跨方言声调生成及变调处理

Automation of cross-dialectal tone generation and sandhi manipulation

为获得具有通用性的声调处理系统,该文运用计算语言学方法,基于管辖音系学理论提出跨语言声调处理算法。经分析天津话和丹阳话变调算法结果,可得出结论:基于管辖音系学的变调自动化处理模型具有跨方言算法内核一致性、代码精简性以及可移植性等特点。这些特点均源于该音系学理论特有的声调表征模式以及声调音系过程解释机制。

An algorithm was developed to automatically generate tones across dialects within the government phonology framework to achieve universal tone processing.An analysis of the results of the sandhi algorithms for Tianjin and Danyang Chinese dialects concludes that the automated algorithm based on government-phonological tonal theories is very effective.The code is simple and concise and can be readily ported to other tonal languages or dialects.All these come from the unique explanatory device of the tonal theory of government phonology.