SENSE PERCEPTION AS KNOWLEDGE

I mentioned incidentally in the first section that it is conceivable that failure to give adequate consideration to practical judgments may have a compromising effect upon the consideration of other types. I now intend to develop this remark with regard to sense perception as a form of knowledge. The topic is so bound up with a multitude of perplexing psychological and epistemological traditions that I have first to make it reasonably clear what it is and what it is not which I propose to discuss. I endeavored in an earlier series of papers [90] to point out that the question of the material of sense perception is not, as such, a problem of the theory of knowledge at all, but simply a problem of the occurrence of a certain material—a problem of causal conditions and consequences. That is to say, the problem presented by an image [91] of a bent stick, or by a dream, or by "secondary" sensory qualities is properly a problem of physics—of conditions of occurrence, and not of logic, of truth or falsity, fact or fiction. That the existence of a red quale is dependent upon disturbances of a certain velocity of a medium in connection with certain changes of the organism is not to be confused with the notion that red is a way of knowing, in some more or less adequate fashion, some more "real" object or else of knowing itself. The fact of causation—or functional dependence—no more makes the quale an "appearance" to the mind of something more real than itself or of itself than it makes bubbles on the water a real fish transferred by some cognitive distortion into a region of appearance. With a little stretching we may use the term appearance in either case, but the term only means that the red quale or the water-bubble is an obvious or conspicuous thing from which we infer something else not so obvious.

This position thus freely resumed here needs to be adequately guarded on all sides. It implies that the question of the existence or presence of the subject-matter of even a complex sense perception may be treated as a question of physics. It also implies that the existence of a sense perception may be treated as a problem of physics. But the position is not that all the problems of sense perception are thereby exhausted. There is still, on the contrary, the problem of the cognitive status of sense perception. So far from denying this fact, I mean rather to emphasize it in holding that this knowledge aspect is not to be identified—as it has been in both realistic and idealistic epistemologies—with the simple occurrence of presented subject-matter and with the occurrence of a perceptive act. It is often stated, for example, that primitive sense objects when they are stripped of all inferential material cannot possibly be false—but with the implication that they, therefore, must be true. Well, I meant to go this statement one better—to state that they are neither true nor false—that is, that the distinction of true-or-false is as irrelevant and inapplicable as to any other existence, as it is, say, to being more than five feet high or having a low blood pressure. This position when taken leaves over the question of sense perception as knowledge, as capable of truth or falsity. It is this question, then, which I intend to discuss in this paper.

My first point is that some sense perceptions at least (as matter of fact the great bulk of them), are without any doubt forms of practical judgment—or, more accurately, are terms in practical judgments as propositions of what to do. When in walking down a street I see a sign on the lamp-post at the corner, I assuredly see a sign. Now in ordinary context (I do not say always or necessarily) this is a sign of what to do—to continue walking or to turn. The other term of the proposition may not be stated or it may be; it is probably more often tacit. Of course, I have taken the case of the sign purposely. But the case may be extended. The lamp-post as perceived is to a lamp-lighter a sign of something else than a turn, but still a sign of something to be done. To another man, it may be a sign of a possible support. I am anxious not to force the scope of cases of this class beyond what would be accepted by an unbiased person, but I wish to point out that certain features of the perceived object, as a cognitive term, which do not seem at first sight to fall within this conception of the object, as, an intellectual sign of what to do, turn out upon analysis to be covered by it. It may be said, for example, that our supposed pedestrian perceives much besides that which serves as evidence of the thing to be done. He perceives the lamp-post, for example, and possibly the carbons of the arc. And these assuredly do not enter into the indication of what to do or how to do it.

The reply is threefold. In the first place, it is easy—and usual—to read back into the sense perception more than was actually in it. It is easy to recall the familiar features of the lamp-post; it is practically impossible—or at least very unusual—to recall what was actually perceived. So we read the former into the latter. The tendency is for actual perception to limit itself to the minimum which will serve as sign. But, in the second place, since it is never wholly so limited, since there is always a surplusage of perceived object, the fact stated in the objection is admitted. But it is precisely this surplusage which has not cognitive status. It does not serve as a sign, but neither is it known, or a term in knowledge. A child, walking by his father's side, with no aim and hence no reason for securing indications of what to do, will probably see more in his idle curiosity than his parent. He will have more presented material. But this does not mean that he is making more propositions, but only that he is getting more material for possible propositions. It means, in short, that he is in an aesthetic attitude of realization rather than in a cognitive attitude. But even the most economical observer has some aesthetic, non-cognitive surplusage.[92] In the third place, surplusage is necessary for the operation of the signifying function. Independently of the fact that surplusage may be required to render the sign specific, action is free (its variation is under control) in the degree in which alternatives are present. The pedestrian has probably the two alternatives in mind: to go straight on or to turn. The perceived object might indicate to him another alternative—to stop and inquire of a passer-by. And, as is obvious in a more complicated case, it is the extent of the perceived object which both multiplies alternative ways of acting and gives the grounds for selecting among them. A physician, for example, deliberately avoids such hard-and-fast alternatives as have been postulated in our instance. He does not observe simply to get an indication of whether the man is well or ill; but in order to determine what to do he extends his explorations over a wide field. Much of his perceived object field is immaterial to what he finally does; that is, does not serve as sign. But it is all relevant to judging what he is to do. Sense perception as a term in practical judgment must include more than the element which finally serves as sign. If it did not, there would be no perception, but only a direct stimulus to action.[93]

The conclusion that such perceptions as we have been considering are terms in an inference is to be carefully discriminated from the loose statement that sense perceptions are unconscious inferences. There is a great difference between saying that the perception of a shape affords an indication for an inference and saying that the perception of shape is itself an inference. That definite shapes would not be perceived, were it not for neural changes brought about in prior inferences, is a possibility; it may be, for aught I know, an ascertained fact. Such telescoping of a perceived object with the object inferred from it may be a constant function; but in any case the telescoping is not a matter of a present inference going on unconsciously, but is the result of an organic modification which has occurred in consequence of prior inferences. In similar fashion, to say that to see a table is to get an indication of something to write on is in no way to say that the perception of a table is an inference from sensory data. To say that certain earlier perceived objects not having as perceived the character of a table have now "fused" with the results of inferences drawn from them is not to say that the perception of the table is now an inference. Suppose we say that the first perception was of colored patches; that we inferred from this the possibility of reaching and touching, and that on performing these acts we secured certain qualities of hardness, smoothness, etc., and that these are now all fused with the color-patches. At most this only signifies that certain previously inferred qualities have now become consolidated with qualities from which they were formerly inferred. And such fusion or consolidation is precisely not inference. As matter of fact, such "fusion" of qualities, given and formerly inferred, is but a matter of speaking. What has really happened is that brain processes which formerly happened successively now happen simultaneously. What we are dealing with is not a fact of cognition, but a fact of the organic conditions of the occurrence of an act of perception.

Let us apply the results to the question of sense "illusions." The bent reed in the water comes naturally to mind. Purely physical considerations account for the refraction of the light which produces an optical image of a bent stick. This has nothing to do with knowledge or with sense perception—with seeing. It is simply and wholly a matter of the properties of light and a lens. Such refractions are constantly produced without our noting them. In the past, however, light refracted and unrefracted has been a constant stimulus to responsive actions. It is a matter of the native constitution of the organism that light stimulates the eyes to follow and the arms to reach and the hands to clutch and handle. As a consequence, certain arrangements of reflected and refracted light have become a sign to perform certain specific acts of handling and touching. As a rule, stimuli and reactions occur in an approximately homogeneous medium—the air. The system of signs or indexes of action set up has been based upon this fact and accommodated to it. A habit or bias in favor of a certain kind of inference has been set up. We infer from a bent ray of light that the hand, in touching the reflecting object, will, at a certain point, have to change its direction. This habit is carried over to a medium in which the conclusion does not hold. Instead of saying that light is bent—which it is—we infer that the stick is bent: we infer that the hand could not protract a straight course in handling the object. But an expert fisherman never makes such an error in spearing fish. Reacting in media of different refractive capacities, he bases his signs and inferences upon the conditions and results of his media. I see no difference between these cases and that of a man who can read his own tongue. He sees the word "pain" and infers it means a certain physical discomfort. As matter of fact, the thing perceived exists in an unfamiliar medium and signifies bread. To the one accustomed to the French language the right inference occurs.[94] There is neither error nor truth in the optical image: It just exists physically. But we take it for something else, we behave to it as if it were something else. We mis-take it.

So far as I can see, the pronounced tendency to regard the perceived object as itself the object of a peculiar kind of knowledge instead of as a term in knowledge of the practical kind has two causes. One is the confirmed habit of neglecting the wide scope and import of practical judgments. This leads to overlooking the responsive act as the other term indicated by the perception, and to taking the perceived object as the whole of the situation just by itself. The other cause is the fact that because perceived objects are constantly employed as evidence of what is to be done—or how to do something—they themselves become the objects of prolonged and careful scrutiny. We pass naturally and inevitably from recognition to observation. Inference will usually take care of itself if the datum is properly determined. At the present day, a skilled physician will have little difficulty in inferring typhoid instead of malaria from certain symptoms provided he can make certain observations—that is, secure certain data from which to infer. The labor of intelligence is thus transferred from inference to the determination of data, the data being determined, however, in the interests of inference and as parts of an inference.

At this point, a significant complication enters in. The ordinary assumption in the discussion of the relation of perceived objects to knowledge is that "the" object—the real object—of knowledge in perception is the thing which caused the qualities which are given. It is assumed, that is, that the other term of a proposition in which a sense datum is one term must be the thing which produced it. Since this producing object does not for the most part appear in ordinary sense perception, we have on our hands perception as an epistemological problem—the relation of an appearance to some reality which it, somehow, conceals rather than indicates. Hence also the difficulties of "reconciling" scientific knowledge in physics where these causes are the terms of the propositions with "empirical" or sense perception knowledge where they do not even appear. Here is where the primary advantage of recognizing that ordinary sense perceptions are forms of practical judgment comes in. In practical judgments, the other term is as open and aboveboard as is the sensory quality: it is the thing to be done, the response to be selected. To borrow an illustration of Professor Woodbridge's: A certain sound indicates to the mother that her baby needs attention. If she turns out to be in error, it is not because sound ought to mean so many vibrations of the air, and as matter of fact doesn't even suggest air vibrations, but because there is wrong inference as to the act to be performed.

I imagine that if error never occurred in inferences of this practical sort the human race would have gone on quite contented with them. However that may be, errors do occur and the endeavor to control inference as to consequences (so as to reduce their likelihood of error) leads to propositions where the knowledge-object of the perceived thing is not something to be done, but the cause which produced it. The mother finds her baby peacefully sleeping and says the baby didn't make the noise. She investigates and decides a swinging door made it. Instead of inferring a consequence, she infers a cause. If she had identified the noise in the first place, she would have concluded that the hinges needed oiling.

Now where does the argument stand? The proper control of inference in specific cases is found (a) to lie in the proper identification of the datum. If the perception is of a certain kind, the inference takes place as a matter of course; or else inference can be suspended until more adequate data are found, and thus error is avoided even if truth be not found. Furthermore (b) it is discovered that the most effective way of identifying datum (and securing adequate data) is by inference to its cause. The mother stops short with the baby and the door as causes. But the same motives which made her transfer her inference from consequences to conditions are the motives which lead others to inferring from sounds to vibrations of air. Hence our scientific propositions about sensory data. They are not, as such, about things to do, but about things which have been done, have happened—"facts." But they have reference, nevertheless, to inferences regarding consequences to be effected. They are the means of securing data which will prevent errors which would otherwise occur, and which facilitate an entirely new crop of inferences as to possibilities—means and ends—of action. That scientific men should be conscious of this reference or even interested in it is not at all necessary, for I am talking about the logic of propositions, not about biography nor psychology. If I reverted to psychology, it would be to point out that there is no reason in the world why the practical activity of some men should not be predominantly directed into the pursuits connected with discovery. The extent in which they actually are so directed depends upon social conditions.

III

We are brought to a consideration of the notion of "primitive" sense data. It was long customary to treat the attempt to define true knowledge in terms derived from sense data as a confusion of psychology—or the history of the growth of knowledge—with logic, the theory of the character of knowledge as knowledge. As matter of fact, there is confusion, but in the opposite direction. The attempt involved a confusion of logic with psychology—that is, it treated a phase of the technique of inference as if it were a natural history of the growth of ideas and beliefs.

The chief source of error in ordinary inference is an unrecognized complexity of data. Perception which is not experimentally controlled fails to present sufficiently wide data to secure differentia of possible inferences, and it fails to present, even in what is given, lines of cleavage which are important for proper inference. This is only an elaborate way of saying what scientific inquiry has made clear, that, for purposes of inference as to conditions of production of what is present, ordinary sense perception is too narrow, too confused, too vivid as to some quales and too blurred as to some others. Let us confine our attention for the moment to confusion. It has often been pointed out that sense qualities being just what they are, it is illegitimate to introduce such notions as obscurity or confusion into them: a slightly illuminated color is just as irretrievably what it is, as clearly itself, as an object in the broad glare of noonday. But the case stands otherwise when the quale is taken as a datum for inference. It is not so easy to identify a perceived object for purposes of inference in the dusk as in bright light. From the standpoint of an inference to be effected, the confusion is the same as an unjustifiable simplification. This over-simplification has the effect of making the quale, as a term of inference, ambiguous. To infer from it is to subject ourselves to the danger of all fallacies of ambiguity which are expounded in the textbooks. The remedy is clearly the resolution, by experimental means, of what seems to be a simple datum into its "elements." This is a case of analysis; it differs from other modes of analysis only in the subject-matter upon which it is directed, viz., something which had been previously accepted as a simple whole. The result of this analysis is the existence as objects of perception of isolated qualities like the colors of the spectrum scientifically determined, the tones of the scale in all their varying intensities, etc., in short, the "sensations" or sense qualities of contemporary psychology textbooks or the "simple ideas" of sensation of Locke or the "objects of sense" of Russell. They are the material of sense perception discriminated for the purpose of better inferences.

Note that these simple data or elements are not original, psychologically or historically; they are logical primitives—that is, irreducible for purposes of inference. They are simply the most unambiguous and best defined objects of perception which can be secured to serve as signs. They are experimentally determined, with great art, precisely because the naturally given, the customary, objects in perception have been ambiguous or confused terms in inference. Hence they are replaced, through experimental means involving the use of wide scientific knowledge deductively employed, by simpler sense objects. Stated in current phraseology, "sensations" (i.e., qualities present to sense) are not the elements out of which perceptions are composed, constituted, or constructed; they are the finest, most carefully discriminated objects of perception. We do not first perceive a single, thoroughly defined shade, a tint and hue of red; its perception is the last refinement of observation. Such things are the limits of perception, but they are final, not initial, limits. They are what is perceived to be given under the most favorable possible conditions; conditions, moreover, which do not present themselves accidentally, but which have to be intentionally and experimentally established, and detection of which exacts the use of a vast body of scientific propositions.

I hope it is now evident what was meant by saying that current logic presents us not with a confusion of psychology with logic, but with a wholesale mistaking of logical determinations for facts of psychology. The confusion was begun by Locke—or rather made completely current through the enormous influence exercised by Locke—and some reference to Locke may be of aid in clearing up the point. Locke's conception of knowledge was logical, not psychological. He meant by knowledge thoroughly justified beliefs or propositions, "certainty," and carefully distinguished it from what passed current as knowledge at a given time. The latter he called "assent," opinion, belief, or judgment. Moreover, his interest in the latter was logical. He was after an art of controlling the proper degree of assent to be given in matters of probability. In short, his sole aim was to determine certainty where certainty is possible and to determine the due degree of probability in the much vaster range of cases where only probability is attainable. A natural history of the growth of "knowledge" in the sense of what happens to pass for knowledge was the last of his interests. But he was completely under the domination of the ruling idea of his time; namely, that Nature is the norm of truth. Now the earliest period of human life presents the "work of nature" in its pure and unadulterated form. The normal is the original, and the original is the normative. Nature is both beneficent and truthful in its work; it retains all the properties of the Supreme Being whose vice-regent it is. To get the logical ultimates we have only, therefore, to get back to the natural primitives. Under the influence of such deistic ideas, Locke writes a mythology of the history of knowledge, starting from clear and distinct meanings, each simple, well defined, sharply and unambiguously just what it is on its face, without concealments and complications, and proceeds by "natural" compoundings up to the store of complex ideas, and to the perception of simple relations of agreement among ideas: a perception always certain if the ideas are simple, and always controllable in the case of complex ideas if we consider the simple ideas and their compoundings. Thus he established the habit of taking logical discriminations as historical or psychological primitives—as "sources" of beliefs and knowledge instead of as checks upon inference and as means of knowing.

I hope reference to Locke will not make a scapegoat. I should not have mentioned him if it were not that this way of looking at things found its way over into orthodox psychology and then back again into the foundations of logical theory. It may be said to be the stock in trade of the school of empiricist logicians, and (what is even more important) of the other schools of logic whenever they are dealing with propositions of perception and observation: vide Russell's trusting confidence in "atomic" propositions as psychological primitives. It led to the supposition that there is a kind of knowledge or simple apprehension (or sense acquaintance) implying no inference and yet basic to inference. Note, if you please, the multitude of problems generated by thinking of whatever is present in experience (as sensory qualities are present) as if it were intrinsically and apart from the use made of its subject-matter of knowledge.

a) The mind-body problem becomes an integral part of the problem of knowledge. Sense organs, neurones, and neuronic connections are certainly involved in the occurrence of a sense quality. If the occurrence of the latter is in and of itself a mode of knowledge, it becomes a matter of utmost importance to determine just how the sense organs take part in it. If one is an idealist he responds with joy to any intimation that the "process of apprehension" (that is, speaking truly, the physical conditions of the occurrence of the sensory datum) transforms the extra-organic stimulus: the alteration is testimony somehow to the constitutive nature of mind! But if he is a realist he conceives himself under obligation to show that the external stimulus is transmitted without any alteration and is apprehended just as it is; color must be shown to be simply, after all, a compacting of vibrations—or else the validity of knowledge is impugned! Recognize that knowledge is something about the color, whether about its conditions or causes or consequences or whatever, and that we don't have to identify color itself with a mode of knowing, and the situation changes. We know a color when we understand, just as we know a thunder-storm when we understand. More generally speaking, the relation of brain-change to consciousness is thought to be an essential part of the problem of knowledge. But if the brain is involved in knowing simply as part of the mechanism of acting, as the mechanism for co-ordinating partial and competing stimuli into a single scheme of response, as part of the mechanism of actual experimental inquiry, there is no miracle about the participation of the brain in knowing. One might as well make a problem of the fact that it takes a hammer to drive a nail and takes a hand to hold the hammer as to make a problem out of the fact that it also requires a physical structure to discover and to adapt the particular acts of holding and striking which are needed.

b) The propositions of physical science are not found among the data of apprehension. Mathematical propositions may be disposed of by making them purely a priori; propositions about sense objects by making them purely a posteriori.[95] But physical propositions, such as make up physics, chemistry, biology, to say nothing of propositions of history, anthropology, and society, are neither one nor the other. I cannot state the case better than Mr. Russell has stated it, although, I am bound to add, the stating did not arouse in Mr. Russell any suspicion of the premises with which he was operating. "Men of science, for the most part, are willing to condemn immediate data as 'merely subjective,' while yet maintaining the truth of the physics inferred from those data. But such an attitude, though it may be capable of justification, obviously stands in need of it; and the only justification possible must be one which exhibits matter as a logical construction from sense data.... It is therefore necessary to find some way of bridging the gulf between the world of physics and the world of sense."[96] I do not see how anyone familiar with the two-world schemes which have played such a part in the history of humanity can read this statement without depression. And if it occurred to one that the sole generating condition of these two worlds is the assumption that sense objects are modes of apprehension or knowledge (are so intrinsically and not in the use made of them), he might think it a small price to pay to inquire into the standing of this assumption. For it was precisely the fact that sense perception and physical science appeared historically (in the seventeenth century) as rival modes of knowing the same world which led to the conception of sense objects as "subjective"—since they were so different from the objects of science. Unless sense and science had both first been thought of as modes of knowing and then as modes of knowing the same things, there would not have been the slightest reason for regarding immediate data, as "merely subjective." They would have been natural phenomena, like any other. That they are phenomena which involve the interaction of an organism with other things is just an important discovery about them, as is also a discovery about starch in plants.

Physical science is the knowledge of the world by their means. It is a rival, not of them, but of the medley of prior dogmas, superstitions, and chance opinions about the world—a medley which grew up and nourished precisely because of absence of a will to explore and of a technique for detecting unambiguous data. That Mr. Russell, who is a professed realist, can do no better with the problem (once committed to the notion that sense objects are of themselves objects of knowledge) than to hold that although the world of physics is not a legitimate inference from sense data, it is a permissible logical construction from them—permissible in that it involves no logical inconsistencies—suggests that the pragmatic difference between idealist and realist—of this type—is not very great. From necessary ideal constructions to permissible logical constructions involves considerable difference in technique but no perceptible practical difference. And the point of this family likeness is that both views spring from regarding sense perception and science as ways of knowing the same objects, and hence as rivals until some scheme of conciliation has been devised.

c) It is but a variant of this problem to pass to what may be called either the ego-centric predicament or the private-public problem. Sense data differ from individual to individual. If they are recognized to be natural events, this variation is no more significant than any change depending upon variation of generating conditions. One does not expect two lumps of wax at different distances from a hot body to be affected exactly alike; the upsetting thing would be if they were. Neither does one expect cast-iron to react exactly as does steel. That organisms, because of different positions or different internal structures, should introduce differences in the phenomena which they respectively have a share in producing is a fact of the same nature. But make the sense qualities thus produced not natural events (which may then be made either objects of inquiry or means of inquiry into something else) but modes of knowing, and every such deviation marks a departure from true knowing: it constitutes an anomaly. Taken en masse the deviations are so marked as to lead to the conclusion (even on the part of a realist like Mr. Russell) that they constitute a world of private existences, which, however, may be correlated without logical inconsistency with other such worlds. Not all realists are Leibnizian monadists as is Mr. Russell; I do not wish to leave the impression that all come to just this solution. But all who regard sense data as apprehensions have on their hands in some form the problem of the seemingly distorting action exercised by the individual knower upon a public or common thing known or believed in.

I am not trying to discuss or solve these problems. On the contrary, I am trying to show that these problems exist only because of the identification of a datum determined with reference to control of inference with a self-sufficient knowledge-object. As against this assumption I point to the following facts. What is actually given as matter of empirical fact may be indefinitely complicated and diffused. As empirically existent, perceived objects never constitute the whole scope of the given; they have a context of indefinite extent in which they are set. To control inference it is necessary to analyze this complex situation—to determine what is data for inference and what is irrelevant. This analysis involves discriminative resolution into more ultimate simples. The resources of experimentation, all sorts of microscopic, telescopic, and registering apparatus, are called in to perform that analysis. As a result we differentiate not merely visual data from auditory—a discrimination effected by experiments within the reach of everybody—but a vast multitude of visual and auditory data. Physics and physiology and anatomy all play a part in the analysis. We even carry the analysis to the point of regarding, say, a color as a self-included object unreferred to any other object. We may avoid a false inference by conceiving it, not as a quality of any object, but as merely a product of a nervous stimulation and reaction. Instead of referring it to a ribbon or piece of paper we may refer it to the organism. But this is only as a part of the technique of suspended inference. We avoid some habitual inference in order to make a more careful inference.

Thus we escape, by a straightening out of our logic (by avoiding erecting a system of logical distinctions and checks into a mythological natural history), the epistemological problems. We also avoid the contradiction which haunts every epistemological scheme so far propounded. As matter of fact every proposition regarding what is "given" to sensation or perception is dependent upon the assumption of a vast amount of scientific knowledge which is the result of a multitude of prior analyses, verifications, and inferences. What a combination of Tantalus and Sisyphus we get when we fancy that we have cleared the slate of all these material implications, fancy that we have really started with simple and independent givens, and then try to show how from these original givens we can arrive at the very knowledge which we have all the time employed in the discovery and fixation of the simple sense data![97]