Variable and constant errors of perceived angle size

Abstract

Experiments were carried out to characterise the perception of angles in terms of the acuity and constant error of comparisons of acute angle sizes. Both measures, determined by the forced choice method of constant stimululi, were found to increase linearly with stimulus angle size. Constant errors varied systematically with stimulus orientation, following the oblique-effect (Apelle, 1971), while acuities did not. The differential expansion of acute angles was found to decrease with increased stimulus duration, stabilising after about 0.5 second.
Despite the previous success of the hypothesis that perceptual expansion of acute angles is an orientation contrast effect due to lateral inhibitory interactions between channels selectively responsive to different orientations, the present observations proved- inconsistent with predictions derived from this hypothesis concerning both acuities and constant errors, and temporal variation of these measures. Results of adaptation and masking experiments also failed to show meridional anisotropies of the selectivities of orientation channels, which were considered necessary assumptions for the explanation of meridional variation of perceived angular extent by the lateral inhibition hypothesis.
The discrepancy between the present results and previous observations which were consistent with the orientation contrast hypothesis was attributed to the fact that in the majority of previous studies perceived orientation was measured, not perceived angular extent.
The present data, therefore, do not contradict the orientation contrast model, but suggest that this contrast is not a sufficient explanation of the misperception of angular extent. As an alternatives, it was proposed that variation of perceived angular extent results from meridional anaisotropies in the scaling of an orientation metric derived from the integration of outputs from orientation selective channels.