Saturday, December 05, 2015

Slippery Slope Arguments

Will Truman recently argued that slippery slope arguments have a worse reputation than they deserve; and Tod Kelly argued in response that they are in fact quite bad. Kelly's reasons for thinking them awful:

(1) SSAs are largely dishonest and lazy attempts to “magic” away strong arguments against one’s position.
(2) SSAs aren’t really tools for convincing people who disagree with you that you’re correct; they’re simply a way to preach to your own choir — while potentially shrinking it.
(3) SSAs assume a wholly static world where one does not exist.
(4) SSAs are entirely indiscriminate.

I think Kelly's argument actually shows something rather different than he thinks -- namely, that standard responses to even very poor slippery slope arguments are very poor as responses to those kinds of arguments. Part of this is that there is always a bit of a muddle about what a slippery slope argument is. There are at least three different kinds of arguments that get called slippery slope arguments:

(A) Causal extrapolations: These are what the phrase 'slippery slope' often seems to suggests to people's minds. This kind of argument is basically a causal prediction; 'if you keep doing this, you're heading to such-and-such bad consequence'.

(B) Motivational extrapolations: These are the camel's nose or thin-end-of-the-wedge arguments. They could all be summarized by the saying, 'If you give them an inch, they'll take an ell'. Unlike (1), these are estimations of political strategy, not tendencies to effects. The 'Overton Windows Move' mentioned by Truman is a good example of this kind of argument.

(C) Identifications of justificatory imprecision: These are about principles or precedents and conclusions that can be drawn given them; given such-and-such principle or precedent, there doesn't seem to be anything that prevents one from also concluding that such-and-such bad thing is good for the same reason.

There are broad features in common among the three kinds -- they all are directional, they all identify a limit that is designated as bad and thus to be avoided, and so forth -- but they are very different kinds of arguments. And it's very easy to see that Kelly's (3) is utterly irrelevant to (C)-SSAs, for instance.

Another thing that I think muddles the ground is the notion that slippery slope arguments are put forward as proofs. This is explicitly assumed by Kelly:

There is literally no position, no matter how innocuous or righteous, that a SSA can’t “prove” will lead to the end of civilization as we know it. That right there should give you some pause about its inherent worth as a tool of discourse.

But it seems very doubtful that SSAs are ever put forward as proofs. Rather, as I have argued before, they are challenges. The point is not to establish that the bad thing will happen -- (A)-SSAs and at least some (B)-SSAs are about present dispositions, not the future, and (C)-SSAs are about rational consistency, i.e., rational consequences and not causal consequences. The point is to raise the question of whether an opposing position is properly thought out in the first place.

Kelly's objection (1) actually draws on this aspect of SSAs, I think, since argumentative challenges are relatively easy to raise in comparison with a lot of other arguments, so machine-gunning challenges at a position is a relatively cheap way to argue, since a person of even mediocre ingenuity can often come up with challenges in a shorter period of time than it takes even a genius to answer them properly. Lazy reasoners do, in fact, tend to fall back on argumentative challenges. But this is a very different thing, of course, from holding that the challenges don't need to be met. Maybe they do, maybe they don't, but the cheap cost of the argument doesn't tell us anything about that. It also doesn't tell us anything about whether a given situation is one in which an argumentative challenge is perfectly justified.

And when we recognize them to be challenges, we see immediately that many of the purported problems evaporate: Kelly's (3) is simply false of SSAs taken as challenges (I think (3) makes the error of confusing causal tendency with causal result, as well, but even if this is not so, it fails to be a correct characterization of something put forward as a challenge), and his (2) is simply irrelevant (since whether a challenge is justified is not dependent on whether the challenged takes it seriously). His (1) and (4), which are the strongest objections if we treat SSAs as purported proofs, run into obvious problems if we are taking SSAs as challenges. It doesn't matter how strong you think the arguments for trying for A are; if you can't rule out that A makes bad B easier, that weakens your overall argument for trying for A. Practical arguments are not hermetically sealed; I may have a good argument for A, but I can at the same time have a good argument for avoiding A -- and what matters for practical reasoning is the overall disposition of all the arguments. You might have a crowd of good arguments that X would be good to have; but you might just have one deal-breaking argument that gives you sufficient reason to avoid X, however good it might be in other ways, because it shows that you don't actually have a way to get X without unacceptable losses or compromises or commitments.

And (4) likewise runs into problems. Obviously the general structure of an argumentative challenge is going to be indiscriminate. Argumentative challenges raise the question of whether rational standards have been properly met, and every practical policy or plan is going to have to meet rational standards if it is to be a rational policy or plan. That challenges could be raised will be taken for granted by any rational and reasonable person; precisely what makes rational and reasonable people rational and reasonable is that they work to have answers to argumentative challenges rather than just assuming that their claims are reasonable. If someone gives an argumentative challenge, there are only two possible adequate rational answers to it:

(a) We have reason to think the challenge doesn't need to be bothered with (e.g., because the purported bad consequence is not bad at all, or because the challenge is logically incoherent)
(b) We have reason to think the challenge can be met.

It's irrelevant that one can challenge everything indiscriminately; what matters is whether the challenge can in principle be answered. Merely assuming that the challenge doesn't need to be bothered with, or that it can be met, is a sign of stupidity. If, to take an example of Kelly's, you are arguing for same-sex marriage and you really cannot show, even when challenged, that the reasons you are giving for same-sex marriage do not also support the bad consequence of marriage to turtles, which you reject vehemently, you are not very bright, and have no reason to regard yourself as in any way reasonable and rational in your support for same-sex marriage; and you should possibly leave the defense of your position to people who are less stupid than you are. We are not talking a high intellectual bar here; it's literally one of the most elementary rational standards: try to reason things through in a principled rather than merely ad hoc way. If your best response to a challenge claiming that you haven't thought through your argument to the end is to say, "Why would one even think beyond this point?", it is you, not the challenge, that is the problem.

Thus Kelly's conclusion fails:

After all, there’s a word for an argument that requires no effort, little reasoning, has little if any expectation of being persuasive, and is equally good for all positions.

That word is “useless.”

This is not true if the argument in question is a challenge rather than a proof. One could very well have a perfectly rational and correct challenge that is easy to make, is unlikely to persuade the person who is challenged, and could be applied to every position -- to take the clearest and most obvious example, if your argument is holding up a basic rational standard and the person being challenged is being irrational. The existence of irrational people exempting themselves from basic rational standards does not make rational argument useless; they make it all the more necessary.(There's an irony, incidentally, and perhaps a bit of deliberate funning, in the fact that Kelly's objection (4) against slippery slope arguments is quite obviously a slippery slope argument.)

None of this is to say, of course, that all SSAs are good challenges. It's sometimes very difficult to make good challenges to arguments and positions. It is sometimes said of free verse or the fantasy genre that they are full of bad writing precisely because they are easy to write but hard to write well; and argumentative challenges are certainly easy to formulate but often difficult to formulate properly. But I think Tod Kelly's argument against SSAs captures the features of most common response to slippery slope arguments. Which is why I said that I think Kelly has really shown that standard responses to even very poor slippery slope arguments are very poor as responses to those kinds of arguments.