More than a quarter century ago, in “Perfection Salad,” I wrote: “It now seems ludicrous that science should have much to say about cooking… cooking is slowly recovering from ‘domestic science’.” In the 2014 epilogue, I asked: “Has food recovered from domestic science?”
“Domestic science” was rebranded as “nutrition science,” with all the same pathologies. That has yielded zero reliable knowledge.
Despite complete ignorance, nutrition “science” issued and enforced confident recommendations that may have been responsible for millions of premature deaths, plus great loss of health and quality of life.
Meanwhile, industry has developed considerable genuine science of food—oriented to optimizing commercial ends, rather than health and tastiness.
Partial public awareness of these problems has produced a proliferation of pseudoscientific, quasi-religious food subcultures.
Passionate belief in mythical meanings of food probably have a evolutionary origin.
Recently, some pundits have started to suspect that—as I suggested in 1988—no one knows what makes food healthy. Perhaps now the public can begin to resist all claims to authoritative food-knowledge.
There must be a correct diet; there must be a rational way to discover it
There’s no reason to believe either of those; indeed, there’s strong evidence against each of them.1 The power of wistful certainty comes from the unspoken alternative: “otherwise, we would have no control over our health, Science would fail, and the world would be hostile and unfair and we might as well just give in to hopelessnihilism.”
This ploy underlies both obviously silly New Age nutrition pseudoscience and authoritative state-endorsed nutrition pseudoscience.
What makes you call nutrition “pseudoscience?” That seems like a wild claim. It’s true that it has failed repeatedly, but isn’t that the way science works? We can’t demand certainty; science can only say what is most likely based on the available evidence. It’s true that a lot of studies have been done badly, but that doesn’t invalidate the best work; it just means we need to insist on better experimental methods.
There is always uncertainty in science, but real science gradually establishes some stable facts; it eventually strongly supports some theories and conclusively dismisses others. It is typical of pseudoscience that it does not progress.
Nutrition has made no progress. It has discovered no stable facts. Everything nutritionists have said, they have said the opposite ten or twenty years later (if not much sooner). They literally know nothing.2 After a century of countless experiments, the most common, most basic problem they’ve addressed—the optimal ratio of fat, protein, and carbohydrate—is completely unsolved. If they can’t figure that out, anything more sophisticated seems hopeless.
Nutrition is now both scientism and pseudoscience. This is a somewhat rare combination; cognitive science is another example, as I pointed out in “Perfection Salad.” Scientism—the eternalistic distortion of science into an authoritative source of meaning—is most harmful when the science is bogus. Pseudoscience is most harmful when it gets the support of the state and other powerful institutions. Food and theories of the mind probably both strongly affect human well-being, so they are particularly bad subjects to have turned into scientism or pseudoscience.
My point is not that nutrition is bad science. Unquestionably, it is bad science; a competent statistician, looking at the design of most experiments, will immediately say “this is meaningless; you can’t learn anything this way.”
It’s worse than just incompetence, in two ways. First, as the “resignation letter” noted, even the best studies have been useless. There seems to be something fundamentally wrong, such that doing the same sort of science better wouldn’t help.
The second, still worse implication is that worthless pseudoscience can get treated as authoritative for a century, and even now. This is partly due to rationalist eternalism, and partly due to institutional imperatives produced by malign social dynamics.
In which “Science” kills a few million people
Just when I wrote “Perfection Salad,” in the mid–80s, nutrition had its greatest breakthrough. “Scientists” “discovered” that fat (especially saturated fat, and doubly especially cholesterol) was the cause of the two biggest causes of death in rich countries: cardiovascular disease and cancer.3 Cardiovascular disease is caused by fat (especially cholesterol) accumulating in blood vessels. So, obviously, eating less fat will prevent cardiovascular disease. Cancer is caused by oxygen free radicals chemically modifying fat into a form that attacks DNA, creating mutations, so obviously if you eat less fat, that happens less.4 Besides, fat has twice as many calories as protein or carbohydrate, so obviously if you eat less fat, you won’t get unhealthily overweight.
A massive public “education” campaign followed. Perhaps astonishingly, in response, Americans dutifully drastically decreased their fat intake (especially cholesterol). This followed the pattern I described in 1988:
Domestic scientists assumed that the ratios of protein, carbohydrate, and fat were relevant variables, and based diet plans on theories of the “correct” ratios. “Nutrition scientists” are still doing this, and so far as I can tell they still don’t know what the “correct” ratios are, because every five years they confidently declare that they have discovered with complete certainty that we should have twenty-three percent or thirty-seven percent or seventeen percent protein in our diets, and American eating habits obediently shift accordingly. Most likely there are no “correct” ratios, because lots of other variables are involved.
An epidemic of obesity began just around the time this “education” campaign began. Health outcomes have been awful. It seems likely that low-fat diet advice actually caused the diseases it was supposed to prevent.5 In any case, most studies concluded that dietary cholesterol does not increase blood cholesterol and does not cause cardiovascular disease; and the saturated fat evidence varies between weak, zero, and counter to the low-fat theory.
As of early 2015, the establishment is trying to figure out how to retract their anti-fat advice, while doing as little damage as possible to their reputation. They are sending up various wafflytrial balloons, experimenting with PR strategies.
Saving nutrition’s reputation is a matter of self-interest. Also, many in the field are probably just too stupid to realize the magnitude of their failure, and honestly assume that somehow Science must know something.
However, more sophisticated players seem to be thinking:
Admitting outright that we were wrong could discredit nutrition permanently—or even Science as a whole. Even though we know nothing now, with better Science, we’ll probably discover the truth soon. It’s critical to preserve respect until then, so people will listen when we get it right. Otherwise, they’ll fall prey to New Age woo and commercial quack diet faddery.
You can hear this in the background of the waffling. I think it is too late; the public is already losing trust.
There’s another problem: if the advice is not anti-fat, what can it be? Some in the field seem to be trying to establish a new consensus, organizing to make anti-sugar the new message. This would take us back to the 1960s–70s, when sugar was the Big Bad. Maybe it is. Who knows? I’m reasonably sure nutritionists don’t.
Another PR strategy has been to blame wrong dietary recommendations on corruption, due to industrial influence. Probably corruption has, indeed, been a significant factor. However, this is a typical example of the eternalist strategy of explaining away failure as due to extraneous factors, which preserves the illusion of present and future competence.
Why has nutrition science failed? At this point, we can’t know. I believe that all available nutrition research funding should be redirected to answering that question. In the mean time, I’ll speculate:
It may not matter what you eat. For example, Ioannidis has recently argued that reported nutritional effect sizes must be grossly overstated, and diet may not have a significant effect on health after all. On the other hand, the observation that peoples become much less healthy when they start eating Western food does suggest that diet is significant. However, this might be due to simultaneous adoption of some other, as-yet unidentified harmful aspects of the Western lifestyle.
It may be that what makes a healthy diet is so different for different people (due to different genetics and/or lifestyles) that experiments done on mixed populations are meaningless. (I think this is relatively unlikely, for evolutionary reasons, but worth pursuing as a possibility.)
An intriguing possibility is that what you eat matters, but not for you. Until recently, all nutritional research assumed that dietary effects worked via human metabolism. Recent studies suggest that gut bacteria play an important role in human health, and that diet affects them much more than it affects human cells. If this is right, biochemical studies of diet have been looking at irrelevant factors for the past century. (I hope this is right, because it might lead to rapid progress, and also because it’s funny.)
Reverse regulatory capture
Honest nutrition scientists would, as in my satirical “resignation letter,” admit that the field has failed, they know nothing, and they cannot now give any meaningful recommendations. I think this would actually be more likely to preserve public trust, in the long run, than the current attempts at waffling and bluffing and muddling. The field is probably too cowardly for honesty, though. The emperor now realizes he has been seen parading naked, but will pretend not to know, to save face.
Anyway, as the “rejection” reply letter explains, institutional imperatives make it impossible to admit ignorance. There will be nutritional recommendations, even if every nutritionist has to be fired in order to create them. Governments, and the food industry, cannot accept that nothing is known, because they would no longer have any basis for their institutional policies. They do not care much what the policies are; but it is critical that they exist.
Initially “domestic science” captured regulators;7 but then state institutions captured nutritional “science.” Once it was established that there were authoritative facts-of-the-matter about what people should eat, state institutions (schools, prisons, the military) needed stable, simple, crisp guidelines about what they were allowed to feed people. For school administrators, it doesn’t matter what the nutritional theory is, but it is critical that there be an authoritative theory they can demonstrate conformity to, in order to remain blameless. So the power here is mainly in the authority-giving power of rationalistic discourse, not in the institutions (much less individuals).
The processed food revolution
In 1988, most American meals were still cooked from scratch. Now that’s rare. Nearly all American food is the product of intensive industrial engineering systems. These rely on new, genuine food science—about how to reliably extrude optimized food-like products, not about what is healthy or (for the most part) tasty. The capture of food by rationality is therefore essentially complete; but it is rationality optimizing for ends we might not choose.
Since we don’t actually know anything about nutrition, it’s impossible to know whether the new engineered food products are unhealthy. From the food industry’s point of view, uncertainty is good, because nowadays any food can be labelled with multiple supposedly-beneficial qualities, according to assorted competing theories (probably none of which have any relationship to reality).8
The obesity epidemic suggests something has gone badly wrong with the Western diet, in which case it must have something to do with processed food, just because nearly all food is now processed.
One of the trial new messages being tested by the nutrition establishment is “avoid processed food,” which has the big advantage (for their future credibility) that no one is likely to adopt it. Cooking has become an unacceptable hassle.9
Public recognition and resistance
Until recently, public opposition to official food recommendations was mainly religious or “ethical.” The monist counterculture (“New Age”), a quasi-religious movement, has produced a series of opposition diets since the 1970s. Although some of these invoked pseudoscience, and made vague health claims, they were all mainly moralistic. They were anti-scientific and anti-capitalist (as monism typically is). The rise of politically-correct food labelling (“fair trade”) may have been partly in response to increasing public realization of the dubiosity of nutritional claims, but it was mainly explicitly ethical.
An uneasy sense that nutrition recommendations had changed too many times, too quickly, seems to have gradually dawned on the public starting about a decade ago. Up until then, almost everyone simply accepted official pronouncements without question. Early 2000s studies supporting the high-fat Atkins diet seem to have shifted the mood. Intelligent people recognized that nutritional advice is uncertain, and liable to change again soon. So then you have to ask: why bother paying attention to the current guidelines?
There’s been another, dramatic change over the past year (starting late in 2014, I think). Science-savvy members of the commentariat—journalists and bloggers—are finally starting to recognize that there is no there, there: nutrition has no cards to play.10
It’s extraordinary how certain and passionate everyone is about their nutritional beliefs—mainstream or alternative—despite the lack of any basis for them. Religion and politics are the only other domains that commonly inspire such delusional commitment.
Every human culture gives elaborate meanings to food—to hunting, gathering, growing, harvesting, processing, cooking, sharing, and eating it. Every culture has elaborate ideologies of what you should and should not do with food—most of which seem insane to anyone from a different culture. (These constitute the standard example of the eternalist ploy of purity.)
Food is hugely evolutionarily important, so it is not surprising that humans give it such meanings. It’s rather more surprising that something so evolutionarily important should have such divergent meanings attached. Aren’t most of them maladaptive?
A speculation: Perhaps the urge to give foods meanings is a relic of our former hunter-gatherer lifestyle, when keeping track of the edibility, habits, and best use of thousands of species was important. Mythological narratives (“we are forbidden to eat that berry by Flying-Buffalo-Woman, who was tricked with one by Centipede-God”) were valuable as mnemonics encoding cultural knowledge. Often that could be a matter of life and death.
Nowadays, even though the evolutionary purpose is lost, we can’t help making up myths about food, and still feel compelled into believing and enforcing them.
I will discuss the meanings given food again in two later chapters:
The ethics chapter considers the moralization of food. There are legitimate ethical questions, but many claims I find highly dubious: not because they are ethically wrong, but because the issues are not ethical ones at all. I use these as examples of a broader phenomenon: the metastasis of morality into domains where it has no business.
In the history of meaningness chapter, I will describe how the meaning of food has changed over the past few decades, as we’ve moved through the systematic, counter-cultural, subcultural, atomized, and fluid modes of relating to meaningness.
1.There are healthy non-Western populations with diets very different from each other’s. Some of those may be somewhat better than others, but there’s no strong reason to believe so. A century of scientific research has failed to discover any nutritional facts. More and better research might; but we can’t be certain of that a priori.
2.For “literally know nothing,” see for example the recent Ioannidis editorial in the BMJ. There are two exceptions. First, they know you should shouldn’t eat poisons. Arsenic and polychlorinated biphenyls are bad for you. Second, there are some chemicals (vitamins, for instance) that you have to get some of, or else you get a deficiency disease. Neither of these facts are relevant to anyone with a vaguely normal diet.
3.The supposed connection with cardiovascular disease goes back to the work of Ancel Keys in the 1950s. However, avoiding saturated fat and cholesterol for cardiovascular reasons only became the mainstream message around 1980. The supposed cancer connection was new in the mid–80s, and gave further credence and urgency to the anti-fat campaign.
4.This led also to the recommendation that you should eat more antioxidants. That message is still common, although most follow-on studies of specific antioxidants found that they are bad for you.
5.Since nothing is actually known about nutrition, we can’t be sure the low-fat diet caused the obesity epidemic. Correlation is not causation; but it’s quite suggestive in this case.
8.Ioannidis notes that “Almost every single nutrient imaginable has peer reviewed publications associating it with almost any outcome.”
9.Disclosure: I do avoid processed food, and most of what I eat I cook from scratch. This is not advice.
10.The paleo movement has played a major role in this. Paleo is interesting as a subculture that combines romantic rebellion with scientistic rationalism. That potent combination that has made it the most effective anti-authoritarian diet ideology so far. With difficulty, I’ve resisted writing more about that here; this page is already too long.