Full Idea
If sets are particulars, a nominalist may say that 'blueness' is a set of particulars, but which set? If the particulars 'are blue' this threatens circularity - though resemblance is usually appealed to to avoid this.
Gist of Idea
If 'blueness' is a set of particulars, there is danger of circularity, or using universals, in identifying the set
Source
E.J. Lowe (A Survey of Metaphysics [2002], p.355)
Book Reference
Lowe,E.J.: 'A Survey of Metaphysics' [OUP 2002], p.355
A Reaction
This supports my suspicion that nominalism is superficially attractive and 'scientific', but when you dig deep into it the theory won't get off the ground without universals.