Abstract
We propose that the well-known dissociation between egocentric and exocentric distance perception (e.g., Loomis et al., 1992) is based on a distinction between two angular perceptual variables. Errors in exocentric distance tasks, such as the one underlying Gilinsky's (1951) experiments and the perception of depth/width ratio can be explained by biases in perceived optical slant. Errors in egocentric distance tasks, such as egocentric distance estimation, can be understood in terms of biases in the perception of angular declination. Whereas exocentric distance estimates typically demonstrate increasing compression with distance, egocentric distance estimates are normally compressed by a fairly constant amount. Li and Durgin (2010) have shown that exocentric aspect ratio tasks can be used as implicit measures of perceived optical slant which increase logarithmically with viewing distance. Here we further propose that explicit egocentric distance estimation can be treated as an implicit measure of perceived angular declination, which is biased in a linear fashion. To clarify that type of visual information matters more than type of task, we used an egocentric version of the depth/width aspect ratio task. Participants stood at one leg of an L-shape formed by them with two experimenters and were asked to position themselves the same distance from the central experimenter as the central experimenter was from the experimenter at the other leg of the L. Participants set themselves much too far from the central experimenter for all tested distances, consistent with overestimation of angular declination by a factor of about 1.5 (i.e., underestimation of egocentric distance by about 0.7). The egocentric aspect-ratio task provides evidence convergent with other explicit forms of egocentric distance estimation that perceived egocentric distance is underestimated. It also provides implicit quantitative confirmation of explicit evidence for the expanded coding of perceived angular declination, a variable important for motor control and spatial updating.