Richard L. Gregory - Eye and Brain: the Psychology of Seeing

Eye and Brain

Eye and Brain

The Psychology of Seeing

Fifth edition

by

RICHARD L. GREGORY

PRINCETON UNIVERSITY PRESS

PRINCETON, NEW JERSEY

Published by Princeton University Press,

41 William Street, Princeton, New Jersey 08540

Copyright 1966, 1972, 1977, 1979, and 1997 by Richard L. Gregory

All Rights Reserved

This edition has been authorized by the Oxford University Press for sale in the USA and Canada only and not for export therefrom

Library of Congress Cataloging-in-Publication Data

Gregory, R. L. (Richard Langton)

Eye and brain: the psychology of seeing / Richard L. Gregory5th ed

p. cm.(Princeton Science Library)

Includes bibliographical references and index.

ISBN 0-691-04840-1

ISBN 0-691-04837-1 (pbk.)

1. Visual perception. I. Title. II. Series.

[DNLM: 1. Vision. 2. Visual Perception. WW 103 G823e 1979]

BF241.G7 1990 89-72211

152.14dc20 DNLM/DLC

Fourth edition, for the Princeton Science Library, 1990

Fifth edition 1997

Typeset by Footnote Graphics, Warminster, Wilts

Printed in Hong Kong

1 3 5 7 9 10 8 6 4 2

Contents

Pretext

This greatly rewritten edition of my first book started from the happy experience of giving lectures and running practical classes (very important!) at the University of Cambridge, England. It was a privilege to share the excitement of trying to explain what we see and how we see, including the strange phenomena of illusions, with generations of students. Many are now friends and colleagues, continuing to be fascinated by the gradually revealed wonders of eye and brain.

This book was not written for examinations or for any formal teaching course, though it has by now been so accepted, especially for psychology, physiology, the visual arts (including architecture), physics, and philosophy. This should not be surprising for perception is the basis of all our experience and understanding, in science and art and everyday life. The study of perception is indeed central to what used to be called by the splendid nameand still is in ScotlandExperimental philosophy. For questions and speculation enliven all interesting science.

This book is an introduction to the psychology of vision. It is written to be read easily, and to be enjoyed; but this does not mean that its subject is easy or that I have glossed over difficult issues. To take almost the first page: one has to think quite hard just why babies do not have to learn to see things the right way up, though the images in their eyes are upside down. (And why do we see ourselves rightleft reversed in a mirror, yet not upside down?)

This is an introduction for solo take-off to reflect about how we see. No one else can altogether see or think for us. I can only hope that this book will be useful and entertaining.

The question How do we see? may be approached from many points of view. Having approached it in one way we may, through change of insight, come to see it very differently. A classical account is that perception is the passive pick-up of information from the world, the brain having rather little to do. Eye and brain takes the very different view, that the brain (or mind) is highly activeconstructing perceptions from hardly adequate information from the senses. On this view, illusions of many kinds take on remarkable significance, as phenomena well worth studying and trying to understand. Illusions have generally been written off as annoying and sometimes dangerous, but essentially trivial. A main theme of this book is to explain such phenomena of visionwhich bridge art and scienceas a way of discovering quite a lot about how perception works. It turns out that there are several very different kinds of illusions. Some are due to upsets of the physiology of the nervous system; others, very differently, are like incorrect hypotheses in sciencedue to inappropriate assumptions, or misplaced knowledge. The first kind of illusions may be compared with computer hardware errors; the second kind with bugs of softwarethough it does not follow that the brain is just like a digital computer.

How similar perception by machines is, or ever will be, to our own, is a topic of ever-growing interest as this new technology advances. In it, we see modern technology linked to ancient questions of philosophyperhaps to find solutions even to why we are conscious.

Eye and brain first appeared in 1966, as the first volume of an imaginative series, World University Library, conceived by the distinguished London publisher Lord Weidenfeld. Each volume was lavishly illustrated, and translated into a dozen languages. This book owes a great deal to the original artists, Audrey Besterman and Mary Waldron, who drew so intelligently from my back-of-an-envelope doodles. Almost all of these first pictures are retained in this present much enlarged edition, together with others from The intelligent eye (1970), which is now out of print, and a large number of new pictures.

The second, third, and fourth editions (1972,1977,1990) of this book added new discoveries and ideas, though its structure remained essentially unchanged. The last 20 years have seen rapid growth of research. The brain sciences, including many kinds of studies of perception, have become a major international scientific endeavour, which has captured some of the greatest scientists from other fieldsnotably Francis Crick, who with James Watson and colleagues transformed how we think of life itself, through the discovery of the structure and significance of the DNA molecule.

This new edition is rewritten and greatly expanded. I hope it remains readable. It is, indeed, a considerable worry to tamper with a book that has been unusually successful over 30 years. It is a curious, and enjoyable experience, to criticize ones much earlier self and try to make use of the added experience and new ideas of ones later life.

Phenomena of illusion continue to be a major theme. Here is a new attempt to make sense of them, with a suggested classification. For any science, classification is extremely important. It seems high time to classify visual phenomena, by appearances and according to theoretical understanding, and this should be stimulated by seeing connections and differences more clearly. This is very much in the same spirit as Mendeleevs periodic table of the elements.

There is now, in this book, a lengthy discussion on the new work on seeing what babies see, and what and how they learn. This concludes with a short section on forgetting how to see: visual agnosia.

Studies of how motion is seen have always been important, but have now greatly advanced with the introduction of computer graphics. Unfortunately these phenomena cannot be demonstrated in a book. I would like to have included more on these experiments, particularly of Stuart Anstis and V. S. Ramachandran at UCSD in California; but perhaps these will have to wait for a new medium. We have, however, added 3-D redgreen stereo, so that previously hidden phenomena may now be seen.

Attempts to give vision to machines remain of great interest, with a recent change of emphasis from digital computers to analogue processors, especially interactive neural nets. This is still in a state of flux, and it remains unclear just how far electronics can encapsulate visual brain function. As this is inherently technical, and has not yet yielded quite the dramatic results that were hoped for some years ago, regretfully I have not given it much space here.

Over the last few years there has been a remarkable burst of interest on consciousness, especially how sensations (