Ding, Shuang and Joo-Kyeong Lee. 2020. The relative contributions of segments and prosody to L1 identification in transplanted L2 speech. Studies in Phonetics, Phonology and Morphology 26.3. 437-460. This study attempts to investigate how the foreign accent of L2 speech is correlated with the identification of non-native talkers’ L1s (hereafter, L1 identification) and which factor in L2 speech, segment or prosody, contributes more to L1 identification. In an experiment conducted to determine the correlation between perceived accentedness and L1 identification, it was found that both English native and non-native listeners recognized the talkers’ L1 more precisely for L2 speech with a stronger accent. In an other experiment, segments and prosody were transplanted between Chinese and Korean talkers of English. Listeners varied between English native and non-native Chinese and Korean listeners, and non-native talkers’ and listeners’ English proficiency was varied between high and low. The results show that segments had a greater effect than prosody. However, this didn’t hold true when native listeners listened to low proficiency talkers’ speech composed of Chinese talkers’ prosody and Korean talkers’ segments, or when the segments were derived from the talkers of the same L1 as the listeners. Both showed that non-native talkers’ and listeners’ L1 backgrounds played a role in the L1 identification of L2 speech. Another intriguing finding was that of skewed responses toward native English talkers for high proficiency talkers’ speech. As presented in the results of the first experiment, less accented L2 speech did not provide enough information for L1 identification, while less accented speech produced by high proficiency talkers didn’t have a sufficient amount of phonetic errors or deviances to successfully identify non-native talkers’ L1s. These findings show that non-natives’ L2 proficiency played a significant role in L1 identification. Overall, L1 identification seems to be affected more by segments than prosody, yet still has intricate interactions with L1 backgrounds and L2 proficiency of non-native talkers and listeners. (University of Seoul, MA student and Professor)

 

Keywords: L1 identification, TD-PSOLA, transplanted speech, segments, prosody