Costs of switching perspectives in route and survey descriptions

The nature of mental representation of space, particularly the role of perspectives during construction and retrieval of spatial information, has been debated without conclusive evidence. To address this issue, six experiments were conducted to examine the perspective switching costs during comprehension and retrieval of spatial descriptions. Participants read route or survey descriptions of environments line by line. For half of the descriptions, the perspective of the last sentence was switched. True/false verification of sentences from both perspectives followed the descriptions. Switching perspective increased reading times but increased verification times only for survey sentences. This suggests that perspective switching exacts a cost in comprehension, but that the cost dissipates after information retrieval, especially for route descriptions. Asymmetry of perspective switching costs was attributed to a greater role of referential continuity in viewpoint integration of route perspective. To test this theory, participants read continuous, semi-continuous, and discontinuous descriptions (Erlich & Johnson-Laird, 1982) in route or survey perspectives. They comprehended the continuous descriptions faster than the semi-continuous descriptions only for route perspective, due to changes in reference frame location in semi-continuous route descriptions. The third study examined which aspects of perspective, reference frame or referent object, accounted for switching costs by using hybrid descriptions. Switching the reference frame slowed reading times more than switching the referent object. The fourth study tested whether spatial memory is perspective-free or multi-perspective. Switching costs were examined after an acquisition and an extended retrieval from a single perspective prior to the switch. Verification times did not increase, supporting a perspective-free model. Two additional experiments examined the effects of landmark descriptions on perspective switching costs. When visual details of landmarks were described prior to a perspective switch, the switching costs were reduced. However, non-visual landmark descriptions had no effect on the switching costs. Taken together the experiments suggest that switching perspective plays a significant role in comprehension that diminishes with repeated retrieval. They point to a fundamental asymmetry between route and survey perspectives, one that depends on reference frame and continuity. They also suggest that landmark descriptions aid formulation of spatial representation, but only when they contain visual details.