1.

Wars involve the intentional killing of people. Under what conditions, if any, could wars be fought permissibly? Much contemporary work on just war theory begins with the thought that such killing could be permissible if those being killed had made themselves liable to be killed by unjustifiably contributing to an unjust war effort. To say someone is liable to be killed is to say that killing them does not wrong them, because they have forfeited their rights against being killed. The focus then turns to under what conditions people make themselves liable to be killed. In The Mechanics of Claims and Permissible Killing in War, Alec Walen suggests this focus is misguided. While forfeiture may account for why many people may be killed in war, Walen claims others who contribute to an unjust war, who are not culpable for posing the threats they pose, have not done enough to forfeit their rights against being killed.

Take the following example from outside of war, Hologram: ‘Polly projects a holographic gun onto Howard’s hand as he extends [it] to greet Diane. Diane, taking the gun to be real, reaches for her own gun to shoot Howard in self-defence. Howard, seeing the holographic gun too, realizes that Diane will shoot him in self-defence if he does not shoot her first with the real gun he is carrying in a holster’ (4).Footnote 1 Walen holds that Howard is justified in defending himself, while Diane is unjustified in defending herself against Howard. Yet, Walen presses, if we begin with the assumption that to forfeit one’s rights against being killed there needs to be an important moral difference between the threatener and their victim, we cannot explain why Diane forfeits her right not to be killed—after all, her ‘mistake is reasonable, and she acts non-culpably, given what she reasonably believes’ (4). While Walen does not think Diane has done enough to forfeit her right against being killed, he does hold that she has done enough to weaken her claim not to be killed to such an extent that it is permissible to kill her. To accommodate this idea requires rethinking the way rights work at their most fundamental level.

What results, in The Mechanics of Claims and Permissible Killing in War, is a book of two substantive parts. The first concerns rights theory. On the dominant way of thinking about rights, which Walen calls the infringement model, rights provide a weighty protection against others’ actions (or inactions). But they can be permissibly overridden at times, when enough good is on the line. In such cases, we say peoples’ rights have been merely infringed, and leave a moral remainder. On this picture, rights function as inputs into moral deliberation. Walen suggests this is mistaken. Instead, according to his mechanics of claims, ‘people have claims that protect their interests, but whether a particular claim corresponds to a right depends on how all the competing claims interact’ (xvi). So, it is not that people may be killed only when, either, they have forfeited their rights against being killed or their rights may be permissibly overridden. Instead, one may be killed because their claim not to be killed is simply weaker than others’ claims to kill them, resulting in their not having a right against being killed in the first place (7).

The second substantive part of the book brings the mechanics of claims to bear on killing in war. Walen claims that ‘noncombatants who help unjust combatants kill unjustly have at most contingent rights against being eliminatively killed’ (212). This class of unjust noncombatants includes, among others, ‘people who design or manufacture weapons, who manufacture parts for weapons, who provide other equipment needed by combatants such as reconnaissance drones, who do linguistic analysis of information intercepted from the enemy, or who raise money to provide resources for combatants […] They may be targeted wherever they are, whatever they are doing, as long as doing so is necessary to reduce the threat posed by an unjust enemy, is proportional to the good that will be caused, and is not for some other reason morally prohibited’ (213).

The book contains many interesting arguments. It raises important challenges that need to be answered by defenders of the infringement model if they are to hold onto their view. And while some complain that contemporary work on self-defence and just war theory is a little shallow, resting on too much of an appeal to mere intuitions about cases, it is more difficult to raise this complaint with Walen’s book.Footnote 2 Yet, all this comes with a cost. The first half of the book on rights theory requires very careful reading, especially for those uninitiated in rights theory and contemporary work on self-defence. And the second part of the book, on permissible killing, is hard to assess without a grasp on Walen’s mechanics of claims. Ultimately, Walen has underestimated the costs, as against the benefits, of this approach. So I suggest below, the putative insights of the mechanics of claims and how it is used to think about permissible killing could be accommodated by the infringement model without putting the reader through over 100 pages of complicated rights theory.

2.

Instead of having rights as inputs into all-things-considered permissibility, as on the infringement model, Walen suggests that people have claims, which combine and compete with others’ claims. The mechanics of claims is then meant to explain ‘how rights emerge as outputs based primarily on claims as inputs’ (65). Patient-claims are claims that others perform, or not perform, some act. Agent-claims are claims to be free to act or not to act in certain ways. The weight of someone’s claim reflects the magnitude of their interests at stake, as well as the type of claim it is. Among other distinctions, negative claims are stronger than positive claims (58–65), and so-called restricting claims are substantially weaker than non-restricting claims (read here: claims against being harmed in ways that use one as a means are stronger than claims against being harmed in ways that do not use one as a means) (82–97).

How these claims are balanced is a result of three principles. First, we have the autonomy principle, according to which ‘we all have our own lives to lead; the space of rights must give us substantial freedom to do so, while also holding us accountable for our choices.’ Second, we have the equality principle, according to which ‘we are fundamentally equal as members or “citizens” of the space of rights.’ And finally, we have the welfare principle, according to which ‘welfare matters, and that value affects the rights we have’ (47). Walen asks us to think of these principles as axioms, analogous to those in geometry. Were one to focus only on autonomy, one would end up with egoism; were one to focus only on autonomy and equality, one gets libertarianism; and, were one to focus only on equality and welfare, one is left with consequentialism. The mechanics of claims balances these three principles to deliver a moderate form of nonconsequentialism (47).

An example will help. Suppose a trolley is headed towards five people, whom it will kill if not diverted. One person is trapped on a side-track. Bystander is at a switch, which will allow her to divert the trolley from the five towards the one. The infringement model says the one on the side-track has a right against being killed, correlating with Bystander having a (pro tanto) duty owed to her not to turn the trolley. Yet, the good to the five of being saved is sufficient to outweigh this duty, and so it is permissible for Bystander to turn the trolley. (Of course, infringement theorists may disagree about the numbers.) The mechanics of claims sees things differently: Bystander ‘is confronted with six basic patient-claims, and she also gets to consider her own agent-claim. The patient-claims break down into two groups: the five have positive claims to be saved; the one has a negative claim not to be killed. [Bystander] has a negative agent-claim not to have to turn the trolley if she does not want to and logically could have [though Walen argues this is not the case] a positive agent-claim to be free to turn it if she wants to’ (68). And Walen suggests the balancing of these claims gives Bystander a right to turn the trolley.

3.

It might have occurred to the reader that claims on Walen’s picture look like rights on the infringement theorists’ picture, just with different labels.Footnote 3 What’s more, Walen and the defender of the infringement model can agree about why Bystander may divert the trolley—their disagreement concerns only whether this affects whether the one has a right against Bystander turning the trolley on them. (The infringement theorist says ‘No’, while Walen says ‘Yes’.) In fact, Walen and infringement theorists can agree about the why behind every first-order question raised in the book (for example, whether Howard may kill Diane in Hologram, whether those who justifiably harm innocent people may be killed, and so on). This raises the question, why does it matter whether we endorse the infringement model or the mechanics of claims?

Walen claims, about Hologram and similar cases: ‘Though there is no reason to think that they [threateners in these cases, such as Diane] have forfeited their right not to be killed, there is good reason to think that they have weakened claims not to be killed and that it is therefore at least plausible that they may be killed’ (7). It looks like Walen thinks the infringement model and mechanics of claims come apart on first-order implications. However, I am confused about why this is. Whatever Walen holds explains why Diane’s claim is sufficiently weaker than Howard’s to make it permissible to kill her in Hologram, a defender of the infringement model can appeal to ground a liability-based justification.Footnote 4 Why could they not? Ultimately, what’s at stake is whether someone such as Diane in Hologram has a right against being killed in defence of Howard. It doesn’t matter whether she doesn’t have that right because she’s forfeited it, as an infringement theorist will suggest, or because she didn’t have it in the first place, because her claim was sufficiently weaker than Howard’s claim.

Later in the book, Walen suggests that appealing to forfeiture opens up an ‘inappropriate normative gulf’ between those who’re liable and those who aren’t liable: ‘If someone retains the right not to be killed, only a substantial consequentialist improvement or stronger competing right would justify “wronging” that person [on the infringement model]. The flip side of that is that if someone has forfeited a right’, they have lost this protection (111). But contra Walen, shouldn’t we expect there to be a significant normative gulf between those who may be killed and those who may not be killed?

Walen suggests this normative gulf leads to a problem: When threateners have forfeited their rights, it will be permissible to eliminatively kill however many of them is necessary to save only one victim; but doesn’t this seem implausible when it comes to threateners who aren’t fully culpable (111)?Footnote 5 However, this simply doesn’t follow. Just because one has forfeited their rights against being harmed doesn’t mean others have no reason not to kill them—it means only one’s reason not to kill them is greatly diminished in virtue of their not having a right against being killed. And, we can hold, the weight of one’s reason not to kill a liable party depends on how morally responsible or culpable they are for posing the threat they pose, so we have a weaker reason not to kill a liable fully culpable threatener than we have reason not to kill a liable non-culpable threatener. (In fact, we could incorporate whatever account Walen likes for how, on his model, one’s claim is weakened.) This could mean, for example, that one may not kill five non-culpable threateners, as in Hologram, to save only one victim, despite all five being liable, because the remaining reasons not to harm them are weightier than the reason to save the victim. (We could even hold, as Patrick Tomlin has recently suggested, that the number of potentially liable parties that need to be harmed in defence of their victims bears on whether they have forfeited their rights in the first place. There is growing discussion of this as it relates to fully culpable threateners; and whatever is true of our reason not to harm fully culpable threateners will be true, to more of an extent, of minimally responsible or non-culpable threateners.)Footnote 6

Walen thinks there are other reasons to prefer the mechanics of claims (100–115). The principal reason he offers concerns compensation. It has traditionally been seen as a strength of the infringement model that it deals well with compensation (and moral remainders, more generally, such as apology or other restitutive reactive attitudes being owed to wronged parties). In our trolley case from above, since the infringement model says that the person on the side-track is wronged by having the trolley turned on them, albeit permissibly, that can accommodate why something might be owed to them (or their estate) in response. Walen suggests the infringement model struggles with other cases, however. In Involuntary “Punishment”, ‘Ruth has […] committed no crime, but the benefits of “punishing” her for six months would be tremendous, and her suffering would not be too great. She therefore has a duty to submit to it even though she has not volunteered to submit to it’ (104).Footnote 7 The problem, as Walen sees it, is that ‘it’s her enforceable duty to take on such “punishment.” Enforcing such a duty seems to involve no wrongdoing at all. But the infringement model accounts for a duty of compensation by saying that it is to make up for having suffered a wrong’ (104). Walen suggests, instead, compensation is better explained if ‘the bad luck that properly belongs to one set of patients is transferred to another set’ (102); and Walen thinks Ruth absorbs others’ bad luck in Involuntary “Punishment”.

Let me say two things in reply. First, there is nothing to stop the infringement theorist agreeing with Walen that absorbing others’ bad luck grounds rights to compensation, accepting for sake of argument that his is the correct view of compensation. While this might be dialectically disadvantageous, inasmuch as accommodating rights to compensation is usually seen as a strength of the infringement model, it doesn’t itself give us a reason to be sceptical of the infringement model. And since the infringement model is so much simpler in general and everyone’s familiar with it, it’s preferable to stick with it plus Walen’s view of compensation, rather than moving to the mechanics of claims.

Second, in any case the infringement model can accommodate Ruth’s being owed compensation. Here is what the view says. (i) Ruth has a right that others not ‘punish’ her, correlating with others owing her a duty not to ‘punish’ her. We can unpack this duty by saying others have reason not to ‘punish’ her, which is some-what exclusionary in character, such that even if the reason is overridden, it takes a significant amount of countervailing good to even register as an opposing reason and leaves a remainder. At the same time, (ii) others are under a conflicting duty to ‘punish’ Ruth, given the good this will do. Finally, (iii) Ruth herself is under a duty to be ‘punished’, meaning she may not resist others’ attempts to ‘punish’ her; since the duty is presumptively enforceable, and not overridden by conflicting considerations, others may make her not resist. But this does not disable others’ overridden duty not to punish her—it simply outweighs it. Since duties are pro tanto on this picture, I am unsure what the problem is supposed to be. All three claims are consistent, and seem quite natural to my eye. And it’s because of the remainder-inducing pro tanto duty not to ‘punish’ Ruth that means she is owed, among other things, compensation after being punished.

4.

None of the preceding itself speaks against the mechanics of claims. Rather, it suggests the putative strengths of the model are illusory. And since introducing the reader to the mechanics of claims is tough going, and ultimately gets us nothing we couldn’t have had on the more familiar infringement model, this is a real limitation with The Mechanics of Claims and Permissible Killing in War. What’s more, since the rest of the literature is couched in the language of the infringement model, it’s often difficult to compare claims made in the second half of the book, concerning permissible killing in war, with other claims made in the literature. Choice points and possible moves one can make are obscured by moving between the two ways of talking about rights. This is all a shame. There are lots of interesting arguments and challenges in the book. I worry they will be missed—and perhaps justifiably so, with so much other work to read that doesn’t require such a commitment—because readers need to work through the mechanics of claims.

Let me conclude by saying a few things that press on the mechanics of claims on its own terms. Recall that, on the infringement model, whether one has a right against being killed does not settle whether it’s permissible to kill one. Part of the appeal of rethinking the infringement model has to be that we get a tighter connection between rights and requirements. If we abandon the infringement model, when one says ‘I have a right that you act thusly’, this would mean more than merely ‘You’ve a special kind of reason to act that way’. Rather, it could mean ‘You must act that way’. And since the mechanics of claims sees rights outputs, doesn’t it have this virtue? Unfortunately, not.

Walen suggests that ‘the three principles [the autonomy, equality, and welfare principles] and the space of rights they ground do not offer a complete framework for making sense of what is and is not “wrong” to do’ (55). This would be fine if this sense of wrong didn’t bear on what’s morally permissible, as then rights could be concerned with what’s permissible even if not with all that’s wrong. But Walen suggests ‘there may be duties of justice that cannot be explicated by reference to particular claimants who have a right that the agent perform her duty’ (56). Walen offers the examples of duties of fairness, retributive justice, and agent-neutral value. (Note that if many of the puzzles of population ethics are not to be solved using person-affecting views, there will be many important duties of agent-neutral value.) This raises two points. First, it means the mechanics of claims, like the infringement view, doesn’t settle what’s permissible; this removes a great dialectical advantage that the mechanics of claims has over the infringement model. Second, the infringement model is prima facie better placed to answer how these duties interact with rights, since infringement theorists already admit that rights, while an especially significant (and distinctive) input into the all-things-considered status of an act, are one among other inputs.

There’s a second limitation on the mechanics of claims. Walen thinks there is a range of cases in which ‘some sort of threshold is crossed such that it is no longer morally wrong to violate rights’ (116). For example, ‘a parent might feel obligated to turn a trolley away from her child onto another, even though the other has a right not to have a trolley turned onto him for the sake of one person’ (115). In this case, Walen thinks ‘it seems right to say that it is some sense permissible, if not obligatory, to do something inconsistent with the respect presupposed by the space of rights’ (116). He calls these cases of threshold deontology.

This seems to do away with much, if not all, of the dialectical force of preferring the mechanics of claims over the infringement model. As may be apparent from the rest of this review, I think of rights as the infringement theorist thinks of rights. But I admit there are questions that need to be addressed, which Walen rightly raises, about how the good to others sometimes makes it permissible to infringe others’ rights. But by opening this possibility for the mechanics of claims, Walen reintroduces all these lacunae into his view. In defence, Walen suggests that threshold deontology is much more extraordinary than the permissible infringing of rights on the infringement model: ‘It represents a failure of the space of rights to function as it should, and if such failures were common, we could not rely on rights as we do’ (119). But I am not sure contingent rarity should be valued to such an extent. And as Walen notes in a footnote, it seems ‘one virtue of the infringement model [is that], like a willow, rights yield rather than break’ (119, fn. 30). And I much prefer it that the space of rights doesn’t break in the way that Walen thinks it does.