Complex event representation in a typed feature structure implementation of Role and Reference Grammar

Abstract

Role and Reference Grammar (RRG) (Foley and Van Valin, 1984; Van Valin and LaPolla, 1997; Van Valin, 2005) is typologically general, semantically-driven syntactic framework. A major focus of RRG is a fine-grained typology of form-to-meaning mapping, emphasizing the syntactic and semantic structure of complex event descriptions. I introduce a constraint-based typed feature structure variant of RRG which exploits RRG’s semantically-motivated syntactic backbone: the Layered Structure of the Clause (LSC). Each layer in the LSC is represented as a subtype of sign, and the unique combination of syntactic and semantic properties possessed by each layer is captured by a set of constraints on the appropriate sign subtype. Sentences are incrementally built up from predicates (represented semantically as event frames) to propositions using constructions to pass and combine information through the different layers (and juncture types) in the LSC. An English fragment of this constraint-based RRG is implemented using the Attribute Logic Engine (Carpenter, 1992).