Effects of processing load on speech segmentation

The goal of this research project is to improve our understanding of the perceptual and cognitive factors contributing to the segmentation of fluent speech. Speech-segmentation research investigates how listeners identify word boundaries in the ongoing stream of sounds. There is ample evidence that the mechanisms supporting segmentation can be categorised as Lexical-semantic, or knowledge-driven, ie, resulting from expectations based on word knowledge, meaning, and syntax. Sub-lexical, or signal-driven, ie, arising from phonological and phonetic cues at word boundaries. However, the way in which these mechanisms operate in natural environments is largely unknown. In this study, the grant holder sets out to explore the effects of everyday processing demands such as attentional and memory loads on listeners' relative reliance on knowledge- and signal-driven segmentation. A key question is whether the nature of the processing load (lexical-semantic vs. sub-lexical) affects the relative weights ascribed to knowledge-driven vs. signal-driven segmentation or whether such weights are impervious to concurrent processing demands. Thus, this research aims to place the segmentation problem within the larger context of attention, memory and effort, and hence, to bring speech-segmentation models closer to ecological validity.