Processing time effects of short-term exposure to foreign-accented English

Abstract

Non-native speech can cause perceptual difficulty for the native listener, but experience can moderate this difficulty. This study explored the perceptual benefits of brief exposure to non-native speech. A cross-modal word matching paradigm was used to investigate perception of foreign-accented speech by native English listeners during the first moments of exposure. In 5 experiments, processing speed was tracked by measuring reaction times to visual probe words following English sentences produced by non-native speakers. In Experiment 1, RTs decreased significantly over the course of 16 Spanish-accented utterances and by the end were equal to RTs to a native voice. Control groups in Experiments 1 and 2 who heard a native voice for the first 12 trials were significantly slower than the experimental groups in the final 4 trials, ruling out practice and general strategy explanations for the rapid adaptation. The adaptation effect was replicated with a Chinese-accented voice in Experiment 3. Surprisingly, the control groups also adapted to the accented voice when they heard it at the end of the experiment. Post hoc analyses showed the difference between the control and experimental groups' means was large for the first 2 accented sentences, but attenuated as more sentence trials were included in the means, suggesting adaptation can occur within 2 to 4 sentence-length utterances. In Experiment 4, adaptation to one Spanish-accented voice improved perception of a new Spanish-accented voice, indicating that abstract properties of accented speech are learned during adaptation. In Experiment 5, adaptation to a Spanish-accented voice was as large whether the utterances consisted of English words or mostly legal nonwords. This finding suggested that some characteristics of accented speech can be learned without feedback from lexical knowledge. Overall, the results emphasize the flexibility of the human speech processing system and the need for a mechanism to explain this adaptation in models of spoken word recognition.