I find it more than merely suggestive that we call many different kinds of pictures "realistic." As a category label, "realistic" is remarkably elastic. We cheerfully place into the category pictures that are made in strict accordance with the rules of linear perspective, pictures that are at slight variance with those rules but that nonetheless look perfectly "correct" , and pictures made in flagrant contravention of perspective geometry . We accept as realistic pictures that are made in strict accordance with (...) the rules of perspective construction that we could never judge as being similar to anything we might or could ever see . We accept as realistic pictures that are in sharp disagreement with what we now take to be the facts of vision . . . There is something charming and yet nasty about the belief in the special relation of picture to world. It is charming because it allows us to "enter" with ease into pictures and allows them to "extend" into our world. It allows us to think of pictures as "true to life," to use [Ernst] Gombrich's beguiling term, to look at a picture and ask questions of it, as if we were looking at the world through a window. It allows us to treat pictures as substitutes for the objects they represent and so, for example, to buy clothing from an illustrated catalogue, or to analyze architectural styles from pictures of buildings. In brief, it allows us to feel proximity to what is depicted and urges us to conclude that in certain important respects looking at a picture is equivalent to looking at what is pictured. Joel Snyder, associate professor of humanities and of art and design at the University of Chicago, teaches aesthetics, and theory and history of photography. A practicing photographer, he is currently completing a monograph on the photographer Timothy H. O'Sullivan. His contributions to Critical Inquiry include "Photography, Vision, and Representation" and "Reflexions on Las Meninas: Paradox Lost", written with Ted Cohen in the Winter 1980 issue. (shrink)