音质在音位系统的基础性作用考察——以汉语感知实验为例

The Investigation of the Fundamental Value of Segmental Phoneme——Evidences from a Perceptual Experiment in Mandarin Chinese

语音系统中最基本的单位是音位,一般可以划分为音质音位和超音质音位。从理论上讲,音质差异对于言语识别(speech recognition)来说更为基础。而汉语(普通话)为声调语言,声调在区别词义时有重要作用。音质与声调在言语识别中的作用孰轻孰重,无疑是一个有价值的课题。本文旨在通过音系理论分析及感知实验两种手段,探索两者在言语识别中的相对价值。实验结果显示,识别正确率与音质音位保留程度成正相关,此结果初步验证了音质音位对于言语感知具有更大价值的假定。但是在音质受损的情况下,声调是否保留对识别的准确性的影响更为显著,这足以证实音质与声调在言语识别中存在互补的关系。本文以汉语为语料,用理论分析、感知实验两方面的证据表明,音质音位在语音系统中的基础地位和作用不可忽视。

The basic unit in phonetic system is phonemes which consist of segmental phonemes and supra-segmental phonemes. Theoretically, the segmental phonemes are more fundamental to speech recognition. Chinese （Putonghua） is characterized as a tone language that employs tones to distinguish meanings. It is a valuable project to evaluate the roles of segmental phonemes and tones in speech recognition. This research aims at exploring the relative value of segmental phonemes and supra-segmental phonemes by means of phonological and experimental methods. The results illustrate that the correctness percentage is positively correlated with the segmental quality and the preliminary prediction that segmental phonemes are more valuable in determining the perception has been testifed. When the segmental value is filtered, tones will play a key role in contributing the correctness of recognition. Therefore, segmental and tone quality forms a complementary relation in speech recognition progress. Based on the Chinese material and through the theoretical and experimental studies, we can predict that the segmental phonemes are signifcant in sound system.