2. The account

According to the counterfactual comparative account of harm, whether or not an event, e, constitutes a harm for some person, S, depends on a comparison between S's well-being in a certain possible world in which e occurs and her well-being in a certain nearby possible world in which e does not occur. We can formulate the view as follows:

This means that harm is an essentially comparative concept, and that no event can be a harm in a non-comparative sense: the fact that e involves S having negative well-being is compatible with its being a benefit for S, and the fact that e involves S having positive well-being is compatible with its being a harm. It also means that temporal comparisons are not strictly relevant: e might be a harm for S even though she is better off than she was a moment ago, if she would have been even better off had things been otherwise.Footnote ³ Finally, CCA is designed to apply to merely possible events as well as actual ones.Footnote ⁴ Although I did not step on the nail, it seems clear that I would have suffered harm if I had.

It is useful to draw a distinction between instrumental and non-instrumental harm.Footnote ⁵ Some events are harmful for the person in virtue of their capacity to cause or otherwise produce other, separate events, while some events are directly harmful for the person, in a way that does not derive from an ability to produce additional events. For example, stubbing one's toe is an instrumental harm. Although I would have been better off had I not stubbed my toe, this is because stubbing my toe produced an episode of pain, and I would have been better off had I not suffered that episode of pain. If a toe stubbing did not produce any episode of pain, then it would not be a harm. Instrumental harms are genuine harms, but indirectly so, and ultimately derive their harmfulness from a connection with a non-instrumental harm. The episodes of pain that are associated with or caused by toe stubbings, by contrast, are non-instrumental harms. Events such as these are, in some sense, directly harmful – I am better off if I do not suffer the episode of pain, and this is not because the episode of pain is connected to some further event such that I would have been better off had that event not occurred. Although CCA entails that the pain's status as a harm depends on a comparison between my well-being in the world in which I suffer the episode of pain and my well-being in the worlds nearest it in which I do not, there is no further event that is caused or produced by the pain such that it is why I am worse off than I otherwise would have been. We can describe such harm as non-instrumental or final harm.

3. Harms of omission and failure to benefit

This article will focus on an objection to CCA based on cases in which someone refrains or otherwise fails to provide a benefit to someone.Footnote ⁶ Intuitively, it seems possible to refrain from benefitting a person without thereby harming them, but such cases can constitute harm according to CCA. Consider the following example:

On the plausible assumption that Robin would have been better off had Batman given him the golf clubs as planned, CCA implies that Batman's keeping them for himself harms Robin. But this seems incorrect; although Robin is not benefitted, this does not mean he is harmed, and the claim that Batman has harmed him seems preposterous. As Ben Bradley points out, “Merely failing to benefit someone does not constitute harming that person.”Footnote ⁸ Because Batman merely fails to benefit Robin, Robin suffers no harm. Since CCA entails otherwise, this suggests that CCA is false.

There are several proposed responses to this objection in the literature. In previous work, I argued that in cases such as these, our intuitions concerning whether a person suffers harm rather than experiencing a failure-to-benefit are context-dependent and affected by whether the agent of harm plays an active or passive role in bringing the harm about.Footnote ⁹ On this view, we are inclined to judge that harm occurs in cases in which the agent plays an active role in initiating the effect in question, whereas we are inclined to regard cases in which the agent plays a passive role in allowing the effect in question to occur without actively bringing it about as non-harmful failures-to-benefit. We are inclined to make different judgments about structurally identical cases depending on whether the agent played an active or passive role in bringing about the effects. Though I was careful to stress that on my view CCA implies that there is no in-principle distinction between failure-to-benefit and full-blooded harm, the claim is that the active/passive distinction explains why we regard some cases as harms and others as non-harmful failures-to-benefit.Footnote ¹⁰

According to this solution, Batman's passive role in refraining from giving the clubs to Robin inclines us to judge that he has merely failed to benefit Robin, which leads us to conclude that Robin has not been harmed. However, because CCA does not support a distinction in principle between harm and failure-to-benefit, this view entails that, ultimately, failures-to-benefit are in fact harms, and that we are responding to an irrelevant detail when we intuitively deny this (though I also suggested that the active/passive distinction affects our intuitions about what would have happened had things been otherwise).

Nathan Hanna argues that no response that denies that Batman harms Robin is satisfactory. Hanna claims that if we deny that Batman has harmed Robin in Golf Clubs, then we must also deny that the Joker harms Robin – after all, convincing Batman to keep the clubs was the Joker's intended instrument of harm. But it also seems that the Joker has harmed Robin: from his perspective, the plan worked beautifully. Batman's keeping the clubs therefore constitutes a harm for Robin.Footnote ¹¹ Whatever else happens, if Batman doesn't harm Robin, then neither does the Joker, and the Joker harms Robin.

In another recent article, Duncan Purves gives an importantly different argument that my contextualist response is ineffective. He writes:

If CCA entails that there is no in-principle distinction between failures-to-benefit and harm, and CCA controls our judgments about whether harm has occurred, we will judge that cases like Golf Clubs involve harm. The fact that we do not judge that Batman harms Robin means that we are committed to the view that CCA gets it wrong. Purves then develops a sophisticated and plausible variation on CCA that directly incorporates the distinction between doing harm and merely allowing harm to occur. He writes: “the difference between mere failures to benefit and preventative harms is that preventative harms, but not failures to benefit, make a person occupy their actual well-being level.”Footnote ¹³ Purves defends the view that an event, e, harms S just in case e makes it the case that she is worse off than she would otherwise have been; whereas e fails to benefit her just in case she is worse off than she would otherwise have been but e does not make her occupy that well-being level. Purves argues that this account dovetails with an intuitive and plausible account of how and under what circumstances harm can be attributed to an agent. On this view, S₁ harms S₂ just in case S₁ performs an action, a, such that a makes S₂ worse off than she would otherwise have been; whereas S₁ fails to benefit S₂ just in case S₂ is worse off than she would otherwise have been but no action of S₁ makes this be the case.Footnote ¹⁴

Finally, Purves replies to Hanna's objection by claiming that Hanna's second premise is false. According to Purves, although there is some plausibility to the claim that the Joker harms Robin, it is no less plausible to deny that Batman harms him. If we deny that Batman harms Robin, we can resolve the objection by concluding that the Joker also does not harm him.

4. Harm and attributions of harm

My current view is that each of these replies rests on an incorrect assumption. Many philosophers who have contributed to the literature on harm have tended to run together claims about whether someone has been harmed with claims about attributions of causal or moral responsibility for harm, without carefully distinguishing between them, and have not been careful enough to make a related distinction between claims about what harm is from claims about what it is to do harm.Footnote ¹⁵ Bradley's argument based on Golf Clubs moves from the claim that Batman has not harmed Robin to the claim that Robin is not harmed. My earlier reply to Bradley is based on the claim that because Batman's role in making Robin worse off required no intervention on his part, Batman merely fails to benefit Robin, without harming him. Hanna's response to my earlier view moves from the claim that whatever harm to Robin is attributable to the Joker must also be attributable to Batman.

I believe that this is a mistake. The fact that Robin has suffered harm does not entail that Batman harmed him, or that Batman did him any harm, or that Batman is responsible for the harm. Instead, my view is that no such appeal to the distinctions between active and passive agency and/or doing and allowing is necessary to resolve the problem of omission, and that there is a simpler solution available. This solution distinguishes between facts about the existence of harm and facts about attributions of having done harm to persons or other responsible entities. I suggest that we see CCA as a view about what harm is, and not a view about what it is to do harm, or how responsibility for harm is to be attributed to agents or other objects.

Recall, for example, Rachels's famous pair of contrast cases involving Smith and Jones. Smith and Jones each stand to inherit a fortune in the event that his six-year-old cousin dies. Smith, motivated by monetary gain, kills the child by drowning in the bath; Jones, similarly motivated, allows the child to die by drowning in the bath. Rachels explains:

While Rachels saw this pair of cases as a counterexample to the view that actively causing death is generally worse than passively allowing it to occur, it is also useful in showing that attributions of harm do not always track the doing/allowing distinction. For it is clear that Jones's cousin is no less harmed by his death than Smith's cousin is, and it is also clear that the passive nature of Jones's behavior is no barrier to attributing this harm to him. While Smith harms his cousin by actively holding his head underwater, Jones harms his cousin by passively not helping as he drowns. Although Jones does not play an active role in causing this harm, Jones is nevertheless morally responsible for it, and we can attribute the harm to him on this basis.Footnote ¹⁷

Let's see how we can apply this insight to Golf Clubs. CCA entails that Robin is harmed, but does not entail that Batman does him harm. Some find the claim that Robin suffers harm to be counterintuitive, but we can begin to see how the case could be harmful by attending to the details of the possible consequences of the fact that Batman did not give the golf clubs to Robin. It is easy to imagine a possible future in which Robin has been harmed as a result of not receiving new golf clubs at that moment.Footnote ¹⁸ Perhaps this decision came at a pivotal time in Robin's development as a golfer, when having a new, high-quality set of clubs would have made an enormous difference in his skills and his relationship to the game. Suppose that in the nearest world in which Batman gives the clubs to Robin, Robin excels at golf, due in part to a burst of renewed enthusiasm that produces a dramatic increase in practice hours, and in part to the superior quality of the clubs themselves, which makes his practice much more effective and enjoyable. In actual fact, none of this happens and he remains an intensely frustrated mediocre golfer for the rest of his life. Perhaps he looks back wistfully at that time, when he was first developing real skills as a golfer, and reflects upon how his development might have gone differently if he'd had a set of clubs that he really liked and cared about. Robin might not blame Batman, but he might wonder what might have been if Batman had given the nice new set of clubs he bought for himself at around that time to him, instead. When we consider the effect that not getting the golf clubs has on Robin's well-being, we can start to see how this might have come to him as a harm.Footnote ¹⁹

However, even if Robin's not receiving a new set of golf clubs at that time is a harm, we nevertheless need not attribute this harm to Batman. For Batman to have harmed Robin is for Batman to be causally or morally responsible for the harm, and as the case is described Batman does not cause the harmFootnote ²⁰ and there does not appear to be reason to suspect that he is morally responsible for it. We balk at attributing this harm to Batman because it is clear that Batman is not causally responsible for it, and the details of the case do not support his being morally responsible for it (though the details of the case do not rule it out). CCA is a theory of the nature of harm, not a theory about who has harmed whom, nor a criterion for causal or moral responsibility for harm. The force of the Golf Clubs case is derived from the intuition that Batman does not harm Robin; once we see that Robin can be harmed without Batman having harmed him, the purported counterexample loses its force.

It should also be emphasized that CCA does not imply that a person is harmed every time she is not benefitted. CCA implies that failures-to-benefit constitute harms only when the person fails to receive a benefit that she otherwise would have received. It is not the case that CCA implies that I am harmed when I sit and watch TV instead of finding one million dollars in cash in between my couch cushions, because it is not the case that I find a million dollars in between my couch cushions in the nearest possible world in which I do not watch TV: there is no million dollars cash in between my couch cushions to be found, and there is no nearby world where I find it. CCA therefore implies that not being benefitted constitutes a harm only when it is true that I would have been so benefitted had things been otherwise – when I would have been benefitted had nothing happened that prevented it. CCA implies that such preventative harms are genuine harms; it does not imply that failures-to-benefit are harms more generally.

To see this more clearly, consider the following example:

It may seem that CCA implies that this case involves some harm to Jimmy, since Clark could have benefitted him but chose not to. But CCA does not have this implication. CCA does not imply that we suffer harm whenever we could have been benefitted, or might have been benefitted, but are not. CCA implies that we suffer harm when we would have been benefitted but are not. Since Clark, like most people, normally does not go around randomly giving $50 bills to people for no reason, there are no suitably close possible worlds in which Clark gives this $50 bill to Jimmy. In ordinary circumstances, the nearest possible worlds in which it is not the case that Clark decides not to give Jimmy the $50 bill are worlds in which Clark does not even consider giving it to him. Since Jimmy is not worse off in any of the nearest worlds in which Clark does not decide against giving him $50, CCA implies that he is not harmed.

On the other hand, consider the following modification:

In this case, CCA does imply that Jimmy suffers harm, because there are nearby worlds in which the $50 bill is not incinerated, and he is better off in all such worlds. But this implication is defensible: Zod's action has a negative effect on Jimmy's well-being (on the plausible assumption that it would be better for him to receive $50), and therefore constitutes a preventative harm.

5. How to attribute responsibility for harm

If the foregoing is correct, we must distinguish between principles that tell us what harm is, on the one hand, and principles that explain who has done harm, or who has harmed whom, or how causal or moral responsibility for that harm is distributed. Harms of omission are cases in which the victim has suffered some harm, and in which the most plausible immediate agent of harm plays a passive causal role in bringing that harm about by refraining from providing a benefit that would otherwise have been imminent.

In cases involving passive agents of harm, the agents do not initiate or contribute to any causal chain of events whose result is a harm-constituting event for the victim; they merely refrain from interfering with a previously established chain. In such cases, there can be doubt as to whether moral responsibility for the harm that befalls the victim can be attributed to the agent. But if we bracket this question, concerning to whom moral responsibility is attributable, and focus instead on the question CCA attempts to answer, concerning whether there exists a harm at all, the alleged objection to CCA collapses. In each case, CCA correctly classifies what happens as a harm. Questions regarding the distribution of responsibility for these harms are beyond the purview of CCA.Footnote ²¹

This is by design. CCA is intended to be morally neutral, and does not presuppose that causing or otherwise bringing about harm is intrinsically or necessarily morally wrong, or is inherently vicious, or must involve bad intentions.Footnote ²² The question of whether S suffers harm is not identical to the question of who harmed her, or whether anyone is morally responsible for harming her, or what the moral status of any particular action is. Intuitively, if S is drowning and I eat ice cream instead of saving her, then she is harmed – by drowning (for let us suppose that she would have been better off had she not drowned) – but it is not clear that I have harmed her, in that I play no active or direct role in bringing about the harm.

However, it may nevertheless be the case that I am morally responsible for the harm, if I had an obligation to assist in preventing the harm (say, because I was serving as a lifeguard at the time of the drowning) and I failed to satisfy it (say, because I allowed myself to become so immersed in the experience of eating my ice cream that I failed to take notice that she was drowning). CCA, by itself, does not directly entail any particular position with respect to such attributions of harm, because CCA is a theory about the nature of harm and benefit, not a theory of moral responsibility, praise and blame-worthiness, or the causal antecedents of harm and benefit.

In order to generate results concerning the proper attribution of harm, we will need to supplement CCA with separate principles that tell us how such attributions are to be made. I suggest that harm is attributable to an agent when that agent is causally or morally responsible for the harm:

The thought is that one person has harmed another when she is responsible for that harm, in either a causal or moral sense. These types of responsibility come apart, and either kind of responsibility is sufficient to support an attribution of having done harm. If a person is causally responsible, her causal role in bringing harm about can involve her being morally responsible, but it might not. A person might be morally responsible for intentionally causing harm to someone, or for having caused an unintended harm, for example by negligence, and if so the harm is attributable to her; a person might cause harm via a non-negligent, unforeseeable accidental process, and if so then the harm is attributable to her in a causal sense, but not in any moral sense.

The basis for HiR is that causal responsibility and moral responsibility come apart, and that either type of responsibility is sufficient for harming, but neither is necessary. We do harm if one or the other of two conditions is satisfied: moral responsibility, or causal responsibility. So if S₁ permissibly causes harm to S₂, then the causal responsibility condition says this is sufficient for it to be the case that S₁ harms S₂; and if S₁ non-causally does harm to S₂ by (say) wrongfully failing to prevent harm, without causing it, then the moral responsibility condition says this is sufficient for it to be the case that S₁ harms S₂. HiR is therefore designed to reflect the intuition that doing harm may or may not be morally wrong.

We restrict attribution principles to final harm because instrumental harms are harms only in virtue of their connection with some final harm. Instrumental harms are harms only in a derivative sense. If a harm is attributable to someone, this attribution must ultimately be in virtue of a connection with a final harm. A harm can be attributable to an agent in virtue of her relationship to an instrumental harm, but just as the harmfulness of an instrumental harm must ultimately be derived from that of a final harm, an agent's responsibility for an instrumental harm must ultimately be derived from a connection with a corresponding final harm.

6. Resolving the problem of omission

Let us now consider how this set of principles applies to cases of omission. In Golf Clubs, CCA implies that the kind of preventative harm Robin experiences constitutes genuine harm. Batman's role in bringing about the harm to Robin is passive; Batman does not make an active contribution to a causal chain that affects Robin's well-being, and so he does not satisfy CR. Whether he satisfies MR depends on whether Batman had an obligation to give the clubs to Robin, or to otherwise prevent him from suffering this harm, and the case as described gives us no reason to believe that any such obligation exists. It is therefore doubtful that Batman satisfies HiR.

On my proposal, it is therefore doubtful that the harm to Robin is attributable to Batman. The combination of CCA and HiR implies that, although Robin suffers harm, it is not the case that Batman harms him.

Let us now turn to the implications of HiR for the Joker's involvement in the case. The Joker's role is more active than Batman's, in that the Joker actively convinces Batman to change his mind and keep the clubs. But although his role in causing Batman to keep the clubs is active, his role in influencing Robin's level of well-being is not. That is, like Batman, the Joker does not make a contribution to a causal chain that has an impact on Robin's well-being. The Joker harms Robin not by directly affecting his well-being, but indirectly, by disrupting a chain of events that would have benefitted him. According to HiR, this is not sufficient for responsibility. Although the Joker causes Batman to change his mind, and CCA entails that Batman's changing his mind is a harm for Robin – Robin is, after all, better off in all the nearest worlds in which Batman does not change his mind – this harm is instrumental, not final, and CR does not require us to attribute harm in virtue of a causal connection with a mere instrumental harm. According to CR, responsibility requires a causal connection with a final harm.

However, on the plausible assumption that the Joker's motivation for interfering was malicious, and that this interference was not in the service of any legitimate or morally justifying purpose, it is plausible that this interference constitutes a violation of the duty of non-maleficence. If so, then the Joker's behavior satisfies MR. If so, then Robin is harmed, and this harm is attributable to the Joker in virtue of his moral relationship with it. It's not that the harm is attributable to the Joker because his interference caused Robin to be harmed; rather, the harm is attributable to the Joker because his interference violated a moral obligation. If we imagine that this interference was motivated by beneficence, not malice – suppose that it was Superman, not the Joker, who interfered, and that this interference came in the form of a suggestion to donate the clubs to a Justice League charity auction benefitting the hungry, and that Superman did not know that the clubs were meant for Robin – the intuition that the interference supports an attribution of harm is significantly weakened.

HiR, CR, and MR are also useful in explaining why Purves's response fails. Purves claims that CCA falsely implies that Robin is harmed in Golf Clubs. But when the details are clearly spelled out, CCA's implication is correct: a person is harmed when she fails to receive a benefit that she would otherwise have received. Robin is therefore harmed when he does not receive a new set of golf clubs, but this harm is not attributable to Batman, and CCA does not say it is. CCA is silent on questions of who has harmed whom. CCA is therefore compatible with the judgment that Batman does not harm Robin. These principles also explain why Hanna's argument for the conclusion that Batman has harmed Robin is unsound. Hanna argues that if we deny that Batman harmed Robin, then we must also deny that the Joker harmed Robin, and it is absurd to deny that the Joker harmed Robin. But Hanna's conditional premise is false: it is possible for the Joker to have harmed Robin without Batman having harmed him. It is possible, that is, for the harm to Robin to be attributable to the Joker without also being attributable to Batman. HiR reveals that these facts about attribution of harm are independent of one another. It is therefore not the case that denying that Batman harmed Robin requires us to deny that the Joker harmed him.

More generally, the fact that a harm has occurred does not, by itself, entail that this harm is attributable to this or that person, or to anyone at all. The fact that someone has been harmed does not entail that there is anyone who caused it, or who is morally responsible or blameworthy for it. For a person might cause a harm to occur in a blameless way, as in the case of accidental non-negligent harms, or a harm might be caused to occur without the involvement of any moral agent whatsoever. The meteorite that caused the Cretaceous/Paleogene extinction event caused harm to a large number of dinosaurs, but the meteorite is not blameworthy.

7. Objections and replies

The argument of the previous section was that HiR, CR, and MR allow us to resolve the argument from omission, by allowing us to accept the implication of CCA that harm occurs in cases of omission but without requiring us to accept intuitively unacceptable attributions of this harm to blameless agents. In this section, we will consider possible objections to this solution.

One objection claims that CR requires us to attribute harm to Batman in Golf Clubs because Batman's forming an intention or making the decision to keep the clubs counts as his initiating or playing a causal role in Robin's being harmed. For if Batman hadn't formed that intention or made that decision, he would have given the clubs to Robin as planned, and Robin is worse off as a result of his having formed it. Batman's forming the intention to keep the clubs is what caused him to keep them, and Batman's keeping the clubs is what harms Robin. Since Batman plays an active role in making the decision, it seems that Batman satisfies CR.Footnote ³²

We can resist this implication, however. Although Batman forms an intention to keep the clubs, and Robin is worse off in the world in which Batman forms this intention than he is in the worlds nearest it in which he does not, this intention nevertheless has no direct effect on Robin's well-being. Batman merely refrains from intervening: he does not initiate any causal chain that has any effect on Robin's level of well-being, and does not make any causal contribution to any such chain. Batman's involvement in the harm to Robin is purely passive; to claim otherwise is to give up the ghost of the argument from omission. Although Batman was about to initiate a causal chain that would have caused an increase Robin's well-being, at the last moment he refrained from doing so, which means that Batman did not play any role in initiating or contributing to a causal chain that had effects concerning Robin's level of well-being. CR therefore does not support attribution of this harm to Batman.

The friend of CCA should acknowledge that Robin would have been better off had Batman given him the clubs, and, a fortiori, that Robin would have been better off had Batman not made the decision or formed the intention to keep the clubs for himself. That is, the friend of CCA should acknowledge that Golf Clubs involves Robin being harmed, and that the event of Batman making the decision is an instrumental harm. But we can deny that Batman's decision plays a causal role in bringing this harm about, and this leaves us room to deny that Batman does harm to Robin.

A second, closely related, objection claims that CR supports attributing the harm to Batman because Batman keeping the clubs is itself a harm for Robin. After all, Robin is better off in the nearest worlds in which this does not happen. CCA entails that Batman's action in keeping the clubs is a harm, and because Batman himself caused this action, it would seem that CR and HiR entail that the harm is attributable to Batman. One might suspect that, in general, if person S₁ performs an action a, and a harms S₂, then S₁ harms S₂. If so, then it seems that no theory of harm and harming based on CCA can overcome the argument from omission.Footnote ³³

We can see why this objection fails, however, when we notice that HiR, CR, and MR all contain a restriction to final harm: neither HiR nor CR entails that harm is attributable to a person in virtue of a causal contribution to a merely instrumental harm, except insofar as this contribution contributes in a suitable way to a final harm. The thought behind this restriction is that we care primarily about non-instrumental harm in itself, and, because the status of an instrumental harm is parasitic on the existence of a final harm, when we care about instrumental harm at all it is only because of a suitable responsibility-generating relationship with some non-instrumental harm, as spelled out by CR and MR. Batman keeping the clubs is not a final harm, and does not satisfy CR or MR with respect to the occurrence of any final harm. These facts, therefore, do not support attributing this harm to Batman in virtue of his decision to keep the clubs.

Recall the case in which someone is drowning and I eat ice cream instead of saving her, without causing the accident or violating any obligation. This case illustrates that it is possible for an agent to perform an action that constitutes an instrumental harm for someone without the harm being attributable to her, as long as the harm is an indirect result of an action that is otherwise morally permissible. In general, if S₁'s action a instrumentally harms S₂, but a does not make a direct causal contribution to the harm, and S₁'s performance of a did not violate any of S₁'s moral obligations, then it is possible for the harm to S₂ to be unattributable to S₁. If this is correct, then although Batman's keeping the clubs is a harm for Robin, the fact that it is an instrumental harm Batman was not obligated to prevent means that it is nevertheless not the case that Batman thereby does harm to Robin.

Finally, one might object that there is a separate problem that arises in cases like Jughead's Hat, which, unlike Golf Clubs, does not involve the actions of a malicious third party who serves as the agent of harm:

This case involves Jughead being harmed, but does not provide anyone other than Archie to attribute it to. It is, however, plausible to think that if Jughead is harmed, someone or something must have harmed him. Archie is the only eligible candidate, so Archie must have harmed Jughead.Footnote ³⁵

Although the friend of CCA should agree that Jughead suffers harm in this case, we need not agree that CCA entails that this harm is attributable to any person. CCA requires that there be something that harms Jughead, but CCA identifies that thing as the event consisting of Jughead not receiving a new hat at a certain time. It would be different if there were some reason why Archie were obligated to give him a hat. But in the absence of any such obligation, and in light of the passive nature of Archie's involvement, it is open to the friend of CCA to affirm that Jughead suffers harm but to deny that Archie – or anyone else, for that matter – does harm to him.

HiR entails that whether or not the harm to Jughead is attributable to Archie depends crucially on the circumstances surrounding the decision to keep the hat. On the plausible presumption that Archie is not under any obligation to give a new hat to Jughead, MR does not apply. Under the continued assumption that omissions cannot be causes, CR does not entail that Archie's keeping the hat caused harm to Jughead. And since Archie's keeping the hat is an instrumental harm, and Archie does not initiate or contribute to any causal chain that results in a final harm to Jughead, CR does not imply that Archie does harm. The present approach is therefore compatible with the possibility that the harm to Jughead is not attributable to Archie.