Japanese English learners use a variety of articulation strategies to produce English liquids

Abstract

This study examines the production of English liquids /r/ and /l/ by Japanese speakers. Six speakers of varying levels of proficiency were recorded using an NDI Wave Electromagnetic Articulograph. They produced words containing /r/ and /l/ in initial singletons and initial clusters. Two native speakers of English were also recorded for comparison. The resulting articulatory data are analyzed for spatial targets, gestural timing, and the amount of variation between utterances. Results indicate a variety of strategies employed by speakers to achieve the non-native targets; however, more native-like, though distinctly Japanese-accented production, among advanced speakers was found. Lower-proficiency and intermediate speakers showed relatively large amounts of within-speaker variation, indicating the ongoing development of articulation strategies. These findings suggest that the acquisition of foreign language sounds is not merely a straight path from L1 transfer to target-like pronunciation, but a complex process involving many paths, hypothesis testing, and experimentation to find a suitable articulation.