Prosodic characteristics in Japanese speech by Taiwan Mandarin speakers and native Japanese speakers

Abstract

Previous studies (Ishihara et al., 2011; Sato 1995) show that prosody contributes more to native-like accents than segments do. It was also found that compared with errors in timing, errors in pitch accent in Japanese speech were more tolerable to native and non-native speakers. This suggests that non-native speakers pay less attention to pitch accent when speaking Japanese; as a result, their acquisition of correct pitch accent does not progress as their overall Japanese proficiency improves. In this study, Taiwan Mandarin speakers and native Japanese speakers produced Japanese words with different syllable structures, some containing all short syllables and others at least one long syllable. These words are 2 to 4 moras long and have nine pitch accent patterns. Each participant produced each word in isolation and in a carrier sentence. All speech data were acoustically analyzed to measure (1) the highest F0 point in accented syllables and (2) the difference in F0 between an accented syllable and adjacent unaccented syllables. The purpose of this study is to investigate common F0 patterns in pitch accents among native and non-native speakers of Japanese, and common pitch accent errors made by the non-native speakers.