Printer Friendly

Resisting tracing's siren song.

IT IS AN UNFORTUNATELY FAMILIAR FACT that intoxication can lead people to act improperly, even criminally. Consider Elie Joseph Arsenault. (1) In June 1954, Arsenault shot and killed Harriet Hinckley, his drinking partner and intimate. The murder occurred at the end of a days-long drinking binge, after Arsenault and Hinckley had shared 11 pints of liquor accompanied by barbiturates. Arsenault remembered little--not shooting Hinckley, not telephoning the police, not confessing to the crime.

Incapacitated wrongdoers like Arsenault pose a problem for reasons-responsiveness accounts of moral responsibility. Those accounts are powerful and popular--it seems right that moral responsibility depends upon an agent's having the capacities to perceive and act upon moral reasons. But, while Arsenault's intoxication likely incapacitated him at the time of his crime, surely he should not be excused from blame. In order to address such cases, many reasons-responsiveness advocates include a tracing condition to supplement the ordinary conditions of responsibility. The intoxicated wrongdoer is blameworthy despite his incapacitation precisely because he is responsible for becoming incapacitated. We hold him responsible for his intoxicated wrongdoing by tracing back to his responsibility for becoming intoxicated. Arsenault was responsible for becoming intoxicated, and his responsibility for the later murder of Hinckley can be traced back to that earlier responsibility.

But not everyone has accepted tracing. One challenge notes that we do not need tracing to blame culpably incapacitated agents. We can hold these agents accountable for their behavior in becoming incapacitated and for the foreseeable consequences of that behavior, and we can do so without taking on the apparent costs of tracing. I go further--I claim that tracing gets things wrong. To show this, I consider a different sort of case: the Odysseus case. Odysseus incapacitated himself in order to sail safely past the Sirens. Arsenault's incapacity was improper; Odysseus' was not. And had things worked out poorly for Odysseus, had his incapacity led to some later wrongdoing, he would have been unlucky but not blameworthy. The core reasons-responsiveness account agrees, but tracing accounts expose unlucky Odysseus agents to blame. Since reasons-responsiveness responsibility appears to get us what we want (as urged by the first challenge) and tracing gets us verdicts that we do not want (as shown by my new challenge), we should reject tracing.

1. Ordinary Responsibility and the Motivation for Tracing


Tracing is often offered as a supplement to reasons-responsiveness accounts of moral responsibility. Following Gary Watson (1996), we can distinguish between two sorts of responsibility: attributability and accountability. When we hold someone responsible for some action in the attributability sense of responsibility, we make an aretaic judgment about them on the basis of their behavior. But here we are interested in responsibility as accountability. When we hold someone responsible for some action in the accountability sense, we are disposed to react to them in certain ways, in particular with blame or punishment for improper behavior. (2) Reasons-responsiveness accounts of responsibility are offered to explain the conditions of this accountability sort of responsibility.

Reasons-responsiveness accounts associate accountability with the ability to respond in the right way to moral reasons. Many reasons-responsiveness theorists then identify the ability to respond in the right way to moral reasons with two normative capacities: the cognitive capacity to discern moral reasons and the volitional capacity to act upon moral reasons. (3) While there is significant disagreement regarding exactly how to pick out the two capacities, the reasons-responsiveness theorists ground responsibility largely (or wholly) on the possession of these two normative capacities.

The reasons-responsiveness account fits well with thinking of responsibility as grounded in concerns about fairness. R. Jay Wallace (1994) picks out the conditions of blameworthiness by looking at when it would be fair to expect others to heed moral reasons, and David Brink and Dana Kay Nelkin (2013) pick out responsibility by considering the importance of the fair opportunity to avoid wrongdoing. Thinking of fairness can explain why we think that insanity excuses, why we think that duress sometimes excuses and sometimes mitigates, and the like. An agent who lacked the capacities to respond to the moral reasons at hand could not fairly be expected to respond to those moral reasons. But, for agents who are sensitive in the right way to moral reasons, it seems fair to demand that they take account of those moral reasons, and it seems fair to hold them responsible when they fail to do so.

A basic reasons-responsiveness account of responsibility claims that reasons-responsiveness is at least necessary (and maybe also sufficient) for moral responsibility. I will call such an account the ordinary-responsibility account. On an ordinary-responsibility account, an agent is responsible for a wrongdoing only if, when the agent committed the wrongdoing, the agent was able to discern and act upon the reasons that counted against committing the wrong. That is, responsibility requires that the ordinary conditions of reasons-responsiveness be met at the time of the wrongdoing. And, at least in the typical cases, reasons-responsiveness is also sufficient for responsibility. (4) If the agent could have understood that the action was wrong, and if the agent could have acted upon that understanding, then the agent is responsible for failing to do so.

On the ordinary-responsibility account, an agent's responsibility for some bit of wrongdoing depends upon facts about the agent at the time of the wrongdoing. If the agent was reasons-responsive at the time of the wrongdoing, then the agent is responsible for the wrongdoing, and if the agent was not reasons-responsive at the time of the wrongdoing, then the agent is not responsible for the wrongdoing. Of course, that the agent's history does not matter for questions of responsibility for wrongdoing does not entail that the agent's history does not bear on any aspect of the agent's responsibility. John Martin Fischer and Mark Ravizza's (1998) account of responsibility for consequences, for instance, is historicist. To determine whether an agent is responsible for some consequence, we must consider whether the agent had the requisite sort of control at some relevant point prior to the consequence. To be fully developed, the ordinary-responsibility account should have something to say about the conditions of responsibility for the consequences of one's actions, including cases in which one of the consequences of an action is some further wrongdoing. I will return to questions about responsibility for consequences later. But, while the broader account of responsibility might be mixed, the ordinary-responsibility account's treatment of responsibility for wrongdoing is ahistorical. (5)


But many reasons-responsiveness theorists deny that responsibility for wrongdoing is an ahistorical evaluation. Fischer and Ravizza, for instance, consider Max, a drunk driver. Max drinks so much that he is "almost oblivious to his surroundings" (1998: 49). Intoxicated, he attempts to drive home and, unfortunately, strikes and kills a child in a crosswalk. By assumption, Max's intoxication left him non-reasons-responsive, both when he decided to drive and when he struck and killed the child. This suggests that Max might be excused under the ordinary-responsibility account's ahistorical analysis of responsibility. Nonetheless, intuitively, he is to blame. Cases like Max's suggest that ordinary responsibility is explanatorily inadequate.

Fischer and Ravizza explain that drunk drivers like Max are responsible for their intoxicated behavior because they are responsible for becoming intoxicated. Fischer and Ravizza hold culpably incapacitated agents responsible for their culpably incapacitated wrongdoing by tracing responsibility for the wrongdoing back to responsibility for the prior behavior that led to the incapacity. As they explain:
   When one acts from a reasons-responsive mechanism at time T1, and
   one can reasonably be expected to know that so acting will (or may)
   lead to acting from an unresponsive mechanism at some later time
   T2, one can be held responsible for so acting at T2 (50).

Tracing allows us to hold an incapacitated agent responsible for her wrongdoing so long as there was some prior moment when the agent could act to avoid her incapacitation and could reasonably foresee her subsequent incapacity and wrongdoing.

Tracing offers to address the explanatory inadequacy that appeared to threaten ordinary responsibility. Return to Max, Fischer and Ravizza's drunk driver. By hypothesis, Max was not incapacitated at the time he was drinking, and he could reasonably have been expected to know that drinking to excess could lead him to act wrongly while incapacitated. (6) When he did later drive while intoxicated to the point of incapacity, we ground responsibility for the intoxicated driving by tracing back to his responsibility for drinking to the point of intoxication.

In addition to appearing to explain our intuitions about culpable-incapacity cases, tracing tracks the way courts have regarded intoxication as a defense. The Model Penal Code, prepared as an advisory guide for legislatures and courts, provides an affirmative defense to defendants whose intoxication interferes with their cognitive or volitional normative capacities, but it allows this defense only when the intoxication was not self-induced. (7) Applying similar reasoning, the Supreme Court of Minnesota addressed the responsibility of a hit-and-run driver who claimed to have been unwittingly incapacitated by his prescribed medicine. (8) The court ruled that the medicated defendant could be excused from responsibility if two conditions had been met: (a) the medication caused him to be temporarily insane and (b) he neither had known nor had good reason to know that it would have this effect. If, that is, the medication's effect had caught the defendant unaware, then his incapacity would be considered involuntary, and the medication could ground an affirmative defense. But if the defendant had been aware of the risk of intoxication and so been responsible for becoming incapacitated, then the medication's effect would provide no defense.

We might worry about the reasonable-expectation element of the tracing scheme. Tracing's extension of responsibility is usually constrained by foreseeability. We see this in Fischer and Ravizza's tracing scheme, for instance. As Manuel Vargas (2005) objects, this constraint robs tracing of much of its explanatory promise, since many of the cases in which tracing might seem to help are cases in which the later wrongdoing was not foreseeable at the time the agent constrained her own agency. For instance, Vargas describes a manager who, as a teenager, purposefully inculcates a cool but jerky persona, and later, acting on the jerky persona, unreflectively mistreats a number of employees. Vargas denies that we can appeal to tracing to hold the manager responsible because he denies that the mistreatment was foreseeable when the manager was a teenager.

What about the more standard tracing cases? Is vehicular homicide a foreseeable risk of social drinking? Fischer and Neal Tognazzini write that "Drunk-driving cases are unproblematic precisely because everyone knows (or at least should know) that too much alcohol will impair the ability to drive a car" (2009: 532). But its being foreseeable that the agent might become too intoxicated to drive safely is not the same as its being foreseeable that the agent might nonetheless attempt to drive. Fischer and Tognazzini further explain that the foreseeability constraint does not require that you know what your wrongdoing will be "in all its florid particularity," and so I here grant tracing advocates the assumption that the agent's later, untoward behavior was foreseeable at the earlier time, at least in the central culpable-incapacity cases that motivate the addition of tracing. (9)

We also should be careful about the cases we are considering. Fischer and Ravizza's Max is supposed to be wholly incapacitated, rendered functionally insane. However, many of the culpably incapacitated agents we actually confront are only partially incapacitated. As Douglas Husak (2012) explains, the capacities to reason will often be impaired by intoxication but rarely destroyed. If the typical drunk driver is only partially incapacitated, then the typical drunk driver remains partially reasons-responsive. And because drunk drivers are often partially reasons-responsive at the time they drive drunk, they are partially responsible for their drunk driving as wrongdoing even without tracing. If we are not careful, our intuitions about the rare fully incapacitated drunk driver could be influenced by our experiences with much more common partially incapacitated drunk drivers. But I will assume due care in this regard.

I will call the account of reasons-responsiveness responsibility supplemented by tracing the tracing account., noting that the tracing account includes both traced responsibility and ordinary responsibility.


Tracing skeptics like Matt King (2014), Andrew Khoury (2012), and Larry Alexander (2013) argue that the ordinary-responsibility account can account for cases such as Fischer and Ravizza's drunk driver even without tracing. As King explains, becoming intoxicated to the point of incapacity creates risk, and often that risk is unwarranted. When an agent creates an unwarranted risk, the agent is reckless if he is aware of the risk, and he is negligent if he is not aware of the risk but should be. It is a familiar feature of ordinary responsibility that we hold agents responsible for their reckless or negligent conduct and for the foreseeable consequences of that conduct, and so we can hold the culpably incapacitated agents responsible for their reckless or negligent conduct in becoming incapacitated as well as for the foreseeable consequences of their recklessness or negligence. We can hold drunk drivers responsible for acting improperly and for the foreseeable consequences of that improper action without needing tracing, and so ordinary responsibility can avoid the explanatory-inadequacy worry. (10)

But tracing's advocates have insisted that ordinary responsibility is not sufficient. Kevin Timpe contends that "it is hard to see ... how one could account for a drunk driver's being responsible for running over a pedestrian without a tracing clause" (2011: 12). According to Fischer and Ravizza, tracing is a "refinement" developed to address a "problem" for their reasons-responsiveness theory of responsibility (1998: 49). And Fischer later writes with Tognazzini that tracing was a "component [that] must be added to get a plausible theory of moral responsibility" (2009: 532; emphasis added). For tracing's advocates, tracing is a necessary addition to the theory of responsibility. Merely being responsible for the foreseeable consequences of some prior action is not sufficient.

To defend this position, the tracing advocate needs to show both that tracing makes a difference and that we should want our theory of responsibility to include that difference. In the rest of this paper, however, I will argue that tracing's advocates cannot achieve both of these goals. On the most plausible understanding of the sort of difference tracing might make, it is a difference we should reject.

2. Tracing and the Odysseus Cases 2.1

The tracing advocate needs to identify a substantive difference between the tracing account and the ordinary-responsibility account. Start by considering the formal differences. We can identify (at least) two sorts of responsibility that are formally distinguished by their objects--action responsibility and consequence responsibility--though this does not yet require that there be any substantive difference tracking this formal difference. The ordinary-responsibility account holds the culpably incapacitated agent responsible for the original incapacitating act [A.sub.1] as a bit of action (11) and for the culpably incapacitated act [A.sub.2] and any further harms (H) as consequences of the original incapacitating act:


The tracing account adds to ordinary responsibility that the culpably incapacitated agent is also responsible for the incapacitated action by virtue of tracing:


But we should remember that the tracing account's extension of tracing responsibility is supposed to supplement ordinary responsibility's account of responsibility for action. That means that tracing allows us to hold the agent responsible for both the original and the later actions qua actions, both with their concomitant consequences:


We can see that there are two formal differences between ordinary responsibility and tracing responsibility: (a) tracing duplicates some of the ordinary responsibility's objects of responsibility, since [A.sub.2] and H each appear twice in tracing's accounting of responsibility, and (b) tracing, but not ordinary responsibility, allows us to hold the agent responsible for [A.sub.2], the culpably incapacitated action, as an action instead of only as a consequence.

Does the duplication matter? (12) It is hard to decide this without first determining whether we are right to want tracing. Suppose that the duplication leads to increased blame. If tracing's duplication leads to excessive blame, then that tells against tracing. But if it is true that tracing is an appropriate addition to the reasons-responsiveness scheme, then tracing's duplication leads to the right degree of blame in the culpable-incapacity cases, and it is ordinary responsibility that has got things wrong, letting the culpably incapacitated agents off too leniently. And even this line of thinking takes as a given that the elements appearing twice heightens the degree of blame warranted, but that is an open, substantive question. So the mere fact of formal duplication does not tell against tracing.

My argument focuses instead on the second formal difference. Unlike ordinary responsibility, tracing allows us to hold the agent responsible for the culpably incapacitated action both as a consequence of the incapacitating action and as an action in its own right. But this formal difference matters only if there is a concomitant substantive difference between action responsibility and consequence responsibility. It is nearly axiomatic that being responsible for doing wrong can make an otherwise blameless person blameworthy. It can be appropriate to resent someone who has done wrong on the basis of that wrongdoing, even if they have otherwise acted appropriately, and it can be appropriate to punish someone who has done wrong on the basis of that wrongdoing, even if they have otherwise acted appropriately. (13) This exposure to blame is what it means to be accountable for a wrongful action.

What about consequence responsibility? It seems uncontroversial that being responsible for a bad consequence matters. It can obligate an agent to make repair, and it can make it appropriate for an agent to feel a special sort of regret. Being responsible for a bad consequence can also matter for an agent's blameworthiness, though this is more controversial. (14) It can increase the degree of an otherwise-blameworthy agent's blameworthiness, and it can change the scope of an agent's blameworthiness. However, being responsible for a bad consequence cannot render an otherwise blameless person blame worthy. It would be inappropriate to resent someone who has done nothing wrong, even if they have caused harm. That is, responsibility for consequences can affect the degree or scope of blameworthiness, but it cannot by itself affect the fact of blameworthiness. Only responsibility for a wrongdoing as an action can affect the fact of blameworthiness.

Consider a surgeon who performs a risky but appropriate surgery. All surgeries carry the risk that something will go wrong, even if the surgeon takes all appropriate precautions and makes no mistakes. Sometimes things just do not work out. Imagine a surgeon who performs a consented-to, warranted surgery competently, and yet the surgery results in disaster for the patient, even the patient's death. The death was a foreseeable result of the surgeon's behavior in performing the surgery, and the surgeon was responsible for her behavior in performing the surgery. So, in some sense, the surgeon might be responsible for the patient's death. However, the surgeon is not responsible in the accountability sense of moral responsibility. Intuitively, the surgeon is not blameworthy for the patient's death. It would be appropriate for the surgeon to feel a special sort of regret for being involved with the patient's death, and it might be appropriate for the surgeon to make some effort at repair or amends, perhaps toward the patient's family. But it would be inappropriate to blame, resent, or punish the surgeon. By contrast, consider a surgeon who performs a risky and inappropriate surgery. Luckily, the surgery is a success. Nonetheless, it seems appropriate for us to blame the surgeon. It is wrong to perform an inappropriate surgery. Being responsible for the wrongdoing is sufficient to expose the surgeon to blame, even if there are no further harms.

To reprise, one formal difference between tracing responsibility and ordinary responsibility is that tracing responsibility allows us to hold the culpably incapacitated agent responsible for the culpably incapacitated wrongdoing both as a consequence and as an action, whereas ordinary responsibility only allows us to hold the culpably incapacitated agent responsible for the culpably incapacitated wrongdoing as a consequence. But to meaningfully distinguish the two accounts, we need to identify a substantive difference tracking that formal difference. Since action responsibility can ground blameworthiness for otherwise innocent agents, tracing (but not ordinary responsibility) makes foreseeable incapacitated behavior sufficient to hold an otherwise blameless agent blameworthy. This creates the possibility of an extensional difference between ordinary responsibility and tracing responsibility. Next, I argue that tracing responsibility gets things wrong.


Recall Odysseus' encounter with the Sirens. Odysseus and his men were to sail past the Sirens on their return to Ithaca. Circe had warned Odysseus that anyone hearing the Sirens' song would be maddened by a desire to stay, never to return home. Odysseus had his men stuff their ears with wax. But Odysseus, wanting to hear the Sirens' song, had his men bind him to the ship's mast instead. When Odysseus and his men approached the Sirens, Odysseus heard their song, and he was filled with desire to stay with the Sirens. But he was bound to the mast, incapacitated, and his men would not unbind him, so he could not act upon his desire. Odysseus and his men passed safely.

Or consider a case Derek Parfit (1987) developed from Thomas Schelling's The Strategy of Conflict (1980). In that case, an armed robber threatens to kill an agent's children unless the agent unlocks a gold-filled safe. The agent knows that it would be irrational to provide the gold (since then the armed robber would kill the agent and her children to stop them from reporting the crime), and she also knows that it would be irrational to ignore the threat (since that would risk the robber killing one of the children to spur the agent to action). The best choice is to take a drug, "conveniently at hand," which would render the agent irrational. The agent's irrationality would leave the armed robber's threats ineffective, since the irrational agent would no longer be moved by concern for her children. The armed robber would hopefully recognize that and decide that his best remaining option would be to escape (presumably without harming the agent or her children, perhaps to minimize his criminal exposure). As Parfit acknowledges, there is a risk that the irrational agent would harm herself or her children during the period of her irrationality. But Parfit contends that it is still appropriate for the agent to cause herself to become irrational, since that risk is outweighed by the need to defuse the armed robber's threats. As Parfit explains, "On any plausible theory about rationality, it would be rational for me, in this case, to cause myself to become for a period irrational" (1987: 13).

In Odysseus' case and in Parfit's rational-irrationality case, the agents use their compromised agency as a tool. Both agents solve some problem -- how to experience the beauty of the Sirens' song without becoming its victim, and how to defuse the invader's threat--by giving up control. Although giving up control was risky, since things could have worked out poorly, the risk was justified. (15) And because both agents purposefully brought about their own incapacity, it was foreseeable to both agents that their behavior would lead to their incapacity. (16) Both Odysseus and Parfit's parent acted: a) competently, b) in a way that foreseeably led to the agent's own risky incapacitation, and yet c) morally appropriately. Call such cases Odysseus cases. Odysseus agents act in a way that foreseeably (and sometimes purposefully) leads to their own incapacitation, and they do so while they are competent. Accordingly, they are responsible for incapacitating themselves. Unlike the culpablein-capacity cases, however, the Odysseus agents are not blameworthy for incapacitating themselves.

Homer's and Parfit's cases are fantastic and fictional. But there are also ordinary Odysseus cases. Going to sleep presents an Odysseus case. Being asleep is risky. The sleeping agent is unaware of his surroundings, unaware of risks that might present themselves, and unable to react. But, at least in normal circumstances, those risks are slight and outweighed by the benefits of sleep. Similarly, becoming medically incapacitated is risky. Being sedated or anesthetized entails giving up control, and that presents some risk. However, anesthetic and sedation are important and valuable elements of modern medicine, and the risks they present are usually outweighed by the benefits they offer. These agents who tie themselves to masts, take irrationality pills, go to sleep, take sedatives, or the like willingly incapacitate themselves, but they do not do so culpably. So Odysseus cases are a feature of our ordinary lives, not merely a philosopher's construction.


The test case we need to distinguish tracing from ordinary responsibility is a special sort of Odysseus case. In addition to the incapacitation's being non-culpable, two further conditions must be met. First, unlike Odysseus himself, whose incapacitation was external, the test agent's incapacitation should be internal, arising because the agent's normative capacities are compromised. It is easy enough to imagine some medications working this way, such as Parfit's convenient pill or the physician's sedative. Second, again unlike in Odysseus' case, things have to work out poorly. In particular, there has to be some second bit of behavior, occurring during the incapacity, that is wrongful behavior. (17)

We can imagine this sort of case by modifying the case of Parfit's self-incapacitated parent. Imagine that the robber behaves as expected, reacting to the parent's irrationality by making his escape. It takes some time, however, for the drug to wear off. In the meantime, the parent irrationally but purposefully --and therefore wrongfully--harms her children. Is she blameworthy for that wrongdoing?

Or consider a more ordinary case. Imagine a surgery patient, recovering from a desperately needed surgery, slowly emerging from the grip of a powerful anesthetic. Awake but still quite drugged, the patient mistreats the attending nurses, making repeated rude, impatient, and insulting demands. Because the anesthesia was a necessary element of a necessary surgery, the patient was properly incapacitated, even knowing that there was a good chance the patient would act impulsively while recovering from the anesthetic. In fact, the hospital requires its patients to remain under observation for a substantial period after surgery exactly because of the anesthetic's effects on appropriate judgment. Many times patients remain asleep throughout that period, but in this case the patient awoke and acted wrongly. Is the patient responsible for that wrongdoing?

Ordinary responsibility would not render these agents blameworthy. Because their incapacitating actions were justified, there is no blame to be had there. What about the incapacitated wrongdoings? These unlucky Odysseus agents are incapacitated when they act wrongfully. Because reasons-responsiveness is a necessary condition of responsibility for wrongdoing under the ordinary-responsibility scheme, the agents are not responsible for their wrongdoings as a bit of action. The wrongful behaviors were foreseeable in light of the agents' earlier actions in becoming incapacitated, and so they might be held responsible for the wrongdoings as consequences. (18) Hence, it might be that they should feel regret, make amends, or the like. However, as consequences, the incapacitated, wrongful behaviors cannot render the otherwise-blameless agents blameworthy.

Contrast this with tracing responsibility. Because the incapacitated wrongdoings were the foreseeable upshots of the agents' earlier behaviors, we trace responsibility for the incapacitated wrongdoings back to the agent's responsibility for their incapacitating actions. Tracing thus holds the agents responsible for their incapacitated wrongdoings as actions. Since the agents would thus be responsible for a bit of wrongdoing, the tracing account entails that the agents are blameworthy.

By considering Odysseus cases, I have identified cases in which ordinary responsibility and tracing responsibility disagree about whether an agent is blameworthy. The tracing account holds unlucky Odysseus agents blameworthy, and the ordinary-responsibility account does not. Between the two, the ordinary-responsibility account offers the more attractive verdict. Intuitively, the unlucky Odysseus agents are not blameworthy. When I imagine the modified Parfit case, for instance, I lack the intuition that the parent is blameworthy, and I lack the related intuition that she could appropriately be punished. Instead, intuitively, she seems unlucky. It is easy to imagine the parent feeling regret, and it is easy to imagine thinking poorly of her if she does not feel that regret or if she fails to make an effort to address the harms she has caused. And I could imagine feeling terrible for the parent who harmed her own child. But she does not seem blameworthy. I feel sympathy, not resentment, toward the parent.

These intuitions comport with thinking of responsibility as tracking the fair opportunity to avoid wrongdoing. Imagine holding the parent blameworthy. You can imagine her asking what she should have done differently. Should she have refrained from taking the drug, thereby exposing herself and her family to the robber's threats? Given the options available, she had no fair opportunity to avoid the risk of the wrongdoing, and so she did not have a fair opportunity to avoid the wrongdoing. Contrast this with the culpable-incapacity cases in the tracing literature, in which the agent did have a fair opportunity to avoid the wrongdoing. The drunk driver, for instance, had the fair opportunity to avoid the wrongdoing when the drunk driver had the opportunity not to become intoxicated to the point of incapacity. It is ordinarily fair to ask someone not to drink to incapacitation.

In the Odysseus cases, ordinary responsibility gets the verdicts right, and tracing responsibility gets the verdicts wrong. That and ordinary responsibility's ability to ground blame in the original culpable-incapacity cases give us sufficient reason to reject tracing and stick with ordinary responsibility.


The Odysseus cases pose a problem for the tracing account because the tracing advocate appears committed to three propositions: 1) tracing extends responsibility in cases of responsible but non-culpable incapacity, 2) tracing extends action responsibility in particular, and 3) being action responsible for a bit of wrongdoing is sufficient for blameworthiness. If those three propositions are true, then the tracing advocate is committed to holding the Odysseus agents blameworthy, and that tells against tracing. So could a defender of tracing not fend off my attack by denying one of those three propositions? Why not, for instance, limit tracing only to cases of culpable incapacity?

The problem for the tracing advocate is not just that any such limitations appear ad hoc. The problem is that the tracing advocate can appeal to these responses only at the cost of making the tracing account substantively indistinguishable from the ordinary-responsibility account, and that would be to abandon a substantive account of tracing. If the tracing advocate rejects one of these three propositions, then she will run into the strongest version of the objection suggested by skeptics like Alexander, Khoury, and King, that ordinary responsibility gives us everything that tracing gives us.

Start with the possibility of limiting tracing only to cases of culpable incapacity. Tracing's advocates do not permit tracing in all cases of incapacity. For example, Fischer and Ravizza point to Roger O. Thornhill, Cary Grant's character in North by Northwest., who is forced to drink bourbon when his enemies want to stage a driving accident. Although Thornhill drives while intoxicated, he is not responsible for his behavior, because he is not responsible for becoming intoxicated. There is no responsibility to trace back to, and so Fischer and Ravizza limit tracing to cases of responsible incapacity.

It might seem natural to strengthen the restriction and limit tracing's extension of responsibility to cases in which the agent was not just responsible but also blameworthy for her underlying incapacity. Were tracing's application restricted in this way, tracing would not extend responsibility in the Odysseus cases because the underlying incapacitation is not blameworthy in those cases. (19)

But the tracing advocate faces a dilemma here. The tracing advocate does avoid the threat of the Odysseus cases by restricting tracing only to cases in which the underlying incapacity is culpable. But, in doing so, the tracing advocate makes tracing duplicative of ordinary responsibility. Even without tracing, the ordinary-responsibility account can explain why the consequences of an agent's culpable incapacity can heighten her blame, obligate her to make repair, and the like. What is left for the traced responsibility to do in such a case? If tracing is limited only to cases of culpable incapacity, then it adds nothing to the ordinary-responsibility account.

The tracing advocate will face the same objection if he attempts to avoid the threat of the Odysseus cases by denying my dialectical presumption that tracing extends action responsibility. Although the tracing advocates suggest that tracing is intended to supplement action responsibility, the conditions of extending tracing responsibility--control at some earlier point when the later wrongdoing is foreseeable--parallel the conditions required for holding an agent responsible for a consequence. And if we understand tracing as extending consequence responsibility instead of action responsibility, then the Odysseus cases pose no problem. Mere consequence responsibility cannot render an otherwise blameless agent blameworthy, and so if only consequence responsibility is extended, the Odysseus agents will not be exposed to blameworthiness. If the tracing advocate appeals to this response, however, tracing becomes substantively indistinguishable from ordinary responsibility. Even without tracing, ordinary responsibility can explain why the culpably incapacitated agent is blameworthy for the consequences of her responsible agency. If tracing does no more than extend consequence responsibility, then it adds nothing new to the ordinary-responsibility account.

Could the tracing advocate deny that being action responsible for wrongdoing grounds blameworthiness? If being responsible in this way for wrongdoing is not sufficient for blameworthiness, then there is room to hold the Odysseus agents responsible for their wrongdoing without holding them blameworthy. This strategy requires the tracing advocate to take a controversial stand on a foundational question about moral responsibility, and that should make this the least-tempting distinction of the three. I suggested that it is nearly axiomatic that an agent's being blameworthy is entailed by her being responsible for a wrongdoing as an action. But not everyone accepts that responsibility for wrongdoing is sufficient for blameworthiness. For example, in a discussion with Derk Pereboom, Fischer writes, "It is crucial here to keep in mind the distinction between moral responsibility and (say) moral blameworthiness (or praiseworthiness)" (2004: 157). Fischer explains that an agent's history--things like manipulation--could make it inappropriate to hold a responsible wrongdoer blameworthy. And Michael McKenna (2012) argues that reasons-responsiveness and wrongdoing alone are not sufficient for blameworthiness; he requires the satisfaction of a quality-of-will condition in addition. Of course, it might be that Fischer's concern about manipulation and McKenna's concern about quality of will are best understood as telling against responsibility, and only thereby against blameworthiness. In any case, these particular constraints will not help the tracing advocate. There is no reason to think that all Odysseus agents are manipulated agents, and it is easy enough to imagine Odysseus agents who might satisfy a quality-of-will condition at the time of the incapacitated wrongdoing. Nonetheless, we can see the conceptual possibility that responsibility for wrongdoing might not be sufficient for blameworthiness.

The tracing advocate here faces the same bind he faced elsewhere. In order to defend a substantive tracing account, the tracing advocate needs to identify some significant difference between tracing responsibility and ordinary responsibility. I have identified one plausible difference between tracing responsibility and ordinary responsibility, but accepting that difference tells against tracing. If the tracing advocate therefore denies that action responsibility is a distinct type of responsibility (or, at least, is distinctive in the way I have suggested), then the tracing advocate has no grounds for holding that the difference between action responsibility and consequence responsibility is more than merely formal. So the tracing advocate can deny that action responsibility is distinctive in this way only by abandoning the substantive tracing account.

This dooms tracing. The tracing advocate needs to show both that tracing makes a real difference and that we should want our theory of responsibility to accommodate that difference. However, the most promising way to distinguish the tracing account from the ordinary-responsibility account --understanding tracing as extending action responsibility--commits the tracing advocate to holding unlucky Odysseus agents responsible and therefore blameworthy. Since the unlucky Odysseus agents are intuitively not blameworthy, the tracing advocate can distinguish the tracing account from the ordinary-responsibility account only by rendering the tracing account extensionally inadequate. The only apparent ways to defuse the threat from the Odysseus cases amount to abandoning tracing as a substantive addition. Accordingly, the Odysseus cases tell us to reject tracing as a substantive addition to responsibility.

Abandoning tracing as a substantive addition does not mean that there is no room for tracing in our thinking about moral responsibility. Even if tracing is not a substantive addition, it might yet serve as a helpful heuristic. Given the similarities between the conditions of applying tracing and the conditions of applying responsibility for consequences, we might charitably understand the arguments offered by the tracing advocates as intending to draw our attention to the role that responsibility for consequences can play in cases in which some of the consequences at issue are further actions. Indeed, philosophers working on other problems have not always treated the tracing account and the ordinary-responsibility account as distinctive accounts of responsibility. Neal Judisch (2005), for example, moves between Fischer and Ravizza's tracing account and their account of responsibility for consequences in discussing a challenge to their taking-ownership condition. And consider Vargas: "We hold someone responsible for the results of drunk driving, not because of the kind of agent they are when they get behind the wheel, but rather, because of the kind of agent they were when they started to drink" (2013: 273). This line of thinking allows that tracing might be an instance of ordinary responsibility for consequences, not a distinctive sort of responsibility for wrongdoing. Noticing the importance of recognizing the cases in which an important consequence of our wrongdoing is some further wrongdoing would be an interesting result, though it would be a revision of the tracing advocates' arguments, given their insistence that the addition of tracing makes a substantive difference.

3. Considering Three Objections

I have argued that we do not need tracing to blame culpably incapacitated agents like the drunk driver, and I have argued that tracing threatens to hold blameworthy non-culpable agents like Parfit's parent or the surgery patient. But tracing has been persistently attractive, and so here I consider three worrisome objections to abandoning tracing.

First, tracing seems to make good the idea that no one should benefit from their own wrongdoing. (20) Being permitted an excuse might seem good for the wrongdoer; the excuse enables the wrongdoer to avoid blame that might otherwise be appropriate. Incapacity is the sort of condition that can ground an excuse. However, the culpably incapacitated agent brings about his own incapacity, and he does so by acting wrongfully. Permitting the culpably incapacitated agent to point to his own incapacitation as grounds for an excuse might then seem to violate the general principle against allowing wrongdoers to benefit from their wrongdoing. The culpably incapacitated agent would have garnered an ostensibly beneficial excuse, and he would have done so by acting wrongfully. At the extreme, allowing culpably incapacitated agents an excuse for their culpably incapacitated wrongdoing might even seem to give wrongdoers a strategy for insulating themselves against recrimination. As the Maine court explained in Arsenault.
   [T]he defense of insanity should never be extended to apply to
   voluntary intoxication in a murder case. It would not only open
   wide the door to defenses built on frauds and perjuries, but would
   build a broad, easy turnpike for escape. All that the crafty
   criminal would require for a well-planned murder, in Maine, would
   be a revolver in one hand to commit the deed, and a quart of
   intoxicating liquor in the other with which to build his excusable

Accepting tracing, and thereby refusing to grant the culpably incapacitated agent an excuse, can ensure that there is no "broad, easy turnpike for escape."

We can set aside this worry. In order to know whether someone has benefited, we have to know what the relevant comparison is. In the culpable-incapacity cases, the agent's earlier, competent wrongdoing--the improper, self-incapacitating behavior--makes the agent more blameworthy than he otherwise would be. He is blameworthy for that initial behavior, and then he risks being blameworthy for further harms (including his later improper behavior) that result. That is significant blame that the agent could have avoided by not acting improperly from the outset. So the agent is not made better off by way of his wrongdoing (at least not in terms of escaping blame). Nor is the agent better off than someone who is involuntarily incapacitated. Both are similarly excused from action responsibility for their incapacitated wrongdoings, but only the culpably incapacitated agent is blameworthy for being incapacitated. Recognizing this, we can see both why tracing might have seemed attractive in this way and why we need not actually worry about it.

Second, we might think tracing is appropriate because it seems to offer the best explanation of a comparative pattern of blaming we see in both ordinary morality and the law: We blame unlucky culpably incapacitated agents who commit some further wrongdoing more frequently and more harshly than we blame lucky culpably incapacitated agents. Consider again the drunk drivers. Drunk drivers, and especially drunk drivers who cause further harm, are exposed to significant and appropriate blame and punishment. What about agents who drink to the point of incapacity but then, luckily, neither drive while intoxicated nor cause any further harm? They are subjected to less frequent and less severe blame and punishment, both in ordinary morality and in the law.

This comparative pattern--that drunk drivers are punished more often and more severely than the merely drunk--might suggest that culpably incapacitated agents are being held responsible for their culpably incapacitated wrongdoing as a bit of action. Recall that it is responsibility for wrongdoing that is supposed to mark the difference between blameworthy and non-blameworthy agents. Since the drunk drivers and the merely drunk alike drank to the point of incapacitation, and since the merely drunk sometimes appear not to be blamed, then it might appear that drinking to the point of incapacitation is not being treated as action. Hence, the wrongdoing that grounds the blameworthiness of the drunk drivers might seem to be their drunk driving.

That we do blame those who commit vehicular homicide more than mere drunk drivers and mere drunk drivers more than mere drunks does not mean that we should blame in these ways. Our practices are not immune to criticism and revision. For instance, Khoury and Alexander suggest that it is our competent behavior that matters for blame, not whatever follows. And so perhaps we should be blaming those who kill less than we do, though we might still expect contrition, compensation, and the like from them. And probably we should blame those who culpably incapacitate themselves more than we do. After all, drinking to the point of incapacity is ordinarily dangerous behavior. People who are that intoxicated cause a whole range of harms, and drunk driving accidents are merely one such particularly deadly result. (21)

Even if, fortuitously, no further harm results, we should sanction that dangerousness. And we might accept one of these latter conclusions without committing to Khoury and Alexander's strict denial of resultant moral luck. If some such revisionary explanation is available, then the need to explain the extant comparative pattern is gone.

And we could, like many, accept that the results of an agent's behavior can affect the degree of appropriate blame. The results need not reflect any difference in the quality of the agent's will nor any difference in the agent's regard for others. However, the results of wrongdoing--risks imposed and harms suffered--can affect the interests of others. The culpable-incapacity cases often result in serious harms. Think of Hinckley, shot and killed by Arsenault, or think of the victims of drunk drivers. Their deaths are serious harms, and many accept that causing serious harms can render a blameworthy agent significantly more blameworthy. And, as King suggests, even culpably incapacitated agents whose further wrongdoing results in no additional harm--such as a drunk driver who fortuitously makes it home without incident --might be held accountable for the additional risk they have imposed, for the close call they created. If these harms and dangers can increase the degree of an agent's blameworthiness, then the tracing skeptic can explain why we might hold the unlucky culpably incapacitated agent far more blameworthy than we hold the lucky culpably incapacitated agent.

We can also explain why it might be appropriate to blame the drunk driver but not the agent who drinks to intoxication but luckily does not drive. We might conclude that, while both are blameworthy, it is only all-things-considered appropriate to blame the drunk driver. We see something like this in cases of de minimis blameworthiness. (22) Section 2.12(2) of the Model Penal Code, for instance, excuses behavior "too trivial to warrant the condemnation of conviction." The de minimis defense is an element of our criminal practices and almost certainly also of our moral practices. Why might this be? Blaming and punishment are not costless. It takes effort to identify blameworthy agents, and we risk blaming and punishing the innocent. Blaming and punishment impose costs--psychological, financial, interpersonal and otherwise --on the blamer, on the punisher, and on third parties. These costs might be particularly unpalatable if the wrongdoing is fairly widespread. And so we might let some wrongdoings slide, though the agents involved are blameworthy. This means that we might blame drunk drivers even though we do not blame drunks, despite both being blameworthy. The costs of blaming might be worth paying in the case of drunk drivers, while the costs might be too high in the case of merely drunks.

I do not here resolve which response the tracing skeptic should offer to address the comparative patterns in our punishing and blaming practices. However, plenty of philosophical resources can be brought to bear, from criticizing our extant practices to explaining them, none of which need tracing. Since we can comfortably address those comparative patterns without appealing to tracing, they do not pressure us to accept tracing.

Finally, rejecting tracing seems to suggest that incapacitated wrongdoings are just ordinary harmful consequences. But surely this is wrong. Both we and the agent should see the incapacitated wrongdoing as more than merely some untoward event in which the agent played some causal role. If ordinary responsibility commands us to take this impoverished view of the relationship between the agent and the incapacitated wrongdoing, so much the worse for ordinary responsibility. (23)

But this objection to tracing skepticism arises only if we are not careful to distinguish between the many different sorts of responsibility that might be at issue. (24) As I have argued, the incapacitated agent is not responsible for the incapacitated action in a way that could ground blameworthiness. However, the agent can be responsible for the incapacitated action in other ways. We might ascribe responsibility to him in a way that permits us to make judgments about his character. For instance, we might think that the pill taken by Parfit's parent unleashed some improper impulse she otherwise would have restrained. She is not responsible for the incapacitated wrongdoing in the accountability sense, but we might make some judgment of her character because she harbored such an improper impulse at all. She is responsible for the wrongdoing in that sense, even if that is not the accountability sort of responsibility that could make her blameworthy.

Likewise, the culpably incapacitated agent might be responsibility in a sense that makes it appropriate for the agent to feel regret and to make efforts at repair. Think of the lorry driver in Bernard Williams's Moral Luck (1981). The lorry driver faultlessly runs over a child, striking the child despite driving with due care. Though the lorry driver, by hypothesis, has done nothing wrong, we expect the lorry driver to feel a special sort of regret, and we may also think it appropriate for the driver to compensate for the harm caused. In the culpable-incapacity cases, the grounds for regret and compensation should be at least as strong. Williams's lorry driver bears only causal responsibility for striking the child. Fischer and Ravizza's drunk driver also bears causal responsibility for the harm caused; however, unlike Williams' lorry driver, Fischer and Ravizza's drunk driver is not faultless. And so just as it would be inapt for the lorry driver to think no more of the harm he caused than that it was something that happened merely in or through him, it would surely be inapt for the drunk driver to have such thoughts.

Non-blame reactions like regret deserve greater philosophical attention. However, we should distinguish them from the guilt and indignation that the reasons-responsiveness theorists and the tracing advocates take to mark accountability and blameworthiness. If we are not careful to distinguish the ways in which agents can be responsible, we might think culpably incapacitated agents are responsible simpliciter for the incapacitated wrongdoing. That could make us think we need tracing to account for culpably incapacitated agents' responsibility, and this would return us to worries about nonculpably incapacitated agents' responsibility--the Odysseus case objection. If we are careful to distinguish between the various sorts of responsibility at issue, however, we can see that tracing is not needed.

4. Conclusion

Tracing's advocates contend that the reasons-responsiveness account of moral responsibility needs to be augmented to account for the blameworthiness of culpably incapacitated agents. However, the ordinary-responsibility account can give us the right explanation in those cases: The culpably incapacitated agents are blameworthy for culpably incapacitating themselves. As other tracing skeptics have suggested, this defuses one motivation for adding tracing, that ordinary responsibility initially appeared to be explanatorily inadequate.

Defusing that motive, however, does not tell us that tracing gets things wrong. To show that tracing is wrong, I have offered a new argument against tracing. The addition of tracing is typically motivated by looking at cases of culpable incapacity, but I have challenged tracing by pointing to cases of non-culpable incapacity, the Odysseus cases. Tracing gets those cases wrong, and ordinary responsibility gets them right. This gives us reason to reject the addition of tracing to the ordinary-responsibility account. And I have supplemented that argument against tracing by offering explanations for tracing's continued popularity, showing how we might have been misled into thinking tracing attractive.

Rejecting tracing is no small matter. Tracing bifurcated the conditions of action responsibility, rendering an agent responsible if either the reasons-responsiveness conditions were met immediately or the tracing conditions were met historically. Rejecting tracing permits us to maintain a univocal condition of action responsibility. And, by rejecting tracing, we eliminate one historical element of the analysis of responsibility. Without tracing, contemporaneous reasons-responsiveness is a necessary condition of responsibility. This is one step toward ascertaining just how and when an agent's history can be relevant to their responsibility for some particular bit of action.

Rejecting tracing also allows us to treat a central sort of criminal wrongdoing more honestly. It might have seemed that tracing was only an exceptional sort of responsibility. However, intoxication is involved in a tremendous proportion of violent crimes. If so many of our most serious crimes involve some degree of culpable incapacitation, it is important that we get the analysis of culpability in those cases correct. So without tracing, what should we say about intoxicated wrongdoers like Max and Arsenault? In answering that question, we will have to wrestle with difficult questions about partial responsibility, about foreseeability, and about just how risky and improper self-incapacitation is. Not all self-incapacitation, not even all intoxication, is alike. Perhaps there is a significant difference between the sort of drinking that agents such as Arsenault have engaged in--drinking far to excess, and in dangerous conditions--and the sort of social drinking that is widespread in our society. Does that difference lead to a difference in culpability? What are we to say about social drinking that unluckily leads to incapacitated wrongdoing? Given the prevalence of the behavior and the stakes of the harm involved, these pressing questions need philosophical investigation. (25)

Craig K. Agule

University of California, San Diego

Department of Philosophy


Alexander, L. (2013) "Causing the Conditions of One's Defense: A Theoretical NonProblem," Criminal Law and Philosophy 7: 623-28.

Brink, D. O. and Nelkin, D. K. (2013) "Fairness and the Architecture of Responsibility," in D. Shoemaker, ed., Oxford Studies in Agency and Responsibility, Vol. 1, Oxford: Oxford University Press, pp. 284-313.

Dworkin, R. (1986) Law's Empire, Cambridge, Mass.: Harvard University Press.

Fischer, J. M. (2004) "Responsibility and Manipulation," The Journal of Ethics 8: 145-77.

--and Ravizza, M. (1998) Responsibility and Control: A Theory of Moral Responsibility, Cambridge: Cambridge University Press.

Fischer, J. M. and Tognazzini, N. A. (2011a) "The Triumph of Tracing," in J. M. Fischer, Deep Control: Essays on Free Will and Value, Oxford: Oxford University Press, pp. 206-34.

--(2011b) "The Physiognomy of Responsibility," Philosophy and Phenomenological Research 82: 381-417.

--(2009) "The Truth about Tracing," Nous 44: 531-56.

Frankfurt, H. (1975) "Three Concepts of Free Action: II," Proceedings of the Aristotelian Society, Supplementary Volumes 49: 113-25.

Husak, D. (2012) "Intoxication and Culpability," Criminal Law and Philosophy 6: 363-79.

--(2010) "The De Minimis 'Defense' to Criminal Liability," in R. A. Duff and S. Green, eds., The Philosophy of Criminal Law, Oxford: Oxford University Press, pp. 362-92.

Judisch, N. (2005) "Responsibility, Manipulation and Ownership," Philosophical Explorations 8: 115-30.

Khoury, A. C. (2012) "Responsibility, Tracing, and Consequences," Canadian Journal of Philosophy 42: 187-208.

King, M. (2014) "Traction Without Tracing: A (Partial) Solution for Control-Based Accounts of Moral Responsibility," European Journal of Philosophy 22: 463-82.

McKenna, M. (2012) Conversation and Responsibility, Oxford: Oxford University Press.

--(2008) "Putting the Lie on the Control Condition for Moral Responsibility," Philosophical Studies 139: 29-37.

Nelkin, D. K. (2011) Making Sense of Freedom and Responsibility, Oxford: Oxford University Press.

Parfit, D. (1987) Reasons and Persons, Oxford: Oxford University Press.

Pomorski, S. (1997) "On Multiculturalism, Concepts of Crime, and the 'De Minimis' Defense," Brigham Young University Law Review (1): 51-99.

Schelling, T. C. (1980) The Strategy of Conflict, Cambridge, Mass.: Harvard University Press.

Shabo, S. (2015) "More Trouble with Tracing," Erkenntnis 80: 987-1011.

Sher, G. (2009) Who Knew? Responsibility Without Awareness, Oxford: Oxford University Press.

Smith, A. M. (2008) "Control, Responsibility, and Moral Assessment," Philosophical Studies 138: 367-92.

Smith, H. (1983) "Culpable Ignorance," The Philosophical Review 92: 543-71.

Timpe, K. (2011) "Tracing and the Epistemic Condition on Moral Responsibility," The Modern Schoolman 88: 5-28.

Vargas, M. (2013) Building Better Beings, Oxford: Oxford University Press.

--(2005) "The Trouble with Tracing," Midwest Studies in Philosophy XXIX: 269-91.

Wallace, R. J. (1994) Responsibility and the Moral Sentiments, Cambridge, Mass.: Harvard University Press.

Watson, G. (1996) "Two Faces of Responsibility," Philosophical Topics 24: 227-48.

Williams, B. (1981) Moral Luck, Cambridge: Cambridge University Press.

Zimmerman, M. (2002) "Taking Luck Seriously," Journal of Philosophy 99: 553-76.

(1) State v. Arsenault, 152 Me. 121, 124 A.2d 741 (Maine 1956).

(2) There is an important ambiguity in philosophers' use of the notion of accountability. For many philosophers, to be accountable for some wrong is to be at fault and culpable for the wrong such that resentment and indignation are appropriate. For other philosophers, to be accountable for some (apparent) wrong is to bear a relationship to the behavior such that it is appropriate for others to demand an explanation. Only if some suitable explanation (such as a justification or an excuse) is not forthcoming would resentment and indignation be appropriate. In both cases, the agent is treated as an accountable agent, an agent upon whom it is generally appropriate to place moral demands. But the two notions of accountability are distinct. On the first sense of accountability, an agent who is excused for some bit of wrongdoing is therefore not accountable for that wrongdoing; on the latter sense, the agent is excused for the wrongdoing as part of the process of holding the agent accountable for the wrongdoing. In this essay, I follow those philosophers who use accountability in the first sense. I thank an anonymous referee for highlighting this distinction.

(3) For John Martin Fischer and Mark Ravizza (1998), for instance, an agent is reasons-responsive, and thus is responsible, when the agent is both reasons-receptive and reasons-reactive. For R. Jay Wallace (1994), an agent has the power of reflective self-control, and thus is responsible, when the agent has the power to grasp and apply moral reasons as well as the power to control his behavior in light of those reasons. And on Dana Kay Nelkin's (2011) rational-abilities view, an agent is responsible when the agent acts with the powers to recognize and act for good reasons.

(4) I have qualified the claim that reasons-responsiveness is seen as sufficient for responsibility. Some reasons-responsiveness theorists are inclined to add a further condition that the agent's character have the right sort of history (or at least not have the wrong sort of history). Such a further condition is often offered to address questions about manipulated agents and agents with a rotten social background. Fischer and Ravizza, for instance, require that an agent have taken ownership for the psychological mechanism upon which she acts. Because conditions like the taking-ownership constraint do not bear on the tracing question, I set them aside.

(5) The ordinary-responsibility account is not the only ahistorical account of responsibility. For example, "mesh" theories of responsibility claim that an agent is responsible so long as there is the right sort of mesh between an agent's preferences or desires and her action. Many take mesh theories to be ahistorical. Harry Frankfurt, for instance, explains that "the questions of how the actions and his identifications with their springs are caused is irrelevant to the questions of whether he performs the actions freely or is morally responsible for performing them" (1975: 122).

(6) In Fischer and Ravizza's mechanism-specific terminology, we would say that Max's drinking was the product of a reasons-responsive psychological mechanism.

(7) Section 2.08(4). We should distinguish two sorts of defenses available at law: elemental defenses and affirmative defenses. An elemental defense contends that one of the crime's constitutive elements is missing; it denies that the defendant has committed the crime. An affirmative defense, by contrast, does not dispute that the crime was committed, but it denies that the defendant should be held accountable for the crime. It is not controversial that intoxication, self-induced or otherwise, can ground an elemental defense. Consider burglary: A defendant commits burglary by breaking into a building with the intent to commit some further crime. If a defendant is so intoxicated that he cannot form the intent to commit some further crime, then no burglary has been committed (though the defendant may be guilty of trespass). However, the tracing question arises when we consider the conditions of responsibility, and denials of responsibility constitute affirmative defenses.

(8) State v. Altimus, 306 Minn. 462, 238 N.W.2d 851 (Minn. 1976).

(9) This debate continues in the literature. Fischer and Tognazzini (2011a) have since offered a revised version of their 2009 paper, again contending that, for each of Vargas's cases, either the agent could have foreseen the wrongdoing in the right sort of way or the agent should not be held responsible. Kevin Timpe (2011) makes a similar argument, claiming that we can defuse Vargas's cases by getting a more precise grasp on the epistemic condition of responsibility. Seth Shabo (2015) has recently offered a further argument in this thread, pointing to cases in which responsibility seems to outpace foreseeability. If tracing requires foreseeability as Fischer and Ravizza suggest, then the cases offered by Vargas and Shabo should give us concern about just how much explanatory power tracing can offer. However, because the Odysseus cases provide independent grounds for rejecting tracing, I leave the foreseeability worries aside.

(10) King also suggests that tracing brings complications we can avoid if we reject tracing, citing Michael McKenna (2008), George Sher (2009), Angela Smith (2008), and Vargas (2005). If foreseeability constrains responsibility, how plausible is it that the ultimate wrongdoing is foreseeable before the agent has become incompetent in the tracing cases? And what about tracing cases in which the agent has incapacitated himself thoughtlessly? Is it reasonable to hold the agent accountable for that oversight even if the agent was never cognizant of the possibility of precaution? But concerns like these cannot tell against tracing. At best, they serve to delimit the scope of tracing to cases in which the ultimate wrongdoing was foreseeable or perhaps even foreseen at the time the agent acted to incapacitate himself or failed to prevent his incapacitation. We might think that at least some cases of agents such as Max and Arsenault are like this. Further, that the tracing account must address these sorts of concerns does not give the ordinary-responsibility account any advantage, since that account must address the same sorts of concerns. Foreseeability and control matter for the ordinary-responsibility account's notion of responsibility for consequences. Responsibility for consequences is how King hopes to explain the responsibility of the culpably incapacitated actor, and so rejecting tracing does not sidestep these problems, which are really questions for accounts of responsibility more broadly. The skeptics' better argument is the explanatory-adequacy argument.

(11) I set aside any distinction between actions and omissions here, recognizing that responsibility for omissions presents a rich set of questions. The important distinction for considering tracing is that between actions and consequences.

(12) An anonymous referee wondered whether the tracing advocate is committed to this duplication. Here we see the problem that the tracing advocate faces throughout. The tracing advocate needs to find a substantive difference between ordinary responsibility and tracing that is worth having. The duplication presents an apparent substantive difference. If the tracing advocate finds some way to avoid the duplication, then the tracing advocate faces renewed pressure to find some alternative way to distinguish the tracing account from the ordinary-responsibility account.

(13) Actually, in both cases, it would be more accurate to say that we have something like a prima facie reason to resent or to punish, not an all-things-considered reason. For example, there might be reasons against resenting or punishing the blameworthy agent--perhaps the costs of resentment and punishment or the harms that might befall third parties--which make it the case that, while there is some reason to resent or to punish, all things considered it would be best not to. I set that difference aside here.

(14) Responsibility for bad consequences can matter for blameworthiness even if we reject resultant moral luck. As Michael Zimmerman (2002) explains, rejecting resultant moral luck does not mean that consequences become wholly irrelevant. Even if we set aside any possible import for blame, being responsible for consequences can have other normative import. Moreover, rejecting resultant moral luck is controversial, and many reasons-responsiveness theorists accept resultant moral luck. Fischer and Ravizza, for instance, have an extended treatment of the conditions of responsibility for consequences. And it is common to see the criminal law as accepting resultant moral luck, punishing completed crimes more harshly than merely attempted crimes.

(15) Or, perhaps more accurately, the cases are presented to us as cases in which we are supposed to take the risk to be justified. Odysseus is supposed to be a hero, and his cleverness is supposed to be his virtue. Some modern readers might find themselves less impressed with his willingness to risk the lives of those loyal to him. And some readers might likewise be unconvinced of Parfit's parent's assessment of the relative risks involved. Even if skeptical readers doubt these particular cases, they should be able to discern the pattern involved and imagine their own cases, perhaps even more fanciful, and I will shortly present more quotidian Odysseus cases.

(16) For both of these agents, risky incapacity was a tool used to achieve some goal. But this is not the key fact. We could imagine an Odysseus agent for whom the incapacity is a foreseeable side effect.

(17) The tracing account extends action responsibility to the later behavior. Extending action responsibility (and not merely consequence responsibility) matters for blaming only when the later behavior is wrongdoing. So, to contrast the tracing account and the ordinary-responsibility account, it is important that the second bit of behavior be wrongful behavior. In this paper, I am agnostic as to the conditions of behavior's being wrongful. However, it is plausible that wrongful behavior requires the possession of certain mental states, and it might be that some of the conditions that mar responsibility also sometimes preclude wrongfulness. For instance, recall the example of the intoxicated burglar from n. 7, where I explained that if the intoxication made it impossible for the agent to form the requisite intention, the agent did not commit burglary. Set aside those cases, and focus on cases in which an incapacitated agent can still act wrongfully.

This limitation marks a significant difference between culpable incapacity and culpable ignorance. The ultimate behavior in the culpable-incapacity cases is wrongdoing, and we are asking whether to hold the agent responsible. On a common understanding, the ultimate behavior in the culpable-ignorance cases is not ordinary wrongdoing, precisely because the agent is ignorant of some fact that bears on the behavior's inappropriateness, and we are asking whether to treat the behavior as wrongdoing nonetheless. For a clear treatment of culpable ignorance invoking a distinction parallel to that between the ordinary-responsibility account and the tracing account, see Holly Smith's "Culpable Ignorance" (1983).

(18) It might seem strange that it is true both that the later, wrongful behavior was a foreseeable consequence of the agent's earlier behavior and that the agent's earlier behavior was not wrongdoing. However, there is nothing improper about this. Lots of behavior runs risks, and so long as we think that some risks can be justified, there is room to think that a bit of behavior might not be wrong even when it results in a bad outcome. Why would anything change about this just because the bad outcome involves a risked bit of wrongdoing?

(19) We might think of this limitation in inheritance terms. Because the traced responsibility is rooted in the responsibility for the underlying incapacity, it might not be surprising if the traced responsibility inherited some of the features of the underlying incapacity. In the Odysseus cases, the underlying incapacitation is justified. It might seem natural in those cases to think that the traced responsibility would inherit the normative effect of the justification of the underlying incapacitation. And if that's so, then the tracing does not extend blameworthiness-grounding responsibility. I thank an anonymous referee for pointing out that this response can be thought of in inheritance terms.

(20) This general principle is a feature of American law familiar to many philosophers from Ronald Dworkin's discussion of the New York case Riggs v. Palmer, 115 N.Y. 506 (1889) in Law's Empire (1986). In that case, Elmer Palmer murdered his grandfather to ensure that the grandfather died before writing Palmer out of his will. As the court explained in ruling against Palmer, New York's probate law was to be interpreted in light of the general principles within the law, including the principle against allowing wrongdoers to benefit from their wrongdoing.

(21) As an anonymous referee noted, drinking to intoxication need not always be reckless. For instance, it is possible to imagine someone who, prior to drinking to the point of incapacitation, takes precautions to preclude later incapacitated misbehavior, such as arranging for a designated driver, handing over the car keys, or the like. Whether these precautions are sufficient to obviate the culpability for becoming intoxicated to the point of incapacity is not clear to me. However, it is true that it is possible for an agent to become intoxicated to the point of incapacity without thereby raising the sorts of risks that ordinarily render such behavior culpable. For such cautious agents, their incapacity would not be culpable, and we should not blame them, regardless of whether any harm results.

(22) What counts as de minimis wrongdoing and whether we should withhold blame in those cases are questions I do not fully address here. For more substantive treatments, see Husak (2010) and Stanislaw Pomorski (1997).

(23) I thank an anonymous referee for raising this objection forcefully. This objection invites tracing's advocates to say more about the boundaries of the reactions that accountability licenses, a rich area for further discussion.

(24) There is a significant literature on the many kinds of responsibility that might be at stake, from Watson's seminal "Two Faces of Responsibility" (1996) (which yields talk of attributability and accountability) to Fischer and Tognazzini's recent "The Physiognomy of Responsibility" (2011b) (where they identify a number of different attributability questions, a number of different accountability questions, and matters of responsibility that lie between the two).

(25) I have received invaluable feedback from many people on this and earlier drafts. I would particularly like to thank Saba Bazargan, Amy Berg, David Brink, Jeremy Dickinson, Stephen Galoob, Joyce Havstad, Gil Hersch, Matt King, Alexandre Marcellesi, Michael McKenna, Per Milam, Dana Nelkin, Ben Sheredos, Michael Tiboris, and Danny Weltman. I would also like to thank the attendees of a talk based upon an earlier version of this paper at the 2014 Pacific Division Meeting of the American Philosophical Association as well as several anonymous referees. Finally, I would like to thank the editorial staff of the Journal of Ethics & Social Philosophy for patient assistance in preparing this article for publication.
COPYRIGHT 2016 Journal of Ethics & Social Philosophy
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2016 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Agule, Craig K.
Publication:Journal of Ethics & Social Philosophy
Date:Jan 1, 2016
Previous Article:I might be fundamentally mistaken.
Next Article:Reducing reasons.

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters