In constructing a viable theory of universal grammar, it is becoming increasingly clear that grammatical relations such as subject and direct object must be considered as theoretical primitives and that they play a central role in the formulation of syntactic rules and constraints_ This position, recently advocated by the proponents of a syntactic theory known as relational grammar, is motivated in part by the facts such as: l) Chomsky's derivative definitions of 'subject-of' and 'object-of' may not be applicable universally, 2)the universal properties of certain syntactic processes cannot be adequately captured unless alternations in grammatical relations that accompany transformations are properly expres' sed, and 3) a series of universa...