Printer Friendly

A plea for omissions.

In an attempt to identify elements that all criminal prohibitions have in common, many theorists, most of whom have been deemed adherents of "orthodox criminal law theory," have proposed a number of formal requirements for criminal liability. One of these requirements--most often called the "act requirement'--is that there must be, among other things, an overt act on the part of the offender in order for that offender to incur criminal liability. (1) Though there is considerable disagreement over what should count as an "overt act," one strategy taken by many theorists is to define an act as a "willed bodily movement." In an attempt to resuscitate the act requirement in the face of criticism of the orthodox position, Michael S. Moore has reaffirmed a similar understanding of the act requirement that defines an act as a volition causing a bodily movement. (2) As a corollary to this understanding, Moore develops a pair of arguments--one metaphysical and the other normative--against liability for most omissions. In this essay, I argue that Moore's treatment of omissions is flawed and that its flaws stem from Moore's broader theory of action, particularly from his commitment to defining an act as a volition causing a bodily movement. In particular, I contend first that Moore's metaphysical claims about omissions are incompatible with his own normative argument against omissions liability, and second that in both his metaphysical and normative claims concerning omissions, Moore fails to recognize the significant and legitimate role that positive duties--requirements to perform specific actions--play within the criminal law. In the end, even though I agree that there is an act requirement in the criminal law, it is my position that the legitimacy of positive duties within the criminal law necessitates an understanding of the act requirement that is incompatible with volition-based accounts like the one Moore defends.

I Moore's Two Claims Against Omissions

One line of criticism used by theorists against the orthodox account of an act is that it implies that the act requirement necessitates overt actions--actual bodily movements--for criminal liability. If this is true, then the orthodox account would seem to preclude, among other things, the possibility that any kind of omission could be held criminally liable. Precluding liability for omissions is problematic primarily because there are many criminal proscriptions of omissions to be found in the present practice of Anglo-American law, very few of which are controversial. Of course, present practice may be theoretically inconsistent or even internally contradictory, but there seem to be good reasons for many of the criminal prohibitions of certain omissions. Few would deny, for example, that the protection of children is a legitimate and important reason behind criminal statutes proscribing the failure of parents to take adequate care of their children. Similarly, statutes prohibiting tax evasion help to guarantee the maintenance of infrastructure and the provision of (at least minimal) social services. As such, a theorist arguing against such prohibitions must demonstrate that laws proscribing omissions either appeal to some principle that deeply contradicts other established principles that underlie the criminal law as a whole or appeal to no principle at all, and thus should be rejected.

Moore attempts to demonstrate just such a contradiction (even though he does concede that criminal prohibitions against failing to take adequate care of one's children constitute justifiable exceptions to his view). (3) Moore argues against treating omissions in the same ways we treat positive actions within the criminal law. His argument stems from the orthodox approach to the act requirement, requiring a willed bodily movement, but strays from orthodoxy in that he is unwilling to acknowledge the importance of the proscription of some omissions in the criminal law that some of his predecessors in orthodoxy (Austin, Bentham, and Salmond) accept quite readily. (4) Moore's treatment of omissions is based on two claims, one metaphysical and the other normative.

(1) Moore's metaphysical claim

His metaphysical claim is an outgrowth of his definition of an act, a definition similar to and based upon that of many of his orthodox predecessors: a bodily movement caused by a volition. Since omissions are not bodily movements caused by volitions, omissions are not actions in Moore's view. In metaphysical terms, Moore claims that omissions are "literally nothing at all." (5) It is important to note that part of Moore's point in making this metaphysical claim is to suggest that it is not only the case that omissions are not actions, but also that omissions are not, strictly speaking, events or states either. Thus, for Moore, omissions share no characteristics or qualities with things in the world around us (and by "things" I meanboth physical objects and the "objects" of metaphysics--those concepts that characterize our everyday experiences but are not readily reducible to their physical components, such as thought, action, consciousness, cause, effect, belief, and so forth). Moore sees omissions as nothing more than deceptive labels that attribute eventlike or cause-like characteristics to an individual's failure to act, when in fact the event or cause in question is something else: a force of nature, the normal course of events, or the actions of another that the omitter (had he acted) would have kept from causing harm. (6) This position allows Moore to argue that omissions cannot be causes, because only events and states, properly construed, can be considered causes (and actions are events). (7) Without causal efficacy, omissions cannot meet the causal requirement of orthodoxy (because they do not cause the results proscribed in the actus reus of the crime), and thus, according to Moore, should not be held criminally liable.

It is not clear that Moore's discussion of causation is unproblematic, and he may be left with a questionable premise about what do and do not count as causes. (8) Still, the proper definition of causation is not the central issue with regard to the place of omissions in the criminal law. Moore himself seems to say as much by addressing the normative claim underlying his view on omissions very early in Act & Crime, long before he develops the metaphysical claim. Moore offers this metaphysical claim primarily as support for his focus on negative duties within his normative claim.

(2) Moore's normative claim

Moore's normative claim presupposes a standard argument against imposing liability for omissions, the general lines of which are easily found in the literature and are fairly straightforward. This standard argument depends upon the values of fairness and liberty and upon a central distinction in the debate over the legitimacy of criminalizing omissions: the distinction between negative and positive duties, and the relative moral significance of each in the criminal law.

The standard argument takes the following form: The law tends to prohibit certain specific positive actions, such as stealing (thus setting up negative duties, that is, duties not to do those actions), rather than require that individuals perform specific positive actions, such as rescuing strangers (which would result in the establishment of positive duties). The reasoning behind this tendency is based on the claim that it impinges more on an individual's personal liberty and is less fair to require her to perform a specific action than it is to prohibit her from performing a specific action, because the latter leaves her with a great deal of choice about what to do (everything which the law does not prohibit) and offers her much more opportunity to avoid criminal sanction. In contrast, requiring her to perform a specific action leaves her no choice whatsoever and, depending on the circumstances in which she happens to find herself, far less opportunity to avoid sanctions. Insofar as statutes proscribing omissions impose positive duties on individuals and statutes proscribing positive actions impose negative duties, statutes proscribing omissions must, the argument goes, be (at least prima facie) less desirable and thus require greater justification in order to be legitimate parts of the criminal law.

Moore's own normative claim follows the general lines of this standard argument against liability for omissions, but also introduces a further value beyond those of fairness and liberty:

Normatively, the retributive value of punishment justifies why we should punish actions and that value (sometimes in conjunction with the values of fairness and of liberty) also justifies why we should punish only actions and not character, emotions, or omissions. The only exception to this is for those omissions that violate our duties sufficiently that the injustice of not punishing such wrongs outweighs the diminution of liberty each punishment entails. (9) The difference between the argument from liberty and fairness and Moore's view presented here derives from his emphasis on the distinction between moral obligation and supererogation. Moore concedes that the actions one takes to make the world better may indeed garner moral praise, but he argues that the purpose of the criminal law is not to enforce (or even encourage) morally praiseworthy behavior, but to prohibit and penalize moral wrongs. This is because retributive justice is not achieved when we do the former, but only when we do the latter.

Moore's argument seems to rely upon at least two dubious assumptions, namely, that moral systems requiring one to maximize utility ask too much of moral agents, and that retributivism is a defensible method for determining how we ought to treat criminals. (10) Nonetheless, despite the potential stumbling blocks resulting from Moore's particular approach to this argument against treating omissions in the same way that the criminal law treats positive actions, the argument in its more general form fails (as does Moore's particular version of it) to attend to the different forms that omissions take and to note the relevant differences between them--differences that warrant very different kinds of theoretical treatment. After all, not all crimes of omission require that some harm result from the omission, thus not all crimes of omission require that the omitter make the world a better place.

II Fletcher on Moore and Omissions

George Fletcher, in a response to Moore's argument against omissions, distinguishes two different kinds of omission: commission by omission and offenses of failing to act. (11) Commissions by omission are omissions that result in some harm usually proscribed under some statute, where that statute is typically understood to prohibit positive actions that cause the specified harm. Offenses of failing to act are "non-result" omissions in that the offender incurs liability by failing to act in a certain situation or by a certain time, regardless of the consequences of that failure to act. (12) Not only does this distinction capture the difference between result and non-result omissions, but it also captures the difference between requiring certain specified actions (proscribing omissions explicitly) and merely treating omissions as positive actions for the sake of punishing someone under a statute proscribing a positive action or a result that typically follows from a positive action.

This latter aspect of Fletcher's distinction leads back to the negative/positive duties distinction, but adds a twist to Moore's treatment of positive duties. Fletcher's distinction here brings to light two different kinds of positive duty required by laws proscribing omissions: duties to act in certain very specified ways at certain very specific times (such as filing a tax return, registering for the draft, and so on) and duties to act in certain general ways at all times toward specific people, that is, duties derived from special relationships (feeding one's newborn child, calling for help for one's sick spouse, and so forth). In his discussion of omissions, Moore focuses more on the latter sort of duty than on the former, primarily because he thinks it particularly problematic to extend statutes proscribing positive actions so that they also cover omissions that have the same prohibited result. However, Moore's position on omissions would seem to apply just as readily to the specific duties as it would to the general ones, because failures to act are, like commissions by omission, "literally nothing at all," and because the imposition of positive duties to act at specific times will infringe on liberty much more than merely requiring that one refrain from some specific action.

III Moore and Tax Evasion

The fact that Moore's argument against omissions would seem to extend as easily to failures to act as it does to commissions by omission is problematic, however, because offenses of failing to act in certain specified ways are fairly uncontroversial and have well-established justifications, only a few of which Moore's argument countenances and none of which it is able to undermine.

Take, for example, failing to file a tax return. A typical justification given for statutes compelling us to pay taxes is that we will all be better off if we pay our taxes because the government can use those funds for projects that we could not otherwise afford, individually or privately, to finance. These projects have a value to each of us beyond the value of the money we pay in taxes. Even if this typical argument is problematic, Moore does not address or provide an alternative to this kind of justification.

It is unfortunate that Moore ignores this sort of justification for punishing a failure to act, because one could construe statutes prohibiting tax evasion as proscribed omissions of the kind for which Moore explicitly allows (omissions for which the injustice of not punishing them outweighs the liberty gained by not proscribing them). This construal would lead to serious difficulties for Moore, however, since he does not seem to want to allow very many crimes proscribing omissions in the criminal law. One can easily understand tax evasion as a sort of free-riding on the system. In this way, it can be seen as an omission that involves taking unfair advantage of the tax-paying members of society. As such, tax evasion strikes me as an injustice of some weight. (13) This fact, coupled with the fact that one loses relatively little liberty when one pays her taxes on time, could lead us to say--and rather easily--that the injustice of not punishing tax evaders outweighs the liberty lost by proscribing tax evasion. (14) Furthermore, for the same reasons we might want to punish the failure to pay one's taxes using Moore's argument, it would seem that we would also have reason to punish most of the omissions proscribed in existing statutes (consider the enormous benefits we, through the police, derive from the registration of motor vehicles, as well as the small inconvenience that registration requires from each of us).

Even though Moore provides no argument for his excepting certain omissions, applying the exception yields a curious result: it would seem to exclude very few, if any, current Anglo-American criminal statutes proscribing omissions (that are offenses of failing to act). If one uses the same kinds of reasons given with respect to tax evasion, it is not clear at all which omissions proscribed by the criminal law Moore is arguing against. As such, it seems that Moore's normative argument fails at what it set out to do: to show that omissions are suspect instances of criminal liability.

IV Moore's Normative and Metaphysical Claims Considered Together

I noted above that the metaphysical question of what causation is does not seem to be the primary issue with respect to whether or not the criminal law should proscribe omissions. Still, insofar as Moore offers his metaphysical claim as support for his emphasis on negative duties within his normative claim, it is important to examine the issue of causation and its bearing on Moore's normative claim against omissions. When one considers his metaphysical claim that omissions are "literally nothing at all" alongside his normative argument against liability for most omissions, it is not clear why, in light of the metaphysical claim, Moore would even bother to make the normative claim. Given his metaphysical stance on omissions, Moore would have to believe that there are no interesting normative questions about omissions since they cause nothing. After all, if omissions cannot be causes, then the "injustices" that might be attributed to not punishing omitters (which he mentions in his exception to his normative claim) must be caused by some other metaphysical entity and cannot be properly attributed to the omission or to the omitter. As a result, from this perspective, it seems that no criminal statute proscribing an omission could satisfy the exceptional condition in Moore's normative claim. If this is an accurate application of Moore's metaphysical claim to his normative claim, then the exception to his normative claim is superfluous.

The superfluity of Moore's exception to his normative claim is especially remarkable, however, in light of the point that I made in the previous section, namely, that Moore's normative exception for some omissions would seem to exclude few, if any, criminal omissions in Anglo-American law. If my analyses in this and the previous section are correct, then Moore's exception for some omissions would both exclude few, if any, omissions and exclude all omissions. Aside from the fact that these are contradictory implications, the real difficulty is that neither implication is one that Moore would endorse. Finally, this conflict between Moore's metaphysical and normative claims effectively undermines whatever support he believes that the metaphysical claim provides for the normative claim.

V Moore's Metaphysical Claim at the Level of Basic Acts

Beyond its adverse impact on his normative argument against omissions, Moore's metaphysical claim against omissions has its own difficulties, most of which can be traced back to his broader theory of action. Moore's broader theory of action, as I indicated above, relies heavily on volitions causing bodily movements as the essential feature that all actions share. He talks of action at two different levels of description: "basic acts" and "complex actions." Relying on the work of Oliver Wendell Holmes, Jr. and J.L. Austin, Moore identifies a basic act as an act that can be the means to another end, but which is not an end toward which any other act can be directed: we do not do basic acts by doing something else; when we do basic acts, we simply do them. (15) Basic acts are those bodily movements we perform, in the simplest possible terms, in order to do something else. Complex actions are richer descriptions of the things people do, and comprise basic acts along with the circumstances in which they are performed. The circumstances of a complex action include facts about the setting and situation within which the basic act is performed (for example, whether the basic act of moving one's hand occurs when one is holding a lit match, whether or not the moving brings the match close enough to a fuse of a stick of dynamite to light it, and whether or not others are anywhere nearby when the dynamite explodes), including the mental state of the agent performing the basic act (whether she acted intentionally, knowingly, and so forth). For example, when one performs the complex action of killing another (by means of a gun), one performs the basic act of moving a finger (to pull the trigger) under certain circumstances (when another is present and when one has a loaded gun aimed at this person). Moore denies that muscular contractions are more basic than bodily movements because it is most often the case that we do not know how simply to contract muscles, or that we do not think of contracting muscles in order to perform a basic act like moving a finger; we just move a finger.

Implicit in this definition of a basic act is the identification of basic acts with "volitions causing bodily movements." This identification supports Moore's claim that omissions are, metaphysically speaking, "literally nothing at all," and that the only instances of acting are volitions causing bodily movements. Moore might say that volitions could be behind some omissions--when one wills oneself not to move, or something of the sort--or that bare intentions are behind them, but inasmuch as those volitions or intentions fail to cause bodily movement, such omissions cannot be instances of action. As such, Moore's theory of action at the level of basic acts would exclude all omissions, intentional and otherwise. One serious challenge that faces a theory of action such as Moore's is the fact that many omissions--especially intentional ones--are clearly instances of human agency, if not instances of action. This is precisely the challenge Fletcher offers to Moore. (16) Fletcher's arguments force Moore to defend his own view more carefully, and Moore's defense makes clearer many of the fundamental problems in his own theory of action. However, even though Fletcher's arguments on this point are interesting, I do not think that they go far enough in demonstrating the deep flaws in Moore's overall theory of action. As such, I will not recount Fletcher's arguments here, but will focus instead on Moore's response to them.

In an attempt to rebut Fletcher's arguments for a unified view of acts (one that includes positive acts and omissions because both evince human agency), Moore points out that in cases of negligent omissions, the omissions are not intentional. He then proceeds to claim that negligent omissions do not involve any human agency, (17) and concludes that since the unified view of actions cannot account for negligent omissions, it fails to account for the majority of the omissions it seeks to incorporate into its theory of action.

Moore is right to note Fletcher's tendency to focus on intentional omissions. He is wrong, however, to claim that Fletcher's unified view cannot account for negligent omissions, because, after all, an omission need not be intentional to be an instance of human agency. And it is strange that Moore would conflate mere agency and intentional action in this way, since in Act & Crime he takes care to argue that volitions are distinct from general (or "further") intentions, and that general intentions cannot do the work that volitions do in action theory. (18)

Even if we set this point aside, do negligent omissions actually pose the serious difficulties for Fletcher's unified theory of action that Moore alleges? In order to answer that question, however, we need to be clearer about what Moore means when he claims that negligent omissions are not instances of human agency. It appears that Moore's negligent omissions objection amounts to something like the following: negligent omissions require no volition whatsoever, whereas all actions, even those performed negligently, require that a volition be directed toward the performance of some act that produces the proscribed result, even if the offender did not intend his act to produce that result. For example, in negligently administering an anesthetic to a patient, a doctor performs the positive act of administering the anesthetic (and in so doing, has a volition that causes his body to move in such a way as to anesthetize the patient), but does so negligently or without taking due care. By contrast, in negligently omitting to pay one's taxes, one may fail to pay one's taxes without having any volition whatsoever--one could do literally nothing at all for a year (though this seems highly improbable). (19)

If this is an accurate portrayal of the reasoning behind Moore's negligent omissions objection, then, in essence, the objection merely asserts that negligent omissions are not acts under Moore's definition of an act. Moore's objection presupposes his account of an act as necessarily entailing a volition causing a bodily movement. But surely Fletcher (and others) would deny that particular understanding of an act, and employ a different account of an act instead--one that does not limit instances of agency to bodily movements. If that alternative definition can account for negligent omissions, then Moore's response falls flat. Furthermore, a definition of an act that can account for negligent omissions such as tax evasion has a distinct explanatory advantage over Moore's definition: it can account for the fairly uncontroversial criminal prohibitions of negligent omissions like tax evasion (which, as I argued in the previous section, Moore's normative claim, oddly, might not exclude).

(1) "Excluder" definitions of voluntariness

One possible alternative definition of an act centers on what Moore calls an "excluder" definition of voluntariness, of the kind advanced by H. L. A. Hart. (20) For Hart, the act requirement defines an act as a voluntary bodily movement. But Hart argues that the concept of voluntariness operative in the criminal law does not refer to a property or feature that all acts have in common, but instead to the exclusion of a set of heterogeneous conditions including somnambulism, hypnosis, unconsciousness, and reflexes. (21) This view takes voluntariness to be an essential characteristic of all responsible conduct, a necessary condition for a bodily movement being an act or for an omission to be punishable under the criminal law. It is a characteristic the absence of which renders questions of mens tea moot. (22)

Since his excluder definition requires that various negative conditions fail to obtain (that one not be unconscious, that one not be under the influence of hypnosis, and so forth), Hart suggests that one could apply it to failures to move one's body as well as to bodily movements. As such, a proponent of a unified theory of action (covering both acts and omissions) could employ the excluder definition as an alternative to Moore's definition of an act. One could then identify "acts" with voluntary behavior or conduct and not just with voluntary bodily movements. (23) Applying the excluder definition in this way, and without tracing out a more complete theory of action, one could argue that since voluntariness is an essential feature of action, any voluntary behavior or conduct (and not necessarily just that including voluntary bodily movements) will be an instance of human agency.

Armed with this (admittedly scant) view of action, one could easily defend a unified theory of action against Moore's negligent omission objection (which is, again, that there is something problematic about the fact that one can commit a negligent omission without having a volition). Consider the fact that one could do literally nothing at all for a year and thereby fail to pay his taxes because he was not able to do anything at all for that year, including paying his taxes (imagine that he has been unconscious and fed intravenously for the past year). It would be very odd to categorize such a case as one of negligent omission. In such a case, a person's failure to pay his taxes would be neither negligent nor an omission that we can properly attribute to him, because his failure to pay his taxes would be involuntary. By contrast, simply forgetting to pay one's taxes when one could easily (or even not so easily) do so would seem to be voluntary under the excluder definition because none of the "excluder" or involuntary conditions obtained. If voluntariness is an essential feature of action, and if it makes sense to say that negligent omissions are voluntary, then negligent omissions must be actions or, at the very least, instances of human agency, and an adequate theory of action should be able to explain that--or explain how we should come to understand negligent activities as something other than voluntary.

One may object that it sounds odd to say that an omission is voluntary when the omitter simply forgets to perform the required action. I suspect, however, that it sounds less odd to say that an intentional omission is also voluntary. These disparate reactions to the possibility of different kinds of omissions being voluntary most likely derive from a tendency to associate voluntariness with positive acts. Since positive acts are usually directed toward some end, it is easy to think of them as teleological. Intentions are also teleological: they are plans, they look forward in time, they take some future state of affairs as their objects. This is why the voluntariness of intentional omissions does not sound odd at all: omissions that are intentional are explicitly directed toward some end. But notice also that attributing voluntariness explicitly to negligent positive actions really seems no less odd than does attributing voluntariness to negligent omissions. This is because "the essential fault in negligence is one of failure by the defendant to take reasonable care: It does not seem to matter very much whether that failure manifests itself in a doing [positive action] or a not-doing [omission]." (24)

Negligent omissions, however, are more difficult, because they require no clear active directing of oneself toward anything. But why is responsibility for omitting to do something that the omitter has a duty to do uncontroversial when the omitter could have easily performed the required act? The answer has to do with the presence of a positive duty to behave in a specified way and adequate opportunity to behave in that way. Ascriptions of responsibility for positive acts depend upon similar conditions, namely, the presence of a negative duty not to behave in a specified way and the opportunity to avoid behaving in that way. Under the excluder definition of voluntariness, the concept of voluntariness serves to exclude those conditions that deprive the agent of the opportunity to avoid moving her body in a particular way. But these conditions (somnambulism, hypnosis, unconsciousness, and so on) could also deprive an agent of the opportunity to move her body in a particular way. As a result, under the excluder definition (applied in the way I have suggested here), the concept of voluntariness would serve to exclude those conditions that deprive the agent of either the opportunity to avoid moving her body in a particular way or the opportunity to move her body in a particular way. Under the excluder definition of voluntariness, if one has a duty to perform a positive act, omitting to perform that positive act would be voluntary as long as one has adequate opportunity to perform it.

Of course, the practical application of such a definition of voluntariness would require standards of "adequate opportunity" for criminal statutes proscribing positive acts distinct from those for statutes proscribing failures to act. The standards will be distinct because, with respect to positive acts, the standard will be relevant only at those particular times at which one actually behaves in a prohibited manner, whereas with respect to failures to act, the standard will have to take account of the entire time that the positive duty to act is in place. Just how much opportunity is adequate in cases of failure to act is also going to depend, I would think, upon how important and how difficult the required positive act would be to perform. So, for example, when X kills Y at time t, we ask whether or not X had adequate opportunity to avoid moving his body as he does at t. However, when A fails to register her car within thirty days of moving into a new state, we ask whether or not A had adequate opportunity to register her car during those thirty days. If it turns out that A is hospitalized for those thirty days, the failure to register would be involuntary. Were A hospitalized for, say, twenty of those thirty days, the decision would likely be different. Exactly where the lines are drawn for what should count as "adequate" opportunity with respect to individual statutes would have to be determined either explicitly by the legislature or at common law.

(2) Voluntariness versus volitions

Despite the fact that some negligent omissions would be considered "voluntary" under an excluder view like the one I have suggested here, Moore's metaphysical view of action would nevertheless seem to exclude those omissions from liability because he is committed to a volitional account of action. As I noted above, negligent omissions require no volition whatsoever, and for Moore all acts are volitions causing bodily movements (even if Moore's normative claim might allow liability for those omissions for which the injustice of not punishing them would outweigh the loss of liberty that punishing them would entail).

But given the fact that a volitional theory cannot account for some uncontroversial cases of omission liability (such as certain kinds of tax evasion), why does Moore develop his theory of action around volitions and not around a broader concept like (an excluder version of) voluntariness? There seem to be two related reasons why Moore chooses this direction. First, Moore seems to think that the more clearly causal an account of action is, the better that account will be able to track and justify the ascriptions of responsibility we make within the criminal law. Second, Moore's overall aim is to establish action as the unifying concept of criminal liability, and volitional accounts of action tend to be univocal or homogeneous, making them well-suited to serve as unifying concepts. By contrast, it would be much harder to defend action as the unifying concept of criminal liability if the concept of action were "heterogeneous" in character, as it would be under the excluder definition.

Unfortunately for Moore, these two reasons fail to provide sufficient justification for a volitional account of action within the criminal law. In terms of the first reason, causation is not central to all ascriptions of responsibility: we typically hold others morally responsible not only for the bad things they cause (even if one construes the "bad things" here very broadly, as Moore does), but also for certain good things they fail to cause or do.

But perhaps part of Moore's point is to show that the ascriptions of responsibility we make in the criminal law do not (or should not) track our moral responsibility for the good things we fail to cause or do. If so, the questions then become: what aspects of moral responsibility should the criminal law track, and why? Moore attempts to answer these questions by appealing to his normative claim against omissions, and the argument from liberty and the retributive value of punishment it presupposes. Essentially, Moore's normative claim precludes punishment for emotions, character, and omissions, the one exception being those omissions "that violate our duties sufficiently that the injustice of not punishing such wrongs outweighs the diminution of liberty each punishment entails." (25)

As an explanation of how closely the criminal law should track moral responsibility, the most interesting feature of Moore's normative claim is the fact that it allows omissions as exceptions, while disallowing emotions or character as exceptions. Moore does not say explicitly why it is that omissions can "violate our duties sufficiently," while emotions and features of character cannot. The most obvious reason is that it is often the case that we cannot control our emotions or our characters, and consequently, that it would be unfair to punish someone for something she could not reasonably avoid (in order to be able to avoid sanction). If this is the reasoning behind Moore's exception, then Moore seems to believe that we can control our omissions or, at the very least, that we can violate our duties by omitting actions. If this is correct, then Moore must indicate why he uses this fact about omissions to justify the exception to his normative claim, while maintaining at the same time that, causally, omissions are "literally nothing at all." Either we can violate our duties by omitting or we cannot violate our duties by omitting (because our omitting can itself cause nothing for which we can be held responsible). Moore cannot use the fact that omissions can violate our duties as a justification for punishing omissions in one place and then dismiss liability for omissions as merely exceptional in another. If in fact Moore tries to do both at the same time, the most charitable thing we can say of his view is the following: the causal explanations Moore takes to be constitutive of ascriptions of responsibility are only part of the story.

The problem with the second reason that Moore chooses to advance a volitional theory of action (because he sees it as instrumental to establishing action as the unifying concept in the criminal law) is that the theory of action it yields cannot account for significant and uncontroversial parts of the criminal law. Part of the point of action theory within the criminal law is to provide some metaphysical explanation for the various ascriptions of responsibility we make. However, since the conditions under which we make those various ascriptions of responsibility (both in morality and in the law) are heterogeneous in their character, a metaphysical view that portrays the conditions under which we ascribe (either moral or legal) responsibility as homogeneous will be inaccurate or incomplete. Of course, the existence of metaphysical inconsistencies in present practice can be the basis for reformatory arguments. In such circumstances, a theory's descriptive lack of "fit" with present practice is justified by the need for theoretical consistency. But metaphysical heterogeneity is not metaphysical inconsistency, and though unity in metaphysical principles is desirable, it is not required for the criminal law to retain legitimacy.

Thus, the main reasons behind Moore's commitment to a volitional theory of action fail to provide adequate justification for that theory, especially in light of the apparent heterogeneity of the essential features of action and in light of alternative theories of action that can account both for omissions and the heterogeneity of action. A theory of action based upon the revised excluder definition of voluntariness I outlined above does just this. Of course, in making the case for voluntariness over volitions, I may be appealing implicitly to an "ordinary language" justification for taking omissions to be instances of action. In other words, one might object that my arguments for including omissions under the definition of an "act" amount to appeals to how we normally use the language of action, because when we typically talk of someone omitting, we tend to talk of that person omitting in much the same way that we talk of someone performing some positive act. Moore argues against such attempts to justify a theory of action (and simultaneously seems to concede the weakness of the metaphysical underpinnings of his own view) when he claims: "Volitions may not be the right answer, but no amount of discovery of near synonyms for 'action' can show that the question to which volitions purport to be the answer is a question that need not be asked." (26)

However, I do not think that I am guilty of appealing--consciously or unconsciously--to ordinary language in arguing for the plausibility of understanding omissions as instances of action or, at the least, human agency. I take it that, for Moore, volitions purport to provide the answer to the question of what the proper criteria are for making ascriptions of responsibility. As such, if it makes sense to ascribe responsibility to individuals for some omissions, then in order to argue for criteria that do not include those omissions, Moore needs to undermine the plausibility of responsibility for those omissions in his account. Beyond appeals to metaphysical simplicity and elegance (which Moore believes his volitional account exemplifies) and highlighting worries about holding omissions liable under statutes requiring a positive act, Moore does not provide any arguments against the plausibility of an account of moral or legal responsibility for some omissions such as the one I have provided here (however scant and incomplete it may be). I have argued above that having a positive duty to perform some specific positive act, along with having adequate opportunity to perform that specific positive act, together constitute one plausible set (among many others one could suggest, depending on the kind of conduct in question) of sufficient conditions for making ascriptions of responsibility. Furthermore, the development of these conditions does not depend upon an ordinary language analysis of "action," but rather upon whether or not we would have reason to hold someone responsible in those circumstances for failing to act. In the end, the excluder definition of voluntariness accounts for more of the diverse criteria we employ in ascribing responsibility than can Moore's volitional account. But then again, perhaps Moore has a different question in mind to which volitions purport to be the answer, though I cannot imagine what other question (or purpose for volitions) he might have in mind.

VI Moore's Metaphysical Claim at the Level of Complex Actions

Even if Moore could answer these objections and still preserve his volitional theory at the level of basic acts, it is not clear that his theory is defensible at the level of complex actions. One problem with Moore's view at the level of complex actions is that even if his volitional theory of basic acts holds, it may not be the case that every complex action requires a volition causing a bodily movement. In other words, there could be circumstances within which one's negligence or even willful inaction leads to a result that in other circumstances we would think appropriately described in terms that attribute some complex action to that person. We might, for example, find it appropriate in certain circumstances to refer to a mother's failure to feed her child as her starving the child.

Moore attempts to address this problem by defending the "identity thesis," which claims that all complex actions require and are identified with basic acts (volitions causing bodily movements). For Moore, all complex actions are basic acts, just more richly described. In other words, every complex action is partially identical to some basic act; they are only partially identical because, as complex actions, all of the attendant circumstances are included in the action description as well. It is important to note that Moore sees the act requirement (that criminal liability may not attach unless the offender has performed an overt act) as a sub-requirement or prerequisite of the actus reus requirement (that criminal liability may not attach unless the offender has performed some action prohibited by some specific criminal statute), and, as such, any offender satisfying the actus reus requirement must also satisfy the act requirement. Since criminal statutes proscribe types of complex actions, the identity thesis guarantees that all complex actions identified under the actus reus requirement will also satisfy the act requirement, because each complex action is partially identical to some basic act. Because it ties together Moore's metaphysical claims about basic acts with the explicit prohibitions of criminal statutes, the "identity thesis" turns out to be central to Moore's formulation of the actus reus requirement and to his overall account of action in the criminal law. However, even were the identity thesis true, identifying the basic act to which a given complex action is identical turns out to be either very difficult or irrelevant in many cases.

Moore notes that many theorists have argued that omissions constitute counterexamples to the identity thesis, and he formulates the objection in the following way:

[I]n certain circumstances we cause death by intentionally refraining from preventing a death, and if we cause a death by our intentional refraining, we kill.... Ergo, the desired conclusion: some complex actions can be done without the presence of a volitionally caused bodily movement because they can be done by an omission. (27)

Moore then proceeds to attack the necessary-condition conception of causation. Originating from the sine qua non test of cause-in-fact, the necessary-condition conception of causation holds that if a result would have occurred but for one's intervention, then failing to intervene is causally equivalent to bringing about the result, because it was a necessary condition of the result occurring that one failed to intervene (otherwise, the result would not have occurred). (28) This is the only conception of causation that Moore believes might allow us to say that we can cause certain effects by refraining from preventing those effects when other forces would bring those effects about without our intervention.

Rather than recounting Moore's arguments against the necessary-condition conception of causation, I will say only that Moore presents a convincing case against the necessary-condition conception of causation, and I tend to agree with him that omissions, metaphysically speaking, cannot be causes (under Moore's definition of a "cause"). But even if one grants Moore this claim, several difficult questions remain. The most pressing are the following: How do we characterize intentional omissions? If they are not complex actions because the basic acts with which they are identified are "literally nothing at all," then what are intentional omissions? Do we merely project them onto the world? Is the criminal law (along with our common understanding of our own agency) just grossly mistaken about how to characterize omissions properly?

It is not clear what Moore's answer to this last question would be. At times, he takes a hard line against omissions counting as complex actions, but only with respect to holding people liable for omissions under statutes that require one to cause (directly or actively) the proscribed result, usually by performing a positive act (for example, construing D's having allowed V to die as a direct cause of V's death). (29) At other times, though, he is willing to treat a few kinds of omissions as if they were complex actions for the sake of holding them criminally liable. However, Moore tends to allow only omissions of a certain sort to serve as exceptions to his version of the act requirement, namely, those omissions that are failures to uphold some legally significant positive duty to act, usually originating from an earlier assumption of responsibility for another person (in some form or other). (30)

Fletcher and others actually agree with Moore that there are problems with treating omissions as though they were positive actions under statutes that require the causing of a specific result. Fletcher notes "that commentators would be shocked by the wholesale breach of legality in the judicial development of duties that supplement the law of homicide" if they were not more "concerned about condemning the injustice of not punishing immoral omissions." (31) Given that others share Moore's concern that holding people liable for some omissions violates the principle of legality, it would seem that we could grant Moore his point here, but nonetheless demand that new statutes be written to criminalize omissions resulting in certain harms. Oddly enough, Moore does not seem opposed to that alternative. Moore suggests that it would be desirable to enact omissions statutes because

1) such statutes would command actions (prohibit omissions) that are now punished despite not being anywhere by statute required; and 2) such statutes would attach lesser penalties to the omission to save than to active killing, a lesser punishment juries and judges now typically give anyway but without statutory guidance or authorization. (32)

Such statutes would avoid the metaphysical difficulty of holding people liable for omissions whose results they could not possibly cause and would avoid the normative difficulty of violating the principle of legality. At the same time, however, Moore's other normative claims would seem to rule out almost all such statutes as unduly infringing on individual liberties, the clearest exception being the failure to care adequately for one's own children. As such, even if we can solve the metaphysical problems via more carefully worded statutory prohibitions of some omissions, it seems that Moore's normative arguments would still undermine most of these statutes' claims to legitimacy. (33)

Similarly, even though Moore seems to support liability for some failures to uphold one's legally significant positive duties (primarily those duties that stem from one having explicitly or implicitly assumed responsibility for another person or persons), it seems that, given his normative claims, there would be very few such positive duties whose enforcement would not unduly infringe upon individual liberties. So why does Moore readily accept liability for omissions to uphold certain positive duties when his normative arguments against omissions would seem to rule out the enforcement of those positive duties? The answer, I think, is that Moore takes the implicit or explicit assumption of a positive duty (by voluntarily taking responsibility for another person or persons) as a kind of exception to his normative argument against omissions: it is an unwarranted infringement on individual liberty to require an individual to perform some specific positive act (as opposed to requiring her not to perform some specific positive act) unless the individual in question has voluntarily assumed for herself the positive duty to perform that positive act and failure to uphold that duty results in a sufficiently serious injustice. (34)

If these are the criteria Moore that employs in determining whether or not to make an exception and hold a given omission criminally liable, then one has to wonder why he is not more careful to make this exception explicit, especially when it appears to cover most of the omissions that Moore would want to hold criminally liable. Part of the reason is that he does not take some of these kinds of omissions to be exceptions to his view of action: consider, for example, Moore's treatment of creating "conditions of peril." By creating a "condition of peril" for another, a person makes it her duty to rescue the other, and failures to rescue in such situations are omissions for which Moore would hold a person liable. These sorts of omissions do not constitute exceptions to Moore's view of action because by creating the condition of peril, the individual in question performs some basic act that is partially identical to some complex action prohibited by a criminal statute (proscribing the harm that would result from the condition of peril). Were one to be charitable to Moore here, it would seem that assuming positive duties of any sort would have to work in much the same way: when one assumes a positive duty, one performs some basic act. For example, when one assumes the duties of a parent, one performs a basic act when one takes the baby home and therefore decides not to give it up for adoption. In this way, the identity thesis could be said to hold, even with the omissions Moore would want to criminalize.

However, characterizing these omissions in this way leads to serious difficulties. First, consider offenses following the creation of a condition of peril. Even though it is generally uncontroversial to hold individuals liable for failures to rescue that follow the creation of a condition of peril, the way in which Moore characterizes those failures to rescue as complex actions is more problematic. Essentially, Moore's position is that we should hold individuals liable for the positive act of creating a condition of peril, but only in the circumstance (which is part of the complex action) in which one does not then rescue the victim. I say this because, for Moore, the creation of the condition of peril is the basic act that is partially identical to the complex action that satisfies the actus reus of a crime. The positive act that creates the peril is the basic act; the failure to rescue cannot itself be a basic act, but only a circumstance within which the basic act is performed, and a circumstance that changes or qualifies the character of the complex action that is partially identical with the basic act of creating the peril. But in failures to rescue so characterized, the mens rea of the crime will occur much later than the basic act of the crime. In such cases, the mens rea will (not in Moore's view) be either the intentionality involved in omitting to rescue at the moment immediately before the harm actually befalls the victim, or the negligence or recklessness involved in omitting to rescue for all the moments leading up to the harm befalling the victim. Either way, it is telling that the mens rea is simultaneous with the omission, with the failure to rescue, and not with the creation of the condition of peril.

Another problem is that characterizing failures to rescue as Moore does precludes the possibility of a given failure to rescue being involuntary, even if one creates the peril voluntarily. For example, one could voluntarily set a car in motion on a crowded street by intentionally pressing down on the gas pedal, thereby creating a condition of peril for the pedestrians in the street and thus creating and assuming for oneself the positive duty to rescue those pedestrians by stopping the car before it hits one of them. If, immediately after voluntarily setting the car in motion, a person were to suffer a massive, incapacitating stroke, the resulting failure to rescue would have to be involuntary, or at least not attributable to her. However, in Moore's view, since one voluntarily pressed the gas pedal within the circumstance of then failing to rescue those she endangered, the action description would be complete and would be properly attributed to her. However, because she was incapacitated and could not rescue, an excluder definition of voluntariness could explain the intuitive claim that she did not properly fail to rescue, even though she created the condition of peril.

Notice also that both of these objections to Moore's treatment of failures to rescue after creating a condition of peril apply just as readily to failures to act on a previously assumed positive duty. In fact, because there is often a longer period of time between the act of assuming the positive duty and a possible violation of that duty, the first objection (that on Moore's view of creating conditions of peril, the mens rea and basic act do not occur at the same time) seems to have more force in the case of failures to act on a previously assumed positive duty. Consider, for example, a parent failing to take adequate care of his three-year-old child. The basic act of assuming the duty to take care of the child as a parent occurs some three years prior to the failure to act, and someone holding a view like Moore's would therefore have to explain how it could be that the mens rea and the basic act of the crime occur three years apart. But even were one to accept Moore's characterization of some omissions as exceptional sorts of complex actions, such a characterization would identify this species of omissions too closely with the acts of creating conditions of peril or the acts of explicitly assuming duties and not closely enough with the expressions of agency that we associate with the wrongdoing these omissions represent. The result is that it would fail to capture much of what we think is wrong about many of the omissions that the criminal law prohibits.

Still, as my objections here and at the level of basic acts indicate, the most serious deficiency in Moore's theory of action is its failure to recognize explicitly the role that positive duties play in legitimate ascriptions of responsibility, both in morality and in the law. Moore fails to appreciate fully that the presence of a positive duty can allow us to treat the omission as we would a positive action directed toward some foreseeable harm (the performance of which would be prohibited by a negative duty). This may be no different, one might say, from the way in which we treat contractual obligations: in certain relationships (contractual or otherwise) certain positive duties are implicit that would not be there but for the existence of the relationship. It is not difficult to think of failures to uphold certain positive duties as failures having as much force (both moral and legal) as any failure to uphold a negative duty. (35) It is not difficult to think this way because positive duties stemming from special relationships carry a suppressed negative duty: one ought not to abandon such duties once one has assumed them. One might respond that contractual obligations deserve different treatment because one undertakes them voluntarily. However, a person also accepts voluntarily, though perhaps only implicitly, all positive duties imposed on her by the criminal law (except perhaps the duty to pay taxes, when one has no means to go elsewhere, and perhaps the duty to be a good Samaritan--unless that duty is well-established and well-known).

One reason why Moore may neglect the positive duties that exist within the criminal law is that the existence of such duties would allow for some criminal prohibitions to be enacted without a specific basic act (again, a volition causing a bodily movement) being identified with the prohibited complex action. But consider the vast array of rules and obligations that play prominent roles in our descriptions of complex actions, in light of which precise basic act descriptions appear to be irrelevant or at least indeterminate. For example, running a stop sign is a complex action that Moore would seem to identify with pressing down on the gas pedal with one's foot as one drives up to a stop sign. Nevertheless, there does not seem to be any reason, given the complex action description, not to claim that the basic act is failing to press down on the brake pedal with one's foot as one drives up to a stop sign, or even removing one's foot from the gas pedal in order to scratch one's toe. It is underdetermined to which basic act the complex action is identical, and in the eyes of the law, as long as the basic act is voluntary, it makes no difference which basic act we choose. Of course, the basic act description that constitutes an omission (failing to press the brake pedal) would not even be a candidate basic act for Moore: he fails to see how we could take the omission, along with its context, to constitute a complex action when the offender's expression of agency (refraining from pressing the brake pedal) does not actually cause the car to run the stop sign (the car's momentum does). But once again, it is the fact that we count the positive duty to stop at stop signs as a relevant part of the circumstances of the complex action that allows us to consider an agent's willed (or even negligent) failure to press the brake pedal in those circumstances as the complex action of running a stop sign. In the end, however, it does not really matter which version of the story we choose: the complex action is the same and the person would still be held liable. This last claim would certainly dismay Moore because, were it true, the criminal law would lack any clear metaphysical grounding in direct causation. The problem for Moore is that this last claim would be true not only of determinations of liability in the criminal law, but also of everyday ascriptions of responsibility.

But whether or not the criminal law (or even any everyday ascription of responsibility) really lacks a clear metaphysical foundation, the debate over the legitimacy of statutes proscribing omissions does not seem to center on issues of causation, but on the relative merit of enforcing positive duties in the criminal law. What I have tried to show is that issues of causation with which Moore concerns himself are largely irrelevant: Moore may, in fact, be able to defend his definition of basic acts against the objections I raise above, and he may be correct that omissions do not cause anything, but these facts do not prevent us from treating some omissions as if they do cause things, namely, those omissions that mark a failure to uphold a positive duty imposed by the criminal law. My argument is normative, not metaphysical. It is, moreover, a normative argument that depends little, if at all, on a particular account of the metaphysical nature of omissions at the level of basic acts.

VII Toward an Alternative Approach

Moore's attempt to reconstruct the orthodox act requirement is only a part of his larger project to defend the broader orthodox actus reus requirement. With this larger project, I am much more sympathetic, but it is not at all clear to me that a defense of the actus reus requirement must depend crucially on a simultaneous defense of the orthodox act requirement. Still, in order to offer an outline of how one might pursue a defense of the orthodox actus reus requirement while giving up on the orthodox act requirement, it is helpful to see exactly how Moore's defense of actus reus via a defense of the orthodox act requirement goes astray.

Consider Moore's definition of the actus reus requirement, which is roughly the orthodox definition:

The actus reus requirement is that no penal liability may attach to any individual unless that individual has done an act that instantiates a type of action prohibited by some statute in force at the time the act was done. (36)

For Moore, the act requirement plays a crucial role within the broader actus reus requirement: by requiring each complex action prohibited by the criminal law to be identical to a particular "volition causing a bodily movement," Moore claims that his act requirement provides a metaphysical basis for the broader actus reus requirement. The issue for Moore here centers on the generality of the actus reus requirement and its ability to apply to the entire criminal law. Under Moore's definition of actus reus, there are as many possible actus reus requirements as there are different types of complex action, so without something more basic which is common to all complex actions, the actus reus requirement is not an interesting general principle of criminal liability.

But we have seen that there is little reason to believe that all that various complex actions have in common is some identification with "volitions causing bodily movements." Indeed, one could argue, as I have suggested above, that there are complex actions that are not even properly (or determinately) identified with any "volitions causing bodily movements." More significantly, however, Moore's theory of action leads to a variety of odd implications and even a few absurd conclusions, some of which are so indefensible (Moore's neglect of positive duties in the criminal law is prominent among these) that they undermine any plausibility his larger project might have. The bottom line is that there is much more to human agency than Moore suggests, and, as Fletcher points out, the act requirement does seem to be about human agency. (37) The act requirement, if we are going to retain one, must identify those key features of human agency common to every complex action the criminal law proscribes (provided that one does not abandon Moore's definition of the broader actus reus requirement as well), so that actus reus may be an accurate, general principle of the criminal law.

Identifying the key features of human agency common to every complex action the criminal law proscribes is no easy task, however, because an account of human agency that covers both action and inaction (in both intentional and negligent forms) will be either heterogeneous in character (in that it would offer, among other things, different treatments of positive actions and omissions) or so broad as to be unhelpful as a formal requirement of criminal liability. For an example of the latter sort of account, consider the alternative reading of the act requirement Michael Corrado offers in response to Moore's understanding of the act requirement. Corrado suggests that the act requirement be understood as tracking very closely a key condition for responsibility: the act requirement for Corrado limits criminal liability to those offenders who have the ability to choose otherwise at the point at which they commit their crimes. (38)

Though I think there is a lot to be said for Corrado's reading of the act requirement, it is so general and tracks moral responsibility so closely that it does nothing more than assert that punishment should be limited to those who were responsible agents at the time of their offenses. But in orthodoxy, the act requirement is more than a requirement of moral responsibility: a traditional justification for the orthodox act requirement is that it precludes punishment for thoughts. Against such a claim, Corrado contends that the no-punishment-for-thoughts justification is not a deep, principled reason for an act requirement in the criminal law, but rather a consideration that derives from the fact that we do not currently see the harms that thoughts can cause worthy of the loss of liberty that would be required to prevent those harms. (39) I take the act requirement to have a deeper, more principled purpose, namely, to allow individuals reasonable opportunity to avoid criminal sanctions. The act requirement serves this purpose in two distinct ways. First, it limits liability to expressions of agency (positive actions and omissions) and disallows liability for thoughts, since it is sometimes (or even often) the case that individuals cannot control what thoughts pop into their heads. Second, it allows, even encourages, individuals to change their minds about engaging in criminal activity. Even if one has initiated some criminal activity, the law provides incentive to change one's mind: consider, for example, the fact that abandoning an attempt at a crime substantially mitigates the guilt of the attempter. If I am right about the purpose of the act requirement, the act requirement really is about human agency after all (and it seems, at least on its face, easily defensible), while Corrado's reading of the act requirement is about a prerequisite for human agency, not about actual expressions of human agency. Still, although Corrado's suggested reading of the act requirement tracks a crucial requirement of moral responsibility, it can not achieve the ends orthodoxy set out for the act requirement. Furthermore, Corrado's act requirement is so general that it is better understood as a basic moral principle and, as such, has little to offer as a legal principle (or as a requirement of criminal liability and not simply morality).

The better approach is to find an account of human agency that is consistent with the purpose of the act requirement, as orthodoxy understands it. That account would most likely be heterogeneous in character. It would have two main sub-headings: positive action and omission. Omissions would be understood as parasitic upon positive duties, and the relevant standard for determining whether or not an individual meets the threshold of the act requirement (of having expressed or employed agency) is one of adequate opportunity to have performed the positive action that the positive duty requires. Positive actions would be understood as something along the lines of "voluntary bodily movements." I do not have a clear idea of how exactly one should define "voluntary" here (if I did, it would be beyond the scope of this last section to explain and defend it), though I do think it preferable to avoid "dark references" to volitions--whatever volitions are--in defining the term. Of course, the better definitions will be those that are not inconsistent with the understanding of omissions I suggested just above.

Consider briefly how an excluder-style account of voluntariness or agency might work within the broader actus reus requirement as Moore defines it. Under such an account, there would be more than one set of sufficient conditions for agency or voluntariness (there would be at least two: one for voluntary bodily movements and one for omissions, along the lines of those conditions I suggested in the previous section). Then, once that threshold was met, one would ask whether the expression of agency in question instantiated a type of action prohibited by some statute in force at the time. This latter move again underlines just how many different expressions of agency can satisfy the particular actus reus of a given criminal prohibition. But this latter move is also important because it helps to account for many different kinds of liability that a view like Moore's struggles to handle, especially liability for negligence and recklessness. And interestingly, the resulting (heterogeneous) account is more universal, more univocal overall than is Moore's more self-consciously univocal, homogeneous account: one would say, under an excluder-style account, that liability attaches provided that the offender performed an action that instantiated a type of action prohibited by statute and that no excluding conditions obtain (such that we would say that the action was properly attributable to him, or was voluntary). This account would be true of negligent behavior as well as intentional behavior, but the excluding conditions would vary depending on the kind of behavior in question.

Granted, an excluder-style account is a broad one that would categorize as voluntary many things that action theorists may be uncomfortable thinking of as "acts" or "actions." But given the act requirement's role as a sub-requirement within the broader actus reus requirement, such worries seem much less relevant than they might otherwise be. Consider drumming one's fingers on a table as one reads. One would probably not think of such behavior as voluntary, especially if one is not fully aware that one is doing it. But under an excluder definition of voluntariness, such behavior would be voluntary, since no excluding conditions obtain in that situation (the drumming was not the result of an epileptic fit or somnambulism, for example). Even if one is extremely uncomfortable calling such drumming of fingers "voluntary," the voluntariness of that activity is really irrelevant unless one could somehow construe the drumming to be an instantiation of a type of action prohibited in the criminal law at the time.

But consider this possibility: the person drumming her fingers is doing so on a control panel hooked up to electrodes attached to a volunteer's head, such that the drumming causes the volunteer to receive a lethal shock (that is not part of the testing for which the subject volunteered). In such a case, even if the person doing the drumming was not fully aware that she was drumming her fingers on the control panel, one could argue that she should have been aware of it, or at least should have taken due care to prevent herself from possibly doing it--in order to avoid a manslaughter (or some similar) charge. In such a scenario, the question of the voluntariness of the one drumming her fingers is very relevant and seems open to a different standard from the one that one might have at first thought: given her presence at the control panel (not to mention her role as a researcher or whatever), did she have opportunity to stop herself from drumming her fingers on the panel, even inadvertently? If so, then the threshold of voluntariness would have been met as far as the excluder account is concerned (and, as far as our normal ascriptions of moral responsibility are concerned), despite the fact that such an understanding of voluntariness might make some theorists very uncomfortable. In this way, should the activity in question instantiate a type of action proscribed by the criminal law, then whether or not the activity was expressive of the agency of the defendant will matter, and one would need to select among the several, heterogeneous sets of sufficient conditions for determining whether or not a given activity is properly attributable to an individual.

Such a heterogeneous account of human agency within the criminal law would offer significant advantages over narrower approaches to action (like Moore's) and broader ones (like Corrado's). Unlike approaches that focus solely on "volitions causing bodily movements" (or even the more traditional "willed bodily movements"), a heterogeneous account would give positive duties their due in the criminal law and would be able to explain why prohibitions against some omissions are well-established and generally uncontroversial parts of the criminal law. Furthermore, unlike approaches that focus on very basic conditions of responsibility, a heterogeneous account would not focus on general abilities in the abstract, but rather on expressions of agency within the contexts of previously assumed positive duties and existing negative duties. The primary advantage of this contextualized focus is that it gives one an omission or a positive action to which mens rea can attach as a qualifier. The criminal law is interested in mens rea only as it qualifies or attaches to some expression of agency: merely negligent thinking (if such a thing is possible), merely having knowledge, reckless thoughts, or "bare intentions" cannot qualify the ability to choose otherwise, they can only qualify actual expressions of agency. Overall, a heterogeneous account of human agency would reflect, explain, and justify the workings of the criminal law much better than would the alternatives I have considered in this essay.

Despite its advantages, though, the heterogeneous account of human agency I have begun to outline here would not allow one to provide a metaphysical theory of action within the criminal law that is univocal in the sense of being based on a single, homogeneous, all-encompassing metaphysical picture of action that links criminal actions to criminal harms in a direct way. In general, however, I do not think that is such a bad thing: action, as we understand it, is anything but univocal, even though we can profitably break down the concept into a number of complementary categories. I think that it is too much to expect a theory of action within the criminal law to be univocal, that is, to present us with a single, unifying metaphysical concept on which to base all of our understanding, but it is not too much to expect a theory of action to account for most, if not all, of the key features of human agency that should and do play prominent roles in the criminal law. (40) Moore's position fails to live up to either expectation, but his failures are instructive and point toward a better direction for action theory within the criminal law: one that recognizes the heterogeneity of action within the criminal law.


(1) By "act requirement" I do not mean the broader actus reus requirement, or the requirement that, as Moore puts it, "no penal liability may attach to any individual unless that individual has done an act that instantiates a type of action prohibited by some statute in force at the time the act was done." Michael S. Moore, Act & Crime: The Philosophy of Action and Its Implications for Criminal Law, Oxford: Clarendon Press, 1993, p. 189. The act requirement, as I define it here, is, it seems to me, a sub-requirement of the broader actus reus requirement or a part of that broader requirement.

(2) Id., pp. 44-45.

(3) Id., pp. 33-34.

(4) J. Austin, Lectures on Jurisprudence, 5th ed., London: John Murray, 1885; Jeremy Bentham, An Introduction to the Principles of Morals and Legislation, reprint, ed. by J. H. Burns and H. L. A. Hart, London: Athlone Press, 1970; John Salmond, Jurisprudence, 11th ed., London: Sweet & Maxwell, 1957; and P. J. Fitzgerald, Salmond on Jurisprudence, 12th ed., London: Sweet & Maxwell, 1966.

(5) See Moore, Act & Crime, p. 28, then discussed in terms of causation on p. 277.

(6) Despite Moore's claims to the contrary, it does seem that individual failures to act do have some metaphysical status: we can identify and distinguish them from other events and non-events and even understand them as effects of certain mental causes, such as thoughts, motivations, or negligence.

(7) See Moore, Act & Crime, ch. 3.

(8) Michael Corrado, for one, wonders whether or not acts, as Moore defines them, can properly be called "events," because acts are "causings," and it is not clear that "causings" can themselves "cause" other things. If they can, Corrado thinks it is unclear whether or not they are the kinds of events that can cause the kinds of results prohibited by the criminal law. See Michael Corrado, "Is There an Act Requirement in the Criminal Law?" University of Pennsylvania Law Review, 142 (1994), p.1533, n. 22.

(9) Moore, Act & Crime, p. 59.

(10) In fact, there seem to be many reasons for thinking that retributivism is untenable, or at least highly problematic. See Russ Shafer-Landau, "The Failure of Retributivism," Philosophical Studies 82 (1996), pp. 289-316.

(11) George Fletcher, "On the Moral Irrelevance of Bodily Movements," University of Pennsylvania Law Review 142 (1994), pp. 1443-53.

(12) Moore argues that all actions entail causation of some state of affairs, and as a result, there are technically no "conduct" or "non-result" crimes. See Moore, Act & Crime, pp. 218-25. Of course, since he does not believe that omissions can cause anything, it is not clear that Moore would make a similar argument with respect to cases of omission liability.

(13) And this wrongdoing, it would seem, would be the end of the story for most retributivists, since retributivism, unlike other theories of punishment, does not "weigh out" the fact that the offender deserves punishment against anything else. I thank Harry Silverstein for pointing this out to me in commenting on an earlier version of this paper.

(14) I am not sure how one might determine a perfect balance of injustice and liberty, but cashing out this "balancing" does seem to bring out a defect in Moore's argument. My sense is that if one must use Moore's balancing test at all, it would have to be used in much the way I have used it here: by appealing to our intuitions about the relative moral weights of injustices and liberties, but only when there is enough of an obvious imbalance that fine measurements need not be made (because Moore offers no mechanism for making such fine measurements and because the value of justice (as Moore understands it) and that of liberty appear to be fundamentally incommensurable).

(15) Moore, Act & Crime, pp. 79-80.

(16) Fletcher, "On the Moral Irrelevance of Bodily Movements."

(17) Michael S. Moore, "More on Act & Crime," University of Pennsylvania Law Review, 142 (1994), pp. 1774-75.

(18) Moore, Act & Crime, pp. 117-23.

(19) The fact that one could negligently omit to pay one's taxes by doing absolutely nothing at all over the course of the year, however improbable that may be, shows that interpreting tax evasion as having a "hidden act requirement"--as a prohibition on the positive action of earning income and failing to pay taxes on it--is not without its problems. Moore appears to interpret prohibitions against tax evasion in just this way. See Moore, Act & Crime, p. 32. After all, one can earn income over the course of a year without having a volition cause a bodily movement that is in any way related to the earning of that income. Furthermore, one can earn income over the course of a year without doing anything whatsoever that would necessitate a volition.

(20) Moore, Act & Crime, p. 7.

(21) H. L. A. Hart, "Ascription of Responsibility and Rights," Proceedings of the Aristotelian Society, 49 (1949), pp. 171-94.

(22) H. L. A. Hart, Punishment and Responsibility, Oxford: Oxford University Press, 1968, p. 90.

(23) Id. at 106. It is important to note, however, that Hart thought that omissions warranted a treatment different from, but closely related to, that of positive acts. In what follows, I will offer a treatment of omissions based on the one Hart suggests. Even though I agree with Hart that there are important theoretical differences between positive acts and omissions, I think that one can profitably categorize both as "acts" in a very broad sense--as expressions of agency that serve as fundamental component parts of the complex actions proscribed under the actus reus requirement. The application of the excluder definition of voluntariness that I suggest here lumps positive acts and omissions under this broader understanding of "acts" in just this way.

(24) A. P. Simester, "Why Omissions Are Special," Legal Theory 1 (1995), p. 327. The terms "positive actions" and "omissions" inserted here in brackets are the ones from my discussion that approximate most closely Simester's terms "doings" and "not-doings".

(25) Moore, Act & Crime, p. 59.

(26) Moore, "Moore on Act & Crime," p. 1821.

(27) Moore, Act & Crime, p. 267.

(28) Id., pp. 267-8; for Moore's attack on this understanding of causation, see pp. 267-74.

(29) Perhaps Moore's strongest statement to this effect appears in his response to his critics in the University of Pennsylvania School of Law Symposium. There he responds to Fletcher by arguing that omissions should not incur liability under statutes that require the causing of some result. He states: "Fletcher is right about our law; our law is simply wrong about the metaphysics, as is Fletcher. Anglo-American criminal law cannot make omissions be causes when they are not." Moore, "Moore on Act & Crime," p. 1785.

(30) Moore, Act & Crime, p. 34; and Moore, "More on Act & Crime," p. 1781.

(31) Fletcher, "On the Moral Irrelevance of Bodily Movements" p. 1449.

(32) Moore, "More on Act & Crime," p. 1795.

(33) Were Moore to disagree with this claim, it would seem that the existence and legitimacy of such statutes would weaken considerably his overall claim that action is a central and unifying concept in the criminal law.

(34) I take it that the seriousness of the injustice is not to be measured by the relative seriousness of the consequences that would ensue as a result of failure to uphold the duty, but rather by the amount and degree of responsibility one assumes in taking on the positive duty in question. Thus, failure to take adequate care of one's children is sufficiently serious not because harm to children is in itself any worse than harm to adults, but rather because children are especially vulnerable and dependent upon their parents.

(35) Moore seems to grant something similar to, but short of, this claim when he claims that the duty to care for one's own children is as stringent a duty as just about any other negative duty.

(36) Moore, Act & Crime, p. 189.

(37) Fletcher, "On the Moral Irrelevance of Bodily Movements," p. 1444.

(38) Corrado, "Is There an Act Requirement in the Criminal Law?" pp. 1556-61.

(39) Id., p. 1560.

(40) Here, I disagree with action theorists such as Jennifer Hornsby, who maintain that action theory is so underdeveloped that we should not be surprised that it does not do for us what we expect it to do. See Jennifer Hornsby, "The Poverty of Action Theory," Philosophical Inquiry 21 (1999), p. 17. I think instead that given its role in the criminal law, especially at the level of basic acts and voluntariness, action theory is no more developed than it need be--and it need not be terribly well-developed to address nearly all of our questions of voluntariness, provided the act requirement is understood as I have suggested here that it be understood, namely, as a sub-requirement of the broader actus reus requirement.

Stephen Mathis is assistant professor of philosophy at Wheaton College, Norton, MA.
COPYRIGHT 2003 Institute for Criminal Justice Ethics
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2003 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Mathis, Stephen
Publication:Criminal Justice Ethics
Geographic Code:1USA
Date:Jun 22, 2003
Previous Article:Between madness and death: the medicate-to-execute controversy.
Next Article:Policing the brass: a case study in command malfeasance.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters