Contrarians, Crackpots, and Consensus
I.
Last week we discussed whether Gary Taubes gets to be admitted to the small but prestigious pantheon of correct contrarians. And the strange part was that there was a lot less argument about how correct he was than about how contrarian he was.
Taubes’ main theory – that low-carb diets could solve the obesity epidemic – hasn’t fared the test of time very well. But some of his supporting points have. Large parts of mainstream nutrition science have eased up on dietary cholesterol, dropped the recommendation against fat, gotten tougher on sugar, and accepted that the science should focus on how to regulate complex satiety mechanisms rather than just counting calories. Given how hard it is to fight the scientific consensus and win, even those few minor victories would potentially be remarkable.
The counterargument is that these are other people’s ideas and he gets no credit for them. Suppose David Icke says that the Queen is a lizard person, and also that the royal family is secretly descended from German nobles and isn’t British at all. A hundred years from now his readers celebrate his genius: although he got the lizard part wrong, the Saxe-Coburg-Gotha theory was remarkably prescient!
If Icke’s book spends just as much time arguing for the Saxe-Coburg-Gotha theory as for the lizard theory, all while inveighing against some supposed consensus of anti-Saxe-Coburg experts, we can imagine the far future finding him pretty impressive, since they might not have done the historical work necessary to realize everyone knew of the Queen’s foreign origins all along. Even though the Queen’s German descent sounds shocking, and even though the average British person probably isn’t aware of it, and even though it’s something nobody really likes to talk about – doesn’t mean Icke “discovered” it in any interesting way. He’s not prescient, he just sometimes reads Wikipedia in between his bizarre ravings.
Thing is, even though the 1990s were like twenty years ago and pretty well-documented, people have had a surprisingly hard time coming to agreement on how novel Taubes’ ideas were back then.
They certainly weren’t perfectly novel – Taubes himself tries to claim he’s just relaying ideas from scientists and researchers to the public, and even his most controversial theories come from other people like Dr. Atkins. And they certainly weren’t perfectly well-known – everyone has a story of their doctor or dietician or something telling them “just eat low-fat foods, cut back on cholesterol, and count calories”. But there seems to be a lot of room in between those two poles.
This is starting to remind me of another debate I got stuck in recently – my argument with Rob Wipond about the serotonergic theory of depression. Wipond argued that psychiatrists irresponsibly promoted a narrative in which depression was a simple serotonin deficiency and so taking Prozac would quickly and elegantly solve the problem. I told him that actually, no, the psychiatric community wasn’t saying that at all, which was why every single example he thought he could find of that turned out to be a garbled out-of-context quote which when investigated honestly was clearly saying the opposite. I got a lot of angry comments that no, people were very sure their doctor had told them that depression was a simple serotonin deficiency.
I think a lot of things are getting obscured by the term “scientific establishment” or “scientific consensus”. Imagine a pyramid with the following levels from top to bottom:
FIRST, specialist researchers in a field. So for example the people doing studies on the effect of dietary cholesterol, or the people dissecting monkey brains to see how much serotonin is in them. These people always have the latest cutting-edge experimental results and a good knowledge of the issues involved in the field.
SECOND, non-specialist researchers in a broader field. Nutrition scientists in general. The guy who is interested in Vitamin B, but goes to the same conferences as the guys studying cholesterol. The research psychiatrist working on schizophrenia, but who maintains a keen interest in what her colleagues over in the depression lab are doing. They know enough about the broad principles of the field to be able to understand and evaluate new ideas more quickly than everybody else, but they still only learn about them the same way everyone else does – by waiting for the specialist researchers to tell them.
THIRD, the organs and administrators of a field who help set guidelines. The head of the USDA who’s in charge of looking over the Food Pyramid to make sure it’s accurate. The APA Committee for deciding exactly what wording to use in the guidelines on depression treatment. The head of Harvard Medical School who has to decide what to put in the curriculum. The editor of the New England Journal of Medicine, who has to decide what gets published.
FOURTH, science journalism, meaning everyone from the science reporters at the New York Times to the guys writing books with titles like The Antidepressant Wars to random bloggers.
ALSO FOURTH IN A DIFFERENT COLUMN OF THE PYRAMID BECAUSE THIS IS A HYBRID GREEK PYRAMID THAT HAS COLUMNS, “fieldworkers”, aka the professionals we charge with putting the research into practice. In nutrition this is doctors and dieticians, who directly inform their patients what to eat. In education research this could be teachers and principals who directly decide how classes will get taught. In sociology it might be the police chief trying to institute a new crime-fighting program. Et cetera.
FIFTH, the general public.
A lot of these issues make a lot more sense in terms of different theories going on at the same time on different levels of the pyramid. I get the impression that in the 1990s, the specialist researchers, the non-specialist researchers, and the organs and administrators were all pretty responsible about saying that the serotonin theory was just a theory and only represented one facet of the multifaceted disease of depression. Science journalists and prescribing psychiatrists were less responsible about this, and so the general public may well have ended up with an inaccurate picture.
Likewise, when Taubes published his book, the ideas he wrote about (at least the correct ones) seem to have been accepted by some specialist researchers, known only as vague inklings among non-specialist researchers, poorly reflected at all in the official dietary guidelines, totally new to the world of journalism, totally new to doctors (who mostly still haven’t gotten the message), and totally new to the general public.
This whole process gets even more complicated when you consider enemy action. In psychiatry, drug companies have established defensive chokeholds at various points on the pyramid, trying to promote pro-pharmaceutical results and sink anti-pharmaceutical ones. This isn’t a far-out conspiracy theory – practically every psychiatrist agrees it’s true to some degree, which is why there are so many conflict-of-interest laws to try to minimize the damage. The only debate is whether we’ve successfully contained it to a small effect, or whether it’s hopelessly contaminated the entire process (I tend to lean more toward the optimistic side; for a true pessimist, read Dr. Nardo). The same is true in nutrition, where a lot of studies are sponsored by groups with names like ‘The United Dairy Farmers Council’ or ‘The League For Wheat’. Even when there aren’t official lobbyists, political opinion plays a big part: the social science journals are full of studies that very competently show that certain politically popular ideas are bunk; by the time they reach the ears of voters and policymakers this has mysteriously been transformed into “scientists agree with you that these politically popular ideas need much more funding”. When you have a block in the process like this, the specialist researchers, the non-specialists, the guideline-makers, the fieldworkers, and the public can all remain on totally different pages for a surprisingly long time.
Taubes – and some of the people making the most noise about the serotonin theory of depression – seem to be people trying to transfer knowledge from the highest levels of the pyramid all the way down to the base, skipping over the levels in between. Does that make them contrarians playing Galileo to a hidebound establishment, or responsible science journalists relaying the establishment’s ideas more faithfully and efficiently than their predecessors? Your opinion probably depends on what narrative suits your purpose at any given time.
II.
I think of some of the contrarians who seem to have their heads screwed on straight. Irving Kirsch and Robert Whitaker
I don’t agree with all these people, I’ve even written long rants against some of them. But they seem to be of a different breed than crackpots like creationists and parapsychologists and anti-vaxxers. It’s hard to specify how. It’s not just credentialed expertise. Michael Behe and Daryl Bem are both professors, and Andrew Wakefield was an MD who’d done previous published immunological research, but their work falls squarely in the ‘crackpot’ column.
But one thing I do notice about these virtuous contrarians – their reception is surprisingly quiet. We know creationism is wrong partly because half the evolutionary biologists in the world have written books about why creationism is wrong, which they advertise prominently on their blogs about why creationism is wrong. Where are all the developmental psychologists shouting down Judith Rich Harris? I’ve seen a few very specialized psychiatrists argue against Kirsch, but never very heatedly, and usually while granting many of his points. The majority of the profession? Never heard of him and don’t care.
When I first became interested in AI risk around 2007, people told me that no legitimate AI experts were seriously worried. I checked and at the time that was mostly true. On the other hand, no legitimate AI experts were specifically not worried either. AI risk just wasn’t their area, and they were perfectly happy to ignore it and concentrate on things that were. There are two types of “no evidence”, and this was the entirely neutral one. It seemed like a very different situation than vaccines causing autism. There, too, experts in the field aren’t worried – but they’re not worried because every single one of them has an opinion and the opinion is “NO”.
(Sure enough, since then a lot more AI researchers have become interested, in exactly the sort of sea change I don’t expect to see mirrored in the autism field.)
The crackpots seem to be met with violent opposition. The virtuous contrarians seem to be met with – well, almost boredom. No one is particularly interested in adopting their ideas, but no one is particularly interested in arguing against them either.
(on the other hand, Time Cube Guy is met with boredom by serious astronomers, either because he’s too small to be noticed or too small to be worth refuting, so it’s not like it’s a great heuristic)
A while ago, I was reading some stuff about the role of choline in the brain, and I thought: “I wonder if anyone has ever used this to treat bipolar disorder”. Well, I searched the literature, and there was one very small study from 1996 in which choline apparently demonstrated excellent effect treating rapid-cycling bipolar disorder, which is otherwise quite difficult to treat. The study isn’t obscure – it seems to have been cited 110 times – but no one’s followed up on that and you could easily go your whole life studying psychiatry without running into any kind of choline-bipolar connection. It seems like a potentially important idea, which has small but nonzero evidence behind it, but which everyone nevertheless ignores, because it isn’t anybody’s business in particular. There are hundreds of things like this scattered across the literature in pretty much every field.
If I were to announce that small-minded scientists were ignoring the result of their own research and covering up The Truth About Choline, possibly at the behest of lizard-people…well, I could certainly do it in a crackpottish way if I wanted to. But I’d be interesting. I wouldn’t be trivially wrong in the same sense as the homeopath who doesn’t care that all biologists disagree with them.
Thomas Kuhn categorized scientific progress into everyday advances and “paradigm shifts”, the latter being major reconceptualizations like the one between geocentrism and heliocentrism, or from classical mechanics to quantum mechanics.
If I understand right – everyone is doing science, and occasionally they come up with something that doesn’t make sense. Whatever. Half the time people come up with things that don’t make sense, and it usually just means your neutrino speedometer is miscalibrated. They either refuse to publish, because no point in publishing nonsense, or they publish, everyone says “Huh, that’s funny”, and they continue doing what they’re doing. The view of science presented to students in the field, and the one that the luminaries of the field think in most of the time, is the one made up of all of the nice consistent results that make sense, with the noise abstracted away.
Then somebody looks a little closer and sees a pattern in the noise. The studies and ideas everyone else was ignoring actually tell a consistent story which is more plausible than the grand narrative of the field which everyone else is working off of. They propose a new paradigm. There is some fighting and eventually it is determined to be superior to the old one, which is jettisoned in its favor.
Someone does a study on Tibetans and says “it looks like they’re adapted to their mountain environment, but that would require really fast evolution, which we all know practically never happens, so it’s probably some weird fluke.” Someone does a study on Indo-Europeans and says “it looks like they have unique lactose tolerance, but that would require really fast evolution, which we all know practically never happens, so it’s probably some weird fluke.” Someone does a study on Ashkenazi Jews and says “it looks like they have higher-than-average intelligence, but that would require really fast evolution, which we all know practically never happens, so it’s probably some weird fluke.” Then Cochran and Harpending and a few others take a sweeping view of everything, and say “OR WHAT IF REALLY FAST EVOLUTION IS HAPPENING PRETTY MUCH ALL THE TIME?!” They’re not exactly pulling this discovery ex nihilo, but they’re taking what might be the private opinion of a couple of isolated specialist researchers who might not have known one another, synthesizing all the evidence together, and saying the thing nobody else wanted to mention.
One way to be a contrarian without being a crackpot seems to be trying to start these sorts of paradigm shifts. Indeed, I notice that they are often people with enough expertise to understand a field who nevertheless acquired that expertise outside of the field itself. For example, Kirsch is a psychologist, as opposed to the psychiatrists and biochemists who usually deal with antidepressant drugs. Cochran is a physicist by training, although he somehow ended up as an anthropology professor. Harris was pursuing a psychology PhD but quit for health reasons and did most of her research independently. Bostrom is a philosopher, and so has license to stick his finger in pretty much whatever pots he wants.
(Whitaker is a journalist. So is Gary Taubes – and, for that matter, Steve Sailer. Science journalism seems like a good example of how somebody can learn a lot about a field while still having an outsider perspective on it)
At their best, these people can look at a field, find ideas that have been excluded from the narrative, and create a new narrative around them. Sometimes this goes horribly wrong – this is how I think of Graham Hancock
So this can sort of be a red flag. But it’s a much less glaring red flag than when people like homeopaths or anti-vaxxers believe they have discovered a new effect, and continue to maintain it exists despite real scientists’ insistence that it doesn’t.
And these are the people who are most likely to get caught in the trap mentioned in Part I. If they’re doing their job right, all they’re doing is calling increased attention to certain results in the field. They’re not the first people to mention that there’s some evidence for recent human evolution. They might not even be the first people to publish a review paper collecting a bunch of different examples of recent human evolution in one place. They’re the first people to be jerks about it, the first people to say “HEY, YOU WITH THE PARADIGM, YOU SUCK” and force all the lower levels of the pyramid – the non-specialists, the administrators, the fieldworkers, the journalists, and the public – to confront the new possibilities head-on.
But shouting “YOU SUCK” doesn’t win anybody any friends. Even if their side triumphs in the end, there will be many much more sober academics who were pushing it almost as effectively. And the same perversity of spirit that led contrarians to challenge the field where it was wrong will probably make them overshoot and challenge the field where it is right. Thus, Taubes not only says that fat has been unfairly demonized, but goes further and says that fat is great for you and you can stay at whatever weight you want just by eating fat. Kirsch and Whitaker not only say that antidepressants were less effective than previously believed, they say they’re worthless for most people and will poison you and psychotherapy is great. Judith Rich Harris not only says that quirks of parenting style don’t matter, she also minimizes the effects of divorce – which I think goes too far.
The likely outcome is pretty much what we’ve got. Even when contrarians win, they lose. Members of the field will be celebrated for being the ones who helped usher in the new paradigm. And the contrarians will be remembered as partisans of crazy false ideas, who happened to gain a thin veneer of credibility by also repeating some true stuff already known by domain experts.
(I’m very happy that brilliant AI researchers like Stuart Russell have joined the fight against AI risk. But I fully expect future textbooks to say that Russell is a great hero for discovering AI risk single-handedly, and also there were some weird guys in Berkeley who gained superficial credibility by parroting Russell’s theories, but were really just silly people writing fanfiction.)
John Baez’s Crackpot Index offers thirty points for “fantasizing about show trials in which scientists who mocked your theories will be forced to recant.”
And if you think you’re a true genius who will have the last laugh, the joke’s on you. You won’t get your show trials even if you’re right.