Arguing for assistance-based responsibilities: are intuitions enough?

ABSTRACT Millions of people in our world are in need of assistance: from the global poor, to refugees, from the victims of natural disasters, to those of violent crimes. What are our responsibilities towards them? Christian Barry and Gerhard Øverland’s answer is plausible and straightforward: we have enforceable duties to assist others in need whenever we can do so ‘at relatively moderate cost to ourselves, and others’. Barry and Øverland defend this answer on the ground that it best fits our intuitions in a variety of hypothetical rescue scenarios. I argue that, although Barry and Øverland’s view is intuitively appealing, appeal to intuitive cases is insufficient to vindicate it satisfactorily. Intuitive cases alone do not allow us to establish: (i) what costs count as moderate and (ii) whether assistance-based responsibilities are, in fact, enforceable. These considerations suggest that Barry and Øverland’s defence of their preferred answer to the assistance question may be incomplete.


Introduction
In Part I of their thought-provoking book, Responding to Global Poverty, Christian Barry and Gerhard Øverland discuss assistance-based responsibilities, namely the responsibilities that we have to help others in need when we are not responsible for their plight. Barry and Øverland defend a highly intuitive view, which they call 'Moderate': Moderate: if we can prevent something (very) bad from happening at relatively moderate cost to ourselves, and others, then we ought to do it (Barry and Øverland 2016, 21). 1 Moderate is said to be superior to rival views because it best fits our intuitions in a variety of small-scale rescue scenarios, along the lines of Peter Singer's 'drowning-child' case. The conclusions reached in these hypothetical scenarios are then extended, with qualifications, to the large-scale case of global poverty (Barry and Øverland 2016, chap. 4). In the present piece, I discuss Moderate only in relation to one-on-one easy rescue scenarios, setting global poverty aside.
As it happens, the official formulation of Moderate somewhat understates the strength of Barry and Øverland's view, precisely as it applies to easy rescue cases. As their subsequent CONTACT Laura Valentini l.m.valentini@lse.ac.uk London School of Economics and Political Science, London, UK 1 Note that this is a full quotation from the book. discussion makes clear, Barry and Øverland believe that, in those cases, (i) a bystander's duty to assist a needy person is owed to her (hence, it is correlative to her right) and (ii) costs may be imposed on the duty bearer to compel her to discharge her duties (Barry and Øverland 2016, 33, 37-38 respectively). These observations suggest that, at least in one-onone rescue scenarios, Moderate is more accurately characterized as follows. 2 Moderate: the needy/imperilled have an enforceable right to assistance, if we can assist them at relatively moderate cost to ourselves, and others. This is the version of Moderate on which I will focus here. Challenging it may look like a daunting task. After all, Moderate nicely captures what common sense tells us about our duties to assist the imperilled. It seems obvious that you are not morally required to lose a limb or risk your life to save a drowning child with whom you have no prior relation. And it is equally obvious that you are required to risk getting a flu and being off work for a week (with the associated income loss) to save that child's life. It also seems obvious that you would wrong the child (hence violate her rights) if you failed to rescue her, and that some costs could be permissibly imposed upon you to compel you to perform the rescue.
Although Moderate is intuitively appealing, appeal to intuitive cases is insufficient to vindicate it satisfactorilyor so I will suggest. This is because intuitive cases alone do not allow us to establish: (i) how we should measure the costliness of rescue and (ii) whether responsibilities to assist in one-on-one rescue scenarios are a matter of enforceable right. Since Moderate asserts that assistance-based responsibilities correspond to enforceable rights in those scenarios, and since it can only be action-guiding if supplemented by a criterion for measuring costs, Barry and Øverland's exclusive reliance on intuitive cases is insufficient fully to vindicate their preferred view. If this is right, despite its many virtues, their case for Moderate remains incomplete.
I proceed by first showing that resort to intuitions is insufficient to determine how the costs of assistance should be measured, and then turn to the idea that duties of assistanceat least in one-on-one rescue scenariosare correlative to enforceable rights. I conclude with some broader methodological reflections.
Determining the costs of assistance: why intuitions are not enough As anticipated, Barry and Øverland's defence of Moderate relies heavily on intuitive cases. I consider two. One is explicitly offered by Barry and Øverland (2016, 23-24), the other is my own, but is in line with their style of argument.
Barry and Øverland remark that, intuitively, Bob does not seem to be under an obligation to transfer all of his pension savings. Why? Because the costs of doing so would be too high. Now consider another scenario. 3 Amanda and the trolley: Amanda is in the proximity of a trolley track, where a trolley is heading towards a child. She could easily save the child by throwing her new Gucci handbag on the rail-lines, thereby stopping the trolley.
Here, I am sure that Barry and Øverland would conclude that Amanda is under an obligation to sacrifice her fancy handbag. In fact, Barry and Øverland seem to think that someone like Amanda may even be required to lose a finger to save the child, though perhaps not a whole arm (Barry and Øverland 2016, 27).
Taken together, these rescue scenarios give strong support to Moderate. Completely emptying one's bank account to save the life of an unknown child is asking too much, but giving up one's fancy handbag, or even a finger, for the same cause is not. To their credit, Barry and Øverland also consider some cases that appear to support a different conclusioni.e., that assistance may demand one to shoulder more than moderate costsbut they skilfully explain away our intuitions in those cases. 4 Even so, I am not persuaded that their intuition-pumps suffice to establish a fully satisfactory version of Moderate. 5 To see this, I shall indulge in a little bit of intuition pumping myself, offering modified versions of Bob's internet banking and Amanda and the trolley.
Bob's internet banking*: The scenario is identical to the previous one, except for the fact that Bob* does not care about his pension savings at all. In fact, he sees that money as a burden. He wishes to lead an ascetic life, a bit like Saint Francis of Assisi. He had previously put money aside intending to provide for his seriously ill daughter. But the daughter's untimely death has made that money worthless to him. For Bob*, giving away any of his daughter's old clothes is more costly than giving away the entire content of his savings account.
Amanda and the trolley*: The scenario is identical to the previous one, except for the fact that the aforementioned Gucci bag contains a framed photo of Amanda*'s mother. The woman died at a young age, in a tragic accident: she threw herself on a track to save Amanda* from being hit by a trolley. Unfortunately, Amanda*'s Gucci bag would only succeed in blocking the trolley if it contained the framed picture. For Amanda*, sacrificing the picture is incredibly costly. That picture is Amanda*'s most precious possession.
These modified versions of the two scenarios cause my intuitions to partly shift: while I am no longer sure that Bob* may refuse to transfer his savings to Jimmy's family, I still think that Amanda* ought to sacrifice her Gucci bag, though she could perhaps be excused for not doing so. The reason for this partial shift in intuitions is clear: in the revised scenarios, the subjective costs of assistance are made salient, and diverge from what we might describe as its objective costs. From an 'objective' perspective, given some broadly plausible conception of the good, someone's life savings are very valuable, while an old photograph in a fancy handbag is not. Yet, there might be individuals from whose perspectives things are exactly the other way round, like Bob* and Amanda*. Providing information about subjective costs unsettles my original evaluations of the cases, and in a way that introduces inconsistencies. In the case of Bob* I seem to intuitively privilege subjective costs, in that of Amanda* I give priority to objective ones.
This suggests that, if we want Moderate to help us determine what we ought to do in tricky assistance situations, we need to adopt a principled metric to calculate costs. Throughout their discussion, Barry and Øverland invoke intuitive cases where objective and subjective costs tend to coincide, thereby sidestepping this further question. But until we know which costs are the relevant ones, we do not know what Moderate actually demands in cases where subjective and objective costs come apart. 6 So, on what basis should costs be calculated? Barry and Øverland (2016, 32) say little about this, suggesting that costs should be understood in terms of wellbeing. 7 Reference to wellbeing, however, does not give us the answer we want, because wellbeing too can be understood either objectively or subjectively (or as a mixture of both): so the same problem resurfaces under a different label. 8 The reason I raise this matter is that I am sceptical about the possibility of settling the question of how costs should be measurede.g., whether by reference to an objective or a subjective account of the goodby relying on intuitive evidence alone. As I hope the scenarios I have offered show, intuition doesn't really push us one way or the other: both objective and subjective metrics have their intuitive virtues and vices. In some cases we tend to privilege subjective metrics (e.g., Bob*), in other cases objective ones (e.g., Amanda*); or, at least, I do. Appeal to some broader theoretical considerations may be necessary to establish whether we want costs to be calculated relative to an agent's own conception of the good (i.e., subjectively) or not. 9 One such broader theoretical consideration, for instance, may be whether the duty to assist is a matter of beneficence or one of justice (Buchanan 1987). In the former case, it would amount to a duty to use one's own resources to promote what one sees as the good of othersof course, provided this is consistent with respect for their rights (Barry 1991;Valentini 2015). Duties of beneficence, at least on a popular view, are established by reference to the duty-bearer's own conception of the good: this determines both where the duty-bearer directs his or her beneficence and which costs count as moderate. By contrast, if the duty to assist is one of justicenamely a duty that is both legitimately enforceable and correlative to rightsthen its existence conditions, including what costs count as moderate, should be determined relative to a 'public' conception of what is valuable, rather than relative to the duty-bearer's private conception of the good (Rawls 1999). 10 For instance, someone's complaint that paying taxes is too costly to him because of his unusual attachment to money makes no difference to our conviction that cheating on one's taxeswhen the tax system is fairis unjust. By failing to pay one's taxes, one violates the rights of one's fellow citizens. Furthermore, tax evaders may paradigmatically be forced to 6 Note that this 'metric' problem affects any moral view the prescriptions of which rest on cost-benefit calculations. 7 In a footnote (p. 21, number 27) they point to another unpublished manuscript where the question of costs is addressed. Since I am unfamiliar with that manuscript, I cannot comment on it. 8 For an overview, see Crisp (2016). 9 Appeal to such broader theoretical considerations would thereby 'widen' our reflective equilibrium process. For discussion, see Daniels (2013). 10 See especially Rawls's reference to primary social goods as 'all purpose means' that people may wish to have independently of their particular conceptions of the good.
comply with their duties, even if compliance is subjectively very costly to them. 11 What justice requires is (arguably) not dependent on the subjective preferences or favoured conceptions of the good of the parties involved. Yet, if a misogynistic man refused to donate part of his (rightfully owned) income to the 'Stop Domestic Violence Association', and funded the 'Veterans Support Association' instead, we would lack grounds for criticizing his actionsas opposed to his character. He would be doing the right thingi.e., acting beneficentlyeven though in part for the wrong reasons, insofar as his choice of charity would be somewhat driven by misogynistic beliefs. Similarly, it would not be contrary to the demands of beneficence for someone to volunteer in a bookshop, but to refuse to volunteer in a hospital, say, due to their fear of blood or their religious convictionseach of which would make the costs of volunteering in a hospital more than merely moderate.
Although Barry and Øverland do not explicitly connect assistance-based duties, justice and beneficence, they do so implicitly, in Chapter 3. They argue that duties of assistanceat least in one-on-one rescue scenariosare both enforceable and owed to others, hence a matter of justice in my terminology. 12 This, in turn, points in the direction of the relevant costs being calculated by reference to some public/objective conception of wellbeinghowever thin this may be. From this perspective, idiosyncratic preferences or tastes should not influence the existence conditions of duties to assist. Someone's unusual lack of attachment to money, and someone's unusual attachment to a picture, should arguably not be taken into account when considering whether a duty to assist exists or not.
This conclusion may involve some bullets to bite (to the extent that we may sometimes find subjective costs intuitively relevant, as with Bob*), and partly undermines the claim that Moderate succeeds in accounting for all of our considered judgements. A fully specified version of the view may well not. More broadly, if I am right that intuitions cannot do all of the work and reference to broader theoretical considerations is neededthe distinction between justice and beneficence being one possible candidate herethen Barry and Øverland's argument-strategy in defence of Moderate misses something.
Establishing whether duties of assistance are correlative to enforceable rights: why intuitions are not enough As I noted at the start, Moderate does not simply set out duties to assist, it states sufficient conditions for one to have an enforceable right to be assisted. For Barry and Øverland, the duty to assistat least in the rescue scenarios discussed up to this pointis owed to its beneficiaries, who may claim its performance and enforce it.
Again, at first sight, intuitions seem to be entirely on the side of Barry and Øverland. Consider, for example, Singer's (1972) drowning child scenario. Most of us would intuitively wish to say that the child has a right to be saved by the bystander, that the 11 Of course, in situations of genuine duresse.g., self-inflicted povertywe may want to excuse certain instances of non-compliance. Still, duress doesn't seem to cancel the underlying duties of justice; it only mitigates the wrongness involved in not complying with them. 12 Note that this does not rule out the possibility of there being duties that are not owed to others (hence not correlative to rights) but which are nonetheless enforceable. Indeed, Barry and Øverland defend such duties in more complex (not one-on-one) assistance scenarios. It is just that, in my terminology, those would not be duties of justice.
bystander would therefore wrong the child by not performing the rescue, and that the bystander may be forced to rescue the child. But of course, reporting our intuitions on this case is not enough to vindicate them. We need a principled explanation for what makes it the case that, in a scenario like Singer's classic drowning child (or Amanda and the trolley), those in need of assistance acquire an enforceable right to assistance. Note that grounding an enforceable right to assistance is not the same as grounding a duty to assist. A duty to assist is a simple moral ought. An enforceable right to assist, by contrast, is a moral ought coupled with an enforceable claim to its performance (Feinberg 1970). By analogy, my duty to respect the environment is a simple ought. It is a further question whether (i) this duty is owed to anyonesuch that he or she has the standing to demand its performance and would be wronged if I failed to complyand (ii) whether anyone may legitimately force me to perform it. So, what may ground the rather demanding normative structure typical of enforceable rights in cases of assistance?
Barry and Øverland again say little about this. They rely on intuitive evidence from stylized rescue cases. This need not be such a serious drawback, though, since there is a well-known theory of rights that can vindicate intuitions such as those motivating their arguments. This is the interest theory of rights. On the interest theory, A has a right against B if, and only if, A's interest (i.e., a dimension of her wellbeing) is weighty enough to place a duty on B (Raz 1986, chap. 7). The weight of an interest, in turn, is measured in relation to how costly it would be for B to attend to it. Moderate perfectly fits the interest-theory model, and may give Barry and Øverland a principled explanation for why duties of assistance in easy rescue cases are correlative to enforceable rights.
Is this principled explanation satisfactory? Is it plausible to suggest, as the interest theory does, that whenever B is required to do something (X) for A's sake, A has a right to X against B? The answer is affirmative, and trivially so, if one defines 'a right' in the way the interest theorist does: as a duty grounded in someone's interests. For the interest theorist, the definition and grounding of rights coincide, because a right simply is a duty grounded in a particular way (Preda 2015). But note that, in the foregoing, a right has been defined as more than a simple duty. A right is a duty owed to another. Moreover, we have been speaking of enforceable rights specifically. And once we focus on enforceable rights so understood, it becomes less obvious whether an interest-based grounding suffices to vindicate them. 13 The following scenario suggests that it does not.
Sam and the exam: Sam has an important exam tomorrow. Passing the exam will open up a world of possibilities for him. Failing the exam will predictably throw Sam into severe depression and destroy his chances of finding a decent job. Tonya classmate with whom Sam never really spoke or struck a friendshipis aware of the situation and knows that he is in a unique position to help Sam. He is top of the class in the relevant subject and has unparalleled pedagogical skills. An hour of tuition from Tony would virtually guarantee Sam's success in the exam. As it happens, Tony has nothing better to do that afternoon.
This case is, of course, unrealistic for many reasonsespecially the level of knowledge of the parties involved. But, taken at face value, it seems to me clear that Tony would have an 13 For critiques of the interest theory, see Hart (1955); Kamm (2002). obligation to help Sam. 14 The cost of helping Sam is objectively (and, let us assume, subjectively) minimal, and the gain to Sam huge. In fact, this is an instance of someone being able to prevent something very bad from happening at very little cost to themselves. But while I am prepared to accept that Tony has a duty to help Sam, I am also convinced that Sam has no enforceable right to be assisted by Tony. Sam may politely ask for Tony's assistance, but I do not think that he may demand it, let alone force Tony to provide it. Tony does not owe it to Sam to help him, although he ought to do so.
If this is correct, here we have a case in which the conditions for enforceable-rights generation set out by Moderate (and the interest theory) are met, yet no enforceable right appears to be in play. This is a problem to the extent that the Barry and Øverland's defence of Moderate entirely rests on its ability to fit intuitive cases. Absent a broader theoretical rationale telling us why we should accept Moderate and revise our judgements in a case like Sam and the Exam, Barry and Øverland's case for Moderate is once again incomplete.

Conclusion: balancing theory and intuitions
In my critical remarks, I have tried to show that Barry and Øverland's argumentstrategy in support of Moderate exhibits some lacunae. The lacunae, it seems to me, stem from Barry and Øverland's excessive reliance on intuitions in making their case. While I agree with them that an important virtue of a general principle is its ability to 'fit the evidence'and, in moral theorizing, the evidence is provided by intuitive judgements 15 -'fit' is not enough. This is for two reasons, one specific to moral theorizing, the other to theorizing generally.
First, intuitions are an 'odd' kind of evidence: much less reliable and much less understood than empirical evidence from, say, the natural sciences. (Barry and Øverland, I should say, are fully aware of this.) This is also why, in moral theorizing, we often consider 'biting the bullet' an acceptable practice. 'Biting the bullet' is a term of art for going against some evidence, by defending principles whose implications clash with some of our intuitions. Fit with evidence, then, has weight in moral theory, but not absolute weight. Since our intuitions are not so reliable, general principles are not automatically invalidated for being unable to account for some of them. 16 Second, a good moral principle must not only sufficiently fit the evidence, but also explain it. It is not enough to note that, every time we can help someone at moderate cost, we think we are obligated to do so. We also need an account of where that obligation comes from, of what grounds or explains it. And if the obligation is correlative to rights and enforceable, we need an explanation for these further moral phenomena, too.
My impression is that Barry and Øverland's defence of Moderate is not fully sensitive to these broader considerations. This, in turn, makes it hard for Barry and Øverland to tell us how the cost-parameter within their framework should be filled. And without an 14 It may be objected that our intuitions in this case are unreliable because it is not one we are likely to experience in the real world. This objection, however, is not one Barry and Øverland could invoke to rebut my critique, since such stylized, somewhat unlikely cases are central to their methodology (e.g., Bob's internet banking). 15 Though, of course, there is a further question about what kinds of intuitive judgments can be trusted. 16 See the discussion of 'reflective equilibrium' in Rawls's A Theory of Justice.
account of how that parameter should be filled, their framework remains underspecified. I have then suggested that Barry and Øverland's claim that duties to assist (at least in one-on-one cases) are correlative to enforceable rights may indirectly provide an answer to the 'cost measurement' question. But that claim, too, stands in need of explanation. Barry and Øverland, however, do not offer a systematic defence of that claim, beyond reliance on intuitive judgements. The obvious move of invoking the interest theory of rights for explanatory purposes, I have shown, won't do. Furthermore, contrary to first appearances, not all verdicts issued by Moderate are in fact consistent with our intuitions about people's rights to assistance in one-on-one scenarios. Once again, then, intuitions alone cannot provide conclusive support for Moderate.
Nothing in my discussion shows that Moderate is to be rejected. As I said, it has much intuitive appeal, and Barry and Øverland's treatment of it is remarkably clear and insightful. Still, my discussion does suggest that more work needs to be done to defend Moderate. In particular, a full defence of it may need to rely on theoretical considerations that are independent of the particular intuitions we are trying to account for: the kinds of considerations we appeal to when we 'widen' our reflective equilibrium. The present discussion, then, may be seen as an invitation to explore this 'wider' approach.