Perceptual Feedback in the Detection and Correction of Errors

Abstract

Self-inhibition is a central concept in the node structure theory. In addition to explaining the many sources of evidence discussed in Chapter 8, self-inhibition is needed to explain how perceptual feedback is processed during an ongoing action and to explain how errors are detected and corrected. Self-inhibition also makes sense of the way speech production becomes disrupted when normal speakers hear the sound of their own voice amplified and delayed by about 0.2 s. Self-inhibition even contributes to “node commitment,” as I call the process of forming new or functional connections between nodes. However, I have developed the last two topics elsewhere (Chapter 10, and MacKay, 1987). The present chapter discusses only feedback processing and the detection and correction of errors. I first examine how errors are detected and corrected and the constraints these phenomena place on theories of the relation between perception and action. I then take these constraints into account in developing a theory of mechanisms underlying error detection and correction and the processing of self-produced feedback in general.

Self-repairs are… rather complex phenomena… they involve quite disparate phonetic processes, such as self-monitoring, the production and detection of phonetic, lexical and other types of speech errors, self-interruption, prosodic marking of the connection.

(Levelt, 1984, p. 105)