Investigating the theory of intentional disfluency production: Do we mean to be disfluent?

Abstract

Disfluencies in speech are attributed to problems speakers encounter at the planning level. However, it is not yet known if they are produced with a purpose. We reviewed a theory proposed by Herbert Clark, which assumes that disfluencies are intentionally produced by the speaker to signal such problems to the listener. To do so, we investigated if the speaker’s inability to attend to the listener lead to a reduction in disfluency. We employed a dual task paradigm that would theoretically reduce the speaker’s ability to attend to their listener. Although our experimental manipulation was not effective, analysis using mixed effects models demonstrated that participants were no more likely to be disfluent if they were also attending to their listener or not. We concluded that disfluencies are not intentionally produced, but also note that our design requires modification before we can assert our conclusion’s validity.