This relates to the sugar question and was raised by Stephan Guyenet (a year and a half ago) in his not surprisingly negative review of my sugar book. As I’ve been thinking about black swans lately and what evidence would constitute a refutation or a falsification of ideas/hypotheses, and specifically those in my books, it seemed I should address this one as well. So here’s Stephan:
A well-studied Tanzanian hunter-gatherer tribe called the Hadza gets 15 percent of its average year-round calorie intake from honey, plus fruit sugar on top of it. This approximates US sugar intake, yet the Hadza do not exhibit obesity, cardiovascular disease, or any of the other disorders Taubes attributes to sugar (10, 11). In fact, many hunter-gatherer groups relied heavily on honey historically, including the Mbuti of the Congo whose diet was up to 80 percent honey during the rainy season (10). Yet they do not exhibit obesity or insulin resistance (12).
The implication is that this is compelling evidence against the case I make in my book that sugar is the fundamental trigger of the obesity and diabetes epidemics that are washing over the modern world. What follows was originally written for an email exchange I had with Stephan shortly after he wrote his critique. I think it’s worth noodling.
So here’s the question and it’s a purely logical one: Because the hypothesis is that sugar is the factor in a nutrition transition that causes populations to undergo epidemics of obesity and diabetes, can a population or a few of them even that consume significant amounts of sugar but have not manifested obesity and diabetes epidemics serve to falsify it?
Answer: I don’t think so. (Of course not, critics might be thinking here, but let’s go on…) In short, when a population or a person fails to get fat or diabetic on a sugar-rich diet that does not necessarily mean that sugar isn’t the cause for those people or populations who do get fat and diabetic. Assuming I’m right about this sugar-ibesity/diabetes connection, this would clearly have to be the case for individuals, at least. It’s manifestly obvious that a significant (albeit perhaps shrinking) proportion of people can eat and drink sugar to something close to their heart’s content and stay lean and apparently healthy. Just as a significant proportion of the folks who smoke never get lung cancer. That doesn’t mean that cigarette smoking doesn’t cause lung cancer. Only that it doesn’t cause it in everyone.
As I see it, meaningful evidence against the hypothesis in the sugar book — a black swan — would be a population that does experience an obesity and diabetes epidemic, and yet consumes either very little sugar or has not recently upped its sugar consumption considerably. Let’s go back to the cigarette and lung cancer analogy: the black swan in that hypothesis would be a population with a lung cancer epidemic that doesn’t smoke cigarettes (or work in asbestos factories). Such a population would force us to ask what other factors in the environment can cause lung cancer epidemics and so whether those factors are also operating here, which could imply that cigarettes are, at the very least, not the only cause. And it could be the case that this other cause is more fundamental and that if that factor didn’t exist in our environment, too, cigarettes would not be carcinogenic. Maybe cigarettes accelerate the lung cancer problem, but the other factor is the necessary-and-sufficient one. All these scenarios would now be possible.
Now, I like to think with analogies because they allow us to simplify situations. So let’s say we find a population that smokes tobacco, but is lung cancer free — say the Native Americans and their tobacco smoking in the 19th century. If they didn’t get lung cancer, would this falsify the cigarette-lung cancer hypothesis? In this analogy, the Native American smoking habits are analogous to the honey that the Hadza eat, and the lung cancer epidemic to the obesity and diabetes epidemics.
I’d say no, it doesn’t falsify, because we’d legitimately have to wonder 1) whether there are differences between how the Native Americans smoke and what the Native Americans smoke and how we smoke and modern cigarettes. And this would be equivalent to wondering whether there are differences between how the Hadza consume honey and whether there are meaningful differences between honey and sugar (and high fructose corn syrup) that might be relevant to their effect on insulin resistant, obesity and diabetes. For instance, we consume sugar all day long and drink it all day long and all year long and so do our children, and maybe the honey consumed by the Hadza is only eaten (not consumed in beverages), and eaten only seasonally, and maybe mainly by adults (although having kids myself, it’s for hard for me to believe that Hadza adults would be any more successful keeping sweets away from their children than I am keeping sweets away from mine). And maybe honey, and the fact that the Hadza wouldn’t drink it (lacking hot tea to put it in) has different physiological effects than sugar, which is and can be consumed in liquid form and so can be digested far more quickly and hypothetically do far more harm. Maybe honey, because of the viscosity, is relatively harmless compared to sugar. Anything is possible. That would be nice to know for those of us who miss sweets and would like to find at least one sweetener that’s natural and harmless.
Lastly, back to tobacco and lung cancer, maybe because the Native Americans have been smoking tobacco for hundreds of years or longer, we might be dealing with a population that has evolved to tolerate the carcinogens in the tobacco they’re smoking (or maybe the tobacco they’re smoking in the form they’re smoking it is not carcinogenic). And, of course, the analogy is we don’t know how long the Hadza have been consuming honey. We know it is not part of a nutrition transition of the kind on which we blame obesity and diabetes epidemics.
This last point is crucial. In discussing evolution in GCBC and The Case Against Sugar, I quote the British Naval researcher Peter Cleave as a way to put it in historical context. Here’s the relevant paragraph from TCAS:
Cleave invoked what he called the “Law of Adaptation,” based on his reading of Darwin, to explain the epidemics of chronic disease that Campbell and others were beginning to document around the world: species require “an adequate period of time for adaptation to take place to any unnatural (i.e., new) feature in the environment, so that any danger in the feature should be assessed by how long it has been there.” To Cleave, the refining of sugar and white flour and the dramatic increase in their consumption since the mid-nineteenth century were the most significant changes in human nutrition since the introduction of agriculture roughly ten thousand years before. “Such processes,” he wrote about the refining of sugar and wheat, “have been in existence little more than a century for the ordinary man and from an evolutionary point of view this counts as nothing at all.”
So if the honey isn’t “an unnatural (i.e., new) feature in the environment,” maybe they’re adapted to it in a way that non-honey consumers would not be adapted to sugar or the amount of honey the Hadza consumed. Maybe the Hadza had their obesity and diabetes epidemics when it was new? We don’t know. What would be interesting is to find out whether members of the Hadza people migrated into urban centers and experienced the classic nutrition transition, now consuming sugar and sugary beverages, rather than eating honey. Do they get fat and diabetic? (Although even if they do, we wouldn’t know whether it was the sugar or some other aspect of their personal nutrition/lifestyle transition.)
So I acknowledge the Hadza and their honey consumption are fascinating but I don’t think these populations falsify the sugar-ob/db epidemic hypothesis, They don’t satisfy the requirements necessary to do so. No nutrition transition, no ob/db epidemics. However, that said, had we many such populations, well-documented, they would surely make me wonder what else is going on.
Egregious (and embarrassing) error correction: In earlier versions of this post, I spelled both the Hadza name incorrectly and Stephan Guyanet’s. My apologies to both.