Speech Characteristics of Japanese Speakers Affecting American and Japanese Listener Evaluations

Abstract

The study examines what pronunciation features (i.e., segmental and suprasegmental features, as well as other acoustic properties such as speech rate, intensity, pitch, and pitch range) affect the intelligibility of Japanese learners of English, when judged by native-speaking (NS; American) and non-native-speaking (NNS; Japanese) listeners. Kashiwagi and Snyder (2008) concluded, based on both statistical and interview data, that intelligibility and accentedness judgments were quasi-independent of each other, and that segmental features in the speech samples of Japanese learners were perceived both by NS (American) and NNS (Japanese) listeners to be more problematic than suprasegmentals. The present study is designed to verify these conclusions by re-examining the data with further statistical procedures. Additional data on speech rate, intensity, pitch, and pitch range were also added to the statistical analyses to explore the issue of intelligibility. The resulting data suggest that non-native pronunciation of segmentals, especially of certain vowels, affect the judgments of intelligibility most strongly, and that intelligibility scores and accentedness ratings are controlled by different sets of factors. Speech rate, intensity, pitch and pitch range are also found to have some effects. The data also suggest that NS (American) and NNS (Japanese) listeners are affected differently by certain pronunciation features.