Reducing the Ponzo illusion by manipulating contextual information: Support for a contextual theory of visual illusions

John Gardai, 03jggard@alma.edu, Alma College
Anne Dorf, 03amdorf@alma.edu, Alma College

Gregory (1963) and Fisher (1967, 1968, 1973) have proposed competing theories of the perceptual processes behind the Ponzo illusion. Their theories involve inappropriate size-constancy scaling and contextual effects, respectively. Subsequent research has seemed to support Fisher's theory. The Ponzo illusion has been manipulated various ways in order to remove the depth information, and an illusion has remained. This would suggest that the contextual factors suggested by Fisher (1967, 1968, 1973) are at least a significant part of the illusion. This experiment attempted to create a Ponzo figure in which the apparent depth remained but the contextual factors that had previously given an illusion were removed. Fifteen students from a small, mid-western liberal arts college participated in the research by adjusting line segments on Eye Lines Software (Beagley, 1990). The modified figures greatly reduced the illusion, and may support Fisher's theory. The results may suggest the importance of contextual factors in visual illusions.

Two major theories have been proposed to account for the Ponzo illusion. The first was a size constancy mechanism proposed by Gregory (1963). Under his theory, the converging lines of the Ponzo figure suggest a three-dimensional spatial array with depth. A size constancy mechanism is therefore inappropriately invoked, resulting in the observed illusion. Since the upper line appears to be farther away, it is perceived to be longer than it really is. A second theory, however, was proposed by Fisher (1967, 1968, 1973). According to Fisher's theory, the Ponzo illusion is caused by contextual factors. The proximity of the upper horizontal line to the converging lines results in an assimilation effect and causes the horizontal line to appear longer than it really is. Subsequent findings have tended to support Fisher's theory.
This experiment was designed both to test the two theories and to ascertain the relative strength of each effect in creating the Ponzo illusion. By creating a stimulus with the normal depth information but no contextual effects, we hope to show which effect is stronger.

Methods

The participants were 15 Alma College students who volunteered to participate in the experiment. Each participant had normal or corrected to normal vision. The stimuli were presented on a computer running Eye Lines Software (Beagley, 1990). There were four sets of stimuli. (figs. 1-4)


Figure 1. Control


Figure 2. Standard Ponzo


Figure 3. Modified Ponzo


Figure 4. Double Ponzo

Results

In the control condition, the adjustable line was adjusted to a mean of 45.98mm, giving a difference between the comparison line and the adjustable line of 0.98mm. The standard Ponzo condition had a mean difference of 2.623mm, the double Ponzo condition had a mean difference of 0.562mm and the modified Ponzo condition had a mean difference of 0.261mm. (See figure 5) Adjusted for the error found in the control condition, the magnitude of visual illusion was 1.643mm for the standard Ponzo condition, -0.418mm for the double Ponzo condition, and -0.719mm for the modified Ponzo condition.



Figure 5. Magnitude of illusion for each condition

Discussion

In this preliminary investigation, we evaluated the strength and effect that different manipulations of the Ponzo figure had on the magnitude of the illusion. The double and modified conditions both created a negative illusion when compared with the control. This suggests that contextual factors play a more significant role in the Ponzo illusion than apparent depth. The findings in this study demonstrate support for Fisher's contexual theory.



More Posters

Web Posters from other courses and projects

Alma College Psychology Department