福建闽中三明方言与沙县方言短语层面的连读变调——韵律音系学的视角

Phrasal-level Tone Sandhi in Sanming and Shaxian Dialects of Central Min:A Perspective of Prosodic Phonology

本文围绕短语层面连读变调这一问题,对三明方言与沙县方言这两种福建闽中方言进行了考察与对比。本文的论析表明,三明方言与沙县方言虽然在单字调与连读变调的一般规律这两个方面既有相同之处又存在差异,但是这两种闽中方言在短语层面的连读变调情况却高度一致。本文在方言调查所得新材料的基础上,揭示了三明方言与沙县方言的连读变调在短语层面与词层面的不同表现。与词层面相比,这两种闽中方言的连读变调在短语层面的运用情况更为复杂,是一个典型的音系-句法接口现象:连读变调在某个短语中是否发生不仅是由音系条件决定的,更要受到短语内部句法结构的影响。在韵律音系学的理论框架下,本文对三明方言和沙县方言的短语层面连读变调进行了分析,证明这两种闽中方言短语层面的连读变调辖域是音系短语,而非其他韵律单位。在此基础上,本文进一步探讨了这两种方言中的音系短语构建问题,并比较了基于边界的方法和基于关系的方法这两种主要的音系短语构建方法在这两种方言中的运用,证明利用基于边界的方法(包括其最初的边界参数设定以及在优选论框架下发展出的同界/包含理论和匹配理论)构建出的音系短语无法准确地预测这两种方言中的短语层面连读变调,而基于关系的方法只需稍作修改就可以合理地解释这两种方言中的短语层面变调问题。

This paper compares Sanming and Shaxian dialects,two Central Min dialects in the Fujian province,from the perspective of tone sandhi at the phrasal level.Previous studies on tone sandhi in these two dialects are confined to the lexical level,while phrasal-level tone sandhi phenomena have not yet been well studied.This study fills this research gap by examining the application of tone sandhi at the phrasal level in Sanming and Shaxian,and demonstrates that,although both similarities and differences can be found between these two dialects in terms of their citation tones and regular tone sandhi patterns,they share the same properties with respect to their phrasal-level tone sandhi.Based on the primary data from fieldwork,this paper shows that tone sandhi at the phrasal level in these two dialects is much more complicated than that at the lexical level.In both dialects,tone sandhi obligatorily applies at the lexical level,while their phrasal-level tone sandhi is a typical case of the phonology-syntax interface.It is found that whether tone sandhi occurs in a given phrasal-level construction is not only decided by the phonological condition(i.e.,the underlying tones),but also determined by the internal syntactic structure of the construction.Under the theoretical framework of prosodic phonology,this paper argues that the phrasal-level tone sandhi domain in these two dialects is constituted by the phonological phrase,rather than other prosodic constituents.It is thus hypothesized that,only when the components of a given phrasal-level construction are grouped within the same phonological phrase after the construction of the phonological phrase domain,can tone sandhi apply between these components.In contrast,if the components of a given phrasal-level construction are separated by domain boundaries between two phonological phrases,then tone sandhi will be blocked between components across the boundaries.This paper also discusses the construction of the phonological phrase domain in Sanming and Shaxian dialects through the comparison between the End/Edge-Based Approach and the Relation-Based Approach,two of the most important approaches to phonological phrasing proposed within the framework of prosodic phonology.It is demonstrated that,the End/Edge-Based Approach,including its original parameterized algorithms and its subsequent development(e.g.,Align/Wrap Theory,Match Theory)within the Optimality Theory,fails to make correct predictions about phrasal-level tone sandhi in these two dialects,while the slightly revised Relation-Based Approach can nicely account for all the phrasal-level tone sandhi phenomena that have been found so far.