Why We See What We Do: An Empirical Theory of Vision
Dale Purves and R. Beau Lotto
2003
260 pages, 105 illustrations
paper
About This Title
This provocative book reviews a broad range of evidence leading to the conclusion that the visual system is not organized to generate a veridical representation of the physical world, but rather a statistical reflection of the visual history of the species and the individual observer. Thus, what humans actually see is a reflexive manifestation of past rather than a logical analysis of the present. The idea that the images we consciously entertain represent the historical significance of visual stimuli follows from the inability to decipher ambiguous retinal information analytically, and has far-reaching consequences not only for vision but brain function generally. The immediate benefit of this approach is that it provides a framework by which to understand a variety of fundamental visual illusions that are otherwise difficult, if not impossible, to explain.
With its straightforward style, Why We See What We Do can be understood by individuals with little or no background in neuroscience or vision. It includes chapter introductions and summaries that make the overall argument easy to follow, over 400 bibliographic citations, and a complete glossary.
See an animated version of the cover image. (QuickTime, 1.9M)
About the Author(s)
Dale Purves is the George Barth Geller Professor for Research in Neurobiology, Professor of Psychological and Brain Sciences, and Chairman of the Department of Neurobiology at Duke University Medical Center. He earned his B.A. from Yale University and his M.D. from Harvard Medical School. His previously published books include: Principles of Neural Development (with J. W. Lichtman, Sinauer Associates, 1985); Body and Brain: A Trophic Theory of Neural Connections (Harvard University Press, 1988); Neural Activity and the Growth of the Brain (Cambridge University Press, 1994); and two editions of Neuroscience (written with collaborators at Duke, Sinauer Associates, 1997 and 2001). The focus of research in the Purves laboratory is visual perception and its neurobiological underpinnings, and, more recently, the neurobiological basis of music.
R. Beau Lotto is a Lecturer at University College London in the Institute of Ophthalmology's Vision Research Department. He earned a B.S. from the University of California, Berkeley and a Ph.D. in developmental neuroscience from Edinburgh University Medical School. Research in the Lotto laboratory combines ecological, behavioral, and computational neuroscience to investigate the general principles that describe the causal relationship between the past (experience) and the present (adaptation) in biological systems, focusing primarily on the enigmatic realm of color perception and behavior.
Reviews and Commentary
"Why We See What We Do ... will remind vision researchers of James Gibson's and David Marr's seminal efforts, and it may prove as influential."
—Vincent A. Billock, Science
"... they bring [the problem] to life using a series of computer-generated illustrations that delight the eye and edify the mind."
—Nature Neuroscience
Table of Contents
Preface
Acknowledgments
1. The Basic Challenge of Vision
The provenance of any light stimulus, and thus its significance for subsequent action, is unknowable directly. The reason is that any of the elements of a visual stimulus could have arisen from many—indeed, infinitely many—different objects and conditions. Since vision, to be useful, must deal effectively with the real-world sources of visual stimuli, the inherent ambiguity of the patterns of light that fall on the eye is the fundamental challenge for which the functional structure of the visual system evolved to solve.
2. Some Pertinent Facts about the Visual System
Many aspects of the structure and function of the visual system are necessarily raised in considering how the quandary of stimulus ambiguity is resolved. Understanding at least some of the basic features of visual anatomy and physiology is essential before considering in more detail the major visual qualities that humans experience, and how these qualities can be explained in terms of an empirical theory of vision.
3. Perceiving the Intensity of Light
A good starting place in exploring the merits of any theory about vision is the perceptions of lightness and brightness generated by the overall intensity (luminance) of light stimuli. These sensations are arguably the most fundamental of the visual qualities that humans and other animals perceive. Moreover, the discrepancies between the physical world and what we see are readily apparent in perceptions of lightness and brightness. The nature of these discrepancies provides a key to understanding how the uncertain relationship between stimulus and source is resolved.
4. Further Challenges to Rationalizing Brightness
To be convincing, any explanation of the percepts elicited by light intensity must also account for the diversity of well-known perceptual enigmas that entail luminance. Several key illusions of perceived light intensity are considered here. In each instance, the perceptual consequences of these intriguing stimuli can be rationalized in terms of the probability distribution of their past sources.
5. Color and Its Significance
Because the provenances of spectral power distributions are just as ambiguous as the sources of luminance, it is reasonable to imagine that perceptions of color are, like the perceptions elicited by the overall intensity of light stimuli, generated probabilistically according to past experience. To understand the merits and potential importance of this idea, it is necessary to discuss in more detail the way in which light of different wavelengths affects the nervous system.
6. Perceiving the Spectral Quality of Light
Like all of perception, a major difficulty in rationalizing color vision in terms of the known physiology and anatomy of the visual system is the remarkable influence of context on the colors we see. The evidence described in this chapter suggests that these contextual effects derive from the empirical significance of spectral stimuli, in much the same way that perceptions of lightness and brightness appear to be determined by probability distributions of the sources of light intensity.
7. Perceiving Spatial Relationships
In addition to conflating the physical parameters that determine the quantity and quality of light reaching the eye, the parameters that define the location and arrangement of the sources of light are also inextricably intertwined in the retinal image. As a result, the spatial relationships of the sources of visual stimuli, like the sources of luminance and the distribution of spectral power, are always uncertain. This further aspect of stimulus ambiguity also appears to be resolved empirically.
8. Perceiving the World with Two Eyes
It has been known since the mid-19th century that binocular information can be used to generate a special sense of depth called stereopsis. This chapter considers evidence that sensations of stereoscopic depth, like the appreciation of spatial relationships that are apparent with only one eye, can be rationalized in wholly empirical terms. The evidence reviewed implies that monocular and binocular perceptions of depth rest on the same statistical basis.
9. Perceiving Motion
The relationship between objects, circumstances, and observer are rarely, if ever, stationary. Even when there is no physical motion in a scene, the retinal projection generated by objects in the world is moving continually with respect to the mosaic of retinal receptors. Thus any general theory of vision must deal with changing spatial relationships, and with the sensations of motion that are usually—but not always—elicited by the ensuing sequence of images. The purpose of this chapter is to assess whether the peculiarities of perceived motion can also be explained empirically.
10. Implications for the Mechanics of Vision
The gist of the previous chapters is that visual percepts are determined by behavioral success (or failure) in response to stimuli whose inevitable uncertainty cannot be resolved by the application of logical and/or intuitive rules. What, then, does this wholly empirical strategy of vision imply about the structure and function of the visual neurons on which the percepts are based, and how might the relevant probability distributions be realized in visual circuitry?
11. A Summary
The evidence reviewed in these chapters indicates that the visual system is not organized to generate a veridical representation of the physical world, but rather a statistical reflection of the visual history of the species and the individual observer. What we see is thus predicted by the probability distributions of the possible sources of the stimulus. This statistical derivation of visual percepts rationalizes the many otherwise puzzling discrepancies between what observers see and the physical measurements of real-world objects and/or their retinal projections. This argument about the nature of vision also suggests a way of deciphering the nature and purposes of the underlying visual machinery.
Glossary
References
Index
Pricing and Options
| Titles | Product Code | Price (USD) | ||
|---|---|---|---|---|
|
Why We See What We Do: An Empirical Theory of Vision | 0-87893-752-8 | $53.95 | Purchase | Request Exam Copy |
| View Cart | ||||
| Outside the U.S., Canada, or the Caribbean? Check our ordering page for local distributors. | ||||
Home ||
Contact Us ||
About Ordering ||
List by Author
List by Subject Area ||
List by Title ||
View Shopping Cart