Recently, various types of Computer-Assisted Language Learning (CALL) system have been developed using highly-skilled computer technologies[9]. The primary goal of these systems is to help learners to develop native-like pronunciation in a second language (L2). In this chapter, we purport to implement a CAPL for Japanese speech rhythm by non-native learners of Japanese. In Japanese, speech rhythm is more important than accent or intonation. Many statistical studies show that unaccented words outnumber accented words in standard (Tokyo) Japanese, and that unaccented words do not show any sudden pitch drop and thus the existence of unaccented words leads Japanese to have many unaccented sentences[32][33].
In terms of speech rhythm, Japanese is a syllable-timed language, more
specifically a mora
-timed language, as opposed to
many other languages such as English, are stress-timed languages that
stressed syllables tend to be lengthened and others
shortened[19]. In a syllable-timed language like
Japanese, rhythmic alternation plays an important role in conveying a
speaker's intention or meaning[34]. Particularly,
Japanese has its own peculiar rhythm called Tokushuhaku and it
prevents non-native learners of Japanese from speaking native-like
pronunciation.
Our CAPL system aims to improve a mora-timed speech rhythm as well as to train many rhythm patterns effectively for non-native learners of Japanese. The word utterances of 20 native male speakers were used to examine their rhythm patterns and define rhythm pattern templates by calculating their means and standard deviations for each mora rhythm. Each mora rhythm was relatively measured by the ratio of its length to the word length, not absolutely by millisecond. A tightly-defined template (Hard Template) for the manner of articulation and a loosely-defined template (Soft Template) for the presence or absence of the voicing were used as a template. Finally, we introduce the example of mora rhythm instruction using the rhythm pattern template and discuss its feasibility.
