Volume 16 Number 3 361 Jared Bernstein, Amir Najmi, and Farzad Ehsani © 1999 CALICO Journal Subarashii: Encounters in Japanese Spoken Language Education Jared Bernstein Ordinate Corporation Amir Najmi Stanford University Farzad Ehsani Sehda, Inc. ABSTRACT Automatic speech recognition is used in a system to offer first-level, com- puter-based exercises in the Japanese language for beginning high school students. Implementing a preprototype version of the Subarashii system identified strengths and limitations of continuous speech recognition tech- nology in supporting open dialogue practice, in which utterance choices are not written on the screen. Lessons learned from this implementation have led to novel methods in the development of materials for computer- based interactive spoken language education. KEYWORDS Speech Recognition, Spoken Dialogue, Japanese, Authoring System, User Testing and Evaluation INTRODUCTION The Subarashii system offers beginning students of Japanese the oppor- tunity to solve simple problems through (virtual) spoken interactions with monolingual Japanese natives. Subarashii is an experimental computer- based interactive spoken language education (ISLE) system designed to understand what a student is saying in Japanese and to respond in a mean- ingful way in spoken Japanese. The computer system poses problems in written English and offers occasional support to the student in the form of written reminders, but problems can be solved only by speaking and un-