The Measurement of Sensation: A Critique of Perceptual Psychophysics
Selective History of Theories of Visual Perception: 1650-1950
The Intelligent Eye
Eye and Brain: The Psychology of Seeing
The World Through Blunted Sight
If any part of psychology has enjoyed undisputed respectability it is the ritual of getting people to lift two small weights one after the other and say whether the second is heavier or lighter than the first. Or they can watch lights or listen to tones and make a judgment of relative brightness or loudness. Endless man-hours have been spent in these devoted exercises, all supposedly establishing numerical relations between physical dimensions like the amplitude of sound waves and sensory or psychological dimensions like loudness. As an impatient undergraduate putting in my stint on the treadmill, I comforted myself with the belief that this activity was a vestige, a hangover from the nineteenth century, soon to be displaced by real psychology. Today it still flourishes, claiming notable advances in method, statistical elaboration, and conceptual sophistication.
And now Mr. Savage, a philosopher at UCLA, denies that psychophysics measures psychological dimensions at all. Without questioning the value of what has been done he argues that the nature of the undertaking has been misconceived; his title is an irony since his theme is that what has been measured is neither sensation, as Fechner supposed in 1860, nor the psychological magnitudes that S. S. Stevens of Harvard has claimed to measure a hundred years later by different methods from Fechner’s.
It may be surprising that such an apparently arid topic still gains attention, but it forms part, perhaps the basis, of the interest in perception, and especially visual perception, which is very much alive at the present time among artists as well as philosophers and psychologists. Of course psychophysical measurement serves a diversity of purposes. Like any other academic pursuit it provides a medium in which reputations and careers are made. It produces, as spin-off, some refinements of practical techniques for assessing individuals’ sensory equipment, mainly sight and hearing. But eventually its nourishment is drawn from a perennial interest in the processes by which physical events such as light waves or the pull of gravity produce experience or, if we prefer, discriminatory response in a living organism.
Nicholas Pastore provides an extremely useful critical introduction to three centuries of observation and thinking about visual perception, especially the question whether our perception of depth, solidity, and distance has to be learned through touching, grasping, reaching—muscular exploration in general—or comes as an immediate visual datum, as immediate as two-dimensional shape. Associated with this question is the problem of the “constancies,” the fact that we recognize objects as the same even though differences of distance or tilt or illumination produce sharply different retinal images. The view that all these things depend on inference and learning prevailed throughout the greater part of the period Mr. Pastore surveys and has been maintained in one form or another by influential modern psychologists, whether in the conditioning theories of Hull and Hebb or in the “transactionalism” of Ames, whose work comes just at the end of the three centuries.
Ames’s demonstrations are striking and well-known, especially the trapezoidal window which rotates continuously in a clockwise direction but seems to alternate between clockwise and counterclockwise movement, and the distorted room which completely disorganizes the observer’s judgment of relative sizes and distances. The window is perhaps treated at non-conscious levels as if it were rectangular like most windows, and the room Ames constructed is certainly assumed to be rectangular (and looks so from the viewing point) when in fact its angles and the lengths of its parallel sides have been altered. In both cases assumptions that work well in familiar circumstances totally mislead us in unusual conditions. Yet such demonstrations with observers long habituated to the normal features of their environment provide no proof that our initial perceptions of distance or of movement in the third dimension have to be learned.
From time to time in the period under review this dominant opinion that perception of depth and solidity and distance had to be acquired, a view associated mainly with Berkeley, was challenged. It is illuminating to see how little headway the critics could make against orthodoxy, though much of the essential evidence for a different view was available; it was cited for instance by Samuel Bailey, of whose writings in the 1840s Mr. Pastore gives a useful account. Everyday observations of infants and the young of animals showed it was extremely unlikely that their perceptions of depth and distance had to be acquired by exploration. Evidence derived from the famous case reported by the surgeon Cheselden in 1728 in which a boy of thirteen, almost blind as long as he could remember, was given sight by being “couched” for cataract, though constantly quoted in support of Berkeley’s theory, could be interpreted differently, as Bailey showed. But the orthodox view was too strong:
It would appear that the selection and interpretation of evidence was guided by the conviction that Berkeleyan theory was “unshakable,” as Mill said in his 1842 review of Bailey, and “inexpugnable,” as Croom Robertson was to say in reply to James in 1888.
Later experimental work has shown that babies and young animals are capable of depth perception so early in life that it requires a determined act of skepticism to believe they have to learn it.
The Berkeleyan conviction, however, was supported by the highly reasonable Molyneux Premise, enunciated by Molyneux in his textbook on optics in 1692:
For distance of it self, is not to be perceived; for ‘tis a line (or a length) presented to our eye with its end toward us, which must therefore be only a point and that is invisible.
The parallel is with a blindfolded man being touched by the point of a rod; for all he knows the rod may be six inches long or six feet—the touch can give him no clue to length. It was around this unquestionably rational premise that the available evidence was interpreted, with selective emphasis upon what seemed consistent with it.
The history of the controversy is relevant to the question how much factual information philosophers need. In a broadcast interview (reprinted in The Listener, January 21, 1971) Gilbert Ryle was reminded that when his book The Concept of Mind appeared, Bertrand Russell said, “It was pointless to write about matters like sensation and perception if you weren’t thoroughly familiar with what neurologists, opticians, psychologists, and other empirical observers had said about them.” Ryle replied that everyday experience made one “very much at home in the field of perception,” but as for rods and cones in the eye, “I don’t pretend to know anything about them and, if I may speak a bit rudely, I don’t want to.” He could no doubt find support for his position in the fact that the essentials of the case against the Berkeleyan view of depth perception were readily available in the eighteenth century, and in the further fact that with the much greater accumulation of evidence now provided by experimental psychology the question still remains in some doubt.
For those who do want to know the recent evidence on visual perception from experimental psychology R.L. Gregory’s Eye and Brain gives a compact survey of the main facts, rather meager in its treatment of color vision but offering abundant material on the visual illusions, the constancies, and the projection of three-dimensional figures onto a flat surface. He has himself formulated a theory of the illusions—attractively comprehensive though not accepted by all psychologists—according to which they depend on our unwittingly taking the angles drawn on a flat surface as representations of the third dimension, other lines then being seen as if at greater or less distances and consequently being scaled as longer or shorter in accordance with the constancy phenomena.
Mr. Gregory’s gift for persuasive and vivid exposition, his ingenuity in devising apparatus, and his resourcefulness in finding and presenting illustrations made him an ideal choice for giving the Christmas Lectures of the Royal Institution, on which his more recent book The Intelligent Eye is based. These lectures, started by Faraday, were to be “adapted to a juvenile auditory,” but although Mr. Gregory’s book bears occasional traces of that origin it is meant for any intelligent layman and is in fact at some points rather difficult reading.
It goes much beyond the earlier book in speculation and the suggested range of its relevance, though a good deal of the same area is covered. There are discussions of the importance of pictorial representation, written language, and the use of measuring instruments in human development, leading up to the theme that more and more of our scientific thinking takes us beyond anything accessible to perception. A step on his way is the following example:
It is not difficult to demonstrate the fallibility of thinking based on perception, when the situation is beyond the range of sensory situations…. In imagination take a sheet of tissue paper a few thousandths of an inch thick and as large as you like. Now fold it in half. Now fold it again, so that the double thickness becomes four thicknesses. Repeat this folding fifty times. Now, how thick is the folded paper? People generally give an answer between one-half inch and three feet. But the answer is, that the thickness of paper would be about the distance of the sun! It seems quite absurd that all we have to do to reach the sun is fold a sheet of tissue paper fifty times, and yet this is indeed so. The mathematical notion of physical quantities increasing by a power law evidently is not well represented in our perceptual models, and intuition then fails dramatically.
From this he proceeds to the position of the physicist who, he says, has “great difficulty in thinking ‘intuitively’ in his subject” and must rely more and more on mathematical formulation and computer simulation. Measurement with instruments, moreover, gives us access to a great range of phenomena (for instance X-rays, radio waves, infrared rays, and non-sensory information about the stars provided by the radio telescope) which are not perceptible with our natural equipment, “so that we are now even more separated from our biological origins.” He concludes:
We have to learn to live with non-sensory data, and the resulting non-perceptual concepts of physics. We are left with a question: how far are human brains capable of functioning with concepts detached from sensory experience? Our future depends upon the answer.
It may. Some very remote crisis in the existence of the species might be surmounted by the resources that only this kind of thinking can tap. Nearer at hand it looks as though survival is more likely to depend on our capacity to use sensibly and decently the knowledge we already have. But this is not Mr. Gregory’s field and we must not complain that he directs our attention to more stimulating intellectual prospects.
After the gusto and enthusiasm of Mr. Gregory’s exposition Mr. Savage’s seems to belong to a more leisurely century, though he is in fact completely up to date and certainly not less exciting in his basic thinking. His preferred way of writing extends each stage of the argument with the utmost care to make its every detail explicit; points of view opposed to his own are phrased as scrupulously as they would be by their advocates, then rebutted, then presented in a revised form which might seem to meet his objections but which in turn is shown to be untenable. Every link in the chain of thought is exposed to inspection and its strength tested; summaries, recapitulations, and frequent references back are part of the technique of exposition. It makes for a very long, sometimes tedious but exceedingly honest book.
Basic to his thinking is the rejection of the commonly accepted distinction between physical and psychological dimensions. The physical scientist, dealing for instance with sound waves and their amplitude and frequency, leaves “loudness” and “pitch” to the psychologist:
Loudness, hardness, fragrance, and the like, are thus designated as psychological dimensions…. Remove perceivers, and loudness, hardness, and fragrance would cease to exist. These qualities are thus dependent on a subject, more precisely, on a nervous system (if one takes up a physicalist orientation), or on a mind (if a mentalist orientation is adopted). They are “subjective” or “psychological” dimensions, and they are the concern of the scientist of subjective phenomena, that is to say, the psychologist.
But this usual view, with its dichotomy between the “primary, physical, or real” dimensions and the “secondary, psychological, or apparent,” Savage finds unjustified.:
If we say that color and pitch are subjective, in virtue of their dependence on nervous systems or minds, consistency requires that we say the same of space, time, and mass. The reason for saying that color and pitch are subjective is that there would be no perception of, statements about, or knowledge concerning color and pitch without subjects. But it is equally true that there would be no perception of, statements about, or knowledge concerning space, time, and mass without subjects possessing nervous systems or minds. If a broad metaphysical distinction between “Appearance” and “Reality” is legitimate (it is not, in my opinion), there is just as much reason to place space and time in the category of Appearance as there is to place color and pitch there.
And in a very close discussion of measurement and of possible methods of measuring loudness he maintains that loudness “is measurable in the narrow, classical sense, that is, by a method employing a physical operation of addition.” Pitch, too, he believes to be measurable, though by different methods, and he concludes that:
Loudness and pitch are, indeed, measurable; but they are not psychological dimensions. Their measurement will produce physical laws, not psychophysical laws, not laws of mind and body.
Needless to say, the modern psychophysicists have discarded Fechner’s aim of establishing a numerical relation between physical and “mental” events. They would endorse Mr. Savage when he writes:
What perceptual psychophysics studies is not relations between dimensions of different types, but relations between perceptual responses and perceived stimuli. Consequently, it is misleading to describe psychophysics as the science of mind and body, because this description suggests that the science spans a gulf between two different worlds: the private mental world and the public physical world. When the distinction between response and stimulus is made the fundamental one, there is no suggestion of different worlds. The response, like the stimulus, is publicly observable and capable of scientific treatment. There is no gulf between response and stimulus.
This may seem like an ordinary behaviorist position, readily acceptable to Stevens and other present-day psychophysicists. But Mr. Savage has doubts. He sees Stevens as a Fechnerian in disguise, for Stevens speaks of “psychological magnitudes”:
Even when psychological magnitudes are behavioristically defined, they become substitutes for introspective magnitudes, for magnitudes of sensations, and take on many of the objectionable features of the latter. Stevens, like Fechner, still believes that psychophysics is the science of mind and body, though mind has become in his theory an uncertain and pallid replica of its old original self.
Stevens of course denies this, though it looks as if the challenge has led him to define his position more carefully and to move in Mr. Savage’s direction.
But Mr. Savage’s distinction between his own and the usual behaviorist position seems not to be stated as fully and clearly as most of his arguments, though it may be of fundamental importance. If we take literally his assertion “There is no gulf between response and stimulus,” we seem to go beyond the ordinary behaviorist account with its distinction between the organism and its environment. If there is no gulf here, then we seem to be treating the organism as an environmental event, possibly along the lines A. N. Whitehead suggested in 1925 in Science and the Modern World.
As far as psychophysics is concerned, Mr. Savage believes that the old difficulties disappear if we replace the concept of a psychological dimension with the concept of a perceptual ability:
Thus conceived, psychophysics does not require “psychological units” or any special methods of “psychological measurement.” Every procedure for measuring perception is then seen to consist in (a) measuring a group of physical stimuli, (b) eliciting perceptual responses to these stimuli from observers, and (c) processing the results by statistical methods. We may call such a procedure “psychological” if we choose, although it does not, strictly speaking, involve any procedure of measurement over and above the measurement of physical dimensions. The measurement of psychological dimensions is not involved.
And what then? Does any of this matter? Psychophysics will continue to flourish, undergraduates will be trained in it, Ph.D.s produced, publication lists extended. It might seem that Mr. Savage has been splitting hairs with infinite precision. But it may also be that he has made one of those fundamental distinctions that produce a decisive shift in outlook; if we fully grasp that the psychophysical methods, old or new, relate physical quantities to perceptual abilities without some intervening psychological dimension we may have altered our view of the kind of structure that the living organism is.
In comparison with the other books, Mr. Trevor-Roper’s collection of facts and suppositions about the effects of defective sight is bedtime reading, agreeable and pleasantly illustrated. It ranges widely, from the possible influence of short sight on children’s personality development, through the stages of perceptual disturbance produced by mescaline, to the symbolic relations between sex and sight, and the likelihood that St. Hildegard’s visions developed from the fortification figures and scintillating scotomata at the onset of a migraine attack.
The most interesting material, however, concerns painters with known or inferred defects of sight. The old story that El Greco’s elongation of his figures was due to astigmatism is analyzed skeptically, but the possibility that cataracts affected the colors used by several painters in their later work—Turner, Verrio, and Monet among them—is taken more seriously. Mr. Trevor-Roper is an ophthalmologist and the book is authoritative when he speaks from his professional knowledge; on wider issues the reliability of his sources varies, but at least he gives them all and we can judge for ourselves.
Psychophysical Research October 7, 1971