The analysis of characteristics of spontaneous speech
The study of hesitation phenomena that interrupt the fluency of spontaneous speech can only be successfully performed on large speech databases. BEA is suitable for several kinds of such studies. Measurement-based analyses are being carried out to determine the production correlates of repetitions and restarts and the way self-monitoring processes operate during speech.
Data on certain expletives occurring in spontaneous speech reveal a synchronic (ongoing) language change. A functional shift can be confirmed to be in progress on the basis of acoustic phonetic data, with respect to the uses, functions, and pronunciations (temporal properties, formant structures) of certain words.
Corrections of false starts and of errors of lexical access are due to covert and/or overt self-monitoring. For the discrimination between those two types of self-monitoring, phonetic analysis of a large spontaneous speech material is indispensable.