| Abstract | Acoustic continua changing from speech to nonspeech were examined. The continuum used in experiment 1 was synthesized on the Pattern Playback & varied in the bandwidths of the first three formants in equal steps, from the vowel /(script a)/ to a nonspeech buzz. The second continuum used in experiment 2 was created via digital synthesis & varied in bandwidths of the first five formants, from the vowel /ae/ to a buzz. Identification & discrimination data were obtained in experiment 1a (N = 16) to determine the status of the continuum relative to categorical perception; in experiment 1b (N = 8), adaptation effects on the placement of the category boundary were examined. Experiments 2a (N = 8) & 2b (N = 8) followed the same procedures as 1a & 1b. Category boundary shifts for both continua were similar to those previously found for natural speech. Results indicate that perceptual adaptation of speech sounds does not derive from either phonetic or auditory feature detectors. It is suggested that the basis for adaptation might be expressed in terms of "sensitivity to the higher order values inherent in acoustic pressure fluctuations, without decomposition into features." |
|---|
