Identity Preservation in Hakha-Lai Tone Sandhi

Abstract

本篇論文的主旨在於討論藏緬語中，哈卡倈語的連讀變調現象。深究的重點是哈卡倈語連讀變調現象中難以預測的變調方向性及變調方向性背後之意義。本篇論文所採用的理論模式是優選理論。本文認為，哈卡倈語的變調方向基本上是由右而左。因為這個方向性可以使得輸出聲調和其參考的聲調（base）比較相同。當這個方向性會衍生出在顯著位置（prominent position）有聲調變化或有標形式時，連讀變調方向就會發生改變。因此，變調的方向性可以由位置信實制約（positional faithfulness constraint），位置音韻制約（positional markedness constraint）和輸出－輸出信實制約（OO-faithfulness constraint）之間的排列順序而得到預測。

This paper examines the tone sandhi phenomenon in the Tibeto-Berman language of Hakha-Lai, which is special for the complex interactions among its elementary rules. The rule interactions in Hakha-Lai lead to both left-to-right and right-to-left rule directionalities in sequences of three or more tones. The rule application directionalities, however, appear to be ungoverned, as none of the principles proposed to date that may contribute to determining directionalities can account for them. In this paper, I argue that the tone sandhi operation directionalities in Hakha-Lai are by no means ungoverned. Normally tone sandhi operates from right to left for identity reasons. This is forced by the IDENT-BOT constraint. The right-to-left direction is sacrificed only when such direction would result in output forms that contain marked sequences or tonal changes at the prominent position, which are forbidden by U[AGREE-t and IDENT-IO-T-L respectively. Thus, the directionalities of tone sandhi operation in Hakha-Lai are naturally predicted by the interaction of the IDENT-BOT constraint, the U[AGREE-t constraint, and the IDENT-IO-T-L constraint, where IDENT-BOT must be dominated by the latter two constraints.