Herbert Feigl, The "Mental" and the "Physical": The Essay and a Postscript (1967).

III. Requirements and Desiderata for an Adequate Solution of the Mind-Body Problem. A Concise Statement of the Major Issues

If the title of this section were not already a bit too long, I should have added, "as 1 view these requirements and desiderata, and as I conceive the adequacy of a solution." All I can say by way of extenuation of my personal biases in this matter is that I have concerned myself seriously and repeatedly with the problem for about thirty-six years; that I have studied most of the contributions from thinkers of many lands in modern and recent philosophy and science; and that this is my fourth published attempt to arrive at an all around satisfactory clarification. There have often been moments of despair when I tried ineffectively to do justice to the many (apparently) conflicting but impressive claims coming from ever so many quarters. It is, then, with a heavy sense of intellectual responsibility and not without some misgivings that I proceed to enumerate the following requirements, desiderata, and considerations which seem to me the conditions (or at least some of the conditions) that may serve .as criteria of adequacy for a solution of the problem; a solution that is to be satisfactory from the point of view of contemporary science as well as in the light of modern philosophical analysis. I concede unblushingly that in some respects I share here the attitude of some of the (shall I say, epistemologically not too naive) metaphysicians who have wrestled with the problem and have tried to provide a solution that is synoptic in that it would render a just, consistent, and coherent account of all relevant aspects and facets of the issue.

Here, then, is my list of requirements and desiderata (or "conservanda" and "explicanda"):

1. The terms "mental" and "physical" are precariously ambiguous and vague. Hence a first prerequisite for the clarification and the adequate settlement of the main issues is an analytical study of the meanings of each of these two key terms, and a comparative critical appraisal of the merits and demerits of their various definitions and connotations. Due attention will also be given to the (partly) terminological question as to whether to include under "mental" beside the directly experienced and introspectible also the unconscious states and processes of depth-psychological theories (Freudian or Neo-Freudian). All this will be undertaken in the next section of this essay.

2. In the light of what was said in the preceding section about the scientific (empirical) components of the mind-body problem, an analysis of the mind-body relation is to be sought which does justice to the arguments for the sort of mind-body unity which impresses itself increasingly upon the majority of psychologists, psychophysiologists, and psychiatrists of our time. Although the question of evolutionary as well as of logical "emergence" cannot be decided by a priori philosophical considerations, vitalistic and interactionist doctrines appear on empirical and methodological grounds as suspect and undesirable. Just what the alleged facts of parapsychology (telepathy, clairvoyance, precognition, psychokinesis, etc.) may imply for the mind-body problem is still quite unclear. Here too, it seems to me, any speculations along the lines of interactionism are -- to put it mildly -- premature, and any theological interpretations amount to jumping to completely unwarranted conclusions. My own attitude in regard to the experiments (statistical designs) on extrasensory perception, etc. is that of the "open mind." The book by Soal and Bateman (325) and its discussion by M. Scriven (305) present evidence and arguments which can not lazily or cavalierly be shrugged off. The chances of explaining the "facts" away as due to experimental or statistical error, let alone as outright hoax or fraud, seem now rather remote. But even granting these facts, I think that efforts should be made to explain them first by revisions and emendations in the physical theory of behavior before we indulge in speculations about immaterial souls or selves. These remarks clearly reveal my bias in favor of a naturalistic, if not monistic, position. That and how this position differs from "crass materialism," the bugbear of idealistic and spiritualistic metaphysicians, will be explained later on.

3. Any solution of the mind-body problem worth consideration should render an adequate account of the efficacy of mental states, events, and processes in the behavior of human (and also some subhuman) organisms. It is not tendermindedness or metaphysical confusions, I trust, which impel this repudiation of a materialistically oriented epiphenomenalism. Admittedly, the testimony of direct experience and of introspection is fallible. But to maintain that planning, deliberation, preference, choice, volition, pleasure, pain, displeasure, love, hatred, attention, vigilance, enthusiasm, grief, indignation, expectations, remembrances, hopes, wishes, etc. are not among the causal factors which determine human behiavior, is to fly in the face of the commonest of evidence, or else to deviate in a strange and unjustifiable way from the ordinary use of language. The task is neither to repudiate these obvious facts, nor to rule out this manner of describing them. The task is rather to analyze the logical status of this sort of description in its relation to behavioral and/or neurophysiological descriptions. In the pursuit of this objective it will of course be necessary to avoid both interactionism and epiphenomenalism; and it will moreover be desirable to formulate the solution in such away that it does not presuppose emergentism (in the sense of physical2 indeterminism), although the door to a scientifically formulated emergentism need not be closed.

In this same connection justice should be rendered to what is meaningful and scientifically defensible in the notion of free will or free choice. If our personality-as-it-is at the moment of choice expresses itself in the choice made; if our choices accord with our most deeply felt desires, i.e., if they are not imposed upon us by some sort of compulsion, coercion or constraints such as by brute physical force, by other persons (or even only by components of our personality which we do not acknowledge as the "core" deemed centrally our "self"), then we are "free" in the sense that we are the doers of our deeds, the choosers of our choices, the makers of our decisions. In other words, it is in this case that our central personality structure is a link in the causal chain of our behavior, predominantly, even if not exclusively, effective in the determination of our conduct. This sort of freedom (in the superb formulation of R. E. Hobart-Dickinson Miller) "involves determinism and [is] inconceivable without it."1

4. A most important logical requirement for the analysis of the mind-body problem is the recognition of the synthetic or empirical character of the statements regarding the correlation of psychological to neurophysiological states. It has been pointed out time and again2 that the early reductionistie logical behaviorism failed to produce an adequate and plausible construal of mentalistic concepts by explicit definition on the basis of purely behavioral concepts. (In the less adequate material mode this might be put by bluntly saying that mind is not identifiable with behavior.) For a long time, however, I was tempted to identify, in the sense of logical identity, the mental with the neurophysiological, or rather with certain configurational aspects of the neural processes. It was in this sense that I (103) suggested a double-language theory of the mental and the physical. But if this theory is understood as holding a logical translatability (analytic transformability) of statements in the one language into statements in the other, this will certainly not do. Interlinguistic translations like "Il ne fait pas beau temps" into "The weather is not fine" are analytic if the respective meanings are fixed with the help of syntactical and semantical metalanguages common to both French and English. Similarly the geocentric description of the pure kinematics of the planetary system is analytically translatable into the corresponding heliocentric description, precisely because we avail ourselves here of transformation rules in a four-dimensional geometry (i.e., kinematics).

But the question which mental states correspond to which cerebral states is in some sense (to be analyzed epistemologically later on) an empirical question. If this were not so, the intriguing and very unfinished science of psychophysiology could be pursued and completed by purely a priori reasoning. Ancient and primitive people had a fair amount of informal and practical psychological knowledge, but the fact that mental , states are closely associated with cerebral states was unknown to them. Aristotle held that the seat of our feelings and emotions is the heart (and this has survived in the traditions of poetic discourse). But to say that Aristotle was wrong means that we have now empirical evidence which proves that the emotions are linked to brain processes. It is therefore imperative to preserve the synthetic character of the assertion of this knowledge claim, whatever specifically may prove to be its most clarifying formulation.

If any of my readers should be hard-boiled behaviorists or "crass" materialists, it will be difficult to convince them that there is a problem at all. I can do no more than to ask them such persuasive or ad hominem questions as, Don't you want anesthesia if the surgeon is to operate on you? And if so, what you want prevented is the occurrence of the (very!) raw feels of pain, is it not? If you have genuine concern and compassion for your fellow human beings (as well as perhaps for your dogs, horses, etc.), what is it that you object to among the consequences of cruel treatments? Is it not the pains experienced by these "others"? It could not be merely their physical mutilation and consequent malfunctioning. Moral condemnation of wanton cruelty presupposes the meaningfulness of the ascription of direct experience to others. Subjective experience in this sense cannot be logically identical with states of the organism; i.e., phenomenal terms could not explicitly be defined on the basis of physical1 or physical2 terms.

It should be noted that we repudiate the logical translatability thesis not because of the possibility, definitely contemplated, of a one-many-ψ-φ correspondence. One could always formulate such a correspondence with the help of a general equivalence between statements containing single ψ-predicates on the one side and disjunctions of statements containing several and various φ-predicates on the other. It is rather the logical necessity of the equivalence which is here rejected. The equivalence must be construed as logically contingent.

5. Consonant with the spirit of the preceding discussions, but now to be stated explicitly, are three very closely related epistemological requirements. To list them first very briefly, they are:

  1. the need for a criterion of scientific meaningfulness based on inter-subjective confirmability;
  2. the recognition that epistemology, in order to provide an adequate reconstruction of the confirmation of knowledge claims must employ the notion of immediate experience as a confirmation basis (the "given" cannot "be entirely a myth!); "Acquaintance" and "Knowledge by Acquaintance," however, require careful scrutiny;
  3. the indispensability of a realistic, as contrasted with operationalistic or phenomenalistic, interpretation of empirical knowledge in general, and of scientific theories in particular.

(ad a) It is generally agreed that scientific knowledge claims must not only be intersubjectively communicable (intelligible), but also intersubjectively testable. The following considerations will illustrate the point. If the stream of my conscious experience continued beyond the death and decay of my body, then this may be verifiable by me (in some, none too clear, sense of "me"; but I shall let this pass for the moment). If such survival were, however, not even extremely indirectly or incompletely confirmable by others; if it were in no way lawfully connected with, and thus not inferable from, any feature of life (mine or that of others) before death, then, while the statement in question may be said to have subjective meaning, it could not become part of science in the sense in which "science" is commonly understood.3

(ad b) Recent behavioristic and physicalistic arguments to the contrary notwithstanding, I am still convinced that purely phenomenal statements make sense and are the ultimate epistemic basis of the confirmation (or disconfirmation) of knowledge claims. By this I do not at all wish to suggest that phenomenal statements are infallible ("incorrigible"), nor that they necessarily have a higher degree of certainty than intersubjectively confirmable statements about the ordinary objects of our common life environment. I grant that, especially for the purposes of the philosophy of science, it is more useful to choose the physicalistic thing language for the confirmation basis of knowledge claims. But when I judge, e.g., that a certain pain is increasing, or that I hear a certain ringing sound (no matter whether this sound-as-experienced is causally due to a doorbell, a police car siren, to "buzzing in my ear," or to a hallucination), then that certain it which may later find its place in the causal structure of the world is first of all, and taken by itself, a datum of direct experience. Whether I get to it "post-analytically," or whether I simply have it, pre-analytically; that is to say, whether I arrive at it by a kind of analysis starting from "seeming," "appearing," "looks like" ("sounds like," etc.) sentences; or whether I can by simultaneous introspection (self-observation) or immediate retrospection, ascertain the occurrence of a certain datum,) I have no doubt that talk about phenomenal data and phenomenal fields makes sense; and that in a rational reconstruction of the confirmation of ordinary observation statements, we can (if we wish) penetrate to this deepest level of evidence.4

I have not been convinced by the arguments of Popper (258) that the search for "hard data" is doomed to failure, that the "given" is like a bottomless swamp. Nor am I convinced that a purely private language5 is inconceivable. Of course, if by "language" one means an instrument of interpersonal communication, then the idea of an absolutely private language is self-contradictory. But, granting that in the normal case the capacity for using a language is acquired by education, it is not logically inconceivable that a child growing up in complete solitude might devise his own symbolism not only for the objects and events in its environment but also for the raw feds of its direct experience. Such a child might well come to use terms for various aches, pains, itches, tickles, moods, emotions, etc. I do not for a moment deny that the use of such subjective terms, in the usual and normal case, is acquired through trial and error learning, and in this process largely inculcated in the child by other persons who tell him, e.g., "now you are tired," "now you are glad," "you must have an awful pain." Such tellings by others are guided by the facial expressions, vocal emissions, posture, etc., i.e., generally by the observable behavior of the child (and by test condition → test result sequences in its behavior, involving both environmental stimulus situations and a variety of responses).6

In sum, I believe that there is an indispensable place for "acquaintance" and "knowledge by acquaintance" in a complete and adequate epistemology. A more detailed account and analysis of the meanings of these terms will be given in the two subsequent sections of the resent essay.

(ad c) The last epistemological requirement, to be briefly discussed here, is that of a realistic, rather than phenomenalistic or operationalistic, reconstruction of knowledge. With the current liberalization of the criterion of empirical meaningfulness7 the narrower positivism of the Vienna Circle has been definitively repudiated, and is being replaced by a ("hypercritical") realism. No longer do we identify the meaning of a statement with its method of verification. Nor do we consider the meaning of a concept as equivalent with the set of operations which in test situations enable us to determine its (more or less likely) applicability. Instead we distinguish the evidential (or confirmatory) basis from the factual content or reference of a knowledge claim. Early and crude forms of behaviorism identified mental states with their (sic!) observable symptoms. Embarrassment might then mean nothing but blushing. But refinements and corrections were introduced in due course. Mental states were considered "logical" constructions based on observable behavior; and statements about mental states were considered logically translatable into statements about actual or possible behavior, or into statements (or sets of statements) about test conditions and ensuing test results concerning behavior. Mental traits were considered as correlation clusters of their (sic!) symptoms and manifestations, and so forth.

But even such a refined or "logical" behaviorism is now rejected as an inadequate reconstruction. It was realized that those behavioral test condition -- → test result conditionals are to be derived from the laws and postulates regarding central states. Such derivations or explanations have been eminently successful in the physical and in some of the biological sciences. In the atomic theory, or in the theory of genes, for example, it is becoming increasingly possible to derive the macro-regularities, regarding, e.g., chemical compounding, or Mendelian heredity from lawlike postulates and existential hypotheses. The central states of molar behavior theory (or the "factors" in the factor analysis of personality traits) are, however, unspecified as regards their neurophysiological basis. This is comparable to the early stages of the atomic theory when nothing was known about the mass and the structure of individual atoms, or to the early stages of the theory of heredity when Mendel's "units" were not as yet identified with the genes, located and spatially ordered in specific ways, within the chromosomes of the germ cells.

There is little doubt in my mind that psychoanalytic theory (or at east some of its components) has genuine explanatory power, even if any precise identification of repression, ego, superego, ego, id, etc. with neural processes and structures is still a very long way off. I am not in the least disputing the value of theories whose basic concepts are not in any way micro-specified. What I am arguing is that even before such specifications become possible, the meaning of scientific terms can be explicated by postulates and correspondence rules (cf. Carnap, 73), and that this meaning may later be greatly enriched, i.e. much more fully specified, by the addition of further postulates and correspondence rules.8

After the recovery from radical behaviorism and operationism, we need no longer hesitate to distinguish between evidence and reference, i.e., between manifestations or symptoms on the one hand, and central states on the other; no matter whether or not central states are micro-specified (neurophysiologically identified).

The meaning of scientific statements consists indeed in their truth conditions. But "truth conditions" does not mean the same as "confirming evidence". (The only possible exceptions to this are the directly and completely confirmable singular statements regarding immediately observable situations.) A theory is required to tell us which observations form confirming evidence for scientific statements about matters inaccessible to direct observation. It is in the light of such theories that we can then specify how much support a given bit of evidence lends to a specified hypothesis.

In section V, I shall return to the crucial questions of reduction and identification. There I shall discuss the logical nature of the relation between mentalistically, behaviorally, and neurophysiologically characterized central states.

No elaborate arguments should here be required for a realistic interpretation of the statements about the "physical" objects of everyday life or of theoretical physics.9 In the explanatory context (or the "nomological net") concepts pertaining to the unobservables are related to, but not identifiable with, the observables which constitute the evidential data for the confirmation of statements about the unobservables. For example, spectral lines, cloud chamber tracks, scintillations on screens, Geiger counter indications, etc. are the evidential data which, in a complete logical reconstruction, must be conceived as nomologically connected with the aspects of atomic and subatomic particles which they confirm. Less exciting, but logically analogous, is the analysis of statements of common life about ordinary (partly or wholly observable) objects. Here the perceived perspectives of mountains, trees, clouds, etc., or the instrument indications of air pressure, wind currents, air moisture, etc., are to be interpreted as evidence related to what is evidenced, by the geometrical-optical laws underlying the projections in visual perception, or the physical laws which explain the operation of barometers, anemometers, hygrometers, etc.

6. The "meat" of an adequate solution of the mind-body problem will consist in a specific analysis of the characteristics and the relations between the attributes of the mental (especially the phenomenal) and the physical (specifically the neurophysiological). It should be clear from the outset that, if a complete solution of these problems is ever going to be achieved, it will arise out of a combination of the results both of scientific research and of philosophical analysis. In all these questions the two components are so intimately bound up with one another, that neglecting either of them seriously jeopardizes the whole endeavor. The philosophical aspects will be given a further analysis in the next section where I shall try to sort out the various meanings and the attached connotations of the terms "mental" and "physical". The most controversial, tangled and perplexing questions concern, of course, the distinctions made rightly or wrongly in the Cartesian and in the subsequent dualistic tradition between the mental and the physical in terms of the various alleged criteria listed in the accompanying table.

Mental

subjective (private)
nonspatial
qualitative
purposive
mnemic
holistic
emergent
intentional

Physical

objective (public)
spatial
quantitative
mechanical
non-mnemic
atomistic
compositional
"blind"; nonintentional

Practically all the perennial perplexities of the mind-body problem center around the listed contrasts. The dualists make prima facie an excellent showing. The more enlightened monists have always realized that any argument in favor of an identification (in some sense!) of the mental and the physical is faced with serious difficulties. Small wonder then that many of the more sophisticated analytic philosophers of the present age either embrace some form of dualism (usually parallelism), or else declare the issue between monism and dualism a pseudoproblem engendered by logical or terminological confusions. I do not share this outlook. In the following section I shall prepare the ground for an "identity" theory, and I shall present my formulation as well as my arguments in section V.


Notes

1. Cf. R. E. Hobart (157).

2. Cf. F. Kaufmann (175), N. Jacobs (163), C. I. Lewis (196), E. Nagel (230), A. Pap (243), et al.

3. For a fuller discussion of the scientific meaning criterion cf. my articles (103, 105, 109, 110, 114, 116) and Carnap (64, 67, 73). For stimulating discussions of the meaning of "disembodied minds" see Aldrich (6) and Lewy (199).

4. For persuasive arguments along these lines, cf. B. Russell (284, 287); H. H. Price (264); C. I. Lewis (195, 197, 198); Ayer (12, 13, 18); N. Goodman (135, 136, 137). For an incisive critique of the "incorrigibility" arguments, cf. K. R. Popper (258); R. Camap (62, 64); H. Reichenbach (273, 276); M. Black (38); J. Epstein (98).

5. Cf. the symposium by Ayer and Rhees (16, 278).

6. Cf. Camap (62, 63); Skinner (320, 321); Wittgenstein (357).

7. Cf. Carnap (64, 73); Hempel (149, 151); Feigl (105, 106, 109, 110, 112, 114, 116); Ayer (18); A. Pap (243, 246, 248). Also Grünbaum (139); Feyerabend (119).

8. For a defense of psychological theory without explicit reference to micro-levels, cf. Lindzey (200). The logic of theoretical concepts in psychology has been discussed in some detail by McCorquodale and Meehl, (213); Feigl (113); Cronbach and Meehl (79); Ginsberg (133, 134); Maze (212); Seward (317); Rozeboom (283); Scriven (306).

9. Cf. B. Russell (288); R. B. Braithwaite (48); Kneale (179); L. W. Beck (24); Feigl (110, 111,114).