Discrimination of the shape of the masked inducing figure precedes perception of the illusory triangle.
Using the Kanizsa triangle in a free-inspection paradigm, Gellatly has shown that subjects report seeing an illusory triangle while apparently not phenomenally aware of the inducing areas. It is argued that Gellatly's procedure may induce response processes which camouflage early sensory processes. By forcing subjects to choose between two response alternatives, it was shown that when they could just perceive the illusory triangle, they could also correctly discriminate the shape of the inducing elements. Under the free-inspection procedure, the duration threshold for seeing the illusory triangle was about 9 ms while the threshold for discriminating the shape of the inducing elements was about 2 ms. A static representation of the dynamic procedure used by Gellatly has been derived, which shows a novel effect of contour completion across boundaries resulting in a ghostly white triangle.