ELSEVIER Journal of Pragmatics 30 (1998) 485-496 journal of Filled pauses as markers of discourse structure Marc Swerts* Centrum voor Nederlandse Taal en Spraak, Universitaire lnstelling Antwerpen (UIA), Universiteitsplein 1, B-2610 Wilrijk, Belgium and IPO, Center for Research and User-System Interaction, P.O. Bo-¢ 513, NL-5600 MB Eindhoven, The Netherlands Received 2 May 1997; revised version 4 December 1997 Abstract This study aims to test whether filled pauses (FPs) may highlight discourse structure. This question is tackled from the perspectives of both the speaker and the listener. More specifi- cally, it is first investigated whether FPs are more typical in the vicinity of major discourse boundaries. Secondly, FPs are analyzed acoustically, to check whether those occurring at major discourse boundaries are segmentally and prosodically different from those at shal- lower breaks. Analyses of twelve spontaneous monologues (Dutch) show that phrases fol- lowing major discourse boundaries more often contain FPs. Additionally, FPs after stronger breaks tend to occur phrase-initially, whereas the majority of the FPs after weak boundaries are in phrase-internal position. Also, acoustic observations reveal that FPs at major discourse boundaries are both segmentally and prosodically distinct. They also differ with respect to the distribution of neighbouring silent pauses. Finally, a general linear model reveals that dis- course structure can to some extent be predicted from characteristics of the FPs. 1. Introduction One of the typical differences between a well-prepared lecture and daily small- talk is that the latter speaking mode generally contains comparatively more disfluen- cies. These may to some extent be language-specific such as the hesitation mor- phemes 'eeto' and 'anoo' in Japanese. Spontaneous speakers of many other languages, like English, regularly produce filled pauses (FPs), though the sounds of these may differ between various dialects or languages (Cruttenden, 1986). This Marc Swerts is also affiliated with the Fund for Scientific Research - Flanders (FWO) in Belgium. The study reported on in this paper was presented in October 1996 at the International Conference on Spoken Language Processing in Philadelphia (PA). Thanks are due to Liz Shriberg for inspiring ideas, to Robbert-Jan Beun for providing the speech materials, to Anne Wichmann for fruitful discussions, and to Rend Collier, Hanny den Ouden, and two reviewers of the Journal of Pragmatics for comments on an earlier version of this manuscript. * Corresponding address: IPO, Center for Research on User System Interaction, Spoken Language Interfaces, P,O. Box 513, NL-5600 MB Eindhoven, The Netherlands. 0378-2166/98/$19.00 © 1998 Elsevier Science B.V. All rights reserved PII S0378-2166(98)00014-9