At the beginning of this article we posed several questions that we felt had not been sufficiently addressed by previous research: What is the proper balance between segmentals and suprasegmentals in pronuncia- tion pedagogy? What defines comprehensible pronunciation and the pronunciation threshold? What are the pedagogical priorities for teachers
of pronunciation? We have attempted to address these questions by considering a series of studies (Riney & Flege, 1998; Riney & Takagi, 1999; and the current study) that have focused on different aspects of the pronunciation of the same 11 Japanese EFL speakers over a 42-month interval. Although we did not directly address comprehensible pronunciation because we were aware of no established measure for it, we consider it likely that the substitution of oneJapanese sound ( [r] ) for two contrasting English sounds (/J/ and /1/) contributes to low comprehensibility and increased ambiguity of speech. It is the responsibility of pronunciation instruction to try to minimize this. At a segmental level, at the very least, EFL speakers should be able to make a distinction between two separate phonemes instead of substituting one LI phoneme for both L2 phonemes. Although the study reported here focused on segmentals, we support the current trend in pronunciation pedagogy that emphasizes both segmentals and suprasegmentals. We also hope that future studies will try
to pinpoint, forJapanese EFL speakers and for EFL speakers of other LI backgrounds, what particular aspects of the suprasegmentals most contribute to global foreign accent and problems in comprehensibility. Only a pronunciation curriculum that is based on a careful identification of learners' needs can expect to articulate the proper balance between a
focus on suprasegmentals and a focus on segmentals.