1 Introduction

In “Control, Responsibility, and Moral Assessment” Angela Smith defends her rational relations view of moral responsibility against the criticism that, in the absence of voluntarism, judgments of moral responsibility are shallow (2006). On Smith’s account, moral responsibility for both the voluntary and the nonvoluntary is to be accounted for in terms of two considerations pertaining to moral criticism: 1) an agent’s rational activity, and 2) the nature of moral demand (18). Regarding the first, a competent moral agent’s desires, emotions, beliefs, and character traits, as well as her actions, reveal her rational activity, and this is a proper basis for judgments of moral responsibility. As for the second, moral criticism addresses moral demands to its targets. The upshot is that, on Smith’s view, what matters is how an agent judges, not how she acts. Of course, one way in which it is revealed how an agent judges is through action, so Smith accounts for moral responsibility for voluntary undertakings under the same umbrella as she does for the nonvoluntary. The salient point, however, is that judgment is explanatorily basic for Smith, and actions are candidates for responsibility only to the extent that judgment is revealed in them.

Smith’s view is similar to other attributivist views in that her account grounds moral responsibility in the attribution to an agent of salient moral qualities even in the absence of any voluntary undertaking.Footnote 1

In this paper, I shall restrict my attention to Smith’s formulation, though much of what I say will be applicable more generally.

2 Voluntarism & the charge of shallowness

Smith offers her view in opposition to voluntarism. In characterizing voluntarism, Smith sometimes speaks of what an agent has voluntary control over (2). At other points, she characterizes the view as restricted to what an agent has chosen (3), and still, elsewhere, to what an agent has voluntarily chosen (4). At one point she characterizes the view in terms of deliberate choices (19). Finally, in a footnote, she suggests that voluntarism is the view that responsibility is restricted to whatever an agent has direct control over (n.42, 34). Depending upon one’s theory of action, Smith’s varying descriptions might capture very different views. For example, not every choice needs to be understood as deliberate, especially if one takes “deliberate choice” to mean a choice that is the upshot of deliberation. Also, what one has direct control over might differ from what one has voluntary control over. (It could be argued that one has voluntary control only over that which involves an exercise of voluntary agency, but that one has direct control over anything that could be influenced by an exercise of voluntary agency.) Furthermore, the range of actions for which an agent is morally responsible could be construed to be much wider than the range of actions that involve choices.

I do not call attention to this variance in Smith’s characterizations of the voluntarist position as a criticism of her treatment of it. Her target is a constellation of views that are loosely associated with a thesis about control, agency, and moral responsibility.Footnote 2

Thus, she has placed in her sights a paradigmatic formulation of the view that treats the voluntary as involving a species of direct control over action (presumably when the control is properly functioning), that takes the voluntary to be located in choice, and that conceives of choice as involving deliberate undertakings. She then applies her arguments to this paradigmatic formulation. Her arguments, however, are applicable to a less restricted form of voluntarism, and I think it is useful to consider her efforts in light of a wider conception of it. Hence, in the remainder of this paper, I shall understand voluntarism to be the following thesis:

The only objects of direct moral responsibility are free actions, where free entails all that is required for the control condition for moral responsibility.

I include amongst free actions, mental acts such as choices and decisions. But I do not restrict free actions to mental acts.Footnote 3 Also, I understand direct moral responsibility to differ from derivative moral responsibility in that objects of direct moral responsibility are not to be explained in terms of non-deviant causal relations with other items for which an agent is morally responsible. My formulation of voluntarism here is more inclusive than the one Smith seems to have in mind, but I think it shares the central features that are at the heart of the dispute between Smith and her adversaries. In essence, voluntarism about moral responsibility is a thesis about control and its scope, which is limited to intentional actions.Footnote 4

To defend her non-voluntarist rational relations view against the charge of shallowness, Smith examines Susan Wolf’s indictment of real self views (1990, 40–44). Wolf rejects real self views since 1) the responses they license involve only descriptions—mere grading—and not a basis for morally judging; 2) they provide no basis for credit or fault; and 3) they do not legitimize judgments that morally reactive attitudes are appropriate.

According to Smith, Wolf’s indictment applies to some real self views, such as J.J.C. Smart’s (1961), but not to hers. Each of Wolf’s complaints about the real self view misses the mark when applied to Smith’s rational relations view. Consider a case in which it is granted that there is nothing that an agent does, but in which her attitude towards others reveals a moral judgment of hers, one that fails to live up to what we owe to others. Smith gives the example of E.M. Forester’s Margaret and Henry from Howard’s End (1910). Smith quotes Margaret’s exasperated indictment put to Henry:

Not any more of this!” she cried. “You shall see the connection if it kills you, Henry! You have had a mistress – I forgave you. My sister has a lover – you drive her from the house. Do you see the connection? Stupid, hypocritical, cruel – oh, contemptible! – a man who insults his wife when she’s alive and cants with her memory when she’s dead. A man who ruins a woman for his pleasure, and casts her off to ruin other men. And gives bad financial advice, and then says he is not responsible. These, man, are you. You can’t recognize them because you cannot connect. (1910, 243–4)

According to Margaret, Henry’s moral fault is his failure to recognize his hypocrisy and his own moral shortcomings. His blameworthiness, Smith explains, is grounded in his moral judgments, not in anything he has done (21).

Smith calls to our attention how frequently we hold people morally responsible for things that are not the result of any voluntary undertaking: forgetting an anniversary, unrecognized insensitivity, humor found in the discomfort of others, and so on. In all of these cases, Smith maintains, we can account for an agent’s moral responsibility without tracing that responsibility to some distil free action. And in doing so, we do not fall prey to Wolf’s three complaints about (other) deep self views. So, for example, 1) Margaret’s response to Henry goes beyond mere grading and involves morally judging since her response is to a moral standpoint Henry adopts, and so the depth found in his responsibility is located in his rational agency; 2) Henry can be seen to be at fault for the moral judgment implicit in his regard for others since, as a rational agent, he is sensitive to moral reasons; and 3) it is perfectly fitting for Margaret to respond to Henry’s moral stance as one that falls short of recognizing reasonable moral demands.

Smith considers and rejects Gary Watson’s efforts to attenuate Wolf’s indictment of at least some real self views (Watson 1996). Wolf holds that on a real self view, when an agent’s action reveals a moral quality, simply reporting on that moral quality is a shallow manner of evaluation unless it is tied to an agent’s control over that very feature of the agent’s self (Wolf 1990, 41). It is shallow, Wolf holds, since it is a mere report, a description of the person. Watson argues to the contrary that such appraisals of character, aretaic appraisals, are not shallow but rather deep since they involve self-disclosure (Watson 1996, 230). According to Watson, when we call to attention features of an agent’s character as revealed in her action, such as an act that is shoddy, we are expressing a judgment of moral responsibility that involves depth, since we are acknowledging a moral feature of the agent’s self. This manner of moral evaluation, Watson holds, involves a “face” of moral responsibility that differs from the face that concerns accountability. Accountability, according to Watson licenses full-blooded blame in that it legitimates certain ways of treating the agent held to be morally accountable, especially by way of censure (Watson 1996, 230-7). Aretaic appraisal falls short of this threshold.

Smith’s rejection of Watson’s proposal is telling, since it casts further light on her commitments to the nonvoluntarist position. Watson’s resistance to Wolf’s indictment of real self views will not suffice for Smith since the moral evaluations highlighted on an aretaic construal will not justify blame. As Smith sees it, an agent can be fully morally responsible even for nonvoluntary features of her character, where this involves the propriety of moral blame no less than the propriety that is fitting for any free action (14–16). Indeed, Smith wishes to claim that Watson’s notion of reactive entitlement is fully consistent with her nonvoluntarist position. As Watson understands reactive entitlement, one holding another morally responsible is entitled to a particular style of moral reaction in response to the blameworthy agent (1996, 230). On Smith’s account, the very acknowledgement of a moral fault in a person carries with it the depth involved in any judgment of moral responsibility. One’s being entitled to hold another morally responsible is a matter of one’s being in an appropriate position to morally evaluate another by, for example, characterizing the other’s action as shoddy. If one is so situated, she is entitled to her response insofar as she can express moral demands to the other regarding what we owe to each other.

3 Evaluating Smith’s reply to the challenge of shallowness

Does Smith offer an adequate reply to Wolf? Let us keep our sights restricted to cases where, as Smith would have it, there is no voluntary undertaking, but the agent is morally responsible, such as the case of Henry and Margaret. In the spirit of voluntarism, one might resist Smith as follows. Smith is entitled to her claims about responsibility for the nonvoluntary, but only when the relevant nonvoluntary object of responsibility was the upshot of distil causes that included prior free actions. So, for example, it is right to indict Henry for his hypocrisy—that is, to judge him—only if Henry acted freely earlier in his life in ways that lead to his hypocrisy.

To test the voluntarist reply as set out above, one might vary the case. Suppose Henry* came to his moral outlook by way of a history of extreme psychological abuse as a child, an abuse that settled for him the values and character traits that account for his current moral blind spot. Suppose also that Margaret* was aware of this history. Now is it so clear that Margaret*’s response to Henry* is fitting? The voluntarist will say that it is not, and that this shows that in other cases of the sort Smith mentions, implicit in our judgments of moral responsibility is the background belief that an agent’s responsibility is anchored in a history in which it is his or her fault for coming to be the person who is responsible in these nonvoluntary contexts.

Aware of this objection, Smith offers the following reply:

When we criticize a person for being unforgiving, for example, it seems to me we are responding to something implicit in the attitude itself, not to facts about its possible origin in a person’s prior voluntary choices. (21)

But the voluntarist need not be committed to the thesis that when we hold responsible for a morally objectionable attitude, what we are responding to is the history as well. Our response, just like on Smith’s rational relations view, is to the attitude. It is just that, on the voluntarist view, so construed, a necessary condition for the propriety of our response is that there be this sort of history in place. And, Smith’s opponent might continue, the case of Henry* and Margaret* is the test that proves the point.

So, I find Smith’s response to the voluntarist insufficient to silence their objection to her view. Regardless, I think her position is compelling, and I am inclined to explore it further, despite the fact that it conflicts with a long-held view of my own. To this end, I shall briefly consider two reasons to support Smith against the voluntarist thesis. First, in my estimation, most voluntarists are not particularly realistic about the practical details of how choices earlier in life can have repercussions later in life.Footnote 5 Let us call these theorists tracing voluntarists. The tracing voluntarist, fixated on the control condition for moral responsibility, hopes to link a credible basis for our judgment about attitudes and character traits to earlier free actions. But another condition on moral responsibility is an epistemic one. Consider the young boy who consciously chooses to cultivate his aggression and thick skin in order to survive in the locker room during his junior varsity football days. He will have very little reason to expect that it will someday be the source of his coolness and tragic distance from his own children. In general, so far as various aspects of our own character traits are concerned, we are often the hapless victims of our own unwitting earlier free choices. We do not possess the foresight of gods, and so who we will become, from the vantage point of who we are, is sometimes just a crap shoot.

The tracing voluntarist can reply with consistency here and simply say that in all of these cases, we are not responsible for the matters at issue. The tracing voluntarist never meant to claim that only a history rooted in free action was required. Indeed. But now the tracing voluntarists’ view is highly restrictive, since many cases in which an agent is regarded as responsible for attitudes and character traits arise in the manner to which I am calling attention here. At this point the tracing voluntarist has a choice: either commit to a modest revisionist account of the proper objects of moral responsibility (denying that many of our moral responsibility judgments about attitudes and character traits are properly grounded), or instead give up the voluntarist requirement.

Here is a second point in Smith’s favor. She can account for some cases that seem to confirm the tracing voluntarists demands. In some cases, Smith can hold, a person can be morally responsible for making herself into the person who now harbors these morally objectionable attitudes. In these cases, the person is morally responsible for the history that resulted in the acquisition of these attitudes. But, Smith can continue, in other cases, most cases I am suggesting, even if a person is not morally responsible for making herself to be as she is, it does not follow that she is not responsible for the relevant character traits, attitudes, and so on. Thus, Smith can distinguish between responsibility for the attitude and responsibility for the acquisition of it.Footnote 6

The tracing voluntarist will retort: “What about the case of Henry* and Margaret*? Doesn’t this case speak against the view I am offering in Smith’s support?” No. The case is under-described. Was Henry*’s history such that he is now impaired in such a way that he is unable to assess his moral outlook rationally, even when it is called to his attention? If so, then, consistent with Smith’s thesis, Henry* is not able to engage in rational activity in this domain, and so is not responsible. If, on the other hand, Henry* was forced to have these attitudes or character traits through this history but he is still able to assess them rationally, to engage in rational activity with respect to them, then the judgments that are revealed in them are subject to Smith’s treatment, and Henry* remains responsible.

4 Moral criticism and moral blame

Smith appears to have a thoughtful reply to the charge of shallowness as it issues from her voluntarist opponent. And, in my estimation, she has some footing in our intuitions to support the thesis that we often do hold people morally responsible for more that what is within the confines of the voluntarists’ scope, such as racist attitudes, inconsiderate regard for spouses or other loved ones, and desires to witness the suffering of others. However, in her zeal to defend her nonvoluntarist position, I believe that Smith has reached farther than she needs to reach to make her case. If I understand her correctly, on her view, moral criticism entails moral blame. This is why, for Smith, there is no good sense to be made of Watson’s efforts to distinguish between an aretaic notion of responsibility and an accountability notion. Aretaic judgments (in the negative case) involve moral criticism, and moral criticism carries all that is involved in blaming.

Smith adopts a similar posture toward R. Jay Wallace when considering Wallace’s (1996) rejection of a near cousin to Smith’s theory of moral responsibility, which is T.M. Scanlon’s theory (1988). Wallace argues against Scanlon that moral blame should be limited to choices or decisions, and so the special force of blame cannot be explained in what an agent believes, but in what she does as this arises from (presumably free) choice or decision (Wallace 1996, 80–81). In response to Wallace, Smith holds that whether an agent is responsible for some X in the manner that she or Scanlon have in mind is a matter of whether or not the person can legitimately be asked to justify that X for which she is being assessed (23). At this point, and elsewhere in her paper, Smith does not offer much in the way of an account of when it is and when it is not legitimate to ask or demand a justification. But it seems that her view involves the claim that a person’s morally objectionable judgment is sufficient for the legitimacy of such a request or demand.

The above reading of Smith is suggested in her resistance to the relevance of the distinction Neil Levy has highlighted between bad agents and blameworthy agents (2005). To illustrate when ‘bad’ is applicable and ‘blameworthy’ not, Smith gives examples of, for example, children before the age of mature reason, or the paranoid schizophrenic in a moment of severe delusion. But when it comes to any agent whose conduct can be morally evaluated and accurately described in any negative light, the mere negative moral description, Smith holds, involves making moral demands of the person so described, and this demand, Smith writes, “by its very nature implies responsibility” (28).

Here I think Smith commits to more than she should. If she wants her view to jibe with a relevant range of cases, she should be able to account for a person capable of a level of moral understanding so that the person can be fully aware of a morally evil deed, but yet be impaired in some manner so that the person violates some condition of morally responsible agency. Not all bad agents who are not morally responsible agents can be accounted for by lumping them in with children or hopeless schizophrenics. (Think, for example, of the character Hannibal Lecter.) Some will come out to have sufficient moral understanding to display morally objectionable qualities, but will display some defect of agency that will render them non-responsible. My criticism of Smith’s treatment of morally bad agents is not intended as an indictment of her nonvoluntary position. It is merely meant to show that there is a wrinkle in her view that needs to be worked out, a point to which I will return, if only briefly.

5 Limiting voluntarism: a compromise view

So far, I have tried to show that Angela Smith’s rational relations view can handle Wolf’s charge of shallowness without falling back on the assumption of voluntarism. But is Smith right that the voluntarist is wrong? Smith claims that her rational relations view is anchored in an agent’s rational activity, something that, it is suggested, is revealed in, for example, Henry’s judgments. But is this activity the sort that involves mental acts, perhaps, choices or decisions about what to think? If so, it is unclear that her view differs from a version of voluntarism.

Smith, however, certainly does not have in mind by rational activity anything as full-blooded as mental action. She writes:

…I do not think legitimate moral criticism presupposes that an agent has voluntarily chosen (or has voluntary control over) that for which she is criticized. What matters, rather, is whether the thing for which an agent is criticized reflects her judgments, because the justificatory demand implicit in moral criticism is a demand to reassess, modify, and in some cases apologize for those judgments. (16)

Given this passage, it is clear that the rational activity that Smith finds in the judgments at issue do not arise from mental action. In some cases, this is just transparent. Some people, through sheer negligence, simply passively accept the racist views of their community without having ever “chosen” to accept them, without ever having done anything that brought them to that state. What, it seems to me, Smith has in mind by rational activity is that an agent’s attitudes and character traits are subject to her own rational assessment. Rational activity can be brought to bear on them. Still, we can press Smith here: What is meant by rational activity in the case in which it is engaged, is active and not merely an untapped disposition? Does it fall shy of action? It must, or Smith gives up the game to the voluntarist.

I will not pursue the question as to whether rational activity falls shy of action. It seems to me that, in many cases, certainly it does. It is natural in discussion of these issue to invoke differences regarding “direction of fit” between belief and desire. In acting, we seek to arrange the world so as to fit our desires. But in believing, at least in the standard case, we seek to alter our beliefs so as to accord with the way things are. An agent whose rational activity in the sphere of morality involves efforts to believe correctly and so judge correctly stands to the world in a different relation than an agent stands when she wishes to bring something about on the basis of what she desires. Furthermore, when our rational capacities are engaged, in at least a large range of cases, our coming to believe correctly is not regard as a voluntary matter. One does not decide to believe truths of mathematics, or a truthful report of some simple empirical fact, such as how many tomatoes are on the counter. One simply makes oneself disposed to the relevant information and, so to speak, comes to have the relevant beliefs. These considerations suggest that in numerous typical cases, rational activity is very different from action. But it is a delicate matter whether coming to proper judgments about moral matters fits so well with this picture of rational activity. I leave this as an open question, though I think it reasonable to grant that there is good sense to be made of a kind of rational activity that differs from full blooded action. This is all that Smith needs.

In closing, I propose an Angela Smith-inspired compromise position. On my proposal, one saves all that should be saved from the voluntarist view, which is a requirement of control. An agent can be morally responsible for her attitudes, character traits, or any other nonvoluntary object of responsibility only if she stands within the scope of her rational control. Here, rational control has two components. One involves the possibility of rational activity (that, let us grant, falls shy of free mental acts). A second involves a standing capacity to perform a free mental act of deciding or choosing to evaluate one’s moral standpoint(s). Although I will not develop the point, I would like to suggest that with this second condition Smith might be able to explain properly Levy’s distinction between morally bad and morally blameworthy agents. This might also give her room to grant to Watson that some agents can satisfy aretaic conditions of moral responsibility, even if they fall shy of accountability standards for moral responsibility.

My proposal, inspired by Smith, puts the lie on the control condition for moral responsibility. We are directly morally responsible on this view for states of our character or attitudes that we did not bring about through any free actions, and so, as Smith insists, we can be directly morally responsible for what we judge and not just what we do. And we can be morally responsible for things that are not the product of any exercise of free agency—that is, any exercise of what Smith might characterize as voluntary control. But a unifying requirement on moral responsibility is that control comes in somewhere. In the case of direct moral responsibility for the nonvoluntary, it comes in indirectly, via a (sometimes unexercised) capacity to decide freely to evaluate one’s moral standpoint (as regards the relevant nonvoluntary object of responsibility). Still, control does not come in everywhere. And this is a point Smith was at pains to defend.

Many of us working on issues of free will and moral responsibility are “control freaks” (me included). We want to inject control everywhere in all aspects of our moral responsibility judgments. But control has its place, and for us to inject it in certain places, such as in the voluntary acquisition of our moral attitudes and our character traits, is to set the bar too high for creatures like us. It reveals an unrealistic set of expectations about what we must be like. Lowering the bar in an important way, we can see that our moral responsibility judgments sometimes are directed at features of our own selves, features that we did not bring about through any exercise of active control in the form of free actions, and that are not themselves directly implicated in anything else we do, but only reveal how we regard others. So, on this point, Smith is correct. Not everything that we are directly morally responsible for involves control exercised by way of a voluntary undertaking. But everything that we are directly morally responsible for does involve our control, or at least that is my contention.