Multiple Representation of Japanese Complex Predicates : A Lexicalist Analysis of Subject and Object Honorific Forms

MULTIPLE REPRESENTATION OF JAPANESE COMPLEX PREDICATES: A LEXICALIST ANALYSIS OF SUBJECT AND OBJECT HONORIFIC FORMS Akira Ishikawa Sophia University, Dept of English and Studies 7 Kioi-cho, Chiyoda-ku, 102 JAPAN Email: ishikawa@hoffman.cc.sophia.tc.jp Abstract Lexicalist approaches to syntax can benefit from multidimensionality of representa- tion as exhibited by Ailohanan(1997). Subject and object honorific forms in Japanese are shown to be complex predicates with its second component grammaticalized and triggering argument-sharing of the two components of the predicate. Multidimensional representa- tion makes it possible to account for the Gricean inference mechanism working during the argument-sharing process of these honorific forms. 1 Introduction Lexicalist approaches to syntax can benefit from multidimensionality of representation as exhibited by Mohanan(1997). By multidimensionality of representation(MR) is meant a rep- resentational scheme in which more than one representation can be associated with a liguistic unit in the lexicon at a particular point of its generation. The lexicon as the generator of wordforms is rather like a knowledge base whose function as a repository of lexical information Comes not only in its static capacity of serving relevant, information at the demand of syntax but also in its dynamic capacity of modifying and updating the overall structure of the lexicon in terms of the relationships between the items of information. It is uncontroversial that lexical information has to reflect various levels of representation in syntax as well as those of semantics and pragmatics. Thus the idea of MR in the lexicon is germane to the nature of the lexicon. Mohanan(1997)'s treatment of NV complex predicates in Hindi is an instance of the MR approach to the lexicon. In the paper, complex predicates are defined as follows: A COMPLEX PREDICATE construction is one in which two semantically predicative elements jointly determine the structure of a single syntactic clause.(p.432) Mohanan uses the following example to illustrate how complex predicates are formally rep- resented using two levels of representation: the semantic structure (Sem Str) and the argument structure (Arg Str). (i) raam-ne apnaa homwark kiyaa. Ram-E self's homework-N do-PF Ram did his homework. (ii) raam-ne mohan-par bharosaa kiyaa Ram-E Mohan-L reliande/trust-N do-PF Ram trusted/relied on Mohan. The verb kar 'do' in (i) is a full verb winch has the following representation, —131-- (iii) [ 'doer' 'done thing'] DO� SEM STR I� I o� o� kar < ARG1� ARG2� > PRED ARG STR whereas in (ii) it acts as a light verb, which comprises a complex predicate together with the noun bharosaa, whose predicatehood is represented as follows: (iv)� 'truster' 'trusted'] TRUST� SEM STR bharosaa < ARG1� ARG2� > PRED ARG STR Semantically, the complex verb (CP) bharosaa kar 'trust do' has at least two arguments, the truster and the trusted. It has also to be shown that the truster, which is an experiences, takes ergative case, just as the doer does. Moreover the nominal bharosaa behaves as an argument ill the syntactic level. These properties of the CP is captured by the following representation. (v)� 'doer'� 'done thing'] DO� SEM STR I�[ 'truster' 'trusted'] …

Content truncated to 500 words