Everybody’s favorite nerdy stick-figure comic ran a joke about Karl Popper, the philosopher of science best known for pushing the idea that science works on the falsification of theories. Popper’s work is widely regarded by philosophers as superseded, but he remains popular with scientists. He may be particularly popular with physicists, or it may just be that I mostly read and follow physicists, but if you hear a random physicist asked to explain how science works, odds are pretty good you’ll get a description in a Popper-ish vein: that we make theories, and test them against observation by looking for places where they fail, so a theory can be “proven false” but never “proven true.” This is usually accompanied by teeth-gnashing from nearby philosophers.

I’ve watched this process repeat many times, often followed by lamentations about why physicists cling to these outmoded ideas when philosophers have known for decades that there are numerous places where Popper’s model fails. In the end, I think he continues this shambling philosophical afterlife because while it might not be logically watertight, the notion of falsification as the core of science is useful in practical terms in a way that many later theories are not.

That is, Popper’s sketch of the process of science provides a model to aspire to, even if it doesn’t perfectly describe everything. If you’re a working scientist, it gives you a generally useful sense of how to proceed: you take a model of whatever phenomenon you’re observing, and you try to poke holes in it. Even when you’re the one who invented the model, and are thus invested in its success, you look for places where it might fail as a pre-emptive measure because your rivals are going to do exactly that, and you need to have counter-arguments ready for them. Popper’s falsification theory is not a properly complete model of How Science Works, but it survives among working scientists because it’s a useful heuristic for identifying what to do next.

That also probably accounts for the relative unpopularity of “social construction” theories of science. It’s not that scientists are just too ignorant to notice the social elements– nobody who’s dealt with Reviewer 2’s demands for marginally relevant citations would be foolish enough to claim that there isn’t a social element to the process of science. But it’s not clear what you’re supposed to do with that information, other than gnash your teeth and add the citations the reviewer asked for.

Popper’s ideas lend themselves to direct application in clear and simple ways that the ideas of Kuhn and the rest do not. And that’s (my arrogant physicist’s view of) why scientists keep trotting Popper out and giving philosophers heartburn.