Herbert Feigl, The "Mental" and the "Physical": The Essay and a Postscript (1967).
C. The Cognitive Roles of Acquaintance.
Various meanings of "acquaintance" and of "knowledge by acquaintance" were sorted out in section IV A. Our present concern is with the roles of acquaintance and of knowledge by acquaintance in the enterprise of science, especially in psychology. The first question I wish to discuss concerns the cognitive "plus," i.e., the alleged advantages of knowledge by acquaintance over knowledge by description. We may ask, for example, what does the seeing man know that the congenitally blind man could not know. Or, to take two examples from Eddington (93, 94), What could a man know about the effects of jokes if he had no sense of humor? Could a Martian, entirely without sentiments of compassion or piety, know about what is going on during a commemoration of the armistice? For the sake of the argument, we assume complete physical (1 or 2) predictability and explainability of the behavior of humans equipped with vision, a sense of humor, and sentiments of piety. The Martian could then predict all responses, including the linguistic utterances of the earthlings in the situations which involve their visual perceptions, their laughter about jokes, or their (solemn) behavior at the commemoration. But ex hypothesi, the Martian would be lacking completely in the sort of imagery and empathy which depends on familiarity (direct acquaintance) with the kinds of qualia to be imaged or empathized.
As we have pointed out before, "knowledge of," i.e., "acquaintance with," qualia is not a necessary condition for "knowledge about" (or knowledge by inference of) those qualia. A psychiatrist may know a great deal about extreme states of manic euphoria or of abject melancholic depression, without ever having experienced anything anywhere near them himself. In this case, of course, it must be admitted that the psychiatrist can get an "idea" of these extreme conditions by imaginative extrapolation from the milder spells of elation or depression which he, along with all human beings, does know by acquaintance. But the case is different for observers who are congenitally deprived of acquaintance with an entire modality of direct experience. This is the case of the congenitally blind or deaf, or that of our fancied Martian who has no emotions or sentiments of any kind. But I think it is also the case of human beings endowed with the entire repertory of normal sensory and emotional experience, when they introduce theoretical concepts in their science, such as the electromagnetic or gravitational fields, electric currents, and nuclear forces. We are "acquainted" with the perceptible things, properties, and relations on the relevant evidential bases which suggest the introduction of these concepts into the system of science generally, or which justify their special application in particular instances of observation.
In the context of the present discussion it does not matter very much whether we use the narrower, philosophical notion of direct acquaintance (restricted to the qualia of raw feels) or the wider commonsense or physicalistic notion of acquaintance (which includes the directly observable properties and relations of the objects in our everyday life environment). I think it does make sense to say that we do not know by acquaintance the "nature" of electric currents or of the forces within the nuclei of atoms. And although the congenitally blind have no acquaintance with color qualities or visual shapes, they may nevertheless come to have knowledge by inference at least of the neural correlates among the processes in the occipital lobes of the brains of persons with eyesight. The "intrinsic nature" of those neural processes remains unknown by acquaintance to the blind scientist, just as the "intrinsic nature" of electric currents remains unknown to scientists who have eyesight, and who have seen electrical machines and wires, hare been tickled or shocked by electric currents, have seen electric sparks, have felt the heat produced by electric currents, have read voltmeters and ammeters, have observed the chemical and magnetic effects of electric currents, etc.
I trust my readers will not charge me with obscurantist tendencies. I do not at all share the view (e.g., Bergson's) that genuine knowledge is to be found only in direct acquaintance or intuition. Bergson, in his Introduction to Metaphysics, paradoxically claimed that metaphysics -- the intuitive knowledge of intrinsic reality -- is "the science that dispenses with symbols altogether." I wish to assert, quite to the contrary, that genuine knowledge is always symbolic, be it knowledge by acquaintance as formulated in direct introspective report sentences, or be it knowledge by description as, e.g., in the hypotheses of modern nuclear physics. If we knew all about electricity, magnetism, nuclear forces, etc., i.e., if we had a complete set of laws concerning those matters -- this would be all we could possibly wish to know within the scientific enterprise. Anything added to this by way of "acquaintance" would be cognitively irrelevant imagery. Such imagery might be welcome from a poetic or artistic point of view. It might occasionally be helpful heuristically or didactically, but even in this regard it amounts only to pictorial bywork, and is often dangerously misleading. "Thou shalt not make unto thee any graven image . . ."
Our world, being what it is, can of course be known by description, in any of its parts or aspects, only on the basis of a foothold somewhere in direct acquaintance. This, it seems to me, is one of the cornerstones of any empiricist epistemology, old or new. But the new empiricism of recent times has come to recognize that it matters little just which areas of acquaintance are available or actually utilized for the "triangulation" of facts or entities outside the scope of direct acquaintance. The congenitally blind-deaf person, I stress again, could in principle construct and confirm a complete system of the natural sciences (including astronomy!) and the social sciences (including the psychology of vision and hearing, as well as the psychology of art and music appreciation!). It should go without saying that such a person, like Helen Keller, would normally depend upon information received from persons endowed with visual and auditory perception.
But, supposing such a human being could survive a long time as a solitary observer and was equipped with supreme intelligence and ingenuity, then one can well conceive of various modern instruments and devices (involving photoelectric cells, amplifiers, electromagnetic indicators, etc.) he could invent which would serve him in the detection of the stars, the chemical constitution of various substances, the behavior of animals, and so on -- all accessible to him ultimately through, e.g., tactual pointer readings of one sort or another. All this is merely a picturesque way of saying that the "nomological net," i.e., the system of scientific concepts and laws, may be "tacked down" in a variety of alternative ways, either in several sense modalities (as in the normal case), or even in only one of them. To be sure, "triangulation of entities in logical space" is much easier and much more secure in the normal case. But, as we have pointed out, normal, unaided perception by itself is also quite insufficient for the confirmation of our knowledge regarding radio waves, infrared, ultraviolet, gamma radiations, cosmic rays, the molecular and atomic structure of matter, the motions and other physical and chemical characteristics of stars and galaxies, etc. Intricate instruments and ingenious theoretical constructions are indispensable in the case of normal (multimodal) perception as well. The difference between persons equipped with all normal sense organs and the deaf-blind is only one of degree, or of the speed with which they would, respectively, attain knowledge about the world in which they are embedded and of which they are parts.
Similar considerations apply to the advantages held by fully equipped persons in regard to psychological and linguistic or descriptive-semantical knowledge. If I have been trained by normal education to apply phenomenal terms (like "red", "green", "lilac fragrance", "rose fragrance", "sweet", "sour", etc.) to qualia of my own direct experience, then I can predict much more readily the application of these terms by other persons in the presence of certain specifiable visual, olfactory or gustatory stimuli. But predictions of this sort are based upon analogical inference; and they are in principle dispensable, because the discriminatory and verbal behavior of other persons is open to intersubjective test. Moreover, if we had a complete neurophysiological explanation of discriminatory and verbal responses we could derive these responses from the cerebral states which initiate them, and which, in turn were engendered by sensory stimulation. Analogously, whatever reliability empathetic understanding in common life, or "clinical intuition" in the psychologist's practice, may have is ultimately to be appraised by intersubjective tests. But the speed with which empathy or intuition do their work depends upon the breadth and the richness of the "experience" of the judge. It also depends upon his use of critical controls.
If the psychologist's personality type is radically different from that of his subject, he will have to correct (often to the point of complete reversal) his first intuitions. For example, an extremely extrovert person will find it difficult to "understand" an extreme introvert, and vice versa. If, however, the personalities are very similar, intuition may "click" readily, and it may even be frequently quite correct. The role of direct acquaintance in all these cases simply amounts to having in one's own experience features and regularities with which one is quite familiar, and which are hence speedily projected and utilized in the interpretation of the behavior of other persons. I conclude that the advantages of direct acquaintance pertain to the context of discovery (cf. Reichenbach, 273) and not to the context of justification. All the examples discussed do not differ in principle from the obvious examples of persons with "wide experience" as contrasted with persons with "narrow experience," in the most ordinary meaning of these terms. Someone thoroughly familiar with the weather patterns of Minnesota, or with the conduct of business in the Congress of the United States (to take two very different illustrations of the same point) will have the advantages of much speedier inferences and (usually) more reliable predictions than someone who has had no opportunity of long range observations in either case.
The philosophically intriguing questions regarding acquaintance are, I think, of a different sort. They are best expressed by asking, e.g., What is it that the blind man cannot know concerning color qualities? What is it that the (emotionless) Martian could not know about human feelings and sentiments? If we assume complete physical (i.e., at least physical) predictability of human behavior, i.e., as much predictability as the best developed physical science of the future could conceivably provide, then it is clear that the blind man or the Martian would lack only acquaintance and knowledge by acquaintance in certain areas of the realm of qualia. Lacking acquaintance means not having those experiential qualia; and the consequent lack of knowledge by acquaintance simply amounts to being unable to label the qualia with terms used previously by the subject (or, by some other subject) when confronted with their occurrence in direct experience. Now, mere having or living through ("erleben") is not knowledge in any sense. "Knowledge by acquaintance," however, as we understand it here, is propositional, it does make truth claims; and although it is not infallible, it is under favorable circumstances so reliable that we rarely hesitate to call it "certain." It remains in any case the ultimate confirmation basis of all knowledge claims.
In many of the foregoing discussions we have suggested that what one person has and knows by acquaintance may be identical with what someone else knows by description. The color experiences of the man who can see are known to him by acquaintance, but the blind man can have inferential knowledge, or knowledge by description about those same experiences. After all, this is true as regards an individual color experience even if the other person is endowed with eyesight. The other person does not and could not conceivably have the numerically identical experience (see p. 30f above). Why should we then not conclude that the behavioristic psychologist can "triangulate" the direct experiences of others? I think that indeed he does just that if he relinquishes the narrow peripheralist position, i.e., if he allows himself the introduction of theoretical concepts which are only logically connected with, but never explicitly definable in terms of, concepts pertaining to overt molar behavior. These acquaintancewise possibly unknown states which the behaviorist must introduce for the sake of a theoretical explanation of overt behavior, and to which he (no longer a "radical" behaviorist) refers as the central causes of the peripheral behavior symptoms and manifestations, may well be identical with the referents of the phenomenal (acquaintance) terms used by his subject in introspective descriptions of his (the subject's) direct experience. As remarked before, in ordinary communication about our respective mental states, we make this assumption of identity quite unquestioningly. It took a great deal of training in philosophical doubt for learned men to call this assumption into question.
But philosophical doubt, here as elsewhere,1 while stimulating in the search for clarity, is ultimately due to conceptual confusions. We have learned how to avoid these confusions, and thus to return with a good philosophical conscience to (at least some of) the convictions of commonsense. We have learned that philosophical doubts, unlike ordinary empirical doubts, cannot be removed by logical or experimental demonstration. What can be demonstrated logically is only the exploitation of certain misleading extensions of, or deviations from, the sensible and fruitful use of terms in ordinary or scientific language. Thus to doubt whether we can at all have knowledge about the "private" experience of other persons is merely the philosophical extension of the ordinary and quite legitimate doubts that we may have in specific instances, for example, when we ask "Is he really as disappointed as his behavior would seem to indicate?" This is to confuse practical difficulties of knowing with (allegedly) basic impossibilities. Once one becomes fully aware of the disease of philosophical skepticism, it becomes possible to cure oneself of it by a sort of self-analysis (logical analysis is what I have in mind here; but in certain cases psychoanalysis may help too, or may even be indispensable).
Granting then that the referents of acquaintance terms and physical1 theoretical terms may in some cases be identical, this does not by itself decide the issue between monism and dualism. As we have seen in the previous subsection, the inference to other persons' raw feels can be logically differentiated from the inference to their central nervous processes. Dualistic parallelism or epiphenomenalism is entirely compatible with the assertion of the identity of the subjectively labeled mental state with the intersubjectively inferred state which is needed for the explanation of molar behavior. The mental state is logically distinguishable from the "correlated" neurophysiological state. Indeed (as pointed out in section III 4), it makes no sense to talk of correlation, or in any case not the usual sense, if the relation of "correlation" were that of identity. We shall tackle this crucial point in the next two subsections.
Before we proceed to the discussion of identity and identification, let us however summarize some important conclusions from our discussion of acquaintance. The data of direct experience function in three roles: First, in the use of typical patterns and regularities of one person's data for the intuitive or empathetic ascription of similar patterns and regularities of direct experience (or even of unconscious processes) to other persons, these data suggest, but by themselves are never a sufficiently strong basis of validation for knowledge claims about the mental life of other persons. Further clinical, experimental, or statistical studies of the behavior of those persons are needed in order to obtain a scientifically respectable degree of confirmation for such inferences. Second, nevertheless, and this is philosophically even more important, the first-person data of direct experience are, in the ultimate epistemological analysis, the confirmation basis of all types of factual knowledge claims. This is simply the core of the empiricist thesis over again. But third, the data are also objects (targets, referents) of some knowledge claims, viz. of those statements which concern nothing but the occurrence of raw feels or whatever regularities (if any!) can be formulated about raw feels in purely phenomenal terms. For examples of the latter, I mention the three-dimensional ordering of color qualia according to hue, brightness, and saturation; the regularities regarding the gradual (temporal) fading of intense emotions like joy, rage, exultation, embarrassment, regret, grief, etc.; the lawful correlations between, e.g., the experienced contents of daydreams and the attendant emotions of hope or fear. In all these cases, no matter whether the raw feels are our own or someone else's, they are the objects of our knowledge claims or the referents of certain terms in the sentences which describe them. I emphasize this point because recent empiricist epistemologies in their concern with the confirmation bases of our knowledge claims, and with observation statements which formulate the confirming (or disconfirming) evidence, have tended to neglect consideration of those cases in which the target of the knowledge claim is a state or a regularity of direct experience. Evidence and reference coincide only in the case of statements about the immediate data of first-person experience. But they are clearly distinct in all other cases, such as those in which the object of reference is a state of affairs in the world outside the observer (or else anatomically physiologically inside his own skin), no matter whether it be the state of inorganic things, or processes in organisms. Even the direct experience of oneself at a time distinct from the present moment, and of course the direct experience of other organisms or persons are numerically distinct from the data of the confirming evidence. In short, the data of immediate experience function either as verifiers or as referents of knowledge claims.
Notes 1. As, e.g., in the problems of induction, the trustworthiness of memory, the veridicality of perception, etc.