Thursday, July 2, 2009

The Selfish Gene

The Selfish Gene, by Richard Dawkins
1976, 352 pp. Revised edition 1989.

Richard Dawkins has recently resigned from the Simyoni Chair for Public Understanding of Science at Oxford, so it is safe to say now, without fear of future contradiction, that he has been exceptionally good at his job. His name has become ubiquitous, well beyond notorious, reviled by some and admired by many; certainly he is the best popularly known academic in his field. I was prepared to call The Selfish Gene his magnum opus, but he assures the reader that it is not: that honor is reserved for The Extended Phenotype. Instead The Selfish Gene is his manifesto, his opening broadside into the world. It details the theory he has spent his career elaborating and explaining, and is certainly one of the most exciting popular science books I've ever read.

The Selfish Gene was first published in 1978, to rave reviews. My copy is the revised edition from 1989, which sports 65 pages of closely-typed endnotes. You must read them. Some are refutations of claims in the original text, some are elaborations, some detail interesting new experiments and theories, some conduct long-running arguments, and some are responses to critics. They are an example of a scrupulously honest scientist in critical engagement with his own work, and they offer quick, incisive intellectual pleasures in ways the text, with its burdens of exposition and clarity, sometimes does not.

The first several chapters consist of Dawkins laying out his theory and the analytical machinery with which it can be applied and analyzed. I will try to explain it in brief, but Dawkins' words really are the best way of expressing the theory. Genes are the unit of natural selection, because they can replicate themselves. Individuals are essentially "survival machines built [and programmed] by a short-lived confederation of long lived genes." Genes interact with one another, so what matters is how well genes are able to survive in the context of all the other genes in an individual, and get themselves passed on to another generation. Dawkins is at pains to be clear that he is not a genetic determinist, especially when it comes to human beings: "it is perfectly possible to hold that genes exert a statistical influence on human behavior while at the same time believing that this influence can be modified, overridden or reversed by other influences. Genes must exert a statistical influence on any behavior that evolves by natural selection." In humans, the most prominent influence is culture, in animals there is obvious variation in individual behavior. But quite a lot of behavior is statistically influenced by genes, which get passed along to future generations depending on the success of the behaviors they promote. Behaviors which can sustain themselves in a population over time are "evolutionary stable strategies" (ESS) which Dawkins proves, using game theory pioneered by Maynard Smith, is at the heart of most animal (and possibly human) behaviors.

The next several chapters consist of Dawkins applying the theory to sets of observed behavior: aggression, family planning, generational conflict, gender conflict. The first of these is fascinating, since it is the first time we get to see the theory in action and get walked through the game theory process to analyze behavior. The next three wear on quite a bit, and left me restless enough to start getting critical. Essentially, it seems to me that there is a falsifiability problem here. These middle chapters consist of the same pattern: Dawkins considers a puzzling or contradictory behavior in some species of animal. He then demonstrates using some simple game theory (only ever the Prisoner's Dilemma, by the way, never extensive-form games, or imperfect information games, or anything more advanced, probably due to the non-specialist audience) how exactly that puzzling behavior is really the dominant or stable strategy from the perspective of selfish genes, and explains that genes for that behavior must therefore have become prevalent in the gene pool. This struck me as the biological equivalent of Hegel's old chestnut about the real being rational and the rational being real. So how can his theory be falsified? Any behavior can be explained away on the grounds that "well, it exists, so it must be an ESS, and therefore caused by a selfish gene."

Now to some extent, this is unfair. Dawkins is writing for a popular audience, and therefore doesn't lay down any actual math, any formal modeling, or any of the other specialist machinery I know perfectly well he has at his disposal. It may also be the case that the selfish gene theory is just that good. I do not have the biological expertise to evaluate the problem, but in the interest of reviewing The Selfish Gene in particular rather than the selfish gene in general, I must admit that I finished the book unconvinced that there was any behavior Dawkins couldn't explain away, and I found this disquieting.

I also have a substantive complaint. Chapter 7, on family planning is interesting when dealing with animals, but an unmitigated disaster when dealing with human beings. As I so presciently noted recently, it is a sad example of an intellectual who begins to wind himself up on the issue of "overpopulation" and immediately deploys a series of unexamined, exploded old Malthusian fallacies. "Individuals who have more children than they are capable of rearing are probably too ignorant in most cases to be accused of conscious malevolent exploitation," he writes, after detailing what he sees as an endless and stupid population explosion. This comes perilously close to Huxley's idiot, paranoid phrase about the "teeming illiterates." But then Huxley seemed to be a reprehensible bigot derived from a long line of reprehensible bigots. Dawkins is neither reprehensible nor a bigot; indeed, he is a very good scientist--too good to make such blunders. Population and demography and their relations to poverty and development is a deeply researched field of development economics, and Dawkins ought to have at least acquainted himself with some of their findings. The poor, especially the rural poor, do not--I repeat, DO NOT--have lots of children because they are stupid, lazy, carnal, illiterate, greedy, gluttonous, lustful animals. This is apparently impossible for comfortable, otherwise educated and sensitive intellectuals to grasp. The poor have lots of children because a) children start work very young and therefore are a productive asset in poor, rural places instead of a drain on food and finances as they are in the First World, b) children will grow up to provide for their elderly parents, who will not have the benefit of a social safety net, c) high infant mortality lends itself to having more children, to make up for the unknown number which will die, d) women who do not work either through cultural stupidity or a simple lack of jobs face fewer opportunity costs for having children and less power to make family planning decisions, and e) poor families in general face a lower opportunity cost for having children. For a rich couple the choice is between a yacht and another child; for the poor it is a gamble that enough of your children will survive that you will be able to live off of them in your old age. Dawkins thankfully does not trot out that great shibboleth of First World neo-Malthusianism: that it is bad to save people, help them get medicine and healthcare, or provide food aid because people who survive will just have vast broods of progeny and you will end up back at square one. It was a relief that he had the sense not to be that stupid, but it was still discouraging to see such fallacies and ignorance from such a methodical scientist and keen intellect.

I stalled on that chapter, but persevered, and in the end was very glad that I did so. My complaints aside, Dawkins is a delightful writer. He has a gentle, urbane sort of tone which barely restrains the obvious excitement and passion he has for the subject. He seems very happy to be writing the book, and very happy to have us read it. He has an endearing habit of adding a decorous, dainty exclamation point at the end of a sentence to indicate that he is indulging in a little joke. He speaks both of himself in the first person, to explain his thinking, of the reader in the second, to establish a rapport and keep attention, and of both of you together to bring you along by the hand through the world of evolutionary biology. There is no sense of Dawkins the bloody-knuckled rhetorical pugilist who turns up lacerating pseudo-science books in the Times or reducing Bill O'Reilly to stunned silence. He takes a few cheerful potshots at religion, mainly in the endnotes, and is not shy about making his opinions clear on theories he finds unconvincing. But The Selfish Gene is not a polemic. It is a book by someone who is very good at what he does talking about his absolute favorite thing, and therefore it is a terrific read. He has a terrific command of biological and zoological information, and I learned quite a bit from his examples and analysis. He is also a great popularizer and communicator of the abstruse, complex ideas of W.D. Hamilton and John Maynard Smith, among others. For people (like me) who are utterly unfamiliar with the discipline, this is an excellent place to start. He also clearly knows that he is presenting a scientific theory which is subject to test and revision. At one point, he challenges the reader outright: "If anybody does not want to admit that parental care is an example of kin selection in action, then the onus is on him to formulate a general theory of natural selection that predicts parental altruism, but that does not predict altruism between collateral kin. I think he will fail."

Even more intriguingly, he ends with two chapters which provide a sort of link between the gene, a tiny molecule, and the necessity of the welfare state. No, you did not mis-read that sentence. He argues throughout the book that his selfish gene theory, with all of its implications against altruist behavior, and against group-selection theory, predicts behavior but does not advocate or condone it. His conclusion is this: "If there is a human moral to be drawn, it is that we must teach our children altruism, for we cannot expect it to be part of their biological nature." He devotes an entire chapter to explaining how reciprocal atruism is an evolutionarily stable strategy and how rebellion against our selfish genes is an inherently and beautifully human act. I was very glad he did this, since many times earlier in the book he made the point that altruist behavior is not stable, since it is open to exploitation by selfish individuals. I spent a lot of time wondering why he was dodging the social science implications of that conclusion, and was glad he returned to it from the beginning.

Now, I have not done an elaborate job of summarizing his points. Perhaps you have thought of an exception to his theory. Social insects, for instance, or parental sacrifice, or the dangeorus warning calls birds give when predators appear. Dawkins deals with most of these, in greater detail than I am willing to summarize here. Read the book. It goes quickly and is extremely interesting, especially for people preoccupied by the social sciences, since it functions as a sort of exhilarating (and exhilarated) introduction to a whole world of knowledge and controversy we otherwise ignore, at our own peril.

No comments: