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

II. The Scientific and the Philosophical Strands in the Mind-Body Tangle

A first indispensable step toward a clarification of the issues is to separate the scientific from the epistemological questions pertaining to the relations of the mental to the physical. Epistemology is here understood in the modern sense of a logical analysis of concepts and statements and of the closely related logical reconstruction of the validation of knowledge claims. Some of the pertinent statements themselves are, however, essentially of a scientific nature in that they fall under the jurisdiction of empirical evidence. It is right here where we find a fundamental parting of the ways. Biologists, psychologists (and with them, many philosophers) hold deep convictions, one way or another, on the autonomy or non-autonomy of the mental. The strongest contrast is to be found between those who hold interactionistic views regarding the mental and the physical, and those who reject interactionism and hence espouse either parallelism (e.g., in its currently favored form, isomorphism) or some emphatically monistic view. Interactionism as well as parallelism is of course a form of dualism. The main difference , and dispute between these two points of view is at present not fully decided by the evidence. But I think this is an issue to which empirical evidence is ultimately and in principle relevant.

Vitalists or interactionists like Driesch, McDougall, J. B. Pratt, Ducasse, Kapp, et al. hold that biological concepts and laws are not reducible to the laws of physics, and hence -- a fortiori -- that psychological concepts and laws are likewise irreducible. Usually this doctrine is combined with a theory of the emergent novelty of life and mind. But there are others who restrict emergence to the mental, i.e. they hold a reducibility view in regard to the biological facts. "Reducibility" is here understood to mean the same as "explainability''; and has no necessary connection with the introducibility (empirical anchorage) of biological or psychological concepts on the basis of physicalistic observation terms. As Carnap (67) has pointed out clearly, the thesis (his thesis) of the unity of the language of science does not in any way prejudge the issue) of the unitary explainability of biological and psychological facts (or laws) on the basis of physical theory. Philosophers should certainly not assume that such a basic scientific issue can be settled merely by logical analysis. It is logically conceivable that biological, psychological, and social phenomena (as well as their regularities) may not be explainable in terms of those physical or physicochemical laws (and theoretical assumptions) which are sufficient for the explanation and prediction of inorganic phenomena (and their regularities).

Logical parallels to such irreducibilities are clearly evident even within physics. The "mechanistic" (Newtonian) premises of explanation are now viewed as entirely insufficient for the explanation of electromagnetic radiation, of the dynamics of intramolecular and intra-atomic processes, and of the interaction of electromagnetic radiation and the particles of matter. Nineteenth century physics added the fundamentally new concepts and laws of electromagnetics; and these in turn were drastically modified and supplemented by the relativity and quantum theories of our century. It is conceivable that homologous emendations may be required for the explanation of the phenomena of life and mind. Contemporary dualists, be they vitalists, emergentists, interac-tionists, or parallelists, maintain that such an enrichment of the conceptual system of science will be indispensable. Their arguments are based primarily on the traditionally captivating evidence of teleological processes, purposive behavior, psychosomatics, and the mnemonic and intentional features of perception, cognition, thought, desire, and volition. And some apparently very persuasive arguments point simply to the existence (occurrence) of immediate experience, i.e., the raw-feels or hard data of the directly given. They maintain that these data, though related to behavior and neurophysiological processes, are not reducible to, or definable in terms of, purely physical concepts; and that their occurrence is not predictable or explainable on the basis of physical laws and physical descriptions only.

At this point the distinction between the scientific and the philosophical aspects of the mind-body problems becomes imperative. "Irreducibility" may mean non-derivability from a specified set of premises; but in other contexts it may mean non-translatability (non-synonymy, non-equivalence in the logical sense). To illustrate: many physical phenomena of sound or heat are derivable from the kinetic theory of molecular motion. In this sense certain parts of acoustics and of thermodynamics are reducible to mechanics, with a high degree of approximation at least within a certain limited range of the relevant variables. But the phenomena of heat radiation (and similarly those of optics, electricity, magnetism, and chemistry) are not reducible to mechanics. Whitehead speculatively maintained that the laws pertaining to the motion of electrons in living organisms differ fundamentally from the laws of electrons in the context of inorganic lifeless bodies. In a similar vein the physicist Elsasser (95, 96, 97), following some suggestions contained in Bergson's views on organic life and memory, regardsthe physical laws as special or limiting cases of biological laws. This is a drastic reversal of the "Victorian" outlook according to which macro-regularities are (usually) explainable in terms of basic micro-laws.1

As a student of the history and the methodology of modern science^ and impressed as I am with the recent advances of biophysics, biochemistry, and neurophysiology, I am inclined to believe strongly in the fruitfulness of the physicalistic research program (involving micro-explanations) for biology and psychology. But qua analytic philosopher my intellectual conscience demands that I do not prejudge the issues of reducibility (explainability) in an a priori manner. Beyond the sketchy empirically oriented arguments which I am going to submit presently, I shall address myself later on primarily to the logical and epistemological aspects of the mind-body problem.

Along empirical lines I believe there are differences, in principle capable of test, between parallelism and interactionism (and/or emergentism). Psycho-neurophysiological parallelism is here understood as postulating a one-one, or at least a one-many, simultaneity-correspondence between the mental and the physical. Parallelism as customarily conceived clearly rules out a many-one or a many-many correspondence. This latter type of correspondence, if I may speak for a moment about the motivation rather than the evidential substantiation (confirmation), is generally unpalatable to the scientific (especially the "Victorian") point of view, because it would obviously limit the predictability of mental events from neurophysiological states of the organism. But given a "dictionary," i.e., more properly speaking, a set of laws correlating in one-one or many-one fashion physical and mental states, physical determinism is not abrogated.

Two important qualifying remarks are in order here: (1) By "physical determinism" I mean, of course, that degree of precise and specific in-principle-predictability that even modern quantum physics would allow as regards the macro- and some of the micro-processes in organisms. (2) By "physical" I mean2 the type of concepts and laws which suffice in principle for the explanation and prediction of inorganic processes. If emergentism is not required for the phenomena of organic life, "physical" would mean those concepts and laws sufficient for the explanation of inorganic as well as of biological phenomena. In accordance with the terminology of Meehl and Sellars (221), I shall henceforth designate this concept by "physical2" in contradistinction to "physical1", which is practically synonymous with "scientific", i.e., with being an essential part of the coherent and adequate descriptive and explanatory account of the spatio-temporal-causal world.

In view of what was said above about the empirical character of the interaction and the emergence problems, the concepts of mental states might well be physical1 concepts, in that they could be introduced on the basis of the intersubjective observation language of common life (and this includes the observation language of science). Just as the concept of the magnetic field, while not denoting anything directly observable, can be introduced with the help of postulates and correspondence rules (cf. Carnap, 73), so it is conceivable that concepts of vital forces, entelechies, "diathetes" (cf. Kapp, 172, 173, 174), and mental events might be given their respective meanings by postulates and correspondence rules. Of course, the question remains whether such ("emergent") concepts are really needed and whether they will do the expected job in the explanation and prediction of the behavior of organisms, subhuman or human. My personal view, admittedly tentative and based on the progress and partial success of physicalistic micro-explanation (implemented by Gestalt and cybernetic considerations), is to the effect that physical2 laws will prove sufficient. But, having abandoned the all too narrow old meaning criteria of the earlier logical positivists, I would not for a moment wish to suggest that the doctrines of emergence or of interactionism are scientifically meaningless.

Let us then return to the empirically testable difference between interactionism and emergentism on the one hand, and parallelism on the other. An obvious and picturesque analogy or model for the interactionist view may be suggested here to provide a more vivid background. Billiard balls are in motion on a billiard table, and their motions are, we assume, predictable on the basis of mechanical laws (Newton's, supplemented by the laws of friction and of partially elastic collision). But imagine now a mischievous boy standing by, once and again pushing this or that ball or lifting some ball from the table. The mechanical laws, combined with a statement of initial conditions for the balls and the table, at a given moment, will then no longer suffice for the prediction of the course of the balls. The system in this case is of course an open one. If we could proceed to a larger closed system including the boy, with information about his shifting desires and so forth, deterministic predictability might be restored. (Since prediction of the boy's actions is precisely the issue at stake, I shall not beg any questions here and shall leave the boy's behavior unexamined for the moment.) This model is merely to illustrate a good clear meaning of "interaction". The boy watches the balls and his actions are in part influenced "by their momentary distribution and motions on the board. The events on the board are in turn influenced by the boy's actions. From the point of view of ordinary usage, it is proper to employ the word "interaction" perhaps only when we deal with causal relations directed both I ways between two continuants (things, organisms, persons, etc.).

But even a theory of emergence, such as the one suggested, though not definitely endorsed, by Meehl and Sellars (221), is conformable in principle by showing that physical2 determinism does not hold. Mental states or raw feels, be they regarded as states of an interacting substantial mind (or soul) or as values of emergent scientific variables, would in any case entail a breach in physical determinism. The system of neurophysiological events inasmuch as it is describable in physical terms would have to be regarded as open not only in the usual way, i.e., in regard to the extraneural, let alone extradermal, events, but it would also be open in regard to the set of mental events with which they are assumed to be causally (functionally) related in a way that would make them radically different from a set of mere epiphenomena. Now, while it is admittedly difficult at present to test for the implied breach in physical2 determinism, the idea is not metaphysical in the objectionable sense that empirical evidence could not conceivably confirm or disconfirm it.

Much depends in this issue upon just how the "interactors" or the "emergents" are conceived. Traditional vitalism, culturally and historically perhaps a descendant of more primitive forms of animism, stresses the capricious nature of vis vitalis and of anima. (In our model the boy by the billiard table is assumed to exercise "free" choice.) But interaction need not be indeterministic in the wider system. The wind and the waves of the sea genuinely interact; even if the wind's influence is quantitatively greater, the waves do have some effect upon the air currents nearby. But though precise prediction of detail is practically extremely difficult because of the enormous complexities of the situation, this type of interaction is in principle deterministically3 analyzable in terms of the functional relations of the two types of variables. Even the individual "free" or "capricious" momentary choices of our boy might be predictable in principle; but here the practical feasibility is far beyond the horizon of current psychology. At best only some statistical regularities might be formulated.

Determinism, inasmuch as it is allowed for by current physical theory, is also the presupposition of the sophisticated conception of emergence as presented in the essay by Meehl and Sellars. Here we have no interacting things or substances, but scientific variables intertwined in such a way that certain values in the range of one set of variables are functionally so related to the values of the variables in the other set, that the relations in the second set are nomologically different from what they would be if the values of the first set are zero. More concretely, once mental states have emerged, their very occurrence is supposed to alter the functional relations between the neurophysiological (physical2) variables in a manner in principle susceptible to confirmation. While my (scientific) predilections are completely incompatible with this ingenious and fanciful assumption, I do consider it scientifically meaningful. I just place my bets regarding the future of psychophysiology in the "Victorian" direction. And I admit I may be woefully wrong.4

With the foregoing remarks I hope to have indicated clearly enough that I consider these basic issues as essentially scientific rather than philosophical. But a full clarification and analysis of the precise meanings and implications of, respectively, parallelism, isomorphism, interactionism, and the various forms, naive or sophisticated, of emergentism is a philosophical task. I shall now develop the philosophical explication of the factual-empirical meaning of these assorted doctrines a little further and bring out their salient epistemological points. Parallelism and isomorphism, now that we have recovered from the excesses of positivism and behaviorism, are generally considered as inductively confirmable hypotheses. Reserving more penetrating epistemological analyses, especially of the "immediate experience" and "other minds" problems, until later, I assume for the present purpose and in the vein of the recent positions of Ayer (15, 18) and Pap (243, 248) that the ψ-φ (i.e., psycho-neurophysiological) relations or correspondences can be empirically investigated; and that mental states (raw feels) may by analogy be ascribed to other human beings (and higher animals), even if in the case of those "others" they are inaccessible to direct confirmation.

Parallelism, then, in its strongest form assumes a one-to-one correspondence of the ψ's to the φ's. It is empirically extremely likely that these correspondences are not "atomistic" in the sense that there is a separate law of correspondence between each discernible ψ1 and its correlate φ1. It is quite plausible that, for example, different intensities of a phenomenally given tone (e.g., middle C), at least within a given range, are correlated with corresponding values in a limited range of some variable(s) of the neural processes in the temporal lobe of the brain.

Isomorphism as understood by the Gestalt psychologists (Wertheimer, Köhler, and Koffka) and the cyberneticists (Wiener, McCulloch, Pitts, etc.) assumes an even more complete one-one correspondence between the elements, relations, and configurations of the phenomenal fields and their counterparts in the neurophysiological fields which characterize portions of cerebral, and especially cortical, processes. As mentioned before, this sort of approach would also countenance a one-many correspondence of ψ's and φ's. In that case, mental states would (with the help of the ψ-φ "dictionary") still be uniquely inferable from neurophysiological descriptions. But many-one or many-many correspondences, even if expressed in terms of statistical laws, would seriously restrict such inferences from specific φ's to specific ψ's. I know of no good empirical reasons for assuming anything but one-one correspondence; or one-many if very exact and detailed φ-descriptions are used, and if account is to be taken of the limited introspective discernibility of the ψ's from one another.

Interactionism, as I understand (but reject) it, would entail a many-one or many-many correspondence. Arthur Pap (242, p. 277), however, argued that there is no empirically confirmable difference between parallelism and interactionism. This, he thought, is because lawful relations or functional dependencies are the modern scientific equivalent of the cause-effect relation. Temporal succession, he maintains, is not a criterion of causal connection. While I admit that the most general conception of the causal relation is simply that of a (synthetic) sufficient condition,5 and is thus free of any connotation regarding the temporal succession of cause and effect; and though I also agree that in the case of ψ-φ relation it would seem rather fantastic to assume anything like a time difference, I think that the interaction hypothesis differs in its empirical meaning from parallelism or isomorphism in that it entails a breach of physical determinism for the φ's. This, if true, could in principle be confirmed by autocerebroscopic evidence. For example, the experience of volitions as directly introspected would not be correlated in one-to-one (or one-many) fashion with simultaneous cortical states as observed (really inferred) by looking upon the screen of a cerebroscope,6 and regularly succeeded by certain processes in the efferent nerves of the brain, ultimately affecting my muscles or glands, and thus ensuing in some act of behavior. This is the sort of most direct evidence one could ever hope for, as regards the confirmation of ψ-φ action. If the idea of interaction, i.e., action both ways between the ψ's and the φ's, is entertained, then there should be sensations (produced by the chain of processes usually assumed in the causal theory of perception, but) not strictly correlated with the terminal cortical events.

Characteristically, philosophers have been emphasizing much more the action of "mind on matter" -- as in voluntary behavior, or in the roles of pleasure, pain, and attention -- than that of "matter on mind." This asymmetrical attitude usually comes from preoccupation with the freewill puzzle, or related to this, from some remnants of theological ideas in the doctrines of an ideal ("noumenal") self. But the freewill puzzle -- even if some details of its moral aspects still await more clarification -- has in its scientific aspects been satisfactorily resolved by making the indispensable distinctions between causality and compulsion (and indeterminism and free choice). The perennial confusions underlying the freewill perplexity, truly a scandal in philosophy, have been brilliantly exposed by empiricist philosophers.7

The main reasons why most psychophysiologists (and along with them many philosophers) reject the hypothesis of ψ-φ-many-one or many-many correspondence are these:

1. Normal inductive extrapolation from the successes of psychophysiology to date makes it plausible that an adequate theory of animal and human behavior can be provided on a neurophysiological basis. Most physiologists therefore favor ψ-φ parallelism or epiphenomenalism. Parallelism, I repeat, is here understood as the assertion of the one-one (or, at least, one-many) ψ-φ correspondence, and not, as by Wundt and some philosophers, as the doctrine of double causation, i.e., involving parallel series of events with temporal-causal relations corresponding (contemporaneously) to one another on both sides. Causality in the mental series is by far too spotty to constitute a "chain" of events sufficiently regular to be deterministic by itself. Epiphenomenalism in a value-neutral scientific sense may be understood as the hypothesis of a one-one correlation of ψ's to (some, not all) φ's, with determinism (or as much of it as allowed for by modern physics) holding for the φ-series, and of course the "dangling" nomological relations connecting the φ's with the ψ's. According to this conception voluntary action as well as psychosomatic processes, such as hysteria, neurotic symptoms, and psychogenic organic diseases (e.g., gastric ulcers) may ultimately quite plausibly be explained by the causal effects of cerebral states and processes upon various other parts of the organism; only the cerebral states themselves being correlated with conscious (or unconscious8) mental states.

2. While the cultural and historical roots of the epiphenomenalist doctrine may be the same as those of traditional materialism, we can disentangle what is methodologically sound and fruitful in the materialistic point of view from what is cognitively false, confused, or meaningless. The fundamental methodological reason for the rejection of interactionism, or the (equivalent) adoption of ψ-φ-one-one (or one-many) correspondence as a working hypothesis or research program, however, is this: If the ψ's are not inferable on the basis of intersubjectively accessible (observed, or usually, inferred) φ's, then their role is suspiciously like that of a deus ex rnachina. The German biologist-philosopher Driesch admitted this candidly, and thereby gave his case for vitalism away. He said that the intentions of the entelechy could be inferred only post factum, but could not be predicted from antecedent physical conditions. This is just like the case, in our crude analogy, of the capricious boy at the billiard table. After he has removed a ball we may say that he intended (perhaps!) to avoid a collision of the red ball with the white one. According to the vitalist interactionist doctrine, the volitions of the boy are in principle unpredictable on the basis of any and all antecedent conditions in his organism and the environment. Interactionism so conceived assumes causal relations between the elements in the series of mental states, the series of physical states, but also some crossing from the set of mental states to the physical ones and vice versa. In the model of the wind and the waves, we have precisely this sort of schema exemplified. But notice the crucial difference. A closed system (or a system with known initial and boundary conditions) is here conceivable in which all relevant variables are ascertainable intersubjectively and antecedently to the prediction of later states of the system; whereas in the case of ψ-φ interaction, intersubjective and antecedent confirmation of the ψ-states is ex hyporhesi excluded.

The flavor of the theological arguments from design and of primitive animistic explanations of nature and human behavior permeates interactionistic explanations. They are at best ex post facto explanations. This sort of explanation, while not as satisfactory as explanations that also have predictive power, is nevertheless quite legitimate and is frequently the best we can provide in complex situations. Earthquakes are notoriously unpredictable (i.e., practically unpredictable), but once we observe a certain case of large scale destruction, its explanation in terms of an earthquake is perfectly legitimate even if the precise location of each piece of rubble in the shambles is far from predictable. Biologists are satisfied with evolutionary (retrospective) explanations of the emergence of a new species, even though they could never have predicted this emergence in any specific detail. Given the species in the Cambrian epoch, and given the principles of genetics and of Neo-Darwinian evolution, nobody could inductively infer the emergence of the chimpanzee or of the orchid; nevertheless, the very partial explanations of the theory of evolution are scientifically significant, acceptable, and helpful. Explanations of historical phenomena like wars, revolutions, and new forms of art furnish another illustration for the same type of ex post facto explanations. Finally, for an example in the psychological domain, if we find that a man has written dozens of letters of application for a certain type of job, we infer that he was impelled by a desire for such a job, even if we could not have predicted the occurrence of this desire on the basis of antecedent and intersubjectively confirmable conditions.

It is important, however, to notice again the decisive difference between explanations for which it is at least in principle conceivable that they could be predictive (as well as retrodictive), and those which ex hypothesi are only retrodictive. Scientists are predominantly interested in enlarging the scope of predictive explanations. The opposition against vitalism then stems from a reluctance to admit defeat as regards predictability. And the opposition against ψ-φ interactionism stems furthermore from the reluctance to admit antecedents which are only subjectively accessible into the premises (regarding initial conditions) for predictive inferences. Expressing the same idea positively, we may say that it is part of the methodology or of the over-all working hypothesis of modern science that prediction, to the extent that it is possible at all (taking account of the basic quantum indeterminacies), is always in principle possible starting from inrersubjectively confirmable statements about initial conditions. Scientists have, on the whole, adjusted themselves to the limitations involved in statistical prediction and probabilistic explanation. Very likely nothing better will ever be forthcoming in any area except in the few where classical determinism holds with a high degree of approximation. Of course, a logical distinction should be made between those cases in which the restriction to probabilistic predictability is a consequence of the complexity of the situation, and those in which the theoretical postulates of a given domain are themselves formulations of statistical laws. Although one can never be sure that this distinction is correctly drawn or that the dividing line will remain in the same place during the progress of science, the distinction can be drawn tentatively in the light of theories well confirmed at a given time.

But scientists are radically opposed to the admission of purely subjective factors or data (conceived as in principle inaccessible to intersubjective confirmation) as a basis for prediction or explanation. This would indeed be scientifically meaningless, if not even statistical relations of subjective states to antecedent or consequent intersubjective observables could be assumed. If they are assumed, then the subjective states are not purely subjective or "private" in the radical sense intended by some interactionists. The "emergent" raw feels in the interpretation by Meehl and Sellars are of course subjective only in the sense that they can be the objects of direct introspective verification, but they are also intersubjective (physical1) in the sense that they can be assumed (posited, inferred, hypothetically constructed) by scientists who do not have the same sort of raw feels in the repertory of their own direct experience. This is so, for example, in the case of a congenitally blind scientist equipped with modern electronic instruments who could establish the (behavioristic) psychology of vision for subjects endowed with eyesight. The blind scientist could thus confirm all sorts of statements about visual sensations and qualities -- which in his knowledge would be represented by "hypothetical constructs." But if ex hypothesi all connections of the subjective raw feels with the intersubjectively accessible facts are radically severed, then such raw feels are, I should say by definition, excluded from the scope of science. The question whether discourse about such absolutely private raw feels makes sense in any sense of "sense" will be discussed later.

The upshot of this longish discussion on the difference between the scientific and the philosophical components of the mind-body problems is this: If interactionism or any genuine emergence hypotheses are sensibly formulated, they have empirical content and entail incisive limitations of the scope of physical determinism. Interactionism is more difficult to formulate sensibly than is the (Meehl-Sellars) emergence hypothesis. In one form it requires substances (things, continuants or systems of such) for a normal use of the term "interaction," and in this form there seems little scientific evidence that would support it. I have read a great many arguments by metaphysicians attempting to support the idea of a totally (or partially) immaterial "self." But I have never been able to discern any good cognitive reasons beneath their emotionally and pictorially highly charged phrases.'Whatever role the self (in Freudian terms perhaps the total superego, ego, and id-structure) may play in the determination of human conduct, it may yet very well be explained by a more or less stable structure of dispositions due to some constitutionally inherited, maturationally and environmentally modified, and continually modulated structure of the organism (especially the nervous and endocrine systems.).

In another form interactionism (without a self) would require "spontaneously" arising mental states, i.e., an indeterminism not even limited by statistical regularities, and this again is neither supported by empirical evidence, nor advisable as a regulative idea for research. Nor is it required for the solution of the freewill problem, or for an account of the causal efficacy of mental events in the course of behavior. As regards the emergence hypothesis (a la Meehl and Sellars), this clearly makes sense, but whether it is really needed for the explanation of behavior is an open question. In the spirit of the normal procedures of scientific induction and theory construction I remain conservative in thinking that the rule of parsimony (Ockham's razor, or Newton's first regula philosophandi) warns us not to multiply entities (factors, variables) beyond necessity. If the necessity should become evident in the progress of research, I shall cheerfully accept this enrichment of the conceptual apparatus of science; or, ontologically speaking, this discovery of new entities in our world. In the meantime, I remain skeptical about emergence, i.e., optimistic about the prospects of physical2 determinism. And, as I shall argue from the point of view of epistemology in sections IV and V, the sheer existence of raw feels is not a good reason for holding an emergence doctrine.

Another philosophical issue which needs careful separation from the scientific problems among the mind-body tangles is that of the "intentionality" of the mental. (For expository reasons the discussion of this issue will be reserved for section IV F.)


Notes

1. I have dealt elsewhere (106, 108, 112, 113, 115, 116) with the logic and methodology of such explanations. See also the important articles by E. Nagel (230, 232); Hempel and Oppenheim (152); Kemeny and Oppenheim (177); Oppenheim and Putnam (in Minnesota Studies in the Philosophy of Science, Vol. II).

2. In this context only; other meanings of "physical" will be listed and discussed in sections IV and V.

3. Again it is only to the extent that hydrodynamics and aerodynamics for macro-processes are (approximately) deterministic.

4. In his earlier formulations of the general theory of relativity Einstein endorsed the so-called Mach principle, according to which centrifugal and inertial forces are the effects of accelerations relative to the total masses of the fixed-stars-universe. But, impelled by what he considered cogent physical and mathematical arguments, he later ascribed those effects to a relatively independently existing "Führungsfeld" (guiding field). I mention this merely as a somewhat remote logically parallel case from an entirely different domain of science. Naturally, my expectation here is that something of Mach's principle, even if in strongly modified form, will be salvaged. Powerful inertial forces as effects of a self-existent metrical field seem extremely implausible to me.

5. And in the laws of classical mechanics and electrodynamics of sufficient and necessary condition.

6. This, for the time being, of course, must remain a piece of science fiction (conceived in analogy to the doctors' fluoroscope) with the help of which I would be able to ascertain the detailed configurations of my cortical nerve currents while introspectively noting other direct experiences, such as the auditory experiences of music, or my thoughts, emotions, or desires.

7. Hobbes, Locke, and especially Hume, Mill, Sidgwick, Russell, Schlick (301); and Dickinson S. Miller, cf. the superb article he published under the pseudonym "R. E. Hobart" (157). See also C. L. Stevenson (329); University of California Associates (339); A. K. Stout (330); and Francis Raab (271).

8. The terminological question whether to speak of the unconscious as "mental" will be discussed in sections IV and V.