Abstract
Some relations - like supervenience and composition - can appear very much like identity. Sometimes, the relata differ only in modal, or modally-involved features. Yet, in some cases, we judge the pairs to be identical (water/H2O; Hesperus/Phosphorus), while in others, many judge one of the weaker relations to hold (c-fiber firing/pain; statues/lumps). Given the seemingly same actual properties these pairs have, what can justify us in sometimes believing identity is the relation, and sometimes something weaker? I argue that it can only be knowledge of differences of individuative criteria that we know a priori because they are built into the meanings of the words/concepts involved. Possible metaphysical conclusions are considered.