The Perception of Pictures - Part 1 - Ecological Optics?


[Vermeer - Delft Street]

I have been reading for a course on Perception this term and have found myself enamored of the theory of James J. Gibson, a psychologist who proposed a radically different theory of perception -- The Ecological Theory. In particular, I have been reading his account of pictures -- what they are, how they are perceived and what they can/should do. I want to work through some of these findings here.

And since I don't really want to detail Gibson's general theory of ecological optics, here is a brief summary by Robert Schwartz in his introduction to Gibson's work in the2004 text Perception:

Gibson's criticism of the tradition goes much further than that found in Gestalt writings. He argues that it is necessary to: (1) replace talk of 'sense organs' with that of 'perceptual systems,' (2) abandon all explanatory appeal to, if not the very idea of, 'sensations,' and (3) reconceive the notion of a 'stimulus.' Stimuli should be thought of in terms of the information contained in higher-order invariant properties of arrays of ambient light, particularly those that result from movement. Adopting this framework, Gibson believes, dissolves the traditional problem of explaining how the mind turns impoverished sensations into rich perceptions. The information in the stimulus is not impoverished.

Schwartz later continues his summary:

Although there were always serious dissenters, the consensus in vision theory well into the twentieth century was that perception of the spatial layout was, in one way or another, a multi-stage process. [...] According to Gibson once the concept of 'stimulus' is properly expanded, it can be shown that the invariant properties of arrays of light contain all the information needed to perceive the layout veridically. The stimulus, in and of itself, accurately reflects the way the world is. There is no need to derive or infer environmental properties from sense-registered cues. We immediately and directly see how things are without the aid of intervening mental states and processes.

Gibson allows that in many instances we may have to learn to see. The expert can see things and properties the untutored cannot discern. Learning, though, is not a matter of bringing in past experience to alter or enrich the interpretation of the sensory data. Perceptual learning is the honing of an ability to pick up directly the information in the light array (Gibson and Gibson 1955).

To conclude the briefest of summaries, Gibson is entirely opposed to every establishment theory of how visual perception happens. He rejects the sensation-perception distinction, he rejects the notion of unconscious inference, he rejects the traditional focus on the retinal image (citing the inevitable homunculus theory that falls out from it), and he rejects the notion of perception as the result of some background or underlying processing.

So why does this interest me? I think Gibson gets so much right that I want to see how far he can actually go with his theory. He must provide an account for what we perceive when we perceive a picture, and I want to see what he says about that. I also want to see if I agree with him. What I think comes out of his discussion of pictures is a very interesting and oft-ignored confusion that occurs in discussions of aesthetics. Gibson's theory shows that much of what has been said in aesthetics and art history about what a picture is and how we perceive it is wrong. I also think that many of Gibson's critics misunderstand what he tells us about the artist's effort and the resulting picture. I want to now work through this mass of debate.