So speaking of awkward silences, I have been wondering whether the theoretical existence of black swans matters all that much. If I’m surrounded by white swans as far as the eye can see, it seems like a rational (if not logical) decision to bracket the assumption that swans are white, and proceed with my cosmically important swan studies. If my other option is to abandon further swan studies, go to bed and turn my face to the wall until it’s possible to deduce or not the existence of statistically significant set of swans in every possible color, I choose the logically false, but rationally defensible assumption that all swans are white. Meanwhile, I continue to get up in the morning.

The original example of the black swan was used by Hume to demonstrate what’s wrong with inductive reasoning (plenty!). Popper took it further and used it declare that induction simply can’t be a part of scientific methodology. That’s why falsification is the basis of the modern scientific method, i.e. a hypothesis can only be tested and falsified, but not proven. In terms of philosophy, it’s more complicated than that, but in terms of the popular and practical meaning of science—not so much. It’s all pretty logical and defensible in the context of say, theoretical physics or maybe even macroeconomics (is it?), but the holes become obvious when you stand it up against the life sciences. Mervyn Susser, in his excellent and melodramatically titled 1989 essay “Epidemiology Today: A Thought Tormented World,” gives the example of Harvey’s discovery of the circulation of blood.

William Harvey may well have arrived at his hypothesis that the blood circulates by a large leap of the imagination (so Walter Pagel has argued), a leap that would thereby exemplify Popper’s requirement for hypothesis formation. But a reading of de Motu Cordis shows that Harvey’s hypothesis was both preceded and followed by an extensive process of induction. Imagination need not act to the exclusion of inductive reason. By whatever process Harvey reached his hypothesis, he then subjected it to what can be described as hypotheticodeductive tests, being so far again in conformity with Popper. However, he neither refuted his hypothesis nor aimed to do so, but verified it.

So, the ruling out of induction in the biomedical sciences is either a polite consensual bullshit or medicine doesn’t get to be a science. But if medicine doesn’t really get to be science, how much less epidemiology with all it’s wooly sociopolitical context? But no, epidemiology is mostly practiced as though it were abstract science and as though there is no sociopolitical context at all and as though epidemiologists certainly aren’t making any value-based judgments. None of which is true, obviously.

Susser describes this as a conflict between the logical and the rational, and so it is.

I started thinking about all this when I began to do design research on citizen scientists (the beloved users of the beloved thing I am making these days.⁰¹ ) I found that one of their greatest common characteristics is an ill-defined concern that they and their loved ones may be insufficiently protected by science (really epidemiology). It comes down to Susser’s conflict between the logical and the rational. By definition, epidemiology, because it is practiced as though it were a strictly logical Science,⁰² must favor falsely negative associations over falsely positive associations.

It’s only when you apply this protective logic to situations involving the health and quality of life of real, individual humans, that it appears truly grotesque. Its application to something like a potential environmental contaminant involves shocking decisions about value. As a scientist, do you prioritize human health by favoring favor false associations, or the status quo, by favoring negative associations? Real human lives might be lost or destroyed long before there is sufficient evidence to approach the problem deductively or satisfy scientific standards of proof.⁰³ That’s the definition of a value decision. Yet epidemiology is Science and is therefore logical, and so has to deny that value or sociopolitical context enters into it. If you counter the logical with the rational, you are Unscientific.

This is why intelligent, rational people are intuitively skeptical when Studies Prove X Is Completely Safe for Everyone Everywhere. They know you have to be a little bit crazy to be able to reconcile the conflict between the logical and the rational.⁰⁴

If Science denies that scientists must make decisions about value and excludes the inductive reasoning that that motivates so many productive lay researchers (and honestly should motivate biomedical researchers more), maybe it’s time to redefine science.

~

And another thing!

Thing A: I was talking to Mitsu about how everyone’s big concern is that lay researchers can’t be trusted not to make up crank theories and allow their personal bias to blind them to the falsifications popping up everywhere in their data. That’s a problem of course, but it’s not unique to individuals. The bias toward the status quo is just as powerful as any personal crank theory, and in the context of orthodoxy, it’s plugged into wide ranging systems of power. The most useful question is the validity of the science, not to whom the science belongs.

Thing B: All of the above is not even taking into account what happens when you try to make sense of really big data without induction. You have to take to your bed. It can’t be done.

~

Guys? Guys? Where is everyone?