Toward a Unified Grammar of Reference

Abstract The unification of traditionally distinct and apparently unrelated objects of inquiry under common more abstract principles is one of the most welcome results of empirical science. This article proposes to draw together some insights of Longobardi (1994, 1996, 2001) into a unified theory of object- and kind-reference viewed as a single grammatical phenomenon, though crosslinguistically parametrized. The present account aims to improve both in accuracy and explanatory force over those outlined in the articles just cited. To do so, the combined leading intuitions of such works are first spelt out, in section 6, into a deeper generalization about the form/meaning relation in nominals and later deduced from a more principled mapping theory, proposing that a syntactically specified position, traditionally labeled D, is responsible in many languages for one of human fundamental linguistic abilities, reference to individuals (Topological Mapping Theory). After the unification of the syntactic mechanisms available for reference to individuals, virtually all other distinctions simply follow precisely from that between the two varieties of such entities (kinds and objects) previewed in Carlson’s (1977a) ontology, indirectly confirming its continuing heuristic power, and from widely accepted economy conditions of recent syntactic theory.