Notes on J.J. Gibson Part II [posted 3/22/00]


Notes and thoughts on J.J. Gibson's Ecological Approach to Visual Perception. Part II

        In the second half of the book (part 3 on) Gibson presents an overview of some of the experimental work he and others have done which support his theories and raise some interesting questions. He elaborates and develops the concept of "affordance" and the pickup of invariant information from the optic array. After thoroughly re-arranging the perspective on perspective, he goes on to present a theory of 2D representation, and of movie editing.

        The experiments are intriguing, although in most cases it would be nice to have had a little more detail in methodology and data, but I guess that is not really within the scope of this book. He continues to stress the importance of considering the perceptual system as a whole, not just an hypothetical stationary camera-eye. I found his descriptions of phenomena vivid and powerful, and most of all consistent with my own visual experience. I feel like I actually see things slightly differently now.

        And in going back to type up these notes, I'm realizing that I probably only got about half of the book. I should read it a couple more times...


p. 156 Gibson is giving some examples of interesting experiments which he claims show that we do not really have depth perception in the sense which it is usually described. The distinction is subtle and I'm not sure I fully get it. To me the experiments indicate a fairly strong understanding of depth and the and the assignment of 3D properties (even when they don't exist). So far it seems that although his explanations are interesting (and consistent with the experiments) they still don't rule out more traditional explanations.

I think the distinction he is making is that we don't actually perceive the 3RD dimension directly. Infact, thinking about perception in relation to traditional Cartesian coordinates is misleading. We don't see things in 2D either. What we perceive is surface layout. We cannot see depth, but we can understand the spatial relationships between things from how they are arranged and how they change when we move.

In many ways this is so obvious that I've never noticed it before. If we actually saw space equally in all directions, than this diagram probably wouldn't make any sense.
(it is, after all, flat on a 2D screen). We certainly understand space, but Gibson's optic array seems like a pretty good explanation of what we see.

p160. "Distance as such cannot be seen directly but can only be inferred or computed. Recession along the ground can be seen directly." "These invariants are not cues, but information for direct size perception." (discussion of stake height estimates in a plowed field)

p162. Gibson's (illuminating) hypothesis that a major component of visual size estimation is textural occlusion seems like it could be tested. An artificial environment could be constructed where the textural elements did actually vary in size as they approach the horizon. Would there be corresponding changes in the estimation of objects' sizes?

Constructing a real physical environment could be a little time consuming. Perhaps the experiment could be repeated in a virtual computer environment? (this might raise some other interesting questions..)

p166 discussion of "mediated perception theory" -includes the trapezoid room illusion I mentioned earlier.

p171 "The stimulus information for motion is the change of pattern, and the information is the same for an intermittent change as for a continuous change. "

p178. The deformations of the elastic array in this experiment may not be the same as those of a tilting array. (there may be some sideways movement of elements to distribute tension in the grid they were streaching)

p195 "...although one can become aware of the seen-now and the seen-from-here.... what one perceives in an environment that surrounds one, that is everywhere equally clear, that is in-the-round or solid..."

p196 Comment about learning to see the world as a picture. Gibson's ideas might explain why it is so difficult for most of us to actually draw a 2D representation of the 3D world - a task which ought to be easy if we were looking at a screen in our heads.

- interesting to use children's and untrained drawings as a means of getting at people's perception

p198 Concept of vistas opening up as a person moves along a path through a place. "One vista leads to another in a continuous set of reversible transitions" "... Each vista is its own 'landmark'..."

I don't agree with (or maybe don't understand) his dismissal of the cognitive map, although it is definitely not the only means of locomotion/wayfinding. I would argue that it is a higher order structure. - I like his description, but I'm not sure how it supplants a cog. map.

-Wayfinding as a "mapping" from vistas to a cognitive construct?

200 "...same invariants under optical transformations..." Implication that Invariant is approximately equal to Topological equivalence?

-Identity ( in the mathematical sense) can be shown by application of transformations? Principal of Reversibility implies transformations have inverses which can be applied to regain the original set.

p206 Clear elaboration between visual field and visual world:

"When I distinguished, years ago, between the visual field as one kind of experience and the visual world as a radically different kind (Gibson, 1950b, ch.3) , I was elaborating on Koffka. The visual field, I suggested, consists of a patchwork of colors something like a picture, whereas the visual world consists of familiar surface and objects one behind another. The visual field has boundaries, roughly oval in shape, and it extends about 180 degrees from side to side and about 140 degrees up and down. The boundaries are not sharp, but they are easily observed when attended to. The visual world, however, has no such boundaries; it is unbounded, like the surface of a sphere extending all the way around me. The visual field is clear in the center and vague in the periphery - that is, less definite towards the boundaries - but the visual world has no such center of definition and is everywhere clear. The oval boundaries of the visual field sweep across the array whenever I turn my head and wheel over the array whenever I tilt my head, but the visual world is perfectly stationary and always upright. The patchwork of the visual field deforms as I move and , in particular, flows out ward from a center when I move in the direction of that center, but the phenomenal surfaces of the world are always perfectly rigid. "

p.258 "Perceptual seeing is an awareness of persisting structure" "The abstracting and perceiving of invariants are what happens in both perceiving and knowing."

p.258 talking about transmission [culture traits] "But they are not in themselves knowledge, as we are tempted to think. All they can do is facilitate knowing by the young."

- I've realized that I've been confused by the term "invariant" as it seems to imply absolutes and Platonic Ideal Forms. Now I'm starting to think about it as a fuzzy variable, a meaning cloud. The invariant can just be the object being looked at

p253. "Knowledge of the environment, surely, develops as perception develops, extends as the observers travel, gets finer as the learn to scrutinize, gets longer as they apprehend more events, gets fuller as they see more objects, and gets richer as they notice more affordances. Knowledge of this sort does not "come from" anywhere; it is got by looking, along with listening, feeling, smelling, and tasting. The child also, of course, begins to acquire knowledge that comes from parents, teachers, pictures, and books. But this is a different kind of knowledge."

- as if he is saying that the mind isn't linked to the perceptual system, it is the perceptual system. Not just visual objects, but ideas, relationships can be perceived. Perhaps we are perceiving machines rather than meme, or gene machines? - yehhck, getting to philosophical

p254. I thought his analysis of time and memory in perception was very nice. Perception is smeared out over time. It doesn't always end when the stimulus stops, and we often fail to perceive something for quite a while after its stimulus is available. It is not yet possible to draw a clear line between perception and memory. He doesn't really present a clear answer, just some clear thinking.

p255. "Note how the definition includes within perception a part of memory, expectation, Knowledge, and meaning - some part but not all of theses mental processes in each case."

p263. Gibson's ways of knowing:
-Direct Perception
-By means of instruments, tools
-by means of language
-by means of pictures

by means of tradition?

------------- Dang, I guy could write a whole thesis just on ideas brought up in this book!









Commentary: -none-

So far, nobody has had anything to say about this text. If you would like to comment, please email skye.