I argue that lexically primitive proper nouns (PNs) are systematically polysemous in the sense that they can be systematically used in different linguistic contexts to express at least three formally distinct yet analytically related extralinguistic concepts. Given independently plausible assumptions about the nature of these concepts, and of the lexically encoded meanings of primitive PNs, my overarching aim is to show how this framework offers a theoretically attractive way to explain Frege’s puzzle regarding the intersubstitution of so called “coreferential” PNs.