8.1 Stratton’s experiment, in which he saw himself suspended in space before his eyes, in a mirror. He went for country walks wearing this arrangement.
Learning how to see
An ancient question in philosophy is: how do we come to know the world? Are we born with inherited, or innate, knowledge—or do we have to learn everything we know? Do newborn babies have to learn how to see?
Philosophers are divided into those often termed metaphysical, who hold that we do have some knowledge of the world apart from any sensory experience, and on the other hand empiricists claiming that all knowledge is derived from observations, experiments, and measurements. To the metaphysician, it seems clear that by thinking sufficiently hard and in the right kind of way it is possible to make ‘contingent’ discoveries, such as the number of planets, without having to look. To empiricists, perception is basic for all knowledge; though, it must be admitted, knowledge and assumptions can affect perception.
The psychologist is concerned with development rather than with what is inherited at birth. But how much of a child’s development is pre-programmed from inheritance? The present view is that the individual behaviour of the child is very important; so development is complicated results of interactions with the world of objects, in social situations, very much depending on the child’s initiatives and available opportunities for play and discovery. Child’s play is not unlike the exploratory play of scientists in their nurseries—their laboratories.
Much innate, immediately available knowledge is found in animals, most dramatically ‘releasers’ such as fledglings cowering from the profile-shapes of dangerous birds of prey. Many animal species seem to know a lot about the world of objects by instinct before they experience it. Insects play successful hide-and-seek with predator and prey before they have time to learn. Migrating birds use patterns of the stars to guide them over featureless oceans though they have never seen the sky before. How is this possible, if the empiricists are right that all knowledge is derived from experience?
Very fast learning is easily confused with immediately available innate knowledge. Many authorities on language development believe that much of the ‘deep’ structure of natural language must be inherited, because it seems impossible for a child to learn so much so soon. The crucial, not ethically accepted experiment was attempted by King James IV of Scotland (1473–1513), by marooning two infants on the island of Inchkeith, to be looked after by a dumb woman. It was reported (if not universally believed) that the babies spontaneously spoke Hebrew with a Scottish accent!
For 2000 years metaphysicians upheld their claim to innate knowledge by pointing to mathematics—especially to the spatial properties of Euclidean geometry—where new facts were discovered not by observations or experiments, but from studying diagrams and juggling with symbols. Only since the great German mathematician Karl Gauss (1777–1855) has it become clear that mathematical discoveries are of a special kind; constituting knowledge not of objects but rather of possible and impossible patterns. Gauss was among the first to realize that there is not only one possible geometry—Euclidean—but that other consistent geometries can be invented. So it is an empirical question to ask which axioms of suggested geometries best fit our world. (And it has been suggested that visual space is non-Euclidean.) The principal support for a priori knowledge of objects in space collapsed with the invention of alternative geometries. It is now believed that neither mathematics nor logic give new facts about the world, in the sense that facts are discovered by observation. In short, knowledge starts from experience, though mathematics and logic are extremely important for testing and drawing conclusions from mental models which might fit reality.
Animals respond appropriately to many objects and situations upon first encounter—but this does not make them metaphysicians. For they are heirs, by inheritance, to knowledge won through many generations of ancestral disasters, by natural selection, which benefits individuals with the most appropriate behaviour, who survive to hand on useful knowledge through their genes. Genetic coding can become modified by natural selection to give inherited knowledge to individuals, but what the parents learn individually in their life-time is lost.
Animals very low on the evolutionary scale rely almost entirely on unlearned reflex reactions; but their range of behaviour is restricted, and they respond in stereotyped ways. Some insects, especially bees, do show perceptual learning; for example the location of the hive and where nectar is plentiful—which could not be known innately. Patterns of petals leading to nectar became built into the inherited bees’ brain, and those lacking flower-vision die for lack of honey. So here is a mixture of innate and learned knowledge.
It is hard to establish what knowledge human infants are born with and what has to be learned. The difficulties are that there are strict limits to experiments that can be tried on human babies, and that infants have extremely limited co-ordinated behaviour. Until recently, almost all we knew of learning how to see has came from young animals. Now, however, there are safe and effective techniques for learning from babies what they can see. We will consider these new techniques and findings a little later, after looking at some physiological effects of experience on animals—and at some other matters—including what it is like to recover when adult from infant blindness.
Physiological changes
Recent experiments have been aimed at whether physiological ‘feature detectors’ (Figure 4.7) are simply given innately, or whether early experience affects them. Kittens have been reared in environments of vertical stripes, then tested for vision of vertical and other orientations. It has been found, especially by Colin Blakemore, that kittens living in a world of only vertical stripes appear to be blind to horizontal lines—and they lack horizontal feature detectors. Similarly, kittens denied vertical stripes do not have well organized vertical feature detectors. This suggests that feature detectors are not completely laid down at birth; but are developed—or ‘tuned’—by visual stimulation encountered by the individual. This is important for considering optimal environments for babies, especially as it has been found that some innately given neural mechanisms degenerate with lack of stimulation. This is clearly so for the ability to see depth stereoscopically. In childhood there are ‘critical periods’ for learning how to see, and without suitable experience at the right time such visual skills can be lost forever. Early visual environment of babies is highly important for adult vision—so nursery wallpaper should be considered!
Adaptation to disturbed images
Displaced images
To discover mechanisms of perceptual learning, we may look at experiments on animals and humans fitted with optical systems of various kinds to modify the retinal image, and see whether eye and brain compensate or adapt to the changed input. This was first tried at the end of the nineteenth century in famous experiments by the American psychologist G. M. Stratton, on himself. But first, let’s look briefly at animal experiments of this kind.
Inverting goggles placed on a monkey had the effect of immobilizing her for several days: she simply refused to move. When finally she did move it was backwards—a point of some interest as these inverting goggles tend to reverse depth perception. Similar experiments have also been tried in chickens and hens. Right–left reversing prisms were attached to the eyes of hens by M. H. Pfister, who observed their ability to peck grain. The hens’ behaviour was severely disturbed, and they showed no real improvement after three months wearing the prisms. The same lack of adaptation has also been found in amphibia, by R. W. Sperry. With vision rotated through 180°, it was found that they would move their tongue in the wrong direction for food, and would have starved to death had they been left to fend for themselves. Similar results were also obtained by A. Hess with chickens wearing wedge prisms which did not reverse the images, but shifted them by 7° to the right or to the left. He found that the chickens would always peck to the side of the grain, and that they never adapted to the shift of the image caused by the wedge prisms. Hess concluded:
Apparently the innate picture which the chick has of the location of objects in its visual world cannot be modified through learning if what is required is that the chick learns to perform a response which is antagonistic to its instinctive one.
It seems clear from the various experiments that animals show far less adaptation to a shift or reversal of the image than do human observers. Indeed, only monkeys and humans show any perceptual adaptation to these changes.
Now let’s look at the classical work of G. M. Stratton on inversion of the retinal image for a human observer. He wore inverting goggles for days on end—and was the first man to have retinal images that were not upside down! He devised a variety of lens and mirror systems including special telescopes mounted on spectacle frames so they could be worn continuously. These generally inverted both vertically and horizontally. Stratton found that when a pair of inverting lenses was worn giving binocular vision the strain was too great as normal convergence was upset, and this did not adapt to the situation. He therefore wore a reversing telescope on just one eye, keeping the other covered. When not wearing the inverted lenses he would keep both eyes covered, or stay in a dark room. He slept in the dark.
At first, objects seemed illusory and unreal. Stratton wrote (1896–7):
… the memory images brought over from normal vision still continued to be the standard and criterion of reality. Things were thus seen in one way and thought of in a far different way. This held true also for my body. For the parts of my body were felt to be where they would have appeared had the instrument (the inverting lens) been removed; they were seen to be in another position. But the older tactual and visual location was still the real location.
Later, however, objects would look almost normal.
Stratton’s first experiment lasted three days, during which time he wore the ‘instrument’ for about 21 hours. He concluded:
I might almost say that the main problem—that of the importance of the inversion of the retinal image for upright vision—had received from the experiment a full solution. For if the inversion of the retinal image were absolutely necessary for upright vision … it is difficult to understand how the scene as a whole could even temporarily have appeared upright when the retinal image was not inverted.
Objects only occasionally looked normal, however, and so Stratton undertook a second experiment with his monocular inverting arrangement, this time wearing it for eight days. On the third day he wrote:
Walking through the narrow spaces between pieces of furniture required much less care than hitherto. I could watch my hands as they wrote, without hesitating or becoming embarrassed thereby.
On the fourth day he found it easier to select the correct hand, which had proved particularly difficult:
When I looked at my legs and arms, or even when I reinforced my effort of attention on the new visual representation, then what I saw seemed rather upright than inverted.
By the fifth day, Stratton could walk around the house with ease. When he was moving around actively, things seemed almost normal, but when he gave them careful examination they tended to be inverted. Parts of his own body seemed in the wrong place, particularly his shoulders, which of course he could not see. But by the evening of the seventh day he enjoyed for the first time the beauty of the scene on his evening walk.
On the eighth day he removed the inverting spectacles, finding that:
… the scene had a strange familiarity. The visual arrangement was immediately recognised as the old one of pre-experimental days; yet the reversal of everything from the order to which I had grown accustomed during the last week, gave the scene a surprising bewildering air which lasted for several hours. It was hardly the feeling, though, that things were upside down.
One has the impression when reading the accounts of Stratton, and the investigators who followed him, that there is always something queer about their visual world though they have the greatest difficulty saying just what is wrong with it. Perhaps, rather than their inverted world becoming entirely normal, they cease to notice how odd it is until their attention is drawn to some special feature, when it does look clearly wrong. Thus writing appears in the right place in the visual field, and at first sight looks like normal writing, except that when they attempt to read it, it is seen as inverted, or at least it appears odd.
Stratton went on to perform other experiments, which though less well known are just as interesting. He devised a mirror arrangement, mounted in a harness (Figure 8.1), which visually displaced his own body so it appeared horizontally in front of him at the height of his eyes. Stratton wore this mirror arrangement for three days (about 24 hours of vision), reporting:
The different sense-perceptions, whatever may be the ultimate course of their extension, are organised into one harmonious spatial system. The harmony is found to consist in having our experience meet our expectations … The essential conditions of the harmony are merely those which are necessary to build up a reliable cross-reference between the two lenses. This view, which was first based on the results with the inverting senses, is now given wider interpretation, since it seems evident from the later experiment that a given tactual position may have its correlated visual place not only in any direction, but also at any distance in the visual field.
8.1 Stratton’s experiment, in which he saw himself suspended in space before his eyes, in a mirror. He went for country walks wearing this arrangement.
Several investigators have followed up Stratton’s work. G. C. Brown used prisms to rotate the field of both eyes through 75°, and found that this reduced the efficiency of depth perception; but there was little or no evidence of improvement with experience, though he and his subjects did find that they got used to their tilted world. Later, P. H. Ewert repeated Stratton’s experiment using a pair of inverting lenses, in spite of the strain on the eyes found by Stratton. Ewert’s work has the great merit that he made systematic and objective measures of his subjects’ ability to locate objects. He concluded that Stratton somewhat exaggerated the amount of adaptation that occurred. This led to a controversy that is still unresolved.
The problem was taken up by J. Paterson and J. K. Paterson, using a binocular system similar to Ewert’s. After 14 days they did not find complete adaptation to the situation. Upon re-testing the subject of the experiment eight months later, they found that when the subject wore the lenses again, he immediately showed the various modifications to his behaviour which he had previously developed while wearing the reversing spectacles. It seemed that the learning consisted of a series of specific adaptations, overlying the original perception, rather than a reorganization of the entire perceptual system.
The most extensive recent experiments on humans have been carried out by T. Erisman, followed by Ivo Kohler at Innsbruck. Both Stratton’s and Kohler’s experiments rely on verbal reports. Kohler stresses the ‘inner world’ of perception, following the European tradition which we find in the German Gestalt writers, and in the work of Michotte on the perception of causality (Chapter 4). This emphasis is foreign to the behaviourist tradition of America, and it is unfortunate that little precise recording of the subject’s movements during the experiments was attempted. From the verbal reports it is difficult to imagine the ‘adapted’ world of the experimental subjects, for their perceptions seem to be curiously shuffled and even paradoxical. For example pedestrians were sometimes seen on the correct side of the street, when the images were right–left reversed, though their clothes were seen as the wrong way round! The suggestion is that having to avoid bumping into people produced re-learning of their positions on the pavement, but not of which side the buttons were on their coats. Writing is one of the more puzzling things observed. With right–left reversal a scene would come to look correct, except that at least sometimes writing remained right–left reversed and hard to read.
Touch had important effects on vision: during the early stages of adaptation objects would tend to look suddenly normal when touched, and they would also tend to look normal when the reversal was physically impossible or highly unlikely. For example, a candle would look upside down until lighted, when it would suddenly look normal—the flame going upwards, instead of downwards. Touch, even with a long stick, would switch the world the right way up.
There is later evidence, mainly from the work of Richard Held and his associates, particularly Alan Hein at Massachusetts, to show that for compensation to displaced images to occur, it is important for the subject to make active corrective movements. Held considers that active movement is vital for perceptual learning in the first place, as well as for compensation. An experiment with kittens is particularly ingenious and interesting. They brought up a pair of kittens in darkness; they could see only in the experimental situation, in which one kitten served as a control for the other. The two kittens were placed in baskets attached to opposite ends of a pivoted beam, which could swing round its centre, while the baskets could also rotate. It was arranged that a rotation of one basket caused the other to rotate similarly (Figure 8.2). With this ingenious device both kittens received much the same visual stimulation, but one was carried passively in its basket; the other, whose limbs could touch the floor, moved the apparatus around actively. Held and Hein found that only the active kitten gave evidence of perception, the passive animal remaining for a time effectively blind. But is this ‘blindness’ the absence of correlations built up between its vision and its behaviour? Or could the kitten indeed be seeing, but be unable to let us know that it sees?
Richard Held also undertook experiments on humans using deviating prisms, finding that active arm movement (striking a target with the finger) is necessary for effective adaptation. Is this adaptation perceptual, or is it proprioceptive—in the control system of the limbs? The principal supporter of proprioceptive adaptation is C. S. Harris. This cannot apply to adaptation to some kinds of distortion of vision, but in these cases the role of feedback from experience is less clear.
Distorted images
We have considered experiments on inverting and tilting the eye’s images, but other kinds of disturbance can be produced, which are important because they involve internal reorganization in the perceptual system itself, rather than relatively simple changes in the relation between the worlds of touch and vision. These can be investigated by wearing lenses which distort, rather than displace, the image on the retina.
8.2 Apparatus designed by Held and Hein to discover whether perceptual learning takes place in a passive animal. The kitten on the right is carried about passively by the active kitten on the left. They thus have similar visual stimulation. Following visual experience limited to this situation, only the active animal is able to perform visual tasks—the passive animal remains effectively blind, until some time after the experiment.
J. J. Gibson found, while undertaking an experiment wearing prisms to deviate the field 15 degrees to the right that the distortion of the image, which such prisms produce in addition to the shift, gradually became less marked while he wore them. He went on to make accurate measures of the adaptation to the curvature produced by the prisms, and found that the effect diminished although his eyes moved about freely. Surprisingly, the adaptation was slightly more marked with free inspection of the figure with eye movements than when the eyes were held as still as possible. This adaptation does not depend on touch, so may be different from seeing where things are.
There is another kind of adaptation which at first sight seems similar to Gibson’s with his distorting prisms, but is almost certainly different in its origin and its significance to the theory of perception. These effects are known as figural after-effects. They have received a great deal of experimental attention, though they remain somewhat mysterious. Figural after-effects are induced when a figure is looked at for some time (say half a minute) with the eyes held very still. If a curved line is fixated, a straight line viewed immediately afterwards will, for a few seconds, appear curved in the opposite direction. The effect is similar to Gibson’s, but for figural after-effects it is essential that the eyes should be still. This seems to be a visual ‘normalizing’ process, useful for correcting standing errors such as astigmatism.
Using television to study visual disturbance
There are limitations to the kinds of inversion possible by simple optical means. A technique has been developed by K. U. Smith and W. M. Smith, using a television camera and monitor arranged so that the subject sees his own hand, displaced in space or time. It is a simple matter to give either right–left or up–down reversal of the image, and eye and hand movements are not affected. In this arrangement, the hand is placed to the side of the subject, behind a curtain so it cannot be seen directly. (Since the apparatus was far from portable these studies were limited to short experimental sessions.) In addition to reversals, the camera may be placed in any position, giving a view displaced in space. Using various lenses and camera distances, the size may be varied, and distortions may be introduced (Figure 8.3).
These techniques show that pure up–down reversal generally proves more disturbing than left–right reversal, and combined up–down and left–right is less disturbing than either alone. Changes in size or position had practically no effect on ability to draw objects, or on handwriting.
Displacement in time
An elaboration of this television technique makes it possible to displace retinal images, not only in space but in time. The original method was to introduce a videotape recorder with an endless tape loop, so that there is a time delay between the recording from the camera and play-back to the monitor. The subject thus sees his hands (or any other object) in the past, the delay being set by the gap between the record and play-back heads (Figure 8.4).
This situation is not only of theoretical interest but is also of practical importance, because controls used in flying aircraft and operating many kinds of machine have a delay in their action. If delay upsets the control skill this could be a serious matter. It was found that a short delay (about 0.5 seconds) made movements jerky and ill co-ordinated, so that drawing became almost impossible and writing quite difficult (Figure 8.5). Practice gives little or no improvement. Present-day video equipment would make such experiments much easier to do.
8.3 K. U. Smith and W. M. Smith’s experimental arrangement using a television camera and monitor to vary the viewing position and size of the subject’s own hands. He or she can draw or write very well despite effectively displaced eyes.
Links of adaptation
A remarkable discovery was made by Ivo Kohler, wearing glasses which did not distort, but which were coloured half red and half green, so that everything looked red while looking to the left, and green when looking to the right (Figure 8.6). He found that the colours gradually weakened, and when the glasses were removed things seen with the eyes directed to the right looked red, and to the left green. This is quite different from normal after-images, for the effect is not related to the position of images on the retina, but to the position of the eyes in the head—so it must be due to compensation taking place in the brain.
8.4 Introducing time delay between action and seeing. The delay was given by the tape loop of the video recorder. This would now be done electronically.
This might be related to a striking effect discovered by Celeste McCollough in 1961. When lines of one orientation are repeatedly shown as red, and alternated with a green grating at another orientation, black and white lines of the same orientations appear coloured. The induced colours seen in the black and white lines are the complementaries of the adapting colours, for each orientation. This also works with movements, and to other visual stimuli, though not for non-visual stimuli such as sounds. So this is only within-vision learning.
Such contingent after-effects, as they have come to be called, show many of the characteristics of classical Pavlovian conditioning. They build up gradually with repeated stimuli; they decay rapidly when elicited by the ‘unconditioned stimuli’ (the differently oriented black and white lines). Left to themselves, the effects can persist for many hours, and even days or longer. It has also been suggested that links between related stimuli given by ‘double duty’ cells are teased out by this procedure. But it is now clear that highly ‘artificial’ and extremely unlikely pairings of stimuli can occur; so this is not a plausible explanation, for we should not expect cells to be pre-wired for highly unlikely combinations of stimuli. It is most likely that this is a kind of perceptual learning. Why, though, should it always be ‘negative’? It seems to be recalibrating vision—to produce constancy of perception against irrelevant repeated sensory signals. This may be similar to compensating for new spectacles and such other visual learning. The McCollough effect is a dramatic and an experimentally useful phenomenon—illustrating how vision scales and adapts itself, generally avoiding its own errors, to make sense of the world. But generally useful processes can sometimes generate curious and revealing illusions.
8.5 Drawing with a time delay. Left to right: normal; with TV but no delay; with TV and delay. The delay provides an insuperable handicap, though displacement in space can be compensated. (The result is of practical importance, since many controlled tasks, such as flying, do impose a delay between action and result.)
8.6 Demonstrating contingent after-effects. After directing the eyes through the green filter to one side, and then the red filter to the other, they become adapted and compensate to the filters for each viewing position. When the filters are removed, the side which was green looks red and vice versa. This adaptation must be in the brain, not the eyes.
Do people brought up in different environments come to see differently? The Western world has visual environments with many straight parallel lines, such as roads, and right-angular corners of buildings and furniture and so on. These are strong, generally reliable perspective cues to distance. We may ask whether people living in other environments where there are few right angles and few long parallel lines have somewhat different perception. Fortunately several studies have been made of people living in such environments.
Those who stand out as living in a non-perspective world, were the Zulus. Their world has been described as a ‘circular culture’—traditionally their huts were round, they did not plough their land in straight furrows, but in curves, and few of their possessions had corners or straight lines (Figure 8.7). They are thus ideal subjects for our purpose. It is found that they experienced the Muller–Lyer arrow illusion (Figure 10.16) to only a small extent, and were hardly affected at all by other such distortion illusion figures.
8.7 Circular culture of the Zulus. They experience few straight lines or corners, and are not affected by ‘perspective’ illusion figures to the same extent as people brought up in a ‘rectangular’ Western culture.
Studies of people living in dense forest have also been made. Such people are interesting in that they do not see many distant objects, because they live in quite small clearances in the forest. When they are taken out of their forest, and shown distant objects for the first time, they see these not as far away but as small. (They have even reported that cattle look like insects.) People living in Western cultures experience similar distortion when looking down from a height. From a tall building, objects look much too small. Steeplejacks, and men who work on the scaffolding and girder structure of skyscrapers, are reported to see objects below them without this distortion. Again, active movement and handling of objects seem to be very important for calibrating the visual system.
Jan Deregowski has found that Zulus who do not experience illusions also see little or no perceived depth in these figures (using the depth measuring technique of Pandora’s box (Figure 10.23)). So there seems to be rather clear evidence for cultural factors in these distortions, related to distance cues as available in the environment. (For a fuller discussion see Chapter 10.)
Recovery from blindness
It might seem a simple matter to bring up animals in darkness—to deny them vision for months or years—and then discover what they see given light. Pioneering experiments of this kind were undertaken by A. H. Reisen. He found severe behavioural losses; but some of these might be due to degeneration of the retina (which was found to occur in darkness, though less so with diffusing goggles) and also to the remarkably passive state of animals, especially monkeys, reared in the dark. It was difficult to infer specific perceptual changes, or losses, because of the general lack of behaviour of these animals. It is not socially possible to bring up human babies in the dark; but there are cases of adult recovery from blindness. Can these tell us how human perception develops?
Perception of the blind was described by Descartes in the Dioptrics (1637), as discovering the world by poking with a stick:
… without long practice this kind of sensation is rather confused and dim; but if you take men born blind, who have made use of such sensations all their life, you will find they feel things with such perfect exactness that one might almost say they see with their hands.
The implication is that this kind of learning might be necessary for the normal child to develop his or her world of sight.
The English philosopher John Locke (1632–1704) received a celebrated letter from his friend William Molyneux, posing the question:
Suppose a man born blind, and now adult, and taught by his touch to distinguish between a cube and a sphere of the same metal, and nighly of the same bigness, so as to tell, when he felt one and the other, which is the cube, which is the sphere. Suppose then the cube and the sphere placed on a table, and the blind man made to see; query, Whether by his sight, before he touched them, he could now distinguish and tell which is the globe, which the cube? To which the acute and judicious proposer answers: ‘Not. For though he has obtained the experience of how a globe, and how a cube, affects his touch; yet he has not yet attained the experience, that what affects his touch so or so, must affect his sight so or so.’
Locke comments (in Essay concerning human understanding, 1690) as follows:
I agree with this thinking gentleman, whom I am proud to call my friend, in his answer to this problem; and am of the opinion that the blind man, at first, would not be able with certainty to say which was the globe, which the cube.
Here was a suggested psychological experiment—with a guessed result. George Berkeley (1685–1753), the Irish philosopher, also considered learning how to see in this kind of way:
In order to disentangle our minds from whatever prejudices we may entertain with the relation to the subject in hand nothing is more apposite than the taking into our thoughts the case of one born blind, and afterwards, when grown up, made to see. And though perhaps it may not be an easy task to divest ourselves entirely of the experience received from sight so as to be able to put our thoughts exactly in the posture of such a one’s: we must, nevertheless, as far as possible, endeavour to frame conceptions of what might reasonably be supposed to pass in his mind.
Berkeley goes on to say, that we should expect such a man not to know that anything was:
high or low, erect or inverted … for the objects to which he had hitherto used to apply the terms up and down, high and low, were such only as affected or were some way perceived by his touch; but the proper objects of vision make new sets of ideas, perfectly distinct and different from the former and which can by no sort make themselves perceived by touch.
Berkeley gives his opinion that it would take some time to learn to associate touch with vision. But guessing, even by the most distinguished philosophers, is no substitute for observation and experiment. There have been several actual cases of the kind imagined by Molyneux and discussed by Locke and Berkeley. The most famous is that of a 13-year-old boy, described in 1728 by the celebrated eye and general surgeon, William Cheselden. (Cheselden was spokesman of the London surgeons. He attended Sir Isaac Newton in his final illness.) Although there are many reported cases, few have sufficient evidence of lack of vision in infancy, and early operations for removing congenital cataracts did not give a good retinal image for weeks or months, if then. The first reported case dates from AD 1020. There are a few recent cases where sight has been restored to an adult who was effectively blind from early infancy or birth, giving a good retinal image immediately.
Some of the reported cases are much as the empiricist philosophers expected. The patients could see but little at first, being unable to name or distinguish between even simple objects or shapes. Sometimes there was a long period of training before they came to have useful vision, and indeed in many cases it was never attained. Some gave up the attempt, reverting to a life of blindness, often after a period of severe emotional disturbance. It is important to note, however, that the reported cases do not all show extreme difficulty or slowness in coming to see. We should also remember that the early operations disturbed the optics of the eye, so there could not be a useful image until the eye had considerable time to settle down. This is particularly important in the case of removal of the lens for cataract of the lens, which constitutes all the earlier cases. The other kind of operable blindness—opacity of the cornea—involves less change or damage and far more rapidly gives an adequate retinal image. We shall now discuss in some detail such a case, which I had the good fortune to investigate, with my colleague Jean Wallace, starting in 1961. Following a life of blindness, S.B. was given successful corneal transplants when he was 52.
The case of S.B.
S.B. was an active, intelligent man, who spent a lot of time and energy imagining the sighted world. For many years he tried to get corneal grafts, but they were in short supply before corneal banks were set up, so it was not considered worth the risk. This is why he was blind from early infancy (or birth) until middle age, which makes this case most unusual.
While blind, S.B. would go for cycle rides, with a friend holding his shoulder to guide him; he would often dispense with the usual white stick, sometimes walking into obstructions such as parked cars, and quite often hurting himself. He liked making things with simple tools in a shed in his garden. All his life he tried to picture the world of sight: he would wash his brother-in-law’s car, imagining its shape as vividly as he could. He longed for the day when he might see. Finally the attempt was made when corneal banks were set up, and it was successful. But though the operation was a success his story ended tragically.
When the bandages were first removed from his eyes, so that he was no longer blind, S.B. heard the voice of the surgeon. He turned to the voice, and saw nothing but a blur. He realized that this must be a face, because of the voice, but he could not see it. He did not suddenly see the world of objects as we do when we open our eyes. But within a few days he could use his eyes to good effect. He could walk along the hospital corridors without recourse to touch; he could even tell the time, from a large wall clock, having all his life carried a pocket watch, with no glass, so that he could feel the time by touching its hands, as he demonstrated to us with great skill. In the hospital he would get up at dawn, and watch from his window the cars and trucks pass by. He was delighted with his progress.
When S.B. left the hospital we took him to London and showed him many things he never knew from touch, but quite soon he became curiously dispirited. At the zoo, he was able to name most of the animals correctly, having stroked pet animals when a boy, and enquiring how other animals differed from the cats and dogs he knew by touch. (We learned of his early life from an older sister.) S.B. was also, of course, familiar with toys and models. He certainly used his previous knowledge from touch and reports from sighted people to help him name objects by sight, which he did largely by seeking their characteristic features. But he found the world drab, and was upset by flaking paint and blemishes. In particular, he looked at a lamp post, walked round it, studying it from different aspects, and wondered why it looked different yet the same. But at that time, he said that he noted more and more the imperfections in things, and would examine small irregularities and marks in paintwork or wood, which he found upsetting, evidently expecting a more perfect world. He liked bright colours, but became depressed when the light faded. His depression became marked and general. He gradually gave up active living, and three years later he died.
Depression in people recovering sight after many years of blindness is a common feature of these cases. Its cause is probably complex, but in part it seems to be a realization of what they have missed—not only visual experience, but opportunities to do things denied them during the years of blindness. Some of these people revert very soon to living without light, making no attempt to see. S.B. would often not trouble to turn on the light in the evening, but would sit in darkness.
We tried to discover what his visual world was like, by asking him questions and giving him various simple tests. While still in the hospital, before he became depressed, he was most careful with his judgements and his answers. We found that his perception of distance was peculiar, and this was also true of earlier cases. He thought he would just be able to touch the ground below his window, with his feet, if he hung from the sill with his hands; but in fact the distance down was at least ten times his height. On the other hand, he could judge distances and sizes quite accurately when he knew the objects by touch and, it seemed, especially distances from walking.
Although his perception was demonstrably peculiar, he seldom expressed surprise at anything he saw. He drew the elephant (Figure 8.8) just before we showed him one at the zoo; but upon seeing it, he said immediately: ‘There’s an elephanty’, and said it looked much as he expected it would. On one occasion he did show real surprise, and this was an object he could not have known by touch—the moon. A few days after the operation, he saw what he took to be a reflection in a window. (He was for the rest of his life fascinated by reflections in mirrors, and would spend hours sitting before a mirror in his local public house, watching people.) But this time what he saw was not a reflection, but the quarter moon. Thinking it was a reflection in the window, he asked the Matron what it was. When she told him—he said he had thought the quarter moon would look like a quarter piece of cake! He seldom, however, appeared surprised by such anomalies—including his dog-like drawing of an elephant.
8.8 SB’s drawing of an elephant. He drew this before having seen one. Half an hour later we showed him a real elephant, at the London Zoo. He was not at all surprised by it.
S.B. never learned to read by sight (he read Braille, having been taught it at the blind school), but we found, in the hospital, that he could immediately recognize capital letters and numbers by sight. This so surprised us we could hardly believe he had been blind. It turned out (from the school records) that he had been taught upper case, though not lower case, letters and numbers by touch at the blind school. The children were given inscribed letters on wooden blocks, to learn by active touch. The children were not given lower case letters to learn, presumably because they were seldom used for inscribed brass plates, and so on. Although he read upper case letters immediately by sight, it took him a long time to learn lower case letters, and he never managed to read more than simple words. Now this finding that he could immediately read letters visually, which he had already learned by touch, clearly showed that he was able to use previous touch experience for his new-found vision. This is interesting to the psychologist, for it indicates that the brain is not so departmentalized as sometimes thought. There seems, indeed, to be a general knowledge-base available to all the senses. But this makes cases such as S.B.’s hard to apply to the normal case of a human infant coming to see. For the blind adult knows a great deal about the world of objects, through touch and hearsay, and can use his or her knowledge to identify objects from the slightest cues. It is also necessary to trust the unproved, not altogether reliable new sense, which means giving up the habits of many years and risking danger and humiliation. S.B. was teased by his friends for his mistakes, especially as he could not identify people at all reliably from their faces. So his situation was really very different from a child learning to see.
S.B.’s use of early touch experience comes out clearly in drawings which he did for us, starting while still in the hospital and continuing for a year or more. The series of drawings of buses (in Figure 8.9) illustrate how he was unable to draw anything he did not already know by touch. In the first drawing, the wheels have spokes, and spokes were a distinctive touch feature of wheels. The windows seem to be represented as he knew them by touch, from the inside. Most striking is the complete absence of the front of the bus, which he would not have been able to explore with his hands, and which he was still unable to draw six months or even a year later. The gradual introduction of writing in the drawings indicates visual learning: the sophisticated script of the last drawing meant nothing to him for nearly a year after the operation, although he could recognize capital letters while still in the hospital, having learned them previously from touch. It seems that S.B. made immediate use of his earlier touch experience, and that for a long time his vision was limited to what he already knew.
We saw, in a dramatic form, the difficulty that S.B. had in trusting and coming to use his vision when crossing a busy road. Before the operation he was undaunted by traffic. We were told that previously he would cross roads alone, holding his arm or his stick stubbornly before him, when the traffic would subside as the waters before Christ. But after the operation it took two of us, on either side, to force him across: he was terrified, as never before in his life.
When he was just out of the hospital, and his depression was but occasional, he would sometimes prefer to use touch alone when identifying objects. We showed him a simple lathe (a tool he had wished he could use) and he was very excited. We showed it him in a glass case at the Science Museum in London, then we opened the case so that he could touch it. With the case closed, he was quite unable to say anything about it, except that the nearest part might be a handle (which it was—the transverse feed handle), but when he was allowed to touch it, he closed his eyes and placed his hands on it, when he immediately said with assurance that it was a handle. He ran his hands eagerly over the rest of the lathe, with his eyes tight shut for a minute or so; then he stood back a little, and opening his eyes and staring at it, he said: ‘Now that I’ve felt it I can see’.
More recently some patients have had their sight restored by acrylic lenses. Rejection by the eye’s tissue was avoided by placing the lens in a tooth extracted from the patient, and implanting the tooth in the eye with the lens in a hole drilled through the tooth. There are now suitable materials for lenses which are not rejected.
Many of our findings with S.B. have been confirmed with these new cases. The extraordinary operations, from which people walk around with a tooth in their eyes as in a Greek myth, were performed in Italy, one of the patients being a philosopher. The results were reported in 1971 by the Italian psychologist Alberto Valvo. He, too, found that earlier experience of letters by touch gave immediate visual recognition. There are now cases of children in Japan who are being intensively studied, and Oliver Sacks has recently described a new case remarkably similar to S.B.
Although there are many philosophers and psychologists who think that these cases can tell us about normal perceptual development in children, I am inclined to think that they tell us rather little directly. As we have seen, the difficulty is essentially that adults, with their great store of knowledge from the other senses and reports from sighted people, are very different from the infant, who starts with no knowledge from experience. It is extremely difficult, if not impossible, to use these cases to answer Molyneux’s question in terms of what babies see. The cases are interesting and dramatic, but when all is said they tell us little about the world of the infant, for adults with restored vision are not living fossils of babies. The philosophers did not guess that there would be transfer of knowledge, gained by touch, to new-found vision. So we have learned something surprising and interesting from these rare dramas.
8.9 (a) SB’s first drawing of a bus (48 days after the corneal graft operation giving him sight). All the features he drew were probably known to him by touch. The front, which he could hardly have explored by touch, is missing, and he could not add it when we asked him to try. (b) Six months later. Now he adds writing. The inappropriate spokes of the wheels have been rejected, but he still cannot draw the front of the bus. (c) A year later, the front is still missing. The writing is sophisticated, though he could hardly read.
Perception and behaviour
We know far too little about relations between perception and behaviour. For example, do visual distortions correspond with behavioural errors? Recent experiments show that they can be separate. Thus the visual illusory size change of circles in the Titchener illusion (Figure 8.10) does not correspond to how discs are grasped with the hands.
8.10 The Titchener illusion with touch. The central circles are the same size; but the circle enclosed by the large circles looks smaller. For the subject’s hands, picking them up guided by vision, there is no distortion. Visual illusions do not always correspond with errors of behaviour.
Here vision is distorted though hand control is not. As David Milner and Melvyn Goodale said, in 1995, in The visual brain in action (pages 163–4):
The primary purpose of perception is to identify objects and places, to classify them, and to attach meaning and significance to them, thus enabling later responses to them to be selected appropriately. As a consequence, perception is concerned with the enduring characteristics of objects so that they can be recognised when they are encountered again in different contexts.
So vision of the object world is not necessarily tied to the observer’s body. But to be useful the neural processes for behaviour must be geared very closely to the body, and to the here and the now. So perception and behaviour do not occupy quite the same worlds.
What do babies see?
Of the most immediate interest is what infants—babies in the first two years of life, without speech—see immediately and in a few months come to see. It is hard to get reliable data because at first there is little motor co-ordination and no language. Recent experiments have, however, shown that babies at birth do have considerable vision. Within hours they can copy facial expressions of the mother, especially sticking out the tongue, so they must start out with some ability to see. Through, and following, the pioneering work of the Swiss developmental psychologist Jean Piaget (1896–1980), there are now several safe and effective techniques for revealing what babies see and understand through the first months and early years of human life. Piaget invented this experimental epistemology.
The first question is: do babies see objects as permanent things, in a world shared by other children and adults?
The Bishop Berkeley effect—permanence of objects
We have stressed throughout that perception is far more than simply responding to stimuli. Rather, it is acting appropriately to assumed sources, or causes, of stimuli. So we must consider not only pattern recognition but also object recognition—objects having lawful pasts and futures, and unsensed characteristics which often are vital for behaviour.
A primary characteristic of objects is their permanence, though they appear but fleetingly from time to time, as we glance at them, or as they reappear from being hidden by nearer objects. How do infants cope with objects disappearing? We may ask first: how do philosophers cope with disappearing objects? Bishop George Berkeley (1685–1753) suggested that objects only exist while they are perceived. This is unforgettably expressed in Ronald Knox’s famous limerick and its reply:
There was a young man who said, ‘God
Must think it exceedingly odd
If He finds that this tree
continues to be
When there’s no one about in the Quad.’
God replies:
Dear Sir,
Your astonishment’s odd:
I am always about in the Quad.
And that’s why the tree
Continues to be.
Since observed by
Yours faithfully GOD
The English empiricist philosopher John Stuart Mill (1806–73) described objects as ‘permanent possibilities of sensation’. Is this how infants see them?
Piaget suggested that very young babies do not have any notion that objects continue to be when not observed. Then they discover that when part of an object is revealed the rest will be there, to be found, and finally that whole objects continue to exist when they are not seen. Piaget held that these are discoveries made by active exploration, though this has been questioned because baby behaviour is so clumsy that such experiments must be hard to carry out in the first few months of life. (The alternatives—that the information is out there to be ‘picked up’, or that there are innate principles for completing missing objects—seem, however, more implausible.)
It is found that young babies recognize objects in spite of changes of view. They show ‘object constancy’; so they do not respond to appearances, but rather to objects as existing and remaining the same in spite of considerable change of appearance. Only slowly do they come to see themselves as moving in a room. Recognition of self is not developed before about two years of age. Part of the evidence for this, is behaviour to their own image in a mirror. When a baby touches his or her own face—not the mirror—it seems that the baby has some notion of self.
Techniques for studying infant vision
How are stages of learning how to see discovered? Techniques include looking for prediction, surprise, boredom.
Prediction
This can mean simply expecting what has happened to continue happening, as for Pavlovian conditioning or it may indicate cognitive understanding, including predicting others’ behaviour in novel situations. It is said that children develop ‘theories of other minds’, tested by predictions which may or may not be confirmed. Perhaps the strongest evidence for other minds is fibs; such as telling a child there are no cookies in the tin, but he or she finds that there are! The fib, or lie, is very different from the behaviour of inanimate objects; though babies only slowly come to distinguish between living and non-living.
Surprise
A key sign of prediction, and understanding is surprise. Cognitive perception is inherently predictive—depending on knowledge and assumptions—revealed most dramatically by failed predictions, with accompanying surprise. There are various signs of surprise in children before speech, such signs as a sudden agitation, and increased heart rate. Almost from the start, babies have expectations of objects. Expectations, and so surprises, become richer and more varied as perception develops. Surprise is, indeed, the principal reward of science.
The Scottish experimental psychologist T. G. R. Bower, projected images of objects stereoscopically (as in Figure 10.24) to appear a foot or so away, though there was nothing to touch. (It should be noted, however, that effective stereoscopic vision is not present before about three months.) Bower found that young infants show surprise when they reach out to grasp images having no substance. No doubt this is like our surprise at successful conjuring. When an interesting object such as a teddy bear is passed behind an opaque screen, and so disappears, the baby will move her eyes to the further side of the screen, evidently expecting it to re-appear. If it does not appear she is likely to get upset. What happens if the teddy gets hidden behind the screen, then another object, such as a toy fire engine, appears instead? Very young babies will show no surprise at a teddy turning into a fire engine; but at a year or so, there is surprise at such a transformation. So at about this age the baby possesses the knowledge that objects such as teddies do not turn into other objects such as fire engines.
In general, surprise at the unusual reveals knowledge of the usual. This is useful for seeing what babies see—and for reading minds.
Boredom
Signs of almost the opposite—boredom—are also useful. Babies get bored (become habituated) with repeated stimuli. But suppose the stimuli change but what they mean remains the same—would such repetition also produce boredom and habituation, so that interest flags and behaviour ceases? The trick is to discover whether the meaning of one pattern or shape is equivalent to another by habituating the baby (boring it) with the first, and noting whether it remains bored or habituated to the second.
There are other techniques for looking into baby vision and understanding.
Suck it and see
The American psychologist Jerome Bruner made imaginative use of one of the infant’s few well organized behaviours, by getting infants to focus a slide projector by sucking on a special teat. In this way they would bring pictures of interest into focus.
What do babies prefer?
Even very young babies have quite well co-ordinated eye movements. These indicate what the baby is looking at, and so what is seen and preferred. Robert Fantz, the pioneer in using infants’ eye movements to see what babies see, presented pairs of pictures and noted which was preferred from the times the baby looked at each picture. This technique has proved to be very useful, though a snag is the baby may find neither picture of interest; so much of what can be recognized or discriminated may be missed by the experimenter.
If one picture screen is blank, and the other has stripes (often called gratings), the infant will look more at the grating. There is a general preference for complexity. By making the grating more and more closely spaced until there is no preference, the baby’s visual acuity can be measured. Infant acuity is considerably less than adult’s, probably because the retina is not fully developed. It used to be thought that infants lack accommodation to different distances, but this is not too serious an experimental problem. Oliver Bradick and his wife Janette Atkinson carry out extensive clinical tests, using a technique based on the ‘red eyes’ seen in flash portraits when the sitter’s eyes are not accommodated to the distance of the camera (cf. pages 42–44).
Very young babies can discriminate between orientations of coarse gratings, and between various simple outline shapes—cross, circle, triangle, square, acute and obtuse angles—with a general preference for patterns with curves.
Presented with a simple face-like picture, and the alternative of a jumbled face, Robert Fantz discovered that babies spend about twice as long looking at the normal face (Figure 8.11). This may imply innate recognition of faces; but early learning is not ruled out, as the mother’s face had not been hidden from the baby. Figure 8.12 shows what happened.
Babies tend to prefer solid objects to flat representations of the same objects; so possibly they have some innate appreciation of depth. This is suggested by reaching and grasping of objects, some authorities believing that very young babies open their fingers to the size of an object before touching it. The famous visual cliff shows that babies at about four months are aware of potentially dangerous drops.
The visual cliff
Eleanor Gibson, while picnicking on the rim of the Grand Canyon, wondered whether a young baby would fall off the edge. This thought led her to a most elegant experiment, for which she devised a miniature and safe Grand Canyon. The apparatus is shown in Figure 8.13, which shows a central ‘bridge’ with on one side a normal solid floor, and on the other side a drop, covered by a strong sheet of glass.
8.11 (a) Fantz’s apparatus for observing babies’ eye movements while they are shown various designs, or objects. Here the baby is shown an illuminated ball, while the eye positions are photographed. (b) A simple face and a randomized face-like design, which were shown to very young babies. They spent longer looking at the typical face picture (as judged by their eye movements).
8.12 Results of Fantz’s eye movement experiments on babies. The horizontal bars show the relative times they spent looking at the designs shown on the left of the diagram.
An infant (or in other experiments a young animal) is placed on the central bridge. The question is: will the baby crawl over the drop? The typical answer is that the baby will not leave the bridge for the drop, and cannot be enticed over it by his mother shaking his rattle, though he will crawl quite happily on the normal floor. So it seems that babies at the crawling stage can appreciate a drop, probably from motion parallax when the baby moves its head, with innate knowledge of the danger of falling.
As adults, we cannot remember what it was like to enter the world and learn how to cope. Very soon we learn to control things; either by handling them, or by persuading others to do our bidding. So we may suppose technology derives from picking up dropped mugs, and using powers of symbols for persuading other minds (extended to controlling nature by magic) becomes how, as adults, we see and do from our early experiences as infants.
8.13 The ‘visual cliff’. In this famous experiment designed by Eleanor Gibson and R. D. Walk, babies or young animals are tempted to a visual drop, protected by a strong glass sheet. The baby refuses to crawl on the glass over the drop, so evidently sees the depth, innately associated with danger.
This chapter is called ‘Learning how to see’. We have considered several aspects, ending with the wonderful potentialities of human infants—as babies learn to see and understand and, far more than any other species, come to control the world. Sadly, some people forget how to see. This is so for what are called visual agnosias.
The word ‘agnosia’ was coined by Freud, meaning lack of knowledge for perception. It occurs with damage to regions of the brain involved in relating sensory signals, which may be entirely normal, to object knowledge. Agnosia may be only for vision, or for touch or hearing. Agnosias have great psychological interest, for they apply to inability to recognize objects though the eyes, or the ears or whatever, are working normally. It is meanings of sensory signals that are lost.
There can be specifically visual, or auditory, or touch agnosia—when objects cannot be recognized by one of the senses though they can by the others. Agnosias are associated with more or less specific brain damage, (behind the central sulcus). The key is loss of connecting knowledge for object recognition; so it is hardly misleading to call visual agnosias ‘forgetting how to see’.
The onset can be sudden. The classical case was described by Lissauer in 1890: an 80-year-old man (G.L.), who in a storm was blown against a wooden fence, knocking his head. He found that everything looked unfamiliar, and he mis-identified familiar objects—such as seeing pictures on the wall as boxes, and confusing his jacket with his trousers. He could, however, copy drawings; so his vision was normal, except for loss of meaning.
The most dramatic and insightful accounts are those of the neurologist Oliver Sacks, especially in The man who mistook his wife for a hat. His patient Dr P. was a gifted music teacher, who though retaining his musical gifts, gradually forgot how to see. After Dr P. confused his own foot with his shoe, Sacks showed him pictures in a magazine:
His responses were very curious. His eyes would dart from one thing to another, picking up tiny features, individual features, as they had done with my face. A striking brightness, a colour, a shape would arrest his attention and elicit comment—but in no case did he get the scene as a whole. He failed to see the whole, seeing only details, which he spotted like blips on a radar screen. He never entered into relations with the pictures as a whole—never faced, so to speak, its physiognomy. He had no sense whatever of a landscape or scene.
He would also ‘see’ things not in the picture: a river, a terrace, and coloured parasols which were not there. Yet he could see and recognize abstract shapes without difficulty. Faces were confused and expressions meaningless, though voices immediately identified the speakers. Then:
… he seemed to think he had done rather well. There was a hint of a smile on his face. He also appeared to have decided that the examination was over, and started to look round for his hat. He reached out his hand, and took hold of his wife’s head, tried to lift it off, to put it on. He had apparently mistaken his wife for his hat! His wife looked as if she was used to such things.
The various agnosias of vision, touch, hearing, and so on, confirm that the senses are separately processed by specialized regions of the brain. Yet normally the senses come together, to give unified perceptions of objects. How this happens is far from understood. The S.B. case of recovery of blindness (pages 153–159) seems to show that there is transfer of knowledge from one sense to another; in this case it was from touch to vision, as following the eye operation for blindness he could immediately identify letters, and tell the time from earlier touch experience of letters and the hands of his watch. He had learned how to see before he had sight. People with visual agnosia have sight but have forgotten how to see.
