Abstract
It has been proposed that there are separate pathways for visually guided action and for visual awareness of the world (Goodale & Milner, 1992). The former may depend on body-scaled informative structures. Distance, for example, is specified by the horizon-distance relation (Sedgwick, 1973) which relates the observer's eye-height to the angle subtended from the base of the target to the horizon. We compared verbal distance estimates to the action measure of walking to the same targets by blindfolded observers. Observers wore either 10 diopter base-up or base-down prism glasses (or none) which shifted observers' angle of gaze elevation from normal while leaving visual angles themselves unchanged. If visual angles are related to the body of the observer (through eye height and oculomotor information for gaze inclination, Grutzmacher, Andre & Owens, 1997) then distances should appear longer in the base-down and shorter in the base-up condition. We obtained radically different results for the action measure compared to verbal estimates. Observers' verbal estimates of distances 30, 60 and 120 feet were short, and accuracy was not affected by distance: estimates were 45, 45.1 and 47.4% of the true distances, respectively. In contrast, observers came very close to the target distances when walking, and accuracy was significantly affected by distance: observers walked 109.4, 96.8 and 84.9% of the true distances. The obtained pattern of results for the prisms was consistent with body-scaling of horizon-related visual information although the effect was very small on both measures, and not significant. (Target objects were far and relevant visual angles were small.) In conclusion, action measures of perceived distance are more accurate than verbal reports, and accuracy is affected by distance to the target. We suggest that this discrepancy is further evidence of separate visual pathways for visually-guided action and for the verbal report of perception.