A Passion for Evolution

Most science popularizers are not controversialists. Most see their task as a simple explanation of science that already sits, tedious and unread, in textbooks. The job of making science appealing to the layman often encour-ages breathless tales of high-tech adventure (a genome project, say) and almost always entails a good deal of dumbing down. Richard Dawkins’s work has never fit this mold. His books on evolution have featured ideas that were only then making their way into textbooks and he has never hesitated to offer his own (sometimes radical) extensions of these ideas. But most important, his books have emphasized ideas, and they have been offered in their fullness, without dumbing down. In the course of his work, Dawkins has also not shied from controversy. He has publicly battled both fellow scientists and religious leaders, and he has made enemies.

Although Dawkins is an extraordinarily popular and prolific writer—his books include The Selfish Gene (1976), The Extended Phenotype (1982), The Blind Watchmaker (1986), and Climbing Mount Improbable (1996)—his shorter pieces have never been collected until now. The thirty-two essays bundled in A Devil’s Chaplain appeared originally in newspapers or magazines, or as forewords to books or as eulogies; a few are previously unpublished. Because these essays span nearly all of Dawkins’s career so far—the earliest appeared in 1978 and the latest in 2003—they provide a sort of chronicle of his thinking. And because each group of related essays is preceded by a short new introduction, we learn where Dawkins now stands on a number of issues.

A Devil’s Chaplain reveals several things about Dawkins, some surprising and some not. Not surprisingly, the book confirms his reputation as a superb prose stylist, perhaps the best popularizer of science working. Whether you agree with him or not, you are never unsure of his meaning and his writing is, in places, stunning. Also not surprisingly, A Devil’s Chaplain includes hard-hitting pieces that attack several of Dawkins’s traditional targets: pseudoscience (New Age crystals get it here), postmodernist obscurantism (he’s fond of Alan Sokal’s Social Text hoax), and religion.

As for the surprises, the biggest is Dawkins’s breadth. For those who know him only as a champion of Darwinism, it’s news to see that he has written a number of beautiful pieces on the Africa of his youth, as well as passionate essays decrying educational systems that sacrifice love of learning for test performance. And his position on several controversial issues may not be what you expect. Though Dawkins is sometimes viewed as soft on sociobiology, A Devil’s Chaplain includes a superb polemic against the errors inherent in human genetic determinism and another against the evils of social Darwinism. The book also features several surprisingly moving eulogies to departed colleagues, including one to Stephen Jay Gould, with whom Dawkins often sparred.

Despite his breadth, Dawkins is surely best known for three things: his defense of the selfish gene view of biological evolution, his invention of the selfish meme view of cultural evolution, and his animosity toward religion. A Devil’s Chaplain takes up each of these themes, some more convincingly than others.

1.

Dawkins has spent much of his career defending a particular view of Darwinism. This so-called selfish gene view grew out of work in the 1960s by George Williams and William Hamilton. While Darwin argued that evolution involves a kind of survival of the fittest, Hamilton, Williams, and their heirs argued that it’s the fittest gene that matters, not the fittest organism. To see what this means, consider an example. When a small bird spots a hawk overhead it will often issue an alarm call, warning its flock-mates of the predator’s presence. The odd thing is that this behavior—which we’ll assume is instinctive, that is, genetically based—is “altruistic.” By sounding the alarm, a bird may well save its flock-mates but it simultaneously calls attention to itself, increasing the odds that it will be attacked by the hawk. How could such a behavior evolve?

If you think of Darwinism in traditional terms—as competition among different organisms—the answer isn’t obvious. A bird who sounds a call (and so perhaps gets eaten) is unlikely to have more offspring than a bird who keeps quiet (and so probably avoids getting eaten). And having more offspring is what Darwinism was supposed to be all about. But if you think of Darwinism in selfish gene terms—as competition among different genes—the answer is clearer. A gene that makes a bird emit an alarm may decrease the odds that the calling bird survives but it can increase the odds that the gene for alarm-calling survives. The reason is that the flock-mates who are saved by the alarm are, like all flock-mates, likely to be related to the caller; and relatives, by definition, tend to carry the same genes, including the gene for sounding the alarm. In effect, then, the alarm-call gene is warning—and saving—copies of itself. Those copies just happen to reside in other organisms. The counterintuitive conclusion is that a gene that sometimes causes an organism to sacrifice itself can increase its frequency by natural selection. The alternative kind of gene—one for not emitting an alarm call—can decrease in frequency, since such genes are on average less likely to be passed on to the next generation.¹

To Dawkins and other advocates of the selfish gene view, such examples reveal something deep about Darwinism: natural selection acts at the level of competing genes, not competing organisms. It is genes that are engaged in a struggle for existence and we can, therefore, expect them to “selfishly” do whatever they can to increase their representation in the next generation. (The quotes emphasize that genes are not consciously selfish; it’s just that their dynamics look like those of consciously selfish agents.) In the end, the selfish gene view suggests that the properties we see in organisms are those that maximized the survival of genes, not the welfare of organisms. Taken to its logical conclusion, genes begin to seem like manipulators who build organisms in whatever way best serves the genes’ “interests,” whether or not these coincide with the organisms’ interests. Indeed Dawkins spoke of organisms as mere “vehicles” for genes or as “lumbering robots” that were “blindly programmed to preserve the selfish molecules known as genes.” It was this view that he popularized in his first book, The Selfish Gene, and that he continues to defend, though in modified form, in A Devil’s Chaplain.

Dawkins’s defense was (and mostly still is) against Gould. Dawkins and Gould engaged in an extended dispute over “units of selection”: Is the gene the fundamental unit on which natural selection acts or can selection act at other levels too? Gould championed “hierarchical selection,” the idea that selection can act at any level in the biological hierarchy, from single genes to entire ecosystems. He was particularly fond of “species selection,” in which whole species are the target of natural selection. (Species that split into new species faster than, or become extinct slower than, other species will become more common.) To Gould, there was nothing particularly special about the gene level and any claim to the contrary reflected a regrettable misunderstanding of Darwinism. Though Dawkins and Gould disagreed about many other matters, the units of selection debate was particularly protracted and noisy.

By the time of Gould’s death, however, the debate had begun to subside. To some extent, this reflected the fact that evolutionists grew tired of it. To a considerable extent, though, it reflected the fact that Gould grew less vague and Dawkins grew less strident. Gould and his colleagues finally offered a precise definition of species selection, and most evolutionists now accept that natural selection can in principle act at any level in the biological hierarchy. Dawkins admits as much in A Devil’s Chaplain (though in a roundabout way: “Indeed it may, after a fashion…”). But no matter, Dawkins’s real admission came earlier, in The Extended Phenotype, in which he rejected the claim that the selfish gene view represents the right way to think about evolution and took up the claim that it represents a powerful, though not necessarily privileged, point of view. Though this may seem a retreat from the selfish gene line—and in strictly philosophical terms I suppose it is—in scientific terms, it is not.

Powerful new points of view are what science is all about. Most scientists would trade a technically sound but intellectually sterile point of view for a provocative and fertile one any day. And on these more pragmatic grounds there can be no doubt that Dawkins had the edge: the selfish gene view has been exceptionally fertile, providing unexpected insight into everything from the organization of insect societies to the spread of “junk” DNA that has no function. Indeed selfish gene thinking is now orthodox in evolutionary biology and, among many evolutionists, represents a near reflex. It is certainly true that Dawkins’s early rhetoric was sometimes extreme. But it is more true that selfish gene thinking has delivered a number of important insights. The same cannot be said for hierarchical selection, as Gould himself lamented in his final major publication, The Structure of Evolutionary Theory. Indeed while many of us suspect that higher-level selection occurs, the evidence for it is, so far, frustratingly weak.

2.

Dawkins did not stop with biological evolution. At the close of The Selfish Gene he pushed the idea of a selfish replicator in a radical new direction: the Darwinian logic of selfish genes may extend beyond genes. In particular, Dawkins argued that cultural entities like ideas, tunes, and fashions behave in a way that’s analogous to genes: some of these entities, which he called “memes,” spread through a culture faster than others. Certain advertising jingles, for instance, are (annoyingly) catchy and spread among people at a remarkably fast rate: I hear a jingle, I hum it; you hear me hum it, you hum it; and so on. This jingle is therefore a fit meme: it is good at propagating among minds and so can be thought of as increasing in frequency in a population of brains by a kind of cultural natural selection. (Note that a meme, like a gene, can be fit whether or not it’s good for the organism. A jingle isn’t fit because it does anything for me; it’s fit because it spreads fast. Note too that memes have no necessary connection with biologi-cal fitness: a fit meme needn’t boost your production of children.) Dawkins boldly suggested that we might think of cultural evolution as a struggle among replicating memes in the same way that we think of biological evolu-tion as a struggle among replicating genes.

The idea of memes is interesting and there’s probably something to it, at least when taken loosely, as a provocative analogy between biology and culture. But A Devil’s Chaplain makes it clear that Dawkins takes the idea more seriously than this. Though Dawkins seemed at one time to soften on memes, he now insists that he is as enthusiastic as ever. Indeed he is “delighted” that others are now making the meme “a proper hypothesis of the human mind.” Dawkins seems particularly fond of the philosopher Daniel Dennett’s suggestion that memes played a key role in the evolution of the human mind.² But this stronger version of the meme idea—the notion that memes provide a basis for a real science of culture or mind—has been sharply criticized.

One of the best-known criticisms is that the analogy between genes and memes cannot be taken seriously since the gene is digital while the meme is not. Genes are digital in the sense that they are made of DNA, which is a string of chemical letters called A, T, G, and C; computer code, by analogy, is digital because it’s a string of ones and zeroes. The digitalness of genes helps them replicate with high fidelity: if you carry a T at some position in your DNA, you can be 99.9999999 percent sure that one of your parents did too. Memes, on the other hand, are not digital; instead, they seem fairly fluid: ideas blend and blur at alarmingly high rates as they pass through multiple minds. (This is why we don’t trust fifth-hand gossip.) Memes thus seem to replicate with low fidelity. This is important as natural selection works well only with high-fidelity replicators. (Think how hard it would be to select for bigger apples if in every generation most genes for bigness turned into genes for smallness.) Memes would not, then, seem to provide particularly good material for Darwinian evolution.

Dawkins now thinks he sees a way around this problem. Because his attempted solution is important—indeed, it potentially revives hopes of a serious science of “memetics”—it’s worth looking at in some detail.³

The key point is that memes might be copied with high fidelity even though they’re not digital. To show how, Dawkins contrasts two games. In the first, a child copies a drawing and a second child copies the first child’s drawing, and so on. By the time the twentieth child makes his drawing, it will probably barely resemble the first. This is a low-fidelity meme. In the second game, a child teaches a second child how to make a paper origami, the second child teaches a third, and so on. In this case, the origami made by the twentieth child will probably closely resemble that made by the first. This is a high-fidelity meme. Why the difference? The reason, Dawkins says, is that in the first game the child is “copying the product” while in the second he is “copying the instructions” for how to make the product.

The distinction is important because when a teacher folds the paper slightly incorrectly, the student will typically figure out that, say, the “instructor intended to fold all four corners into the exact centre of a perfect square.” The child is smart enough, that is, to copy the “inferred instruction.” This inference makes all the difference as it ensures that accidental deviations from the intended task don’t get copied, yielding a high-fidelity meme. Dawkins concludes that “these considerations greatly reduce, and probably remove altogether, the objection that memes are copied with insufficient fidelity to be compared with genes.”

But I’m not so sure. For one thing, the low fidelity of memes is a simple observation: we all know that ideas change as they pass through many minds. This brute fact is unchanged by clever argumentation about how memes could be replicated with high fidelity; the fact is they often aren’t. To put it differently, Dawkins’s attempted fix is at best relevant to a subset of memes: the existence of game two does not, after all, make game one disappear. Moreover I suspect that most memes behave like game one. When I hum a jingle, for instance, I’m not copying the instructions for the jingle; I’m copying the jingle. Finally, Dawkins’s fix only seems to work because he’s smuggled in a battery of mental processes like inference, intuition, and idealization: the child figures out the “inferred instruction,” intuits what the “instructor intended,” and correctly identifies certain “idealized tasks.”

The problem is that it’s all this inference, intuition, and idealization that does the heavy lifting in Dawkins’s scenario, not memes. It’s hardly surprising that if every child infers the same implied task, all children will pass on instructions for the same task. But this leaves wholly unexplained why and how each child infers the same thing—and this is the source of high-fidelity copying in Dawkins’s scenario. While I wouldn’t claim that this objection is fatal, it at the least suggests that, if you want to understand the mind, you’re probably better off trying to understand inference, intuition, and idealization than memes.

But there’s another problem, one that has little to do with the gene-meme analogy but that’s at least as serious: unlike the selfish gene view, the selfish meme view hasn’t led anywhere. Where are the puzzling phenomena that have been explained by memes? Dawkins provides no examples and I suspect there aren’t any. The truth is that the meme idea, though a quarter-century old, has inspired next to no serious research and has failed to establish a place for itself in mainstream cognitive science, psychology, or sociology. Though laymen often have the impression that scientific ideas die in decisive experiments, far more often they die because they didn’t suggest many experiments. They failed, that is, to inspire a rich research program. Though I could obviously be proved wrong, and while I have no problem with the notion that some science of cultural change may be possible, I’m far less confident than Dawkins that memes will play an important role in any such enterprise.

3.

Dawkins’s passion for evolution is perhaps matched only by his hatred of religion. Indeed Dawkins has railed so often against religion that his reputation as a God-basher may now nearly rival his reputation as a science-booster. A Devil’s Chaplain leaves little doubt that the reputation is well deserved. Arguing that those who have masked their contempt for religion must speak out, Dawkins lets loose. He announces that religion is a “dangerous collective delusion” and a “malignant infection.” Acknowledging that this position may seem “contemptuous or even hostile,” he insists that “it is both.” Asked why he is so hostile to organized religion, he answers that he’s not particularly fond of disorganized religion either. Indeed: “I think a case can be made that faith is one of the world’s great evils, comparable to the smallpox virus but harder to eradicate.”⁴

I suspect most readers will remember A Devil’s Chaplain more for its broadsides against religion than for its defenses of Darwinism. What readers may not realize is that Dawkins’s position on religion reflects another of his long-running disagreements with Stephen Jay Gould. Gould believed that science—in its frequent excursions into ethics and its vague denunciations of religion—had overstepped its proper bounds. He also believed that religion—in its sporadic pronouncements on empirical matters—had overstepped its bounds. Gould’s solution, which he championed at book length, was simple: we must distinguish the legitimate sphere of science (the physical universe) from the legitimate sphere of religion (meaning, value, and ethics) and we must ensure that neither intrudes on the other.

I must admit that I didn’t find Gould’s defense of religion entirely convincing.⁵ To my surprise, though, I find Dawkins’s attacks on it even less so. The problem is not that Dawkins’s conclusion is necessarily wrong; the problem is that the arguments that lead him there are surprisingly shaky.

His first argument is that religion is just plain false. The events spoken of in sacred texts—six days of Creation and three days to Resurrection—did not occur. So why then do so many people believe them? Dawkins thinks he knows the answer: religious beliefs are viruses. Religious ideas, like all ideas, are memes; but religious memes, unlike many, are useless “mind parasites.” Religious beliefs exploit the fact that children’s minds are machines that copy—and believe—anything they’re told. So just as desktop computers can be infected by computer viruses, young human minds can be infected by religion viruses.

But if religious ideas are viruses, then why aren’t scientific ideas? After all, we foist all manner of scientific ideas onto young minds and they soak them up with astonishing efficiency. But Dawkins insists that scientific ideas are not viruses. They are more like mental adaptations:

Scientific ideas, like all memes, are subject to a kind of natural selection, and this might look superficially virus-like. But the selective forces that scrutinize scientific ideas are not arbitrary or capricious. They are exacting, well-honed rules, and they do not favor pointless self-serving behavior. They favor all the virtues laid out in textbooks of standard methodology: testability, evidential support, precision, quantifiability, consistency, intersubjectivity, repeat-ability, universality, progressiveness, independence of cultural milieu, and so on.

I confess that I find this argument astonishing. Why in the world should conformity to scientific criteria decide what counts as a “good, useful” meme? Why aren’t good, useful memes the ones that make you happy, or give you a sense of belonging, or increase the odds of having cooperative friends about? If anything, these criteria would seem more natural than Dawkins’s. But the deeper point is that there are no natural criteria. The whole point of memes is that a good meme is one that increases in frequency, period. Now we, as armchair memeticists, are free to partition successful memes into those that are “useful” vs. those that aren’t, but someone has to decide: useful for what? For describing nature? Science is a useful meme. For building community? Religion is a useful meme. In the end, Dawkins’s religion-is-a-virus argument comes perilously close to tautology.

Dawkins’s second problem with religion is that it’s the root of much evil. A Devil’s Chaplain is run through with the murder and mayhem that follows on the heels of religion, and the Crusades never seem far from his mind.⁶ But Dawkins’s history seems curiously Victorian. In his drive to show that religion is the source of so much evil, he must obviously confront the awkward fact that the twentieth century was largely a chronicle of secular evil. Stalin, Mao, and Pol Pot were atheists and Hitler wasn’t particularly pious. Dawkins deals with the problem in an especially simple way: he ignores it. Except for a mention of Hitler, he sidesteps what is arguably the key lesson of the twentieth century—that secular ideologies, including atheist ones, can inspire atrocity and genocide as readily as any religious creed. And Dawkins’s treatment of Hitler is remarkable: arguing “please don’t trot out Hitler as a counter-example,” he notes that Hitler never renounced his Roman Catholicism and quotes from an obscure speech in which the future Führer emphasized that he was a good Christian boy. Dawkins’s normally robust skepticism seems to fail him here and he’s silent on the obvious interpretation—that Hitler knew how to manipulate a Catholic crowd.

The point is not that religious views don’t sometimes lead, directly or indirectly, to evil. Of course they do. The point is that they have no monopoly: nationalist views (Italian fascism), economic views (child labor), and even scientific views (eugenics) have all had horrid consequences. Now in the last case Dawkins would surely argue that it was the abuse of science that led to acts of evil (forced sterilization, racist immigration policies). And I would agree. But if you allow this kind of move for science, it’s a bit unclear why you don’t allow it for religion too: Did Jesus really intend the Crusades?

Though he’s less explicit about it, Dawkins’s third problem with religion seems to be that it’s been a thorn in the side of science; it has stood in the way of truth. He’s fond, for instance, of pointing to the Galileo affair and he never tires of ridiculing Rome’s late reconciliation with Darwin. But these jabs are potentially misleading. The popular impression of long warfare between Church and science—in which an ignorant institution fought to keep a fledgling science from escaping the Dark Ages—is nonsense, little more than Victorian propaganda. The truth, which emerged only from the last century of scholarship, is almost entirely unknown among scientists: the medieval Church was a leading patron of science; most theologians studied “natural philosophy”; and the medieval curriculum was perhaps the most scientific in Western history.⁷ There were of course some skirmishes between Church and science (the Condemnations of 1270 and 1277, the Galileo affair) and the Church made a number of stupid decisions, but it’s not entirely clear that these were more egregious than those made by secular institutions (e.g., Lysenko’s suppression of Mendelian genetics).

Now you might argue that this is all beside the point anyway. You might argue that what conflicts did occur between science and religion were due to misunderstandings of one or the other. Indeed you might argue that Dawkins’s belief that science and religion can conflict reflects a misconstrual of the nature of religious belief: while scientific beliefs are propositions about the state of the world, religious beliefs are something else—an attempt to attach meaning or value to the world. Religion and science thus move in different dimensions, as Gould and many others have argued.

Dawkins is dead set against this argument. It is, he claims, a popular trick among theologians and amounts to nothing more than evasion, a cynical and “dishonest” dodge designed to sidestep the findings of science. Dawkins also believes the different-dimensions move is a modern invention:

Don’t fall for the argument that religion and science operate on separate dimensions…. Religions have historically always attempted to answer the questions that properly belong to science. Thus religions should not be allowed now to retreat from the ground upon which they have traditionally attempted to fight.⁸

But, again, I’m not sure Dawkins has his history right. Theologians didn’t concoct the view that there are different dimensions to dodge Darwin’s evolution or Hubble’s cosmology. The view (or something close to it) dates at least to Aquinas in his adjudication of the proper roles of theology vs. Aristotelian philosophy, including natural philosophy. And the Church fathers early on rejected the idea that the Bible could be read as a scientific document describing empirical events: Augustine, in the fifth century AD, forcefully argued against such literalism.⁹

In the end, Dawkins’s refusal to grapple seriously with the different-dimensions view is perhaps the least satisfying aspect of his treatment of religion. For one thing, it’s unclear why a scientist (as a scientist) should be bothered by those who surrender all pretense of describing the physical world. (Conversely, it is clear why he should be bothered by idiotic creationists or by those who merely feign such surrender.) For another, Dawkins runs some risk of sounding like a man who denounces novels because they’re all made up. When told that novels present a different kind of truth, one that doesn’t depend on a story’s literal factuality, he smells sophistry. Now I’m not sure this analogy holds well enough to sustain a different-dimensions defense of religion. But I do suspect that matters are considerably more complex—and considerably more subtle—than Dawkins’s arguments admit.