I’m going to start this long-overdue series of posts with a bit of a shaggy dog story, a lengthy preamble (“amble” perhaps being the operative word) before I get to the meatier issues.
One of my supporters in mainstream medical research is Allan Sniderman, a professor of cardiology and medicine at McGill University in Montreal. Since the mid-1980s, Sniderman has been arguing that Apo-B (the protein component of low and very low density lipoproteins) is a far better predictor of heart disease, which it surely is, than the cholesterol that happens to be contained in these lipoproteins. He’s also a co-discoverer of the hormone ASP — acylation stimulating protein — which plays a role, however controversial, in fat storage. Sniderman read Good Calories, Bad Calories shortly after it was published in September 2007 and then invited me up to lecture at McGill. He later described his feelings about GC,BC this way:
I had spent some years studying adipose tissue metabolism but it is fair to say I learned more from [Taubes’s] book than I had from my experiments. He restored a sense of how our ideas about obesity and vascular disease had developed and how a number of them had gone off the track. I did not agree with everything he wrote but I did learn a huge amount and much of what I learned is now core to my thinking about the relations of obesity and metabolic disease.
In May 2008, I decided to take a more proactive approach to motivating obesity researchers to test the critical arguments in GC,BC. Step one was to induce them to read the book, or at least get the arguments on their radar screen. One way to do that, I figured, was to give seminars at institutions that had the requisite capabilities, expertise and experience to do the experiments. I had been lecturing at medical schools and in nutrition departments and even the National Institutes of Health – twice by that time, and once since – but other than the NIH, which had a new metabolic ward facility, none of these institutions had the resources necessary to do the experiments. (And the NIH people told me that they would be happy to do the experiments, if I raised the money from outside sources, something I am working on at the moment.)
One institution that did was the Pennington Biomedical Research Institute in Baton Rouge Louisiana. The PBRC is probably the most influential academic obesity research center in the U.S. if not the world. It has a large metabolic ward facility that can house volunteers for the requisite weeks to months, and it has the equipment to measure body composition, energy expenditure, substrate oxidation, and anything else I could imagine being useful. Its director is Claude Bouchard, who like Sniderman, spent his research career in Quebec – in Bouchard’s case, at Laval University. I guessed that Sniderman probably knew him well, which he did, and so I asked him to put a word in for me, which he did. Just a few days later I received an e-mail from Bouchard. “I had a nice conversation recently with Allan Sniderman,” Bouchard wrote, “who suggested that we should have you at the Pennington Biomedical Research Center for a seminar and perhaps a series of meetings with our scientists interested in obesity, diabetes and metabolism.”
I gave the seminar in January 2009 (after having it postponed once by Hurricane Gustav) and it led to two moments that captured perfectly the challenge of what I’ve been trying to accomplish for the last four years. I’ll discuss the first in this post, and the second will come at the end of this series, when I get around to discussing experimental tests of the competing hypotheses.
The PBRC auditorium was packed — standing-room-only — and I presented a slightly more technical version of the lecture I’ve given frequently – “Why We Get Fat: Adiposity 101 and the Alternative Hypothesis of Obesity.” (For those, who haven’t seen it, you can find a recent version here, from last April at the Ohio State University Medical Center.)
I argued in the lecture that obesity research had made little to no progress in the years since the Second World War (and for anyone who thinks progress has been made see “obesity epidemic” as a counter-argument) and the reason is because the researchers had been laboring under the wrong paradigm. And by paradigm, I didn’t (and don’t) mean how the word is often bastardized nowadays to imply virtually any shift in thinking or technology, no matter how minor. I meant what the philosopher of science Thomas Kuhn defined as “scientific achievements that… provide model problems and solutions for a community of researchers.” By a paradigm shift, Kuhn meant, for example, the shift from Ptolemaic astronomy, in which the sun circles the Earth, to Copernican astronomy, in which the opposite happens, not the replacement, say, of traditional standard-definition televisions with HD.
At the PBRC, as in my books, I argued that obesity researchers had come to universally conceive of obesity as a disorder of energy balance – we get fat because we consume more calories than we expend — when they should have been thinking of it as a disorder of excess fat accumulation as the pre-WW2 Europeans had come to do. Not an issue of overeating or calories in being greater than calories out, but one of how the calories were partitioned in the body – stored as fat or muscle, oxidized for fuel, etc.
If the post-WW2 generation of researchers had simply defined obesity as a disorder of excess fat accumulation rather than one of energy balance, I argued, they would have naturally asked the question, what hormones and enzymes and other factors regulate fat accumulation. And that’s what they would have and should have been studying for the past sixty years. But, with precious few exceptions, that’s precisely what they have not been doing.
Instead, they’ve focused on what factors control eating behavior and physical activity, and they’ve considered the actual regulation of the fat tissue and fatty acid metabolism irrelevant. (As I describe in GC,BC, a handful of influential researchers in the 1970s worked diligently to achieve this state of affairs, removing any discussion of adipose tissue regulation from discussions of obesity itself, largely because they didn’t like the implications.) At the end of my lecture, I proposed the experiment that I thought the PBRC researchers could do, and I explained why this experiment would serve the critical purpose of differentiating between the two paradigms, and why it was precisely the type of experiment they should be doing. (And this is what we’ll discuss at the end of this series of posts, along with iconic moment number two.)
Now for the punch line to my shaggy dog story – i.e. iconic moment number one. In the Q&A session following my hour-long presentation, a member of the PBRC faculty, a distinguished-looking gentleman who I’d guess was in his mid to late sixties, raised his hand and said, “Mr. Taubes, is it fair to say that one subtext of your talk is that you think we are all idiots?”
Is it fair to say that I think they are all idiots? A surprisingly good question.
Certainly one subtext of my talk (and my work) is that a journalist is getting it right and sixty-odd years of nutritionists and obesity researchers got it wrong (with maybe a half dozen exceptions who were marginalized for their beliefs.) So, yes, it was fair to say that I think a large body of otherwise very smart people, Ph.D.s and M.D.s all, were operating with suboptimal intelligence. Certainly in a pursuit — science — in which the one goal is to get the right answer, getting the wrong answer on such a huge and tragic scale borders on inexcusable.
That isn’t, of course, how I responded at the moment. I smiled, and I said, no, what I believed was that researchers of his generation – those who would have started their careers in the 1970s – had inherited a paradigm of obesity from the generation that preceded them. And this paradigm seemed so obvious (we get fat because we take in more calories than we expend) that they never thought to question it. Indeed, I didn’t think to question it myself until 2003 or 2004 when my research took me to the pre-WW2 European ideas about obesity (for which I owe the late Alfred Pennington – no relationship to the Pennington of the PBRC – for showing me the way) and to researchers in the U.S. who were studying fat accumulation in animals for reasons other than necessarily understanding human obesity.
But this was still just a kind way of saying that researchers of this fellow’s generation, and all those who have followed had simply missed the point. They hadn’t done their job. They should have questioned the obvious, even if it was obvious. And had they done so, all the bad science that followed might never have happened, and maybe we wouldn’t be having an obesity epidemic and a diabetes epidemic along with it.
So why didn’t they do it themselves?
Well, one obvious reason is maybe it’s wrong. They thought about it, looked into it, rejected it. That’s what Stephan Guyenet recently argued on his Whole Health Source blog, and we’ll address that directly later on this series of posts — part III, as it looks at the moment. (It was Stephan’s post, not surprisingly, that finally motivated me to find the time to blog again and put aside, momentarily, my other obligations — raising a family, participating in my marriage, earning a living, and trying to influence the 99 percent or so of the medical and public health establishments that aren’t reading any of these blogs. For that, I’m occasionally grateful.)
Assuming for the moment that it’s not wrong – that I’m not wrong or at least not completely wrong – the other obvious explanation is that they didn’t do it because they were living inside the energy balance paradigm and couldn’t see beyond it. Everything they did, all their conversations, every research question they asked and even the funding they received to answer those questions, their ability to move up in the hierarchy of their field, to succeed, in a word – to become an assistant professor and then a professor, to edit journals, and serve on prestigious committees, to thrive, support a family, pay their lab techs, etc. – all existed within the same belief system. And so they had little to no reason to see outside it and little motivation to overturn it. Seeing a reason to challenge the existing paradigm was not only exceedingly difficult from within the paradigm, but following through on this challenge could seriously jeopardize an individual’s ability to succeed — even to be taken seriously from day to day.
This is why, as Kuhn explained in The Structure of Scientific Revolutions, his seminal thesis on paradigm shifts, the people who invariably do manage to shift scientific paradigms are “either very young or very new to the field whose paradigm they change… for obviously these are the men [or women, of course] who, being little committed by prior practice to the traditional rules of normal science, are particularly likely to see that those rules no longer define a playable game and to conceive another set that can replace them.”
So when a shift does happen, it’s almost invariably the case that an outsider or a newcomer, at least, is going to be the one who pulls it off. This is one thing that makes this endeavor of figuring out who’s right or what’s right such a tricky one. Insiders are highly unlikely to shift a paradigm and history tells us they won’t do it. And if outsiders or newcomers take on the task, they not only suffer from the charge that they lack credentials and so credibility, but their work de facto implies that they know something that the insiders don’t – hence, the idiocy implication.
This is why a common and understandable response to any challenge to the existing paradigm – to the conventional wisdom, in effect – from an outsider is this: “who the hell are you (or am I) to be questioning us? You’re not a member of the priesthood. Not an upper wizard of the stratosphere. You haven’t trained in the field. You haven’t proven yourself. You haven’t done, in effect, what we have done; you haven’t learned what we have learned. You didn’t have the necessary apprenticeship in the relevant arts. Bug off!” (Although, this is by no means a universal response, as this paper – “Obesity and Energy Balance – Is the Tail Wagging the Dog?” — published in July in the European Journal of Clinical Nutrition demonstrates, taking my ideas and those of Robert Lustig’s and exploring the implications.)
This knee-jerk rejection is indeed a valid response, because most outsiders who challenge the conventional wisdom are dead wrong. Some huge majority are quacks – assuredly greater than 99 percent and we can probably add a few more 9s to this percentage and still be on the safe side.
When I wrote for Discover magazine in the 1980s, one of my beats was high energy particle physics, which was also the subject of my first book. Every time I wrote a story on the new theory or elementary particle de jour, I would receive letters written in crayon (implying, as it was explained to me at the time, that the letters had been written by prison inmates who weren’t allowed to use sharp objects like pencils and pens for the purpose). These letters would typically explain why Einstein or Dirac or the latest generation of theoretical physicists had been wrong, and they would propose a new theory of quantum mechanics or relativity or just a theory of everything – known in the jargon as a T.O.E. — that the authors invariably had absolute confidence was correct. And, for all I know, one or more of these incarcerated amateur physicists might have been dead on. Maybe they had created a working theory of everything, but into the waste paper basket these letters went. The odds against them being right were astronomical and time is short. And the odds wouldn’t have been significantly better had the letters been written by fellow journalists or even science journalists using IBM Selectric typewriters or even the few desktop computers that had begun slipping into the offices of the era.
A good rule of thumb is that outsiders challenging establishment science are invariably wrong. And we don’t want our experts and authorities wasting their time vetting every last crackpot theory that arrives over the transom. Crayon or not. We have better uses for their time.
And here’s the challenge to both the scientist working in the field and the lay observer following along: how do we tell the difference between the one in a million times, say, that an outsider comes along and gets it right, and the other 999,999 quack-driven attempts? The numbers alone tell us that the best idea is always to bet against the outsider, that we’re always best served by ignoring him or her and getting back to science as usual (what Kuhn called “normal science”). The odds are enormously in our favor if we do so. But, still, when a paradigm is shifted, it’s going to be an outsider who does it, so keeping an open mind is a reasonably good idea, particularly when the evidence suggests such a shift is in order (see aforementioned obesity epidemic).
This leads to a second major problem with making these assessments – who’s right or what’s right. As Kuhn explained, shifting a paradigm includes not just providing a solution to the outstanding problems in the field, but a rethinking of the questions that are asked, the observations that are considered and how those observations are interpreted, and even the technologies that are used to answer the questions. In fact, often the problems that the new paradigm solves, the questions it answers, are not the problems and the questions that practitioners living in the old paradigm would have recognized as useful.
“Paradigms provide scientists not only with a map but also with some of the direction essential for map-making,” wrote Kuhn. “In learning a paradigm the scientist acquires theory, methods, and standards together, usually in an inextricable mixture. Therefore, when paradigms change, there are usually significant shifts in the criteria determining the legitimacy both of problems and of proposed solutions.”
As a result, Kuhn said, researchers on different sides of conflicting paradigms can barely discuss their differences in any meaningful way: “They will inevitably talk through each other when debating the relative merits of their respective paradigms. In the partially circular arguments that regularly result, each paradigm will be shown to satisfy more or less the criteria that it dictates for itself and to fall short of a few of those dictated by its opponent.”
So this can be considered a warning. I’m about to launch into a discussion of two hypotheses of obesity that exist in competing paradigms – the food reward/palatability hypothesis that Stephan Guyenet has revived, which lives firmly within the energy balance paradigm (calories-in>calories-out) and the carbohydrate/insulin hypothesis, which I’ve been pushing and which lives in the fat accumulation disorder/fuel partitioning paradigm.
In explaining my problems with food reward and palatability as a viable hypothesis of obesity, I’m going to repeat many of the arguments I made in my books for why the energy balance paradigm itself seems to be such a failure. (Not all of them because life is short, but many.) And these, of course, will also provide the rationale for why something like the carbohydrate/insulin hypothesis is necessary. It doesn’t imply that the carbohydrate/insulin hypothesis is right, but that something very much like it almost assuredly is. I’ll also explain why I find many of the observations and some of the experiments used to support the hypothesis meaningless and inconsequential. I hope, as I did with my books, to create what Kuhn called a “playable game.”
But here’s another catch: This map-making exercise can be perceived as a justification for cherry-picking of the data, which, in a way, it is. But I’m arguing that such selective interpretation of the data is a fundamental requirement to make progress in any field of science, and particularly one as off the rails as that of obesity and nutrition. It is inherent to the process that Kuhn described as “map-making,” to taking a non-playable game – a dysfunctional paradigm – and making it playable.
This was a point the physicist Richard Feynman made indirectly back in 1965 in The Character of Physical Law, the book version of a series of lectures he gave the year before at Cornell University. (The lectures themselves are available on line and are worth viewing for many reasons, one of which is the experience of listening to one of the great thinkers of the 20th century express himself in a thick New Yawk/Queens accent.) Feynman was talking about how physicists find a new law of nature, and this is what he said:
In general we look for a new law by the following process. First we guess it. Then we compute the consequences of the guess to see what would be implied if this law that we guessed is right. Then we compare the result of the computation to nature, with experiment or experience, compare it directly with observation, to see if it works. If it disagrees with experiment it is wrong. In that simple statement is the key to science. It does not make any difference how beautiful your guess is. It does not make any difference how smart you are, who made the guess, or what his name is — if it disagrees with experiment it is wrong. That is all there is to it.
But then he added the caveat:
It is true that one has to check a little to make sure that it is wrong, because whoever did the experiment may have reported incorrectly, or there may have been some feature in the experiment that was not noticed, some dirt or something; or the man who computed the consequences, even though it may have been the one who made the guesses, could have made some mistake in the analysis. These are obvious remarks, so when I say if it disagrees with experiment it is wrong, I mean after the experiment has been checked, the calculations have been checked, and the thing has been rubbed back and forth a few times to make sure that the consequences are logical consequences from the guess, and that in fact it disagrees with a very carefully checked experiment.
And this is the point. Experimental results and observations have to be rubbed back and forth a few times to see if the interpretations that first come to mind are really justified, and whether the experiment, for that fact, is a “very carefully checked” experiment. And what we want to know is whether the result really disagrees or agrees with the predictions. Or is something else going on? Not just dirt in the equipment, but maybe another interpretation entirely – an alternative hypothesis? What was missed in the interpretation? Artifacts in the experimental apparatus? Confounding factors that might explain the observational evidence?
Asking these questions, indeed, leads to all kinds of cherry picking of the data, what a Scottish physician once described to me as “Bing Crosby Epidemiology” – i.e., accentuate the positive, eliminate the negative. And the paradigm in which we live, not surprisingly, will determine how we define positive and negative and so what we accentuate and what we eliminate. Depending on our paradigm or our preferred hypotheses, we’ll put more or less effort into the rubbing back and forth process based on whether the experimental results agree with our notions or don’t.
As I’ve said before in various venues, at one time in the writing of Good Calories, Bad Calories I had a 400,000 word unfinished draft. I couldn’t complete it because it was obviously far too long already – twice as long as it should be — and yet I had important chapters yet to write. I solved the problem by giving it to my editor to read with the suggestion that maybe we could make it two books. He read it in its entirety (one of many acts of editorship that earned my undying devotion) and said, no, one book. We proceeded to cut the document by more than half, so I could then write the chapters that still had to be written and end up with a book that was under 200,000 words (bibliography and endnotes, not included).
Much of what was removed was the rubbing back and forth. I would present an observation – high levels of insulin, for instance, in obese subjects first observed in the early 1960s – and then I would explain how it was interpreted to support the conventional wisdom (we get fat because we overeat and being fat then causes insulin resistance and so increases insulin levels) and why that wasn’t necessarily the correct interpretation and how the same observation supported alternative hypotheses as well. And I would go back and forth with arguments and counterarguments.
My editor pointed out that this wasn’t necessary; that my job was to present my interpretation of the evidence and if someone wanted to challenge it later, so be it. I could provide the arguments and counterarguments, the rubbing back and forth, then.
What I always found amusing once the book was published (okay, amusing in an irritating way) were the critics who would first complain that GC,BC was too long – I go “on and on about experiments old and new,” as Gina Kolata put it in the New York Times – and then upbraid me for leaving something out that they considered important. And so when Kolata pointed out that “definitive” experiments by Leibel and Hirsch should have been in my book because they refuted my arguments – thus accusing me, in effect, of the supposedly heinous crime of cherry picking — I was left to point out in a letter to the editor that the experiment (no “s” at the end, as Kolata had it) was poorly done, didn’t address the salient issues, that Kolata got many of her facts wrong, and that her use of the word “definitive” left much to be desired and that “ambiguous” was a far more accurate description.
So Kolata read the Leibel/Hirsch experiment in a way that supported her beliefs and didn’t bother to rub them back and forth. (She had just published a book a few months earlier that adhered closely to the conventional wisdom.) And I did, because of the implication that the experiment refuted my arguments. I had to see if it did indeed do what Kolata claimed and concluded (not surprisingly, considering my bias) that it didn’t. Or at least that it couldn’t be used, as she had used it, to refute my arguments.
This selective interpretation of the evidence is human nature, as Francis Bacon pointed out almost 400 years ago. But it’s a necessary part of science. For a paradigm to shift, a significant proportion of experimental results will have to be reinterpreted – meaning the interpretation in the new paradigm and the significance is going to be different than it had been under the old. Some significant portion of experiment results will be deemed irrelevant, on the basis that they don’t shed meaningful light on the subject. And, of course, how meaningful is defined is dependent on the paradigm.
So we’re back to the tricky business of assessing who or what is right in such a situation – in determining where to place our bets?
The ultimate determination should indeed be based on data, but not just any data or any experiment that seems relevant. A controversy would not exist if it were not possible for most experimental results and most observations to be consistent with both hypotheses, both paradigms. The key to making progress is to identify observations in nature or generate them by experiment that are consistent with the predictions of only one of the competing paradigms or hypotheses, not both — or not all, if there are more than two. (Thus invariably prompting proponents of the unsuccessful paradigms/hypotheses to evoke what philosophers and historians of science would call “epicycles” to rationalize away the negative evidence.) The problem with the Hirsch/Leibel experiments, as I pointed out in my letter to the Times, is that the results were consistent with both hypotheses, and so the solution was not to conclude on the basis of a popularity contest which was right, but to advocate for better experiments.
What we ultimately want, as Feynman suggested, is an experiment or an observation that can unambiguously — i.e., rubbing back and forth gets us as close to nowhere as we can get — differentiate between hypotheses or paradigms. The competing hypotheses/paradigms predict different results and only one of the predictions holds up. Meaningful experimental results or meaningful observations are those that refute one hypothesis but not the other. Anything less doesn’t help us and doesn’t answer the question of what or who is right. So a constant reminder in this business is to ask ourselves whether the observations or experimental results we’re discussing serve this purpose: can they differentiate between the two hypotheses? If they can’t, let’s move on and find (or fund) ones that can.
In the next post, I’ll begin by defining the questions I think have to be answered by any viable hypothesis of obesity, and how this is relevant to the salient issue of food reward and palatability vs. insulin signaling and carbohydrates. And these questions will speak directly to observations and experiments that I believe can be used to establish the validity of one (mine, of course) and not the other. It may take me awhile to finish the series, as the aforesaid obligations are bound to crash back in, so please bear with me and stay tuned.