Publication ｜Sophia Linguistica
Kaoru Koyanagi
Second language research has utilized basic units of analysis such as T-unit and c-unit
to segment phrases, sentences, and utterances in order to analyze learners’ production.
However, using such units originally developed for first language data has caused
problems in applying to second language data, since nonnative speakers’ performance,
in particular, in oral production, involves errors, incomplete sentences, false starts,
and hesitations. Recent trends in task-based second language research where learners’
performance is often assessed in terms of complexity, accuracy and fluency (the socalled
CAF) need an agreement as for basic units of analysis before estimating CAF. In
this line of research, the AS (Analysis of Speech) – unit proposed by Foster, Tonkyn &
Wigglesworth (2000) –accompanied with detailed guidelines on application procedures
has become the norm to analyze oral data in English. However, adopting the AS-unit for
Japanese requires a number of modifications due to differences in language structures
between English and Japanese, while maintaining the main conceptualizations in common,
which allows to compare with studies in other languages. In this paper, after reviewing
previous research using units of analysis in English and those in Japanese, new guidelines
to apply the AS-unit to oral data in Japanese is discussed.
