Japanese language learners’ production of short/long vowel and single/geminate consonant contrasts: A longitudinal development.

Abstract

This longitudinal study examined the development of native English speakers’ production ability to distinguish between single/geminate consonant and short/long vowel pairs in Japanese over 4 months abroad in Japan. Seven participants, who had formal Japanese language classes for 2 years in the US, recorded rika(a) and ka(k)ko pairs in a carrier sentence three times each, at three speaking rates, before and after the study abroad (pre-/post-Japan). Seven measures, including the duration of contrasting segments, words, and sentences, were taken and various ratios were also calculated. Of the native Japanese speaker criteria used to evaluate participants’ data, the vowel-to-word and closure-to-word duration ratios of 0.38 and 0.35 were used as boundary ratios between short and long segments [Hirata (2004); Hirata and Whiton (2005)]. Using these criteria, participants’ data were classified at 66% accuracy at pre-Japan but improved to 85% at post-Japan for vowel distinction. In contrast, they improved from 73% to 74% accuracy for consonant distinction. Analyses of other durational measures also supported that participants made significant improvements in vowel distinction, but not in consonant distinction. These results support the claims of Toda (1997) and Han (1992) and extend to include intermediate language learners into their respective studies.