Spatial affordances trigger spontaneous visual perspective-taking even in the absence of other agents

Poster Presentation: Sunday, May 18, 2025, 8:30 – 11:30 am, Pavilion
Session: Action: Grasping, reaching, pointing, affordances

Robert Walter-Terrill1, Brian Scholl1; 1Yale University

A central goal of vision is to recover information about local environments that is not tied to a single perspective ("What does it look like from here?"), but is generalizable ("What's out there?"). Most work on such themes involves either deliberate imagery ("What would it look like from over there?") or social cognition -- as when taking another agent's perspective ("What would it look like from her shoes?"). Here, in contrast, we show that spontaneous visual perspective-taking occurs even in the absence of other agents -- triggered just by the spatial affordances of the environment itself. Observers viewed two buttons on a table and pressed keys with their right or left hands to indicate when those buttons haphazardly changed to particular colors. The two buttons were vertically aligned from the observers' viewpoint (one closer, one further away), such that adopting a perspective from the left or right side rendered responses spatially congruent or incongruent (as in the Simon effect). Critically, this different perspective was afforded by an environmental regularity in the scene that influenced whether the buttons were reachable. For example: (1) One side of the table had a chair, with the other side flush against a wall. (2) One side of the table had a normally-oriented chair, while the other had a backwards-facing chair. (3) One side of the table had no obstruction, while the other had a translucent screen. Or (4) one side of the table was on solid ground, while the other stood over a sheer cliff face. Each case yielded robust spontaneous visual perspective-taking: despite the task-irrelevance of these manipulations, observers responded faster when the afforded perspectives yielded congruent spatial button/response mappings. These results show how spatial affordances spontaneously promote a type of generalizability during scene perception, and how visual perspective-taking does not require other agents.

Acknowledgements: This project was supported by the Wu Tsai Institute, Yale University.