概念-句法接口处的题元系统

Theta system at the concept-syntax interface

讨论概念的词汇化以及词汇语义与句法之间的互动,离不开对题元及其映射问题的研究。虽然题元这一概念在当代句法理论中起着举足轻重的作用,却一直有一个定义不清的问题。不同的理论不仅概括出来的题元数量不同,就是抽象出来的题元等级也不尽相同。这自然使得题元遭遇了来源不明的难题,即原生性问题。于是,句法学家试图通过句法结构或者事件结构的分解(或词汇分解)并辅以功能语类的假设来推导题元结构。这一努力很有一点儿生成语义学回潮的味道,虽然表面上看这种分析很符合人们对事件的民俗心理学认知,但它实际上模糊了句法与词库之间的界线,以致无法应对来自理论和经验事实的挑战。在这一背景下,Reinhart另辟蹊径,坚持词汇主义的立场,以特征分解法,将题元分解为c(致使变化)和m(心理活动)两个特征,通过特征组配,形成不同特征丛,并借助普遍映射规则的运作,对论元的分布做出更为科学的预测。Reinhart创立的这套题元系统,不仅具有很强的可操作性,而且为题元的定义确立了概念基础,从而加深了我们对词汇语义以及概念-句法接口等问题的认识。

Although thematic roles play an important part in the interpretation of the lexical information carried by verbs and the regulation of the distribution of syntactic arguments,they have the problem of definability as their semantic content and inventory as well as their prominence hierarchy with respect to their linking to syntactic positions have never been clearly defined.To solve the problem,syntacticians attempt to eliminate theta roles from the lexicon and derive them instead from the decomposition of syntactic or event structures and the postulation of functional categories.This approach,downplaying the status of the lexicon as an active module and blurring the boundaries between the lexicon and syntax,is actually an echo of the analysis proposed by Generative Semantics,which has been abandoned by the mainstream generative linguistics after Chomsky(1970) because of its conceptual and empirical problems that can hardly be solved.Against this background,Reinhart takes a rather different approach in her analysis of theta roles by decomposing them into two features,i.e.c(cause change) and m(mental state),and establishes a universal mapping mechanism to account for the interaction between lexical semantics and syntax.We think that the Theta System proposed by Reinhart is a substantial step toward the understanding of lexical semantics and the concept-syntax interface.In this paper,we present a critical survey of the Theta System as proposed by Reinhart(1991,1996,2002,etc.) and discuss its advantages and limitations in terms of theoretical considerations and empirical evidence.