A computer-assisted instruction system for self-teaching of discriminating Chinese tones based on their acoustical properties

Abstract

A computer-assisted instruction (CAI) system for self-teaching to discriminate Chinese tones was developed, and provided through the internet to Japanese college students of Chinese language class. The data of acoustical analysis and perceptual judgment on the Chinese tones played an important role in designing the system. The result of observation on characteristic errors in discriminating Chinese tones by the Japanese students was also utilized. This system offered three kinds of use for self-teaching: testing, training, and reviewing. The original word lists consisted of 15 bisyllabic words with every combination of the four tones of Standard Chinese, involving different rates of unvoiced consonants, which disturbed the tone discrimination. In the partial word lists, only words having tone 2 and/or tone 3, which were difficult for the beginners, were excerpted from the original lists. Their speech samples with both female and male voices were accommodated. Temporal changes in fundamental frequency were extracted from the speech samples and displayed as visual cues in the training. It was also possible in the training to present only the words answered incorrectly in the preceding testing. By selecting suitable uses and word lists with the progress of self-teaching, the required time for achieving the goal was reduced.