The Effects of Sentential Context on the Perception of Assimilated Speech by L2 Listeners

The aims of this study were to examine the effect of sentential context on perceptual compensation for assimilation and to compare compensation patterns between English and Korean listeners with a high proficiency level in English. To these ends, we conducted two experiments involving English coronal place assimilation. In the discrimination experiment, two types of stimuli (i.e., compound words and sentences) were presented. In the identification experiment, a target token including one of the two types of codas (i.e., coronal and non-coronal consonants) was embedded in a semantically neutral sentence. The results showed that in the discrimination experiment both listener groups demonstrated higher detection rates in sentences rather than in words. However, the Korean listeners were not as sensitive as the English listeners to phonetic differences of coda consonants in the unviable change context, and they showed more variations in detection rates than the English listeners. The results of the identification experiment presented a significant effect of coda type of target tokens on both listener groups. In sum, the L2 advanced learners were able to use sentential context to perceive assimilated speech, as were the L1 listeners.