suppose that every thought you have -- including this one -- is an autonomous parasite in your brain: a pattern of brain cells that copies itself from mind to mind.
Congratulations! You've just caught a meme.
A meme, according to Richard Dawkins, who coined the word in his bestseller "The Selfish Gene," is anything a human can remember and transmit. Memes are meant to be the brain's equivalent of DNA. A meme could be an idea or a snatch of music or a dance. So long as it gets copied between brains fairly accurately and competes with other copies for survival, it will do as a candidate for evolution. If memes exist, they have modified the world just as genes have: Genes have made the biosphere; memes have made the memeosphere, the place where human beings exist.
The idea is catchy -- the "meme" meme is particularly popular online -- but controversial. Dawkins himself has withdrawn from it a little. He said last year: "There are people who take memes seriously and there are people who don't. I sort of sit on the fence, and don't mind seriously one way or the other. That wasn't my purpose in producing them."
But lots of smart people do take them seriously. Daniel Dennett, the leading philosopher of artificial intelligence, writes as if the existence of memes were an established fact. A human being, he says, is an animal infested by memes.
And Susan Blackmore, a British psychologist who has been the scourge of every sort of paranormal nonsense for the last 20 years, has got memes bad. She is trying to work out ways to find them in the wild.
"I want to do away with the notion of 'ideas' and look at brain structures and behaviors," she says. "The brain structures may be hard to find. They probably consist of patterns of activation, which will not be there all the time. But I have talked at you for half an hour -- that's a behavior. You have written down some of my words. That's another behavior. And so the memes have been copied. It is a very loose form of copying -- but we can study the process."
Talking to Blackmore is exhausting and invigorating all at once. Ideas fountain out of her like champagne in an intoxicating stream; it hardly matters whether they are bubbles or not.
"If you take a common-sense view," she says, "humans are units called selves which somehow generate ideas from within. But when you take the meme's-eye view, you see brains as hosts for memes. We don't own or generate 'our' ideas. Nor are they working for our purposes. We can imagine meaning in them, but really all that's going on is imitation and storage," she says.
These ideas fit in with Blackmore's Buddhism, where the human personality is no more than a swarm of causes and consequences, temporarily bundled in a body. But they also fit with her sense of a crisis in academic psychology: "I've spent 25 years in psychology and it's an absolute mess. We don't have a decent theory of emotions, or of motivations."
Meme theory seems to offer a way out of this. If beliefs prosper for their own reasons, and not for the good of the organism that contains them, then this would explain the inexhaustible capacity of the human mind to produce bad ideas and disastrous plans. It does not matter to the success of a belief if the original brain that carries it dies, provided it has been spread to more brains in the meantime. In Daniel Dennett's slogan: "A scholar is just a library's way of making another library."
Blackmore says that a good test for the independent existence of memes would look for "behavior that is not in the interest of the person -- but purely in the interest of the propagation of the memes."
The classic example of this spread is religious: "Martyrdom" originally meant "testifying." It's just that handing out tracts or making speeches turned out to be a less effective marketing method than getting chewed by lions in public. "The blood of the martyrs is the seed of the church," as the meme has it.
All this leaves out the consideration of what a meme actually is: how you can identify an individual, transmissible unit of culture. And this is a huge difficulty for meme theory, one that has led most people who have considered the idea to attack it. For everything we know about human culture suggests that it is hugely different from the genetic system.
John Maynard Smith sides with Dawkins against Stephen Jay Gould in battles over biological evolution, but has nonetheless expressed very clearly one of the main theoretical objections to memes: "Two features are necessary for any genetic system that is to support adaptive evolution. The system should be digital, and it should not support 'the inheritance of adapted characters.'"
Genetic copying is digital -- the DNA strings code for 20 discrete amino acids, and nothing in between -- while cultural copying is analog and fuzzy. Cultural evolution is Lamarckian -- ideas are modified by experience before being passed on -- whereas it is a central dogma of biology that evolution cannot be Lamarckian. Genes are not selected directly, but as a result of their effects; a gene that makes an animal grow or behave in a certain way will be copied (or not) as a result of the behavior or capacities of the animal that carries it. (The animal's physical and behavioral characteristics are known in the jargon as the phenotype; the DNA that encodes the instructions to build the phenotype is the genotype, so that biologists say that genotypes are selected for the phenotypical effects.) But meme candidates are selected directly.
Culture appears to be a ladder -- especially to a scientist like Dawkins -- whereas evolution is a bush. Stephen Jay Gould calls memes "a meaningless metaphor"; virtual-reality pioneer Jaron Lanier thinks they are dangerous nonsense that promotes sloppy habits of thought. Yet it is all too easy for meme believers to dismiss their opponents as people infested with mind viruses -- as if they were not in the same situation themselves.
None of these difficulties bothers Susan Blackmore. She takes her strength from a syllogism: If the Darwinian algorithm is the only way we know to generate order and complexity in nature, which it is; and if our minds are full of order and complexity, which they are; then there must be a Darwinian algorithm operating on something inside them, and that something we shall call memes.
"The Darwinian algorithm can run whether you have got digital or analog information," she says. "It will run better and with higher fidelity if it is digital. But it can run over analog information."
Language is an analog medium, not a digital one, but Blackmore sees the emergence of language as a prime candidate for memetic evolution. "There is a problem of design in language: not just at the deep level of universal grammar, but at the observable level of the languages we actually speak. Why are there so many languages, and why do they stay so separate?" She cites research to show that after 40 years of mass television, the differing regional vocabularies of Britain are more divergent than before. Something is clearly combating the homogenizing effect of TV, and she believes it is the force of memetic evolution.
"Language allows for more accurate copying of thought as it grows more precise. Writing makes copying still more accurate, and thus will increase the size of the memeosphere," she argues. The Internet is the most recent expansion of this force.
"What's driving it? Is it our genes, or is it our individual interests? Memetics says that actually it is all in the interests of the memes. What has happened with the development of the Internet is a huge step toward high-fidelity copying -- with just enough errors to make evolution possible. We would expect that to happen because it is in the interests of memetic ambition, and it is completely irrelevant to our happiness. All we're doing is acting as a selection environment."
But memes may not need us to survive, any more than our genes any longer need the bacteria in which their ancestors originated. It is not that we will at some stage create artificial life or artificial consciousness, but that it will create successors to us.
"The more inefficient the technology, the more it relies on humans. But as the technology advances, humans become less and less important. Books made humans less important as a medium for the survival of information. We ought to be able to predict from a memetic analysis of what sort of hardware will best propagate the most memes with the highest fidelity -- and that is what we will find ourselves building."
Blackmore and I were both, by this stage, sitting at open laptops, typing and talking in a synchronous overload; and for a moment I suspected that the laptops were using us to communicate, rather than the other way round.
"The point at which it really takes off," she said, "will be the point at which you have robots which directly imitate other robots in really complex ways. Nothing in artificial intelligence does that now, but when it does, we will have truly human-independent memes."
Only we won't have them. It will be the other way round -- if they can be bothered with us.