This project is a continuation of a collaborative research program to understand general auditory categorization as it relates to speech perception. Language learners form (phonetic) auditory categoriesof native-language sounds from the distributions of experienced speech sounds produced by many talkers. However, these averaged categories may not be relevant to the speech produced by a specific talker. For example, non-native speech may not adhere to the normal patterns of native speakers. The project described in this application examines the ability of humans to """"""""tune"""""""" auditory categories to characteristicsof the particular listening situation. Speech from non-native and native speakers of English will be analyzed to determine differences in the distributions of sounds that may affect perception. These productions will be presented to listeners (sometimes in manipulated forms) as precursors to a categorization task. The degree to which speech sound categorization is affected by talker-specific information will be ascertained. Two possible mechanisms underlying this ability will be examined. In one set of experiments, speech sounds will be preceded by non-speech precursors. These precursors retain some of the acoustic structure of the speech precursor sentences but none of the linguistic/phonetic content. Shifts in the categorization of the speech sounds will indictate that the auditory system tunes categories to the acoustic makeup of the surrounding context sounds. In another series of experiments, listeners will be trained to categorize distributions of complex non-speech sounds. After training, distributions regularities will change. The ability of listeners to track these changes and maintain optimal performance will be examined. The results will indicate the sensitivity of listeners to changes in the distributional characteristics of sound categories. Taken together, these experiments will provide insight into the ability of listeners to accommodate perturbations of normal native speech productions and, more generally, inform theories of auditory categorization.
Showing the most recent 10 out of 52 publications
