Numerical identity across instantiations cannot ground the kind of similarity relation observed between distinct particulars.

The point is that relations of similarity between two particulars cannot be self-explanatory; they must have some explanation (some “real foundation”) in the things that are similar. This is amongst other things an argument against resemblance nominalism. But it is also supposed to be an argument against the view that a common nature could be numerically identical in each instantiation. As Paul Vincent Spade helpfully comments in loc. (ftnt. 8), “for mediaeval authors … a thing x is not (qualita