5th International Conference on Spoken Language Processing

Sydney, Australia
November 30 - December 4, 1998

Spoken Word Identification by Native and Nonnative Speakers of English: Effects of Training, Modality, Context and Phonetic Environment

Debra M. Hardison

University of California, Davis, USA

Several experiments explored the contribution of visual information (lip movements) to spoken word identification by Japanese and Korean learners of English, and native speakers (NSs), and its interaction with sentence context, phonetic environment and, for learners, perceptual training (involving /r,l,p,f,theta,s/ using minimal pairs). The gating technique was applied to videotaped stimuli presented audiovisually (AV) or audio(A)-only to both groups. Stimuli were familiar, bisyllabic words beginning with the following visual categories: bilabial (/p/), labiodental (/f/), /r/, /l/, and nonlabials (/s,t,k/) combined with high, low and rounded vowels. Test (pretest--posttest), initial consonant-vowel (CV) sequence, modality of presentation (AV vs. A), and condition (context or excised word) were independent variables. Groups of NSs were presented with the same stimuli. NNS results revealed word identification was significantly earlier after perceptual training, in AV vs. A-only presentation, in context vs. excised word condition, and varied significantly with initial CV sequence. NS results also revealed significant effects of modality (earlier identification in AV vs. A-only), context, and initial CV sequence. Findings indicate the transfer of perceptual training from segment identification to the process of word identification in connected speech, and are consistent with a multiple-trace model of spoken language processing incorporating visual input.

Full Paper

Bibliographic reference.  Hardison, Debra M. (1998): "Spoken word identification by native and nonnative speakers of English: effects of training, modality, context and phonetic environment", In ICSLP-1998, paper 0120.