Talking is preceded by planning the thought to be conveyed on the one hand, and by assigning a grammatically, phonologically, phonetically, and pragmatically appropriate actual linguistic form to it on the other. Speech planning processes normally take place in parallel, so much so that the speaker is unaware of the individual operations. However, occasional disharmony may arise in them, leading to various types of disfluency that disrupt the natural flow of spontaneous speech. - A self-monitoring mechanism of the speech production process is available during speaking, ready to recognize and repair any faulty messages that may occur. Overt monitoring is responsible for corrections that are observable in speech production. Covert monitoring, on the other hand, takes place at the various levels of speech planning. Temporal patterns in speech help us in locating disfluencies, as well as sites of overt or covert error detection, with high probability. The length of pauses preceding corrections has a predictive function. - This paper surveys some relevant models and then demonstrates the operation of self-monitoring via an analysis of Hungarian data
Financed by the National Centre for Research and Development under grant No. SP/I/1/77065/10 by the strategic scientific research and experimental development program:
SYNAT - “Interdisciplinary System for Interactive Scientific and Scientific-Technical Information”.