| Abstract | [Introduction]
We can listen to speech at many levels. We can listen selectively for meaning, sentence structure, words, phones, intonation, chatter, or even, at a distance, Auden’s “high, thin, rare, continuous hum of the self-absorbed.” This paper is concerned solely with phonetic perception, the transformation of a more-or-less continuous acoustic signal into what may be transcribed as a sequence of discrete, phonetic symbols. The study of speech perception, in this sense, has in recent years begun to adopt the aims, and often the methods, of the information-processing models of cognitive psychology which have proved fruitful in the study of vision (Neisser, 1967; Habere, 1969; Reed, 1973). The underlying assumption is that perception has a time-course, during which information in the sensory array is “transformed, reduced, elaborate” (Neisser, 1967:4) and brought into contact with long-term memory (recognized). The experimental aim is to intervene in this process (either directly or by inference) at various points between sensory input and final precept, in order to discover what transformations the original information has undergone. The ultimate objective is to describe the process in terms specific enough for neurophysiologists to search for neural correlates. |