Some issues on A-not-A questions
关于汉语正反问句的一些问题
John Xiang-ling Dai 戴相陵

Abstract 摘要
In the classical approach, an A-not-A question in Chinese is derived from its corresponding alternative question by transformational deletion. Huang (1988) challenges the traditional view with these claims: 1) The alternative question is syntactically different from the A-not-A question, and so they belong to different question types. 2) One type of A-not-A questions and information questions show certain similar syntactic behavior and undergo the same phonological realization process, and therefore they should belong to a single question type. While we basically agree with Huang’s first claim, i.e., the necessity to separate the A-not-A and alternative questions, we will in this paper demonstrate that his arguments for the formal syntactic distinction between the two are either faulty or unconvincing. Specifically, the Lexical Integrity (LI) and Preposition Stranding (PS) tests Huang (1988) provides for such a distinction prove to be so superficial that they can be covered by an independently motivated principle in the phonology of Chinese, i.e., the bisyllabification rhythm rule which applies in the prosodic domain formed on certain syntactic focus constructions. The principle in turn illustrates the fact that the two components of morphology/syntax and phonology may influence and constrain each other. Besides the LI and PS tests, the only seemingly acceptable evidence Huang presents for the distinction between the two types of questions, their distributions with regard to island constraints, turns out to be unconvincing, once pragmatic, phonological and parsing factors, along with further data, are taken into consideration. Huang's second claim, that A-not-A and information questions belong to the same question type, will be refuted. In particular, the theoretical assumption on which his first argument is based, i.e., that there is a [+Q]-triggered phonological rule generating both types of questions, lacks independent support, and therefore the argument is logically circular. His second argument, the distributional similarity in some islands between the two question types, is invalidated by further data.

黄正德（1988）提出汉语正反问句与选择问句在句法上应分别归类，而正反问句与特指问句在句法上类似，应归为一类。本文将论证黄正德为正反问句和选择问句在句法上的不同所提供的论据是片面的，没有说服力，从而不能作为两类问句在句法上区别的依据。两类问句在“词语自主律” (Lexical Integrity) 和“介词悬空” (Preposition Stranding) 上的表面句法区别可以从汉语音系学的双音节节奏规则中找到解释。两者在“孤岛限制” (Island Constraints) 上的差异，也由于语用学，音系学及处理分析 (Parsing) 方面的考虑和反例的出现而变得没有说服力。黄正德的正反问句和特指问句同属一类的论点也不能成立。因为理论上的缺陷，由统一的语音形式产生两类问句的论证在逻辑上是循环论证；两者在“孤岛限制”上的相同，由于反例的出现而不能作为两类问句同归一类的依据。

<< Back 返回

Readers 读者