But in the last chapter of his book, «Memes: the New Replicators», Dawkins took a step further. Dissatisfied with the usual Darwinian explanations of human behaviour in genetic terms, he postulated the existence of a unit of cultural transmission, analogous to the gene, which he termed meme. Like genes, memes would be replicators, and the mechanism by which they produced copies of themselves would be imitation:
Ever since, the study of memes -- memetics -- has strived to attain a scientific status. Scott Atran has described the steps to be taken in order to achieve this goal: the initial stage would be specifying whether and how the gene / meme analogy stands under verifiable scrutiny. Were the analogy to hold, it would then lead to «significant and surprising discoveries about specific causal structures». Were the analogy not to hold, the whole endeavour would have to be discarded as unscientific, although Atran acknowledges the likelihood that
The question is thus whether the notion of meme is to have a merely heuristic value, or whether on the contrary it may make possible the development of a scientific programme, be it in its original Dawkinsian form or in some modified version.
Literature on memetics tends to divide into two main branches, according to the analogy on which it is based. The most popular interpretation of memetics, sometimes known as the «meme-as-germ» interpretation, sees memes as similar to disease agents -- as reflected in the titles of well-known popularisations and / or vulgarisations of memetics: Virus of the Mind , Thought Contagion , etc. This point of view lays great emphasis on the memetic answer to the key question in evolutionary theory, which Daniel Dennett has formulated in Latin as «Cui bono?», «Who profits?» Whereas traditional social science has always assumed that cultural behaviour must ultimately benefit the individuals displaying it, according to memeticists it is not brains, individuals or societies that profit from evolutionary cultural dynamics, but memes themselves. The meme-as-germ approach thus makes great play of Dawkins's description of memes as leaping flea-wise from head to head, extending to cultural units his pithy description of genes as immortality-seeking parasites which use and then discard their hosts.
The meme-as-germ approach to the study of cultural transmission takes epidemiology as its model science. Memes would thus be the cultural equivalents of flu bacilli, transmitted through the communicational equivalents of sneezes. As a result of this epidemiological analogy B -- until recently, the predominant strain in memetics -- memeticists have tended to focus on how memes are transmitted from individual to individual. However, despite its popularity, the viral analogy tends towards distortion, endowing memes with a certain unwarranted malign character (which can be easily avoided by simply conceiving of memes in their more traditional guise as concepts or representations).
The second main approach to memetics, also known as the «meme-as-gene» interpretation, exploits in more depth Dawkins's original analogy between genetic and cultural transmission. This school of thought takes evolutionary genetics as its model science. It thus adopts a «meme's eye view» of cultural processes, similar to the «gene's eye view» which has characterized the development of evolutionary theory in the last years. It is this trend which is currently prevailing in memetics, and its aim is to precisely detail the ways in which cultural transmission is evolutionary. In order to do so, it has pushed the analogy with the gene to its logical extremes, seeking cultural equivalents for the main evolutionary genetic concepts, such as genotype, phenotype, transcription, code, etc. The main concept taken by evolutionary memetics from genetics is however that of replication: just as DNA strands replicate by producing identical copies of themselves B with an inevitable rate of mutation, which allows for evolution -- so memes would replicate themselves in order to be transmitted from bearer to bearer. Standard evolutionary memetics also argues that the means by which memes replicate themselves is the form of learning known as imitation. In the next section, we shall examine these two concepts more closely.
The idea of applying Darwinian evolutionary theory to the study of culture is of course nothing new. The aim of sociobiology and its latest offspring, evolutionary psychology, is to close the gap between the natural and the social sciences, bringing about a general -- and, according to their followers, long overdue -- «Darwinization» of the study of the human being. These schools of thought consider all human behaviour as the consequence of the interaction of evolved physiological and psychological variables with the natural environment. Thus, human culture would ultimately be biologically determined by the evolutionary history of the species. This stance leads sociobiologists and evolutionary psychologists to regard certain widespread behaviours, «backfiring» from the evolutionary point of view (such as the use of contraceptives), as the maladaptive result of the clash between the Stone Age instincts hardwired into our brains and our current technological environment.
The extreme reductionist claims of sociobiologists and evolutionary psychologists have been somewhat tempered by the cultural selectionism approach, which admits the existence of a dual system of inheritance in the human species: genetic transmission and cultural transmission. (This approach is thus sometimes also termed gene-cultural coevolution). In this respect, memetics can be considered a subcategory within cultural selectionism, differing from mainstream coevolutionism in its insistence on replication as the mechanism of cultural inheritance (cultural selectionism merely posits an unspecified inheritance mechanism).
The idea that cultural units replicate as genes do assumes that discrete, definable units can be distinguished within culture. Which is, of course, a rather strong claim to make. The memetic stance has been criticized, mainly by anthropologists, on the grounds that culture constitutes a continuum, and any units within it will necessarily be arbitrary constructs of the observer.
But even if cultural units can be properly distinguished, the memeticists' insistence on replication as the only mechanism for cultural inheritance has also been criticized on the grounds that replication is the exception rather than the rule in processes of cultural transmission -- the rule being almost always transformation. Such obvious examples as linguistic change or the often bewilderingly varying versions of rumours and urban legends indicate that mutation is indeed the default case in processes of cultural diffusion. As Dan Sperber has put it , when actual replication does takes place, it can be seen as the limit case of zero transformation.
A further argument is that the evolutionary character of cultural transmission is not compromised by the definition of its inheritance mechanism as other than replication. Indeed, as Joseph Henrich and Robert Boyd have shown  -- and as cultural selectionists have always emphasized -- evolutionary processes can take place without replication. Replication is only a particular instance of inheritance, peculiar to DNA.
Thus, it seems increasingly obvious that the only reason for the maintenance of replication as the mode of transmission is the adherence to the original gene / meme analogy (or, as Francisco Gil-White jokingly terms it, the «fetishism of the gene analogy»). Even Dawkins himself softened his initial views in The Extended Phenotype:
Sperber has suggested that the «directedness» of cultural transmission may be due to cognitive mechanisms characteristic of the human species. Thus memes would not be replicators, but units subject to transformation in the course of constructive processes of an inferential nature.
Another difference between cultural selectionism and memetics is that cultural selectionism considers that any means of non-genetic transmission can take part in cultural transmission, and thus admits a wide array of possibilities: imprinting, classical conditioning, operative conditioning, observation, imitation, direct instruction. By contrast, standard memetics is more restrictive in that it defines imitation as the only learning process which makes cultural diffusion possible. According to one of the leading memeticists, Susan Blackmore , the reason for this emphasis on imitation is that it is the only means of transmission which makes possible the accumulation of modifications over generations -- what is known as the «ratchet effect», whereby cumulative changes eventually become irreversible. This would make imitation the distinguishing feature of the human species. In her book The Meme Machine , Blackmore also put forth the hypothesis that the «memetic drive» which had originally led to the development of the capacity for imitation in mankind would also account for such odd phenomena peculiar to the human species as hypertrophied brains, speech, or the tendency to engage in non-altruistic acts, even in large groups of non-kin.
The concept of the «memetic drive» is, according to Blackmore, unique to the mimetic perspective and what sets it apart from alternative evolutionary approaches to culture. However, its restriction to imitation has been questioned by philosopher David Hull , who argues that the consequent restriction to the human species seriously reduces the span B and interest B of memetics. An exclusively human memetics, argues Hull, cannot explain such general evolutionary trends as the increase of intelligence within some animal families.
As Robert Aunger has pointed out , a potential memetic science faces three main questions:
(1) Whether culture can be seen as composed of independently transmitted information units;
(2) whether the process whereby these units are transmitted is necessarily one of replication; and
(3) whether a Darwinian or selectionist approach is the most adequate form for a science of culture to take.
The answer given by standard memetics to all these three questions is «yes». However, the objections mentioned in this essay are but a sample of the complications which such an unqualified answer raises. In the future, memeticists must meet these complications, revising and redefining their current conceptions of replication and imitation, in order to provide a consistent theory, if memetics is to attain a truly scientific status.
 Dawkins, Richard. The Selfish Gene (revised ed.), pg. 192. Oxford: OUP, 1989.
 Atran, Scott. «The Trouble with Memes: Inference versus Imitation in Cultural Creation», in Human Nature 12(4): 351-381, 2001.
 Brodie, Richard. Virus of the Mind: the New Science of the Meme. Seattle: Integral Press, 1996.
 Lynch, Aaron. Thought Contagion: How Belief Spreads through Society. New York: Basic Books, 1996.
 Sperber, Dan. Explaining Culture. A Naturalistic Approach. Oxford: Blackwell, 1996.
 Henrich, Joseph & Boyd, Robert. «On modeling cognition and culture. Why cultural evolution does not require replication of representations», in Journal and Cognition and Culture (forthcoming).
 Dawkins, Richard. The Extended Phenotype, pg. 112. Oxford: Freeman, 1982.
 Blackmore, Susan. «The memes' eye view», in Aunger, Robert, Darwinizing Culture, Oxford: OUP, 2000.
 Blackmore, Susan. The Meme Machine. Oxford: OUP, 1999.
 Hull, David. «Taking memetics seriously», in Aunger, Robert, Darwinizing Culture, Oxford: OUP, 2000.
 Aunger, Robert. «Introduction», in Aunger, Robert, Darwinizing Culture, Oxford: OUP, 2000.