).
In memetic terms, all that happens – whether in science or art – is selective imitation. The emotions, the intellectual struggles, the subjective experiences – these are all parts of the complex system that leads to some behaviours being imitated and others not. And it is because imitation lets loose a second replicator that ideas begin to ‘have a life of their own’. In this way, memetics provides a mechanism for the evolution of scientific ideas that Popper’s three worlds cannot.
Although Popper did not use the idea of a replicator, his views directly gave rise to the new field of evolutionary epistemology, which does. Evolutionary epistemology began in 1974 with a critique of Popper by Campbell, and applies Darwinian thinking to the evolution of knowledge (Hull 1988
a
,
b
; Plotkin 1982). The American philosopher David Hull studies the way scientific ideas develop over time in lineages rather as species do. He treats scientific ideas as the replicators and scientists as the interactors (he prefers the term ‘interactor’ to Dawkins’s ‘vehicle’ because of its more active connotations). Plotkin considers science as not only ‘the product of a “Darwin machine” ’ but ‘a special form of culture that is transformed in time by evolutionary processes’ (Plotkin 1993, pp. 69, 223). According to evolutionary epistemology, biological adaptations are one form of knowledge, and science is another; both are produced by the processes of blind variation and selective retention (Campbell 1975). This approach is firmly based in Universal Darwinism and does not bring everything back to genetic advantage.
Whose advantage?
We can now see that many theories of cultural change use evolutionary ideas but they are not the same as memetics. There are two fundamental differences. First, most do not distinguish general evolutionary theory from the specifics of biological evolution. This means they are unclear about the relationship between biology and culture and easily fall foul of the obvious differences between genetics and cultural evolution. Second, they do not introduce the idea of a second replicator such as the meme. This means they do not see cultural evolution as proceeding in the interests of a selfish replicator.
This last issue is most important and I want to pursue it. The whole point of memetics is to treat the meme as a replicator in its own right, operating entirely for the benefit of its own selfish replication. If there is no second replicator, and you are a committed Darwinian, then somehow or other everything must come back to the genes – to biological advantage. If there are two replicators (or more) then there will inevitably be conflicts of interest – circumstances in which the interests of the genes pull in one direction and those of the memes in the opposite direction. These examples are very important for memetics because they would not be predicted by a purely genetic theory. If they occur, they prove that we need a theory of memes – or at least a theory involving some kind of second replicator. This is what distinguishes memetic theory from other theories of cultural evolution.
Dennett (1995) makes the same point when he asks ‘
Cui bono
?’, who benefits? He says ‘The first rule of memes, as it is for genes, is that replication is not necessarily for the good of anything; replicators flourish that are good at … replicating! … The important point is that there is no
necessary
connection between a meme’s replicative power, its “fitness” from
its
point of view, and its contribution to
our
fitness (by whatever standard we judge that)’ (Dennett 1991, p. 203, italics in the original).
Dawkins explains:
As soon as the primeval soup provided conditions in which molecules could make copies of themselves, the replicators themselves took over. For more than three thousand million years, DNA has been the only replicator worth talking about in the world. But it does not necessarily hold these monopoly rights for all time. Whenever conditions arise in which a new kind of replicator
can
make copies of itself, the new replicators
will
tend to take over, and start a new kind of evolution of their own. Once this new
evolution begins, it will in no necessary sense be subservient to the old (Dawkins 1976, pp. 193-4, italics in the original).
Of course, memes could only come into existence when the genes had provided brains that were capable of imitation – and the nature of those brains must have influenced which memes took hold and which did not. However, once memes had come into existence they would be expected to take on a life of their own.
Dawkins argued that biologists had so deeply assimilated the idea of genetic evolution that they tended to forget that it is only one of many possible kinds of evolution. He complained of his colleagues that ‘In the last analysis they wish always to go back to “biological advantage” ’ (Dawkins 1976, p. 193). In other words, they might accept the idea of memes, or some kind of unit of cultural evolution, but then still believe that memes must always act somehow for the benefit of the genes. But this is missing the whole point of the second replicator. If memes are replicators, as I am convinced they are, then they will not act for the benefit of the species, for the benefit of the individual, for the benefit of the genes, or indeed for the benefit of anything but themselves. That is what it means to be a replicator.
I am labouring this point because I am now going to review some theories of cultural evolution that
have
introduced the idea of a second replicator – or at least some kind of new cultural unit. (Durham 1991 provides a more thorough review.) At first sight these may all appear equivalent to the idea of the meme, but they are not. There are many similarities and differences but the most important point to look for is whether the new unit is really being treated as a replicator in its own right. If it is not then the theory is not equivalent to memetics.
In 1975, just before Dawkins proposed the idea of memes, the American anthropologist F. T. Cloak wrote about cultural instructions. He pointed out that whenever we see any behaviour being performed we assume that there is some internal structure in the animal’s nervous system that causes that behaviour. All animals have such instructions but humans, unlike other animals, can acquire new instructions by observing and imitating others. Cloak suggested that culture is acquired in tiny, unrelated snippets that he called ‘corpuscles of culture’ or ‘cultural instructions’.
Furthermore, he distinguished very carefully between the instructions in people’s heads and the behaviour, technology or social organisation that those instructions produce. The former he called the ‘i–culture’ and the latter the ‘m–culture’.
He was absolutely clear about the status of cultural instructions, even though he did not use the replicator concept. He said that the ultimate function of both i–culture and m–culture is the maintenance and propagation of the i–culture. Therefore, he concluded, we should not be surprised to find some m–culture features that perform functions that are irrelevant, or even destructive, to the organisms who make or do them. He compared cultural instructions to parasites that control some of their host’s behaviour – a bit like a flu virus that makes you sneeze to get itself propagated. He concluded ‘In short, “our” cultural instructions don’t work for us organisms; we work for them. At best, we are in symbiosis with them, as we are with our genes. At worst, we are their slaves’ (Cloak 1975, p. 172). Quite clearly, Cloak had seen the implications of having a second selfish replicator – even though others subsequently argued that cultural instructions are not replicators at all (Alexander 1979).
In
The Selfish Gene,
Dawkins mentions Cloak, saying that he wants to go further in directions being explored by Cloak and others. However, Dawkins lumps together both the behaviours and the instructions that produce them, and calls them all memes, while Cloak separates the two -a distinction that is somewhat analogous to the distinction between the genotype and the phenotype in biology. Later, Dawkins (1982) makes the same distinction as Cloak and defines a meme as ‘a unit of information residing in a brain’. I shall return to consider the importance of this difference later on. For now we need only note that Cloak’s cultural instruction is, like the meme, a true second replicator.
Sociobiology and culture on a leash
While Dawkins was writing
The Selfish Gene,
the new science of sociobiology was being established – studying the genetic and evolutionary basis of behaviour. There was, at the time, a great outcry against applying sociobiology to human behaviour. Some of this came from sociologists, anthropologists and others who argued that human behaviour was almost entirely free from the constraints of the genes and could not be understood by what they saw as (horror of horrors) ‘genetic determinism’. The genes, they claimed, only give us a ‘capacity for culture’. Some came from ordinary people who rejected the idea that their cherished beliefs, decisions and actions were constrained by their genetic make–up -what about ‘free will’?
This reaction reminds me of the antagonism to Newton, to Copernicus
and to Darwin himself. Sociobiology seemed to push human beings further off their self–created pedestal – to undermine their sense of free will and autonomy. As we shall see, memetics takes a further big step in this direction and so will probably reap the same antagonism. Still – as Cloak put it’… if we
are
the slaves of some of “our” cultural traits, isn’t it time we knew it?’ (Cloak 1975, p. 178).
Much of the antagonism to sociobiology has died down, perhaps because of the increasing evidence for the evolutionary basis of human behaviour, and perhaps because of a better understanding of the way genes and environment interact. The old image of genes as providing a blueprint or wiring diagram for building a body is clearly wrong. A better analogy is with a recipe, though it is still not a close one. Genes are instructions for building proteins, and the results of this protein synthesis are influenced at every stage by the available raw materials and the nature of the environment. Nothing is purely genetically
determined
and nothing purely environmentally
determined.
We human beings, like all other creatures, are a complex product of both – and this is true of the way we behave as well as the shape of our legs.
In spite of the antagonism, sociobiology made great progress but, as its founding father Edward O. Wilson complained, it had little to say about the individual human mind or the diversity of cultures. In 1981, Wilson teamed up with the physicist Charles Lumsden to develop a theory of gene–culture coevolution and introduced the concept of the ‘culturgen’ as ‘the basic unit of inheritance in cultural evolution’ (Lumsden and Wilson 1981, p. x). They hoped their new theory would lead right through from genes to mind to culture and developed mathematical treatments of how different culturgens would affect genetic fitness. However, they always came back to the genes as the final arbiters. If maladaptive culturgens are sometimes selected this is because their harm is not immediately apparent and so there is some lag before the system adapts. Ultimately, the genes will win out. As they put it – ‘the genes hold culture on a leash’.
The ‘leash principle’ is a more memorable way of expressing what Dawkins meant about his colleagues wanting ‘always to go back to “biological advantage” ’. It also provides us with a helpful image. If Lumsden and Wilson are right then the genes are always the owner and the culturgens are the dog. The leash can sometimes get longer – even extremely long – but it is still a dog at the other end. According to memetics, the genes may turn into a dog and the memes become the owner – or perhaps we should enjoy the spectacle of two dogs, one on either end – each running like mad to serve their own selfish replication.
The Stanford geneticists, Luigi Cavalli–Sforza and Marcus Feldman (1981), developed a detailed model of cultural transmission based on the ‘cultural trait’ as the unit. Cultural traits are learned by imprinting, conditioning, observation, imitation or direct teaching (note that this is a broader range than for memes which, by definition, have to be passed on by imitation and cannot be acquired by imprinting or conditioning). They clearly distinguish cultural selection from Darwinian or natural selection and they use the concept of ‘cultural fitness’ – that is, the fitness for survival of a cultural trait itself- a concept that is useful in memetics. They also introduced the distinction between vertical transmission – such as from parent to child – and horizontal transmission – such as from child to child or adult to unrelated adult. We shall see later how important this is for understanding life in an age of predominantly horizontal transmission.