Selective perception, perceptual modes, and automaticity in first- and second-language speech processing

Abstract

A common phenomenon reported by experienced, late learners of a second language (L2) is that comprehending L2 speech, especially under non-optimal conditions (e.g., in noisy rooms, when driving, when the speaker talks rapidly) is more effortful than processing L1 input. A variety of paradigms have documented this phenomenon experimentally. Many current theories of L1 and L2 speech perception invoke concepts of learned (language-specific) patterns of selective attention or attunement to characterize the processes by which native speakers rapidly and efficiently extract sufficient phonetic information from complex acoustic signals in order to recover phonological sequences (words or word-forms). It is suggested here that there are two modes of language-specific speech processing available to adult listeners, a phonetic mode (requiring attentional resources) and a phonological mode (automatic), that are tapped in the laboratory to different degrees as a function of complex interactions of subject’s linguistic experience, stimulus characteristics, and task demands. Exemplary experiments of L1 and L2 listener’s perception, using perceptual assimilation (cross-language identification) and speeded discrimination tasks, as well as electrophysiological indices of discrimination, illustrate some of these interactions, within the framework of the automatic selective perception model of speech perception. [Work supported by NIH-NIDCD and NSF.]