To see illusion, imagine ‘duck eats rabbit’

U. ILLINOIS (US) — Looking at a classic psychology experiment in a new way suggests that our visual systems are more flexible than previously thought.

One of the oldest experiments in psychology involves the famous duck-rabbit ambiguous figure first described by psychologist Joseph Jastrow in 1899 and based on earlier drawings.


Other ambiguous figures have been used in experiments over the years, yielding the same results produced with Jastrow’s figure. Psychology researchers Kyle Mathewson and Melinda Jensen of the Beckman Institute at the University of Illinois tried a new take on these experiments.

“While the duck/rabbit figure itself is ambiguous, people viewing the figure make a single interpretation,” Mathewson says. “They see a duck or a rabbit but not both possibilities at the same time. Later examples of similar ambiguous figures confirm this distinction; while the figure is ambiguous, our perception is not.”

The new study began following a lab meeting in which they found that a prompt can allow people to see both figures as distinct at the same time, a discovery that provides some interesting insights into the human visual system.

They wrote about their project investigating ambiguity in visual perception for the journal Perception.

Just as “most observers cannot see a single figure as both a duck and a rabbit, most cannot see one figure in a pair as a duck and the other as a rabbit even though the two figures and their features are spatially distinct,” the researchers write. “Is this inability to see both interpretations an inherent limitation of the visual system or is it just due to differences in top down processing?”

Using a simple prompt “immediately allows people to see both interpretations, to their own surprise.”

What was the prompt? Jensen discovered it at the lab meeting, when the duck-rabbit figures were presented to the group by Beckman researcher Alejandro Lleras.

“Someone drew them on the board and Melinda said ‘Oh, I can see one as a duck and one as a rabbit.’ But none of us could,” Mathewson says. “She said, ‘picture one eating the other’ and then we all could see it.”

To test this new perspective on an old experiment, Mathewson and Jensen devised their own experiment using the duck/rabbit ambiguous figure, as well as figures from 1915 (old woman/young woman) and 2006 (kangaroo/whale). They used 64 subjects who viewed a single ambiguous figure and were given both interpretations; the subjects then viewed pairs of the figures with relational phrases such as “imagine the duck is about to eat the rabbit” or “the kangaroo is about to bounce on the whale.”

The subjects were asked after each viewing of the pairs to attempt to see them as different figures at the same time. Providing the prompt, or cue, the researchers write, “allowed the majority of these non-seers to perceive opposing figures when they could not before,” with many subjects reporting an ‘aha’ experience upon seeing both interpretations.”

The results suggested an answer to the question they posed about whether the inability to see both interpretations was an inherent limitation of the visual system or due to top down processing differences.

“We think it’s probably more the latter; it’s top-down processing,” Mathewson says. “We’re allowing people to use higher level processes in a better way, in a new way. So it’s not that their eyes aren’t letting them see, or that your visual system doesn’t let you see a duck beside a rabbit, but that everything you thought about ducks and rabbits before didn’t let you see the duck and the rabbit. So we’re giving you a new lens to look at the figures.”

As for the theoretical implications, Mathewson and Jensen write that “the effectiveness of the relational cue illustrates the strength of top-down influences over our conscious perception” and that “our ability to see opposing interpretations highlights the flexibility of our perceptual systems. We are not limited either by bottom-up processing or by our top-down settings to make only one interpretation of the same kind of stimulus at the same time. . . . At least some viewers are able to see both interpretations.”

More news from the University of Illinois’s Beckman Institute: