On the enforceability of poverty-related responsibilities

ABSTRACT We argue that even if an agent’s initial responsibilities are not very demanding, it can become permissible to burden her with significant costs if she culpably fails to discharge those responsibilities. In particular, we defend the claim that even if our responsibilities to assist others are not initially very demanding, our failure to live up to them can make us liable to possibly burdensome enforcement costs. Christian Barry and Gerhard Øverland (2016) disagree. They claim that other things equal, fewer costs may be imposed on an agent if she culpably fails to live up to her assistance-based responsibilities as opposed to her responsibilities not to contribute towards harm. Their thought is that our responsibilities to assist others are less demanding than our responsibilities not to contribute towards harm, and they assume that this asymmetry is matched by an asymmetry in the enforceability of the two types of responsibility. We agree with Barry and Øverland (2016) that our assistance-based responsibilities are less demanding than our contribution-based responsibilities. We argue that autonomy-based reasons support this asymmetry. Pace Barry and Øverland (2016), we claim that there is no reason to think that the two types of responsibility differ in their enforceability.


Introduction
It is relatively uncontroversial that as materially well-off individuals who live in liberal democracies, we have certain moral responsibilities towards the world's poor. But what happens if we culpably fail in these responsibilities? Is it morally permissible for others to enforce our responsibilities by imposing on us the costs necessary to secure the moral goods that our responsibilities were meant to protect? At least within deontological circles, it is standardly assumed that we need to clarify the nature of the responsibilities that we fail to live up to before we can start to answer these questions. More specifically, if we are culpably failing in what Christian Barry and Gerhard Øverland call our assistance-based responsibilities to help the poor, thenso the thought goesonly small to moderate costs may be imposed on us to enforce our responsibilities. 1 By contrast, if we are culpably failing in our contribution-based responsibilities not to contribute towards harming the poor, then potentially very significant costs may be imposed on us to enforce our responsibilities. We call this view on the different enforceability of assistance-versus contribution-based responsibilities the Asymmetry View. In this paper, we argue that the Asymmetry View is false, and that a Symmetry View should be put in its place instead. More precisely, we argue that whether a responsibility is contribution-or assistance-based is irrelevant to the level of costs that may be imposed on an agent who culpably fails to live up to it.
It is not entirely clear whether Barry and Øverland are sympathetic to the Symmetry View that we defend. Their discussion in chapter 3 suggests that they are. In chapter 3, Barry and Øverland investigate what happens when an agent fails in her assistancebased responsibilities. They argue that when an agent could easily assist a person in dire straits, but chooses not to do so, then the agent can later be required to bear significant costs to assist the person if assisting her has in the meantime become much costlier (Barry and Øverland, 2016, 33-4). In contrast with this, in chapter 9, Barry and Øverland introduce Force Implication, a principle according to which our stringent contribution-based responsibilities are highly enforceable, if necessary through the use of physical force. As formulated by Barry and Øverland, Force Implication applies only to contribution-based responsibilities, which suggestsin tension with the arguments that they present in chapter 3that Barry and Øverland endorse an Asymmetry View. Read as a comment on Barry and Øverland's thought-provoking book, our paper thus provides an argument for widening the scope of Force Implication to include assistance-based responsibilities, and in this way to resolve the just described tension in favour of a Symmetry View.

The demandingness of assistance and contribution-based responsibilities
Some of the responsibilities we have towards the world's poor are assistance-based. These responsibilities arise because many of the worlds' poor are striving to meet minimal subsistence needs, whereas we in rich parts of the world are generally living affluent lives. Moreover, we are in a position to assist the poor: there are things that we can do to help alleviate their suffering at little or modest cost to ourselves.
Peter Singer (1972Singer ( , 2009) is a vocal advocate of the view that most of us are failing in our assistance-based responsibilities towards the world's poor. As a utilitarian, Singer thinks that our assistance-based responsibilities towards the poor are quite demanding, 2 but he makes it clear that most of us are failing in our responsibilities even if these responsibilities are actually much more modest than he thinks they are. From a utilitarian point of view, it is plausible to assume that our assistance-based responsibilities towards the poor are quite demanding, as it would plausibly increase the sum total of welfare if we started to make significant personal sacrifices to help alleviate and prevent poverty-related harms. But many philosophers -Barry and Øverland includedresist the claim that our assistance-based responsibilities are possibly quite demanding.
2 In this paper, we call a responsibility demanding to the extent that an agent may be required to accept costs in order to live up to it. What we refer to as demanding Barry and Øverland (6-7) refer to as a constraining. They call a responsibility demanding to the extent that an agent who fails to live up to it has to bear costs to secure the moral good the responsibility was meant to protect. We refer to this latter aspect of a responsibility as its enforceability.
Their basic thought is that we have obligations to help others whose misfortunes we are not responsible for only if their plight is serious, and only if we are able to provide effective help at no more than moderate cost 3 to ourselves. Like Peter Singer, Thomas Pogge (2005, 2008 is convinced that most of us are failing in our responsibilities towards the global poor. But while Singer focuses on assistancebased responsibilities, Pogge concentrates on our responsibilities to refrain from contributing towards harming the poor. According to Pogge, most of us contribute towards poverty-related harms by acting in ways that help establish or further entrench 'institutional arrangements that foreseeably result in more severe or more widespread poverty or human rights deficits than would feasible alternative institutional arrangements' (Barry and Øverland, 2016, 122). Pogge (2008, 15) himself puts it more boldly and claims that we are outrightly 'harming [the poor by] inflicting seriously unjust rules upon […] them' (see also Pogge 2005, 60). 4 Unlike assistance-based responsibilities, contribution-based responsibilities are commonly thought to be possibly very demanding. To see the difference between the two types of responsibility, consider the following simplified cases (where 'ABR' and 'CBR' stand for assistance-and contribution-based responsibilities, respectively): Mudslide ABR. A mudslide is threatening a village. If youan otherwise uninvolved bystanderdon't do anything, two villagers will die. If you run to the village and warn the villagers, all villagers will live, but the option of warning the villagers comes at a certain cost to you.
Mudslide CBR. A mudslide is threatening a village. If youan otherwise uninvolved bystanderwait for a while before crossing a specific passage, the mudslide will remain contained, and it will not kill anyone. If you cross the passage right away, the mudslide will gather momentum, and it will kill two villagers.
If Barry and Øverland are right that assistance-based responsibilities are at most moderately demanding, then the costs that you are morally required to bear to warn the villagers in Mudslide ABR are limited. If warning the villagers means that you will miss a movie date, then you will have to miss your date. But if warning the villagers means putting yourself in harm's waymaybe because there is a real risk that you will be buried by the mudslide if you run to the village to issue a warningthen you are not morally required to help the villagers.
By contrast, in Mudslide CBR, the costs that you are morally required to bear to keep the mudslide from gathering momentum are generally thought to be potentially very high. Suppose that a vicious murderer is pursuing you, and that he will catch up with you and kill you if you don't cross the relevant passage right away. Most moral philosophers agree that even if your life is at stake, it is morally impermissible to cross the passage at once. If you cross the passage at once, you sacrifice two innocent villagers to save your own life, and it is generally accepted that you lack the moral licence to do so (see e.g. Thomson, 1991, esp. 296;Quong 2009, 507-8).
The enforceability of assistance-vs. contribution-based responsibilities For the purposes of this paper, we accept the asymmetry with respect to the demandingness of assistance-vs. contribution-based responsibilities. What we deny is that a similar asymmetry obtains with respect to the enforceability of the two types of responsibility.
We call a responsibility enforceable just in case an agent who has culpably 5 failed to live up to it is not wronged if costs are imposed on her to secure the moral good that the responsibility was meant to protect. The enforceability of a responsibility increases with the size of the costs that an agent can be required to bear if she has culpably failed to discharge that responsibility. To get a better grip on this definition, consider again Mudslide CBR. Suppose that you decide to cross the tricky passage right away, maybe to get to your movie date on time, or maybe to outrun your killer. Whatever your reason, by deciding to cross the passage right away, you culpably fail in your responsibility not to contribute towards harm. The moral good that your responsibility protects is the lives of the two villagers who will die if the mudslide gathers momentum. If we are morally permitted to impose certain costs on you to ensure that the villagers will liveif we may for example restrain you to keep you from crossing the passageand if we do not wrong you by doing so, then your contribution-based responsibility is enforceable. The more we may morally permissibly do to you in order to keep you from crossing the passage, the more enforceable is your responsibility not to contribute towards harming the villagers.
As mentioned, it is usually thought to be permissible to enforce contribution-based responsibilities through the use of significant force. For instance, it is usually assumed that if an agent threatens to culpably inflict harm on a victim, then the victim, or a third party acting on the victim's behalf, is permitted to defend the victim against the impending harm provided two conditions are met. First, the costs imposed on the attacker must be proportionate to the harm averted. If an attacker threatens only to punch his victim, then it is disproportionate to kill him to keep him from doing so. Second, the costs imposed on the attacker must be necessary to avert harm to the victim. If it suffices to forcibly restrain an attacker to keep him from hurting his victim, it is impermissible to kill him, even if killing him would be proportionate.
In support of the view that contribution-based responsibilities may permissibly be enforced through the use of potentially significant force, Barry and Øverland defend the 'conditional claim that if we possess stringent contribution-based responsibilities [towards the world's poor], then we are liable to the use of defensive force' [if we fail to discharge these responsibilities] (Barry and Øverland, 2016, 173, original emphasis). Barry and Øverland call this conditional claim Force Implication. The fact that they formulate Force Implication only with respect to contribution-based responsibilities suggests that in their view, a responsibility's enforceability rises with its demandingness. After all, the main difference between the two types of responsibility is that contribution-based responsibilities, unlike assistance-based responsibilities, are potentially very demanding. If Barry and Øverland are correct to suggest that the scope of Force Implication extends only to contribution-based responsibilities, then the Asymmetry View is correct, and assistance-based responsibilities are never permissibly enforceable through the use of significant force.
Although our aim is to argue against the Asymmetry View, we grant that it is prima facie plausible. To see what lends the Asymmetry View its initial plausibility, consider again the two mudslide cases. In both cases, the moral good at stake is the same, and it is highly significant. If you fail to live up to your responsibilities in either case, two innocent villagers will die. Now in Mudslide CBR, we said that your contribution-based responsibility is highly enforceable: if you attempt to cross the relevant passage right away, others can kill you without wronging you if this is necessary to keep you from crossing the passage. In Mudslide ABR, by contrast, if you culpably refuse to warn the villagers because it would be inconvenient for you to do so, it seems counterintuitive that others would not wrong you if they inflicted serious harm on you to keep the villagers safe. After all, you had no assistance-based responsibility to suffer serious harm to keep the villagers safe, so why should others now have the right to inflict this harm on you to guarantee the villagers' safety?
The Asymmetry View is not only prima facie plausible; it is also the view that is standardly held by moral philosophers working on the distinction between our positive duties to aid and our negative duties not to harm. 6 By contrast, the Symmetry View that we defend in this paper does not seem to be explicitly endorsed by anyone. 7 According to the Symmetry View, contribution-and assistance-based responsibilities are similarly enforceable if we hold fixed the moral good that the responsibilities protect. As long as the costs imposed on the culpable agent are proportionate and necessary, the culpable agent is liable to have these costs imposed on her. It follows that on the Symmetry View, whatever costs you are liable to have imposed on you if you culpably fail in your responsibilities in Mudslide CBR, you are also liable to have imposed on you if you culpably fail in your responsibilities in Mudslide ABR. Hence if you make yourself liable to be killed if you culpably fail to discharge your responsibility in Mudslide CBRas we think you do, provided killing you is necessary to keep the villagers safethen you can become similarly liable to be killed in Mudslide ABR.
At first blush, the Symmetry View may seem implausibly harsh. But once we consider what we take to be the most plausible rationale behind the view that assistance-based responsibilities are at most moderately demanding, it becomes doubtful that we should accept the putative positive correlation between the demandingness and the enforceability of a responsibility that Force Implication intimates.
In our view, the most compelling justification of the claim that assistance-based responsibilities are at most moderately demanding appeals to the value of autonomy. 8 If our responsibilities to help others are limited in nature, this prevents morality from becoming overly burdensome. Extensive and possibly very demanding assistance-based responsibilities would interfere with our ability to lead a self-determined life, as they would dramatically reduce our freedom to pursue a variety of plans and projects that we find personally meaningful. By contrast, if assistance-based responsibilities are only moderately demanding, this will generally leave us with an agentially meaningful and valuable set of permissible options for action.
Unlike assistance-based responsibilities, highly demanding contribution-based responsibilities do not similarly conflict with the value of autonomy, as they usually leave agents with a range of significant options for action. The general thought here is that contribution-based responsibilities only constrain our behaviour by excluding certain options from the set of actions that we may permissibly perform; they do not dictate to us particular ways in which we ought to allocate our time, our money, or our attention. To illustrate, if I am morally required never to kill anyone, this is not a meddlesome requirement: I can generally respect it while living my life in any of a number of meaningful ways. Whereas if I have a duty to try and ensure that no one dies a premature death, I have my work cut out for meif I want to live up to this duty, this will to a large extent determine what shape my life can take.
Suppose, as we do in the context of this paper, that this autonomy-based rationale in favour of only moderately demanding assistance-based responsibilities is correct. If so, it suffices to protect the value of autonomy that our assistance-based responsibilities are limited in their demandingness. There is no need to have similar limitations built into their enforceability. On the view that we defend, whatever enforcement costs that exceed the moderate costs that an agent was initially required to bear are always costs that the agent had a prior chance to avoid. Moreover, they are costs that the agent would have avoided had she lived up to her moral responsibilities (cf. Tadros 2014, 365). The fact that moderately demanding responsibilities are highly enforceable therefore does not prevent an agent from living a self-determined life. As long as the agent sticks by moderately demanding moral rules, she gets to pursue plans and projects of her own.
In sum, there needn't be any constraints on the enforceability of our assistance-based responsibilities to protect the value of autonomy. 9 Yet this in turn raises the question what the appropriate considerations are that determine the extent to which our responsibilities are enforceable. In their discussion of this question, Barry and Øverland (2016, 38) refer to Jeff McMahan's account of liability to defensive harm. If an agent is liable to a particular harm, she is not wronged if others inflict that harm on her. According to McMahan (2002), an agent's liability to defensive harm is grounded in considerations of local distributive justice. More precisely, McMahan claims that an agent becomes liable to defensive harm just in case she is morally responsible 10 for a threat of unjustified harm towards someone else. McMahan's basic idea is that it is generally least unjust if the person who is responsible for a threat of unjustified harm is burdened with the costs that will necessarily accrue if the threat of unjustified harm is to be averted (McMahan 2002, 401-5). While McMahan spells out this rationale only with respect to contribution-based responsibilities, we think that a similar rationale can 9 Tadros (2014, 367) argues that our right to live self-determined lives would be 'too fragile' if our sometimes quite understandable 'difficulties [to conform to our] positive obligations' could make us liable to be killed. We agree with Tadros that it can be difficult to conform to one's assistance-based responsibilities, for instance because it is not always clear what these responsibilities amount to. But to the extent that such difficulties are genuine, they will tend to excuse agents for their failures to live up to their assistance-based responsibilities, and an excused agent is not culpable for her transgressions. 10 On McMahan's use of the term, an agent can be morally responsible for a threat of unjustified harm without being blameworthy for it. We disregard this complication here.
be formulated also for assistance-based responsibilities. More precisely, it seems to us that when an agent culpably fails in her assistance-based responsibilities, it is generally least unjust if she has to bear the costs that are necessary to secure the moral good that the agent had a moral responsibility to protect. After all, the agent had an initial moral responsibility to protect the good in question, and she could have avoided being burdened with significant costs simply by living up to this responsibility. In sum, while the demandingness of assistance-based responsibilities is plausibly constrained by the value of autonomy, their enforceability is governed by considerations of justice, and considerations of justice speak in favour of burdening a culpable agent with potentially significant enforcement costs. Methodologically speaking, we believe that Barry and Øverland have done a valuable job at identifying different normative characteristics of responsibilities, in particular by drawing attention to the difference between a responsibility's demandingness and its enforceability. But we also believe that more work needs to be done to explore the relationship between these different characteristics. It seems to us that a fruitful way of doing this is by investigating what considerations might ground or constrain the different normative characteristics.

Conclusion
To be poor is to be deprived of the material security that is needed to lead a flourishing life. For a poor person, seemingly minor disturbances can quickly become lifethreatening. It follows that if we culpably fail to live up to our responsibilities towards the world's poor, we fail to secure important moral goods. On the Symmetry View, this means that we become liable to the infliction of necessary and proportionate harm, regardless of whether our responsibility was assistance-or contribution-based. This suggests that at least in principle, redistributive wars fought by the poor, or fought on behalf of the poor, could be morally permissible (for defences of such a view, see Lippert-Rasmussen 2013; Fabre 2012).
While the Symmetry View has this implication in principle, its practical implications are most likely far less dramatic. According to the Symmetry View, those who culpably fail to discharge their poverty-related responsibilities only become liable to harm if such harm is both necessary and proportionate to the reduction in poverty-related harms that is thereby achieved. Yet in the world as we know it, it is difficult to see how the use of violence could help reduce poverty. Physically harming affluent people in the fight against poverty sounds like a futile undertaking at best; at worst, it sounds like an ill-conceived idea that would inevitably backfire (cf. Pogge 2013). Needless to say, futile or counterproductive use of force can never be justified, and would wrong those who are harmed.
Ways of imposing cost that do not include the use of physical violencethink hacking the bank accounts of relevantly culpable affluent individualsare more plausible candidates for the permissible enforcement of our poverty-related responsibilities. But even with non-violent measures, it seems doubtful that they could sustainably reduce or put an end to poverty if they are not also accompanied by institutional reform. And it is extremely difficult to see how institutional reform could ever be pushed through, violently or not, in the absence of affluent countries' willing cooperation.
These considerations suggest that even if poverty-related responsibilities are enforceable through violent means in principle, they are not enforceable through violent means under present circumstances. In light of this fact, it is all the more important that other avenues for combating poverty are explored, and that sustained philosophical and political attention is directed at how we can best live up to our responsibilities towards the global poor.