The role of spectral, temporal, and dynamic cues in the perception of English vowels by native and nonnative speakers

Abstract

Previous cross‐language studies have demonstrated that different languages may utilize different phonetic features to signal phonemic distinctions. This study investigated which phonetic features are utilized by language learners perceiving speech in the target language and if the acoustic parameters that specify the features are processed in the same way by native and nonnative speakers. Native speakers of English and students of English as a second language (ESL) identified American English vowels that were electronically modified to vary systematically in the type of acoustic information available. The role of three types of acoustic information was investigated: vowel targets available in quasisteady‐state syllable nuclei; dynamic information contained in rapidly changing formant transitions; and intrinsic duration information. Results indicated a remarkably similar error pattern across conditions for both native and nonnative subjects, except for one less‐advanced subgroup of nonnatives who identified vowels in certain modified conditions more poorly than vowels in an unmodified control condition, suggesting that less‐advanced nonnatives may need all possible sources of information to identify English vowels correctly. [Work supported by NINCDS.]