Perception of relative velocity: a revision of the hypothesis of relational determination.

The conventional formulation of the hypothesis of relational determination asserts that the perceived relative velocities of two or more objects depend upon their relative rates of displacement with respect to stationary referents in the visual field. Experiment 1 showed that this formulation is too restrictive by demonstrating the transposition-of-velocity effect under conditions in which two light spots moved in the absence of static visible contours and traversed unequal path lengths. Experiment 2 showed that angular velocity per unit of relative angular extent, and not relative linear or perceived extent, of the respective motion paths influences perceived relative velocity in nonarticulated space. The retinal dimensions of static visible frameworks were shown in experiment 3 to influence perceived relative velocity in a direction consistent with the conventional formulation of the relational-determination hypothesis, but the weight of the evidence suggested a reformulation along the following lines: the perceived relative velocities of two objects are significantly affected by the proportions of the retinal projections of the respective movement fields traversed by the two objects in the same unit of time, even when the motion fields consist only of the objects' motion paths.