Guest post: There is no algorithm for truth

Originally a comment by Bjarte Foshaug on So many? Like five?

Back in my Movement Skeptic days, before the “deep rifts”, for a while my thinking was heavily influenced by Steven Pinker’s The Blank Slate. While I still think Pinker made some valid points*, in the light of everything that’s happened since, I have grown much more sympathetic towards (or at least understanding of) the reluctance among certain feminists to talk about innate cognitive or psychological differences between men and women. As I remember there was a certain “gotcha” that was very popular among “anti-blank-slatists” at the time:

So what you’re saying is that if there were differences in the distribution of interests and talents between men and women that were not entirely attributable to culture, discrimination wouldn’t be wrong after all?

Indeed, I was almost certainly guilty of occasionally using this “gotcha” myself. I now think this is a strawman. It’s not that the supposed innate differences justify discrimination, it’s that they’re too often used to explain away discrimination (Michael Shermer’s infamous “more of a guy thing” comment being a prime example).

There is a tendency among movement skeptics to talk as if claims are either “supported by evidence and sound logic” or not, when, in fact, things are almost never that clear cut. As they say, there is no algorithm for truth. The data is usually at least somewhat ambiguous and open to interpretation, no method is ever infallible, and no study is ever without flaws, so if we’re motivated to reach certain conclusions and avoid others, we can always find reasons why studies that lead to an inconvenient conclusion are “fatally flawed”. For studies that lead to a desirable conclusion we don’t look so hard for flaws and don’t ascribe such fatal consequences to the ones we do notice. We can be biased and guilty of intellectual double standards even without contradicting any established facts and without making any obvious logical fallacies or methodological errors.

So while I don’t deny that there are real differences in the distribution of interests and talents between men and women, my general impression is that right-leaning people tend to be very quick to embrace biological explanations for things like the under-representation of women in positions of power and status without exposing them to the same level of hypercritical scrutiny as the alternatives. They also seem very quick to conclude that if different innate preferences enters into the explanation at all, there is no need to look any further: That’s all there is to it, and sexism has nothing to do with it. And to be fair, at the risk of engaging in false equivalence and bothsiderism, leftleaning people are almost certainly guilty of the opposite double standard.

While I have some issues with Jonathan Haidt, I think he is right to say that the merit of science is not that it makes individual scientists immune to bias, motivated reasoning, intellectual double standards etc. It’s that, at its best, it allows the competing biases of different scientists to “cancel out”, at least to some degree, which is why viewpoint diversity is so vitally important in science. I also think this is a major part of the reason it’s so important to determine in advance (i.e. before the data are in) what is going to count as a positive result. You should always be prepared to bet your pet hypothesis on predictions that haven’t yet been confirmed or disconfirmed. Because once the facts are in, it is always possible to retrofit the data to a desired conclusion.

* Indeed several critics of gender ideology, including Helen Joyce, have made the point that the tendency among many feminists to downplay and minimize the importance of biology has put them in an awkward position when it comes to defending the need for female only spaces.

10 Responses to “Guest post: There is no algorithm for truth”