Mere moral failure

Abstract

When, in spite of our good intentions, we fail to meet our obligations to others, it is important that we have the correct theoretical description of what has happened so that mutual understanding and the right sort of social repair can occur. Consider an agent who promises to help pick a friend up from the airport. She takes the freeway, forgetting that it is under construction. After a long wait, the friend takes an expensive taxi ride home. Most theorists and non-theorists react to such cases by either judging the agent's action as a violation of her obligation to help or as having satisfied the only obligation she really had, namely to try to help. However, as I show, there are serious difficulties that arise from categorizing this agent's action as satisfying or violating her obligation – difficulties that are avoided if we instead add “mere moral failures” to the basic categories for moral evaluation. An agent merely fails when she neither satisfies nor violates her obligation. She is responsible for what she has done, yet without thereby having done wrong. Moreover, there is a recognizable (though nameless) reactive attitude reflecting the agent's responsibility that falls between blame and mere regret, and is also importantly different from agent-regret. What I show is that we need not reach for blame as our only way of registering that an agent has morally failed another. While thus far overlooked, mere moral failure is by no means a rare occurrence, but rather a regular part of life among friends, family, investors and clients, police and citizens, doctors and patients, and many others.

Keywords::

• moral obligation
• violation, wrongdoing
• blame
• failure
• agent-regret

Acknowledgements

I want to thank Stavroula Glezakos, Barbara Herman, Richard Kraut, Dana Nelkin, the blind referees at the Canadian Journal of Philosophy, and audience members who gave critical feedback on earlier drafts presented at University of Illinois, Urbana-Champaign, University of California, Riverside, California State University, Northridge, University of California, Irvine, University of California, Santa Barbara, Bay Area Forum for Ethics and Law, Ninth Southern California Philosophy Conference, First Annual Society for Ethical Theory and Political Philosophy at Northwestern University and the Work in Progress Group at the Claremont Colleges. I also want to thank David Beglin for proofreading assistance.

Notes

^1. Although I am treating “obligation” as interchangeable with the terms “requirement,” “ought,” “should” and “supposed to,” an alternative approach can be found in Simmons (1979, chapter 1).

^2. There are, of course, many categories of evaluation, such as “permissible” and “supererogatory,” but these forms of evaluation are not at issue in this article.

^3. In this article I focus on moral obligations and aim to establish the category of mere moral failure, though the term “moral” is often left implicit. However, I also believe that one can merely fail with respect to non-moral obligations.

^4. This is not a case involving an implicit promise (and hence a version of the first example), but rather a situation where an agent is set on avoiding harm to another, and thus focused on one feature necessary for doing so, but who forgets to consider a second feature necessary for avoiding harm.

^5. In the first section I provide further details to persuade the reader that aiding in the second case is obligatory, not supererogatory. The reader may, in the end, find the other cases more convincing. Since no one case may speak equally strongly to all audience members, and any one case suffices for my purposes, I present a range of cases so that readers can focus on the one they find most compelling.

^6. To say that an emotional reaction is justified is different from claiming it is typical, understandable, useful, or required. It might be typical for LA drivers to become enraged when cut off in traffic, understandable for parents to become extremely fearful when their child faces a minor danger, and useful to be irate with the customer service representative since this makes it more likely one will get reimbursed. However, if a feeling is justified, then the content of the thought that constitutes or is associated with the emotion is correct (as well as the emotion's duration, intensity and so on). See, for example, Aristotle's Nicomachean Ethics, Books III–V, D'Arms and Jacobson (2000), and Tannenbaum (2007). Fear, for example, involves the thought that the object is dangerous and this emotion is justified only if the object is in fact dangerous (there are complications here that I do not have space to address such as when one's belief that something is dangerous reasonable but false). Lastly, note that a feeling can be justified without being required: being nervous may be justified when performing in the Olympics, but one makes no sort of mistake if one instead has only feelings eagerness or competitiveness.

^7. For some readers, the key aspect of this article will be (a) and (b), namely that there is a reactive attitude reflecting the agent's responsibility for morally failing that falls between blame and mere regret, and, as I later show, is importantly different from agent-regret.

^8. In fact, the landscape is even more complicated since in some cases an obligation is taken care of by another before one has chance, is transferred to another, etc. But I am setting such cases aside since they are not relevant to the issue at hand.

^9. Aristotle relied on emotional reactions to inform, among other things, his account of the counter-voluntary, the virtues and the difference between the enkratic and virtuous agent. Strawson (1962) highlighted emotional reactions such as resentment and focused on the sort of considerations that mitigate or remove it. He then used this as a basis for critiquing certain theories about moral responsibility and free will. Williams (1981) criticized a Kantian account of justified action by appeal to an emotional reaction he dubbed “agent-regret.” Many others (e.g. Stocker [1976]), have also appealed to emotional reactions to critique and broaden moral theorizing.

^10. This article will not address the question of what sort of repair may be called for in cases of mere moral failure. Rather, its aim is to establish that there is such a moral category.

^11. There is a large literature on whether to accept the slogan and how to interpret it. For a recent discussion, see Scanlon's (2008, 45–52) objection to Judith Jarvis Thomson's use of an “objective ought.” Nothing in this article hangs on accepting or rejecting the slogan. Here I am only showing that accepting the slogan, at least under one reasonable interpretation, is consistent with claiming that the agents described earlier have an obligation to help/not harm.

^12. While the drowning person would be no worse off were the agent on the pier to do nothing, this does not lessen responsibility of the person on the pier. This is akin to the way in which doctors have obligations to help their patients even when the patients would have been no worse off were the doctors to do nothing. Insofar as one thinks that some cases of the imperfect duty to aid can be obligatory, (e.g. Singer's [1972] example of a drowning child), one should fashion such a case in one's mind.

^13. Some epistemic obligations have theoretical, others practical sources, and within the practical, both moral (e.g. generated by making a promise) and nonmoral (e.g. generated by a desire for contentment) sources.

^14. What makes a lapse simple has nothing to do with whether it is extraordinary, uncommon, or unexpected, but rather with the explanation of the lapse – whether the causal explanation is such that it makes a difference to our moral evaluation. For further examples of such lapses, see M. Smith (2004). Not only can lapses be simple or basic but so too can mistakes. See Austin's (1979, 200, note 1) discussion of mistaking the hot for the cold water tap. In Section 2.1 I discuss whether simple lapses ground blame or serve as an excuse from blame.

^15. One might have this worry: these agents could not have recalled through an act of will, and thus they cannot recall. This would then lead us to doubt that they have an obligation to help or avoid harm. Is voluntary control, however indirect, as opposed to mere reason-responsiveness, a necessary condition for the ability to recall? See, for example, Fischer et al. (2007) and Hieronymi (2009a, 2009b). For purposes of this section I am assuming that there are situations in which an agent could have remembered, but did not, while leaving open what the correct account of this metaphysical possibility is.

^16. It is an interesting question on what basis we categorize some situations as those where one could have remembered but did not as opposed to those where one could not have remembered. In both cases there are events in the brain and questions of freedom and its connection to agency arise. As I noted earlier, I am assuming that there are situations in which an agent could have remembered but did not, without providing an account of this type of metaphysical possibility.

^17. Thomson (1989) endorses a similar claim.

^18. Later I will consider whether F's failure to pick up her neighbour is blameworthy.

^19. And so F's epistemic lapse does not reflect an objectionable evaluative judgment or commitment. For a discussion of epistemic lapses that do reflect objectionable evaluative judgments or commitments see A. Smith (2005).

^20. Similarly, when my friend bangs my child's head into a lamp because he forgets to look up, his lapse is not due to a blameworthy orientation or any other blameworthy condition.

^21. Dahl (1967), among others, claims that not doing what one should is sufficient for being blameworthy. F's case challenges Dahl's claim as applied to epistemic obligations (later I show that F's case challenges Dahl's claim as applied to obligations to act). F has a morally derived epistemic obligation (obligation to know certain facts and to bring that knowledge to bear in the right way in her reasoning). She does not meet this epistemic obligation, she is not blameworthy for it, and she is not excused from it, all of which indicates that she neither violates nor satisfies this epistemic obligation.

^22. Utilitarians who allow talk of obligations might accept such a claim. And some Deontologists hold a modified version of this claim. For example, Thomson (1991) implicitly accepts this view: if an agent has an obligation to help and this obligation is not overridden, cancelled or otherwise neutralized, then not helping violates this obligation. A contractualist such as Scanlon is also likely to accept it: “What makes an action wrong is the consideration or considerations that count decisively against it,” (2008, 23) and I think he would agree that there are considerations that count decisively against F taking the freeway to pick up her neighbour from the airport.

^23. Perhaps this is a good place to reveal my naming scheme. V is so-called because she violates her obligation, whereas F merely fails with respect to her obligation. S, who I will discuss later, is so-called because she satisfies her obligation.

^24. Or consider this version: V prefers to take the side streets for the scenery, and she knows this entails leaving the house early. But she does not leave the house early because she wants to watch the end of her TV show.

^25. However, there is another set of reactive attitudes that are justified, and they involve the thought that F is on the moral hook and has failed her neighbour (see Section 3).

^26. An additional concern is whether my own past behaviour is the right kind of reason for refraining from blaming another. See Hieronymi (2004).

^27. I take no stand on whether blameworthiness and blaming come apart, but show that even if they do, it would not help us properly distinguish V and F's actions.

^28. For example, according to Rosen “we are hardly ever culpable for the wrong we do” (2002, 62). “If [one] has been neither negligent nor reckless in the management of [one's] opinion – then [one's] ignorance is blameless and so is the act done from ignorance” (63). Sher puts forward this condition for blameless wrongdoing: “an agent is not responsible for an act of whose wrongness or foolishness he is unaware unless (1) his failure to recognize the act as wrong or foolish falls short of satisfying some applicable standard, and (2) his failure to satisfy that standard is in turn accounted for by some combination of his constitutive attitudes and traits” (2009, 21, see also 87–90). F's action is not “explicable in terms of the interactions among the innumerable desires, beliefs, attitudes, and dispositions, many of them unconscious, that together make him the individual he is” (21). If Sher is using “responsible” to mean “blameworthy,” then Sher would characterize F's action as blameless wrongdoing. If, instead, he sees these as conditions for responsibility, then, contrary to Sher, I maintain that F is nevertheless morally responsible and that certain negative moral reactive attitudes (but not blame) are justified – more on this in the next section.

There are two other ways one might pull apart blame and wrongdoing. One could appeal to Scanlon's account in which wrongness is determined by whether one acted contrary to principles that could not be reasonably rejected whereas “the blameworthiness of an action depends, in ways that wrongness generally does not, on the reasons for which a person acted and the conditions under which he or she did so” (2008, 124) My response to this view is discussed in the second part of endnote 34. Alternatively, one might loosen the connection between the evaluation of an act and the evaluation of the agent who did the act, and claim that while F's action is wrong, F is not to blame, since blame-judgments are a matter of an agent's character. Thomson (1991) and Ross (1939) adopt a view like this. However, I do not think this strategy will be of much help since one should judge not only V and F differently but also one should judge their actions differently. For an additional criticism see Scanlon (2008, 127).

^29. The examples are drawn from Thomson (1990, 229 and 1991).

^30. It is unclear whether these last two agents really act, but, according to Thomson, “agency is no more required for violating a right than fault is” (1991, 302).

^31. Those who accept “ought implies can” will reject the claim that these agents have violated their obligation, and hence will also reject categorizing their actions as blameless wrongdoing. If an agent is obligated to act, then, they will claim, she must be able to reason to the conclusion to do the act (where this includes omissions). If one applies this lesson to the obligation not to harm, then the harms one is obligated to avoid are limited to those it is reasonable to require an agent to know about and so reach a conclusion to avoid. It is not reasonable to demand, for example, that the agent know that flipping the switch will kill her neighbour. And so, holders of “ought implies can” would maintain that the harm she inflicts is not the harm that she has an obligation to avoid.

^32. Blame certainly implies wrongdoing, but if wrongdoing also defeasibly implies blame, then one might accept this defeasible biconditional: an action is wrong if and only if it is blameworthy. Accepting this biconditional is consistent with maintaining that wrongdoing is the more fundamental concept or explanatory notion.

^33. Thus, the following way of carving up the conceptual terrain is problematic: when one has an obligation to act, one either satisfies or violates that obligation; if the latter, then either blame or some other reactive attitude is justified. The problem is that blame and this other reactive attitude are indistinguishable since they have the same constitutive thought: “the act was wrong.” Some take guilt to involve not only the thought that one has acted wrongly and a feeling (feeling bad about what one has done), but also the thought that one deserves to feel bad. And so one might try to distinguish F's feeling from V's by claiming that F does not have this extra thought, while V does – this is why what V feels is guilt, while what F feels is not guilt. This is a controversial interpretation of guilt and it is not clear we should accept it. In any case, an explanation is owed as to why it would be inappropriate for F to have the thought that she deserves to feel bad for what she has done.

^34. The same challenge applies to those who might describe F's action as objectively wrong but subjectively right (see Graham [2010]) or as “justified wrongdoing” (see Buss [1997]). Note that, in any case, F's action does not seem to be an example of “justified wrongdoing.” Buss' case and F's case have this in common: the agents are blameless and yet accountable. But her agent is blameless because justified, whereas F is not justified. F can be reasonably expected to know both what would help and that F has an obligation to help, whereas in Buss' case we “suspect he was justifiably ignorant” (338).

The challenge facing a view like Scanlon's is different. Blame, according to Scanlon, is not tied per se to wrongdoing. The latter is a matter of doing an action where considerations count decisively against it and not a matter of the agent's reasons for action (2008, 23, 124), whereas blame is about the “meaning of the action,” which is determined by whether the action shows something about the agent's attitude towards others that impairs the relations that others can have with that agent (128, 141). Applying Scanlon's view, only V, not F, can justifiably blame herself, since only V has a problematic attitude towards others. Moreover, since the meaning of an action depends on the agent's attitudes or reasons for action (55), then it would seem that F's action does not have any special meaning for her relationship with her neighbour. And yet the moral relationship between F and her neighbour is impaired as is clear from the fact that F is justified in having some sort of negative self-reactive attitude and for initiating repair (see Section 2.2). This impairment is not because of F's attitude but because of what F did. Moreover, how is F to think about what she did, except, on Scanlon's view, as wrongdoing? In which case wrongdoing itself does have moral meaning, or, alternatively put, the meaning of an action is not exhausted by the agent's attitude, as Scanlon claims, but can be solely a matter of whether she did what she was not supposed to do.

^35. The Stoics, Kant and Kantians are often interpreted as holding such a view. See, for example Herman's claim that “good intentions plus adequate care are enough (whatever the outcome, the agent has done what she ought)” (1993a, 98). Some virtue theorists too hold such a view. Foot claims, “it is not, of course, necessary that charitable actions should succeed in helping others” (1978, 169, note 12). The view can be found beyond the two camps just mentioned. For example, according to Prichard, “an obligation must be an obligation, not to do something, but to perform an activity of a totally different kind, that of setting or exerting ourselves to do something” (1949, 35) and Rabbi Tarfon claims, “It is not your duty to complete the task, but you are not free to desist from it” (1962, 2:21).

^36. In what follows I assume that adopting and pursuing the end of helping is appropriate. It may be that one must also intend to do one's duty, or to act well, while seeing helping as instrumental to, an instantiation of, or constitutive of doing one's duty or acting well. The arguments in this article do not hang on the resolution of this issue. Also, when I say that an agent acted with the intention of helping, I mean this to be different from, and inconsistent with, for example, acting with the intention of helping so as to later inflict harm (e.g. reviving someone in order to then torture him).

^37. One can adopt the end of helping without ever pursuing this end (perhaps because the pursuit of this end just so happens to always be overridden by other ends whenever the occasion to help arises). And even if adopting an end entails taking the means in a certain situation, it is yet a stronger claim to say that the means must be pursued with this intention. Moreover, perhaps trying to help requires very little effort whereas acting with the intention of helping requires more effort (e.g. the expression of trying seems applicable for a child who just dips his toe into the cold water. “I tried,” he can truthfully claim).

^38. An obligation merely to try to help is a scaled back in comparison to an obligation to help since much less is required to satisfy it compared to the obligation to help

^39. It is an interesting question when one has an obligation to try to help, for example, rather than an obligation to help. We certainly would criticize a person for not trying if she reasonably but falsely believed there was a way to easily help another. Perhaps this reveals two conditions of when we have scaled back obligations: (i) it is through no fault of the agent's own that Xing is impossible and (ii) the agent would be morally criticizable if she did not try to X. An alternative view to consider: one always has an obligation to try and sometimes also has an obligation to help. Though perhaps we should reject this view for the same reason we reject this view: one always seems to see and sometimes also sees.

^40. Similarly, when the person on the pier first comes across the drowning stranger, it is possible for her to save him, and it is reasonable to require her to know that what would best help is throwing the rope where it will not be swept away immediately by the currents. And in the case of the friend carrying my child it is possible to step, and reasonable to require him to know to step, where it is dry and there is no low hanging lap.

^41. See Williams (1981) and Tannenbaum (2007).

^42. Wolf (2001) considers two equally negligent acts only one of which causes death to another. She suggests that it is a virtue to take responsibility for the consequences of one's action (13) and that the negligent killer should blame himself, which is determined not merely by whether the agent is blameworthy but also by whether agent-regret would be appropriate. In this article I am developing an account of responsibility that does not appeal to blame or agent-regret (unlike Wolf's account of the negligent killer). F's case is importantly different from those considered by Wolf, since F is not negligent (see Section 2.1), and hence not blameworthy, and, as I have just explained, agent-regret is also not an appropriate response for F to have towards her action.

^43. Granting that an apology can repair less than successful activity does not commit one to a “balance sheet” view of action. One cannot do wrong with the intention of later apologizing, as if this would make everything come out “even.” It is another question – not one addressed in this article – whether and when an apology itself can or should be enough to make up for or repair what one has done.

^44. Herman (1993a, 97–98) suggests that agents motivated by duty will adopt maxims of response in such a situation, but it is unclear why doing so is a form of repair if the agent has previously acted on well-formed maxims and if acting on such maxims is sufficient for satisfying one's obligation.

^45. If F is missing an appropriate intention, is she therefore, on this “intention only” view of obligations to act, blameworthy? That has traditionally been the response. But, given the emotional phenomena that are our starting point, F is not blameworthy. So one who holds the “intention only” view now must join with me in rejecting the dichotomy (violation vs. satisfaction) of the basic categories of evaluation and accept the category of mere failure. In addition, there is still room for disagreement about this: what obligation did F fail – merely an obligation to act with the appropriate intention or, what I suggest in Section 3, an obligation to help? The last challenge considered in this section helps push us yet farther away from an “intention only” view of obligation.

^46. See Herman (1993b, 28–29, 1993a, 98).

^47. According to Sher there are four different contexts in which we attribute responsibility. We do so whenever we (1) blame an agent for acting wrongly; (2) praise an agent for exceeding his duty or doing the right thing under difficult circumstances; (3) criticize an agent for acting foolishly or self- destructively; or (4) give an agent credit for his success in achieving some aim” (2009, 14). If Sher thinks that these are the only categories of responsible action, then our views conflict. As F's action demonstrates, there are actions that are not wrong, not blameworthy and neither foolish nor self-destructive, but for which the agent is nevertheless morally responsible.

^48. Thomson (1991) claims that self-defence is justified in order to defend one's rights from violation and that innocent threats (that kill) do violate one's right not be harmed or killed. However, an innocent threat does not threaten due to a blameworthy orientation or condition. And so, were an innocent threat to harm, it would not violate an obligation not to harm. If violation of this obligation is necessary for a rights violation, and a rights violation is necessary for justified self-defence, then an alternative justification must be offered for self-defence against innocent threats.

^49. See, for example Fischer (1986), Watson (2003) and Sher (2006). An exception is Strawson (1962).

^50. Based on an example from Thomson (1991, 293).

^51. Is the combination of failing to help or avoid harm and a blameworthy orientation or condition of the agent (where the latter explains the former) a sufficient condition for the agent's violating an obligation to help or avoid harm? This will depend on, among other things, whether and in what sense the ability to do otherwise is a necessary condition for violating these obligations, which I am not addressing in this article.

^52. That is, if one has an obligation to help or not harm, then one's action must result in the relevant person being helped or not harmed. But if, for example, an agent reasonably but falsely believes road X is the only way to the airport (perhaps an otherwise reliable radio announcer mistakenly directs listeners to road X, which is in fact blocked), then she only has an obligation to try to pick her neighbour up from the airport, not an obligation to pick her up (see endnote 39). And if someone beats her to the punch and picks up her neighbour first, then she has no obligation to help (see endnote 54) or to try to help.

^53. Notice that the jet skier, Alfred, and F fail to satisfy their obligations. However, the jet skier and Alfred violate at least one obligation (of attitude) while F does not.

^54. Moreover, the existence of an obligation also depends, in very particular ways, on the facts of the situation. If sharks suddenly appeared on the scene in the midst of hauling the drowning stranger to safety, and devoured the stranger in mere seconds, then the agent's attempt to save would not have failed her obligation. In such a case the obligation to help dissolves. Likewise, if someone had jumped in the water and saved the drowning stranger before the person on the pier had a chance to throw the rope, the rope thrower's obligation to help would have dissolved.

^55. See Nagel (1979).

^56. I have not by any means discussed the full range of lapses in thought that can lead a person to fail to satisfy her obligation. For example, I have not discussed errors in reasoning that lead an agent to undertake an unsuccessful course of action. Consider, for example, an engineer who makes a simple multiplication error while calculating the force necessary to stop an empty trolley from barrelling into people trapped on the tracks. As a result, his use of force was inadequate to stop the train.

^57. While lapses are inevitable (they will happen), it does not follow that they are unavoidable (they have to happen; they cannot be avoided). And so one still has an obligation to help or not harm, as opposed to an obligation to help most of the time or merely to try to help. Consider a different sort of case, namely weakness of will. It too is inevitable but not unavoidable. It can be true that I will succumb to temptation, while also being true that I can and should resist temptation.

Additional information

Notes on contributors

Julie Tannenbaum

Julie Tannenbaum is an associate professor in the philosophy department at Pomona College. She works on philosophical issues that arise in moral psychology, ethical theory, metaethics and bioethics.