Abstract

Japanese sentences read by Japanese‐descent Brazilians (L2) and native Japanese speakers (L1) were rated by native Japanese speakers, and Portuguese sentences read by native Japanese speakers learning Portuguese (L2) and Japanese‐descent Brazilians (L1) were rated by Brazilians. The raters evaluated three types of samples: (a) sounds without spectral properties (representing prosodic features); (b) F0‐flattened sounds (representing segmental features); and (c) unmodified sounds. In both languages, the perceptual scores of L1 and L2 samples were separated most clearly in (c), followed by (b) and (a). Both the scores of (a) and (b) showed strong correlations with (c) in L2. These suggest that both prosodic and segmental features play a role in the detection of foreign accent, but the latter has a greater effect. F0 characteristics varied between speaker groups rather than languages. The speaking rate in L2 was correlated with the perception scores. The effect of segmental features was greater in L2 Portuguese than in L2 Japanese, i.e., greater in less fluent speech. Foreign accents in both languages share common properties although having some differences. [This work is the revised version of Komatsu and Kimoto (2008) and Komatsu (in press), and partially supported by KAKENHI (20242010) and HSUH Kotaisa grant.]

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call