Well, it’s nice to see Scientific American doing a little philosophy of science for a change, even if it’s really old philosophy of science. As Stephen Turner said in another post, nice to see that they finally caught up with Quine. Philosopher’s joke – that’s the 1970s we’re talking about.

I’m always a bit worried when Sci Am talks philosophy, because they often get it wrong, and they often start from that sneering “scientists are explorers while philosophers are just tourists” kind of stance (I had a physicist tell me that once – we had a lovely argument for an hour or so). And indeed, right on cue in this article they quote Dirac saying more or less that. But what’s motivating this is not so much a stab at philosophers, but at contemporary anti-science types who use some bad version of falsification to claim that they’ve taken apart the science behind vaccines, or evolution, or climate change, or whatever they are on about.

And sure, their reasoning is terrible, for more or less the reasons that the article outlines. Science isn’t about falsification, but about preponderance of evidence. So yes, let’s knock the feet out from under those arguments. But a couple of things about the Sci Am piece:

1. Popper came up with falsification for a specific reason – to combat verificationism. That was the view that science proceeded by finding confirming cases for something. And he was right to push against that – that makes for lazy science. And we’ve known that for a long time, at least back to Francis Bacon, whose empiricism had us looking for similarities between things, but also differences. It became all too easy in the 19th and early 20th century to become enamored with a complex account of things, and then find confirming cases and declare that you have explained it.

2. It’s worth noting that Popper didn’t just say “falsify instead of verify” and leave it at that. I mean, the whole of “The Logic of Scientific Discovery” is written to nuance what falsification actually means. He dumps verification, but instead he argues for corroboration, which looks like preponderance of evidence to me. In other words, going back to what Popper actually said would be useful. But hey, what do philosophers know?

3. And even past the issue of preponderance of evidence, we have the much more difficult problem of theory change. That’s what the anti-science folks really want, right? You have a theory of evolution that has no place for God, I want to dump your theory for one that has God in it. So the real question is theory change – how does that happen, under what circumstances does it happen, when should it happen?

The go-to person on this for a long time has been Thomas Kuhn, mentioned in this piece in passing, and I think in a useful manner (the author points out that Kuhn talks about the fact that people held on to Newton’s Laws of Motion even though they were contradicted by some planetary motion, and Einstein eventually produced a more adequate theory). Kuhn is known for the idea that theories change not for rational reasons, but because of irrational leaps, mostly things like a generation of scientists dying out and a new group coming into activity and holding a different theory.

Sure, fine, paradigm shifts happen sometimes. They’re just as often a smug stance taken by someone who thinks they’ve advanced further than current science, and can see the true nature of things. I think they’re great for looking at the history of science, and seeing the move between Newtonian laws and Einsteinian ones, or seeing the shift from the idea that land masses were stationary to Alfred Wegner’s outrageous idea that they moved (which we now know as plate tectonics, and is more or less universally accepted because, by the way, there was corroboration just like Popper said we should look for).

I’m more interested in someone like Gaston Bachelard’s idea of the problematic. The idea got adopted by Althusser and used in a Marxist context as an “epistemological rupture”, a moment when one way of seeing the world became inadequate and another started to take over. That is clearly useful in a Marxist story about historical change, but it wasn’t what Bachelard had in mind. He wanted to move from a focus on claims that we were defending, and deciding on whether they were true or not, to questions that formed the space in which those claims made sense.

He was convinced that the question had to come first, but that we often lost sight of it in the pursuit of scientific verification. This isn’t a Kuhnian paradigm shift, which is a non-rational moment within an otherwise rational process. It’s not a shift of world-views either (much beloved by evangelicals, who think that if only one becomes saved, i.e. changes a world view, then everything will work out). There’s no seeing the light here.

The point about questions is that they have a life-span. A well asked question yields possible research areas. We see it over and over again in the proliferation of scientific disciplines – almost without exception they come into being because there is a new question asked, and along with it concepts that are tailored to that question, allowing it to flourish.

How do new theories come into being? Often they don’t. Often a new question takes research down a different path. That example of Newton and Einstein? Newton wasn’t proven wrong, he was proven provincial, that is, his laws held at a particular inertial frame of reference. If you go very large or very small, you reach the limits of their applicability. That was not obvious when he formulated them, but it became obvious over time. No one cancelled Newton, the questions changed so that it was apparent where he was applicable and where he wasn’t.

Sometimes theories are proven wrong. Sometimes it’s because concepts become untenable. We needed ether as a medium for light to travel through – until we didn’t. The point for Bachelard is that there are multiple ways of asking new questions, but a truly new question will result in an epistemological rupture, a re-assembly of what we thought we knew. We still see it now, as we come to grips with complexity theory and systems theory in a world still in thrall to linear mechanistic explanations. Lots of stuff doesn’t work with linear causes and effects, including our genomes, the weather, evolution, economies, and plenty else.

Anyway, the point is that this stuff is really complex. The anti-science folks are almost never this complex – they have a desired goal, and then go casting about to figure out how to support it. That, among many other things, makes it bad science.

But this Sci Am article, even if it’s good science, isn’t great philosophy. Popper already had a response. And, if the real question is theory change and development, and how science actually conducts itself, there’s a lot more going on than even most scientists realize, much less the lay public. That’s one reason that philosophers are not just tourists in this enterprise, and having philosophers around to think about how to ask better questions, and what the implications of that might be, is not a bad idea.