The clustering of discourse markers and filled pauses: A corpus-based French-English study of (dis)fluency

This article presents a corpus-based contrastive study of (dis)fluency in French and English, focusing on the clustering of discourse markers (DMs) and filled pauses (FPs) across various spoken registers. Starting from the hypothesis that markers of (dis)fluency, or ‘fluencemes’, occur more frequently in sequences than in isolation, and that their contribution to the relative fluency of discourse can only be assessed by taking into account the contextual distribution of these sequences, this study uncovers the specific contextual conditions that trigger the clustering of fluencemes in the two languages. First, the contexts of appearance of DMs and FPs are described separately, both in English and French, focusing on their distribution, position and co-occurrence patterns. Then, the combination of DMs and FPs in sequences and their different configurations (DM+FP, FP+DM, etc.) are investigated. Overall, it appears that FPs function differently depending on whether they are clustered with DMs or not, and this difference consists in either maintaining or erasing inter- and intra-linguistic contrasts.