The trouble with the trouble with memetics

Skeptical Inquirer, Volume 32, No.2, March/April 2008, pp 65

(Note: This version is very slightly different from the published, edited, version)

Massimo Pigliucci’s objections to memetics (“The Trouble with Memetics”, Skeptical Inquirer, September/October 2007) mostly misconstrue the basic idea of memes. So I’d like to explain where I think he has gone wrong and why memetics really could have a bright future.

Several of his arguments take the following form:

Memes are not like genes in respect of X

Therefore the analogy between memes and genes is false

Therefore memetics is false.

But analogies between memes and genes need not be close. To see why we must go back to the origin of the term ‘meme’. As Pigliucci himself explains, Dawkins invented the term to refer to a cultural replicator; that is, to information that is copied from person to person, or person to artefact. This information (whether ideas, skills, habits or stories) varies, and the variant forms are subject to selection, so this counts as a replicator. Genes and memes are both replicators and therefore should have much in common, but they are very different kinds of replicator, so we should expect many differences too. This means that analogies may help us in deriving hypotheses about the way the new replicator works, but could also lead us astray if we expect them to be too close.

Pigliucci’s first “X” is that “there doesn’t seem to be any distinction between memes themselves and the phenotypes they produce.” Agreed – in many cases but not all. More interestingly I think that memes are actually evolving this distinction right before our eyes. Think about it this way. Genes have been evolving for about 4 billion years. Starting from very simple self-replicating molecules they have ended up packaged inside elaborate vehicles, and with fabulously high-fidelity, effective copying machinery, involving accurate transcription, random variation, and a separation between the germ line (which is copied) and its phenotypic expression (which is not). Memes have been around for, at most, 2 million years but now they are catching up very fast indeed, and are rediscovering such “good tricks” as separating out the replicator itself from the products it makes possible; a trick that, among other advantages, avoids errors accumulating and allows for easier redesign of phenotypes.

This direction of change can be seen all around us. Suppose someone sings a song and someone else copies it – there is no split, and errors accumulate. Compare this with the song being written down in musical notation and lots of perfect copies being printed. The number of copies made depends on the popularity of the song, just as in biology the number of genes passed on depends on the success of a phenotype. Or think about the difference between someone watching someone else build a grass hut or a wattle fence and then trying to do it themselves, and the building of a block flats in which instructions for making bricks, windows, pipes, roof tiles, etc. are accurately copied in computers or on paper, and then the block of flats is built but not itself copied. You can think through similar examples such as cars produced in factories, books printed in presses, clothes fashions spread by competition between factory produced items. Then there are computers. Think about Microsoft Word. This software is copied with perfect fidelity in billions of computers around the world, but its success depends not on anyone seeing the code, but on the success of the documents it produces – a germ line/phenotype split if ever there was one.

Memes are already overtaking genes in their evolutionary innovations. Whereas until now human brains (the original, slow and imperfect meme machines) have done most of the copying, varying, and selecting of memes, this is changing. Already computers, especially on the Internet, do much of the copying, and they are beginning to take on the tasks of producing variants and even of doing the selecting (think viruses, crawlers, automated essays, but especially search engines). I believe that we can only understand what is going on here by taking a meme’s eye view and being realistic about our own, diminishing, role in the process.

Among Pigliucci’s other Xs is that memes have no obvious physical basis, but this is simply wrong. They are information that is copied – whether as variations in the airwaves of human speech, or as something hard to pin down like a dance or a vague idea, or (increasingly) as digital information stored in physical systems.

I won’t pretend that memetics is easy, but these objections will not do. I think we’ll find that memetics’ greatest strength lies in its vision of culture as a vast parasitic system evolving increasingly fast, and using we human meme machines as a resource for its own inevitable expansion. The way it sucks up the planet’s resources without care of the consequences is now our greatest challenge.