Sixth International Conference on Spoken Language Processing
(ICSLP 2000)

Beijing, China
October 16-20, 2000

Japanese Spoken Language Learning System Using Java Information Technology

Jouji Miwa, Hiroshi Sasaki, Kazunori Tanno

Department of Computer and Information Science, Iwate University, Japan

This paper describes a system for Japanese spoken lan- guage learning using Java information technology (IT). The system is automatically evaluated Japanese speech ut- tered by trainee as computer assisted language learning (CALL) on the Internet. For the Internet, the system can be used at any time, in anywhere and for anyone so that it is very suitable for busy learners in the world. Phonetically, Japanese is a mora-timed language but

English is a syllable-timed language. Also Japanese is a pitch-accented language but English is a stress-accented language. So many learners of Japanese as a second lan- guage can hardy hear and speak for special mora and word accent such as super-segmental feature.

Japanese speech uttered by a learner for special mora is automatically scored as 0 to 100 degree and a type of Japanese accent for speech uttered by a learner is automat- ically decided by the system. If learner's score is less than 50 degree and/or learner's accent type is wrong, learner must train until the score goes up with the system.


Full Paper

Bibliographic reference.  Miwa, Jouji / Sasaki, Hiroshi / Tanno, Kazunori (2000): "Japanese spoken language learning system using java information technology", In ICSLP-2000, vol.3, 578-581.