In writings prior to the publication of The Principles of Mathematics (PoM), Russell denies that relations “in the abstract” ever relate and holds instead that only particularized relations, or relational tropes, do so; however, in PoM section 55, he argues against his former view and adopts the view that relations “in the abstract” are capable of a “twofold use” – either as “relations in themselves” or as “actually relating”. I argue that while Russell rightly came to recognize that rejecting his (...) earlier view is necessary for avoiding the Bradleyan view that complex wholes are unanalyzable, his later view can appear as an ad hoc means of avoiding Bradley's argument against “relational thought”. (shrink)