Despite his dilettantish forays into mushroom foraging and wild-boar stalking, Michael Pollan is primarily a journalist, and he's at his best when he's parsing the esoteric political and social forces that have shaped the modern American diet. From the 1970s food-policy overhaul that flooded the market with cheap commodity corn (in the first and most rigorous section of The Omnivore's Dilemma) to the frontier thirst for alcohol that catapulted Johnny Appleseed to fame (in the first and most rigorous section of The Botany of Desire), Pollan is a consistently surprising, compelling reporter. But he also has a tendency toward self-indulgence—think of Dilemma's hunter-gatherer episode or Botany's self-conscious attempt to describe the subjective pleasures of a marijuana high. His new book is billed as a "manifesto," which doesn't bode well for the proportion of fact to flab. If anything, it's worse than it sounds: In Defense of Food is profoundly anti-intellectual.

If you're looking for advice about how to eat, you don't even have to crack the book open. It's right there on the cover (and in his "Unhappy Meals" essay, printed in the New York Times Magazine about a year ago): "Eat food. Not too much. Mostly plants." Hey, thanks! It's good advice, and so pithy you can chant it under your breath at the supermarket (or a year-round farmer's market, as Pollan and his merry band of Berkeley locavores would prefer). The next 200 pages are devoted not to advancing the argument but to whining about the way nutrition is studied, reported, and politicized in the United States.

Pollan starts off by attacking science, of all things. He focuses his ire on what he calls "nutritionism," a pejorative term for scientific research that attempts to isolate the impact on one's health of, say, saturated fat or beta-carotene. He's saying, in essence, that a little knowledge is a dangerous thing: Food marketers seize on soft correlations, then plaster reductive claims all over their products, and then, occasionally, take everything back. (The trans fats in margarine are not better for you than the saturated fats in butter; they're actually much worse.) Add to this the potential for bizarre food fads (low-carb spaghetti!), and you've got a compelling case for greater regulation of nutritional claims, better-designed research, and smarter reporting. But Pollan seems to have it in for the scientific method itself: No research that asks questions about a single variable can be worthwhile.

At the same time he's scapegoating science for changes in the way we eat, Pollan neglects other enormous developments that occurred over the same period—those social forces that were supposed to be his specialty. He writes, "The 'What to eat' question is somewhat more complicated for us than it is for, say, cows. Yet for most of human history, humans have navigated the question without expert advice. To guide us we had, instead, Culture, which, at least when it comes to food, is really just a fancy word for your mother." Entirely apart from that dumb uppercase C, there are a couple of unexamined assumptions here: First, that we should bemoan the loss of rigid ethnic food cultures instead of celebrating the fact that we now borrow from each other freely; and second, that we should be nostalgic for the time when "mothers" were in charge of the kitchen (if nothing else). In this context, advice like "Don't eat anything your great-grandmother wouldn't recognize as food" starts to sound slightly reactionary. For my great-grandmother, food was what she cooked all day long for her several children. We can't eat her kind of "food" every day—it's worth neither the time nor the human resources.

As in The Omnivore's Dilemma, Pollan is on firmer ground when he focuses on food policy. With the exception of sucrose, he notes, individual nutrients don't have powerful lobbies. So when the government attempts to set guidelines for nutrition, as a Senate committee headed by George McGovern did in 1973, advice that annoyed organized lobbies ("reduce consumption of meat") were rewritten to offend only invisible substances ("choose meats, poultry, and fish that will reduce saturated fat intake"). This kind of clear-cut, useful reporting makes Pollan worth reading. Let's hope his next book drops the hostility toward science and the barely veiled nostalgia for a prefeminist age and focuses on what it would actually take to change American eating habits. recommended

Michael Pollan reads from In Defense of Food on Wed Feb 13, Town Hall, 1119 Eighth Ave, 652-4255, 7:30 pm, $5.