Horgan, Terence [1997] 'Brute Supervenience, Deep Ignorance, and the Problem of the Many', Philosophical Issues 8, pp. 229-36.

I argue that the epistemic approach to vagueness faces the following version of the generic difficulty that Peter Unger dubbed the problem of the many: among the numerous equally eligible-looking candidate boundaries for vague terms and concepts, we cannot conceive or imagine any explanatory basis that would single out some unique candidate over against its competitors. This apparently forces the epistemicist to embrace one or the other of two implausible claims: (1) that although the needed explanatory principles exist, humans cannot grasp them, or (2) that supervenience facts linking use- patterns to semantic boundaries are metaphysically sui generis.