1 Introduction

Over the last decade, advancements in speech technology have opened up new possibilities in the field of Computer-Assisted Language Learning (CALL)[1]. New experimental methods have yielded insights into various CALL systems [2][3]. And more recently, interest has been focused on automatic systems that are intended to develop native-like pronunciation of non-native learners in a second language (L2). Traditionally, the training of L2 has seldom focused on pronunciation. However, nowadays the importance of acceptable pronunciation has been acknowledged and the use of efficient training methods has been accentuated.

In Language Learning (LL), it is generally said that human speech perception and production patterns become language-specific quite early in the course of language acquisition so that humans acquire the phonetic sounds of their first language (L1) without difficulty. However, it is not always easy for them to acquire that of a second language (L2) after the sound system of L1 is established [4]. More interestingly, they can not hear the points at which their pronunciation does not correspond to that of L1 due to their L1 filter so that they are often unable to produce the new sounds. This is especially true when they cannot move their mouth in the particular way required to pronounce certain L2 sounds [5]. This explains why an effective method is so desirable, indeed needed for L2 learning.

Previous CALL systems largely depended on the contrastive visualization of acoustic features between (native) model speech and (non-native) learners' speech, whereas other current new types of systems have been also criticized due to retrograde and are also incompatible with various theories of linguistics[6]. We believe that speech recognition technology is the key to implementing an effective CALL system. In this thesis, for the segmentals (vowel and consonant sounds) and the suprasegmentals (rhythm) of Japanese, our novel methods to the CALL system are described and demonstrated with the cooperation of a linguistic expert and a panel of language teachers.

• 1.1 Computer-Assisted Pronunciation Learning
  • 1.1.1 Advantages
  • 1.1.2 Interdisciplinary Problems
  • 1.1.3 Our Approach
    • Training Methods
    • Theoretical Basis
• 1.2 Technical Considerations
  • 1.2.1 Speech Recognition
    • Large Vocabularies
    • Speaker Independence
    • Continuous Speech
  • 1.2.2 Multi-Media
  • 1.2.3 Platform
    • Portable LL on NoteBook PC
    • On-site LL with Virtual Reality
    • Cyber LL on the World Wide Web
• 1.3 Thesis Outline