In their 1984 article "Can Indirect Realism be Demonstrated in the Psychological Laboratory?," Stephen Wilcox and Stuart Katz utilize the organizational structure of a clever reductio ad absurdum argument (first provided by E.B. Holt, 1914) to demolish the internal logic of the old "representative" doctrine of indirect realism.
More importantly, the article also indicates that "modern" (naive realist) adherents of indirect perception theories are equally vulnerable to the reductio because they assume the same underlying self-contradictory double epistemology -i.e., that while psychological researchers have unhindered access to their empirical apparatus and can make judgments about the veridicality of knowledge, their subjects are claimed to have no such direct access.
"the psychologist is now juggling with two epistemologies, not one: an indirect realism for the perceiver who cannot know the world directly, and a direct realism for himself who can. This has two important consequences. First, it refutes the thesis that indirect realism is an empirical matter, and second, it shows that the thesis is ultimately paradoxical" (p. 153).
The main argument presented in the article is that direct realism "stands on its own" philosophical merits as an epistemological position for the very reason that all indirect realist positions are self-refuting. The authors do acknowledge, however, that James J. Gibson's Ecological Approach to Visual Perception (1979) presents a psychological theory of direct perception which is both consistent with direct realism (philosophically speaking) and accounts for "how the ambient [visual] array specifies the environment" (see Footnote 2 on p. 156). In order to further contextualize this important footnoted comment, we need only mention parenthetically that Gibson's under-recognized Senses Considered as Perceptual Systems (1966) had also done the same for all of the other senses too.
Wilcox, S. & Katz, S. (1984). Can Indirect Realism be Demonstrated in the Psychological Laboratory? Philosophy of the Social Sciences, 14, 149-157.
Can Indirect Realism be Demonstrated in the Psychological Laboratory?[1]
STEPHEN WILCOX, Psychology, Franklin and Marshall College
STUART KATZ, Psychology, University of Georgia
"The psychological experimenter has his apparatus of lamps, tuning forks, and chronoscope, and an observer on whose sensations he is experimenting. Now the experimenter by hypothesis (and in fact) knows his apparatus immediately, and he manipulates it: whereas the observer (according to the theory) knows only his own 'sensations', is confined, one is requested to suppose, to transactions within his skull. But after time the two men exchange places: he who was the experimenter is now suddenly shut up within the range of his 'sensations', he has now only 'representative' knowledge of the apparatus; whereas he who was the observer forthwith enjoys a windfall of omniscience. He now has an immediate experience of everything around him, and is no longer confined to the sensations within his skull. Yet, of course, the mere exchange of activities has not altered the knowing process in either person. The representative theory has become ridiculous.... In plain fact the experience of both experimenter and observer is at all times immediate. The real objects, and no 'sensations' thereof, are their two experiences. When the observer says that he has a 'sensation' of so-and-so, he means merely that it is so-and-so much, certain portion, and not another, of the objects that lie about him at the moment, which is in his experience.... In short, there is no sensation of an object. Experience presents no object once as outer and again as inner fact, and no content of knowledge that is other than its object...." (Edwin B. Holt, The Concept of Consciousness, 1914, pp. 149-150).
What exactly can the psychologist tell us about epistemology? It is becoming fashionable to believe that he can tell us a great deal. And what he can apparently say is this: there are certain sorts of scientific data that prove the truth of a particular theory of knowledge called indirect or representative realism. The assertion is certainly not new; it is as old at least as modern science or modern philosophy, and in essentials probably much older. But we shall only be concerned here with its most recent expressions. From those of our contemporaries who have been especially [p. 150] concerned with the relation of psychology to epistemology there have been firm pronouncements in support of this thesis. Quine, for example, believes that the theory of knowledge has, at long last, been 'naturalized'. Epistemology, or something like it', he argues, 'simply falls into place as a chapter of psychology and hence of nature science' (1969, p. 82). Piaget named a new discipline genetic epistemology, which, he said, '... attempts to explain knowledge... on the basis of its history, sociogenesis, and especially the psychological origins of the notions and operations on which it is based' (1970, p. 1). The neurophysiologist Maturana has said of cognition, it '...is a biological phenomenon and can only be understood as such; any epistemological insight in the domain of knowledge requires this understanding' (1970, p. 5). McCulloch (1965) also named a discipline, and he called it 'experimental epistemology, whose purpose is the mechanization, and presumably the understanding, of the knowing process.
It would be a mistake, however, to think that it is only a select few who have brought psychology into epistemology, or to put it exactly, have converted epistemology into psychology. Many psychologists today, though probably unaware of Quine's assertion, are engaged in the conduct of research whose purpose it is to elaborate the particular theory of knowledge we have named above. It is our purpose in this paper to show the difficulties to which this view leads. We will require three steps to do this. The first is a preliminary, where we summarize for the reader the various classes of scientific fact that are taken as compelling evidence for indirect realism. In the second we explain the paradoxical nature of the thesis that facts prove this epistemology. And in the third we suggest how this difficulty can be avoided. The epigraph [from E. B. Holt, 1914] with which we have introduced this paper, written nearly 80 years ago, shall be our guide. It, too, has a tripartite structure of statement, paradox, and solution. It will be our task to fill out and embellish that structure.
1. THE PSYCHOLOGICAL EVIDENCE FOR INDIRECT (REPRESENTATIVE) REALISM
"The psychological experimenter has his apparatus of lamps, tuning forks, and chronoscope, and an observer on whose sensations he is experimenting. Now the experimenter by hypothesis... knows his apparatus immediately, and he manipulates it: whereas the observer... knows only his own 'sensations', is confined... to transactions within his skull."
What facts does the psychologist-epistemologist concern himself with? These can be divided into two types: (1) those that arise from the study of the behavior of organisms, what perhaps we might call psychological' facts proper, and (2) those that arise either from the study of the morphology of sense organs or from the study of other aspects of the world (i.e., the environment). Let us cll these facts 'structural. The [p. 151] psychological facts include illusions, the relativity of perception, perceptual constancy, imagination, dreams and hallucinations. The structural facts include, in the example of vision, the camera-like structure of the vertebrate eye (specifically the existence of the retina) and the characteristics of the physical media to which perceivers are said to have immediate' access, such as waves, photons or other 'primary' physical entities, to use Locke's term.
Both types of facts work together to demonstrate a discrepancy between what is perceived and what, the psychologist assumes, the perceiver directly senses. Consider one class of phenomena, the perceptual illusions. When experimental subjects are shown contrived two dimensional configurations, some lines are reported to be longer, or more bowed, or more tilted than they really are. Stereoscopically presented random patterns give the appearance of depth, when, of course, there is not depth. Some configurations give the appearance of subjective contours, others generate ambiguous figures, still others impossible figures, none of which are what they seem to be. There are even ingeniously designed rooms that when peered into make objects look much larger or smaller than they really are. These experiments, and hundreds like them, are taken to mean that perceivers make errors and therefore cannot perceive that world as it is.
Anther class of phenomena are the so-called perceptual constancies, where perceivers report that the shape or size of an object appears relatively unchanged under rigid transformation. Since the perceived size or shape of the object remains unchanged, relative to the object itself, the perceiver is adjudged accurate; but, as in the case of illusions, the theorist still assumes that there is a discrepancy between what is perceived and the sensuous data (in this case the retinal images) to which the perceiver has immediate access.
Whatever the particular psychological phenomenon, however, an attribution of discrepancy could not be made without the second of the two sorts of observations we have mentioned, namely the structural ones. Scientists since Descartes have been impressed by the anatomical and physical acts which have appeared to constrain what is the immediate or proximate cause of perception. The facts vary in importance, depending on what level of analysis the scientist chooses to begin with. For example, among the anatomical facts of vision, the proximal cause of the visual experience may either be the two dimensional image on the retina, or streams of photons striking a mosaic of photoreceptors. In both cases, the proximal stimulus is something less than the complete object of the world of experience. This viewpoint is given further credibility by a conception of the world derived from eighteenth and nineteenth century physics wherein macro-objects are thought to e conveyed to the perceiver in a dismantled form, either as waves, or streams of photons, or small mechanical forces, or clouds of molecules. [p. 152]
These structural constraints set the task for the perceiver; he must reconstruct the world of macro-objects if he is to perceive veridically. The psychological facts to which we have already referred seem to confirm this. The Necker cube, which is actually a two dimensional configuration in the world, and thence on the retina, is seen as having depth, and is seen in depth in one of two ways. This must be a reconstruction, the argument goes, though a reconstruction in error. An object translated from near space to a more distant location is perceived as roughly the same size in either location, which of course it is. But the image shrinks on the retina and, so the argument goes, the perception must therefore be a reconstruction, since the shrinking image is all the perceiver apprehends directly.
These are but a small sample form a vast and ever-growing body of scientific knowledge about perception, but what we have said will suffice for our purposes. Structural fact, the thesis goes, give us a picture of what entities in the world the perceiver confronts immediately, and what anatomical features in the perceiver make contact, as it were, with those entities. The psychological facts suggest to us what the perceiver does with the information he has. Since that information is incomplete, the perceiver must add to it, i.e., he must reconstruct perceptual world commensurate with the external world of macro-objects and events. This conclusion is in fact what we were looking for, for it is precisely the widely held theory of knowledge called indirect realism. According to it, the perceiver comes to know the world 'indirectly', or 'mediately' by using a collection of atomic elements to reconstruct the world through some form of quasi-rational inference or schematism. Virtually every modern theory of perception or cognition can be thus described. We have 'unconscious inference' (Helmholtz, 1867), 'schema' formation (Bartlett, 1932), 'hypothesis' testing (Gregory, 1977), 'perceptual atom theory' (Julesz and Schumer, 1981), 'analysis by synthesis' or later the 'perceptual cycle' (Neisser, 1967, 1976), 'assimilation/accommodation' (Piaget, 1963), and uncountably many variations on these. But having now arrived at our sough-for theory of knowledge we have also arrived at our sought-for thesis, since indirect realism has been established 'scientifically'.
In our view indirect realism has serious defects (e.g., see Hirst, 1959; Ryle, 1949), but in what follows, we shall be concerned not with these defects per se, but with the implications of the theory for the things psychologists do, or claim to do, and the conclusions they draw about what they do -or in other words with the thesis that indirect realism is scientifically demonstrable.
2. THE THESIS IS PARADOXICAL
"But after time the two men exchange places: he who was the experimenter is now suddenly shut up within the range of his 'sensations', he has now only [p. 153] 'representative' knowledge of the apparatus; whereas he who was the observer forthwith enjoys a windfall of omniscience. He now has an immediate experience of everything around him, and is no longer confined to the sensations within his skull. Yet, of course, the mere exchange of activities has not altered the knowing process in either person. The representative theory has become ridiculous...."
According to indirect realism, perceivers are epistemically isolated, for they can only know the external world indirectly through sensation, which is itself fragmentary and incomplete. This is, so to speak, the subjective side of the doctrine. But there is also an objective side. The doctrine tells us that even if the perceiver arrives at knowledge of the world indirectly, he nonetheless arrives at it, guesswork or not, occasional error or not. Now we should like to know why the theorist would wish to make this claim, for there is nothing in the theory itself which permits that assertion. The perceiver, after all, reconstructs the world knowing nothing of what it is like in actuality. How then is it known that he arrives at it in actuality? The answer, of course, lies with the psychologist himself. He knows that the perceiver apprehends the world veridically because he himself makes that judgment. He knows when the perceiver commits an error, for he can detect the discrepancy between the perceiver's point of view and the world as it is. He knows when the perceiver is correct, for he will then find no discrepancy between the point of view of the perceiver and the world as it is.
But if that is the case, then the psychologist is now juggling with two epistemologies, not one: an indirect realism for the perceiver who cannot know the world directly, and a direct realism for himself who can. This has two important consequences. First, it refutes the thesis that indirect realism is an empirical matter, and second, it shows that the thesis is ultimately paradoxical. Let us look at these in turn.
1. According to the thesis we have been examining, indirect realism is matter to e proved or disproved by scientific investigation. Now, however, we have discovered a case to the contrary, for while indirect realism for the perceiver was arrived at by virtue of psychological and other data, it had to be done by presupposing for the psychologist, without such facts, direct apprehension of the external world. We shall call this second epistemology a direct realism epistemology, and we shall say that it is assumed by the psychologist in order to arrive at an indirect realist epistemology. The thesis we have been examining, then, cannot be correct because an epistemology (ironically the antithesis of the one proved) is assumed in order to interpret the facts and not the reverse. The psychologist has committed the most common of philosophical follies: he has failed to examine his won presuppositions.
2. The acceptance, at one and the same time, of both indirect realism for the perceiver and direct realism for the psychologist is logically absurd. An the reason is that the two actors in this play can 'exchange places with the other; psychologist can become perceiver, and [p. 154] perceiver, psychologist. But if, on the other hand, the psychologist is a perceiver, then he must be under the constraints of indirect realism. He cannot, therefore, know the world directly, and if that is the case, he cannot himself decide what is perceptual error and what is not. And of course, if he cannot do that, then he cannot assert that indirect realism itself is true. In the last analysis, he can say nothing at all, for as a perceiver, he cannot escape from his sensations to the external world. He is himself a victim of the egocentric predicament.
On the other hand, at least some perceivers, namely members of the species Homo sapiens (and perhaps other creatures as well) can each become a psychologist, and even a certain sort of psychologist with philosophical pretensions to indirect realism. But if that is so, each forthwith enjoys a windfall of 'omniscience', for he can now make judgments about whether others perceive what is actual. And if he can do that, he must know a great deal about the external world, so much so in act that he can no longer be a perceiver in the indirect realist sense. Each role then, perceiver and psychologist, contradicts the other, for one and the same knower is asserted to know the world mediately (qua perceiver) and immediately (qua psychologist).
The paradox appears more striking if we consider the opinions of certain psychologists who, on the basis of psychological evidence, argue that the external world cannot even be indirectly known to the perceiver, for it cannot be known at all! Consider the following: Jerison (1976), a brain theorist who has studied the evolutionary development of the nervous system, speculates that in the course of natural history, new species evolved to perceive the world in new ways. On the basis of this, he reaches the surprising conclusion that: 'Reality, or the real world we know intuitively, is a creation of the nervous system... the work of the brain is to create a model of a possible world rather than to record and transmit to the mind a world that is metaphysically true' (p. 99). Von Foerster (1973), citing examples of illusion and neurophysiology, concludes that cognition is a computational activity involving an infinite recursion of descriptions. "This formula', he goes on to say, 'has the advantage that one unknown, namely "reality", is successfully eliminated' (p. 39). And likewise von Glasersfeld (1977), referring to the literature on illusions, asserts 'Nothing is easier than to show that our senses are unreliable and that what we perceive is subjective' (p. 2).
Once again we are led to the paradox, and with a vengeance. On the one hand, the sceptical psychologist urges us that psychological, neurophysiological, and 'paleontological' evidence gives us reason to doubt the existence of the external world. On the other hand, he is most assured of the import of the facts he uses to arrive at this conclusion, and in the authenticity of other persons to whom he communicates these facts. But of course if his thesis is correct, then all facts must be private [p. 155] fancy; his observations of what perceivers do and know, and indeed of the existence of the perceivers themselves, must be his own private doings. He can no more be certain of the truth of his own assertions than can the epistemically isolated perceiver he is studying. And yet he propounds with an authority and certainty what, by his own lights, he ought to doubt.
The problem, once more, is that the sceptical psychologist, like the proponent of indirect realism, has two epistemologies; in this case he assigns a thoroughgoing subjectivism to the perceiver, and accepts a direct realism for himself. The experience of the perceiver is private, but the psychologist asserts this on the basis of knowledge he himself regards as public. Yet as before, we know that the psychologist and the perceiver can exchange roles, and because they can do this we find that one and the same knower can be conjointly ultra-sceptical and absolutely certain about the origins of his experience.
3. RESOLVING THE PARADOX
"In plain fact the experience of both experimenter and observer is at all times immediate. The real objects, and no 'sensations' thereof, are their two experiences. When the observer says that he has a 'sensation' of so-and-so, he means merely that it is so-and-so much, certain portion, and not another, of the objects that lie about him at the moment, which is in his experience.... In short, there is no sensation of an object. Experience presents no object once as outer and again as inner fact, and no content of knowledge that is other than its object...."
Two lessons can be learned form the foregoing discussion. Of the first we can be brief: there is no one-way street from facts to epistemology. Facts themselves require interpretation, and as we have seen, the facts purported to prove indirect realism require direct realism as a backdrop against which the psychologist takes the measure of his observations. In general, it is not possible to employ scientific methods without the presuppositions that allow interpretations of those methods. The psychologist who asserts the thesis we have considered in this paper has left his presuppositions at home, and as n unfortunate consequence, thinks he does not have any.
The second lesson concerns the paradox itself. The thesis, we have seen, leads to the collocation of two incompatible epistemologies in a single knower who can freely play the role of perceiver or scientist as he chooses. How can this be resolved? The obvious solution would be to eliminate one or the other epistemology, for this, prima facie, would eliminate the inherent contradiction. If we reject direct realism, however, and accept indirect realism, we shall have solved nothing, because a psychologist to whom indirect realism applies could not, by our earlier argument, make assertions of error or truth for any perceiver, nor even for himself. If he cannot know what the perceiver cannot know, he can [p. 156] hardly make claims about perceptual verite, for he is in no better position than the perceiver to make such claims.
If we choose direct realism, the consequence are much better.[2] The psychologist may now see himself as sharing with every other knower a common reality which both perceive directly. He need not, therefore, play both seer and blind man, with the accompanying contradiction of that dual role. We have, in sum, one epistemology, not two; one knower, not two, though the knower can play the part of observer or perceiver. The result is a consistency and parsimony we did not have before. As for direct realism itself, there are certainly many questions to be asked about it, and much more to be said concerning it. It is, however, beyond the scope of this paper to enter into such an analysis. For those interested in pursuing this point of view we suggest Gibson (1979) and Shaw, Turvey, and Mace (in press). It would be appropriate, nonetheless, to conclude with the following remark. To assert direct realism is to deny the division between the objects of special private status called 'sensations' or 'sense data' and the properties o the external world which, according to indirect realism, can only be inferred from sensations. There are, in other words, no objects 'as outer and again as inner fact'; there are only the 'objects' themselves. If the perceiver at any one time perceives only a part of what he might perceive, or if he perceives 'incorrectly', it is because discovery of the world itself is always a partial, and never a fully completed process. Direct realism does not imply omniscience, nor a process of discovery that is faultless, as is sometimes suggested. It does imply immediacy and, with due regard for one's point of view, it also implies verity.
REFERENCES
Bartlett, F. C (1932). Remembering: A Study in Experimental Social Psychology. Cambridge.
Gibson, J.J. (1967). 'New Reasons for Realism', Synthese, 17, 162-72.
Gibson, J.J. (1979). The Ecological Approach to visual Perception. Boston.
Gregory, R.L. (1977) Eye and Brain: The Psychology of Seeing. New York.
[p. 157]
Helmholtz, H. von (1867) Handbook of Physiological Optics. London.
Hirst, R. J. (1959) The Problems of Perception. London.
Holt, E. B. (1914) The Concept of Consciousness. New York.
Jerison, H. (1976) 'Paleoneurology and the Evolution of Mind', Scientific American, 24, 90-101.
Julesz, B. and Schumer, R. (1981) 'Early Visual Perception', Annual Review of Psychology, 32, 575-627.
Maturana, H. (1970) 'Biology of Cognition', Biological Computer Laboratory Report number 9.0. Urbana-Champaign.
McCulloch, W. S. (1965) Embodiments of Mind. Cambridge, Mass.
Neisser, U. (1967) Cognitive Psychology. New York.
Neisser, U. (1976) Cognition and Reality. San Francisco.
Piaget, J. (1963) The Origins of Intelligence in Children. New York.
Piaget, J. (1970) Genetic Epistemology. New York.
Quine, W. (1969) Epistemology Naturalized', in W. Quine, Ontological Relativity and Other Essays. New York.
Ryle, G. (1949) The Concept of Mind. London.
Shaw, R., Turvey, M., and Mace, W. (forthcoming) 'Ecological Psychology: The Consequences of a Commitment to realism', in W. Weimer and D. Palermo (eds.), Cognition and the Symbolic Processes II. Hillsdale, N.J.
von Foerster, H. (1973) 'On Constructing a Reality', in W. F. E. Presider (ed.), Environmental Design Research, vol. 2, Stroudsburg.
von Glasersfeld, E. (1977) 'Notes on the Epistemological Revolution', paper presented to the SGSR/AAAS Symposium. Denver.
Footnote-[1]: We would like to thank Edward Reed and an anonymous reviewer for helpful comments on earlier versions of this paper....
Footnote-[2]: As our remarks indicate, we see the belief that epistemology can be grounded in psychology as an obstacle to direct realism. Inorganically, Gibson (1967) presents his psychological theory as a new reason for [direct] realism. The essence of Gibson's agreement, however, is not that he has discovered new facts which lead to realism. Rather, he shows how the old arguments against realism can be refuted by reconceptualizing the problem of specificity (cf. Gibson 1979, p. 115). For the typical indirect realist, as for Johannes Muller before him, the energy array is only capable of specifying the states of peripheral nerve endings. For Gibson, on the other hand, the energy array contains 'higher-order' properties (i.e., invariants) defined over space and time which are specific to the meaningful world. We, perhaps, depart from Gibson and some of his followers in our belief that direct realism stands on its own. In our view, an account of how the ambient array specifies the environment is an interesting adjunct, but not a crucial precondition for direct realism.