Abstract
It has been proposed that a continuous ground surface is important for perception of distance (Gibson, 1950; Sinai, Ooi, and He, 1998). Most studies of absolute egocentric distance perception have only investigated targets placed on the ground. We asked whether distance perception as indicated through blind walking or matching would differ when targets were placed on the ceiling versus the floor. The position of objects on the ceiling is not consistent with our everyday experiences and creates less ground-surface information for depth, potentially leading to increased error or variability. In three experiments, targets were placed 5, 10, 15, and 20 feet from the observer and each subject performed two tasks. In Experiment 1, targets were placed on the ceiling. In one condition, subjects blind walked so that their bodies were aligned underneath the target. In the second condition, they completed a location matching task in which they adjusted a visual marker in depth on the floor so that it was aligned underneath the target. We found accurate blind walking but a significant 10% overestimation in the marker position relative to the target position. Experiment 2 also placed targets on the ceiling and compared blind walking with a frontal matching task in which two poles were moved to set an exocentric width extent to match the egocentric depth extent from the observer to the target. Both measures showed accurate performance. Experiment 3 was the same as Experiment 2 but placed targets on the floor, again showing accurate performance for both measures. The results indicate two novel findings in need of further investigation: 1) blind walking remains accurate even when targets are not on the ground plane and 2) matching a location in depth leads to an overestimation of distance to targets on the ceiling that is not seen in the frontal matching-extent task.
This work was supported by NSF grants IIS-00-80999 and IIS-01-21084.