Objectivism and Interactionism OBJECTIVISM AND INTERACTIONISM* JOSEPH MARGOLIS University of Western Ontario The views of linguistic analysts and objectivists are explored with regard to the question of interactionism. It is argued that the admission of a logical difference between explanation by cause and explanation by motive cannot disqualify causal explanations of human action, cannot be construed as challenging the competence of science, and cannot count against interactionism. It is also argued that objectivist programs for eliminating mentalistic concepts either implicitly admit interactionism or cannot distinguish relevantly between interactionism and parallelism. I wish to consider the possibility of a rapprochement between the accounts of mental life that are provided by adherents of linguistic analysis and of that kind of philosophy of science that speaks of the "eliminability" of mentalistic concepts. I should not wish to be misunderstood, however. For one thing, so-called linguistic analysts are by no means in agreement on the critical issues. For another, there may well be "linguistic" objections pressed against the programs of behaviorism, physicalism, objectivism that, rightly or wrongly, disallow any complete reconciliation. For a third, not all philosophers of science would subscribe to theories of the sorts mentioned. And for a fourth, the rapprochement intended affects no more than a part (however substantial) of the range of issues concerning relations between the mind and the body. I am interested primarily in isolating possible sources of disagreement that promise to be resolved as well as those that threaten to resist treatment more stubbornly; and in doing so, I wish to concede whatever may be conceded to the objectivist program and, at the same time, to challenge the program at a point that may be said to question, in its own terms, the adequacy of all such programs. I have arranged the discussion under two principal headings. 1. Motives are sometimes causes and sometimes not causes. Greed, for example, may be both the reason for an action and the cause, or part of the cause, of an action. One's reasons for an action are not causes of the action; that is, reasons and causes are distinct and irreducible categories but what serves as a reason may serve as a cause. It is true that some linguistic analysts ([6], [8]) hold that motives cannot be causes, in the strong sense that motive and cause are logically independent categories, but I do not find that they often address themselves to the question of whether what may be described under the category of a reason may also be described under the category of a cause. Melden, it is true, holds that to treat a motive as a cause is incoherent ([6], Ch. IX). But he also admits (possibly inadvertently) that a motive is the motive "for some action performed or performable by the agent" ([6], p. 77)1 and this clearly eliminates the alleged incoherence. Also, Peters, it is true, regards motives as "a particular class of reasons" and is very critical of the attempt to construe motives as causes and views causal theories of motivation as confused ([8], Ch. II). Nevertheless, a careful reading of Peters will show that he does not (and could not) disqualify, say, drive theory; he objects, really, not to causal explanations of the "directedness" * Received February, 1965. My italics. 118 OBJECTIVISM AND INTERACTIONISM 119 of, or the motivational patterns of, behavior, but rather to incorporating such ex- planations "in the meaning of 'motive' " ([8], p. 40). Again, I take this to mean (though Peters may have no interest in agreeing with this) that what serves as a motive may also serve as a cause. On the other hand, a commentator like P. H. Nowell-Smith has no difficulty in holding that moral judgments concern "actions that are caused by characteristics that can be strengthened by praise and blame" ([7], p. 56), which suggests that linguistic analysis is by no means uniformly committed to Melden's or Peter's way of speaking. In the same spirit, Anthony Kenny explicitly remarks that "ca single state of affairs [may be] both object and cause of the same emotion" ([4], p. 75), which is in effect to say that, though two conceptual categories may be logically quite different, it is possible that one and the same "state of affairs" may be appropri- ately characterized in terms of both such categories. To admit that what may be a reason may be a cause is not, however, to admit that whatever may be a reason may be a cause. If I designate the purpose of an action, I have not designated what may serve as a cause, even if a cause in some sense corresponding to the given purpose may be specified and even if an action to which we may assign a purpose may be causally explained. It is true that Peters allows causal explanations of actions only in so far as an action reduces to a mere happening, and thus departs from some rule-following model of action (a view which may surely be challenged) ([8], Ch. I). But a close reading of Peters shows conclusively that even he holds that since actions "can never" be specified "exhaustively in terms of movements of the body or within the body" ([8], p. 12), actions can never "be sufficiently explained in terms of causes, though, of course, there will be many causes in the sense of necessary conditions" ([8], p. 12). The upshot is that analysts who distinguish sharply between causes and reasons are not bound to deny (and ought not to deny) that causal explanations are relevant to understanding human actions; they hold at best only that, in that respect in which reasons are not causes, human actions cannot be adequately understood in causal terms alone. I am prepared to admit, I may say, that the issue is not often clearly formulated by the most interested analysts. For example, though Peter Winch says straight- forwardly that he wants "to show that the notion of a human society involves a scheme of concepts which is logically incompatible with the kinds of explanation offered in the natural sciences" ([10], p. 72), the sequel shows (what he admittedly does not see) that the concepts he has in mind are not incompatible but complementary. In fact, he even allows that "it is sometimes possible to predict decisions; only that the evidence on which such predictions are based must be different in kind from that on which scientific predictions are based" ([10], p. 93). This is, of course, a multiple confusion. What Winch wishes to say is that decisions are significant in intentional terms and that prediction would have to accommodate this fact. He fails to see that an objectivist is not unable to accommodate the intentional aspect of actions, that the prediction of decisions is not impossible from the point of view of science, that emergence and novelty do not as such theoretically handicap science, that the difference in explanation by causes and explanation by motives is not one of logical incompatibility but of logically distinct interests. Nevertheless, there is justification for his view that when we speak of understanding human societies, we do wish to attend precisely to motivational and intentional distinctions that are different in kind from causal distinctions. It needs to be insisted that to draw a distinction between reasons and causes does not in the least affect the possibility of an objectivist science of man. For one thing, 120 JOSEPH MARGOLIS the distinction does not imply that reasons are not empirically and publicly accessible. A man's motives and reasons are open to scrutiny, in principle, even if they happen to be secret. It is even possible, contrary for instance to the view of Stuart Hampshire ([3]), that a man may not know his own motives and reasons, while another may know them. Also, a man's actions, qualified by motives and reasons, may be causally explained. Also, the motives and reasons and purposes of a given action need not be, nor need correspond to, specific mental acts. But if, as May Brodbeck for instance concedes, favoring an objectivist psychology, motives are sometimes not causes and explanation by motives is compatible with, but distinct from, explanation by causes ([2], 321), it cannot be maintained, in the sense required, that "talk about causes [may] replace, without residue, talk about motives" ([2], 323). It may be (though it may be debated) that extensionally equivalent expressions "containing only physio- logical, behavioral, and environmental terms" may be provided for any expression about "a person's motives or other mental states" ([2], 323), but this is a question altogether different from that of the replacement of talk about motives by talk about causes. It follows further that, though an adequate causal explanation of an action may be provided, causal explanations of an action may not be adequate in the sense, precisely, that an explanation by motives may be required in order to understand actions as such. In short, on the concessions made, no objectivist view can (or needs to) ([2], 310) hold that motivational talk can be suitably replaced by causal talk. The concession that statements that are (synthetically) extensionally equivalent do not have the same meaning intentionally or referentially (sic) is not quite addressed to the ques- tion of replacing talk about motives by talk about causes ([2], 323). But of course, all this is compatible with the objectivist view that talk about motives (as distinct from causes), though eligible in public discourse, is irrelevant to science. With this in mind, the apparent argument dissolves. Analysts like Peters and Melden and Winch are primarily concerned to insist on the impossibility of eliminating motivational, non- causal talk in the description and explanation of human actions. And objectivists like Brodbeck are primarily concerned to insist that the causal explanation of human actions is compatible with the logical difference between causal and motivational talk. 2. Mental and physical states and events may interact causally. An injection may produce a pain and a pain may cause my body to stiffen. Greed may drive me to larceny and darkness may terrify me. These are commonsense admissions. Nevertheless, objectivism holds that "a mental state can affect the body only through its corresponding physical state. [T]he objectivist term 'anxietyo' is an indirect description of the mentalistic 'anxietym'. They may thus each be used to express the same2 fact [i.e., they express the same fact in that the terms in which they differ are related to each other as direct and indirect descriptions or ways of referring]. In particular, if it is true that anxiety, causes ulcers, we may also truly say that anxietym causes ulcers" ([2], 323). The force of this and similar remarks is not altogether easy to assess. For if the mentalist and the objectivist statements express the same fact, in what sense can it be maintained that mental states affect the body only through their corresponding physical states? The point seems to be that only if a causal statement expressed in objectivist terms is true is there a warrant for the truth of the corresponding mentalist statement. Causal statements are restricted to the physical order but common-sense interactionist statements are allowable on the grounds of the relationship direct and indirect description and reference. As Brodbeck says: "mental events parallel but do not interact with physical events. OBJECTIVISM AND INTERACTIONISM 121 In other words, despite the correlation between mind and body, no mental state causally affects any material state and conversely. In a phrase, the view is that the physical world, including men's bodies, is causally closed with respect to minds" ([2], 313). If I understand this thesis correctly, it can be challenged by showing that the denial of interactionism follows from the assigned meaning of 'cause' rather than merely from empirical facts and laws. Alternatively put, objectivism and interactionism can be shown not to subscribe to the same sense of 'cause' (cf. [2], 314). Suppose I hold that a pain may cause one's body to stiffen. Questions may be raised about the nature of the public criteria by which I identify pains but, short of holding that sensations are brain processes or bits of behavior or dispositions, these criteriological questions do not bear at all on the causal thesis posited. I claim that a pain may cause one's body to stiffen but I concede that I rest my claim on my ability to identify pains by means of physiological, behavioral, and environmental circumstances. Furthermore, I insist that whatever serve as the criteria of pain are not, in any sense, identical with pain. If this is the case, then no argument based on criteriological considerations can possibly bear on the apparently factual claim that pain may cause one's body to stiffen; the rejection of interactionism, in this setting, can only be achieved by defining 'ause in an appropriate way. Now, suppose I succeed in providing an indirect description of pain on the model of construing 'the color with the longest wave length' as a way of indirectly referring to the color red ([2], 323). In that case, I shall have to hold that 'pain,' designates the cause of what 'pain,,' directly refers to, since "light waves cause but are not colors" ([2], 323) I shall, therefore, have to admit interactionism. Suppose, finally, I succeed in providing an indirect description of pain, in physical terms, that is not to be construed causally. Brodbeck does not specify what such a model would be like, but it seems clear that if it is not a causal relationship that is employed it may be an identity of some sort. Brodbeck does speak of 'thinking-about-Italym' and 'thinking- about-Italy,' as having a relationship similar to that holding between light waves and colors ([2], 323), but she cannot have intended this consistently. On the other hand, if it is an identity, then once again interactionism will have to be admitted; that is if Sm is identical with S, and SO is causally connected with To then Sm is causally connected with To. This, for instance, would be the consequence of J. J. C. Smart's efforts ([9], [5]). But if no program of these sorts can disqualify interactionism, a fortiori no arguments based merely on considerations of extensional equivalences (non-causal correlations) can either. That is, a non-causal correlation is either irrelevant to the issue (in so far as it is not addressed to it) or question-begging (in so far as interactionism is denied by linguistic fiat). The issue remains open whether, on objectivist grounds, the difference between these two theories (interactionism and parallelism) can be based on empirical distinctions or reduces merely to a question of defining 'cause'. I do not yet see how the distinction can be maintained. The only other possibility that arises is that, in science, what are described and explained causally are never mental states and events but only physical states and events. This may be true and, if it were, it would considerably complicate the characterization of the alleged quarrel between objectivism and interactionism. For then, it would not be true merely (for an objectivist) that mental states do not enter into causal relations with physical states but rather that mental states as distinct from physical states have no place in science. This would also bear on the strategy of challenging objectivism, since not only could it be said that interactionism is ruled 122 JOSEPH MARGOLIS out by the meaning of 'cause' but also that the very nature of what are admitted to be the facts and laws of science preclude interactionism. Now, this is an argument that threatens to become a mere quibble, though it is not a mere quibble to contrast the order of science (so construed) and the order of commonsense interactionism. That is, consider that, in ordinary discourse, not only are mental states said to enter into causal relations with physical states but also mental states are mentioned and described. The objectivist, who eliminates mental concepts (as far as science is concerned) does not even acknowledge (within his program) the existence of precisely what the interactionist holds a special thesis about. It seems fair to say, on this interpretation, that there can be no possible confrontation between the objectivist and the interactionist. But this is emphatically not to say that, in some sense that may be specified, the objectivist theory of science may be defended (or challenged) as providing a "true" or "correct" picture of science. We may look at this issue in a somewhat different way. Brodbeck says, helpfully: "We say that mental anguish, for instance, anxiety, causes ulcers or that a bodily wound causes pain. Our common language is interactionist through and through. Extensional equivalence and direct versus indirect reference permit us to fit scientific parallelism with this piece of interactionist commonsense. On the objectivist account, a mental state can affect the body only through its corresponding physical state" ([2], 323). The difference between interactionism and parallelism reduces, then, to this. The objectivist holds that the linking of direct and indirect descriptions of the same2 fact is a linking of science and non-scientific commonsense; the interactionist holds that the linking obtains within the boundaries of science itself. Hence, the objectivist, either finding or recommending the meaning of Ccause' in science to be what he claims, translates interactionism into parallelism; the interactionist, holding to a different sense of Ccause', rejects parallelism. As far as I can see, on the grounds actually posited by Brodbeck, the cash difference between the two positions is zero. The fact is that she insists that direct and indirect descriptions can be given for mental events such that the relevant statements "express the same2 fact," in the sense that they refer to the same events. But if this is so, it makes no difference whether we talk about distinctions within science or between science and commonsense: parallelism and interactionism are, in this respect, equivalent. Gustav Bergmann, discussing much the same issue, admits the following: "Mental states cannot be described indirectly unless two conditions are fulfilled. For one, the parallelistic hypothesis must be true. For another, we must know enough of what according to the hypothesis is there to be known, or, as one says, can be known in principle" ([1], 227). He holds that "all the evidence favors the parallelistic hypo- thesis," though he supplies none ([1], p. 227). And he allows that "much of what is now being said about mental facts in the behavior sciences is mental discourse" ([1], 227). If I understand this correctly, then, in practice, mental events do enter into the causal accounts of science but, in principle, indirect descriptions (given the parallelistic hypothesis) may always be substituted for expressions in "mental dis- course." This inclines me to think that there are really two quite distinct kinds of parallelistic theories that are at stake in these discussions: one is the view that mental events never enter into causal relations with physical events; the other is the view that indirect descriptions in objectivist terms may always, in principle, be supplied for mental events. The first, as I have already in effect argued, cannot be a scientific question, concerns the meaning of 'cause', and is perhaps best characterized as a metaphysical question. The second, as I have also insisted, is either indistinguishable OBJECTIVISM AND INTERACTIONISM 123 from interactionism or compatible with it. The telltale clue, in Bergmann's account lies in his speaking alternatively of the "parallelistic hypothesis" ([1], 226) and of the "parallelistic frame of reference" ([1], 228). Let me, then, merely summarize the logic of the quarrel. Linguistic analysts characteristically insist that explanations of human actions cannot be given in causal terms. But they mean by this (in the best and most defensible versions) that explana- tion by motives is logically different from explanation by causes and that an under- standing of human action calls for explanation by motives. They need not deny, and cannot deny, that causal explanations may be given of human action. To the extent that they concede that motives may be causes, linguistic analysts provide for inter- actionism. The objectivist concedes also that explanation by motive and by cause are logically distinct. He cannot then argue that causal explanation may replace the other and he cannot hold that the support of the difference between these two sorts of explanation counts, in itself, as a charge against the competence of science. The fate of possible strategies against interactionism itself may be summarized in this way. Arguments based on criteriological considerations will be irrelevant, since if mental states are distinct from whatever serve as their public criteria, attention to the criteria alone will not bear on the causal question regarding mental states. If the relationship between indirect and direct descriptions of mental states is construed in causal terms, interactionism will be implicitly admitted. If the relationship between mental and physical states is construed as identity, interactionism will be implicitly admitted. If the relationship between mental and physical states is construed as empirical, noncausal correlations, information about such correlations will not yet bear on the causal question. Finally, if an objectivist denies that mental events enter at all in the causal explanations of science, no common ground can be provided for disputes between the two views. There do not seem to be any other relevant possi- bilities. REFERENCES [1] BERGMANN, Gustav, "Purpose, Function, Scientific Explanation," Acta Sociologica 5 (1962), pp. 225-238. [2] BRODBECK, May, "Meaning and Action," Philosophy of Science 30 (1963), pp. 309-324. [3] HAMPSHIRE, Stuart, Thought and Action (London, 1959). [4] KENNY, Anthony, Action, Emotion and Will (London, 1963). [5] MARGOLIS, Joseph, "Brain Processes and Sensations,' Theoria, forthcoming. [6] MELDEN, A. I., Free Action (London, 1961). [7] NOWELL-SMITH, P. H., "Freewill and Moral Responsibility," Mind, LVII (1948), pp. 45-61. [8] PETERS, R. S., The Concept of Motivation (London, 1958). [9] SMART, J. J. C., "Sensations and Brain Processes," Philosophical Review, LXVIII (1959), pp. 141-156. [10] WINCH, Peter, The Idea of a Social Science (London, 1958). 9 Article Contents p. 118 p. 119 p. 120 p. 121 p. 122 p. 123 Issue Table of Contents Philosophy of Science, Vol. 33, No. 1/2 (Mar. - Jun., 1966), pp. 1-198 Front Matter On Theoretical Constructs and Ramsey Constants [pp. 1-13] The Probabilistic Argument for a Non-Classical Logic of Quantum Mechanics [pp. 14-21] Von Neumann's Argument for the Projection Postulate [pp. 22-39] Pragmatic Aspects of Explanation [pp. 40-60] Discussion Dray on Rational Explanation [pp. 61-69] Projecting Unprojectibles [pp. 70-75] The Truth and Falsity of Definitions [pp. 76-79] Book Reviews Review: untitled [pp. 80-82] Review: untitled [pp. 82-84] Review: untitled [pp. 84-85] Review: untitled [p. 85] Review: untitled [pp. 85-87] Abstracts from Dialectica [pp. 88-89] Freedom and the Marxist Philosophy of History [pp. 101-117] Objectivism and Interactionism [pp. 118-123] Objectivism vs. Subjectivism in the Social Sciences [pp. 124-133] Extensionality and Randomness in Probability Sequences [pp. 134-146] On Semantic Pitfalls of Biological Adaptation [pp. 147-153] Discussion Fallacy of the Single Risk [pp. 154-162] Mr. Manier's "Theory of Evolution as Personal Knowledge": A Quasi-Reply [pp. 163-164] Deduction, Prediction and Completeness Conditions [pp. 165-167] A Corrected Model of Explanation [pp. 168-171] Is Geology Different: A Critical Discussion of "The Fabric of Geology" [pp. 172-185] Book Reviews Review: untitled [pp. 186-189] Review: untitled [pp. 189-190] Review: untitled [pp. 190-192] Abstracts from Inquiry [pp. 193-196] Back Matter [pp. 90-198]