The Müeller-Lyer and Ponzo Illusions. In each of the figures at the right, the two horizontal lines are the same length.  When shown Figure 1, the vast majority of viewers (even those who are familiar with these types of illusions) will report that the top line appears shorter than the bottom line.  When shown Figure 2, viewers tend to report that the top line appears longer than the bottom line.

Research Methods. Most people have seen these classic line illusions, but unless you've studied perception you may not be familiar with how they're approached in the laboratory or the theories used to explain them.  One method psychologists use to objectively measure the size of the illusory effect is to have subjects adjust the length of one line (the 'test') until it appears to match the length of the second line (the 'standard').  The amount by which the test line deviates from the standard (e.g., in pixels or mm) provides a quantitative measure of the effect.  Illusions are errors in perception, so an illusory effect is a measure of error. If there is no illusory effect, about half the subjects in an experiment should produce positive errors, and half should produce negative errors.  Or, as is often the case in practice, each subject should make about an equal number of positive and negative errors across many trials.  Thus, an illusory effect is demonstrated when there is evidence that subjects systematically make errors in one particular direction.

Size Constancy and Apparent Depth. These illusory effects are robust across practice, meaning it's difficult to learn how to overcome the illusions, but they do vary somewhat across cultures--and this has important theoretical implications (discussed later).  A popular theory for these (and other related illusions) is based on the idea of apparent depth.  In the Ponzo illusion (Figure 2) the two angled lines induce a sense of depth through linear perspective, a trick routinely used by artists to give the impression that something (such as a road) is receding into the background.  The image these two slanted lines subtend on the eye is precisely what the eye would see if two parallel lines really were receding into the background.  Linear perspective is a simple but powerful cue to depth information in the real world, so the theory starts by assuming that the two angled lines induce a subtle representation of depth in the visual system.

Fig 1. Müeller-Lyer illusion

Fig 2. Ponzo illusion

The error in line-length judgment arises because the visual system is simultaneously trying to solve another type of problem in the display.  The two horizontal lines in Figure 2 subtend exactly the same image on the eye, yet the top line is judged to be farther away than the bottom line, because of the apparent depth induced by the two angled lines.  In the real world, if two objects appear to be the same size but one object is known to be farther away, then the far object must actually be larger.  This is because the size of an image on the eye is inversely related to distance from the observer.  Thus, in order to maintain the apparent depth in the display and still account for the information in the retinal image, the top line is perceived as being larger than it appears to the eye.

That the Müeller-Lyer illusion also arises because apparent depth is taken into account when judging line length is less obvious.  Where is the apparent depth? Again, the apparent depth is induced by linear perspective.  [Tilting your head while looking at Figure 1 may help.]  It has been suggested that the upper configuration resembles the external corner of two walls (what you would see if you were looking at one edge of a room from the outside).  The lower configuration resembles the internal corner of two walls (what you would see if you were looking at the same edge from inside the room).  The idea here is that people in modern 'carpentered' cultures are exposed to so many examples of these geometrical configurations that they automatically evoke a representation of depth.  Using the same reasoning applied to the Ponzo illusion, the lower 'edge' is interpreted as extending away from the observer, whereas the upper 'edge' is interpreted as jutting out toward the observer.  Both lines subtend the same retinal image, so the farther edge must be interpreted as being longer to account for why it has the same length as an edge which is closer.

The idea that line-length is misperceived because of apparent depth may seem too obtuse an explanation for the Müeller-Lyer illusion.  However, I'm aware of at least one study (I'm sure there were more) showing a reduction in the Müeller-Lyer effect with people who were native to regions where carpentered buildings (and straight edges) were scarce.  Certainly a curious finding in its own right.

These illusory phenomena are more than just interesting visual tricks; they reveal something about the behavior and structure of the human visual system, and therefore can be used to better understand the nature of vision and perception.  Thus, psychologists and vision scientists have tried to explain these illusions at a much lower level--in terms of sensory mechanisms within the visual system itself (such as receptive fields and simple feature detectors).  Some psychologists find it useful to model perception as an active problem-solving process, in which the observer uses stored information to generate and test hypotheses about the objects and events encountered in the physical world.  The late Irvin Rock summed up this idea beautifully in the title of his book, "The Logic of Perception."

The proper APA citation for this article is:

DeCaro, S. A. (2003). Classic line illusions and the depth hypothesis. Retrieved [Month Day, Year], from http://psychology.sdecnet.com/illusion.htm.