Chinese Resultative Verb Compounds And The Lexicon-Syntax Interface

Couched against the background of the Lexicalism vs. Nonlexicalism debate, this dissertation articulates a strong argument against the lexicalist "Endo-skeletal" approach. It argues, on the basis of a re-examination of facts, that RVCs pattern with achievements instead of accomplishments in the event type they denote, and that V1in an RVC does not deserve the "head" status. Drawing on Fillmore (1970)’s bipartite classification of verbs and Levin&Rappaport (2010)’s Lexicalization Constraint, it claims that RVCs are the basic units in Chinese that combine a "manner" with a "result", so that the grammatical meaning of RVCs is interpreted as:an entity undergoes a "change of state" in a manner denoted by V1and ends up in the resultant state denoted by V2.Under the nonlexicalist Exo-skeletal approach, the author has managed to realize the grammatical meaning of RVCs syntactically and a new analytical model is proposed as follows:the "small clause", composed of the deep object and the V2, is a vP composed of the maximal projection of the null functional predicate BECOME, which licenses change-of-state events; V1, instead of selecting the "small clause" as its complement, attaches to it as a modifier; the vP, as an unaccusative core (UC), underlies all RVCs, and within it, V1and V2are reduced to bare roots clustering around BECOME as the modifier and complement respectively, and both acquire the categorical label "V" by virtue of the local structural relation they form with the head; quite different from V2, V1is assumed not to have an independent position in the UC so that it hovers over the structure before combining with BECOME; in the making of the UC, V1directly infuses its phonetic form into BECOME though Conflation, and V2, which is predicated of the DP argument at the Spec of BECOME by virtue of the c-command relation, will finally incorporate into BECOME and join with V1to satisfy the mutual need of morphological feature checking. The validity of the UC model lies not only in the fact that it clears a lot of our confusions brought about by the variant behaviors of an RVC in telling its inherent "root" meaning from the extra meanings it acquires in the syntactic structure, thereby unveiling the true colors of RVC s, but also in the fact that in dissolving the puzzles that baffle the lexicalists it accurately predicts a host of grammatical features of RVCs, and properly accounts for their special syntactic behaviors.On Huang’s (1993) assumption that light verbs in Chinese must be phonetically supported, I further assume that a c-commanding functional predicate, such as DO and CAUSE will trigger head-movement from the UC, and that an RVC acquires the agentive or causative meaning exactly though the head-movement. The assumption of structure extension provides a consistent account of all RVCs-involved argument alternations. In the course of the building and extension of the UC, thematic roles are strictly interpreted in terms of configurational relations, namely, a DP gets interpreted as an Agent, a Causer or a Theme by virtue of occupying the Spec of DO, Cause or BECOME. As the landing site of head-movement always depends on the semantic relation between V1and the DP at the Spec of BECOME, overgeneralizations that are likely to arise in structural extensions will be properly curbed by encyclopedic knowledge. It is shown that the UC model I propose for RVCs can be easily carried over to several related constructions in Chinese, such as the De-construction, Ba-construction, verb copying construction. On Sybesma’s (1997) assumption that the verbal le is a resultative predicate, this generating mechanism is also valid for simple sentences with a monomorphemic verb predicate.The UC model proposed, by virtue of incorporating the spirit of Nonlexicalism, has preserved the reasonable kernel of "the result small clause analysis", but is completely free from the limitations of the latter. It is theoretically superior to other current analyses in the following two aspects. First, by way of ascribing the host of grammatical features of RVCs to their underlying structure, it manages to subject as many facts as possible to a single analytical mechanism, thereby exhibiting great explanatory adequacy and consistency; by grounding semantics deeply in the syntactic structure, it obviates the need to resort to structure-external mechanisms for semantic interpretations, thereby achieving the economy of its theoretical assumptions. To conclude, by showing that Chinese RVCs are produced by the interaction between the functional predicates from the Universal Grammar and the verbal roots from the Encyclopedia, this research has not only articulated a strong argument for the nonlexicalist "Exo-skeletal" approach, but also shows that mechanisms in the syntax may be also operative in the lexicon, and that the "lexicon-syntax interface", if any, must be a direct and transparent one.