CHISHOLM AND "THE METAPHYSICAL PROBLEM OF HUMAN FREEDOM"

ARNOLD LEVISON 1.

Approaching the problem

Probably the majority of contemporary analytical philosophers would agree with Donald Davidson's recent assessment of the traditional problem of freedom and determinism. According to Davidson, "Hobbes, Locke, Hume, Moore, Schlick, Ayer, Stevenson, and a host of others have done what can be done, or ought ever to have been needed, to remove the confusions that can make determinism seem to oppose freedom.''l But if the majority of philosophers would agree with Davidson's assessment, Roderick M. Chisholm is an important exception. Chisholm continues to defend the view that there is a genuine conflict between the belief that human beings are sometimes free agents who are morally responsible for their actions, and "a deterministic view of human actions." Indeed, Chisholm sees this conflict as but a segment of a more general puzzle, which he has called "the metaphysical problem of human freedom.''2 My aim in this paper is to discuss the nature of the latter problem, as Chisholm sees it, and the structure of the reasoning whereby he hopes to solve it. Since this is a problem with enormous ramifications, and since Chisholm has written about it extensively, I cannot hope to cover his views adequately in a short paper. Therefore, I shall not attempt to delve into Chisholm's analyses of the relevant concepts employed in his theory of action. What I wish to do is discover the structure of Chisholm's reasoning, i.e. distinguish premises from conclusion, consider available evidence for or against the premises, and try to assess the validity of the inferences. I believe, however, that I will be able to show some areas where Chisholm's theory needs further detailed work if we are to be in a position eventually to evaluate the acceptability of his conclusions about this problem. Students of Chisholm's published works will have noticed that his views about the solvability of the problem of freedom and determinism have undergone an important development, in at least one important respect, since his first published paper on this subject: 537

ARNOLD LEVISON "Responsibility and Avoidability.''a In this brief but impressive treatment, Chisholm expressed scepticism about the possibility of f'mding a satisfactory solution to the problem which would be based on the premise that human beings are sometimes (at least) morally responsible for their actions. If we accept this premise, as ChishoLm does, of course, the problem becomes that of reconciling it with our belief in the so-called causal principle as extended to human actions. The apparent conflict of principles which is in question was well stated by Thomas Reid in the 18th-Century, in his Essays on the Active Powers o f Man. 4 According to Reid, on the one hand: We have, by our constitution, a natural conviction or belief that we act freely. A conviction so early, so universal, and so necessary in most of our rational operations, that it must be the result of our c o n s t i t u t i o n . . . And on the other hand: the notion of causes, and the belief that every event must have a cause which had power to produce it, is found in every human mind so firmly established, that it cannot be rooted out. 9

But these two beliefs, or "natural convictions;' each of which seems self~vident, appear to conflict, as in examples such as the following: If a man trusted by the government with a secret, which it is high treason to disclose, be prevailed upon by a bribe, we have no mercy for him, and hardly allow the greatest bribe to be any alleviation of his crime. But, on the other hand, if the secret be extorted by the rack, or by the dread of present death, we pity him more than we blame him, and would think it severe and unequitable to condemn him as a traitor. What is the reason that all men agree in condemning this man as a traitor in the f'wst case, and in the last,either exculpate him, or think his fault greatly alleviated? And Reid goes on to conclude: If he acted necessarily in both cases, compelled by an irresistible motive, I can see no reason why we should not pass the same judgment on both. Reid does not think that all motives are "compelling" or "irresistible." Reid refers to "the torment of the rack" or "the dread of present death" as "violent motives," while "the love of money, and 538

METAPHYSICALPROBLEM OF FREEDOM of what is called a man's interest," is a "cool motive." He thinks our willingness to excuse the torture victim but not the corrupt official depends on our belief that the motive of the torture victim is more difficult for an ordinary person to "resist" than the motive of the corrupt official. Thus the motive of the torture victim may be said to "compel" the victim to betray the secret; hence, to the degree that he is compelled, to that degree we do not hold him responsible. But the corrupt official would not ordinarily be said to be "compelled" by his motive of "self-interest"; we believe it was in his power to refuse the bribe even though taking it was in his "interest." Thus we do hold the corrupt official responsible for his action (accepting the bribe). Now, by "acting necessarily" Reid does not mean "being compelled to act (e.g. by an "irresistible motive')," - he means "being caused to act," and he means to contrast this with what he takes to be its opposite, viz., "acting freely." If a person "acts freely," for Reid, then he has exercised a power he has "over the determinations of his will": I consider the determination of the will as an effect. This effect must have a cause which had power to produce it; and the cause must be either the person himself, whose will it is,or

some other being. The first is as easily conceived as the last. If the person was the cause of that determination of his own will, he was free in that action, and it is justly imputed to him, whether it be good or bad. But, if another being was the cause of this determination.., then the determination is the act and deed o f that being, and is solely imputable to him. In this passage Reid clearly says that "the person" may be a "cause," and he connects this possibility with the person being "free" and "responsible." It is part o f Reid's theory that when a person causes something, what he causes is "the determination of his own will." This consideration provides the basis of Reid's theory of freedom: By the liberty of a moral agent, I understand, a power over the determinations of his own will. If, in any action, he had power to will what he did, or not to will it, in that action he is free. But if, in every voluntary action, the determination of his will be the necessary consequence of something involuntary in the state of his mind, or of some.thing in his external circumstances, he is not free;he had not what I call the liberty of a moral agent, but is subject to necessity.

539

ARNOLD LEVISON Since Reid speaks of "the determination of the will" as an "effect" of which something must be the cause, his introduction of "the will" creates no logical difficulties of the sort connected with traditional doctrines of "the freedom of the will." It is not the will which is free, on Reid's theory, but the person, the moral agent. Reid's point is that if the person is caused, whether to will or act as he did, then he was not free in that willing or acting. But whether the person is free or not free, his will is caused. If he is acting freely, then it is he that causes his will or his bodily movements or omissions; but if there is something else that causes him or something involuntary in the state of his mind that causes his will, then he is not acting freely. According to the theory that Reid rejects, which he calls "the supposition of necessity" and which we call "determinism," there is a causally sufficient condition for every human action. A free or voluntary action is one that is caused by a person's own desires and beliefs, by his own motives, character or temperament. An unfree action, on the other hand, is one that is compelled. Thus a determinist seeks to distinguish between free and unfree actions by distinguishing between kinds o f causes of actions. Some kinds of causes are excusing, according to him, while others are not. A difficulty for the determinist is to arrive at a principle which will enable him to distinguish systematically between "excusing" and "nonexcusing" causes of actions. For example, a kleptomaniac is said to act under a compulsion when he steals, even though the cause of his action is internal and psychological. One might even be justified in holding that his stealing is "voluntary" in the sense that it is caused by his desires and beliefs. So the determinist must augment his theory by adding that in order for an action to be free or voluntary, it must be caused by a person's desires and beliefs in a rational framework, s In the case of a kleptomaniac, his desire to steal is not subject to change in accordance with ordinary or normal methods of altering motivations. Thus rational considerations of self-interest do not affect the conduct of a kleptomaniac, whereas properly administered drugs along with behavioral modification techniques, or perhaps some form of psychotherapy, would affect his conduct. On the other hand, considerations of self-interest or other rational considerations, apart from any use of drugs or psychotherapy, might prevent an ordinary criminal, such as a politician who sometimes steals for profit, from taking a bribe in a particular situation. Perhaps it is true that neither the corrupt official who did take a bribe, nor the kleptomaniac, could have done otherwise in the very circumstances in which they acted. Even so, the determinist might 540

METAPHYSICALPROBLEMOF FREEDOM say, the corrupt official is properly "condemned" for his action, while the kleptomaniac is not. The former is a fit subject for moral blame and punishment because his motive - which is the cause of his taking the bribe - belongs to a class of psychological events than can be prevented from issuing in behavior by the application of techniques belonging to a certain class, namely those which we denominate "rational considerations;' and which, in the general case, can be applied by an individual to himself. Similar considerations would enable us to distinguish the degree of responsibility of the torture victim from that of the corrupt official. "The dread of present death," or "the torment of the r a c k " as Reid calls them, are compelling motives, in the sense that they belong to a class of causes of behavior that people, on the average, would not resist even if they had strong motives for doing so. But Reid wants to say that if it is always true that there is a causally sufficient condition for human actions, and true that no man ever could have done anything different from what he did do, then no man would ever be free, even if the distinction between acting voluntarily and acting involuntarily can be preserved on the determinist's account. For Reid, the fact that an action is caused by a person's desires and beliefs in a rational framework, is not sufficient for its being free. If a man is caused to act, he is "acting necessarily." It does not matter what causes him to act, nor whether those causes are internal or external, psychological or physical, compeRing or otherwise. If acting voluntarily is being caused to act by one's own desires and beliefs, then acting voluntarily is a mode of acting necessarily. The person, and not merely his states of mind, must play a causal role in bringing about his behavior, if his behavior is a free action. Furthermore, "acting freely" and not merely "acting voluntarily" in the determinist's sense, is a necessary condition of moral responsibility. Chisholm's theory is similar to Reid's, except that Chisholm claims, not only that freedom is a property of the person rather than of the will, but also that there is no "act of wal" involved in his account: 6 Our assumptions do imply that whenever a man does something freely then there is something he is thereby endeavoring or undertaking to do. But undertaking.., does not imply undertaking to undertake. I assume that the person makes certain events happen and hence that agents as well as events may be causes. 541

ARNOLD LEVISON With the understanding, therefore, that neither Reid nor Chisholm regards himself as committed to a theory of "the freedom of the will," and that Chisholm at least does not regard himself as committed to a theory that willing (or undertaking) is an "act" such that one might will to will, or undertake to undertake, we may refer to their kind of theory as "Libertarianism" (or as species of "Libertarianism"). According to our argument as developed so far, there are two assumptions which distinguish Libertarianism: (1)an agent is not morally responsible for an action unless he could have done otherwise, and (2)the formula "he could have done otherwise" implies that there is no causally sufficient condition for his acting as he did. A determinist may challenge either one of these assumptions. We have just discussed how a determinist might challenge (1). He might challenge the second assumption by adopting Donald Davidson's claim, that it is only "confusion" that makes determinism seem to oppose freedom. Davidson alludes to G.E. Moore's well-known attempt to show that moral responsibility is compatible with determinism, by analysing "he could have done otherwise" into "he would have done otherwise i f he had chosen otherwise. "7 Moore's idea was this. Accept the claim that a man is morally responsible for what he does only if, in the circumstances in which he acts, he could have done otherwise. This claim would be compatible with the idea that his acting as he did was nevertheless caused, if "could have done otherwise" meant simply: '~,ould have done otherwise if he had chosen otherwise." But as Chisholm points out, the phrase, ''he would have done otherwise if he had chosen otherwise" could be true, even when it is false that he could have chosen otherwise,s If a person chose to act in a certain way, and acted in that way as a result of that choice, then if he could not have chosen otherwise, he could not have done otherwise. If the choice was not avoidable by the man, then the act, which by hypothesis resulted from that choice, was not avoidable by him either. Since, on the determinist's view, choosing is itself an event, Moore's analysis does not succeed in showing how determinism - the theory that every event is caused by other events - might be compatible with saying of a man who made a certain choice and acted as a result of that choice, that he could have acted differently. Now in the "Responsibility and Avoidability" paper which I referred to earlier, Chisholm had already made clear the grounds of his rejection of the kind of view advocated by Davidson. Being unwilling to accept either an "indeterministic" or "capricious" view 542

METAPHYSICAL PROBLEM OF FREEDOM

of free action, or the fatalistic conclusion that we can't avoid doing what we do, Chisholm proceeded to consider the possibility of constructing "a metaphysical theory about the self" - a theory that would employ terms "designating something to which we apparently need not refer when expressing the conclusions of physics and the natural sciences. ''9 "Active powers," "prime mover;' and "higher level of causality" (than event-causality) are examples of the kind of metaphysical terms Chisholm had in mind. But in reflecting on the possibility of there being such a theory, Chisholm wrote: 10 I believe we cannot know whether such theories enable us to escape our dilemma. For it seems impossible to conceive what the relation is that, according to these theories, holds between the "wig;' "self," "mover" or "active power;' on the one hand, and the bodily events this power is supposed to control, on the other - the relation between the "activities" of the self and the events described by physics. Thus Chisholm, at the time this paper was written, was unable to fred a solution to the over-all problem that satisfied him. However, in his next published treatznent of the subject, viz. his "Freedom and Action" paper, Chisholm's doubts seem to have been resolved. 11 At least, he does not mention his previous doubts concerning "the conceivability" of the relation between "the activities of the self," as required by the metaphysical theory, and the physical events of the internal workings of the body, as described by physics and neurophysiology. Instead, he begins by stating what he calls "the metaphysical problem of human freedom;' posed in terms of the following dilemma: 12 Human beings are responsible agents, but this fact appears to conflict with a deterministic view of human action (the view that every event that is involved in an act is caused by some other event); and it also appears to conflict with an indeterministic view of human action (the view that the act, or some event essential to the act, is not caused at all). He then proceeds to state flatly that in order to solve the problem, "we must make somewhat far-reaching assumptions about the self of the agent - about the man who performs the act." These considerations give rise to the question: what view of the relation between the self of the agent and events in the agent's body (the "events described by physics") did Chisholm arrive at in the years between the publication of the two papers, "Responsibility 543

ARNOLD LEVISON and Avoidability," and "Freedom and Action," that enabled him to regard the "conceivability" problem as solved? Furthermore, what are "the far.reaching assumptions about the self of the agent" that he is referring to here? In order to answer these questions, I shall attempt to reconstruct Chisholm's reasoning in a step-by-step fashion, beginning with his formulation of the dilemma of freedom and determinism. 2.

The structure of Chisholm's argument for freedom In formulating the "metaphysical problem of human freedom," Chisholm asks us, in effect, to consider the following three propositions, their metaphysical import, and their logical relations: (F) Human beings are responsible agents. (D) Every event that is involved in an act is caused by some other event. (I) The act, or some event essential to it,is not caused at all. Chisholm then proceeds to construct an argument which may be presented schematically as follows, using abbreviations as above and stating each controversial premise separately: (1)

(F).

Premise

(2)

Nec (if (F) then noteD)).

Premise

(3)

Nec (if (F) then hOt,I)).

Premise

(4)

Therefore, both not-(D) and not-(I).

From (1)-(3)

In other words, according to Chisholm, determinism and indeterminism must both be false. But as philosophers, usually discuss these views, a deterministic view o f human actions is assumed to be the contradictory opposite of an indeterministic view, i.e. such that at least one of the two views must be true although both cannot be true. On this customary account, if (F) were to conflict with both (D) and (I), as Chisholm claims, then since (D) and (I) could not both be false, it would follow that (F)must be false! But Chisholm refers to (F) as a "fact," as something he "feels certain" is true. 13 What, then, is the nature of the logical relation between (D) and (I), as Chisholm conceives it? We may illustrate the nature of this relation by means of an analogy. Corresponding to (D), we might say, with respect to some human being, that every ancestor of his had a human ancestor. And corresponding to (I), we might say, of that human being, that some ancestor of his had no ancestor at all. For example, some unthinking 544

METAPHYSICALPROBLEM OF FREEDOM people might well suppose that every ancestor of theirs had a human ancestor. On the other hand, the Bible tells us that Adam, the first man according to biblical accounts, had no ancestor at all (although he had a Maker). Clearly, both these propositions couldn't be true; it couldn't be true both that Adam, an ancestor of ours,had no ances. tor at all, and that every ancestor of ours, including Adam, had a human ancestor. But these propositions could both be false. Thus, according to Darwin's Origin o f Species, it is false both that there is some human being every ancestor of whom has a human ancestor, and it is also false that there is any human being some ancestor of whom had no ancestor at all. So Adam, if he existed,had an ancestor who wasn't exactly human, and the same is true of the rest of us. Now apply this analogy to the question of how we shall conceive the relation between (D) and (I), replacing "human beings" by "responsible actions," "ancestor" by "cause," and "human ancestor" by "event-causality." Thus if it were true of some human action, that: (i)

every event involved in that act is caused,

(ii)

there is an event involved in the act that is caused by something not itself an event,

and

then (D) would be false. For (D) says that every event involved in an act is caused by some other event. But this very condition, which, if it were satisfied, would falsify (D), would also falsify (I), since (I) says that there is some action, or some event essential to that action, which is not caused at all. This would be like saying that Adam, the biblical f'Lrst man, had no ancestor at all, not even a non-human one, - which would be very different from saying that he had some ancestor who wasn't exactly (or completely) human. (Of course, someone who accepted the biblical account, and rejected the theory of evolution as applied to the origin of man, should be prepared to say just this.) To sum up, if there is an action involving events at least one of which is caused by something not itself an event, and ff every event involved in that action or essential to it is caused by something or other, then both (D) and (I) would be false. brings us to the next step in Chisholrn's reasoning. If (I) is false, then the act and every event essential to it is caused, and if (D) is false, there is an event involved in the act the cause of which is not 545

ARNOLD LEVISON any other event. But if the cause of this event is not another event, what can it be? According to Chisholm, "the possibility that remains" is: 14 (A)

What causes the event in question is the agent - the human being.

The human being is not an event, nor, according to Chisholm, a logical construction of events or similar entities, such as temporal stages of a body. The human being, or agent, is a self or person, and is . . . persons - such entities as ourselves - are substantival concrete things, in the strictest sense of the term "thing," that persist through time, in the strictest sense of the expression "persist through time." Furthermore: . . . the fact that we are - occasionally at least - morally responsible for what we do, implies that we are causes, that persons are c a u s e s . . . But what kind of causation is this? Clearly, Chisholm is not saying that a person's existence is a sufficient condition for his responsible actions. On the contrary, he denies that a man's responsible acts have any sufficient causal condition. A question we must consider is this. Is (A) intended as a logical consequence o f preceding steps and premises, or is it an independent premise, a new hypothesis, in the argument7 What appears to follow in steps (1) through (4) of the argument is that something not itself an event (and not reducible to events, sets of events, etc.), must be capable of causing at least one event in any action for which an agent is responsible. This "something" could not be an event, or logical construction of events, for the following reason. If the world consisted ultimately of events, or if only events might be terms in causal relations, then Chisholm's (D) and (I) would be contradictories, and not more contraries. For on either of these assumptions, whatever is a cause /s an event. In that case, either (D) or (I) necessarily would be true, and step (4) of Chisholm's argument (that (D) and (I) are both false) would itself be false. It would then follow that Chisholm's (F) would be false. Now Chisholm describes his theory as an attempt to construct a '~/a reed/a, a middle way, between a deterministic and an indeterministic view o f human actions. 1~ Thus it is crucial for Cldsholm's 546

METAPHYSICAL PROBLEM OF FREEDOM

argument that these views of human action, namely determinism and indeterminism, be defined or characterized in such a way that it is logically possible for them both to be false. But as we have just seen, (D) and (I) may both be false, and a via media between determinism and indeterminism possible, only if there are kinds o f things in the world that are not events and not reducible to events, but are such that the things of those kinds may be causes of events. Such things are persons, selves, agents, or human beings. That is why Chisholm remarks that in order to solve the problem posed by the conflict of human freedom with determinism - the "metaphysical problem of human freedom" - we must make somewhat far-reaching assumptions about the self of the agent - about the man who performs the act." For it is to the self of the agent that Chisholm assigns the crucial metaphysical role of being able to cause events without being identical with any event (or logical construction of events). In view of these considerations, it seems best to regard (A), not as a premise of the argument, but as a consequence of a logically prior principle, namely, that everything that is a cause is either an event or an agent, an "agent" being taken to be a "substantial self" which maintains its identity over a period of time or through various changes in events. The argument also assumes that there are two irreducibly distinct kinds of causality, namely event.causality and agent.causality. (Chisholm sometimes refers to agent.causality, after C.D. Broad, as "nonrecurrent causality," in order to differentiate it from event-causality.) Another consideration is this. Nothing in the argument as so far developed excludes the possibility that the agent, in causing the event in question, is himself caused to cause it, either by antecedent events or by other agents, in accordance with laws o f nature. In that case, the argument would not have succeeded in avoiding determ i n i ~ . So, in order to insure that what is agent.caused is a free action, we have to add a premise to the effect that nothing, or no one, caused the agent to cause the events that bring about his free actions. It appears, then, that Chisholm's argument for metaphysical freedom involves the following structure overall (recapitulating the first four steps):

(1) (2)

(F). Nec (if (F) then not-(D)).

Prem~e Premise 547

ARNOLD LEVISON

(3) Nec (if (F) the not-(I)). Premise (4) Therefore, both not-(D)and not-(I). From (1)-(3) Let R be a free and responsible act of an agent S: (5) Then there is an event (one or more) involved in R that is not caused by any event. From (4) *(6) Every event is caused either by another event or else by an agent. Premise (7) Therefore, the agent causes the event(s) in question. From (1)-(6) *(8) Nothing (or no one) causes the agent to cause the event(s) in question. Premise (9) Therefore, in causing an event, the agent initiates a causal sequence of events (in his body). From (1)-(8) I have flagged steps (6) and (8) of the argument in order to indicate that they are premises specially introduced and necessary to derive the conclusion. Step (9), the conclusion of the argument, represents Chisholm's doctrine - his proposal for solving "the metaphysical problem of human freedom." 3.

Some objections to Chisholm's reasoning There are a number of different kinds of objections that might be made to Chisholm's train of reasoning. One of the more obvious ones is that the introduction of additional premises, as in steps (6) and (8), makes the argument begin to look suspiciously ad hoc, as if Chisholm just keeps adding premises to solve difficulties, in the absence of any independent evidence to justify the additional premises. But Chisholm seems to be saying that these assumptions are necessary; we cannot explain the possibility of moral freedom without making them. Let us then consider objections based on considerations of coherence. We can imagine someone who, reflecting on Chisholm's argument, might ~ tempted to reason as follows. "Determinism;' this person might say, is just the view that every human action is caused by something or other, in an unending and continuous chain. It is not essential to determinism that the cause of an event be events only. Furthermore, he might say, "indeterminism" is just the view that something involved in the origination of an action isn't caused by anything at all. It is not essential to indeterminism that what is 548

METAPHYSICALPROBLEM OF FREEDOM uncaused be an event (as opposed, say, to the activity o f a person, or simply a person). Now Chisholm appears to be committed to both "determinism" and "indeterminism" as so characterized. He only seems to have found a via media between thorn because he provides tendentious definitions of the terms. Chisholm is committed to both views because, according to him, the act, or every event involved in the act, is caused by something or other (either by another event or by the agent), and this is a "deterministic" view o f human actions; and because, according to Chisholm, the agent, in originating the action by causing some event involved in or essential to it, is not himself caused to make this event happen by anything or anyone, and this is an "indeterministic" view of human agency (if not of human actions considered as large-scale events). Therefore, our objector concludes, Chisholm has not succeeded in finding a via media, a middle way, between the horns o f a deterministic and an indeterministic view of human action. On the contrary, what seems to follow is that he is impaled on both horns! In his defense, Chisholm could say that he could accept this definition of "indeterminism;' and accept the label as so defined. He could then point out that it is consistent with this version of "indeterminism" to say that the agent's action is caused, and caused by the agent, i.e. that every event involved in his action is caused. It follows then that if the agent caused the action,by initiating a chain of events that comprise the action, then even if nothing caused him to cause those events, still it was he who was responsible for their occurrence. He made them exist. So on the proposed corrected definition of "indeterminism" the latter would not be incompatible with the belief that human beings are responsible agents. So this objection to Chisholm's reasoning is really not conclusive. A similar but more difficult objection has been stated by C.D. Broad. It concerns how we are to understand what it is for an agent to choose, or what it is for an agent to make a certain event happen. Let us assume that there is an event which is, in Broad's terms, "literally determined by the agent or self, considered as a substance or continuant, and not by a total cause which contains as factors events in and disposition of the agent." Broad goes on to say: 17 . . . it is surely quite evident that, if the beginning o f a certain process at a certain time is determined at all, its total cause must contain as an essential factor another event or process which enters into the moment from which the determined event or process issues. I see no prima facie objection to there 549

ARNOLD LEVISON being events that are not completely determined. But, in so far as an event is determined, an essenUal factor in its total cause must be other events. How could an event possibly be determined to happen at a certain date if its total cause contained no factor to which the notion of date has any application? And how can the notion of date have any application to anything that is not an event? Broad concludes that the assumption of agent-causality as a mode of causation irreducibly distinct from event-causation is "self-evidently impossible." Now Chisholm could simply accept what Broad says here by denying that the agent is himself "the total cause" of the event in question. The total cause might include desires and beliefs of the agent, which are dated occurrences in the sense required by Broad (or at least can be construed as such). All that Chisholm needs to insist upon is that the agent's desires, beliefs, etc., are not causally sufficient for the action, if it is a free action. Thus the agent must contribute causally to the outcome; he may be "inclined" to act on his beliefs and desires, but he may not be "necessitated" by them. Still, the problem remains of the conceivability of the relation that is supposed to exist, on Chisholm's theory, between the agent and the brain event that is the first occurrence in the causal sequence leading to the bodily movements (or their omission). Does the agent really have a causal role to play in his total physiological system? In a recent issue of the Journal of Philosophy, Harry G. Frankfurt remarks: ~8 Chisholm offers no reason for believing that there is a cliscernible difference between the experience of a man who mirac. ulously initiates a series of causes when he moves his hand and a man who moves his hand without any such breach of the normal causal sequence. And Irving Thalberg has scathingly characterized Chisholrn's theory on this point as "a priori neurophysiology.''~9 These objecUons have a point so far as Chisholm's theory implies that a free action necessarily is one that has a certain causal history, namely one leading back to a brain event made to happen by an agent. This is reminiscent of the old Cartesian conception of a free action as the upshot of a spirit's influence over an automaton, the automaton being the human body and the spirit being the self, mind or soul. It seems that in Chisholm's theory the role of the agent is 550

METAPHYSICALPROBLEM OF FREEDOM simply to push the button that starts the mechanism, but how the agent accomplishes this remains mysterious. 2~ 4.

An objection to Chisholm's premises Chisholm takes it to be a fact that human beings are responsible agents (F). But is (F) a fact, or, put more circumspectly, are we "certain," as Chisholm professes to be, that (F) is true? In general, we cannot be certain that a proposition is true, if we are certain that it entails another proposition which is logically incompatible with one for which we have considerable independent evidence. Now, according to Chisholm, (F), interpreted as the view that we are responsible agents, and not just as the view that we believe we are responsible, entails that (D) - a deterministic view of human actions - is false. But if Chisholm is right about this controversial claim, then many philosophers might want to deny that (F) is a fact, or at least, to deny that we are "certain" that (F) is true. For they would point out that we have considerable independent evidence that (D) is true. Such evidence, it might be said, is provided by psychological experimentation, neurophysiology, and sources such as Freud's Psychopathology ofF, veryday Life. I am not referring here to Freud's assumption of determinism, but simply to his empirical observations of obsessive and compulsive behavior patterns exhibited by normal persons in their everyday life. Now it may be argued that we also have considerable evidence that (F) is true, viz. the evidence of "practical reasoning" and social life, as well as the evidence of our conscious experience in choosing, deciding, and deliberating. Chisholm might say that our evidence for (D) belongs to the context of "theoretical reason," while our evidence for (F) belongs to "'practical reasoning" and our experience of ourselves as active, thinking beings. In the way Chisholm has stated the metaphysical problem of human freedom, having asserted that (F) is a fact and having rejected (D) as a consequence, he has, in effect, asserted the doctrine of the "primacy" of practical over theoretical reasoning. 21 Let us, in conclusion, try to see just what this comes to. In defense of his procedure, Chisholm argues as follows. First, as we have seen, he establishes the fairly uncontroversiai claim that (F) entails, with respect to those actions for which a human being is responsible, that he could have acted otherwise than he did act. And Chisholm goes on to argue that the proposition, "he could 551

ARNOLD LEVISON have done otherwise," is logically incompatible with the assumption that there was a causally sufficient condition for the man acting as he did. The principle of Chisholm's reasoning in this context might be stated as follows: Where S is a person or human being, and 'x' and 'y' range over events, including human actions, then: (CR) Necessarily, if x is causally sufficient for y, and S is morally responsible for y, then S is morally responsible for x. According to this principle, no one can be morally responsible for an effect if he was not responsible for its cause, and of course, if a person was unable to prevent a certain cause he could not fairly be held responsible for it (or for any of its effects). But if a deterministic view of human actions is true, then by the (CR) principle we are responsible for our choices or actions only if we are responsible for events that happened before we were born or existed. Since this is clearly absurd, then if the (CR) principle is true, and if a deterministic view of human actions is true, it follows, apparently, that no one is morally responsible for his actions. The argument may be applied to the case of choice as follows. Human beings are responsible agents, (F), only if, with respect to their responsible actions, they could have done otherwise; and therefore only if they could have chosen otherwise (assuming that responsible actions are caused by the agent's choices). But they could have chosen otherwise only if there is no set of events, C, such that C is a causally sufficient condition for their having chosen as they did. But it appears that if this conclusion is true, and if it follows from the assumption that human beings are moral agents, together with the (CR) principle, - then, contrary to Chisholm, no one can be justified in feeling "certain" that (F) is true, i.e. that human beings are responsible agents in his sense. The conclusion we should draw from Chisholm's argument is that we are not in a position to assert (F) as a "fact." For (D) as Chisholm as stated it, is not the postulate of universal determinism - a rival metaphysical principle - but simply a deterministic view of human actions, the view that every event involved in an act is caused by some other event. So stated, (D) is neither necessarily true nor necessarily false; it is an empirical claim which might be true and for which there is, as we have already noted, considerable empirical 552

METAPHYSICAL PROBLEM OF FREEDOM evidence. Given, then, that we do have considerable empirical evidence that (D) is true, then by the structure of Chisholm's argument, we have as much reason for denying (F) as we have for denying (D). Thus Chisholm's denial of (D) rather than (F) is arbitrary. Now if (D) were true, and if the (CR) principle is self-evident (and I am inclined to believe that it is), then what follows is that human beings are not responsible agents, although perhaps they might have been. What I am suggesting is that Chisholm's metaphysical theory of human freedom might describe just a possibility, but one that happens not to be exemplified by human beings in this actual world. If it is a contingency whether human beings are sometimes free or always determined, then Chisholm's formulation of the metaphysical problem of human freedom contains a serious flaw. The flaw in his assumption that it is a fact that human beings are "responsible agents," in his sense, i.e. in the sense that their free actions necessarily have a certain causal history. Thus Chisholm seems to be confronted with a dilemma. If it is a fact that human beings are responsible agents, then this fact cannot imply a consequence about which we are justified in having serious doubt. For if it did, the doubt we had about the consequence would shift to the premise. On the other hand, if this alleged fact does not have this consequence, (viz. that (D) is false), then the attempt to connect man's moral freedom with causal freedom is mistaken from the start.

UNIVERSITYOF MARYLANDBALTIMORE COUNTY BALTIMORE, MARYLAND 21228 USA

NOTES Donald Davidson, "Freedom to Act," in Essays on Freedom of Action, ed. Ted Honderich, London and Boston: Routledge & Kegan Paul, 1973, pp. 139-155. 553

ARNOLD LEVISON

Roderick M. Chisholm, "Freedom and Action," in Freedom and Determinism, ed. Keith Lehrer, New York: Random House, 1966, pp. 11-14. 3 See Determinism and Freedom, ed. Sidney Hook, New York University Press, 1958, pp. 145-7. 4 Thomas Reid, Essays on the Active Powers o f the Human Mind, The M.I.T. Press, 1969, Essay IV, Chap. VI, pp. 304ff. Subsequent citations from Reid may be found in Chap. I, pp. 259ff. s See John Hospers, "What Means This Freedom? ," in Hook, op. cit., pp. 126-142. 6 See Chisholm's "Comments on D. Davidson's 'The Logical Form of Action Sentences', in The Logic o f Decision and Action, ed. Nicholas Rescher, University o f Pittsburgh Press, 1966, pp. 113-4. 7 G.E. Moore, Ethics (London, 1912), Chapter V1. s In Lehrer, op. cit., pp. 15, 16. 9 Hook, op. cit., p. 147. l o Ibid. 11 Lehrer, op. cit. t 2 Ibid. 13 Hook, op. cit., p. 147. 14 Lehrer, op. cit., p. 17. Is Chisholm, "Reflections on Human Agency," Idealistic Studies, I, pp. 36-46 (1971). 16 Lehrer, op. cit., p. 21. t7 C.D. Broad, Ethics and the History o/Philosophy (London, 1952), pp. 195-217. I s "Freedom of the Will and the Concept of a Person," Journal o/Philosophy, LXVIII (1971), pp. 5-20. 19 Enigmas o / A g e n c y (London, 1972), p. 44. 20 "Reflections on Human Agency," p. 40. 2t Hook, op. cit., p. 147. 2

554