Recognition of prosodic contrasts in literal and idiomatic utterances by native and nonnative speakers of English

Abstract

Abilities of three groups—native English speakers, fluent nonnative English speakers, and advanced students of English as a second language (ESL)—to identify literal (L) and idiomatic (I) meanings of ambiguous sentences (e.g., “He was at the end of his rope”) were tested. Tape‐recorded utterances produced by two speakers with either intended L or I meanings were presented singly and in contrastive pairs; listeners judged whether each stimulus expressed an L or an I meaning. Native speakers performed significantly better than fluent nonnative speakers on both single and paired presentations, while the ESL speakers performed at chance on both tasks. Linguistic training was associated with greater accuracies in the native but not in the fluent nonnative speakers. Acoustic analyses had indicated that L and I utterances differed significantly on measures of fundamental frequency (F0), number of pitch contours per utterance, lexical and overall utterance durations, and pausing. I utterances were higher in F0, and had fewer contours, shorter durations, and fewer pauses than the L utterances. The results replicate the original finding that native speakers can discriminate L from I meanings from prosodic cues alone [Van Lancker et al., J. Speech Hear. Res. 24 (1981)], and further indicate that performance on the test was not attributable to a listening strategy and that native competency significantly affects abilities to process these prosodic cues.