Word-order and argument-marking: Japanese vs Chinese vs Naxi

Word-order and argument-marking: Japanese vs Chinese vs Naxi Paul Law Department of Chinese, Translation and Linguistics, City University of Hong Kong Tat Chee Ave, Kowloon Tong, Hong Kong paul.law@cityu.edu.hk Abstract. It is argued that varying word-order is related not to the marking on the arguments indicating their grammatical relations but to the syntax of object. The flexible word-order in Japanese and the lack thereof in Chinese may be attributed to the presence or absence of argument-marking, but evidence from Naxi shows that although argument- marking may help identify the semantic roles of the arguments from the processing point view, it is syntax that decides the word-order. Insofar as the relative positioning of the object and adverbs cannot be reduced to argument-marking, the mechanism allowing movement of the object past the verb from a universal order derives both the verb-final property and the varying word-order. Keywords: movement, processing, syntax of object, Universal Order Hypothesis.  I am grateful to Masayuki Hirata and Yurie Hara for sharing their native judgment of Japanese with me, and to He Lijuan, He Wanchuan, He Zhili, Yang Yunfang and Zhao Qinglian for being patient language consultants for Naxi. I thank three anonymous reviewers for insightful comments, constructive criticisms and helpful suggestions. The work reported in this paper is supported by a GRF grant from the Research Grants Council of the Hong Kong Special Adminstrative Region government, for which I am very grateful. I assume responsiblity for any inadequacy that remains. Copyright 2011 by Paul Law 1 Introduction A common view about word-order is that if the arguments of a predicate are overtly marked, e.g., for Case, leading to the identification of their grammatical relations (subject, object and obliques), then the arguments may be ordered in more than one way. The semantic roles of the arguments can be recognized by their grammatical relations in conjunction with the voice of the predicate (active, passive). Conversely, if such marking is lacking, then linear order is the crucial means to identify the grammatical relations of the arguments and their interpretations with respect to their semantic roles. The processing point of view above works well for the comparison of a language like Japanese (or Korean) with one like Chinese (or English). The picture becomes a bit more complicated when Naxi comes into the picture. Naxi is a verb-final language like Japanese and Korean. It too marks the arguments to the effect that their grammatical relations may be identified. But the marking is optional; yet, the varying word-order is largely the same. In this paper I argue that for certain word-orders syntax must be appealed to since they do not fall under the processing account. If the underlying phrase structure is universally the same (Kayne 1994), then the superficial variation with respect to the relative positioning of the object and the verb must be due to movement. This account both derives the verb-final property, which evidently has no bearing on processing, and the varying position of the object. Processing nevertheless has a …

Content truncated to 500 words