Configural effects in perceived pointing of ambiguous triangles.

Five experiments examined the influence of configural variables on perceived pointing of ambiguous (equilateral) triangles. The amount of time required to see the triangles point in specified directions was measured for single triangles and linear configurations of multiple triangles. The results show that perceived pointing is biased toward directions that are coincident with or perpendicular to the configural line. Perceptual interference occurs when the configural line biases pointing away from the correct direction, and facilitation sometimes occurs when it biases pointing toward the correct direction. The magnitude of interference increases (a) when the required direction is perceptually less salient, (b) when the configural line contains more triangles, and (c) when the triangles are closer together. Further results suggest that the spacing effect depends on relative rather than absolute distance and that the number and spacing effects are additive. In addition, substantial differences were found among individuals in terms of their susceptibility to configural interference. These results are discussed in terms of a sequential sampling model of perceived pointing. The samples result from spontaneous alternations in a multistable system of three mutually inhibitory directional subsystems. It is proposed that configural effects arise from low resolution pattern information that influences the behavior of the multistable system. Possible mechanisms are suggested through which directional, attentional, and individual differences might also arise.