Using unfilled pauses to measure (dis)fluency in English-Chinese consecutive interpreting: in search of an optimal pause threshold(s)

Unfilled pauses (UPs) are commonly regarded as an index of interpreting (dis)fluency. Despite their important role, it is not always straightforward in interpreting research to describe and characterize UPs. One potential controversy concerns the selection of a pause threshold to identify UP. A wide range of thresholds, ranging from 0.2–3 s, has been operationalized in interpreting literature. However, there has been little empirical evidence supporting the choice of a particular threshold or informing such selection. The present study, based on a large corpus of 320 English-Chinese consecutive interpretations, intends to generate such evidence by empirically examining the correlations between three acoustically measured, UP-related temporal variables (i.e. number of UPs per minute, mean length of UPs, and phonation time ratio) and rater-generated fluency ratings, for 16 pause thresholds that systematically increased from 0.25s to 1s. Major findings include: a) for both interpreting directions choosing a higher pause threshold generally led to lower correlation coefficients; b) for both directions phonation time ratio had the strongest correlation with subjective fluency ratings; c) stronger correlations were observed in the English-to-Chinese direction than the opposite direction. We discuss these results and provide suggestions regarding the selection of pause thresholds for future research.