Abstract

Phonetic perception is at its essence a categorization task involving multiple imperfectly valid cues to category membership. To examine the general cognitive processes underlying the formation and utilization of complex auditory categories such as speech, we have created a set of stimuli crafted from 300-m bursts of white noise. These stimuli vary in onset duration (or attack) and the center frequencies of two spectral gaps. Categories of varying complexity can be created from this stimulus set and input distributions can be strictly controlled. In addition, category formation can be observed ‘‘online’’ in this microgenetic design as subjects make responses (identification and discrimination) while learning the categories. In one such task, subjects were trained with categories defined by boundary values across three attributes. After 10 h of training, the subjects demonstrated extensive learning despite the fact that the categories suffered from ‘‘lack of invariance;’’ that is, no cue could be relied on exclusively to determine category membership. This demonstrates the viability of this paradigm for examining complex categorization. Implications for speech categorization and phonetic acquisition will be discussed. [Work supported by NSF.]

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call