Reading David Runciman’s essay “Why Not Eat an Eclair?” in the 9 October 2008 LRB (it’s about why people vote despite the fact that their individual vote will not count), I hit this passage (which, among other things, explains the title):
Tuck’s threshold argument is compelling but it skirts around a significant fact about real-world elections that I highlighted earlier: although in theory it only requires one vote to take someone over the top, in practice, the closer you get to that threshold the harder it is to find it, as the mist of political enmity descends. One way to address this is provided by Tuck’s account of other cases where the threshold seems to disappear the closer you get to it. This happens when the borderline between two states of affairs is unavoidably vague, even though the process of change is cumulative. Baldness is a classic example. I go bald by losing my hair one strand at a time, but the loss of no one strand is enough in itself to move me from the category of non-bald to bald. So if I consider the loss of my hair on a strand by strand basis, I can’t go bald, not even if I lose it all. The same kind of reasoning can also apply the other way, say to fatness. No single éclair is ever going to make me fat, so I might as well eat this one. But if no single éclair will ever make me fat then, having eaten one yesterday, I might as well eat another one today, and so on, until I become the thing that one éclair at a time isn’t supposed to make me: fat. These are known as ‘sorites’ paradoxes (the ‘sorites’ being a ‘heap’ of the kind that ought never to arise if you add to it one grain of wheat at a time). It is not easy to say how they should be resolved. But Tuck shows that the best way to think about these puzzles is to consider them as not that different from the problem of voting.
All of this stuff is interesting, but what struck me from a linguistic point of view is the use of the word sorites. It is indeed from Greek σωρείτης ‘fallacy of the heap’ (from σωρός ‘heap’; in English it’s pronounced suh-RITE-eze, and despite appearances, it’s singular), but according to Merriam-Webster, it means “an argument consisting of propositions so arranged that the predicate of any one forms the subject of the next and the conclusion unites the subject of the first proposition with the predicate of the last”—a very different thing. The OED agrees, but has a sense 3 “A sophistical argument turning on the definition of a ‘heap’,” which has one citation, from 1768-74: Abraham Tucker, The light of nature pursued II. 140 “The like attack as was made of old by the Academics and Sceptics against the judgment of the senses, with their sophism of the sorites, or argument of the ‘heap’.”
So my question to those who actually know and use the word is: does that third OED sense, with its lonely centuries-old quotation for support, represent a current sense that will have more citations when they get around to revising the entry, or is Runciman misrepresenting current usage?