Abstract
The watercolor illusion (WCI) was first demonstrated by Pinna and colleagues in 2001. Since then, the strength and spread of the WCI have been tested in numerous static conditions, most recently by Hale and Brown (2021). The aim of the present study was to determine if the WCI could be replicated in spatiotemporally dynamic stimuli, and if so, to gain insight into how global context and local surface and edge properties might constrain or enhance such spreading. To accomplish this, we measured the strength and spread of the WCI using motion-induced contour (MIC) stimuli first described by Klymenko & Weisstein (1981) and recently revisited by Erlikhman & Caplovitz (2018) and colleagues (2019). When an outlined cube is rotated back and forth in depth minus the line representing the front corner, an illusory corner is perceived. Removing the lines representing the cube’s sides and top edges creates a simpler MIC stimulus, two chevrons with illusory contours visible along their sides perceived as two rotating illusory surfaces and the MIC. To test for a moving WCI, simpler versions of MIC stimuli were constructed with purple lines and an inner orange fringe along all or part of the purple lines. We also tested simpler versions where the orange fringe extended out alone, past the purple lines, creating local regions where the WCI should not occur. Participants rated WCI magnitude on a 7-point scale, then indicated where orange color was perceived. For all versions, participants reported the WCI between the orange fringe bordered by purple lines, establishing a moving WCI. Interestingly, the WCI was perceived between the orange fringe in versions where it extended beyond the purple lines, suggesting that the global context influenced the color filling-in to the point of overriding the necessity of adjacent purple lines, contrary to the WCI literature.