Predicate Composition and the Determination of Scope

Proceedings of PACLIC 19, the 19th Asia-Pacific Conference on Language, Information and Computation Predicate Composition and the Determination of Scope Kenji Yokota Faculty of Education and Human Studies, Akita University 1-1, Tegata Gakuen-machi, Akita 010-8502, Japan yokota@ed.akita-u.ac.jp Abstract In this paper we argue that the Japanese causative constructions should be viewed as a complex predicate having two possible c-structure realizations (i.e. a single monoclausal f-structure, and monoclausal and biclausal c-structures). Along with the Optimality Theory architecture, we suggest that the emergence of two c-structures in predicate composition is a consequence of the interactions that regulate the parallel representations of clause structure. The proposed analysis can account for various grammatical phenomena including adverb scope. 1. Introduction There are two levels of “syntactic structure” in LFG; c-structure and f-structure. The one that is built out of words is c-structure; f-structure consists of abstract attributes (features and functions) and their values. The theory of LFG defines f-structure and c-structure as independent, but mutually constraining levels of representation. This makes it possible for a given sentence to have more than one c-structure realization, as long as well-formedness conditions such as completeness and coherence are met at f-structure (Butt 1995, Bresnan 2001). In this paper, we propose that the Japanese (morphological) causatives be considered as a complex predicate having two possible c-structure realizations (i.e. a single f-structure and multiple c-structures). This enables us to provide a full account for phenomena like adverbial scope that cannot be handled in the traditional LFG (Ishikawa 1985, Matsumoto 1996). We further suggest that the emergence of two constituent structures in complex predication is a consequence of the interactions that regulate the parallel representations of clause structure, assumed in Optimality Theoretic (OT)-LFG (Bresnan 2000). 2. Defining characteristics of complex predicates in LFG In LFG, Argument Fusion (or Event Fusion) allows a complex a-structure to correspond to a single PRED at f-structure. Once complex predicate formation is defined at a-structure, the algorithm which relates c-structures to f-structures allows both monoclausal and biclausal c-structures to correspond to the f-structure in (1) (For a more detailed discussion, see Alsina 1993 and Butt 1995). Let us look at how the complex-predicate analysis can be applied to Japanese morphological causatives. e.g., John ga Bill o hashir-ase-ta. John Nom Bill Acc run-Caus-Past ‘John made Bill run.’ (1) f-structure: PRED ‘cause-to-run < SUBJ, OBJ>’ SUBJ [PRED ‘John’] OBJ [PRED ‘Bill’] c-structure: i) [Jon ga [Bill o hashir-]ase-ta] (biclausal c-structure) ii) [Jon ga Bill o [hasher-ase-ta]] (monoclausal c-structure) A-structures are assumed to project skeletal f-structures through lexical mapping principles (Bresnan 2001). Following Butt, Isoda, and Sells (1990), we assume all arguments in upper clause ‘cause’ and the SUBEVENT clause in (2) are mapped onto entities in a simplex f-structure (1). (2) a-structure: REL ‘cause < AGENT, PATIENT, SUBEVENT >’ AGENT [REL ‘John’] PATIENT [REL ‘Bill’] i SUBEVENT REL ‘run < AGENT >’ AGENT i Based on the notion of the complex-predicate formation, the next section provides evidence to support this view. 3. Syntactic/semantic integrity of Japanese morphological causatives …

Content truncated to 500 words