This post contains Part 6 (“Moral Responsibility”) of my dissertation prospectus, written in pursuit of my Ph.D in philosophy at the University of Colorado at Boulder and submitted to my committee in early December 2007. The full prospectus is available in PDF format and as an MS Word file. Comments and questions are welcome. While they won’t change the prospectus, they might be of use as I write the dissertation over the next year.
As already noted, the basic purpose of a theory of moral responsibility is to determine that for which a person is properly moral judged. Since morality presupposes voluntary acts, a theory of responsibility must identify the essential qualities of all voluntary actions. Those criteria were originally defined by Aristotle in the Nicomachean Ethics, Book 3, Chapters 1-5. Aristotle’s explicit purpose in those chapters on moral responsibility is to aid proper moral judgment. He observes that properly bestowing “praise and blame” on “voluntary passions and actions” and “forgiveness and also sometimes pity” on involuntary passions and actions presupposes that we can “distinguish the voluntary and the involuntary.” While wrong or incomplete in some details, Aristotle’s control and epistemic conditions for voluntary action provide the basic outlines of a theory of moral responsibility consistent with the purpose of moral judgment and the nature of human agency.
In this section, we will consider only the requirements of voluntary action. Responsibility for products and character will be discussed with resultant and constitutive luck, respectively. So what does moral responsibility for actions require?
First and most obviously, a person must control his actions to be morally responsible for them. For Aristotle, that control condition means something very specific, namely that the action originates from within the agent himself, such that he has the power to do or not do the action. Voluntary actions cannot be forced upon an agent; they must be the product of the agent’s own powers of self-direction. That control over actions is found in ordinary bodily movements, e.g., answering the phone or not, standing up or not, turning on the television or not. It is also found in cognitive processes: as a self-reflective agent, a person is capable of identifying, evaluating, and directing some of his own mental processes. For example, a person can choose to exert the effort of thinking or not, to think about some issue or not, to accept some argument or not, to confront painful facts or not, to trust gut feelings or not, and so on. Since actions originate in thought, that control over mental processes is the necessary foundation for control over bodily movements.
The most obvious cases of failure of control are movements produced by irresistible external powers, as in Aristotle’s cases of the man spirited away by kidnappers or knocked over by the wind. A person’s uncontrolled bodily movements also might be purely physiological responses, like the secretion of bile by the gallbladder or twitching caused by a brain tumor. Less obvious is the proper analysis of actions commonly described as “forced on a person by circumstances”–such as when a tyrant orders evil acts upon pain of death of one’s family or when a ship captain jettisons his cargo in a storm to save the ship. As Aristotle observes, such actions are properly regarded as voluntary because they are “worthy of choice at the time when they are done” and “the end of an action is relative to the occasion.” So the person’s ultimate regret does not prove his action to be involuntary because the actions were chosen at the time from amongst the alternatives available at the time–and chosen rightly, in the case of the ship captain. In general, since a person’s possible actions are always constrained by his particular circumstances, actions must be judged as voluntary or not within the context of those circumstances. So when a person must choose between two undesirable alternatives–like between total shipwreck and merely lost cargo–he selects his course voluntarily, even if regretfully. That’s why Aristotle observes that the terms “voluntary” and “involuntary” should be used “with reference to the moment of action,” rather than by any comparison to more favorable circumstances.
Significantly, the control condition sketched by Aristotle does not demand power over all aspects of an action–as does Nagel’s control condition. Instead, the morally responsible agent must simply have the power to regulate his own actions, in the sense that he has the power to act or not. Moreover, that control condition is not an intuition from nowhere, as Nagel supposes. It is grounded in basic facts of human agency, particularly that people are capable of regulating some (but not all) of their bodily movements. Where such self-regulation is possible, a person can choose to do or not do anything physically possible to him. He is then justly held responsible for that choice: he can be praised for acting as he ought or blamed for acting as he ought not.
Second and less obviously, a person must act with adequate knowledge of his actions to be morally responsible for them. The agent must be aware of “the particular circumstances of the action,” such as “who he is, what he is doing, what or whom he is acting on, and sometimes also what (e.g., what instrument) he is doing it with, and to what end (e.g., for safety), and how he is doing it (e.g., whether gently or violently).” So if Cindy slaps her friend Joe on the back, not realizing that his shirt hides a sensitive sunburn, her action is not voluntary due to her ignorance of that crucial fact. More precisely, she does voluntarily slap Joe on the back but does not voluntarily slap Joe on his sunburn. Given her ignorance, Cindy could not know that in doing the former she is also doing the latter. That’s why she will be horrified to learn what she’s done when Joe cringes in pain and shows her his sunburn.
The basic justification for this epistemic condition for moral responsibility is that when a person acts on the basis of a faulty understanding of the circumstances, his action is not what he supposes it to be. Since the person thought he was doing X when he actually did Y, he did not do Y voluntarily. While such mistaken actions may be evaluated, those evaluations would not constitute genuine moral judgments. Moral judgments seek to identify the basic principles and values by which a person governs his actions. Yet a person’s actions also depend on commonplace beliefs about the particular circumstances of the action, e.g., whether the shaker contains sugar or salt, whether that’s Joe or John across the room, whether Fanny enjoys or laments teasing about her name. When a person errs in those ordinary beliefs, his outward actions reveal little to nothing about his basic principles and values. Cindy’s ignorance of Joe’s sunburn, for example, makes her back-slap perfectly consistent with her claims of deep affection for him. The same cannot be said if she was aware of his sunburn at the moment of her slap.
However, a person’s ignorance of circumstances does not always render his actions involuntary. That depends on whether the person regrets his action or not, per Aristotle’s critical distinction between “involuntary” and “non-voluntary” acts. Aristotle describes mistake-based actions as “involuntary” only when contrary to the wishes of the agent, such that he regrets the action and would have done otherwise if he’d known the relevant facts.  However, often such factual errors are of little to no practical significance to the agent. If I carry a bag of dog food into the house, the fact that I might think it to be fifteen pounds rather than ten doesn’t make my action involuntary. A person might even be pleased by an error due to some unexpected benefit, such as when a thief steals a silver goblet thinking it to be a tin cup. In such cases, the person does not act according to his particular intention (i.e. voluntarily) nor contrary to his general preferences (i.e., involuntarily), so Aristotle classifies the action as “non-voluntary.” Aristotle never considers the question of responsibility for such non-voluntary actions. However, because the person’s action does not substantially depend on his mistaken belief, the action is properly regarded as near-voluntary. The agent’s lack of regret constitutes an endorsement of the action, so he is properly held responsible for it. Consequently, the epistemic condition should be understood as removing moral responsibility only when the agent is mistaken about some fact of significance to him, such that he would have acted differently if he had known it.
Moreover, the epistemic condition only requires that the agent know the particular facts relevant to his action; ignorance of the relevant general principles (or universals, to use Aristotle’s term) or a mistake in their application does not render the agent’s action less than voluntary. Contrary to the suggestions of Aristotle and Aquinas, the reason is not that ignorance of the proper general principles of morality is always culpable. A person could be innocently ignorant of or mistaken about some general principle relevant to his action, e.g., whether parents should stay in a miserable marriage for the sake of the children. Yet that could only make the action excusable or understandable. It would still be voluntary because the person would not be wrong about the nature of his action, as with ignorance of particulars, but only about the propriety or wisdom thereof. That a person is ignorant of or mistaken about proper moral principles is relevant to our moral assessment of him, even if not always blameworthy in itself.
Worrisomely, the control and epistemic conditions might seem to permit a vicious person to evade responsibility for any bad act whatsoever by deliberately rendering himself ignorant and/or incapable. So the woman who steadfastly refuses to hear her daughter’s desperate hints for protection against the sexual advances of her new husband could not be blamed for leaving them alone together for a week because she wouldn’t know what her husband would do to her daughter. Similarly, a husband might evade responsibility for picking the kids up from school as promised simply by drifting into a nap because he can’t control when he wakes up once he’s asleep. Thankfully, that seemingly straightforward application of the control and epistemic conditions is completely wrong. A person is properly held responsible for his actions when ignorant or incapable–when he voluntary places himself in that condition.
Although usually unnoticed, people routinely render themselves incapable or ignorant in various ways. Many cases thereof are morally blameless if not virtuous: the ignorance or incapacity is an insignificant side effect of the agent’s pursuit of his legitimate ends or a means to those ends. For example, if Mary chooses to study economics rather than psychology, then she might never learn the difference between anorexia and bulimia. Similarly, if Jane doesn’t buy ice cream at the grocery store, then she’s better able to stick to her diet because she can’t indulge in those delicious calories in the wee hours of the night. In these two cases, as in countless others, the ignorance and incapacity are voluntary. The control condition is satisfied because the person has the capacity to do otherwise, e.g., to study psychology, to buy ice cream. The epistemic condition is satisfied because the person knows the circumstances of his action, e.g., that not studying psychology will entail knowing less about the subject, that not buying ice cream will preclude eating it at home. Speaking generally, a person ought to be concerned for the possibilities for future action and for future learning foreclosed by the pursuit of one course rather than another. That’s part of the active concern for the future required for flourishing.
A person’s voluntary incapacity or ignorance can be morally blameworthy, however. For example, a father who breaks his promise to attend his daughter’s basketball game because he chose to leave town on a last-minute fishing trip with his buddies is culpable for his absence, even though incapable of attending once out on the lake. Similarly, a student is properly blamed for his wrong answers to exam questions if he opted to sleep in class and party rather than study. In those cases, the person’s incapacity and ignorance is of his own doing and damaging to his chosen ends–and that’s why he’s properly blamed for his current state and its results. Notably, this analysis of voluntary incapacity and voluntary ignorance parallels the proper understanding of responsibility for outcomes: if I throw a stone at a window, knowing that the glass will shatter if hit, I cannot rightly deny responsibility on the grounds that I was unable to stop the stone after it left my hand. The analysis is also consistent with Aristotle’s understanding of the control and epistemic conditions, particularly with his discussion of culpable ignorance. Ultimately then, a person who voluntarily renders himself incapable or ignorant is responsible for his actions in that state, for better or worse.
Our discussion so far does not exhaust the complexities of moral responsibility. However, it provides a general framework for discussion of the proposed categories of resultant, circumstantial, and constitutive moral luck.
 Aristotle NE, 1109b30-5.
 Aristotle NE, 1109b35-1110a2, 1110a17.
 Binswanger 1991, p. 156.
 Aristotle NE, 1109b35-1110a3.
 Aristotle NE, 1110a4-11.
 Aristotle NE, 1110a11-13.
 Aristotle NE, 1110a13-4. This general principle will be critical to the proper analysis of circumstantial moral luck. Also, dire circumstances may influence the substance of our moral appraisals of actions and agents. So a man tortured by a tyrant may be praised for “endur[ing] something base or painful in return for great and noble objects gained” if he resists or forgiven for doing “what he ought not under pressure which overstrains human nature and which no one could withstand” if he succumbs (Aristotle NE, 1110b1-3).
 Aristotle NE, 1111a3-5, 1111a 24.
 My analysis seems similar to the view that an action may be intentional under one description but not under other descriptions, as in Davidson 2001, pp. 43-51.
 Aristotle NE, 1110b16-24. Broadie uses the more clear term “countervoluntary” in place of “involuntary” (1991, p. 126).
 Bostock argues against the whole category of non-voluntary actions on the grounds that after-the-fact feelings are irrelevant to moral culpability (Bostock 2000, pp. 111-2). He claims that “if the act was due to ignorance, and ignorance which is not itself blamable, then clearly the agent cannot be blamed for it, whether or not he afterward regrets it” (Bostock 2000, p. 111). Yet that would render any action deviating slightly from the agent’s plans involuntary and blameless, so long as the deviation was due to some non-culpable ignorance. For example, if a hit man’s attempted strangulation of his victim caused a fatal heart attack instead of suffocation due to an undiagnosed heart condition, that would render the hit man blameless for the death of his victim. Like Urmson, I regard the “vexation” (1110b20) and “regret” (1111a21) felt after involuntary actions not as retroactively changing the nature of the action but rather showing that the action was directly contrary to (as opposed to merely inconsistent with) the motivating intention (Urmson 1988, p. 46).
 Aquinas 1993, #408.
 Aristotle NE, 1110b16-24.
 Once again, contemporary debates about actions intentional under some descriptions but not others promise to shed some light on the proper analysis of these kinds of cases.
 Urmson 1988, pp. 45-6 has a helpful discussion of this issue.
 Aristotle NE, 1110b33; Aquinas 1993, #411.
 Aristotle NE, 1114a13-22 uses this example in a somewhat different context.
 Aristotle NE, 1110b25-28, 1113b31-1114a3.