Further studies in the perception of a changing shape

Koffka's hypothesis of “invariant relation” holding between tilt and perceived shape of an object is discussed from the standpoints both of logical tenability and of experimental findings.

After discussion of certain shortcomings in the method of investigation devised by Stavrianos, two experiments are suggested, the first, of a simple character, intended to follow out in practice the hypothetical conclusions of Koffka's argument, the second to test the hypothesis under conditions which attempt to ensure the preservation of the perceptual object throughout the arc of tilt. For this purpose subjects attempt to match two shapes, one of which is stationary whilst the other revolves at a constant speed. The comparison shapes cover the range of the possible arc of tilt.

The results of the first experiment, whilst appearing to contradict the requirements of the hypothesis, are by no means conclusive. The second experiment, though not absolutely conclusive, gives little support to Koffka. The main conclusion drawn is that there occur non-systematic variations in perceptual constancy throughout the arc of tilt. These are of a phasic character and are qualitatively different from results obtained by matching stationary shapes. There appears to be no simple function which would result in exhibiting a uniform constancy value for the entire arc of orientation.