Ethics Review Forum: Jonathan Quong’s “The Morality of Defensive Force”

Welcome to our Ethics Review Forum on Jonathan Quong‘s The Morality of Defensive Force (OUP 2020), reviewed by Joseph Bowen.

Below, you’ll find a description of the book, as well as a condensed version of Joseph’s review. Jonathan’s response will appear in the comments. Please join Jonathan and Joseph in continuing the discussion!

Book Blurb:

When is it morally permissible to engage in self-defense or the defense of others? Jonathan Quong defends a variety of novel ideas in this book about the morality of defensive force, providing an original philosophical account of the central moral principles that should regulate its use. We cannot understand the morality of defensive force, he reasons, until we ask and answer deeper questions about how the use of defensive force fits with a more general account of justice and moral rights. In developing this stance, Quong presents new views on liability, proportionality, and necessity. He argues that self-defense can sometimes be justified on the basis of an agent-relative prerogative to give greater weight to one’s own life and interests, contrary to the dominant view in the literature. Additionally Quong develops a novel conception of individual rights against harm. Unlike some, who believe that our rights against harm are fact-relative, he argues that our rights against being harmed by others must, in certain respects, be sensitive to the evidence that others can reasonably be expected to possess. The book concludes with Quong’s extended defense of the means principle, a principle that prohibits harmfully using other persons’ bodies or other rightful property unless those persons are duty bound to permit this use or have otherwise waived their claims against such use.

Excerpt from Bowen’s Review

The Morality of Defensive Force […] will no doubt become a cornerstone of work on the justifications and limits of defensive harm. It strikes a great balance between being accessible to those who want to pick it up for a particular topic […] while [also] being a rewarding, well-integrated read.

[…] Chapter 2 focuses on liability to defensive harm. To say that an individual is liable to be harmed is to say that he has forfeited some of his rights against being harmed, and so harming him would not violate his rights nor would it wrong him. Quong introduces a new account of liability, the moral status account[… ][:] “A is liable to defensive harm for Φ-ing when: (a) the evidence-relative permissibility of Φ-ing depends on the assumption that at least one person, B, lacks a moral right, but (b) B in fact possesses the relevant moral right, and thus (c) B faces a threat, or apparent threat, to her rights” (38). […] Chapter 3 defends that we have agent-relative prerogatives to harm non-liable parties.

Whether a victim may defend herself depends on whether her self-defensive act is proportionate and necessary. Chapter 4 introduces a new account of proportionality, on which the stringency of victims’ threatened rights determines the amount of defensive harm that it is proportionate to inflict (112). […] Chapter 5 introduces a revisionary account of necessity, on which defensive harm is unnecessary if it is inconsistent with victims’ duty to rescue others at reasonable cost to themselves (132).

Chapter 6 argues that rights against harm are determined by asking what we can reasonably demand of others, including the evidence that we can reasonably demand they possess (156-60). But the view is not evidence-relative all the way down—whether someone does not have a right against being harmed because they are liable or have consented to being harmed depends on what they have in fact done, and not on duty-bearers’ evidence of what they have done (161-3). Finally, chapter 7 defends a version of the means principle, on which it is impermissible to harm people in ways that make use of things to which they have prior claims[…], unless they are under a duty to suffer that harm or consent to being used in that way.

Let me offer an extended example of one of the ways the book makes for a well-integrated read, which forms the basis for [my] critical comments. In chapter 2, Quong rejects the dominant account of liability, the moral responsibility account, because it deems minimally responsible threateners liable to defensive harm (26-29). In the […] Conscientious Driver example, a careful and conscientious driver finds herself unavoidably swerving off the road towards Pedestrian, whom she will kill […] [T]he moral responsibility account says that Driver is liable to be harmed by Pedestrian, and so has no right against Pedestrian [killing her in] self-defence. Quong thinks this is the wrong verdict. His moral status account says that Pedestrian is not liable: the evidence-relative permissibility of Driver’s setting out on her drive does not depend on the assumption that anyone else does not have a right against being harmed.

One argument for why Driver is liable is that many find it plausible that Pedestrian may kill Driver in self-defence. And how else to explain this unless Driver is liable? Quong has an answer: […] Pedestrian has an agent-relative prerogative to prioritise herself […] (58). Now, an initial problem with agent-relative prerogatives is that most think that we may not grab a bystander and use them as a protective shield against a lethal projectile (81-85). But if we have agent-relative prerogatives to prioritise ourselves, why may we not prioritise ourselves in [this way?] […]

Recall now though, that if there are no agent-relative prerogatives, we have no explanation for why Pedestrian may defend herself in Conscientious Driver, at least if we endorse Quong’s moral status account of liability. And if we think that Pedestrian may defend herself, we ought to reject Quong’s moral status account […]. But Quong does have an answer for why, even if we have agent-relative prerogatives to prioritise ourselves, we may not [use a bystander] as a protective shield: [namely], the means principle, [which] says that we may not prioritise ourselves when doing so would use others’ bodies or something else to which they have a prior claim, unless they are under a duty to suffer that harm or have consented to that harm.

In the remainder of this review, I examine some of the views featured in the preceding argument. Quong’s moral status account says that Driver is not liable to be harmed in Conscientious Driver because the evidence-relative permissibility of Driver’s setting out does not depend on the assumption that anyone else does not have their usual rights against being harmed. Compare this with Resident (23-24): suppose that the identical twin brother of a notorious serial killer breaks down in a remote area. Unaware that his brother has recently escaped from prison, and is now on a killing spree, the innocent twin (“Innocent”) knocks on the door of the nearest house to ask for help […]. The resident (“Resident”) justifiably believes that Innocent is the killer and has been warned that [they] will kill at their first opportunity. Because of this, Resident lunges at Innocent with a knife […]. Quong thinks that Resident is liable to defensive harm. Rendering these verdicts consistent—that Resident, but not Driver, is liable—has proven difficult in the literature. […] [After all, both Resident and Driver act in ways that would not be permissible, were they in possession of the facts.]

Quong’s moral status account finds a difference between these cases. Resident’s action is evidence-relative permissible only because Resident’s evidence supports that Innocent is liable to be harmed. Driver’s action is not (evidence-relative) justified in this way. Rather, […] the expected benefits weighed against the expected costs of the practice of prudent driving make it that any action-token of the practice is permissible. And this is what matters when it comes to liability: “whether you treat others as if they lack rights against the harms that you might impose” (5). [Quong] continues, “[w]hen you act in this way, you treat others as if they are not entitled to equal concern and respect all persons are normally owed, and so it’s appropriate that you bear special liability for your actions”.

[…] […] I am not sure, if there is a morally significant difference between these cases, Quong has found it. It is crucial on his account whether the evidence-relative permissibility of someone’s act depends on others lacking rights that people ordinarily possess—if so, one can be liable; if not, one cannot be liable. But why should that matter? Evidence about whether our actions will harm others and evidence about whether others have made themselves liable to be harmed is not so different. They place equal demands on one. They make our actions evidence-relative permissible in the same way.

To be fair[…], Quong tries a great deal to motivate the idea that the difference between the cases matters (34-39; 94-95; 163-6). For example, he says of paradigmatic culpable threateners that they “cannot treat others as having diminished moral status while insisting others respect [their] moral status even at the cost of suffering serious harm”—this would be to fly in the face of equality and reciprocity (34). […] I am sceptical how much we can learn from this. Equality and reciprocity are morally loaded terms of art, and using them in our analysis often presupposes what we are trying to prove. […] What is more, while Resident does treat Innocent as having diminished moral status, she is evidence-relative justified in treating her so; and to the extent that I think it is morally significant that paradigmatic culpable threateners treat others as having diminished moral status, it is only because they are neither belief- nor evidence-relative justified in doing so.

In addition, the focus the moral status account places on whether one treats others as if they lack rights ordinarily possessed by others leads to some peculiar verdicts about liability. Quong thinks that we can have lesser-evil justifications to harm non-liable parties […]. When threateners have a lesser-evil justification to harm victims, while victims retain their rights against being harmed, their rights are permissibly overridden for the sake of averting the much greater evil. On this picture, […] rights and (directed) duties are pro tanto considerations—they carry genuine weight but can be outweighed. An implication of this is that when threateners act with a lesser-evil justification, they do not treat their victims as if lacking rights that people normally possess. […] It follows that threateners are not liable to defensive harm when acting with a lesser-evil justification given the moral status account. This leads to a strange result. If a threatener is evidence-relative justified in harming their victim with a lesser-evil justification, but in fact has no lesser-evil justification, then she is not liable to defensive harm. Whereas, if she is evidence-relative justified in harming her victim because the evidence supports that the victim is liable, but in fact the victim is not liable, the threatener is liable to defensive harm.

There is some discussion of this problem as it relates to Resident, but I think its significance is not fully appreciated (94-95). Suppose that some soldiers are successfully fooled by a totalitarian regime into believing their regime is justified, and under repeated threats by their evil neighbours. Their neighbours are in fact a just regime. The soldiers are told there is a terrorist camp over the border. In fact, the ‘terrorist camp’ is a village of innocent civilians. […] In a first iteration of our case, discussed in the book, suppose the soldiers are told the camp is made up of only enemy combatants, who are liable to be killed. Quong’s view implies that the duped soldiers are liable to be killed in defence of the villagers, since the evidence-relative permissibility of their conduct depends on assuming that the enemy combatants have made themselves liable to be harmed (33-35). So far, so good. But now suppose, in a second iteration, the soldiers are told the camp has been vacated by the terrorists, and some civilians have (unknowingly) moved in. However, the terrorists could return, and enough good can be done by destroying the infrastructure of the camp that the soldiers have a lesser-evil justification to destroy it, while foreseeably harming some innocent civilians […]. Quong’s view implies the soldiers are not liable in this case since the evidence- relative permissibility of their conduct does not assume that their victims lack rights that people ordinarily possess. This is deeply counterintuitive. […]

41 Replies to “Ethics Review Forum: Jonathan Quong’s “The Morality of Defensive Force”

  1. I’m very grateful to Joe for his generous and insightful review. He raises a series of difficult challenges for some of the central claims that I defend in the book. I’ll try to say something about each of those challenges here.

    1.
    First, Joe presses a worry about my view of liability. After noting that the moral status account draws a distinction between Conscientious Driver (nonliable) and Resident (liable), Joe says: “I am not sure, if there is a morally significant difference between these cases, Quong has found it. It is crucial on his account whether the evidence-relative permissibility of someone’s act depends on others lacking rights that people ordinarily possess—if so, one can be liable; if not, one cannot be liable. But why should that matter? Evidence about whether our actions will harm others and evidence about whether others have made themselves liable to be harmed are not so different. They place equal demands on one.”

    But I don’t agree that these forms of evidence are not so different, nor that they place equal demands on those who might act upon such evidence. Suppose D has evidence that A is wrongfully threatening to break her arm. Let’s assume this wrongful threat would make it proportionate to break A’s legs in defense. Now compare this with a potential lesser evil case. What kind of evidence would D need to possess before having a lesser evil justification for breaking nonliable B’s legs? The fact that doing so would save D (or anyone else’s) arm from being broken would not be nearly enough. She would need to justifiably believe that a lot of good could be done—probably more than one life would need to be saved for A to have a lesser evil justification for breaking B’s legs.

    These judgments—if sound—are one way of illustrating what a grave matter it is to judge someone liable to harm. Once someone is liable, we can permissibly discount their interests against suffering harm to a great extent, and we can permissibly refuse to bear costs or risks that we would be required to bear for others. Making a mistake about someone’s liability is thus, other things being equal, a much bigger deal than making a mistake about whether harming someone is permissible for reasons unconnected to their liability. When we judge someone to be liable, we discount their interests relative to others—sometimes very dramatically. When D makes this judgment about A and it’s an error, I think D cannot credibly insist that she bears no special responsibility for the threat she poses to A—that A must extend to her the equal moral consideration that she failed to extend to A. But things are different when we make a reasonable mistake about, for example, a lesser evil calculation. In this case we don’t treat anyone as having diminished moral status.

    2.
    As Joe notes—and as the preceding discussion indicates—I appeal to considerations of both equality and reciprocity to motivate the moral status account, but Joe is skeptical. He says “Equality and reciprocity are morally loaded terms of art, and using them in our analysis often presupposes what we are trying to prove. For example, we could just as easily say, in support of the view that mistaken threateners are not liable to be harmed, that equality and reciprocity demand that we not hold others responsible for acting in ways in which we ourselves would act were we in similar situations.”

    I don’t think we can just as easily say this. If I was faced with a lethal threat I think there’s a good chance that I might panic and do something wrong to try and save myself—for example, I might grab a bystander and use him as a shield. But even if we suppose that most people would act in a similar way, this doesn’t change the fact that those who act in this wrongful way can be held responsible. How we—or most people—would act in a similar situation is not the correct test to determine whether equality or reciprocity require that others be held responsible or liable for their actions. Consider Feinberg’s well-worn case of the hiker who must break into a cabin to save himself from a dangerous storm. The fact that most of us would also break into the cabin if we were in the hiker’s position doesn’t entail that the hiker cannot be held liable for compensating the cabin owner.

    The moral responsibility account and the moral status account are in agreement that people can be liable for actions even if it’s true that any reasonable person might perform the same act under those conditions. But the moral responsibility account (whose proponents also sometimes motivate their view by appeal to equality) is much broader than the moral status account—it holds persons liable whenever they foreseeably risk imposing harm on others and the harm turns out to be fact-relative wrongful. So the worry that Joe has expressed above would apply with even more force to the moral responsibility account favored by McMahan, Otsuka, Gordon-Solmon, and others.

    To get an account of liability consistent with the suggestion floated by Joe—that we not hold people responsible for acting in ways that we would ourselves act—we probably need to make culpability a necessary condition for liability. Of course this is a view that some have defended—most notably Kim Ferzan—but that view comes with some well-known and very significant costs (e.g. it entails that the Duped Soldiers in both variants are not liable to harm), and I also don’t think this view can be successfully rooted in appeals to equality or reciprocity.

    3.
    In addition to his skepticism about the rationale for the moral status account, Joe also has worries about its implications in particular cases. He imagines a variant of my Duped Soldiers case where “the soldiers are told the camp has been vacated by the terrorists, and some civilians have (unknowingly) moved in. However, the terrorists could return, and enough good can be done by destroying the infrastructure of the camp that the soldiers have a lesser-evil justification to destroy it, while foreseeably harming some innocent civilians who have moved in. Quong’s view implies that the soldiers are not liable in this case since the evidence-relative permissibility of their conduct does not assume that their victims lack rights that people ordinarily possess.” This, Joe says, “is deeply counterintuitive”.

    I am not sure exactly what Joe finds counterintuitive: it could be the bare claim that the soldiers are not liable, or it could be the practical implications that follow from this fact (e.g. that third parties cannot permissibly defend the civilians from the soldiers).

    If it’s the former, I’ll register modest skepticism that we have many clear and reliable intuitions about the bare notion of moral liability to defensive force. This kind liability is a technical notion that has only gained prominence amongst a small group of philosophers over the last few decades. My own sense of how the literature has developed is that we don’t have clear judgments about when people are liable to harm that we’re trying to capture with an account of liability. Rather, we have clear judgments in various cases about when defensive force is or isn’t permissible, and different philosophers have tried to design accounts of liability that can explain at least many of these judgments while also being rooted in some independently plausible moral ideas (e.g. ideas about distributive fairness).

    So let’s turn to the second interpretation—on this version what Joe finds counterintuitive are the implications that follow if we hold that the soldiers are not liable in his version of the case. Joe expresses two worries along these lines: (1) third parties would not be permitted to intervene on behalf of the civilians, and (2) although on my view the civilians could permissibly fight back, they couldn’t use the soldiers bodies or property in the course of defending themselves.

    I don’t have a lot to say beyond the fact I don’t see these implications as obviously counterintuitive. Someone who acts upon a perfectly justified but mistaken assumption about a lesser evil justification seems to me to be morally on a par with someone who acts upon a perfectly justified but mistaken judgment that their act poses only a minimal and permissible level of risk to others. Neither agent discounts anyone’s claims against harm nor treats anyone as having diminished moral rights. Since I don’t think it would be permissible for a third party to intervene and kill Driver in Conscientious Driver, I don’t think it’s permissible for third parties to intervene in Joe’s version of Duped Soldiers. For what it’s worth, I suspect some people might have asymmetric reactions: they might agree with me about Conscientious Driver but balk at the implication in Joe’s version of Duped Soldiers. I think that divergence—if it exists—might be explained by the fact people find it difficult to genuinely believe that the soldiers could be justified in holding the false belief about possessing a lesser evil justification, whereas it’s familiar and easy to understand how a conscientious driver arrives at the justified belief that their act is extremely unlikely to harm anyone.

    Joe’s second point concerns the means principle, so I’ll turn more directly to the challenges he raises about my account of the means principle.

    4.
    One worry Joe expresses about my formulation of the means principle concerns the focus not only on uses of a person’s body, but also other bits of the world over which they have rightful prior claims. Joe asks us to imagine a standard trolley case, where the trolley can be redirected onto the side track, saving five but killing one, but with the added wrinkle that you can only redirect the trolley onto the sidetrack by using a wrench that belongs to the one. My view entails that this would be a wrongful violation of the means principle. Joe says “This looks implausible to me. I do not need either to inquire whose property the wrench is or to refrain from turning the trolley if I know the wrench is Victim’s”.

    I don’t agree. I do not think that our rights over our bodies are, in some deep sense, categorically different than the rights we have over other parts of the world. Of course many of the rights we have over our bodies are typically much more stringent than rights over parts of the external world, but this is because we are often more vulnerable to serious harm via incursions on the body and our plans and control over our lives can often be more dramatically thwarted by incursions on the body. But this is contingent and not always true. If you steal someone’s oxygen tank or heart medication, this can be as harmful and serious as stealing a vital organ. It’s not credible, I think, to suppose that the means principle could apply to the use of other people’s bodies, but cannot apply to the use of other people’s oxygen tanks or heart medications. Of course under typical conditions, wrenches aren’t very valuable bits of property, and so under typical conditions using someone’s wrench without consent would be a relatively minor transgression of a right. But in the case Joe imagines things are atypical: if you use the one person’s wrench, he will die. I thus don’t find it at all counterintuitive to suppose the means principle prohibits your using the wrench. Doing this would be morally on a par with stealing and using the one person’s heart medication that he needs to survive.

    5.
    Joe also worries about the rationale I offer for the means principle. As he says, what we want from a rationale is an explanation of why, holding the degree of harm constant, moral rights constraining others from using our bodies/property should be more stringent than our rights that constrain others from merely affecting our bodies/property. But Joe doesn’t think what I say about independence and control offers a way to draw this distinction. He thinks that rights against useful harming and rights against non-useful harming similarly “endow B with control over the weight of A’s reasons for and against acting. So, appealing to control, and the independence this affords us, is not on its own helpful in grounding the means principle”.

    I think this is too quick. Although it’s plausibly the case that all claim rights give the right-holder (at least mentally competent adults) some control over the behavior of the duty-bearer, it doesn’t follow that there’s no difference in the degree and kind of control. If the means principle as I understand it is correct, then there’s part of the world that is unavailable for A to use without B’s permission. B is the only person who requires no one else’s permission to make use of that part of the world. So long as other duties/obligations are not in play, B has exclusive authority to use his part of the world as he sees fit. Others may sometimes permissibly act in ways that incidentally damage B’s part of the world or render it less valuable, but only when they can do so by using parts of the world which they have permission to use, and only when there is some proportionate justification. That is a kind of independence we would lack if all we had were rights against mere harm. This kind of independence explains why—beyond reasonable duties of rescue—others cannot appropriate our bodies or personal property even when doing so would bring about very good outcomes.

    In his closing paragraph, Joe reminds us of one way the various parts of the book are connected. He suggests that his worries about the means principle entail that “we ought to pause before endorsing the means principle”, but this should then give us pause about my account of agent-relative prerogatives, and this then gives us reason to doubt the moral status account of liability. But even if you share some of Joe’s worries about my particular formulation of the means principle, note that you might still endorse some other variant of the means principle (or some principle in the neighborhood), which could then be plugged back into the rest of the wider argument without changing too much.

  2. Hi Joe and Jonathan.
    Jonathan states that “[l]iability, on its own, is incapable of providing a justification for the use of defensive force” (6), and I agree. Since the book is entitled The Morality of Defensive Force, Jonathan wants to explain under which circumstances defensive force is morally justified. It is therefore disappointing when he uncritically makes use of “Parfit’s distinction between three different standards of moral evaluation.” He informs us: “An act is wrong in the fact-relative sense when it would be wrong if we knew all the relevant facts. An act is wrong in the belief-relative sense if it would be wrong if our beliefs about the facts were true. Finally, an act is wrong in the evidence-relative sense when our act would be wrong if the relevant facts were what the available evidence gives us sufficient (or apparently sufficient) reason to believe they are.” He applies these same distinctions to permissibility and thus justification. (20) Unless the meaning of the second occurrence of “wrong” in each of these three sentences is explained, the definitions are uninformative. Moreover, if it is to be understood in one of these three senses themselves, then this obviously makes some of these sentences circular and others incoherent (or even unintelligible). If, however, it is to be understood in a completely different sense – namely, according to Parfit, in the “ordinary sense” – then it is safe to assume that this different or “ordinary” sense is the fundamentally important one (after all, it appears in all three definitions), and one would like to know when exactly the use of defensive force is morally justified or permissible in this ordinary sense – that is, morally justified, period.

    In fact, the ordinary sense of justification found in law is a mixed one, and this is also correct from a moral point of view. Moreover, the mixed justification is not one that combines the three Parfitian senses but rather does without them. The reason for this is that you cannot combine the three justifications, primarily because fact-relative justification for the use of defensive force is virtually unattainable (in the real world it is practically impossible, blind luck of astronomical proportions set aside, to stay within the confines of the necessity condition, fact-relatively understood.) Yet Jonathan seems to believe that the three standards can “converge” (20). No, in the real world they cannot (or almost never) in the dimension of permissibility (unlike in the dimension of wrongness). But this means that in the real world real acts of justified self-defense (in the ordinary and relevant sense) will never be justified in the irrelevant sense that they are justified in all three Parfitian senses simultaneously. In short, I do not think that relying on Parfit’s distinctions is a good start in our search for moral justifications for the defensive use of force. (For sustained arguments against the tripartion, for the practical impossibility of fact-relative justification, and for the mixed account, see my Self-Defense, Necessity, and Punishment [SNP], sections 2.2.2.4 and 2.2.5; as well as The Ethics of War and the Force of Law [EWFL], section 3.2.2.)

    But Jonathan is also interested in liability (for his definition, see p. 21). He starts his book by telling us that Albert, who wrongfully assaults Betty, is liable to defensive force (1). His official explanation for this liability is his “moral status account of liability to defensive harm,” which holds: “A is liable to defensive harm for Φ-ing when: (a) the evidence-relative permissibility of Φ-ing depends on the assumption that at least one person, B, lacks a moral right, but (b) B in fact possesses the relevant moral right, and thus (c) B faces a threat, or apparent threat, to her rights.” (38) (It is not entirely clear whether he means this “when” as a mere “if” or as an “if and only if,” but the last sentence on p. 41 would suggest the latter interpretation.)

    It is unclear how this account is supposed to explain Albert’s liability. After all, unless otherwise stated, “we can assume that the three standards [of wrongfulness] converge” (20), so Albert doesn’t actually have a justification for his Φ-ing, that is, for his assault on Betty; and since there is no evidence-relative permissibility in the first place, this non-existent permissibility cannot depend on anything. Thus, if the “when” is meant to be equivalent to an “if,” the account does not imply Albert’s liability, and if it is an “if and only if,” it implies his non-liability. Perhaps some might be tempted here to reply that Jonathan’s account needs to be interpreted in a counterfactual way: Albert is liable to defensive harm when he would only then have an evidence-relative justification for his assault on Betty if he had evidence that she is liable. Unfortunately, this condition is not fulfilled, for Albert would also have an evidence-relative justification for assaulting Betty if he had evidence that this is the lesser evil (or, according to Jonathan, that he has an agent-relative prerogative for his assault). So the counter-factual interpretation does not imply Albert’s liability either (or even implies his non-liability). In short, on both of the two interpretations, factual and counter-factual, Jonathan’s moral status account of liability cannot even explain liability in paradigmatic cases of self-defense.

    (It should be noted that elsewhere Jonathan balks at a counter-factual interpretation of his Resident example. In that example, “Resident has sufficient evidence to believe that there is a liable attacker at the door. Given his evidence, Resident has a liability-based justification but no agent-relative justification … In considering whether a defensive agent’s justification depends on the assumption that a target is liable, we must hold constant the other facts of the case, and only eliminate the evidence that supports the assumption of liability.” (93) That is fine in the case of the Resident, since he does have a justification, and so it is reasonable to ask on what that actual justification is based. Yet Albert does not have a justification, and it is not reasonable to ask on what non-existing justifications actually “depend.” We cannot “eliminate the evidence that supports the assumption of liability” here because there is no such evidence.)

    As an afterthought, I also fail to understand why someone who has an “evidence-relative justification” to harm someone should only become liable to counter-force if his target is not liable to be harmed. After all, Albert might have an evidence-relative justification, believe that he hasn’t, and try to beat up Betty for the mere fun of it (yet she was liable herself as she was about to unjustifiably and severely harm innocent Bob, but in light of Albert’s threat, she punches Albert instead.) While the formulation of his official status account suggests that Albert would not be liable to Betty’s punch, this does not square well with parts of the underlying rationale Jonathan offers in defense of his account. For example, he thinks it would “conflict with ideals of equality and reciprocity” if someone said: “I demand that you respect my moral right, even though I am not extending the same treatment to you.” (34) Of course, Betty did not have a moral right not to be punched here, but Albert did not know that. He wanted to inflict harm on her for no good reason. What “standing to press claims” (37) does he have if Betty reciprocates? If he says: “How dare you punch me for no good reasons,” she might well reply: “Look who’s talking, you wanted to punch me with no good reason. And in fact, I, unlike you, actually had a reason: I saw you were about to punch me. So you don’t have standing to complain, pal.”

    It appears best to set Jonathan’s formulation of the “moral status account” simply aside and focus on some of the more precise claims he makes about liability. In fact, it seems that the main thesis of the book, regarding liability, is this: A is liable to defensive force if and only if he unjustifiably violates a right, where the relevant liability-defeating justifications are the lesser evil justification, the agent-relative prerogative justification, and an actual-liability-based justification (as opposed to a mere evidence-for-liability-based justification). Let us have a closer look.

    His reliance on the prerogative justification plays a great role in his rather dismissive attitude towards accounts that claim that justification does not defeat liability and that non-responsible threats can be liable. To wit, he believes that the civilians in the justified bomber case and the man at the bottom of the well in the Nozickian falling man case are intuitively justified in defending themselves (and I agree with this assessment). That, however, would speak for an account that holds justified attackers and non-responsible threats liable. To block this apparently unwanted conclusion, he purports to explain the permissibility of self-defense in these cases with an alternative explanation, namely precisely with the agent-relative prerogative, according to which a defender may harm or even kill a non-liable person, a person with a right not to be killed, even if the defender thereby only averts equivalent harm or death to himself. On the face of it, this seems to be entirely counter-intuitive and to not be taking rights seriously: we do not think that we can kill another person and thereby violate her right to life only because “we may attach to our own lives” and interests a “disproportionate (when viewed impartially) weight.” (70) (I use the term “violate” as neutral between justified and unjustified rights violation here; Jonathan uses for justified violations the term “infringement.”)

    Jonathan, apparently, thinks otherwise. He claims, for instance, that you may divert a trolley away from your own child towards a single bystander who would then be killed instead. (68) This is an irrelevant example. This act of yours is not justified under an appeal to the value that you attach to your child, but rather by a special obligation towards the child. You do not have such a special obligation towards yourself – you have a liberty-right against yourself to let yourself drown in a shallow pond, for example, but you do not have a liberty-right against your son to let him drown although you could easily save him. Second, Jonathan does not only want to justify diverting existing threats, but also creating new ones, for example by shooting at people with vaporizing ray guns (80). So he needs to provide a clear example where it is permissible, in a one-to-one situation, to kill a non-liable person (non-liable in the sense that the person has a right not to be killed in this way) because you value your live more than his. That won’t be possible, though. His violinist case, for instance, on which he puts heavy emphasis, is question-begging: Jonathan just presupposes that the violinist is not liable to disconnection from the other person’s blood stream, but Thomson, who developed the original example, of course thought that he was so liable in the sense that he had no right not to be disconnected. To simply state that the violinist “has done nothing to constitute a forfeiture of his rights” (72) is insufficient to show that he did have a right not to be disconnected to begin with, for if he has no such right, we don’t need a prerogative-based justification to infringe it. For example, even if Betty’s refusal of Albert’s marriage proposal will cause his death by causing his depression and then suicide, this does not mean that Albert ever had a right against such a refusal. His right to life is restricted by her right to matrimonial self-determination, and thus his right to life never included a right against her that she not cause his death by refusal. Likewise, the violinist’s right to life is limited by your right to bodily integrity and self-determination, and therefore he never had a right that you not disconnect him from your blood stream. Jonathan does nothing to show otherwise. (Moreover, Jonathan obviously also presupposes here that people can only forfeit rights by action, but that is mistaken. I will come back to that).

    In contrast, it is easy to provide counter-examples. You and Betty run away from a tiger, there is an escape route that can only be taken once and by one person. Betty is faster and wears her protective suit, only her head is vulnerable. You shoot her in the head, pass her corpse, and escape. Justified? I don’t think so. In Western jurisdictions this is murder or manslaughter, and that is the intuitively right stance to take. Moreover, Jonathan’s account implies (see his remarks on asymmetry and symmetry on pp. 18 and 71) that Betty would not have a liability-based justification (only a prerogative-based one) to open defensive fire at you, although you started the attack and contravened her right to life by shooting at her. To call this implication “counter-intuitive” would be quite an understatement.

    Thus, Jonathan’s attempt to restrict his incredibly permissive stance towards killing non-liable people for your own good by invoking the alleged inviolability of the “means principle” (already described by Joseph) falls short, for the above killing is not excluded by this principle. He also overestimates it in other ways. For example, he states that he “cannot appeal to agent-relative considerations as a sufficient justification for stealing your vital organs, even if [he] will die without them” (60), and thinks that his means principle makes sure of this restriction. Not quite. His means principle only excludes “opportunistic” organ stealing (where you need the organs as a means to your survival), but not “eliminative” organ stealing (where they are no such means) (the same is true for his other two examples on that page). To wit, if a villain programs a drone in such a way that it will kill me in case it detects you (or someone looking exactly like you) still in possession of your organs, and I can only save myself by sedating you, cutting you open, and stealing your organs, then this does not violate Jonathan’s means principle. After all, Jonathan’s test question for determining whether the principle has been violated or not is whether your sudden disappearance with all your property would have averted the threat against me. If it would have, then you have not been used as a means (85, 190). And of course it would have, for the drone would then not have detected you in possession of your organs and therefore not killed me. Thus, if you don’t think that the existence of that drone should provide me with a justification to cut you open and steal your organs (all the while happily conforming with the “means principle” in all its splendor) for the simple reason that I value my life higher than yours, then you better reject Jonathan’s agent-relative prerogative justification. (To avoid confusions, I have endorsed “act-specific agent-relative prerogatives” myself, see SNP, 2.1.2.6 – but I have not passed them as justifications and denied, of course, that they allow to infringe rights in the one to one situations under consideration here. For further criticism of Jonathan’s prerogative account, see ibid., section 3.1.2.)

    Certainly better than Jonathan’s agent-relative prerogative or rights-infringement account fares a liability account, which (partially) explains the justifiability of defensive measures against justified attackers and non-responsible threats by their violating other people’s rights and thereby becoming liable – that is, their relevant rights need not be infringed because they have forfeited them. As already noted, Jonathan does not deem this view worth of serious engagement on his part. In a mere note, he states that he will not discuss what he there calls “the causal account” (calling it rights-violation or liability account would be more appropriate) – he mentions Thomson, Uniacke, and Leverick as proponents – because it has been “subjected to decisive criticism” (19, n. 2). Well, this allegedly “decisive criticism” has been subjected to decisive criticism itself by, among others, Yitzhak Benbaji, Adam Hosein, and me. He entirely ignores this literature. (I say “ignore,” not just “is not aware of,” because he certainly was aware of my article “Justifying Defense Against Non-Responsible Threats and Justified Aggressors: the Liability vs. the Rights-Infringement Account,” Philosophia (2016, for an updated version, see SNP, 3.1.2), where I directly target his views and thanked him for having provided me with feedback. I also sent him the published version. That article was neither the first nor the last time I defended the liability account.) Since he does not address the objections found in this literature, he has not overcome them. Instead, sometime after that footnote, in another chapter, he simply rehearses (on not even half a page) arguments that have already been dealt with more than a decade ago. To wit, against Thomson’s claim that the falling man violates the right of the man on the bottom not to be killed, he states: “[I]s is not plausible to suppose, as Thomson does, that non-responsible threats are guilty [Thomson doesn’t talk of “guilt”] of threatening to violate anyone’s right. This is because, in order for one person to violate another’s rights, there must be responsible agency. For example, although it would make sense to say that you were killed by a boulder that fell from the top of a cliff …, it would not make sense to say that the boulder violated your right not to be killed. The boulder has no agency, thus cannot be subject to any moral duties, and so cannot violate (or even infringe) your rights.” (62) First, of all, as I have shown elsewhere (SNP, 229), Thomson’s argument can easily be reformulated without relying on the “violation” language. Second, that’s not even necessary, however. Even if we grant that only responsible agents can be duty-bearers, this does not amount to granting that they can violate their duties only through agency. As Yitzhak Benbaji points out: “I violate my duty to stay out of your apartment even if I was brought there while I was sleeping. This is because, as opposed to pianos [or boulders, I may add], I am subject to certain duties by virtue of my being an agent. It does not follow that violation of my duties involves agency. After all, I can, unknowingly, unintentionally, and even without doing anything, violate my duty to pay taxes, for example.” (“Culpable Bystanders, Innocent Threats and the Ethics of Self-Defense,” Canadian Journal of Philosophy (2005), at 591-592.) Indeed. Moreover, it is not actually the case that only responsible agents can be duty-bearers. There is nothing weird about claiming that a criminally insane (and thus non-responsible) agent violates my rights when he tries to strangle me to death. I bet the police officers trying to save me would see it the same way. (And even the criminally insane have to pay taxes if they have income.)

    So far the liability of non-responsible threats or agents. As regards the liability of justified attackers, he chooses Rodin, who denies that justification defeats liability, as his target. Yet Rodin makes his point in passing, on at best two pages (but Jonathan spends almost 3 pages on him, Thomson was less fortunate), while Hosein and I have defended the view that justification does not defeat liability at length and in detail. Jonathan rests his supposed refutation of Rodin on the latter’s responsibility account and his peculiar version of reciprocity. I share neither, nor does Hosein, but perhaps that made us too hard targets to tackle. Again, since Jonathan ignores the relevant literature, his arguments here remain entirely unconvincing and insubstantial. (For my repeated and sustained defenses of the liability of non-responsible threats and my attacks on the justification defeats liability doctrine, see On the Ethics of War and Terrorism (2007, pp. 80-97); SNP, section 3; EWFL, sections 4.1.3-4.1.6. For Hosein, see “Are Justified Aggressors a Threat to the Rights Theory of Self-Defense?” in Frowe/Lang (eds. 2014), How We Fight; and “Responsibility and Self-Defense: Can We Have It All?” Res Publica (2017).

    So far regarding liability. Jonathan also, as mentioned, addresses other topics. His chapter on the means principle does, frankly, what all recent defenses of the DDE, the means principle, and related principles do: try to motivate the principle with pairs of examples that are methodologically entirely inadequate, fail to provide a convincing rationale (see on this also Joseph Bowen’s review of the book), and ignore devastating recent criticisms (he also fails to discuss Jason Hanna’s 2012 and Andrew P. Ross’s 2015 critiques specifically targeted at his use of the means principle). (I have complained about this wide-spread tendency among defenders of these ill-fated principles in “The Secret to the Success of the Doctrine of Double Effect (and Related Principles)” (2018, for a shortened version, see EWFL, section 4.1.7) and “Wild Goose Chase” (2019). My critique of Walen in the latter article should also at least partially apply to Jonathan, who himself notes the similarities between the two accounts.) I have also already criticized before Jonathan’s account of proportionality (see SNP, 2.2.4.2) and of necessity (see https://philpapers.org/rec/STEFAQ as well as my previous PEA soup discussion with him here: https://peasoup.princeton.edu/2019/01/featured-philosopher-jonathan-quong/). These criticisms still apply to what he says in the book, and I have nothing to add for now (but let’s see how the discussion goes). In sum, while Jonathan made quite an effort – and I do commend him for emphasizing the difference between justification and liability as well as for insisting that one needs to provide a plausible rationale for the necessity condition – he has not succeeded in providing a credible account of the morality of defensive force. Unsurprisingly, I recommend SNP instead.
    Best, Uwe

  3. Thanks to Jon for a fascinating book, and to Joe for a super review!

    I have a question for Jon, which tracks issues in the Duped Soldiers discussion above (and which I don’t *think* is discussed in the book).

    On Jon’s view of liability:
    (1) A is liable to defensive harm for posing a threat to B’s right if A has good evidence that B lacks that right, but B does in fact have that right,
    (2) A is *not* liable to defensive harm for posing a threat to B’s right if A has good evidence that transgressing B’s right would be justified as the lesser-evil (i.e. transgressing the right would bring about sufficient good to justify doing so), but in fact there is no lesser-evil justification.

    What does this view say in cases of ‘overdetermined evidence’? These are cases where: A has good evidence that B lacks the relevant right *and* good evidence that transgressing B’s right would be justified as the lesser-evil, but in fact B has the relevant right and there is no lesser-evil justification. For example, I have good evidence that if I turn the trolley towards the 1 then 5 lives will be saved, and I also have good evidence that the 1 maliciously set the trolley rolling towards the 5, but both these propositions are false.

  4. Hi Jonathan, Joe, et al.!

    Thanks for letting me join in. There’s already a lot going on in this thread! Just a cavil from me. So, Joe suggested that it is counterintuitive to think that you cannot permissibly turn the trolley if it means using a wrench belonging to Victim. And you responded by likening the wrench to an oxygen tank or to heart medication – a vital possession.

    But importantly, the wrench isn’t saving Victim’s life. That is, Victim don’t need the wrench to live. He didn’t before, and he doesn’t now. Indeed, if the wrench just disappeared or were annihilated, it would be great for Victim! (Unlike, say, an oxygen tank or heart medication). So, if there is some reason for thinking that you are not permitted to use the wrench, it can’t be that Victim needs it to live.

    Now, you might say that Victim’s life depends on prohibiting me from using the wrench. That’s true. Maybe the means principle extends to such cases. But in ascertaining whether this is so, I don’t think we can analogize such a case to ones in which the item in question is itself life-saving, as is the case with heart-medication and oxygen-tanks.

  5. First off, I wanted to say thanks to Jon for the really thorough reply. And, a big thanks to Jordan, both for organising and for cutting down my review for the blog!

    Rather than focus on all of Jon’s points, I’ll just mentioned two for now.

    On 1., this is all really helpful. So, I think you’re right to say that “Making a mistake about someone’s liability is […], other things being equal, a much bigger deal than making a mistake about whether harming someone is permissible for reasons unconnected to their liability”, namely, because narrow-proportionality (concerning harms to which people are liable) is so much less demanding than wide-proportionality (concerning harms to which people aren’t liable).

    But couldn’t this difference just be reflected in the strength of the evidence that one needs to possess before one gets a liability-based evidence-relative justification to harm someone? If I am mistaken about whether you’re liable, things are going to be really bad, because I get to discount your interests to such a degree. So, my evidence needs to be pretty robust—more so, getting things as close to equal as we can, than it needs to be when I am acting with a lesser-evil justification. But *if* my evidence is strong enough such that I do have a liability-based evidence-relative justification to harm you, it isn’t clear to me why that shouldn’t protect me from being liable, given the way you think evidence-relative permissibility works in other cases. (One thing that perhaps I wasn’t clear enough about in the review: I’m not saying I come down either way on whether evidence-relative permissibility should vitiate liability. I’m more putting pressure on why there would be such a stark asymmetry, when it comes to whether one is liable, between cases in which (i) one’s evidence-relative permissibility assumes someone is lacking a right they usually possess and (ii) one’s evidence-relative permissibility not assuming someone lacks a right they usually possess.)

    Perhaps another way to put the point is this. You say, “Suppose D has evidence that A is wrongfully threatening to break her arm. Let’s assume this wrongful threat would make it proportionate to break A’s legs in defense. Now compare this with a potential lesser evil case. What kind of evidence would D need to possess before having a lesser evil justification for breaking nonliable B’s legs? The fact that doing so would save D (or anyone else’s) arm from being broken would not be nearly enough.” I’m a bit confused by this, since the reason she doesn’t get to break non-liable B’s legs has nothing to do with evidence. It’s not that her evidence isn’t strong enough. It’s that harming B is disproportionate, where it wouldn’t be disproportionate were B to be liable. So, asking what kind of evidence would she need is something like a category mistake (or, at the least, moot).

    On 3., I should’ve been clearer on what I found counterintuitive. I *think* what I find most worrying about this case (in addition to the two disambiguations you offer: that (i) I find it counterintuitive that the soldiers aren’t liable and (ii) I find the practical implications counterintuitive) is the asymmetry between the two versions of Duped Soldiers. (This was what the parenthesis that followed in the review was meant to get across: “We can imagine those of the soldiers who have read The Morality of Defensive Force thinking to themselves, ‘Thank God there aren’t any liable parties in there: only some innocent civilians. Otherwise, we’d be liable if our evidence turns out to be faulty!’”)

    Recall my worry we discussed above (point 1.): “It is crucial on [Quong’s] account whether the evidence-relative permissibility of someone’s act depends on others lacking rights that people ordinarily possess—if so, one can be liable; if not, one cannot be liable. But why should that matter?” On the reading of my worry I am offering here, the pair of cases are meant to highlight why we might think this difference oughtn’t matter.

    I’d be interested to see whether others think that the different verdicts between the two versions of Duped Soldiers is counterintuitive.

  6. Thanks Joe for a great kickstart, and Jon for such a rich book. I wanted to note a path apparently not taken that would strongly separate the Conscientious Driver case from the Duped Soldiers II case that Joe raises, and then invite Jon to comment on why it is dispreferred.

    The way Jon secures the verdict that Drives is not liable is to say that the permissibility-on-her-evidence of driving does not depend on the assumption that the Pedestrian (or anyone, for that matter) lacks a right they in fact have. But the way we get there is by noting that the /practice/ of driving is an objectively justified risk-imposition: that “the expected benefits weighed against the expected costs of the practice of prudent driving make it that any action-token of the practice is permissible.”

    If there are facts about whether a token risk-imposition is permissible (that do not simply reduce to whether it would be permissible to directly bring about the actual token outcome), then rather than appealing to the evidence-relative permissibility, one could say that what makes Driver non-liable is that her action was /in fact/ ex ante permissible. The same cannot be said for the Duped Soldiers (who are duped into believing they are lesser-evil justified in targeting the civilians): while their evidence supports believing that their action is ex-ante permissible, in fact it isn’t.

    Going this route of course incurs the theoretical burden of explaining what makes risks permissible to impose, and distinguishing the uncertainty arising from chancy world-processes from the uncertainty arising from imperfect evidence, so perhaps that is why you prefer appealing to the agent’s actual evidence. But it seems to me to be consistent with the main motivations for the Moral Status Account, and to offer a more satisfying answer to the Duped Soldiers II case, so I’d be interested to hear why it isn’t an ultimately appealing move. (And my apologies if you cover this in the book! I read it some time ago, and didn’t have a chance to refresh before this discussion).

  7. Hi Jonathan!

    Thanks for the very good question. I don’t directly address exactly the example you have in mind in the book, though I do say something about an example in the same neighborhood on p. 93. Here’s what I say about a related case on that page: “In considering whether a defensive agent’s justification depends on the assumption that a target is liable, we must hold constant the other facts of the case, and only eliminate the evidence that supports the assumption of liability”.

    So, in your overdetermination case, I think the evidence-relative permissibility of A’s act doesn’t depend on B’s liability, and thus the moral status account would deem A nonliable. I’m not sure if you find that counterintuitive. My own tentative view is that cases like this are sufficiently complex and unfamiliar that I’m inclined to be skeptical about intuitive reactions to them.

    Here’s a further wrinkle. Consider a case where—given A’s evidence—B is liable to 100 units of harm, but where (again, given A’s evidence) there’s also an independent lesser-evil justification that would justify imposing 90 units of harm on B. Suppose A goes ahead and threatens to impose the 100 units on B but, as a matter of fact, B is not liable nor does A have a lesser evil justification. In this version of the case, the moral status account entails that A is liable to defensive force, but only for the additional 10 units. That strikes me as a plausible result, but again, cases involving both overdetermination and multiple forms of mistakes are probably so complex that intuitive judgments probably aren’t worth too much.

  8. Hi Saba!

    Thanks very much for this. Although I agree that there are some disanalogies between the wrench in Joe’s case, and heart medication or an oxygen tank, I’m not persuaded that they are morally relevant differences in a way that might threaten the formulation of the means principle. You point out that Victim doesn’t need the wrench to survive in two senses: he didn’t need it before (I take it you mean before the whole trolley threat was initiated), and he doesn’t need it now (in the sense that if the wrench disappeared, then you couldn’t turn the trolley toward him).

    But neither of those senses seem morally significant to me. I’m most confident about the first sense being not morally significant. Many objects we rightfully acquire might not be needed or important to us when we rightfully acquire them (obviously including medicine and oxygen tanks) but come to be crucial at a later date.

    The second sense is more interesting. It’s true that if the wrench were to suddenly disappear, Victim would be safe. But that doesn’t seem to me to be a fact that could weaken Victim’s claim over the wrench. I think what matters—as your comment anticipates—is that Victim’s survival depends on his having normative control over the wrench.

    Also note we could apply the same reasoning to some of our body parts. Imagine that the only way you can redirect the trolley toward Victim is by physically grabbing Victim’s arm and using it to flip a switch (the switch is coded to his palm-print). In this case it’s also true that Victim doesn’t need his arm to survive, and he would in fact be safe if his arm were to suddenly disappear. But everyone sympathetic to the means principle is going to agree that you violate the means principle if you manipulate his arm like this.

  9. Thanks to Joe and Jon for the fascinating exchange. There’s a lot I could dig in on–I cover many of the positions Jon takes in my book: The Mechanics of Claims and Permissible Killing in War, but I don’t want to take this as an occasion to rehearse those arguments. I will limit myself to two smallish points.

    (1) I do think it’s important to distinguish different ways of using the property of another. I think JJ Thomson was right when she devised the case, where you use the sidetrack man’s wrench to turn the trolley, to push on the idea that such turning seems permissible, despite the fact that it’s HIS wrench that get’s, in a sense, turned against him. But Jon is right to suggest that sometimes we can’t use the “second property” of another to do something good, if harm would come to the owner–I’ve discussed taking drugs from a drug owner and Jon discusses an oxygen tank. So, I think we need to distinguish ways a person has a claim on his property. I’ve struggled with trying to make that distinction for years, and Matthew Liao and Christian Barry recently took me to task for a failed attempt in their piece on Øverland’s and my versions of patient-centered deontology (Law and Phil, 2020). But I have a forthcoming (in Law and Phil) reply where I try again. I am happy to share the draft if anyone wants to see it.

    (2) I want to second Joe’s response to Jon on liability NOT arising when you make a mistake about whether a lesser-evil justification applies. Partly I resist the distinction between Jon’s two categories of cases because I reject the idea of a lesser evil justification full stop. But I can make sense of the basic idea: a mistake about whether negative restricting claims are outweighed by positive restricting claims
    (lesser evil like), vs a mistake about whether someone has chosen or acted in a way that threatens others without having the right to do so (including the intermediate category of merely Being a threat without the right to do so). Anyway, it seems to me that if one is a threat to others, without having the right to be one, one should be liable. Period. BUT, we should also recognize that liability comes in degrees. The conscientious driver may rightfully be killed (as liable) to save the pedestrian, but proportionality limits apply. If there were TWO drivers who had equal responsibility for the drive, they might have a right not to be killed to save the pedestrian. Contrast a culpable threat, where arguably there is no limit on the number of them who can be killed to save an innocent victim… though level of culpability surely matters for the proportionality analysis: mere negligence might not forfeit as much as intentionally doing what the agent should recognize wrongs the victim. The point is: this implication strikes me as one of the hardest parts of Jon’s theory to swallow, though I do appreciate the attempt he offers to clarify why he holds it.

  10. Hello all,

    Great discussion so far.

    Just a quick comment from me. I’m a bit confused about the discussion of property between Joe and Jon, that Alec chips in on too. I think that ‘X owns T’ just is something like ‘X has a certain bundle of rights, powers, etc with respect to T’. So I think that X’s owning T can’t provide an explanation of the ownership constituting rights, powers, etc that X has with respect to T.

    So when we ask: ‘is it permissible for Y to use X’s property to save 5 killing X’ an argument for the view that Y is, or is not, permitted, cannot normally rest on the fact that the object is X’s property. To say that Y is not permitted because T is X’s property is just to say that one of X’s property constituting rights is that X is not permitted to use T in these circumstances. But X does not have this right because the object is their property; their having it (if they have it) is just one of the things that constitutes X’s ownership.

    What we need is an argument about the scope of the rights over objects that we get through standard modes of acquisition (purchase, gift, etc) under fair or unfair conditions. I’m tempted by the view that acquisition under standard conditions has limited importance in very extreme circumstances where the object is not purchased for, for example, security. So if, in a war, I claim priority to use my car for shelter against falling bombs, the fact that I acquired this particular car, where yours is out of reach, has no significance. What I acquired was a set or rights to use the car to get to work, to sell it on, etc. But did I acquire the right for priority of protection in these very unusual and extreme circumstances? The claim that it is my property isn’t an argument in favour of my having priority – that claim just asserts the conclusion that needs arguing for – that the property constituting rights I gain through acquisition include a right of priority in extreme circumstances like this. An argument that I do should explain why, for example, acquisition of a car under fair or unfair conditions for ordinary purposes like getting to work etc also gives rise to priority of use in these much more extreme conditions. I can’t see what the argument would be though.

    That is why claims about use of bodies and claims about property come apart – the claim that you can’t use my body in a way that harms me isn’t a claim that draws normative conclusions from normative material – its being my body isn’t a normative relationship in the way that something being my pen, for example, is. The claim that another may not use my body to save five in a way that kills me because it is my body may be false, but it is not circular – its being my body is not constituted by the right that I claim to have in virtue of its being mine.

  11. Thanks, Joe, for this very helpful further response, and for clarifying where you intend to put pressure.

    On the first point: I certainly see the appeal of trying to reflect the difference between liability judgments and lesser evil judgments by requiring different evidentiary thresholds to be met before it is ER permissible for an agent to impose harm. But even if A meets the higher threshold required to act on a judgment about B’s liability, I’m not persuaded that this immunizes A from liability should the judgment turn out to be mistaken. And my reasons are the same ones I’ve already expressed. First, it has too many implications that I find counterintuitive (in cases like Resident and Duped Soldiers where we can stipulate the agents meet very high thresholds). The other reason I’m not persuaded is that even when A can say “I made sure the evidence of B’s liability was overwhelming” it remains true that when A acts on this judgment, she treats B as entitled to less moral consideration than that to which people are standardly entitled in order to advance her own interests or the interests of others. Imagine A’s threat to B is in progress and A can’t stop it from occurring, though A now realizes her error. Suppose, though, that C is a third party who can intervene one way or the other. It would be, I think, implausible for A to tell C “Look, I realize B now faces serious harm because I treated him as entitled to less than the equal level of moral consideration that all persons are owed, but it would be wrong for you to redirect the harm away from B and back toward me. Because I made sure the evidence of B’s liability was overwhelming, B and I are morally on a par—neither of us is liable to suffer any of this harm”. This doesn’t sound right to me because they aren’t symmetrically situated: A treated B with less than equal consideration but B did no such thing to A.

    Note also that, although I endorse a particular view about narrow proportionality that ends up being (you might think) harsh with regard to how much harm people who make such reasonable mistakes are liable to suffer, you could jettison my view of proportionality while holding onto the moral status account to arrive at less harsh judgments. On this alternative, you might say that Resident is liable, but his liability is significantly mitigated by his epistemic excuse. To be clear, this isn’t my view, but it’s consistent with the idea there has to be an asymmetry between Resident and the innocent twin at the door, since only Resident treats the twin as lacking certain moral rights he in fact possesses.

    On the second point: thanks for clearing this up! I totally understand why people might find it strange that the soldiers could be liable in my version of the case but not in your (mistaken lesser evil) version. Unfortunately, I don’t have too much to say beyond what I said above.

    But here’s one way of framing the issue. If you think the two versions of Duped Soldiers must be equivalent in terms of liability, then either they are liable in both versions or liable in neither version. Probably the best way of explaining why they are liable in both versions is the moral responsibility account, or something close to it. But as you know, I have various reasons for resisting that account, including the fact that it entails Driver in Conscientious Driver is liable (and lots of other people who foreseeably pose only tiny risks of harm will end up being liable too). So I think the costs of going that way are substantial. Alternatively, we could say the soldiers are not liable in either version. But this would require endorsing something like what I call the evidence-relative account of liability. But that account has so many deeply counterintuitive implications (including in Resident and Duped Soldiers) plus I think it lacks a compelling deeper rationale partly for reasons I expressed in my original reply. In sum, even you think it is counterintuitive to draw a distinction between my version of Duped Soldiers and yours, I think the alternatives have even more counterintuitive implications.

    But thanks again for pressing further on this and helping me see the full force of your objection.

  12. Hello everyone…wonderful discussion. A quick thought on Victor’s comment, followed by a couple of points on rescue killings.

    1. It seems to me that, notwithstanding the fact that my body is not mine in the way my pen is, I can no more answer ‘is it permissible to use my body to save X killing 5’ by responding ‘no, because it’s my body’, than I can answer the same question in relation to my pen. The answer begs the question in both cases, for the question is precisely whether have the relevant right. (Incidentally: I am with Jon in doubting that there are categorical differences between the rights we have over our body and the rights we have over our property. But that point is orthogonal to my comment on Victor’s. ) (Alec, could you send me your forthcoming LP paper please? Thanks.)

    2. Turning to rescue killings. Joe’s Duped Soldiers II example is excellent. Jon says that on his account, (a) third parties are not morally permitted to kill DPII in other-defence anymore, just as (b) third parties are not morally permitted to kill CD in defence of the pedestrian. Jon also surmises that some people will balk at (a). I agree. One reason why Renee’s argument appeals to me is that it would allow, plausibly, for third- party, other-defensive killing in this case – while holding on the claim that third parties are not morally permitted to kill CD in defense of Pedestrian.

    But some might also balk at (b). Suppose that the police are on the scene. It’ s not clear to me that they may not kill CD (granting for the sake of argument that CD is not liable.) Put more generally: perhaps the police are authorised agents who can exercise an agent-relative prerogative of self defence on the pedestrian’s behalf – see Morality of Defensive Force at c. 78.)

  13. Hi Cecile

    Here’s another way of putting the point.

    Suppose that X owns T is some set of rights {r1…rn}, and priority of protection is r-priority. Suppose that r-priority is in {r1-rn}. Then X’s ownership of T is not the ground of X’s right of priority of protection. X has such a right, but not because X owns T. X’s right is just part of the set of rights that is X’s ownership. So X’s ownership cannot figure in an argument for X’s priority of protection. (Compare the claim that an object is in my house because it is a member of the set of things in my house!).

    In contrast, suppose that this body is my body. This is true not because of some rights that I have that others don’t use my body. And this fact – that it is my body – might make it true that others cannot use this body in a way that harms me. Of course, you are right, and I implied as much, that we still need an argument why the fact that a body or body part is mine in this non normative sense gives rise to some set of rights. But we might still have identified the grounds of my rights. The claim that this is the ground of my rights is contestable, but there is nothing problematically circular here.

  14. Hi Renee!

    Thanks for this excellent question. Your proposal, if I’ve understood it correctly, is that we might hold that Driver is ex ante permitted (rather than ER permitted) to engage in the token act of conscientious driving, but the Duped Soliders (in either version of that case) are not ex ante permitted to act, and this might nicely explains why Driver is nonliable but the soldiers are liable.

    But I need to have a better grip on what ex ante permissibility is and how it differs from evidence-relative permissibility. Suppose ex ante permissibility means something like “the predicted benefits of permitting this practice are sufficiently great and the risks to any given individual are sufficiently low that token instances of the risky behavior are permissible”. This would explain why what Driver does is ex ante permissible. But it also looks like it might explain why what the Duped Soldiers do is permissible (or that what Resident does is permissible). The benefits of permitting the practice of people defending themselves or others when they justifiably believe they have a lesser evil or liability justification might be great enough compared to the risks such that token instances are permissible. So we wouldn’t get the asymmetry between the cases that I find intuitive.

    Of course maybe there’s a way of specifying a particular theory of permissible risk imposition where the Driver’s risk imposition is permissible but the soldiers’ imposition in Duped Soldiers isn’t OK, but it won’t be possible to design a plausible theory of risk-imposition that precludes the bare possibility that sometimes people who treat others as liable to attack are engaged in ex ante permissible risk imposition.

    But I’m worried I may not have fully grasped the proposal, so please follow-up if I’m not getting it right.

  15. Hi Jon,
    you haven’t answered to any of my objections to your position. Are you still thinking or already ignoring me (again)? That would a kind of prove my point, though, wouldn’t it?

  16. Hello everyone,

    Congrats on the book, Jon! I realised whilst reading it that I’m at my philosophical happiest whilst engaging with your work. I’ve already had lots of fun conversations discussing it with others. I have a ton of questions but shall (for now) restrain myself to what might be a slight daft question for everyone (not just Jon). First, though, I would like to note my hurt that Jon asked Renee to do the artwork for the cover, even though I know for a fact that he is familiar with my own expertise in this area. Perhaps just something to bear in mind for the next book.

    Jon argues that since Resident in Mistaken Attacker is almost certain to cause Twin serious harm, “the justification for his act must assume the liability or likely liability of the person whom Resident harms.”(48) But what evidence justifies this assumption of liability? Jon says that “Resident has no reason to believe that he faces a non-responsible threat.” (93) But this is hardly evidence that he faces a responsible threat. Resident’s evidence is simply that the person before him is a murderer who has escaped from prison. He might, as Jon suggests, assume, on this basis, that Twin is liable. Or, he might believe that anyone who commits such a terrible crime must suffer from a psychiatric disorder, and assume that Twin is a non-responsible, non-liable threatener. Neither assumption about liability is obviously contradicted or supported by the available evidence, which supports only the assumption that Twin is very dangerous. Even if Twin is indistinguishable from a liable attacker, it’s not clear that this is evidence that Twin is a liable attacker. When two things are indistinguishable, this means that we lack evidence that enables us to tell them apart.

    Jon says that if we remove Resident’s evidence of Twin’s liability, whilst holding the other facts constant, we thereby remove his evidence-relative justification for force. But since there is no specific evidence of liability, it’s hard to know what this means. We could keep all the facts constant in Mistaken Attacker without undermining the assumption that Twin is a non-responsible threatener.

    It’s not as if only Jon assumes that Resident is ER justified in believing that Twin is liable, of course – it’s widely assumed in discussions of the case. Jeff stipulates something similar in his original discussion. So I guess my question is, what do we take to be evidence of *liability* (or, I guess, moral responsibility), specifically? Is the thought just that we assume by default that people are morally responsible, unless we have evidence to the contrary (as Jon suggests)? And if so, is that properly described as an evidence-relative justification?

  17. Hi Alec!

    Thanks very much for these points. On (1): it sounds like we are in partial agreement—you agree that there are cases where an agent uses Victim’s property to redirect or shift harm away from others and toward Victim, and this would constitute a violation of the means principle (or a principle that’s in that neighborhood). It seems that there are then at least two points of disagreement. First, we disagree about what’s permissible in the case involving the wrench. Second, and more generally, you want to draw a distinction between different types of claims people can have over their property in a way—as I understand it—that entails that only some harmful uses of other people’s stuff end up being violations of the means principle. My own view is that any part of the external world to which we have a rightful prior claim can be something to which the means principle applies, since there are all kinds of ways in which control over such bits of the external world can suddenly become essential to protecting ourselves from harm. But it sounds like I need to see your new forthcoming paper in Law and Philosophy, so please do send it my way.

    On (2): here we clearly disagree on multiple dimensions, partly because you reject the idea of lesser evil justifications altogether. But you also say “if one is a threat to others, without having the right to be one, one should be liable.” We disagree about this, since on this view the Duped Soldiers in Joe’s case are posing a threat to the rights of others without a right to do so, and thus they must be liable. I don’t endorse this idea partly because I suspect the notion of [having a right a to pose a threat to others] is less explanatorily fundamental than the question of whether the agent has done something to forfeit her rights or render herself liable. For example, in thinking about Obstructors who block someone’s path to safety, I think in order to know whether they have a right to pose this threat, we need to figure out whether they’ve done something that would render them liable.

    But the preceding is a bit tangential to your main point. I take the point that you and Joe are pressing, namely, that a number of people find the distinction I draw between liability mistakes and lesser-evil mistakes to be pretty counterintuitive.

  18. Hi Victor!

    Thanks for this—I know this issue about the scope of rights over the external world is an issue you have pressed on me before, but clearly I haven’t learned anything!

    You say “I’m tempted by the view that acquisition under standard conditions has limited importance in very extreme circumstances where the object is not purchased for, for example, security. So if, in a war, I claim priority to use my car for shelter against falling bombs, the fact that I acquired this particular car, where yours is out of reach, has no significance.” Obviously there’s disagreement here. I don’t think the scope of a person’s rights over an object should typically depend upon the intended purpose for which she justly acquired the object. This would be an unnecessarily constraining picture of moral rights and distributive justice. I did not buy my car with any thought of using it to shield myself from a hungry lion that has escaped from the zoo. But it seems weird to me to say, if that situation were to arise and multiple different people needed the car, the car is now manna from heaven to be distributed anew with no one having a prior claim. I think there are at least three deeper reasons this seems weird to me. First, I’m skeptical that there’s any principled line to be drawn between extreme and standard circumstances. How would such a line be drawn? We might say, if some dangerous event is sufficiently unlikely to occur then we’re in extreme circumstances. But I think that’s going to be a very awkward line to draw. Some events are extremely unlikely to occur in a given area—earthquakes or giant tsunamis—and yet people will understandably think their rights over external objects persist when the unlikely risk eventuates (e.g. it’s still my car that I have distinct claims to use to escape the oncoming tsunami). Second, I also think it’s going to be extremely hard to draw a principled line around the purposes for which an object has been purchased. One problem with drawing such a line is whether it’s going to treat people differently depending on their subjective reasons for purchase. If A sincerely thinks “I bought this object to be used for an indefinitely wide range of purposes including any life-saving uses it might have in extreme emergencies does that give A wider scope rights over the object than B’s rights over a similar object where B thinks “I bought this purely to use at the barbecue next Saturday”? Third, I favor a certain kind of liberalism about issues of distributive justice. Part of this liberal outlook involves not setting more limits than needed on people’s freedom to use their bodies and their stuff as they see fit, including making brand new decisions about how to use their bodies and their other stuff. I think the picture you’re describing is, at least in part, in tension with this outlook.

    But I want to emphasize some possible level of agreement. I agree that under non-ideal conditions (which is our regular condition), most people haven’t been allocated their fair share of worldly resources, and so there are huge and largely unanswered questions about how to think about the apparent de facto claims people have over various parts of the external world. I also agree that in emergency conditions when some object currently has no prior claimant, it can be a new question of distributive fairness how rights over that object ought to be allocated.

    One last point. I don’t agree with your claim that “its being my body isn’t a normative relationship in the way that something being my pen, for example, is”. I agree with what Cécile says on this—in the case of both the body and the pen we must decide whether a person’s prior claims over the object include control rights over the object in the particular circumstances at issue.

  19. Hi Cécile!

    Thanks very much for these points. On the first point, I think we are in complete agreement.

    Your comments on the second point are great both because they provide more evidence for how counterintuitive at least quite a few people find my view on this point, but also for the interesting suggestion that police officers might be agents authorized to exercise the pedestrian’s agent-relative prerogative. I don’t have a firm view about when, if ever, state officials could be authorized in this way. But one reason to be skeptical that police officers could have this kind of authorization is that in cases like Conscientious Driver such authorization would go both ways, at least if I am right about Driver’s nonliability. The officer would thus apparently be free to permissibly use force against either party on behalf of the other. Not only does that seem counterintuitive, but it also seems like we have more general and independent reasons not to want police officers to have that kind of discretion in deciding against which nonliable person to use serious force.

    One other point: you say Renée’s proposal appeals because it seems to provide a way of avoiding the implication that the soldiers are nonliable in Joe’s variant on this case. But, as I indicate in my reply to Renée, I’m not clear that her proposal actually does yield this result, or at least I need to have a clearer grip on the proposal to see how it does.

  20. Jonathan, I don’t know how you’re keeping up with all these missives. If you’re running out of gas and need to forego a response this, I totally understand. This’ll be my last reply in any case.

    So, I claimed that whether Victim needs the wrench to survive is morally relevant. You suggested that it isn’t. How do we resolve this? The point of my original post was this: the fact that it would be wrong to use Victim’s oxygen tank or heart medication doesn’t show much. After all, in those cases, the Victim does indeed need those items to survive. To show that such a factor is irrelevant (as you suggest) we need a case where it is clear that using Victim’s possession is wrong, despite that Victim doesn’t need the possession to survive. You helpfully suggested the body-part example in your response to me. But that introduces a confounding variable: there is at least one sense in which I am body. By using my body, you are, in one sense, using me. That isn’t the case when you’re using my wrench.

    Honestly, I’m not sure whether needing-to-survive is morally relevant. The fact that I’m pulled in one direction by the wrench-example and in another direction by the oxygen tank / heart medication example, suggests to me that it might be, since that’s a difference between the two class of cases. But my intuitions are pretty weak, in any case.

  21. I love this discussion! Let me just say, to Victor, you assume that property rights, as a category, is just some sort of label we throw on certain types of rights/claims. But I think the order or explanation goes the other way. We need to distinguish certain kinds of claims as PROPERTY claims because they are essential to agency; other kinds of patient claims are not essential to agency in the same way; they are really only about welfare. So there’s a fundamental reason for making the property claim vs restricting claim distinction, and then you can ask about just what has to be part of the cluster of rights that constitute property rights (e.g., which ones can be alienated and it still remain your property; which one’s are inalienable). The physical facts about your body are, of course, normatively significant, but they don’t make your body the ONLY thing that counts as property in this fundamental sense.

    To Jon, we agree on a lot, and disagree on a lot. We agree about the importance of the means principle and the significance of property, but I think we don’t have the same interest in all our property. I’ll send you the paper. As for the priority of liability/forfeiture and having a right to act, I think we disagree in part b/c I don’t accept your notion of AC prerogatives. I think they go too far, and that the limit in terms of the MP isn’t sufficient. But really, to work out who is right, we’d need to line up our theories head to toe, as the differences are pretty systematic. We’re both trying to create accounts that systematically handle a large range of cases, and we take a few different turns along the way. The tradeoffs/choices have substantial implications, but without having all the differences in view, it’s hard to say who is on better footing.

  22. Hi Jon,

    I’m sure it’s me that isn’t learning!

    I wasn’t intending the view that the rights that we gain through acquisition typically depend on the intentions of the person acquiring the property. Rather, I was intending the view that there is a limit on the principle that property becomes equally available for protection in emergency situations where the object is purchased with the intention to improve security in such situations (it would take more spelling out than I have space for here, but the idea is that, within limits, we might want to make a scheme of property rights ambition sensitive for emergencies by making it possible for people under fair conditions to purchase extra security, or to spend money on other things).

    More generally, though, when we acquire property, we acquire more or less expansive bundles of rights. Land rights are limited by rights of neighbours, rights of those who need access to the land, and so on. Governments can requisition property during war for purposes of security. Important works of art must not be destroyed, and some must be publicly displayed. There are limits on whom we can sell property to, for example in the case of monopolies, and so on. So the question is: why should my having purchased a car come with a right of priority of security in emergency situations? Or should those rights be more limited?

    You suggest a liberal view favours people having more rights over their stuff. This is the right kind of argument to look for I think: to find a value in people having the more expansive rights you think they have – it’s not because the object is theirs, but because liberty is best protected by a people having the opportunity to purchase a more extensive set of rights without intending to acquire those rights. I agree that liberty, certainty, the ability to plan, efficiency, and so on, are often advanced by giving people the ability to purchase wide ranging rights over objects for familiar reasons. But I don’t see any general liberal reason to favour the right to gain priority of protection in grave emergencies, where the object has not been purchased for protection. Giving one person more liberty or power over an object just limits the liberty and power that others have over it. If the car owner has the liberty to exclude me when the bombs are falling, I lack the liberty to exclude them, even if I need the car more. And I lose an equal chance of survival. Furthermore, such a right does nothing to advance people’s freedom to realise their ambition to be more secure where the object is not purchased to foster security. So how does the liberal view support this particular right? I think that a similar thing explains why it seems (to me, Joe and perhaps others) so counterintuitive that it is wrong to use a person’s wrench to turn the trolley from the five to the wrench owner – the values that underpin a right to determine who gets to use the wrench through purchase, that provide the foundation for property rights, doesn’t seem to support a restriction on using the wrench in these circumstances.

    The broader methodological point, though, is that in order to argue for or against the permissibility of using the wrench, we need to understand the values that underpin the property rights that we have and their limits, rather than referring to ownership itself. The relevant reason against using the wrench is not that the one person owns it, but that liberty, or some other value, would be set back if the rights of the person acquiring the wrench were limited, permitting it to be used by others in emergencies against the interests of the owner and without consent. I don’t yet see that this is so, but it’s the right place to look.

  23. Hi Helen!

    Thanks so much—I’m delighted to be promoting philosophical happiness. My next book is going to be about legitimate injustice, so send in your stick figure design ideas now…

    I think your question is excellent, and I don’t have much to say about it in the book. I think in most contexts, when A has evidence that some adult person B is threatening her, she also has evidence that B is liable. This is true for the simple reason that the vast majority of adults meet the threshold for morally responsible agency. To be below the threshold for minimally responsible agency is pretty unusual, and so you should not believe that a random person falls below this threshold without evidence that they do. I can see that this might feel odd—we don’t have direct evidence that B is morally responsible, A only has: (a) evidence B is threatening her via some act, and (b) the statistical evidence that the vast majority of adults are above the relevant threshold. But, absent further details, I don’t think it’s objectionable to treat the combination of those facts as evidence of B’s liability. Of course there are all kinds of interesting complications that arise when we start to think about the relevant reference group for (b). If A is currently visiting an institution reserved for people with severe mental illnesses and B appears to be a patient, A has to adjust—it now seems like she’s got some decent evidence that B may not be morally responsible.

    I concede that things can get really tricky here in determining what counts as evidence of moral responsibility, but I don’t see any way around it, and of course similar problems emerge for the moral responsibility account, and the culpability account.

  24. Hi Saba,

    Thanks very much for this follow-up. I take your point that it would be helpful to have a clear example where using Victim’s property is wrong despite the fact they don’t need the possession to avoid harm (that is, assuming you don’t use the possession). I don’t think the body part case really does introduce a confounding variable, but let’s set that aside.

    Of course part of the difficulty here is that you and I are likely to diverge on whether any given example *is* a clear example—any attempt to provide you with an example might just elicit differing intuitions between us. But let me try with a variation.

    A runaway trolley is headed down the main track where it will kill five innocent people. There is a sidetrack where one innocent person is trapped. The only way Bystander can redirect the trolley and save the five is by getting the one person’s car keys out of his jacket pocket, running over to the one person’s nearby car, and then crashing the car into the trolley at just the right moment, thereby turning the trolley away from the five and toward the one.

    I think this is a clearly wrongful use. But of course it’s just like the wrench case, but with a much bigger, more valuable object, and a bit more work on the part of the Bystander. A hypothesis: wrenches are small objects of typically very little value, and this might elicit the reaction that of course wrenches are no big deal, so it can’t be a genuine violation of the means principle. But cars are big valuable things, and so people might find it easier to accept you can’t just use another person’s car to redirect a threat toward that person. Of course I don’t think the typical value of the object matters, but my guess is that more people are going to agree that using the car is clearly wrongful. What do you think?

  25. Alec likes the “discussion.” Perhaps because “revisionists” like to inhabit an intellectual echo chamber in which they discuss minor points among themselves and incessantly congratulate each other on their achievements, which are largely imagined since fundamental criticism coming from the outside has been obstinately ignored and therefore stands unrefuted. It’s a cult.

  26. Hi Jon,

    Like Saba says, no need to reply further at this point!

    I agree – this is a problem for most theories, not just yours. But I wonder if it matters more for you in at least one way. Whereas we can similarly challenge e.g. Jeff’s assumption that Resident has a justified belief that Twin is liable (in his original discussion of the case), this isn’t an objection to Jeff’s broader argument. He believes that it’s impermissible to kill non-responsible threateners to save one’s own life. Thus, he can simply grant that if Resident is not evidence-relative justified in believing that Twin is liable, then he is not evidence-relative justified in using force. On his view, Resident’s evidence-relative justification straightforwardly depends on a justified belief that Twin is liable.

    But on your view, our agent-relative prerogatives permit us to defensively kill non-responsible threateners. So the lack of a justified belief that Twin is liable doesn’t preclude an evidence-relative justification for using force. And it doesn’t seem implausible to me that Resident might treat her evidence as I suggest above: thinking that anyone who has done something so awful must not be in their right mind (perhaps combined with some evidence of high levels of mental illness in prison populations). And so why couldn’t she act on the basis of an agent-relative prerogative in Mistaken Attacker (and hence, contra what you say, avoid incurring liability) when she threatens Twin?

  27. Hi Alec,

    Obviously to properly engage with some of our disagreements, I’ll need to read the new paper you have coming out. But one small thing here. You say “we don’t have the same interest in all our property”. I agree, but I also think, as some of my earlier replies indicate, that we can’t say anything general and systematic about different kinds of property and the levels, or degrees, of interest people have in claim rights with regard to types of property. Typically, our interest in control rights over wrenches is limited, but that can change. The same is true for pretty much any object, I think. As I say in the book, I think the stringency of a right is partly determined by the severity of the harm that would befall the right-holder if the right is transgressed, and that’s something that’s going to depend on the context.

    But as you say, part of what’s going on here are decisions about the competing costs and benefits of different approaches when looking at the entire view.

  28. Hi Victor,

    Thanks for the further response on this. The issue you’re raising is one of the deepest and most important questions.

    What you say in your first paragraph is helpful—that gives me a clearer picture of what you’ve got in mind, and obviously it’s more plausible than how I’d initially construed it.

    On the methodological issue, we’re in agreement. I agree that we need a good answer to the question of scope: should our moral rights over bits of the external world (like wrenches, cars etc..) extend to emergency situations or not? So I agree we can’t just unproblematically move from “A has a robust set of liberal ownership rights over object O under such-and-such conditions” to the conclusion that in all contexts A has those rights. For example, as I argue in the book, I think people have moral duties to rescue others from certain urgent harms when they can do so at modest or reasonable cost, and so when those rescue conditions are triggered, the duty-bearer temporarily lacks the kind of claim rights over her body or other resources that she would otherwise possess.

    So, the issue is whether we have good reasons to hold that there’s a distinct and useful category, something like “emergency situations” [where this is distinct from duty to rescue cases], that delineates a sphere where people no longer possess the claim rights over the external world that they would otherwise possess. I’m skeptical about this for reasons which I partly gestured at in my previous reply. First, I’m just not sure there is a compelling and principled way to distinguish emergency situations from non-emergency situations. Our ordinary, everyday lives are filled with sudden emergencies. A family member is suddenly very ill and needs to be rushed to the hospital. There’s an accident on the road and multiple people are hurt. There’s a fire and several people are in danger. If our normal liberal ownership rights don’t extend to these ordinary cases, that would be very surprising. Part of the reason why it’s important to have control rights over a range of the external world (and why it’s important those control rights be distributed fairly) is precisely because ordinary life comes with lots of risks and a range of different emergencies, and we each have a powerful interest in having a fair share of control over the world to help us plan for, and respond to, such problems.

    We’re also in agreement that at least part of how we answer the question about the scope of rights is by looking for deeper values or reasons within political philosophy, such as individual liberty. But you say “I don’t see any general liberal reason to favour the right to gain priority of protection in grave emergencies, where the object has not been purchased for protection. Giving one person more liberty or power over an object just limits the liberty and power that others have over it.” I agree that freedom is zero-sum in the way you describe—the more control rights over the world A possesses, the less of the world that is available for B to control. But what’s at issue here isn’t whether we should allocate moral rights over the external world equally or unequally, rather what’s at issue is whether we should think about the distribution in a more holistic way, or in a more case-by-case way. Answering this question requires, in part, working up a complete theory of distributive justice, and obviously that’s not something I’ve got in my pocket. But I’ll make two quick points here which I also briefly make in the first chapter of the book. If what we’re worried about is overall distributive fairness—making sure each person gets his or her fair share of external freedom—this almost certainly can’t be achieved by engaging in repeated local redistributions. What I mean by “local” is that we focus only on the specific context, and try to fairly distribute only the harm or risk of harm amongst everyone currently subject to some particular threat. That won’t achieve overall fairness since the focus is too narrow—equally distributing resources to help Albert and Betty stand an equal chance of evading *this* particular threat ignores whatever wider inequalities might exist between Albert and Betty—indeed it might exacerbate those wider inequalities. Second, in many of the emergency cases under consideration, no fair distribution is possible. There’s nothing fair about redirecting a trolley away from five and toward one—the one person bears all the costs and the five reap all the benefits. Partly for these reasons, I think it makes sense to approach questions about rights over external resources a bit differently. I suspect we should be thinking: what’s a fair allocation of rights over the external world that gives each person maximal control over some part of the world (compatible with a similar freedom for all) to shape their lives and insure or protect themselves from various emergencies that might arise? I think this is more likely to respect the values of freedom and equality than an approach that tries to redistribute resources in the local way I was sketching.

    But I concede what I say above is just a sketch of a view, one that requires way more working out than I’ve provided in the book or anywhere else.

  29. Hi Helen,

    You’re right: the issue that you’re pointing to creates some complexities for my view that it doesn’t create for a view like Jeff’s.

    One point I make in the book is this: in many cases (like Resident) a defensive agent’s evidence that they face a liable attacker is also the key evidence that they face any kind of threat at all. In Resident, for example, without the evidence provided by the local authorities that there’s an escaped murderer on the loose, Resident would have no reason to believe the person at the door is liable, and also no reason to believe they pose any kind of threat.

    But I also think there is also an important *advantage* for the moral status account as compared to the moral responsibility account when these two kinds of evidence come apart. On the moral responsibility account, everything is going to hinge on whether an apparent threat is morally responsible. If the threat doesn’t meet that bar, or there’s some strong but not-totally-decisive evidence the threat doesn’t meet that bar, then the defensive agent ought to conclude that they cannot permissibly defend themselves. Because the content of other people’s mental states are largely opaque, there are thus likely going to be a range of cases where defensive agents are virtually certain that they face a serious threat, but significantly uncertain as to whether they are permitted to defend themselves. On the moral status account, by comparison, so long as you have sufficiently compelling evidence that you face a real threat, you can be confident that you act permissibly in using at least some degree of defensive force, even when you are very uncertain about whether the threat meets the threshold for moral responsibility. I think defensive agents are, in general, more likely to have clear evidence about whether they face a threat as compared to clear evidence of whether the threat meets the threshold of moral responsibility. So the moral status account has the benefit of giving defensive agents more confidence about the permissibility of what they can do in a range of situations where they are plausibly likely to find themselves. Of course this doesn’t preclude the problem of bad moral luck—they might be mistaken about whether they face a threat at all—but that’s a problem shared by both views.

  30. Hm. My worry is more about what this means for liability on your view, not permissibility. Resident is liable if she treats Twin as lacking a right that he in fact possesses. She does this if she treats him as morally responsible and hence liable. But if she takes her evidence to be that he’s non-responsible, then she doesn’t render herself liable if she tries to kill him (with the concomitant implications for what third parties may (not) do to Resident). So the uncertainty about whether someone meets the threshold for responsibility seems to have significant implications for your view (as well as posing problems for the MR account, as you say).

  31. Hi Helen,

    I totally agree that increased uncertainty about the liability of an attacker creates additional complexities for my view in terms of the liability of the defensive agent.

    But just to be clear, I don’t hold that it’s up to the defensive agent to subjectively decide whether the evidence she’s got is evidence the threat she faces is liable or nonliable: that’s something that is determined by the balance of available evidence and not the agent’s beliefs about the evidence. I’m not saying you were attributing that view to me, I just wanted to be clear since I think it’s entirely possible it wasn’t clear given what I said earlier!

  32. Thanks very much for the nice reply Jon – though obviously disappointing that you haven’t offered a complete theory of distributive justice in your book on defensive force :). I also just wanted to say thanks for an excellent book, and to you Joe and others for terrific discussion.

  33. Dear Jonathan,

    First, I want to congratulate you for this excellent book. The reading is lucid and clear (which is quite rare!), and the arguments are extremely interesting. Thank you for this book. And thank you to Joseph for a splendid review.

    I was reading the chapter on necessity. In p. 146 you write:

    Consider a case where Albert is wrongfully attacking Betty and she has three options: (i) she can seriously harm Albert and avert his attack, (ii) she can seriously harm a bystander and avert Albert’s attack, or (iii) she can escape Albert’s attack by incurring some minor cost… Betty would wrong Albert if she chooses (i), but not to the same extent that she would wrong the bystander if she chooses (ii)… Internalism, on the other hand, says that Albert is not liable at all, and thus it yields the counterintuitive result that Albert and the bystander … each would be wronged to the same extent.

    I am not sure this follows. It seems that it may well be the case that, in such a case, Albert and the bystander are liable to the same degree due to existence of option (iii). You might be relying, for the intuition you have to the contrary, on the fact that if option (iii) did not exist, then Albert would indeed be liable, and the bystander would not be. More importantly, I am not convinced that Betty may go to the party. It is at least arguable that it is necessary that she should not go to the party (as opposed to, which you agree, that it’s possibly all-things-considered wrong for her to go to the party.) You argue that if we accept this, it would be necessary, for example, for a battered woman not to go outside without her husband’s permission if he threatens to beat her if she does, and that such cases are “outrageous.” Indeed, it is perfectly true that “abusive men cannot demand their partners refrain from doing permissible things” (p. 150). But it seems to me that other accounts of necessity are quite able to handle such a case. What is especially wrong with an abusive husband’s act is not that he threatens his wife, but that he controls her life. This is a significant harm, that may, on both your account and other accounts of necessity, justify a woman inflicting defensive harm an abusive husband. Suppose Betty knows for a fact that this will be *the only time* that she will be threatened with harm and have to harm Albert if she goes to a party. It is at least arguable that in this case she should not go, but instead call the police.

  34. Dear Jon,

    I am coming to this late, and possibly too late for my comment to be published. (I have just started teaching new courses at a new place of work this week, and it has been hectic).

    I wanted primarily to echo some of Helen’s sentiments. I, too, find it extremely rewarding and happiness-inducing to engage with your philosophical views . Congratulations on a great book, and thanks also to Joe for a thought-provoking review.

    One small comment that I don’t think has been made already, though I have not yet had the chance to read all contributions with the attention that they deserve. Have you considered the thought that in Duped Soldiers (your original case), the intuition that “surely the peacekeeping force may intervene” may be influenced or strenghtened by lesser evil considerations? If we buy into the set-up of the case, it will (help) undermine the war effort of an unjust regime if the peacekeeping force intervenes. The expected effects of the intervention thus go beyond “saving (say) 100 non-liable civilians while killing 100 duped soldiers” and these further effects help justify intervention and might render it permissible even if the duped soldiers are not liable.

    I am asking mostly because I’m currently in the process of defending an evidence-relative account of liability (I call it a subjective impermissibility account). At the same time, I very much feel the pull of your Duped Soldiers case. Can we remove all suggestion of the relevance of lesser evil considerations while retaining a firm intuition that intervention is permissible?

    Thanks, and all the best!

  35. Thanks for an excellent book and for a great discussion, sorry for arriving late to the party.

    My question concerns how to distinguish culpable bluffs like Albert in Albert’s Bluff from apparent threats like the twin brother in Resident without appealing to beliefs or culpability to distinguish them.

    Your explanation for why Albert is liable in Albert’s Bluff, as far as I understand, is that he has given Betty sufficient evidence that he is about to murder her (47,221). Consider Evil Twin, who is aware that his brother is on the loose and that the authorities have notified Resident. Evil Twin maliciously decides to scare Resident by walking up to Resident’s house to make him believe that he is facing a lethal threat. Presumably Evil Twin would be liable on the same grounds as Albert is in Albert’s Bluff.

    Consider now Dumped Twin. The authorities have made evidence about their notification of Resident and his brother’s escape available to Dumped Twin, but Dumped Twin innocently ignores the evidence because he was just dumped. How do we distinguish Dumped Twin from Evil Twin? From what I gather, lack of awareness about available evidence is irrelevant on the Moral Status account the same way someone’s beliefs about one’s evidence is irrelevant, though perhaps I’m mistaken about this?

    It is of course a possibility to deny that available evidence can ever be *innocently* ignored, but I’m not sure how this would help the Moral Status account, given that it denies culpability any direct relevance to liability.

    I understand if you’re too exhausted to reply to this after the number of comments you’ve already offered, though I would of course love to hear your thoughts on this.

    Again, I really enjoyed reading the book, Joe’s review as well as this discussion.

  36. Hi Elad,

    Thanks very much for the kind words, and for these good comments.

    Regarding the case on p. 146… I don’t believe Albert and the Bystander are nonliable to the same extent, nor do I think (though of course it’s difficult to be sure) the intuition is explained purely by the fact that only if option (iii) was not available, Albert would be liable. One way to try and tease out our intuitions about my case is to ask how much defensive force Albert would be permitted to use if Betty wrongly chooses option (i), and then compare that to how much defensive force the Bystander would be permitted to use if Betty wrongly chooses option (ii). My firm judgment is that Albert would not be permitted to deploy as much force as the Bystander. Unlike the Bystander, if Betty chooses to wrongfully harm him, Albert cannot credibly complain “this forced choice situation is in no way my doing—I’m not morally on the hook for the fact that someone has to suffer harm here”. Of course Betty is failing to take the low-cost option (iii) and so she wrongs him to some extent, just not to the same extent as the Bystander, who can credibly make the statement above. I think that’s one way to motivate the appeal of partialism in this case. But of course the argument here does appeal in part to one’s intuitive reaction to the case, and I understand maybe not everybody shares this intuition.

    Regarding whether Betty may go to the party and the issue of battered spouses…First, just to clarify (though you also note this), I do not say that Betty may go to the party, I only say that she does not violate the necessity condition if she chooses to go to the party. As you say, I suggest it could be all things considered wrongful for her to go.

    You go on to suggest that other accounts of necessity may also be able to explain why an abusive spouse lacks the right to demand of their battered partner that they refrain from doing certain permissible things. You think this can be true because of the more general and deeply wrongful harm involved in the attempt to coercively control the life of the battered partner. In most cases we must see what happens not as an isolated instance but as a component of a larger pattern of wrongful control, and even some of the accounts of necessity that I consider and reject can then get the intuitive result. I don’t disagree with you about this at all. Where we disagree, I guess, is what the necessity condition implies even in the one-off instance, where issues of a pattern of coercive control are absent. My view is that in such a case (a husband who has never abused or threatened his spouse before but is going to do it this time if she does such-and-such permissible act) the husband does not have a claim right to demand that the partner refrain from performing the permissible act. First time would-be abusers cannot appeal, in my judgment, to the fact that they haven’t yet engaged in the most egregious forms of coercive control as a justification for demanding their partner refrain from doing things that she is permitted to do. When I try and reflect carefully and vividly on such a case, I have the firm judgment that the would-be abuser cannot make this moral demand. I also think there is a good independent explanation why they cannot, namely, that doing so fails the sort of interpersonal test that I borrow from Cohen.

    But I do understand why one might be tempted by the opposing point of view, and you aren’t the first person to press me on this issue. Thanks for pressing me on both these points!

  37. Hi Susanne!

    Thanks very much for this, and great to have you join the discussion.

    I think your worry—that intuitions about my original Duped Soldiers case might be contaminated by lesser evil considerations—is a reasonable one. So let’s imagine a version of the case where this issue is eliminated. Suppose the Duped Soldiers are fighting on the just side in a just war. In this version of the case the soldiers are tricked not by their government, but by their rogue commanding officer. This officer has developed a hatred for the enemy that has warped his moral perspective and so he now wants to kill innocent civilians on the other side. So he tricks his troops (just as convincingly as in the original) to believe they are attacking liable enemy terrorists rather than nonliable civilians. Also suppose the commanding officer’s wrongful conduct has just been discovered by his superiors and they are going to relieve him of command and prosecute him, so there’s no risk he will continue to engage in this behavior. So the only question is whether the nearby third party forces can intervene and kill the duped soldiers to save the civilians. I remain as firm in my judgment that the neutral party can intervene in this way because the soldiers are liable. But it sounds like even when we’ve adjusted the example in this way, you don’t share this judgment?

  38. Hi Lars!

    Thanks very much for this nice set of further cases.

    You are right that on my view Evil Twin would be liable on the same basis as Albert in Albert’s Bluff. With regard to Dumped Twin, if he has been provided with the relevant evidence by the authorities (e.g. they call him and warn him not to go out because his evil twin brother is on the loose and people might mistake him for his brother) but he negligently forgets to heed the warning because he’s depressed, then yes, on my view he is liable when he appears at Resident’s door. This may seem harsh, but there’s no plausible account of liability (I think) that doesn’t have some harsh implications for some people (either defenders or victims of mistaken defense). There’s no way to design an account of liability that entirely avoids all the bad consequences of moral luck.

    But note also a point I made in response to Joe much earlier in the thread: you could jettison my view of proportionality while holding onto the moral status account to arrive at less harsh judgments. Of course I have independent reasons to think we should not do this, but the point is simply that there’s a way of combining the moral status account with a different view of proportionality to lessen the harshness of some of these implications (though doing so will of course increase the burden on the defensive agent).

  39. Hi Jon,
    Thanks for constructing a case to fit my criteria so well. It still seems to me intuitively permissible to kill the duped soldiers, I have to admit. I am not even sure my intuition is weaker than in the original. Maybe a bit.

  40. Thanks for the reply. When you describe Dumped Twin as negligently (i.e. culpably) forgetting to heed warning authorities has in fact made him aware of, I don’t find the implication so harsh. I was thinking of a case where Dumped Twin has not become aware of evidence made available to him by the authorities – say, they sent the information in a letter which he didn’t open – and where his failure to open the letter is innocent.

    But perhaps one can never be *innocently unaware* of available evidence. In that case, the upshot seems to be that apparent threats only become liable if they are culpable (either for ignoring available evidence that they might appear as an aggressor to others, or for intending to appear as an aggressor). This could indicate a disjunctive basis for liability, one for actual threats and another for apparent threats.

  41. Hi Lars,

    Ah, sorry I had not properly grasped what you had in mind with the Dumped Twin case. I do think there can be cases where someone non-culpably does not become aware of some piece of evidence of which they could easily have become aware (e.g. your housemate leaves a flyer on the kitchen counter hoping that you will glance at, and you easily might have, but you don’t and there’s no expectation that you ought to have done so). The case you describe feels slightly different since we might think: (a) the police really shouldn’t be sending such urgent and very serious information by post, and yet (b) there’s some expectation that people open their letters, though it’s unclear when they can be expected to do so. If the Dumped Twin is totally non-culpable for their lack of awareness (like the flyer on the kitchen counter) then I think they are not responsible for providing Resident with any evidence that they are a threat. They are like the twin in the original version of the story.

Comments are closed.