The cost of perspective switching: Constraints on simultaneous activation.

Visual perspective taking often involves transitioning between perspectives, yet the cognitive mechanisms underlying this process remain unclear. The current study draws on insights from task- and language-switching research to address this gap. In Experiment 1, 79 participants judged the perspective of an avatar positioned in various locations, observing either the rectangular or the square side of a rectangular cube hanging from the ceiling. The avatar's perspective was either consistent or inconsistent with the participant's, and its computation sometimes required mental transformation. The task included both single-position blocks, in which the avatar's location remained fixed across all trials, and mixed-position blocks, in which the avatar's position changed across trials. Performance was compared across trial types and positions. In Experiment 2, 126 participants completed a similar task administered online, with more trials, and performance was compared at various points within the response time distribution (vincentile analysis). Results revealed a robust switching cost. However, mixing costs, which reflect the ability to maintain multiple task sets active in working memory, were absent, even in slower response times. Additionally, responses to the avatar's position varied as a function of consistency with the participants' viewpoint and the angular disparity between them. These findings suggest that perspective switching is costly, people cannot activate multiple perspectives simultaneously, and the computation of other people's visual perspectives varies with cognitive demands.