Action and crime: a fine-grained approach.

• Author: Goldman, Alvin Ira
• Position: Symposium: Act & Crime

There is much to admire in Michael Moore's comprehensive and resourceful treatment of the philosophy of action in the context of criminal law.(1) I am impressed by the degree to which he finds classical metaphysical issues in the theory of action to be germane to the interpretation and systematization of the criminal law. Many of his positions strike me as plausible and well-defended, including his causal-volitional approach to the nature of action. The main topic on which he is not convincing is the defense of the coarse-grained view of act-token individuation, or what he calls the "exclusivity thesis."(2)

According to this thesis, all complex act-tokens are identical to basic act-tokens, which are bodily movements. For example, the complex act-token of Smith's killing Jones is identical to the basic act-token of Smith's curling his finger (thereby firing a gun at Jones). It is granted, of course, that the act-type killing is different from the act-type curling one's finger. But a particular act of killing is always identical to some particular bodily movement, according to the coarse-grained theory.

This contrasts with the fine-grained theory, which maintains that the above-mentioned act-tokens of Smith are two different, though intimately related, actions.(3) Although the coarse-grained view is a widely respected contender in the theory of action and is imaginatively applied by Moore to the legal context, it suffers from a number of serious problems. Some of these Moore does not address, and others are more damaging than he appreciates. I shall present a case for the superiority of the (extreme) fine-grained view. In addition to its purely metaphysical virtues, this approach is helpful in the criminal law, especially for handling the notion of an offense.

Six types of problems for the coarse-grained view have been presented in the literature:(4) (1) the problem of effects, (2) the problem of causes, (3) the problem of modifiers, (4) the problem of the by-relation, (5) the time problem, and (6) the spatial location problem. Moore pays a good bit of attention to problems (5) and (6), and he tries to show how they can be handled from a coarse-grained perspective, but he gives little or no attention to the first four problems. I shall exposit and discuss these problems in order.

All six problems have the same general form. Whereas the coarse-grained view is committed to the identity of many "pairs" of act-tokens, there often appear to be properties that one member of a pair possesses and the other member lacks. But if the acts differ on any property, they cannot be identical, according to the Indiscernibility of Identicals principle.(5) Thus, unless the appearance of property disparity can be explained away, the identity claim implied by the coarse-grained view, and hence that view itself, is unsustainable.

1. THE PROBLEM OF EFFECTS

   The first type of problem, the problem of effects, starts with the assumption that acts have causal properties, such as producing certain effects, results, or consequences. Suppose Ned plays the piano, and his playing causes Dolly to fall asleep and Molly to wake up. In virtue of these two effects, Ned can be credited with the complex act of putting Dolly to sleep and the complex act of waking Molly up. It is agreed, then, that there are the following act-tokens: Ned's playing the piano (act A), Ned's putting Dolly to sleep (act A'), and Ned's waking up Molly (act A"). The coarse-grained view holds that these acts are one and the same: they are all identical to certain basic acts of Ned, viz., his moving his hands and fingers at a certain time and place, and hence identical to one another. But can these identities be sustained? By hypothesis, Ned's playing the piano (act A) has the property of causing Molly to wake up. But his putting Dolly to sleep (A') appears to lack this property: it seems false to say that Ned's putting Dolly to sleep causes Molly to wake up, since this wrongly suggests that Dolly's falling asleep is somehow causally implicated in Molly's awakening. That is not the intended scenario. Dolly and Molly are in separate rooms, and their states are unaffected by one another. Analogously, although act A has the property of causing Dolly to fall asleep, act A" appears to lack this property. Hence, A is not identical to either A' or A", contrary to the coarse-grained view.(6)

   Moore gives brief attention, though only in a footnote, to this dilemma.(7) His remarks occur in the context of a larger discussion of one of Judith Thomson's temporal-order problems.(8) Thomson points out the oddity of saying "A's killing of B caused B to die several hours later." Moore insists that the killing does occur before the dying, and he explains the oddity of the quoted sentence as a merely pragmatic feature of the usage. He observes that "[i]t is like saying that the most discussed event of the decade (referring to a large flood) caused more discussions to take place about it over the next ten years than about any other event."(9) When a predicate repeats information already conveyed in a sentence's subject, as in this example, pragmatic oddity results. Moore proposes to diagnose my causal-effect cases in the same way. He acknowledges the oddity in saying that a defendant's killing of a victim caused the defendant's gun to fire, but still maintains that the killing of the victim did cause the defendant's gun to fire. Apparently he would offer the same response to the piano case: Ned's putting Dolly to sleep does have the property of causing Molly to wake up, and his awakening Molly does have the property of putting Dolly to sleep, though it sounds odd, given a full description of the circumstances, to say these things. The problem of effects, he wishes to conclude, is not an insuperable one for the coarse-grained view.

   Moore's explanation, however, simply does not work for the piano example, or for many other such examples. The descriptor "Molly's waking up" does not repeat any information contained in the subject expression "Ned's putting Dolly to sleep." So no appeal to repetition can explain the oddity in question. Of course, Moore may not restrict pragmatic oddity to repetition, but he does not identify any other source of pragmatic oddity that would cover the present case. The situation, then, is as follows: if we take the correctness of the coarse-grained theory as a premise, we can infer that Ned's putting Dolly to sleep does cause Molly's waking up, because Ned's putting Dolly to sleep is identical with his playing the piano, which causes Molly's waking up. It would then follow that the oddity of saying this is not a semantical oddity but merely some sort of pragmatic oddity. But since the correctness of the coarse-grained theory is precisely the issue in dispute, its correctness cannot be taken as a premise. Moore needs some independent explanation of the oddity in some familiar pragmatic terms. In the absence of such an explanation, the statement's falsity is the most natural account of its oddity. If the statement is false, however, as we seem forced to conclude, then there is no complete sharing of properties, and Ned's putting Dolly to sleep is not identical to his playing the piano, or to his moving his hands and fingers at the relevant time and place.

2. THE PROBLEM OF CAUSES

   The second problem arises not from the effects of acts, but from their causes. There are cases in which one member of a pair of putatively identical acts has the property of being partially caused by a specific factor F, whereas the second member lacks the property of being partially caused by F. I originally gave the following case to illustrate this difficulty.(10) John answers the phone and says "hello." He has just been quarreling with his wife and is in a tense emotional state. As a result, he unintentionally says "hello" very loudly. Consider now the following "pair" of acts: John's saying "hello" and John's saying "hello" very loudly. According to the coarse-grained theory, these acts are identical. Hence, they should share all of the same properties. But John's act of saying "hello" very loudly is partially caused by his tense emotional state, whereas John's act of simply saying "hello" is not partially caused by this emotional state. The absence of causal connection in the second case is supported by the absence of a relevant counterfactual, or "but-for," condition. It is false that if he had not been tense, he would not have said "hello." On the contrary, he would have said "hello" anyway, even if he had not been tense or upset. So the act of saying "hello" very loudly possesses a relational property that the act of saying "hello" lacks, and they cannot be identical. A similar example is given by Fred Dretske:

   My tug on the steering wheel of my car, for instance, doesn't cause

   my car to move, much less to move at 63 mph. Rather, it causes

   the 63-mph movement to be in that direction. My heavy foot is

   responsible for the speed, the dirty carburetor for the intermittent

   pauses, and the potholes in the road for the teeth-jarring vertical

   component of the movement.(11) This example is best construed as focusing on an event (rather than an action): the movement of the car. In the coarse-grained view, there is just one event here. The various descriptions "the car's moving," "the car's moving at 63 mph," and "the car's moving in that direction" all refer to one and the same event. But this is problematic for the reasons Dretske indicates. The car's moving in that direction is partly caused by my tug on the steering wheel, but this does not even partly cause the car's moving (simpliciter) or the car's moving at 63 mph. Once again, identities implied by the coarse-grained view seem to be unsustainable.

   Before turning to ways in which a coarse-grained theory might try to handle this difficulty, let us ponder the implications of accepting the nonidentities endorsed by the (extreme)...