This wonderful book (Jablonka & Lamb Reference Jablonka and Lamb2005) has opened up my thinking on evolution and thrown unexamined assumptions into disarray, for which painful intellectual turmoil I am grateful. The sections on memes are, however, confused and inadequate, and it is these I address here.
Unlike the first three dimensions, the fourth is largely restricted to one species – humans – and so its existence raises questions about the origin of human uniqueness. Jablonka & Lamb (J&L) argue that humans are unique because of the complexity and power of culture, and attribute this uniqueness, as have so many before them, to the acquisition of symbols. For the authors, the turning point in human evolution was something like Deacon's (Reference Deacon1997) “symbolic threshold” – once this threshold has been crossed, the argument goes, a species can have symbolic culture that can evolve. This is why they call their fourth dimension the “symbolic inheritance system.”
Memetics, by contrast, does not depend on the notion of symbolism. For memetics, the turning point in hominid evolution was the appearance of imitation. Imitation is a kind of copying, and the information that is copied (memes) varies and is selected, which necessarily creates a new evolutionary process. Once this process got underway, the evolving culture could interact with genetic and epigenetic systems to transform the species in which it arose, resulting in modern humans along with their complex and powerful culture (Blackmore Reference Blackmore2001). Memes are defined as “that which is imitated” (or more generally, whatever is copied; Dawkins Reference Dawkins1976). So although many memes, including words, sentences, or diagrams, are symbolic, others, such as tunes, cars, or hairstyles, are not. Symbolic thought is a result of cultural evolution – not its starting point.
I am suggesting that one advantage of memetics is its ability to explain human uniqueness and culture (as arising from imitation) without getting bogged down in the slippery notion of symbolism, and J&L do get bogged down. Let us consider some of their examples of symbolic culture. They frequently refer to songs and dances, but although words are symbolic, one can copy a dance or tune without any symbols being involved. They also claim that “ideas, artifacts, and institutions are almost entirely based on symbols” (p. 205). But is this so? A lifelike statue is symbolic, but what about an abstract sculpture or painting? An institution such as a bank or hospital has symbolic facets such as paper money or patient records, but buildings are not symbolic, nor are vaults or hospital beds. For J&L, this must surely cause confusion; for memetics, there is no problem. The important questions concern which memes are copied and why – for example, why most hospital beds use a particular design, or why some institutional structures thrive and are copied, even though they are not optimal (Runciman Reference Runciman1998).
Perhaps memetics suffers from far more serious drawbacks. J&L certainly think so, referring to “the dreaded memes” (p. 224). Memetics is, they say, “seductively simple” but flawed: “The flaw stems from the distinction that is made between replicators (memes) and their vehicles (human brains, human artifacts, and humans themselves are all given this role)” (p. 208). The authors refer to Dawkins's treatment of memes in The Extended Phenotype (Dawkins Reference Dawkins1982, sometimes called Dawkins B) but in his initial formulation in The Selfish Gene (Dawkins Reference Dawkins1976, or Dawkins A), Dawkins does not claim such a strong distinction, and nor do others who followed him (e.g., Blackmore Reference Blackmore1999; Dennett Reference Dennett1995). The confusion is compounded by J&L using the example of tits opening milk bottles, which is generally agreed to have spread by stimulus enhancement rather than imitation (Sherry & Galef Reference Sherry and Galef1984), and by their incorrectly describing a “memeplex” as “the set of memes in a brain,” rather than as a group of memes that are copied together.
My own view is that the division into replicators and vehicles was a useful development in biological evolution because copying the instructions for making a product is less prone to accumulated error than copying the product directly. So the very efficient system by which genes (instructions) are copied accurately down the germ line while phenotypes (their product) are not directly copied, itself evolved.
Turning to the evolution of memes, we can see the same development happening. Early memes such as dances, songs, spoken language, or ways of making tools were directly (and inefficiently) copied, but now we have production lines, computer programs, printing presses, and Web servers that make multiple copies of products and accurately copy the instructions for making more. The fact that this development is happening all around us means that memetics can provide insight into how evolutionary processes themselves evolve.
An important question J&L do discuss is whether memes are copied or reconstructed, and the extent to which development, learning, and meaning are involved in the process. I agree that many mathematical models assume a simple copying process and therefore do not accurately reflect the complexity of memetic transmission, but this complexity is no reason to reject memetics. If blind copying is better than the messy processes of development and meaning-sensitive reconstruction, then we should expect to see memetic systems evolving in this direction – and that, I would argue, is precisely what we do see. Compare the spread of gossip through word of mouth with its spread on the Internet. In the former, copying and selection occur together through the low-fidelity processes of listening, learning, recall, and reconstruction. In the latter, the copying is done by high- (almost 100%) fidelity digital processes, and only selection remains meaning dependent. We might expect to see, and may already be seeing, digital selection processes that bypass low-fidelity human processes. Or take J&L's own example of the spread of clothes fashions. Long ago, someone who admired another's dress would have had to find the cloth, draw a pattern, and cut and sew a copy. Now they simply go to the high street and choose from among all the myriad copies accurately produced in factories. All this is much better explained by memetics than by a “symbolic inheritance system,” for what does a new dress symbolize?
I find it odd that J&L, who have so effectively revealed the complex, fluid, and messy nature of other evolutionary processes, do not welcome memetics as another way of showing that not all evolution is clean, tidy, and gene-like. This is why Dawkins invented the concept in the first place, and J&L would seem ideally placed to have taken it on.
