This is the third in a series of Medium posts. They start here. Excerpted from All Natural: A Skeptic's Quest to Discover If the Natural Approach to Diet, Childbirth, Healing, and the Environment Really Keeps Us Healthier and Happier, from Rodale, which comes out January 29th.
Part III: Food structure
Assumption 1: Molecules matter, food is irrelevant
It’s relatively easy for scientists to measure the type and number of molecules of any nutrient (using mass spectrometry for instance) but infuriatingly hard to see how they fit together to form actual food. This is a common problem for science—categorizing and counting the parts of a system is simple (or at least feasible) but understanding the relationships between the parts is difficult. So for a long time many scientists simply assumed that the structure of food was irrelevant. When the early nutritionists thought about food structure at all, it was to plot its destruction. The molecular nutritionists, remember, had won their fame in identifying the nutrients needed to prevent deficiencies, so they were prejudiced against foods that clung obstinately together. They wanted simple foods that digestive tracts could easily absorb. For years, therefore, scientists encouraged processing the complexity out of foods. The results were products like Wonderbread—vehicles for vitamins and minerals that barely required chewing.
“They’re rocket fuel,” German said, “the nutrients have just been atomized—they go into the bloodstream like they’ve been injected.”
Milk suggests that perhaps we should be striving for the exact opposite: Calories bound up in complex structures that break down bit by bit. Milk doesn’t start out in complex chunks; in order to pass through a narrow aperture—the nipple—it has to be fluid. But once through, enzymes in the baby’s stomach trigger a transformation of milk proteins and, like a ship unfolding in a bottle, they open and link together, forming large curds. Put another way, evolutionary trial and error has fixed it so that babies drink milk, but digest cheese. Next, of course, the baby must break down this cheese to extract the nutrients. Evolution would not tolerate the expense of knitting together this complex structure then breaking it down again if it had no benefit. But according to the dominant dietary theory, which holds that food is simply independent molecules, there is no benefit: Chunky milk and fluid milk are nutritionally identical.
It’s unequivocally apparent to German that the structure of foods matters. A simple restacking of identical nutrients was so important, so advantageous in the sink-or-swim test of natural selection, that it made it worth solving the devilish engineering problem of getting cheese through a nipple. The implications are enormous: It means that a nutrient that’s good for you in one food may be bad for you in another. And that makes the nutritional information boxes required on all food packaging almost completely irrelevant: The same type of fat may have different consequences if it arrives in a slice of coconut, a steak, or a scoop of gelato.
 If we are to give credit where credit is due, the honor for the invention of cheese belongs to babies. We still employ the enzyme, rennet, from the stomach of a baby goat, or sheep, to perform this magic trick. It’s likely the first man-made cheese was a happy accident that occurred when someone used a bag made from a calf’s stomach to carry milk.
Excerpted from All Natural: A Skeptic's Quest to Discover If the Natural Approach to Diet, Childbirth, Healing, and the Environment Really Keeps Us Healthier and Happier, from Rodale, which comes out January 29th.