Native and non-native perception of phonemic length contrasts in Japanese: Effect of identification training and exposure

Abstract

Japanese distinguishes between words by the presence or absence of several types of mora phonemes, often realized as a contrast in segment duration, e.g., /hato/ (pigeon) versus /hat:o/ (hat). Several studies suggest that such contrasts are difficult for English listeners to perceive [e.g., R. Oguma, in Japanese-Language Education Around the Globe, 2000, Vol. 10, pp. 43–55]. In this study we investigated the effect on the perceptual ability of both Japanese exposure and perceptual identification training. Three groups of subjects were tested: (1) English speakers with no Japanese experience; (2) English speakers who had spent 1–6 months in Japan; and (3) native Japanese speakers. Subjects participated in a forced-choice identification task in which they heard words and nonwords produced by Japanese speakers and identified which word they heard by choosing among items that minimally differed with respect to these contrasts. Additionally, subjects in the second group underwent five days of perceptual training during which they received immediate feedback, repeating trials until they responded correctly. Results suggest that although the overall performance was relatively high, identification accuracy improved with exposure to Japanese and with perceptual identification training. Implications of this work on theories of second-language learning will be discussed. [Work supported by TAO, Japan.]