Identifying Problematic Segmental Features to Acquire Comprehensible Pronunciation in EFL Settings: The Case of Japanese Learners of English

Abstract

The present study examines how to identify problematic pronunciation features for particular EFL learners, namely native Japanese speakers (NJs) learning English, to acquire comprehensible pronunciation, and tests the appropriateness of the selection. The study comprises two phases. In the identification phase, eight English-specific segmentals, /æ, f, v, θ, ð, w, l, ɹ/, were selected as the most problematic for NJs by drawing on various cross-linguistic analyses (i.e. a remedial approach) as well as a survey in which the advice of 48 experienced NJ English teachers was examined (i.e. an expert judgment approach). In the experimental phase, the relative influence of these sounds on comprehensibility and accentedness was analyzed. Twenty NJ participants read two types of sentences: sentences containing eight English-specific segmentals and sentences without them. Four native English speakers (NEs) subsequently rated all speech stimuli on a rubric of accentedness and comprehensibility. Significant differences were found between NEs’ ratings of the two types of sentences both in the domain of comprehensibility and accentedness. The results indicate that the eight segmentals determine NEs’ speech perception to a great degree, which in turn provides some support for the validity of the identification procedure (i.e. the combination of the remedial and expert judgment approaches).