Armstrong's account explains determinate resemblance via partial identity constituted by shared unit length universals.

One concern with anti-realist accounts has to do with the source of classification of determinates under a determinable term or concept. As above, this is usually taken to involve inexact resemblance, but if inexact resemblance is primitive, then objections of the sort leveled against resemblance nominalist treatments of inexact resemblance may arise, mutatis mutandis (see Fales [1990: 117–20] for discussion, and Rodriguez-Pereyra [2002] for a recent response). An alternative means of classifica