Cost-Lowering Slippery Slopes as Multi-Peaked Preferences Slippery Slopes

Cost-lowering slippery slopes, it turns out, are a special case of a broader mechanism—the multi-peaked preferences slippery slope.

In many debates, one can roughly divide the public into three groups: traditionalists, who don’t want to change the law (they like position 0); moderates, who want to shift a bit to position A; and radicals, who want to go all the way to position B. What’s more, one can assume “single-peaked preferences”: both traditionalists and radicals would rather have A than the extreme on the other side. We can represent the preferences as follows, which is why the preferences are called “single-peaked”:

If neither the traditionalists nor the radicals are a majority, the moderates have the swing vote, and thus needn’t worry much about the slippery slope. Say that 30% of voters want no street-corner cameras (0), 40% want cameras but no archiving and face recognition (A), and 30% want cameras with archiving and face recognition (B). The moderates can join the radicals to go from 0 to A; and then the moderates can join the traditionalists to stay at A instead of going to B. So long as people’s attitudes stay fixed, there’s no slippery slope risk: those who prefer A can vote for it with little danger that A will enable B. {I assume here that the enactment of A doesn’t change people’s preferences; in later posts, I will relax that assumption, and consider the possibility that enacting A might actually alter people’s attitudes about B.}

But say instead that some people prefer 0 best of all (they’d rather have no cameras, because they think installing cameras costs too much), but if cameras were installed they would think that position B (archiving and face recognition) is better than A (no archiving and no face recognition): “If we spend the money for the cameras,” they reason, “we might as well get the most bang for the buck.” This is a multi-peaked preference—these people like A least, preferring either extreme over the middle.

Let’s also say that sh