Wednesday, May 10, 2006
Consider the figure above. It is standardly presented as an example of the "Horizontal-Vertical" (or "Vertical-Horizontal") illusion. (This particular figure is scanned in from a classic text on illusions, Stanley Coren and Joan S. Girgus (1978), Seeing is Deceiving (Hillsdale, NJ: Erlbaum), p. 29.)
My question is: How do we know if there an illusion here? Coren and Girgus state simply that the vertical line is seen as being longer than the horizontal line. Yet I'm not sure I experience it that way. Of course the current context of presentation is somewhat less than ideal -- for example there are text and borders nearby that may compromise the illusion. But even in ideal circumstances, not everyone will swiftly say that the vertical line looks longer. In this respect, the current illusion is unlike more dramatic illusions, like the Poggendorf illusion, that command instant assent in most viewers.
If people are given the chance to choose from among an array of cross-like figures, ranging from those in which the vertical line is obviously longer to those in which the horizontal line is obviously longer, they may err in choosing a figure with the vertical line too short -- suggesting illusion. Or if given the chance to adjust the lengths of the lines until they seem the same size, they may also err. But does this mean that in the normal case, when they are simply presented with a perfect cross and they say that it's a perfect cross and say that it seems to them visually that it's a perfect cross, that they nonetheless are experiencing an illusion? -- an illusion they can't report and that doesn't fool them? That strikes me as rather strange!
On the other hand, maybe it shouldn't be surprising if people are often inexpert in discerning their own visual experience. It may take a bit of effort to learn to judge accurately whether a visual illusion is present or not -- and certainly there are cases (this may be one) in which I feel unsure whether I'm experiencing a visual illusion. If I'm unsure, doesn't that suggest the possibility that I might be making a mistake?
(For further reflections on this issue, see Introspective Training Apprehensively Defended.)
Posted by Eric Schwitzgebel at 8:20 AM