Identification and evaluation of L1 and L2 Chinese accents

This study investigates the identification and evaluation of native language (referred to as L1) and second language (referred to as L2) Chinese accents by Chinese judges. Eleven Chinese speakers, including two Beijing Mandarin speakers, two L1-English L2-Chinese speakers, two L1-Korean L2-Chinese speakers, one L1-Thai L2-Chinese speaker, and five Chinese heritage language (HL) (or American-born Chinese, ABC) speakers, recorded their reading of a passage, The Sun and the North Wind. The first 30 seconds of their recordings were extracted, and then the amplitude normalized. 96 Chinese judges completed the accent identification and evaluation tasks online. In the accent identification task, the judges were asked to judge whether the speaker was a native Chinese speaker first. They were then asked to identify their ethnicity and provide reasons/justifications. In the evaluation task, the Chinese judges were asked to rate the accentedness and comprehensibility of the eleven accents on a 0-100 slider scale. The findings of this study reveal that L1 and L2 accents can be differentiated. Among L2/HL accents, some were identified more accurately than others, such as the American L2 and HL accents. All L2/HL speakers except for the Taiwanese HL accent, were identified less accurately, ranging from 12.5% to 34%. Our findings confirm that the familiarity effect is a crucial factor influencing ethnicity identification. The analysis of comprehensibility ratings shows that L1 accents are more comprehensible than all L2 accents, with no significant differences between different L2 accents. Regarding accentedness, all L2/HLs are more accented than L1s, and some L2s are more accented as opposed to others. These findings align with Munro and Derwing's research (1999), suggesting that accentedness may be independent of comprehensibility, namely, highly accented speech can still be highly comprehensible.