Michael Pollan's shortest and most focused book opens with seven words that have entered the language: "Eat food. Not too much. Mostly plants." The rest of the book is an explanation of why this apparently obvious advice is, in the context of modern nutritional science and industrial food, genuinely radical.
Nutritionism and Its Discontents
Pollan’s central target is “nutritionism” — the ideological framework that reduces food to its nutrient components and then attempts to engineer health by manipulating those components. The history of nutritionism is, he argues, a history of systematic failure: fat is bad, then fat is fine; carbohydrates are good, then carbohydrates are killing us; antioxidants are essential, then supplements don’t work; and so on, indefinitely.
The reason nutritionism fails is structural. We don’t eat nutrients; we eat food. And food is a complex, contextual system in which thousands of compounds interact in ways that isolate-and-test laboratory science cannot model. The reductionist approach breaks down the very thing it is trying to study.
The Western Diet as Pathogen
Pollan’s practical argument is simpler: the populations that eat traditional diets — Mediterranean, Asian, Indian, Latin American — don’t get the diseases of the Western diet: heart disease, type 2 diabetes, many cancers, obesity. The problem is not any specific nutrient; it is the Western diet as a whole, understood as an eating pattern.
Eat food. Not too much. Mostly plants. That's it. Everything else is commentary, and much of the commentary is bought and paid for by interests that profit from your confusion.
— Michael Pollan, In Defense of Food
6 Key Ideas From This Book
By reducing food to nutrients, nutritionism creates confusion that benefits the food industry (which can reformulate products to claim health benefits) and harms consumers (who replace genuinely nutritious foods with engineered substitutes). The ideology serves industrial food, not human health.
Pollan's first rule: don't eat anything your great-grandmother wouldn't recognise as food. If it has more than five ingredients, contains ingredients a home cook wouldn't use, or was invented after 1900, it is probably an engineered product rather than food.
Every traditional diet — regardless of its specific foods — produces better health outcomes than the Western diet. No traditional diet has ever been proven inferior through scientific investigation. This is strong evidence that the accumulated wisdom of food cultures, whatever its mechanisms, works.
When nutritionists isolate specific compounds from whole foods and test them as supplements, they consistently fail to replicate the benefits of eating the whole food. The interactions between compounds within a food — and between different foods in a meal — are too complex to reduce to single-nutrient models.
The most consistent finding across all nutritional epidemiology: populations that eat more plants have lower rates of chronic disease. The specific mechanism is less important than the pattern: a diet rich in diverse plants, with animal products as a complement rather than a foundation, produces better health outcomes.
Food cultures — their rules about what to eat, when, with whom, and in what combinations — encode dietary wisdom accumulated over generations. The abandonment of food culture in favour of nutritional science has been a net loss for human health.
Any Weaknesses?
Pollan’s seven-word manifesto is memorable but imprecise. “Mostly plants” still leaves enormous latitude — a diet of 60% ultra-processed plant foods is technically “mostly plants.” The book is better at critiquing nutritionism than at providing a usable positive alternative.
The advice to eat like your great-grandmother is also class-specific: which great-grandmother? A French peasant’s diet is very different from an American sharecropper’s, and neither is accessible to most people in the modern food environment.
Anyone confused by decades of contradictory nutritional advice who wants a clear, principled framework for thinking about food that doesn't depend on following the latest study.
The Omnivore's Dilemma by Michael Pollan for the food system context that explains why the Western diet exists, and How Not to Die by Michael Greger for the clinical evidence base supporting the "mostly plants" prescription.
Marketing professionals in the food industry. Pollan's analysis of how nutritionism is exploited in food product marketing is simultaneously a critique and a how-to guide.
The seven-word summary is the whole book, compressed. Some readers find the book elaborates the obvious at length. If you already eat well and understand food politics, you may get more from Pollan's other books.
Frequently Asked Questions
Is In Defense of Food worth reading?
In Defense of Food is the clearest and most principled short book about eating available. Its critique of nutritionism is devastating and its positive advice is sound. The most efficient possible introduction to thinking well about food.
Who should read In Defense of Food?
Anyone confused by decades of contradictory nutritional advice who wants a clear, principled framework for thinking about food that doesn't depend on following the latest study.
What is In Defense of Food about in one sentence?
Michael Pollan's shortest and most focused book opens with seven words that have entered the language: "Eat food.
The Verdict
*In Defense of Food* is the clearest and most principled short book about eating available. Its critique of nutritionism is devastating and its positive advice is sound. The most efficient possible introduction to thinking well about food.
→ Find on Amazon