Her talk, "Grammaticalizing case: Some inherent problems", will take place in SS 560A at 3 PM. A reception will follow in the department lounge.
Grammaticalization is a common source of language change. Though most often applied to changes in which lexical items become functional items (e.g. Old English verb willan to Modern English modal will), it also applies to changes among functional items (e.g. perfect to evidential), and to changes that don't involve a change in meaning so much as broadening in the context for a single item. Despite a large literature from typological and functional perspectives (Heine 1993; Hopper and Traugott 1993; Bybee et al. 1994; among others), relatively little work from formal generative perspectives has focused on grammaticalization. But because grammaticalization highlights relationships between categories, it gives us a window into otherwise-abstract representations.
This talk highlights the potential advantages of using grammaticalization to inform syntactic and semantic analysis, focusing on two domains where what grammaticalizes is not necessarily a word or morpheme, but potentially a case pattern. These are possessive modality (the use of possessive morphosyntax to express modal necessity), and correlations between viewpoint aspect and oblique case marking, manifested both by oblique subjects in perfect and perfective clauses, and oblique objects in imperfective clauses.
These are of particular interest because oblique case has been tied, in both traditional and generative analyses, to the semantic interpretation of arguments, i.e. so-called inherent case. But modality and viewpoint aspect involve relations not between individuals, but between sets of worlds or between times. The DPs that surface with oblique case are thus not semantic arguments of the functional head that determines their case — in apparent contradiction of standard views of inherent case as linked tightly to the semantic relation of an argument to the case-assigning head (Chomsky 1981, 1986; Woolford 2001).
This motivates a reconsideration of the mechanisms underlying inherent case, accounting for cases where it is dissociated from thematic interpretation. On the basis of the grammaticalization facts, I argue that inherent case is indeed assigned by functional heads to arguments in their specifiers — but that this should be encoded in terms of a mechanism of feature transfer, and extended to DPs in derived positions. The analysis of grammaticalization thus offers new perspectives on the morphosyntax of case, with implications for continuing debates on the division between structural and inherent case.