From the annals of great-minds-thinking-alike: Erstwhile roomie Glen Whitman emails to point out that he and NYU's Mario Rizzo wrote a paper on slippery slopes in which they make much the same point about precedent and sorites that I did in my Web piece Thursday. Except Glen and Rizzo add the excellent point that the process is highly path dependent. That is (read the column or this won't make sense) whether granular logic leads you to conclude that everything's a heap or that nothing is depends on which end you start from.
When it comes to precedent, that means that, for example, if there had been a spate of early cases where the fact pattern was like Kelo's—takings to hand over to a private developer with "economic growth" as the "public use" rationale—there might well have been an initial body of case law setting out precisely what wasn't good enough to count as a "public use." Perhaps then later cases involving takings to ameliorate urban blight would have been assimilated to that paradigm, leading to the construction of a "heap" in the other direction, in favor of a far narrower takings power.
The problem, of course, is that the government controls the starting-point. They don't start out with attempts at massive overreach that will build limiting precedent. They start with actions that are relatively clearly within the scope of federal power, and then gradually push out from that core. (Of course, one can think of exceptions: the Alien and Sedition acts, say.) There's reason, then, to expect the sorites problem to tend to work in favor of expanded government power, rather than against it. (One can imagine an Epstein-style sorites logic working that way: It's a taking if the government seizes your whole house. Well, what if they only seize 98 percent of the ownership rights? 95 percent? And so on.)