Two theories have been put forth to explain the Necker Cube reversal illusion. Kohler and Wallach ( 1 ) introduced the satiational theory and Rock ( 2 ) purposed a higher order, cognitive view wherein a constructional process mediates this illusion. A total of 51 university students were tested in three experimental sessions, separated by 1-day periods, in a repeated-measures design. Each session involved watching a Necker cube for a 2-min. period with subjects recording the number of reversals experienced. One session involved the plain cube, another had a fixation mark in the cube's center, and a third had a mark which appeared for 30 sec. at each of four positions which were %-in. off center and located on the vertical and horizontal axes of the cube. The mark moved from top location to bottom to left and to right. The order of session presentation was counterbalanced. Students were informed as to the nature of this illusion and were asked to take a passive attitude. Each cube had a 4-in. square front and back with 2-in. long diagonal sides which made a 50 angle with the horizontal base and top of the front square which was done and to the left of the back square. Each cube was projected from behind onto the center of a translucent viewing screen via an 8-mm motion picture projector set at 24 frames/sec. The screen was 18 in. from the student's nose which intersected the central axis of the screen approximately at right angles. If a satiational mechanism is operative, the highest number of reversals should be reported with fixation mark, an intermediate number with plain cube, and the lowest number with sequential fixation. If a purely cognitive, constructional process is involved, the experimental manipulations should have no effect on rate of reversal between the two alternate percepts. A Latin square analysis of number of reversals in the three 2-min. sessions for the 5 1 students yielded nonsignificant effects of group, order, and interaction of the viewing condition X order. Pooling like viewing conditions, a single-factor, repeated-measures analysis of variance showed viewing conditions were significant (Fz.la, = 3.3, f i < .05). Newman-Keuls analysis indicated a significant difference between no fixation ( M = 29.28, SD = 14.08) and sequential fixation ( M = 25.78, SD = 15.11), with fixation ( M = 27.22, SD = 14.98) in between. Such results are inconclusive with respect to the two views as was prior work ( 3 ) . However, number of reversals reported between viewing conditions significantly correlated ( p < .05) .76 for no-fixation and fixation, .83 for noand sequential-fixation, and .76 for fixation and sequential fixation. In a pilot smdy ( N = 51) of the same design, but with only two sessions separated by a 1-mo. period. r was .76. The mechanism of this illusion operates at a fairly constant rate for each person across viewing conditions whether separated by 1-day or 1-mo. periods. This fairly stable rate may indicate the involvement of a lower-order mechanism (satiation) under passive and informed instructions. A higher order mechanism might be expected to fluctuate more in its operation on different days, depending more on cognitive and motivational components.
Read full abstract