Alice's adventures in um-derland: psycholinguistic sources of variation in disfluency production

Abstract

This study tests the hypothesis that three common types of disfluency (fillers, silent pauses and repeated words) reflect variance in what strategies are available to the production system for responding to difficulty in language production. Participants' speech in a storytelling paradigm was coded for the three disfluency types. Repeats occurred most often when difficult material was already being produced and could be repeated, but fillers and silent pauses occurred most when difficult material was still being planned. Fillers were associated only with conceptual difficulties, consistent with the proposal that they reflect a communicative signal, whereas silent pauses and repeats were also related to lexical and phonological difficulties. These differences are discussed in terms of different strategies available to the language production system.

Keywords:

• disfluency
• language production
• discourse

Acknowledgements

We thank members of the Communication and Language Lab, Ellen Bard, J. Kathryn Bock and Matthew Rispoli for their comments on previous versions of this work, and Keturah Bixby, Shelby Luzzi, Dipika Mallya and Amie Roten for assistance with data coding.

A preliminary report of this work was presented at the 12th Workshop on the Semantics and Pragmatics of Dialogue (LONDIAL'08), King's College, London, June 2–4, 2008.

Funding

This work was supported by the United States National Science Foundation [2007053221]; United States National Institutes of Health [T32-HD055272, R01DC008774].

Notes

1. Disfluencies may not always reflect planning problems. For example, speakers may sometimes deliberately use fillers like uh to introduce a dispreferred response (Schegloff, 2011).

2. The disfluency literature is presently divided over the use of the term repair. Some authors, following Maclay & Osgood (1959), use the term to refer only to corrections of material already spoken. Others, such as Levelt (1983), term essentially all disfluencies repairs under the assumption that other disfluencies such as fillers and silent pauses represent covert repairs of speech still being planned. For the present study, we reserve the term repair to refer to overt corrections of material already spoken.

3. Although work continues to investigate what proportion of the effect of lexical frequency on planning should be attributed to the grammatical level versus the phonological level (see Kittredge et al., 2008, and references therein), this issue does not bear directly on the present hypotheses, which concerned differences between planning at the message level and at either of the later levels. What is important for the present study is that the lexical frequency influences some level after the message level, not which of those levels it influences.

4. In an additional analysis, we also considered whether the properties of a word influenced whether a disfluency followed it. Disfluencies were less influenced by what preceded them than by what followed them, and the relationships that did exist were consistent with our other results: silent pauses but not fillers were influenced by lexical properties (lexical class and previous mention), and repeats were more common after a new plot point had already been initiated rather than before.

5. Confidence intervals are symmetric around the log odds but become asymmetric when parameters are back-transformed into odds ratios.

Additional information

Funding

Funding: This work was supported by the United States National Science Foundation [2007053221]; United States National Institutes of Health [T32-HD055272, R01DC008774].