When Is A Science Not A Science?
One day, long ago, I asked Dr Michael Osterholm about the connection between vaccines and autism. He works at the Centre for Infectious Disease Research and Policy at the University of Minnesota, so if anyone would know, I figured it’d be him.
The issue with vaccines isn’t the vaccines themselves, exactly, but rather mixing multiple vaccines together. The claim is that no one’s quite sure what happens then, and the 30% increase in autism rates in the US since 2008 has made a lot of people, both crazy and not-crazy, correlate the two in some giant scientific conspiracy.
Dr Osterholm waved his hands and rejected the correlation, but he suggested something more intriguing. He wasn’t going to dismiss the rise in autism but instead pointed to the skyrocketing use of antibiotics (about 838 antibiotic prescriptions for every 1000 people in the US in 2015, 2.5 times higher than in Sweden) as the more likely cause. He said the data is still coming in to confirm this, but the effect of antibiotics on babies’ neurobiology in utero appears to fit the trend much more closely.
Again, he emphasised that we need more data, which is exactly what any decent scientist should say. But it got me thinking about how easy it is to see patterns in activity and to assume you’re looking at science.
My question is: what is science? Why do we believe in its results? Is it just the latest version of religion, or does it actually make sense?
I’m referring to the Popperian definition: science is a discipline that makes falsifiable, reproducible predictions. Of course, the word predates Karl Popper, but the reason I like the definition is that in the last few decades, science has come to mean “authoritative and indisputable truth” instead and become an arm of the State. Authoritative truth has become Official Truth.
Of course, facts without inferences can certainly be authoritative and indisputable; I’ve no problem with that, so long as they are collected by disinterested people who don’t fiddle with the data. That’s unfortunately very common these days, as is drawing conclusions where they shouldn’t be.
My beef is that these conclusions, as in the case of Margaret Mead, have major effects on the outside world, just like Dr Osterholm’s theory might if he’s proven correct about antibiotics. If Mead had just written a travelogue, no false authority would have been assumed. Instead, she told millions of readers that science confirmed what they wanted to hear, which led to some really bad government decisions about social policy for decades.
Once upon a time, we thought science worked because it was reasonable. After all, it’s difficult for a reasonable person to disbelieve something that makes objective statements about reality that can be confirmed by independent experiments, especially if (as Popper taught us) they can be disproved. Back then, science was considered the newest branch of the Aristotelian tradition of rational philosophy, and it laid the foundations of the modern world.
The nice thing about this view is that it allowed for a whole world of reasons that had nothing to do with dropping balls off the Leaning Tower of Pisa. Science was important to the Enlightenment, but to say it wasn’t all about inductive experiment is historically inaccurate.
Then something strange happened. Science was no longer the domain of aristocratic gentlemen puttering about with ideas, living off their centuries-old estate incomes. Sometime in the 19th century, even the stupidest of them started to realise a new world was being born, and it wasn’t just a restoration of classical civilisation. What did we have that the Greeks and Romans didn’t? In a word – science.
This is the cultural trope that gave us positivism and scientism. People no longer believed that science worked because science was reasonable. They decided that if reason worked, it had to be because reason was a scientific concept. Can you see the change?
What the positivists, empiricists, pragmatists, progressives, etc, who came up with this view didn’t see is that by rejecting reason, they were surrendering pretty much all reality and human experience, which is (like economics) not open to inductive experiment, to the forces of violence and superstition they all assumed were a thing of the past.
For instance, why is cultural anthropology considered a “science?” No controlled experiments can be conducted here, no falsifiable hypotheses can be stated, and no underlying quantitative structure can be imagined. On the other hand, it does end in “ology,” so I guess it has that going for it.
Imagine you were a member of some super species that is as different to humans as humans are from fruit flies. One day, you decide to set out pairs of human cultures that are different in only one variable, holding all others constant. If you could do that, then “humanology” would truly be an experimental science, as it would at least produce testable ideas.
Now imagine talking to someone who claims to be a humanologist and finding out he doesn’t construct controlled experiments at all. Instead, his work is based on one test planet called “Earth,” which he just created and left in the fridge for a while. He says this planet has a bunch of random cultures that developed organically, which he has been comparing in “natural experiments.”
Shouldn’t the first thing you tell this guy be to go read Richard Feynman on cargo cult science? Because if humanology without controlled experiments is a pseudoscience, it’s a pseudoscience. Simple as that. It cannot produce any falsifiable results. It doesn’t matter who is doing the experiments.
An easy way to spot an unjustifiable position like this is that, rather than finding the one or two watertight arguments that justify it and ignoring everything else, the person spreads their efforts across everything that sounds vaguely right. This is often a sign that there are no watertight arguments, only false justifications. A classic case of this was Johnnie Cochran’s defence in the OJ trial. Cultural anthropology follows a similar strategy.
Before the last century’s mass insanity, people did “cultural anthropology” and “social science” in a very different way. They knew perfectly well that they were flawed humans writing about other flawed humans in an entertaining way about how we all live, work, and play. Perhaps it’s interesting, when reading about the Yanomamo Indians, to know the percentage of fathers who have killed other Yanomamo males. But without the ability to treat this tribe like fruit flies, statistics will get you about as close to real science as a good trampoline will get you to Jupiter.
One example of this is the tendency to fit models into existing data. That’s called hindcasting, and because there’s no distinction between hindcasting and data dredging, I think it’s a scientific crime. Every hypothesis must be tested with evidence that was not used in constructing the hypothesis. That’s science 101. Yet this is precisely what happens all the time in anthropology, which, after a century of collecting data on the habits of every Papuan cannibal tribe, is pretty low on out-of-sample data.
The problem is easier to see when looking at history (the study of civilised human beings). For the past 150 years, everyone from Comte to Marx to Peter Turchin has been claiming to invent a science of history. Turchin’s is even quantitative, with, like, models and shit. Yet no one has come close to making a testable prediction. History is always happening. Every day, it provides us with a new, fresh nugget of out-of-sample data. But there’s no usable model in any of it.
The worst thing is that it’s all paid for by the State, meaning it’s paid for by you and me. Everything the State does has to be Official and Important. And the more of both the better. So, as a matter of normal Darwinian funding processes, the simple task of looking at other humans and writing about them becomes larded over with pretentious maths, forty-dollar words and outrageous claims about universal truth.
What good does any of this busywork, this intellectual pork, this mindless and mind-numbing nonsense, produce for anyone? Besides, of course, the legion of academics? I suppose it keeps them from making trouble in the streets, at least most of the time. If history teaches us anything, it is that intelligent people can be dangerous. But if the government reorganised its budget and assigned the academics to, say, copying manuscripts, at least it’d have something it could put on TradeMe. Lord knows it needs the cash.