Trisyllabic tone sandhi in the Changting Hakka dialect
长汀客家方言的三字组变调
Hui-chuan Hsu 许慧娟

Abstract 摘要
"This paper centers around trisyllabic tone sandhi in the Changting Hakka dialect. Many potential solutions will be shown to be unable to explain the data, including modes of rule application, rule relations, and polysyllabic tone sandhi rules. Instead, the complex tone sandhi phenomena in Changting result from the satisfaction and interaction of four constraints, namely the One Step Principle, Tonotactics, Temporal Sequence, and Set Consistency. The One Step Principle prohibits further change of a derived tone. Tonotactics is a well-formedness condition filtering out surface structures with a possible input to tone sandhi. Both the One Step Principle and Tonotactics can predict the directionality of rule application. If there is a conflict, the One Step Principle wins out. Temporal Sequence, which indicates that the default mode of rule application is iterative from left to right, comes into play in cases where the One Step Principle and Tonotactics do not get involved. Set Consistency specifies that trisyllabic strings with the same first two underlying tones share one directionality of rule application. The One Step Principle as a derivational constraint finds a piece of supporting evidence from the Huojia dialect. More interestingly, there exists a discrepancy between tonal and segmental derivation with respect to the One Step Principle in that a segment can undergo phonological rules again and again as long as their structural description is met. This discrepancy invokes some further questions, such as the theoretical motivation behind the difference between tonal and segmental derivation, and the possible cognitive basis for the One Step Principle. As well-known, African tone languages behave quite differently from Chinese dialects in the nature of tone sandhi. The issue of whether the One Step Principle holds in African tone languages is left open for future research."

长汀客家方言的三字组颇为复杂,规律运作形式 (mode of rule application)、规律关系 (rule relations) 和多音节变调规律 (polysyllabic tone sandhi rules) 都不能提供有效的解释。事实上，三字组变调以两字组变调为基础，只需要透过四个限制(constraint)，即一步原则(One Step Principle)、声调组合律(Tonotactics)、语言的时序(Temporal Sequence)、和组内一致性 (Set Consistency) 的交互作用，就可导衍出表层结构 (surface structure)。一步原则要求衍化调 (derived tone) 不得产生变调，而声调组合律则不容许在表层仍有可供变调的语境。一步原则和声调组合律均可以决定规律运作的方向，如果互有冲突，以一步原则为准。如果两者都不起作用，依照语言的时序，变调规律由左向右连续进行。组内一致性是说底层结构 (underlying) 里前两个声调一样的三字组遵循一个规律运作方向。一步原则在获嘉方言找到佐证。此外，音段衍化 (segmental derivation) 并不遵守一步原则。只要语境符合，一音段可连续产生音变。一步原则表现在声调衍化 (tonal derivation) 和音段衍化的差异有否理论上的及认知上的依据都是值得进一步探讨的课题。非洲声调语言和中国方言的变调性质殊异，因此一步原则在非洲声调语言仍否成立又是另一个研究方向。

<< Back 返回

Readers 读者