Electronic Thesis and Dissertation Repository


Master of Science


Health and Rehabilitation Sciences


Dr. Lisa Archibald


Statistical language learning is an implicit process wherein language learners track sequential statistics in fluent speech, and may it facilitate the learning of word boundaries. This process is well studied, however, the cognitive mechanisms supporting it remain poorly understood. The present thesis investigated whether domain-specific or cross-domain explicit working memory engagement would impair implicit statistical learning of word boundaries in fluent speech. Participants (n = 110) were exposed to an implicit statistical word segmentation paradigm while concurrently engaged in no other task (control), or an explicit domain- specific (verbal) or cross-domain (visuospatial) working memory task of either low- or high- demand. Participants in the control task and either visuospatial task (low- and high-demand) reliably segmented words from the artificial language, however those in either verbal working memory condition (low- and high-demand) did not. These findings suggest an interference effect on implicit verbal learning by explicit processing of material from the same domain.