J. Consc. Studies 7, No. 8/9, 2000, special issue Art and the Brain, part 2,
ed. J. Goguen and E. Myin.

Review of

Visual Space Perception: A Primer,
Maurice Hershenson (MIT Press, 1998)

by Joseph A. Goguen


This book is a good survey of a large body of important work in the experimental psychology of visual space perception. Although much has been left out, the quality of author's selection is a strong point of the book. I particularly enjoyed the material on Gibsonian psychophysics, which is still unusual in textbooks. I think workers in many fields could benefit from this book, not least those in consciousness studies, philosophy, and art, many of whom may be weak on the facts of vision.

I would have liked to see a little more on the historical development of the subject (for example, Wilhelm Wundt is not mentioned at all, and Hermann von Helmholtz gets only a parenthetical citation), but this would take up more space, and is perhaps asking too much of the author, since the history of a science is a rather different discipline from the science itself. Understandably, there is nothing on color vision, on art as such, or on consciousness. But there is an extensive bibliography, and enough interesting technical material to keep the serious reader busy for quite some time.

While I applaud Hershenson for using Gibson, I also consider that the way he has used Gibson is a bit too limited. The nature of the problem is illustrated by the very first sentence of Hershenson's book, which says

The study of visual space perception begins with the assumptions that the physical world exists and that its existence is independent of the observer.
Although Gibson would have accepted these assumptions (Gibson 1977, 1979), I think he would also have agreed that they are not the best place to begin the study of visual phenomena. So despite its laudable treatment of several Gibsonsonian topics, this book does not take a Gibsonian perspective. Readers of this volume of JCS will have easy access to the introductory essay by Eric Myin (2000), which contains an excellent short review of Gibson's anti-representationalist approach to vision, which questions the interpretation of many classic experiments, because they produce artificially fragmented perceptions.

My plan for this review is to go further than Gibson himself went, and then use this as a basis for re-examining the starting point of Hershenson's book. While I consider this an exploration of some deeper implications of Gibson's work, I have little doubt that many readers will consider it a significant reinterpretation. I would however emphasize that a strengthened Gibsonian perspective does not undermine the experimental results in Hershenson's book, but only challenges some of their interpretations, especially at the higher, more philosophical levels. Nor do I wish to deny that there is a physical world that is independent of the observer, but instead would advocate trying to proceed without that assumption. One can think of this as a phenomenological approach, although I actually learned this way of thinking from my study of Buddhism; however, I will try to bring out the ideas using a language that is more familiar to students of consciousness.

Let's start with Gibson's notion of "affordance," which can be seen as an ingenious bridge over the infamous subject-object gap, precisely by not assuming the prior existence of a world of objects (or its necessary complementary assumption of a knowing subject). Using a language that is still somewhat dualistic, we might define an affordance to be

a capability for a specific kind of action, involving an animal and a part of its environment.
For example, a cup provides an affordance for drinking. An affordance consists of the sensory-motor schemata (to use the suggestive terminology of Piaget, though with a different flavor) that support its particular class of actions, such as drinking coffee from a cup. One way to see the importance of this, is to notice that instead of constructing, storing, and then using (massively large and complex) representations of objects, such as cups, from the external world, we rely upon the world to provide us with exactly the information we need, exactly when we need it, in order to carry out actions, such as drinking coffee; this can be seen as an "offloading" of representation into the world, providing an enormous gain in efficiency, as discussed in some recent AI research [[1]]. (Rudolph Arnheim (1969), who is also not mentioned in Hershenson's book, may be considered another precursor in merging subject and object).

It is clear that subjects can never know objects "as they are," but only partially, at a particular time, from their own perspective, based on their own capabilities and prior experience. The knowledge associated to an affordance cannot be captured by passive predicates, and even the currently popular paradigm of active perception does not capture its essence; both perception and action are better understood in terms of affordances. Affordances are tacit in an essential way: no one can describe accurately and in detail how they actually drink coffee from a cup, but given a cup with coffee, they can do it. Of course, there can be many different affordances for a cup, and even many different affordances for drinking coffee from a cup. But they are all embodied, not just in individuals, but also in particular social and physical circumstances; for example, we may drink differently from a very hot cup than from a cold one, and differently at a very high-toned party than at breakfast at home.

An affordance only has real existence as a relation between a particular subject and a particular object, at a particular time and place; it does not make sense to separate the subject and object from each other [[2]], or from the relation, because each only shines forth in view of its participation with the others in that particular occurrence. But it does make sense to take the notion of affordance as free-standing, without reference to either subject or object. Gibson (1977) wrote

An affordance cuts across the dichotony of subjective-objective and helps us to understand its inadequacy.
But once we have taken this step, we may also begin to have doubts about the notion of an autonomous subject, who exists independently of the world. For is it really clear that there is some core identity that persists through all the changing affordance relations? Or is what we call the subject just a place holder in those relations?

It was Martin Heidegger (1927) who pioneered the discussion of human being (which he termed "Being-in-the-world") without the subject-object split. His well-known notion of "ready-to-hand" is a close relative of Gibson's affordances, more sophisticated philosophically, but with less empirical content. What I wish to suggest is the desirability of bringing together the perspectives of Gibson and Heidegger.

Finally coming back to space perception, it seems that an approach like that sketched above could have significant implications for spatial perception. A starting point would be the observation that seeing involves the deployment of complex sensory-motor schemata that embody important learned relationships; for this reason, spatial perception should be understood in terms of the affordances that are involved in actually doing spatial perception. This of course is Gibson's own approach.

Despite Hershenson's less than fully convincing attempts at neutrality between (weak) Gibsonian and traditional representation-based theories, I think a thoughtful reading of this book will convince most JCS readers that Gibson was on a promising path, and I hope that this review will also persuade them that it would be worthwhile considering what it might mean to go further down that path than this book will take them.


References

  1. Arnheim, R. (1969), Visual Thinking, Univ Calif Press.
  2. Brooks, R. (1991), Intelligence without Reason, Proc 1991 International Joint Conference on Artificial Intelligence, pages 569-595.
  3. Brooks, R. (1991a), Intelligence without Representation, Artificial Intelligence 47, pages 139-159.
  4. Gibson, J. (1977), The Theory of Affordances, in Perceiving, Acting and Knowing: Toward an Ecological Psychology, ed. R. Shaw and J. Bransford Erlbaum.
  5. Gibson, J. (1979), An Ecological Approach to Visual Perception, Houghton Mifflin.
  6. Heidegger, M. (1927), Being and Time, trans. J. Macquarrie and E. Robinson, Blackwell, Oxford, 1962.
  7. Myin, E. (2000), Two Sciences of Perception and Visual Art, J Consciousness Studies 7, no. 8/9, pages 47-59.
  8. McDermott, D. (1987), A Critique of Pure Reason, Computational Intelligence 3, pages 151-160.
  9. Nehaniv, C. [ed] (1998), Computation for Metaphors, Analogy, and Agents, Springer, Lecture Notes in Artificial Intelligence, vol. 1562.

Footnotes

[[1]]. See articles in (Nehaniv 1998), especially those by Brooks et al., by Dautenhahn, and by Beynon; see also the classic articles (McDermott 1987), (Brooks 1991) and (Brooks 1991a).

[[2]]. In Buddhism, this is summarized in principles called "egolessness of self" and "egolessness of dharmas," or "emptiness of self" and "emptiness of other," and similar phrases.


To my Introduction to Art and the Brain, Part 2
To Eric Myin's Introduction to Art and the Brain, Part 2
To the full table of contents and abstracts for Art and the Brain, Part 1
To my Introduction to Art and the Brain, Part 1
To the Journal of Consciousness Studies homepage
To my homepage