Abstract
Vision and haptics play a central role in perceiving environmental layout to guide future actions. Kinsella-Shaw, Shaw, & Turvey (1992) found that observers can perceive the maximal slope that supports upright standing. Hajnal, Wagman, Bunch, & Doyon (2014) demonstrated that visual perception of stand-on-ability is accurate compared to action capabilities; haptic perception of stand-on-ability reliably underestimates action capabilities. This finding contradicts Gibson's (1979) theory of equivalence in perceptual systems: perception should be equivalent regardless of modality. The current experiment offers a direct comparison between visual and haptic perception of stand-on-ability, i.e., one perceptual system is used to measure the other. Two groups of observers provided affordance judgments, confidence ratings, and matching judgments in two conditions: (1) visual perception and haptic matching of a visual stimulus and (2) haptic perception and visual matching of a haptic stimulus. In condition (1), observers viewed a surface set to a discrete angle then attempted to match it with a continuously adjustable haptic stimulus occluded by a curtain. In condition (2), observers felt a surface set to a discrete angle then matched it with a continuously adjustable visual stimulus. Results indicated that visual and haptic perceptions of stand-on-ability are indeed equivalent: no differences were found between visual, haptic, and action boundaries. Matching judgments were highly correlated, showed high internal consistency, and showed high veridical accuracy for angles that fall within the behaviorally relevant range. The current experiment offers support for Gibson's (1979) theory and contrasts with previous investigations of perceptual correspondence where haptic perception consistently underestimates the action boundary (Hajnal et al., 2014): the correlation in haptic matches is nearly identical to visual matches, but with an intercept offset. The multisensory nature of this task promotes equivalence where unimodal tasks as in Hajnal et al. perpetuate differences across systems revealing nonequivalence.
Meeting abstract presented at VSS 2016