Abstract
Three experiments tested effects of self-similar contexts on the visual and haptic T-illusions (when both lines of a T are equally long, the undivided one appears to be longer than the divided one and is also haptically indicated longer). Experiment 1 used patterns of 4 Ts, Experiment 2 used branching patterns in which 4 Ts had been embedded, and Experiment 3 used patterns of 4 triangles and 4 catenary-derived forms for which the T constituted a skeleton. All patterns were of symmetry group c4, and all were presented at their 2 canonical orientations (all figure or skeleton lines horizontal or vertical or all of them oblique at 45 deg). Figures were scaled in size by factorially crossing 3 lengths of the Ts lines, and patterns in Experiment 1 were also scaled in density. 3 independent samples of 12 observers each had to haptically indicate the lengths of selected target lines and verbally judge relative lengths of both types of lines of target Ts. The haptic illusion vanished with the branching patterns of Experiment 2, but the visual illusion persisted throughout, being stronger when the undivided line was target. Patterns as such and their density did not have great effects, suggesting that the causes of the illusions are local. Results from Experiment 3 suggest that in addition to interactions between orientation-sensitive neurons referring to the Ts T-junction, the illusion may be caused by a discriminative response to different plane angles or different curvatures referred to the Ts endpoints. The asymmetric visual illusion can possibly be explained in terms of priming.
Meeting abstract presented at VSS 2014