Matt Yglesias has a post where he says that slippery slope arguments are often “logically” false but “empirically” correct. This gives me an opportunity to highlight a minor argumentative pet peeve of mine: the conflation of reductio arguments and slippery slope arguments, which are related but distinct—and I think it’s a distinction worth preserving.
The reductio ad absurdum is related to a familiar form of proof in logic: You make an assumption and then derive from it a contradiction or a known falsehood by a series of valid inferences. The assumption is then disproved. A typical reductio shows that the principle according to which one would support, say, policy X is also an equally compelling reason to support policy Y, where policy Y is taken to be self-evidently misguided or horrible. So, for instance, pro-lifers will often argue that if you support late term abortion, that because there’s almost no functional difference between an 8 month fetus and a newborn, you must also support infanticide, which is clearly awful. (I offer this argument as an example; I don’t mean to endorse it.) The argument being advanced is not (usually) that we will one day be killing newborns if we accept late abortion, merely that there would be no consistent reason to oppose the latter and not the former. Usually, though, these arguments serve mostly as starting points or intuition pumps to get us to see what’s wrong with the underlying principle, since if policy Y really is so awful and logically equivalent to X, then it should be at least possible to condemn X on its own terms.
The slippery slope argument, as Matt notes, is not a logical but a sociological claim. We can recognize that one can perfectly consistently support, say, a form of national ID with various forms of privacy protection built in without also supporting various more intrusive measures. Probably almost nobody would regard such a position as incoherent on its own terms. But civil libertarians might nonetheless say that a national ID (or any number of other such measures) create an institutional structure that’s rife for abuse, or cause people to be psychologically acclimated to being tracked that opens the door to more objectionable measures.
I waste this much space on a simple distinction only because I feel as though I occasionally see people talking past each other because it’s not clear which form of argument is being advanced. It might be useful, for the sake of clarity, for people to briefly state, when deploying either form of argument, which they intend. (They may intend both, of course.)