Quite a number of years ago now, I first read Daniel Dennett's book Darwin's Dangerous Idea. This post is not about the central topic of that book, which is evolution (I'm sure I'll get into posting about that on this blog in time, but for now you'll have to read this article on my old site if you want to see where I'm coming from). Instead, I want to talk about one particular claim Dennett makes in his book: that Stephen Jay Gould did not believe in Darwin's dangerous idea, the central premise of evolutionary theory.

Just to recap briefy, Darwin's dangerous idea is that every feature of any living organism that looks like "design", i.e., like it is adapted to a particular purpose, was produced through evolution by natural selection. Actually, a better phrase to describe the process is the one Richard Dawkins first used in The Selfish Gene: "the differential survival of replicating entities." (Dennett calls it a "dangerous" idea because, as he argues persuasively in the book, it works in any domain, not just the evolution of living organisms. For example, cognitive science is converging on a model of the mind in which the same process is at work on several levels.)

Dennett's claim about Gould, then, is simply that Gould does not believe that Darwin's dangerous idea actually explains why living organisms are adapted to their environments. When I first read this claim, I thought: "That's crazy! Gould doesn't believe in evolution?" I will confess here that I liked, and still like, reading Gould's essays; I like his writing style and his way of weaving in baseball and other cultural references with his scientific discussions. So I was, to put it mildly, surprised to find Dennett, whose writing I also like and whose thinking I respect, making such a claim about Gould. Then I read the actual critique, and followed Dennett's argument, and realized why he said what he said, and thought: "He's right! Gould really doesn't believe in evolution. Holy --!"

More recently, I came across a post by Eliezer Yudkowsky on the "Less Wrong" site entitled Beware of Stephen J. Gould. It was only on reading this article that I fully realized just where Stephen Jay Gould really stood in the minds of much of the academic community, or at least the community of evolutionary biology. The post begins thus:

If you've read anything Stephen J. Gould has ever said about evolutionary biology, I have some bad news for you. In the field of evolutionary biology at large, Gould's reputation is mud. Not because he was wrong. Many honest scientists have made honest mistakes. What Gould did was much worse, involving deliberate misrepresentation of science.

As scathing as Dennett's critique of Gould was in his book, he didn't go this far. But on reading through Yudkowsky's post, and reading the further material he links to, I realized that something was certainly afoot. In particular, this letter to the editor of the New York Review of Books, by John Tooby and Leda Cosmides (in response to previous letters of Gould's) got my attention, with its catalogue of ways in which Gould "is giving non-biologists a largely false picture of the state of evolutionary theory", in the words of John Maynard Smith. If these criticisms were accurate, and they certainly seemed to hold water, Stephen Jay Gould was not what I thought he was. But it seemed preposterous, all the same, to think that Stephen Jay Gould, the poster boy for evolution in the eyes of the lay public, was somehow a traitor to the cause.

Not that I thought Gould was always right, or that I always agreed with what he said about evolution. As I noted above, I like reading Gould, and I also am in his debt for introducing me to the world of evolutionary theory. But I did start disagreeing fairly early on with a lot of what I read him saying about it. For example, I never bought his and Lewontin's case for "punctuated equilibrium" as any kind of "alternative" to "standard" Darwinism; as Dennett argues in his book (not that he is the only one to make this argument by any means) "punctuated equilibrium" is just standard Darwinism looked at on the appropriate scale. So even though I liked reading Gould, I felt a vague uneasiness whose root cause I couldn't quite put my finger on.

Dennett's critique in the book goes further than simply discrediting "punctuated equilibrium", though. He attacks and refutes Gould's position on "adaptationism" in general. Part of the attack is to simply be clear about what "adaptationism" is. It is not the claim (which no reputable evolutionary biologist has ever made, as far as I know), that all evolutionary change is the result of natural selection. Now that we can look directly at DNA and other evidence at the molecular level, it is clear that much evolutionary change has no impact on adaptive fitness and is therefore "invisible" to natural selection. In so far as Gould tried to argue that "adaptationism" was committed to the view that all evolutionary change must be explicable as an adaptation, he was simply setting up a straw man.

But Dennett also makes the point that, even if we restrict attention to evolutionary change that is adaptive, we have to be careful not to misunderstand what an "adaptation" is. An adaptation is any feature that provides a selective edge, regardless of how it came about. In particular, an adaptation does not have to be a step in a continuous process of selection for the same function. A feature that has one function in an ancestor can be "exapted" for a very different function in a descendant, and that still counts as an adaptation.

I was surprised to read this part of Dennett's critique, because I was sure I remembered Gould talking in his essays about precisely this point, and being on the same side as Dennett is in the book. For example, in one of his essays (I don't have his books handy to check which one), he says something along the lines of, "Critics of evolution often ask things like, What good is five percent of an eye? We argue that a feature being five percent of an eye is irrelevant because the ancestor who had the feature at that stage did not use it for sight." In other words, exaptation is adaptation, exactly as Dennett says.

But as the examples Dennett tirelessly catalogues in his book make clear, Gould was much more consistently on the other side of the question, arguing for an extremely narrow view of what "counts" as an adaptation, and inventing the term "spandrel" to apply whenever he saw an "adaptationist" being too free with his labeling. Reading this part of the book brought at least one part of my previous vague unease about Gould into focus: Gould was not consistent in his arguments, and often seemed to be more concerned with not agreeing with whoever he was arguing with than getting at the truth.

The bit about "spandrels" also brings up another quality I had often seen in Gould's writing, namely, that while he seemed to feel very strongly that he was "against" something, it wasn't always clear what that something was, or why he was against it, or what positive position he held that drove him to refute whatever he was refuting. You will look in vain in Gould's writing for any precise definition of what a "spandrel" is, and the same goes for other terms that he apparently thought were very important. Indeed, he never really gave a precise definition even of "punctuated equilibrium". And on those occasions when he did spell out reasonably precisely what he was talking about, his argument was obviously bogus, as with his claim that religion and science are "non-overlapping magisteria". (To see why this idea of his is bogus, ask yourself how far you would get with a sincere Christian by telling him that all that business about Jesus being the Son of God is just metaphor and isn't supposed to be actual historical fact, because historical facts are in a different "magisterium" from religion.)

Still, inconsistency and imprecision are not the same as deliberate misrepresentation. Is that more serious charge really justified? Dennett's book, as I noted above, does not make that charge, and does not really discuss the key omission from Gould's writing that justifies it. That's why reading Yudkowsky's article was such an eye-opener for me: because it was only then that I looked back at everything I had read of Gould's, and realized that not once had he made clear to me that evolution, at bottom, is about genes. Gould talked about evolution in terms of individual organisms; he also (in his vague way) talked about "group selection", the (supposed) evolution of groups as separate "units of selection" from individuals. But he never once got across to me the fundamental fact that genes are the primary units of selection.

It is this omission that Yudkowsky discusses in detail in his article. Gould essentially wrote as if the "gene's-eye view" of evolution had never been invented, let alone won over the entire field by its acknowledged superiority in making sense of the data. In hindsight, this should have been a huge red flag to me; after all, you can't read Dawkins, for example, for more than a minute or so without some mention of genes, if not the actual phrase "selfish genes". And I first read Dawkins not long after I started reading Gould; not the entire book The Selfish Gene, but at least the shorter article condensing its material, "Selfish Genes and Selfish Memes", that appeared in the book The Mind's I. So it wasn't as though I didn't know about genes and the role they played in evolution.

Why did it take me so long to catch on to this? One possible reason is hinted at in the closing of Tooby and Cosmides' letter:

Yet in the final analysis, there are genuine grounds for hope in the immense and enduring popularity of Gould. Gould is popular, we think, because readers see in "Gould" the embodiment of humane reason, the best aspirations of the scientific impulse. It is this "Gould" that we will continue to honor, and, who, indeed, would fight to bring the illumination that modern evolutionary science can offer into wider use.

I have already said that I found myself not really buying a lot of what Gould said. Why, then, did I keep on reading him? I think it was because I saw Gould the way Tooby and Cosmides describe readers as seeing him; he was a symbol of science as a force for good in the world, and he managed, among all the vagueness, to include some eloquent statements of that ideal. For example, in The Mismeasure of Man, even though he got a lot wrong when discussing the field of IQ testing (for one thing, he described the field as it was decades before he wrote, and critiqued IQ tests used during World War I as though they were representative of current ones), he still managed, in one sentence, to sum up why people get so concerned about standardized tests, however "scientific" they seem. The sentence was: "You can't judge an individual by a group mean." And however wrong he may have been about the actual scientific status of IQ tests (see for example this review, with Gould's response so that you can see his side of the story, or this article discussing Gould's treatment of "factor analysis", the statistical technique that underlies IQ tests), he was right that, in the eyes of too many non-scientists, IQ tests have been nothing more than a handy excuse to write off whole groups of people.

But if Gould did serve as a symbol of science fighting for good, that only makes it more disappointing that he felt he had to misrepresent his own field to do so. Tooby and Cosmides may have been more kind than they wanted to be in their closing; as Yudkowsky notes at the end of his post, "Many academic writers on Gould could not speak as sharply as Gould deserved." For myself, I keep coming back to the startling claim that I read in Dennett's book, that Gould did not believe in evolution. Gould was also an atheist and a Marxist, so apparently he also did not believe in the faith of his ancestors, Judaism, or the political foundation of his country. Was he just angry because others did not seem to share his particular brand of unbelief? Or was he just another aspiring revolutionary, trying to stir up the masses against the establishment? We'll never know. Perhaps the best lesson we can learn from Gould is that science does not need a "symbol"; it stands on its own, and the best thing we can do to "fight for" it is to understand and use it properly. The "Gould" that the lay public thought it was reading would have agreed with that, whatever the real Stephen Jay Gould might have said.