This research is designed to test a model of how human speech perception and production evolves through environmental experience. The approach is to study the speech behavior of children and adults who have had massive exposure to new phonetic systems, as in a foreign language (L2), and to determine how such learners acquire the phonetic representations unique to English. Three experiments will examine the production and perception of English stops (/p,t,k/ and /b,d,g/) by native Spanish-speaking Ss; four will examine vowel timing by Chinese Ss; and one will quantify the extent to which departures from norms for English vowel duration by non-native Ss affect listener judgments of speech naturalness. In both the VOT and vowel timing experiments we will examine Ss' temporal specification of sound in speech production, the effect of temporal manipulations on their identification of sounds as """"""""voiced"""""""" or """"""""voiceless""""""""; and the effect of phonetic categorization on their ability to mimic continua of sounds differing in duration. Performance by the English Ss will be compared to that of non-native Ss differing in: (a) age of first exposure to English; (b) amount of English-language experience; and (c) authenticity of English accent. These experiments will define the human ability to modify timing behavior. More importantly, they will test a model offered by Flege (1981) which postulates that the seeming limits on how accurately L2 learners produce sounds in a foreign language results not from an acquired inability to modify articulation, but is an indirect result of """"""""equivalence classification"""""""". The hypothesis is that, as a result of classifying acounstically different L1 and L2 cognate sounds as the """"""""same"""""""", L2 learners fail to develop accurate perceptual representations for the sounds being learned in L2. This, in turn, impose a limit on how accurately the L2 sounds can be produced. If confirmed, the model will provide insight into how previous and ongoing experience contributes to the development and maintenance of the phonetic representations needed for speech. It will also provide a framework for diagnosis and treatment of speech pathologies characterized by abnormalities of articulatory timing, and provide insight into any approach to pathologies requiring the use of sensory stimulation for changing established patterns.
