Abstract
To further demonstrate that orientation affects amodal completion (Srebotnjak 1984; Sekuler 1994; Sgorbissa & Gerbino 1999) we studied how variable-duration primes (100, 400, 1000 ms) affect matching of simple patterns (crosses and fish-like polygons). Four primes were utilized: a modal cross; a shape that could be amodally completed as a cross (global completion with maximum shape regularity) or as a fish-like polygon (local completion with good continuation at T-junctions); a modal fish-like polygon; a disk (unrelated to test patterns). According to the microgenetic approach (Sekuler & Palmer, 1992), amodal completion is favored only at long presentation times. Previous findings as well as the field model of visual interpolation (Gerbino & Fantoni, 2000) suggest that perception of a fish-like shape should occur only when orientation along cardinal axes supports the good continuation of T-junction stems. Primes and test pairs were shown in two orientations (× vs. +). A relative facilitation measure was defined to indicate an RT advantage for either global (crosses) or local (fish) test pairs. The effect of orientation on priming varied at different prime durations. Priming effects at short prime durations are consistent with detection and comparison of features, rather than on configural processing. Priming effects at long prime durations, when interpolation is complete, are consistent with the proposition, included in the field model of visual interpolation, that good continuation is more effective when T-junction stems are oriented along cardinal axes. In principle, the relative strength of local vs. global completions might depend on the differential effect of orientation on good continuation alone, regularity alone, or both. Present data do not allow us to reject any of such hypotheses.