"National Bestseller", it screams on the cover of my trade paper edition of this book. It's easy to see why: Pinker writes in an easy style with only occasional bits of technical jargon, and who wouldn't want to read a book that promised to explain everything from how we see shapes to why some people are successful and others aren't to the very meaning of life? Unfortunately, Pinker doesn't live up even to his own hype, let alone that of others who have made this book a bestseller.
Let me start with the very simplest of methodological issues: the title "How the Mind Works" is, at least as far as Pinker's book goes, complete nonsense. This is much more "How the Mind MIGHT Work." With his intellectual roots in the computational theory of mind, Pinker is fond of making arguments of the form:
Bzzt! Wrong answer. This sort of reasoning can provide ideas of how the mind might work, but it does not at all prove that the mind is wired the same as the computer program. Part of this is simply because X, Y, and Z might not describe the entire feature set of the mind. By analogy, suppose you run across a computer program that generates the numbers 1, 2, and 3 as sequential outputs. You discover that you can write your own program that simply counts, and it has the same outputs. "Aha," you say, "that other program is just counting, and its next two outputs will be 4 and 5." Imagine your surprise, then, when the next two outputs of the other program are 5 and 8, because it is not counting, it is deriving each output from the sum of the two previous outputs. Reasoning because of partial analogy fails in this case.
So it may be in all of Pinker's cases. Yes, some parts of the visual processing mechanism in humans act as if we had edge detectors and shape matchers. This does NOT mean that we have edge detectors and shape matchers. It is entirely possible that the entire computational structure that Pinker assigns to vision is simply wrong.
Even so, vision comes near the start of Pinker's book, where he still seems to have good contact with experimental results. As the book progresses, he applies his ideas to more and more abstract realms: stereotypes, altruism, family values, war, art, and, yes, the meaning of life (at least, the chapter title promises the meaning of life, though by my reading he doesn't even make a stab at this one). Throughout, his relentless approach is that minds-are-computers-evolved-by-natural selection. If something can't be explained from this viewpoint, it can't be explained period in Pinker's universe. If someone is not working from that viewpoint, then their ideas are not even worth mentioning, let alone seriously challenging.
And as Pinker moves into more abstract realms, the book seems increasingly like a catalog of his own prejudices. I just about gave up reading when he started discussing sex. In the Pinker universe, of course, we are all trapped by our genes into heterosexual sex, where men have evolved to impregnate as many women as possible and women have evolved to try to trap a single partner. Homosexuals are only interesting in that they prove this is the pattern: all male homosexuals are promiscuous, all female homosexuals are monogamous. Sexual variety is limited to the male drive to have multiple partners. There is nowhere in Pinker's world for such things as BDSM, water sports, or even oral sex. One wonders whether he even knows anything about the variety of human sexual behavior. His explanations of music and art are similarly unsatisfying. Pinker's explanation of music as a set of frequency interactions is as if a physics geek was content to have understood music by being able to record a sonogram. Most of us would agree that there is something more to it than that.
On the whole, I think Pinker fails in his ambitious program of understanding simply because he only has one idea: minds are computers that were formed by natural selection. The feeling of dissatisfaction this leaves me with is, I find, common to most of what I call "total systems": patterns of thought that explain everything by One Big Idea. It's easy to find other systems that fit this pattern: classic Marxism, modern American Libertarianism, Scientology, and many others. The truth, I think, must be more complex than One Big Idea can explain.
I would have been happy to have read this book if it had ended after reviewing what we know about the physiology of vision from experiments. There was some stuff there that I hadn't seen before. But to have to sit through all of Pinker's speculations to reach the end of the book pushed this firmly into the "skip it" pile for me.