Free Will (Again)
For a decade and more, the most active school of philosophy in English-speaking universities has been the “linguistic philosophy” that stems from the later writing and teaching of Ludwig Wittgenstein. An assumption hitherto basic to members of the school has been that traditional philosophical problems and controversies derive mainly from deep-lying misconceptions and consequent misuses and misapplications of “ordinary language.” Ordinary language, it is argued, provides the inescapable matrix of all significant expression and thought. This does not mean that one is never justified in departing from its rules—for example, through metaphorical extension, ad hoc contextual stipulation, or systematic redefinition. But such departures from the customary usages of words, however necessary they may be on occasion to the purposes of the poet or the scientist, are nearly always obfuscating when introduced into philosophical discussions. For unlike poetry, philosophy is not a form of free imaginative literature, interesting to us on its own account; nor, despite the pretensions of some of its professors, is it a science whose terms and distinctions are of necessity closely tied to well-developed experimental practices and operations. Whatever else it may, be (and perhaps one of the essential features of philosophy is the fact that its practitioners eternally disagree about what it is and what it is for), philosophy, first and last, is talk. But it is talk which somehow always feeds on itself and seemingly makes no egress from the “universe of discourse” into the surrounding world of things and of actions. The hallmark of a philosophical argument, so it is said, is that it is never won, and of a philosophical, problem that it is never solved. As one critic of linguistic philosophy succinctly puts it, philosophy, from this point of view, is simply “the pathology of language.” Or, to adapt to present purposes a striking phrase of Wittgenstein himself, in philosophy language is not working but “idling.”
It has been the business of linguistic philosophy to explain how this is so, the hope being that once the philosopher understands the source of his perplexities, his “mental cramps” will disappear—along with his philosophy itself. And in fact, the implication, never far beneath the surface, is that through the therapy of linguistic (or conceptual) analysis, philosophy, conceived as an academic subject or discipline, will, like the socialist state, gradually wither away and die.
Such a view of philosophy and of the role of linguistic analysis, although it understandably scandalizes the school philosophers, is nothing new; it is only northeast from the positions of Kant in the Critique of Pure Reason, and of Berkeley and Hume before him. All of these philosophers regarded most of traditional metaphysics, including what Kant called “rational theology,” as a bone-yard of conceptual antinomies and paralogisms. And they all took it upon themselves as philosophers’ philosophers to help in the removal of this litter. Hume’s celebrated analysis of the free-will problem provides a case in point. “I hope,” said he, “to make it appear that all men have ever agreed in the doctrine both of necessity and of liberty, according to any reasonable sense, which can be put on these terms; and that the whole controversy has hitherto turned merely upon words.” Hume then sought to show that the whole problem of “liberty and necessity” is factitious, the consequence of a radical failure on the part of muddle-heads to distinguish clearly the meanings and to take account of the proper domains of application of the main terms of the controversy. A free act, says Hume in effect, is not causeless, but, rather, a consequence of volitions and choices made by the agent himself. When (as we say) a person does what he wills (or chooses), then he acts freely and is, to that extent, a free agent. Correspondingly, a person is not at liberty and his behavior not free when what happens to him is owing, not to any volition or choice on his part but to some external force beyond his control. But clearly, freedom from external control does not in the least connote an absence of causal determination; on the contrary, a wholly uncaused movement, being something that occurs at random, is precisely not an action for which anyone can be held responsible. Therefore, not only do the essentially ethical conceptions of free and responsible action not entail that of an uncaused and unpredictable happening, but in fact they are inconsistent with it. Thus, if one means by “determinism” simply the principle of universal causation, according to which every state of affairs, including every human action, desire, and choice, has in principle some assignable cause, then determinism is inconsistent, not with free will and free action, but only with the possibility that something may happen for no reason at all. Indeed, as Hume pointed out, the shoe is on the other foot: for it is the indeterminist who, in renouncing the belief that human conduct follows from something “durable and constant” in human nature, is precluded from holding anyone responsible for his “actions” and hence from viewing him as an agent, free or otherwise. Law, government, education, even religion all depend upon the firm assurance that men act from assignable motives and for determining reasons; once that assurance is destroyed, the fabric of human intercourse and of civil society is torn beyond repair.
Such has been the position of many philosophers, including Bertrand Russell and A. J. Ayer, who pride themselves on their respect both for logic and for the prerogatives of scientific explanation. Significantly, the younger linguistic philosophers, judging by the contributions to Freedom and the Will1—a recent collection edited by D. F. Pears, and including pieces by such distinguished writers as Stuart Hampshire, Iris Murdoch, P. F. Strawson, and G. J. Warnock—find it much too simple. For one thing, it offers no clear or sustained account of the meanings of the key terms of the free-will problem. Eschewing metaphysical make-believe, how are we to construe such “complex and problematical” notions as those of efforts and acts of will; for that matter, what, if anything, is “the will” and in what circumstances can we be said to exercise it? As Professor H. L. A. Hart indicates in his acute discussion of “Acts of Will and Legal Responsibility,” the idea of an act of will provides no independent criterion of responsible action; indeed, it seems hardly more than a misleading stand-in for the variegated criteria actually applied in deciding what actions are not responsible. Much the same is true in the sphere of morals. An action is said to result from an act of will when someone can be held responsible for it. But there is no single standard for such an act to which appeal must be made in deciding when this is the case. Nor, evidently, can we establish what standards of non-responsibility are to be adopted merely by a nicer scrutiny of “the facts” of human behavior. For the exigent question remains: what sorts of facts are to be judged relevant or decisive in such matters? This, plainly, is a problem of evaluation which in our time has become increasingly difficult and unsettled.
However, it is not just the terms of the free-will controversy that are, or have become, problematical. The very meaning, or point, of the doctrine of “determinism” seems increasingly unclear. For instance, in the preceding redaction of Hume’s analysis I employed the crucial phrase “in principle” in formulating what I there called “the principle of universal causation.” But, as Mr. B. A. O. Williams points out in his excellent opening essay, “Freedom and the Will,” very great difficulties “surround the slippery phrase ‘in principle.’” Under what conditions, if any, could we know that in principle every event or action has a cause or that its occurrence is predictable? Moreover, how, if challenged, could the meaningfulness of the principle of determinism be established? It is, apparently, no sort of scientific hypothesis, for no one has ever managed to state the observable test-conditions for its falsification. Whatever happens, the determinist may stick to his belief that “in principle” it can be explained; nor can the indeterminist make anything of the occurrence of unprecedented events, for he is in no position to deny that the oddity which he imputes to them is owing merely to limitations of his own experience. Perhaps, as some of the contributors to Freedom and the Will rather mincingly suggest, determinism (whatever its proper formulation should be) is less a verifiable theory about the nature of things than a “program” which (I take it) may be desirable or undesirable, realizable or unrealizable, but not true or false. From this point of view, the questions at issue become: (1) what sorts of plans and practices does the program of “natural” science commit us to, especially within the sphere of human behavior? and (2) whether this program, to the extent that it is realizable, is consistent with the practices involved in moral judgment, agency, and responsibility.
Here it may be too much to hope any longer for a knock-down demonstration either of logical consistency or of inconsistency. One reason is that the terms in which the two programs are framed may no longer have the clarity necessary to such a demonstration. However, logical inconsistency is not the only kind of conflict in which rival (if they are such) doctrines may involve us. For example, it may turn out in practice that the attempt to carry out one program would make it practically difficult or in certain instances impossible to carry out the other. Or, as the program of scientific explanation, including the diagnosis and prognosis of individual or group behavior, is carried through in ever increasing detail and range, it may be that the image of man as a responsible and free agent will come to seem inconsequential.
The contributors to Freedom and the Will do not discuss these questions in any detail; nor do they make clear just what is involved in conceiving of determinism as a program. And in general it must be reported that their discussions, although stimulating, are too often impressionistic and inconclusive. Moreover, they frequently exhibit precisely the same propensity for unsubstantiated and what I shall call, “tendentious” generalization which the linguistic philosophers have rightly criticized in the works of their predecessors. For instance, in the lively fourth chapter, a symposium on “Determinism,” the participants (Warnock, Strawson, and J. F. Thompson) debate the conceptual relationships holding between what may be called “the language of action” (which, at least for most academic Americans, is misleadingly referred to here as “the language of psychology”) and “the language of events and processes” (which they speak of as “the language of physics”). Now two of the contributors, Mr. Strawson and Mr. Warnock, appear to think that the language of action (which essentially involves such characteristic concepts as reasons, intentions, choices, and responsibilities) is not only distinct from the language of events (which, in the human sphere, permits us to speak only of bodily movements or changes) but also largely, or wholly, independent of the latter. Thus, according to Mr. Strawson, not only do accounts of human actions involve a systematically different “vocabulary” and call for a wholly different “dimension of explanation” from those in which (as he puts it) we may discuss “particular bits” of human behavior in terms of “moving parts,” but indeed there are no “effective correlations” whatever between the two. Hence, as I understand him, physical determinism, according to which all human as well as non-human behavior is explainable in terms of purely “physical laws” (whatever they come to) can never logically “impinge” upon any question about action and choice and therefore free will, since such questions are statable entirely in terms of the language of action (or as he calls it “psychology”). But no careful examination or account of either the vocabularies or the “dimensions” of explanation in question is offered. In fact we are left completely in the dark as to the respective, and presumably divergent, functions of the two forms of expression and explanation. Granted that “psychological” explanations have nothing to do with “physical” ones what difference would it make to our lives and to the conduct thereof if we simply dispensed with psychological chat altogether? To this pressing question Mr. Strawson offers not the vestige of an answer.
On the other side, Mr. Thompson contends that if we assume “unlimited success in physical explanations of physical movements,” then it is “by no means clear . . . that people in general would . . . find [the point about absence of correlations between descriptions of actions and descriptions of physical movements] an adequate defense against [physical] determinism.” From his point of view, moreover, only scientific investigations could show whether there are “effective correlations” between what is asserted in “psychological terms” about human actions and what is said in “physical terms” about the movements or changes of human organisms. Unhappily, Mr. Thompson, in his turn, offers no clue as to the way in which such an investigation might proceed, so that his contention, which Mr. Strawson finds palpably absurd, can, at this stage, scarcely even be called a “program.” Nor does he offer any evidence in support of his speculation that “people in general” would take Strawson’s view about a total absence of correlations not only to be an inadequate defense against determinism but in fact a virtual “surrender” to it.
Like Mr. Strawson and unlike Mr. Thompson, I do not think that “correlations” made between statements about actions and statements about purely physical processes can be accounted for exclusively by the procedures of physical science. And this for the reason that it is no part of the function of physical science, as such, to decide what is to be done, to specify or ascribe intentions, to assign responsibilities, and the rest. But like Mr. Thompson and unlike Mr. Strawson, I find it quite fantastic to suppose that there can be no correlations, positive or negative, between statements made in the two “languages” (this we should remind ourselves, is just a manner of speaking; they are in reality merely distinguishable functions of the one familiar language all of us habitually use in the conduct of our affairs) . What would be the consequence for our conduct itself if our thinking about human actions were disengaged from any possible reference to the physical movements or changes that often serve as essential evidences for the occurrence (or non-occurrence) of actions of a certain sort? Imagine a killing attended by no physical changes in the body of the victim! And then imagine the effects upon our moral and legal practices if such changes were always declared irrelevant in attempts to establish the guilt (or innocence) of putative killers!
One does not have to be a materialist, whether metaphysical or analytical, in order to acknowledge that our actions out and are meant to out, a swath in the material world. To hack and to saw, to build and to mold, to form and to fashion, to do and to make something admittedly involve more than the modification or transformation of the sheerly physical aspects of things; but they also, inevitably, involve that too. Mr. Strawson strikes me here as the linguistic counterpart of a metaphysical dualist whose logical bifurcation, if resolutely adhered to, would result in practical futility. As the old Greek philosophers more wisely saw, realization of even the most ethereal of aims or “final causes” involves some assignable “material cause” as well. The most impalpable gesture requires the movement of a muscle, and, as Spinoza long ago perceived, a completely immaterial God would, by that very fact, prove to be inconsequential and inactive. For the linguistic philosopher, interested in the logical issues of determinism, the salient, arguable questions are not whether the language of action and that of physics represent two irreducibly different ways of thinking about the human condition, but what assignable roles these languages have in the customary movements of human thought and conduct; not whether these roles are in principle correctable, but how they are in practice correlated and to what assignable ends. To these questions, I am sorry to report, none of the present contributors makes a trace of an answer.
In his concluding “Postscript,” by some margin the best piece in the book, Mr. Williams briefly considers some of the bearings of psychoanalytic theory, and the forms of “determinism” commonly thought to attend it, upon the problems involved in the free-will controversy. He also mentions, without comment, the highly important fact that many contemporary theories of behavioral and experimental psychology (he might also have included many theories in the so-called social sciences) “seem to aspire to a physicalist basis, and hence not to be very ultimately psychological [sic] . . . .” Unfortunately he does not consider the effects of such theories upon our work-a-day understandings of such concepts as cognition, motivation, learning, emotion, and mind itself. Whether through explicit physicalist redefinitions of such terms or, more informally, by virtue simply of more and more extensive and resolute use of methods of analysis and verification similar to those employed in the physical sciences, the so-called “behavioral sciences” are in process of acquiring conceptual schemes that at least appear analogous to those prevalent in the former disciplines. Particularly in England, where psychology and “social studies” are only grudgingly given the title of “science,” the success of such schemes is arguable. But there can be little doubt that our going psychological concepts, first pulled out of shape by psychologists, psychiatrists, and social scientists, and then hauled and mauled by novelists, critics, and historians, now form a kind of linguistic no man’s land where it is next to impossible any longer to be sure of one’s logical bearings.
But it is not only the behavioral and semi-behavioral sciences of man that play hob with the language of mind and action. So also do the unprecedented products of modern technology, including computers, automation, and missiles; and so, too, do the incalculable effects, human and otherwise, of modern government and industry, modern war and war-prevention, modern art, and, not least, modern ideology, totalitarian and otherwise. Who, any longer, really knows what to say about our modern human and sub-human “doings”? And what are the “proper” words in which to say it? I suspect that one of the primary reasons for the increasing strangeness of so much contemporary imaginative literature and drama, social criticism, and existentialist philosophy and theology is precisely the fact that the familiar, fine, secure old “world” of human agents and actions, obligations and responsibilities, intentions and principles, is simply falling apart. Quite literally, we hardly know what to say and think about one another and so how to take one another; likewise, we hardly know what to think about ourselves and so what to make of or do with the primordial philosophical dictum “know thyself.” What the devil does it mean?
These points have been highlighted with particular poignancy in recent months by the well-nigh intolerably perplexed and ramified discussions of the Eichmann trial and the “problematic” literature that has grown up around it. As these discussions, like those concerning “the bomb,” illustrate, even the merely linguistic aspects of the free-will problem have become deeply, inescapably “moral” and “metaphysical.” Or I at least, in order to humanize them, have an uncontrollable impulse so to call and to dignify them. But of course the knife cuts both ways. For until we can decide how we ought, or want, to talk and think about the Eichmanns (both dead and alive), their leaders, their victims, collaborators, accusers, and judges—and until, therefore, we are able to decide whether it any longer means anything to us to speak to each other about such queerish entities as their choices, actions, intentions, and guilt—we will not know what our “sentiments” about them and their neo-gothic doings really are. Either way, we must decide. And either way, both ways, the problem concerns the emerging image or, if I may use an overworked but still serviceable word, archetype, of the talking animal himself.
One thing at least is certain: we can never return to the halcyon philosophical days of David Hume, not so much because he was simple-minded, but because we no longer speak the same language. What language, then, do we speak? When we have answered this question, or rather when we know at last what sort of question it is, we may find that the free-will problem is a bit more manageable than it now seems. Meanwhile, we philosophers must ask what happens, in such a changing linguistic world as ours, to a philosophy whose only wisdom, even about language itself, is the dictum: “Ordinary language is correct language; cleave to it and forget your metaphysical agonies.” For ordinary language, or at least the vital parts of it that pertain to the life of the mind and of conduct, is increasingly an enigma, a conundrum, and an oracle.
1 St. Martin's Press, 136 pp., $3.75.