InterSpeech 2021

Cross-linguistic Perception of the Japanese Singleton/Geminate Contrast: Korean, Mandarin and Mongolian Compared
(3 minutes introduction)

Kimiko Tsukada (Macquarie University, Australia), Yurong (Inner Mongolia University, China), Joo-Yeon Kim (Konkuk University, Korea), Jeong-Im Han (Konkuk University, Korea), John Hajek (University of Melbourne, Australia)
The perception of Japanese consonant length contrasts (i.e. short/singleton vs long/geminate) by native and non-native speakers was compared to examine the extent to which difficult foreign language (FL) sounds are processed accurately. Three groups of participants had Korean, Mandarin or Mongolian as their first language (L1) and had no experience with Japanese. Unlike Japanese, Mandarin and Mongolian do not use consonant length contrastively. The phonemic status of consonant length in Korean is debatable. Further, unlike Japanese and Mandarin which predominantly use open syllables and restrict the occurrence of consonants in coda position, Korean and Mongolian permit a wide range of consonants in that syllable position. Via the AXB task, the participants’ discrimination accuracy of Japanese consonant length contrasts was assessed and compared to that of a group of 10 native Japanese speakers who served as controls. The Japanese group was at near ceiling with little individual variation. The Mongolian (but not Korean and Mandarin) group did not significantly differ from the control group when the target token (X) contained a geminate. All non-native groups were significantly less accurate than the control group when X contained a singleton. These results were interpreted as reflecting the participants’ L1 quantity system.