It is not always as dismal as you think and, as Susan Blackmore insists, high fidelity is not necessarily synonymous with digital. Suppose we set up our Chinese Whispers Chinese Junk game again, but this time with a crucial difference. Instead of asking the first child to copy a drawing of a junk, we teach her, by demonstration, to make an origami model of a junk. When she has mastered the skill and made her own junk, the first child is asked to turn round to the second child and teach him how to make one. So the skill passes down the line to the twentieth child. What will be the result of this experiment? What will the twentieth child produce, and what shall we observe if we lay the twenty efforts out in order along the ground? I have not done it, but I will make the following
confident prediction, assuming that we run the experiment many times on different groups of twenty children. In several of the experiments, a child somewhere along the line will forget some crucial step in the skill taught him by the previous child, and the line of phenotypes will suffer an abrupt macromutation which will presumably then be copied to the end of the line, or until another discrete mistake is made. The end result of such mutated lines will not bear any resemblance to a Chinese junk at all. But in a good number of experiments the skill will correctly pass all along the line, and the twentieth junk will be no worse and no better, on average, than the first junk. If we then lay the twenty junks out in order, some will be more perfect than others, but imperfections will not be copied on down the line. If the fifth child is ham-fisted and makes a clumsily asymmetrical or floppy junk, his quantitative errors will be corrected if the sixth child happens to be more dextrous. The twenty junks will not exhibit a progressive deterioration in the way that the twenty drawings of our first experiment undoubtedly would.
Why? What is the crucial difference between the two kinds of experiment? It is this: inheritance in the drawing experiment is Lamarckian (Blackmore calls it ‘copying-the-product’). In the origami experiment it is Weismannian (Blackmore’s ‘copying-the-instructions’). In the drawing experiment, the phenotype in every generation is also the genotype -it is what is passed on to the next generation. In the origami experiment, what passes to the next generation is not the paper phenotype but a set of instructions for making it. Imperfections in the execution of the instructions result in imperfect junks (phenotypes) but they are not passed on to future generations: they are non-memetic. Here are the first five instructions in the Weismannian meme line of instructions for making a Chinese junk:
1. Take a square sheet of paper and fold all four corners exactly into the middle.
2. Take the reduced square so formed, and fold one side into the middle.
3. Fold the opposite side into the middle, symmetrically.
4. In the same way, take the rectangle so formed, and fold its two ends into the middle.
5. Take the small square so formed, and fold it backwards, exactly along the straight line where your last two folds met.
…and so on, through 20 or 30 instructions of this kind. These instructions, though I would not wish to call them digital, are potentially
of very high fidelity, just as if they were digital. This is because they all make reference to idealised tasks like ‘fold the four corners exactly into the middle’. If the paper is not exactly square, or if a child folds ineptly so that, say, the first corner overshoots the middle and the fourth corner undershoots it, the junk that results will be inelegant. But the next child in the line will not copy the error, for she will assume that her instructor
intended
to fold all four corners into the exact centre of a perfect square. The instructions are self-normalising. The code is error-correcting. Plato would enjoy it: what passes down the line is an ideal essence of junk, of which each actual junk is an imperfect approximation.
The instructions are more effectively passed on if verbally reinforced, but they can be transmitted by demonstration alone. A Japanese child could teach an English one, though neither has a word of the other’s language. In the same way, a Japanese master carpenter could convey his skills to an equally monoglot English apprentice. The apprentice would not copy obvious mistakes. If the master hit his thumb with a hammer, the apprentice would correctly guess, even without understanding the Japanese expletive ‘ ** **** ** !’, that he meant to hit the nail. He would not make a Lamarckian copy of the precise details of every hammer blow, but copy instead the inferred Weismannian instruction: drive the nail in with as many blows of your hammer as it takes your arm to achieve the same idealised end result as the master has achieved with his – a nail head flush with the wood.
I believe that these considerations greatly reduce, and probably remove altogether, the objection that memes are copied with insufficient high fidelity to be compared with genes. For me, the quasi-genetic inheritance of language, and of religious and traditional customs, teaches the same lesson. Another objection, discussed, like the first, in Susan Blackmore’s illuminating chapter on ‘Three problems with memes’ is that we do not know what memes are made of or where they reside. Memes have not yet found their Watson and Crick; they even lack their Mendel. Where genes are to be found in precise locations on chromosomes, memes presumably exist in brains, and we have even less chance of seeing one than of seeing a gene (though, in an article referred to by Blackmore, the neurobiologist Juan Delius had pictured his conjecture of what a meme might look like). As with genes, we track memes through populations by their phenotypes. The ‘phenotype’ of the Chinese junk meme is made of paper. With the exception of ‘extended phenotypes’, such as beaver dams and caddis larva houses, the phenotypes of genes are normally parts of living bodies. Meme phenotypes seldom are.
But it can happen. To return to my school again, a Martian geneticist,
visiting the school during the morning cold bath ritual, would have unhesitatingly diagnosed an ‘obvious’ genetic polymorphism. About 50 per cent of the boys were circumcised and 50 per cent were not. The boys, incidentally, were highly conscious of the polymorphism and we classified ourselves into Roundheads versus Cavaliers (I have recently read of another school in which the boys even organised themselves into two football teams along the same lines). It is, of course, not a genetic but a memetic polymorphism. But the Martian’s mistake is completely understandable; the morphological discontinuity is of exactly the kind that one normally expects to find produced by genes.
In England at that time, infant circumcision was a medical whim, and the Roundhead/Cavalier polymorphism at my school probably owed less to longitudinal transmission than to differing fashions in the various hospitals where we happened to have been born – horizontal memetic transmission, yet again. But through most of history circumcision has been longitudinally transmitted as a badge of religion (of
parents’
religion I hasten to point out, for the unfortunate child is normally too young to
know
his own religious mind). Where circumcision is religiously or traditionally based (the barbaric custom of female circumcision always is), the transmission will follow a longitudinal pattern of heredity, very similar to the pattern for true genetic transmission, and often persisting for many generations. Our Martian geneticist would have to work quite hard to discover that no genes are involved in the genesis of the roundhead phenotype.
The Martian geneticist’s eyes would also pop out on stalks (assuming they were not on stalks to begin with) at the contemplation of certain styles of clothing and hairdressing, and their inheritance patterns. The black skull-capped phenotype shows a marked tendency towards longitudinal transmission from father to son (or it may be from maternal grandfather to grandson), and there is clear linkage to the rarer pigtail-plaited sideburn phenotype. Behavioural phenotypes such as genuflecting in front of crosses, and facing east to kneel five times per day, are inherited longitudinally too, and are in strongly negative linkage disequilibrium with each other and with the previously mentioned phenotypes, as is the red-dot-on-forehead phenotype, and the saffron robes/ shaven head linkage group.
Genes are accurately copied and transmitted from body to body, but some are transmitted at greater frequency than others–by definition they are more successful. This is natural selection, and it is the explanation for most of what is interesting and remarkable about life. But is there a similar meme-based natural selection? Perhaps we can use the Internet
again to investigate natural selection among memes? As it happens, around the time the word ‘meme’ was coined (actually a little later), a rival synonym, ‘culturgen’, was proposed. Today, culturgen is mentioned twenty times on the World Wide Web, compared with memetic’s 5042. Moreover, of those twenty, seventeen also mention the source of the word, falling foul of the
Oxford English Dictionary’s
criterion. Perhaps it is not too fanciful to imagine a Darwinian struggle between the two memes (or culturgens), and it is not totally silly to ask why one of them was so much more successful. Perhaps it is because meme is a monosyllable similar to gene, which therefore lends itself to quasi-genetic sub-coinings: meme pool (352), memotype (58), memeticist (163), memeoid (or memoid) (28), retromeme (14), population memetics (41), meme complex (494), memetic engineering (302) and metameme (71) are all listed in the ‘Memetic Lexicon’ at
http://www.lucifer.com/virus/mem-lex.html§MEME
(the numbers in parentheses count the mentions of each word on the Web on my sampling day). Culturgen-based equivalents would be more obvious but less snappy. Or the success of meme against culturgen may have been initially just a non-Darwinian matter of chance–memetic drift (85)–followed by a self-reinforcing positive feedback effect (’unto every one that hath shall be given, and he shall have abundance; but from him that hath not shall be taken away even that which he hath’, Matthew 25: 29).
I have mentioned two favourite objections to the meme idea: memes have insufficient copying fidelity, and nobody really knows what a meme physically is. A third is the vexed question of how large a unit deserves the name ‘meme’. Is the whole Roman Catholic Church one meme, or should we use the word for one constituent unit such as the idea of incense or transubstantiation? Or for something in between? Susan Blackmore gives due attention to such questions, but she rightly concentrates on a more constructive approach, developing the positive explanatory power of the ‘memeplex’ – an abbreviation which she prefers over the full ‘coadapted meme complex’, and I shall be surprised if in time her book does not bring about a Darwinian reversal of their numerical fortunes (today, 20 and 494, respectively).
Memes, like genes, are selected against the background of other memes in the meme pool. The result is that gangs of mutually compatible memes – coadapted meme complexes or memeplexes – are found cohabiting in individual brains. This is not because selection has chosen them as a group, but because each separate member of the group tends to be favoured when its environment happens to be dominated by the others. An exactly similar point can be made about genetic selection.
Every gene in a gene pool constitutes part of the environmental background against which the other genes are naturally selected, so it’s no wonder natural selection favours genes that ‘cooperate’ in building those highly integrated and unified machines called organisms. Biologists are sharply divided into those for whom this logic is as clear as daylight, and those (even some very distinguished ones) who just do not understand it – who naively trot out the obvious cooperativeness of genes and unitariness of organisms as though they somehow counted against the ‘selfish gene’ view of evolution. Susan Blackmore not only understands it, she explains the matter with unusual clarity and goes on to apply the lesson with equal clarity and force to memes. By analogy with coadapted gene complexes, memes, selected against the background of each other, ‘cooperate’ in mutually supportive memeplexes – supportive within the memeplex but hostile to rival memeplexes. Religions may be the most convincing examples of memeplexes but they are by no means the only ones. Susan Blackmore’s treatment is, as ever, provocative and revealing.
I believe a sufficient case has been made that the analogy between memes and genes is persuasive and that the obvious objections to it can be satisfactorily answered. But can the analogy do useful work? Can it lead us to powerful new theories that actually explain anything important? This is where Susan Blackmore really comes into her own. She warms us up with some fascinating vignettes which get us used to the memetic style of reasoning. Why do we talk so much? Why can’t we stop thinking? Why do silly tunes buzz round our heads and torment us into insomnia? In every case she begins her response in the same way: ‘Imagine a world full of brains, and far more memes than can possibly find homes. Which memes are more likely to find a safe home and get passed on again?’ The answer comes back readily enough, and our understanding of ourselves is enriched. She pushes on, with patience and skill applying the same method to deeper and more exacting problems: What is language for? What attracts us to our mates? Why are we so good to each other? Did memes drive the rapid, massive, and peculiar evolutionary expansion of the human brain? Along the way, she shows how the theory of memes can throw light on particular areas where she has special expertise from her academic career as a psychologist and sceptical investigator of the paranormal: superstition and near-death experience.
In the end, showing greater courage and intellectual
chutzpah
than I have ever aspired to, she deploys her memetic forces in a brave – do not think foolhardy until you have read it – assault on the deepest questions of all: What is a self? What am I? Where am I? (famous questions posed
by Daniel Dennett long before he became the philosophical mentor of all meme theorists). What of consciousness, creativity and foresight?
I am occasionally accused of having backtracked on memes; of having lost heart, pulled in my horns, had second thoughts. The truth is that my first thoughts were more modest than some memeticists, including perhaps Dr Blackmore, might have wished. For me, the original mission of the meme was negative. The word was introduced at the end of a book which otherwise must have seemed entirely devoted to extolling the selfish gene as the be-all and end-all of evolution, the fundamental unit of selection, the entity in the hierarchy of life which all adaptations could be said to benefit. There was a risk that my readers would misunderstand the message as being
necessarily
about genes in the sense of DNA molecules. On the contrary, DNA was incidental. The real unit of natural selection was any kind of
replicator,
any unit of which copies are made, with occasional errors, and with some influence or power over their own probability of replication. The genetic natural selection identified by neo-Darwinism as the driving force of evolution on this planet was only a special case of a more general process that I came to dub ‘Universal Darwinism’. Perhaps we would have to go to other planets in order to discover any other examples. But perhaps we did not have to go that far. Could it be that a new kind of Darwinian replicator was even now staring us in the face? This was where the meme came in.