ABSTRACT
Two manual choice reaction-time experiments are reported in which subjects identified vowels in consonant–vowel–consonant (CVC) syllables. Specific acoustical relationships were considered between the voices in which the syllables were spoken. For pairs of voices having different average formant frequencies, and hence involving a perceptual adjustment to a different vocal tract size, there was a substantial increase in RT on trials when the voice changed but the response did not. Such increments did not occur for voice differences such as pitch, even though these were perceptually salient; dichotic masking experiments also confirm a distinction between the perceptual effects of these two aspects of voice differences. It is concluded that sequential RT increment measures are capable of providing estimates of the processing demand of decoding operations in perception of speech sounds, but interpretative power derives chiefly from directional predictions for these measures enabled by an understanding and control of stimulus structure.
