Abstract
Whether linguistic prediction requires executive resources is currently unclear. Supportive evidence comes from reduced/delayed prediction in populations with fewer available executive resources, such as children or second-language speakers. However, it is possible that differences in these groups instead stem from differences in linguistic experience. In this study we test prediction using the visual world paradigm, investigating two factors proposed to reduce the availability of executive resources during listening: memory load and hearing loss. Specifically, we apply a visuo-spatial cognitive load manipulation in older adult groups matched in linguistic experience and working memory capacity, but differing in hearing ability. Without a load task, participants with normal hearing were earlier to look to the predictable target relative to those with hearing loss, suggesting a cognitive cost to hearing loss that impacts prediction mechanisms. With a load task, participants with normal hearing were later to look to the predictable target relative to the no-load task, while participants with hearing loss showed no difference by load. Our findings suggest that taxing executive resources via load delays prediction specifically in older adults with normal hearing. While the lack of load effect in the hearing loss group is surprising, it could reflect greater resilience to challenge, altered prediction strategies, or differences in task prioritization. All reported patterns were specific to predictable items and were not seen in neutral items. Importantly, differences between groups cannot be explained by differences in linguistic experience, but rather support findings that executive resources play an important role in prediction.