Few baseball players have ever plumbed the epistemological depths of baseball, few stockbrokers have written groundbreaking works of theoretical economics, and few of those who are good at physics have proved much good at metaphysics. Science simply isn’t the same thing as the philosophy of science, and few real scientists would imagine the two interchangeable.

But Michael Shermer is not a scientist, exactly. The founder of Skeptic magazine and a columnist for Scientific American, he’s more of a popularizer of science. The author of such books as the 1997 Why People Believe Weird Things—an enjoyable book, which summarizes the more scholarly work of others on such topics as the rage of accusations about Satanic abuse in child-care centers in the 1990s—Shermer has now followed up with The Believing Brain.

There are fewer actual case studies here than in Shermer’s previous work. He does mention a number of wild beliefs: alien abductions, economic bubbles, political fantasies, anti-Darwinism, ESP, autism-causing vaccinations, 9/11 conspiracy theories, and, naturally, faith in God. But the book is clearly less about the examples than about the theory Shermer uses to explain them all.

Shermer is not an original thinker, and does not really pretend to be. If you want a better written and better presented set of examples, pick up a copy of Charles Mackay’s 1841 classic Extraordinary Popular Delusions and the Madness of Crowds; if you want a better if still lightweight account of the underlying theory, get Carl Sagan’s 1995 The Demon-Haunted World: Science as a Candle in the Dark. Or skip the middlemen and go straight to Sextus Empiricus’s Outlines of Pyrrhonism and Karl Popper’s Logic of Scientific Discovery.

And yet, for someone trying to understand an intellectual movement like modern antitheism, it can be useful to read books like Shermer’s—for these are often where the ideas of more theoretical writers actually cash out, and where the real consequences of a subtle error sometimes become clearest.

Shermer’s theory looks like this: The human mind is inherently a “belief engine”; we perceive endless bits of information, and we must posit beliefs as ways of organizing and making sense of them. Our techniques for doing so break in two directions: toward attempting to discern a pattern in our perceptions and toward attempting to find a cause or a source of the perception. Shermer labels these mental impulses “patternicity” and “agenticity.” As we move through the world, we are presented with a swirl of emotions, formal ideas, and physical observations, and out of this confusing mélange we build possible explanations of the way things go together and the agency that brought them about. Having found a possible explanation, we then seek confirming evidence and deepenings of the patterns and agents we believe we have discerned. The result, Shermer claims, is that we live much of the time in “belief-dependent realism,” which is to say that our beliefs are shaping what we see in the world, rather than the world shaping our beliefs.

This could just be an attempt to do an end-run around the answer to David Hume’s skepticism that Immanuel Kant came up with—the argument that even space and time are not necessarily objective realities, for they are the required preconditions of perception and thus the subjective lenses through which, true or not, we must perceive the world. Shermer seems to be saying that we should take Kant at his word, recognize that the world is organized for us by our intellectual acts of patternicity and agenticity, and accept that the structure of belief is built into human perception.

But Shermer also believes in a dividing line between benign or helpful beliefs and malignant ones (like religion). In all things, rightly or wrongly, we believe we see patterns in what may be only random noise; and, correctly or incorrectly, we have difficulty imagining that there isn’t a cause or an agent behind the perceptions streaming at us. But Shermer thinks we may be able to discern a difference between valid beliefs and invalid ones by how passionately they are presented, the intolerance that results from them, and the lack of constraint with which they are held.

So what about the passion with which Richard Dawkins, the author of The God Delusion, puts forward his atheism, the intolerant disdain he has for religious believers, and the lack of constraint he shows in his willingness to redefine all terms to defend his believed pattern that religion is the cause of all war? Shermer would give Dawkins a pass because Dawkins is a scientist, and the scientific method is our salvation. Ronald Bailey, a smart science correspondent for Reason magazine, notes that Shermer claims to be an old-fashioned liberal, and he quotes with deep approbation Timothy Ferris’s old line, “Liberalism and science are methods, not ideologies.” One wishes Shermer actually believed it, for science emerges from The Believing Brain as a full-blown ideology, lifted out of its proper realm and applied to all the puzzles of the world.

Shermer is not an unpleasant writer, and The Believing Brain is not an unpleasant book. Though he’s an unreconstructed atheist, Shermer lacks the old monomania that makes most atheistical tracts a drudge to read. But he starts with what the philosopher Gilbert Ryle taught us to call a category error, imagining that the mechanisms of scientific knowing can be made the model of all knowing and that science can be easily lifted up to metaphysics. The Believing Brain reveals the upside-down nature of much that passes today for atheistical argument. Try standing all Shermer’s examples on their heads. The human brain is structured to look for patterns, he says—and therefore, he concludes, there are no real patterns in the world (except the ones that science shows us). Evolution has formed us to seek an agent cause for things—and therefore, Shermer concludes, the world has no agency. Conspiracy theories, from 9/11 Trutherism to UFO cover-ups, arise from our hunger for explanation—and therefore religion is an explanatory conspiracy. As a race, we’ve believed in more than ten thousand gods—and therefore God does not exist.

If it’s human nature, in other words, then it’s false. How science escapes this problem, Shermer doesn’t quite explain, but, more to the point, it’s a skeptical premise held without skepticism. Wouldn’t a more logical inference be that we seek patterns because patterns actually exist for us to see, our brains matched to reality? Couldn’t it be the case that we seek agency because the world itself testifies to an agent? And couldn’t our hunger for gods reflect the truth that is God? It’s more plausible, anyway, than assuming, as Shermer seems to, that we wouldn’t be so inclined to look for something if it actually existed.