Updating Perception and Action Across Real-World Viewpoint Changes

Abstract

According to the action-specific account of perception, performing actions can distort our visual perception of the world. For example, using a reach-extending tool leads people to underestimate the distances to objects. Unlike many real-world situations, these action-specific distortions have largely been studied under stationary viewing conditions. Here, we tested whether these distortions persist when observers move to new viewpoints. Participants stood at one end of a table and viewed as an object was projected onto the table. Participants reached for the object with their index finger or a metal baton, and touched the object if it was within reach. Although all objects were beyond reach of participants’ index finger, all objects were within reach of the baton. After reaching for the object, participants stood in place or moved to a new viewpoint, then estimated the object’s distance from their current viewpoint. In our first two experiments, when participants stood in place, using a tool led them to report shorter distance estimates. However, when participants moved to a new viewpoint, using a tool did not influence their distance estimates. Thus, action-specific distortions did not persist when observers moved to new viewpoints. In our next three experiments, we tested whether these distortions persist when observers produce other types of movement. Specifically, participants rotated in place, took a step backward, or simply walked in place. In all three experiments, using a tool did not influence participants’ distance estimates. Thus, the present findings were not simply due to changes in viewpoint, and could be observed when participants only received sensorimotor feedback from movement. Together, these findings suggest that action-specific distortions may not be as persistent as many theoretical accounts assume. Indeed, because observers are often free to move and observe objects from multiple viewpoints, these distortions may not be observed in many real-world situations.