Non-native listener's linguistic background and use of contextual cues

Abstract

Non-native listeners have more difficulty perceiving acoustically degraded English speech and/or utilizing contextual cues than native listeners. In addition, non-native listeners' performance varies greatly depending on many linguistic factors, such as age of English acquisition, duration of immersion, etc. Recognizing the importance of these factors, the current study obtained a detailed linguistic profile of 30 non-native listeners to account for their performance on the reverberant Speech-Perception-in-Noise (SPIN) test. Six test conditions (2 levels of noise × 2 levels of reverberation × 2 levels of context predictability) were included. Through correlation and stepwise multiple regression procedures, variables related to reading in English (age of acquisition in reading, age of achieving fluency in reading, preference in reading) to be strong predictors of listeners' performance in all test conditions. Furthermore, listeners' accent severity in speaking in English and proficiency in speaking in English also accounted for a significant amount of variances in the model. These preliminary results of this ongoing study showed that age of English acquisition, a widely used variable in research and clinical work concerning bilingual and/or non-native speech perception, may not be optimal in prediction of listeners' performance on complex tasks involving acoustic degradation and context use. © 2009 Acoustical Society of America.