談漢語中的五種蒙受結構

本文旨在分析漢語中的五種蒙受句型，包括類雙賓結構，非賓格蒙受結構，受益有標蒙受結構，讓與有標蒙受結構，和受害有標蒙受結構。受害有標蒙受結構又分為強調與非強調受害有標蒙受結構。　　Larson (1988)以與格轉換理論分析雙賓結構與施用結構。由於漢語中的前置「給」字句為修飾語，無法進行與格移位，因此與格轉換理論並不適用於分析漢語的類雙賓結構。Zhou（2000）將前置「給」詞組分析為修飾介係詞組，但此分析無法解釋為何受害有標蒙受句中的施用賓語，即「給」字後的名詞，可表被動語態。Tsai (2007) 認為漢語中的類雙賓結構與非賓格蒙受結構為中階施用結構(middle applicative construction)，然而中階施用詞組本身是否構成層階（phase）的設定在類雙賓與非賓格蒙受結構中並不一致。本研究的分析採取 Cheng et al.(1996)的結構分析法(decomposition)，Pylkkänen (2000)的高階與低階施用結構理論(high and low applicative constructions) 與Jeong (2007)提出的論元驅動移位理論(theta-driven movement)。鑑於把字句與前置「給」字句在語意上的相似性，本研究分析也比較此兩種結構在語意與句法上的特性。本研究的分析結果如下。第一，類雙賓結構與非賓格蒙受結構皆包含高階施用詞組與低階施用詞組，結構上的差別在於後者的非賓格輕動詞並不投射限定語(specifier)。第二，受害有標蒙受結構分成強調與非強調兩種，後者的深層結構與類雙賓結構相同。前者則只包含高階施用詞組，「給」字由高階施用詞中心語循環移向焦點詞組，因此獲得強調語意。第三，受益有標蒙受結構中的「給」詞組為一介係詞組。第四，讓與有標蒙受結構包含高階施用詞組，「給」字本身由高階施用詞中心語移位至致使輕動詞中心語，因此得到致使語意。第五，典型的把字句與前置「給」字句差別在於前置名詞的句法地位，前置名詞在把字句中是必要論元但在「給」字句中為非核心論元。由於兩種結構皆衍生出非典型變體，因此把字句與「給」字句的用法區別漸趨模糊。

關鍵字

蒙受結構；施用結構；把字句

並列摘要

This thesis aims to analyze the five types of affective sentences in Mandarin: the pseudo-DOC, the unaccustive affective, the benefactive marked affective, the permissive marked affective, and the malefactive marked affectives, which includes the emphatic and the non-emphatic. Larson (1988) proposes the dative shift hypothesis for DOC and the applicative construction, which is inapplicable for the pseudo-DOC in that the preverbal gei-phrase is adverbial and thus unable to undergo dative shift. In Zhou (2000), the preverbal gei-phrase is analyzed as an adverbial oblique PP, which conflicts with the result of the passive test that the post-gei DP of the malefactive marked affective can be passivized. Tsai (2007) suggests ApplmidP for the pseudo-DOC and the unaccusative affective. The disadvantage of the middle applicative (Applmid) approach is that its property of heading a phase is inconsistent in these two types of affectives. The present analysis is based on theories of the decomposition of Cheng et al. (1996), Pylkkänen’s (2000) high and low applicative distinction, and Jeong’s (2007) theta-driven movement approach. Due to the semantic similarity between them, this thesis also analyzes the syntactic and semantic properties of the ba-construction and the marked affectives. The result of this analysis is shown as follows. First, the pseudo-DOC and the unaccusative affective involve both ApplloP and ApplloP except that the defective vP of an unaccusative verb doesn’t project spec-vP. Second, the malefactive marked affectives are classified into the emphatic and the non-emphatic. The latter shares the same deep structure with the pseudo-DOC; the former involves only ApplhiP, from which gei rises cyclically to FocusP. Third, gei-phrase in the benefactive marked affective is a PP. Fourth, the permissive marked affective involve ApplhiP, from which gei is raised to the vcause and accordingly receives the causative interpretation. Finally, the main difference between the representative ba-construction and the marked affectives lies in the syntactic status of the preverbal DP. It is an argument in the former, whereas a non-core argument in the latter. Since both constructions develop the non-typical varieties, the boundary between these two constructions becomes intangible.