What Does Disfluency Tell the, uhh, Listener?
Item statusRestricted Access
Previous research into speaker disfluency has indicated that filled pauses may fulfil a communicative role. This has subsequently been confirmed in more recent studies which show that filled pauses are beneficial to listeners’ comprehension. Furthermore, filled pauses have been shown to have an effect on listeners’ predictions of an upcoming referent. In response to this latter finding the current study aimed to show that filled pauses occurring before a final referent will lead participants to expect the speaker to refer to a new item, whereas a lack of a filled pause will lead participants to anticipate the upcoming referent to be an item that the speaker has recently mentioned. To test this theory, participants’ eye movements were tracked as they experienced 72 experimental trials, consisting of paired audio and visual stimuli. After participants were exposed to a particular item, the current study examined whether they were subsequently more likely to fixate on either the same item, a phonologically similar and equally plausible end referent or a phonologically similar but implausible end referent according to whether the sentence was disfluent or not. Results did not support the above hypothesis as a repeated measures ANOVA showed no within subject differences between fixation probabilities of objects with regards to disfluency. Also there was no interaction between the experimental manipulations of disfluency and final referent. The current studies’ failure to provide support for results found in previous studies is discussed, looking at procedural differences between studies and providing suggestions for future research.