Tone dissimilarity judgments by Chinese listeners
声调异同辨别的测试
Jack Gandour

Abstract 摘要
Fifty listeners from each of three Chinese tone languages (Cantonese, Mandarin, Taiwanese) made direct paired-comparison judgments of tone dissimilarity. Stimuli consisted of 19 different fundamental frequency trajectories (5 level, 4 rising, 4 falling, 3 falling-rising, 3 rising-falling) superimposed on a synthetic speech-like monosyllable. The dissimilarities data were organized into 150 dissimilarity matrices and analyzed by the INDSCAL multidimensional scaling model. The INDSCAL analysis revealed two dimensions which were interpreted primarily as “height” and “direction”. Results from an analysis of variance of dimension weights indicated that Cantonese listeners attached relatively more importance to the “height” dimension than either Mandarin or Taiwanese. Hierarchical clustering analysis indicated that differences in clustering of the stimulus tones can be related to abstract structural properties of listeners’ phonological systems.

对五十个来自三种不同中国方言语系（广东话，标准国语，台湾话）的人做直接的音调异同对比辨别测试。包括十九种不同音调频率刺激（五个高低音阶，四个高音，四个低音，三个低高音，三个高低音）重叠在一个由电脑合成的口语化单音节字上。辨别异同音的资料被组合成为150个异同矩阵，输入电脑中经由音得失卡(INDSCAL) 多层面等级模式来分析。音得失卡(INDSCAL) 分析显示‘高低度’和‘高低音向’为两个主要层面。经由变异分析层面比重的结果显示，广东语系的人相对地较标准国语系和台湾语系的人偏重于‘高低度’这一层面。层集分析指出，对刺激音调归类的不同，可能与测试者语音系统的抽象结构特性有关。

<< Back 返回

Readers 读者