Maureen Donnelly (SUNY at Buffalo) gives a talk for the Metaphysics of Relations Conference, held on 3rd-5th October 2012 in University of London. In some relational claims- e.g., 'Abelard loves Eloise'-the order of the individual terms determines what relational fact is posited in the claim. In other relational claims- 'Abelard is next to Eloise' the order of the terms seems irrelevant to the underlying relational fact. Whereas there seems to be only one possible fact involving Abelard and Eloise in the relation at issue in the latter claim, there seem to be two possible relational facts involving Abelard and Eloise in the relation at issue in the former claim. I assume that there must be some difference among relations which explains why (and how) some, but not all, relations may return distinct relational facts when combined with fixed relata. I take positionalism to be the view that each argument place of a relational predicate is associated with a particular position or role. On this view, the argument place occupied by a term in a relational claim determines what role its referent plays in the corresponding relational fact. A relation R may generate distinct relational facts involving the same relata if fixed objects may play different roles in R-facts. In this paper, I develop a version of positionalism which assumes that certain properties and relations are instantiated only relative a particular object. I show how object-relative properties can explain differences in relational facts involving the same relata and answer objections to positionalism raised by Fine and MacBride.