On the Cognitive Status of Pauses in Discourse Production
This chapter discusses some aspects concerning the cognitive status of pauses in research on discourse production. It starts with a concise review of some of the ‘canonical’ studies on language production in which the pause analytical methodology is adopted. Section 2 discusses methods of collecting pause data and constructing pause databases. Section 3 addresses one key issue: The empirical status of pauses. First, it is shown that the relation between pauses and cognitive processes in discourse production can be maintained by looking at what happens in the immediate neighbourhood of pauses. Based on an analysis of actual transcripts, four kinds of pauses are distinguished: Pauses signalling retrieving, pauses signalling monitoring and pauses signalling repairing processes. Secondly, this section discusses how pause time variances can be interpreted in terms of underognitive processes, and the section concludes with a discussion of how pauses are related to text structural characteristics. The fourth section discusses various statistical methods for analysing pause data, showing the kind of research questions that can be addressed by each method.
KeywordsPauses/pausing pause analysis scope of pauses locations of pauses pause time hesitation text production discourse analysis discourse structure planning retrieving formulating monitoring repairing units of production paragraphs sentences clauses constituents mean length frequencies variance covariance
Unable to display preview. Download preview PDF.
- 1.The examples presented in this chapter are all taken from this corpus. A description of the corpus and sampling methods can be found in Schilperoord (1996). Throughout this paper, I will use transcript examples all taken from this corpus. When possible, I have skipped the original Dutch text.Google Scholar
- 2.This finding was replicated in Schilperoord (1997, 2000). See also Section 4.Google Scholar
- 3.Although the problems of detecting and measuring pauses even grow in complexity with this sampling method (see Van Waes, 1991).Google Scholar
- 5.As this is not the place to report the actual results of these analyses, I stick to mentioning the possibilities. The reader is referred to Schilperoord (1996) for empirical data.Google Scholar
- 6.Almost any cognitive theory of production I know of starts from this distinction. See for example: Levelt (1989) and Flower and Hayes (1981).Google Scholar
- 14.This depends fully on the kind of question one would like to study, or the theory one puts to test.Google Scholar
- 15.For further discussion on this issue, see Schilperoord (1996: 293-299).Google Scholar
- 18.It would imply that words are the only category relevant for a production theory, which is reminiscent to ‘early’ psycholinguistic theories of production (cf. Lounsbury, 1954).Google Scholar