跨方言声调生成及变调处理

• 作 者: 贺俊杰
• 机 构: 陕西师范大学外国语学院,
• 出 处: 清华大学学报(自然科学版)
• 发表日期: 2011年09期
• 关键词: 声调 自动处理 音系学 处理算法 计算语言学 丹阳 可移植性 处理模型 连读变调 主位

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。

In order to obtain a universal tone processing system, this paper uses computational linguistics to propose a cross-language tone processing algorithm based on the governing phonology theory. After analyzing the results of Tianjin dialect and Danyang dialect adjustment algorithm, it can be concluded that the transposing automatic processing model based on governing phonology has the characteristics of cross-dialect kernel consistency, code conciseness and portability. These characteristics originate from the phonetic representation of the phonological theory and the interpretation of the tonal phonological process.