Justice and humanity, actions and omissions

Pogge argues that global poverty is an injustice perpetuated by the rich against the poor, while Campbell argues that global poverty is inhumane, in that the rich fail to alleviate suffering where it is both severe and salient.Footnote 4

Campbell states that he takes justice, as the basis of cosmopolitan duties, to be a moral value roughly akin to fairness, desert and merit, and he takes humanity, as an alternative basis, to be a moral value roughly akin to benevolence, altruism and caring.Footnote 5 ‘Humanity’ as it appears here should not be confused with the more general justification for cosmopolitan theories, which is that we have obligations to one another in virtue of our shared humanity. On the one hand, the kind of humanity at stake in positions like Campbell's (or Singer's) is humanitarian, about obligations generated by the existence of suffering. Pogge, on the other hand, argues for global poverty relief by appealing to a conception of justice. His argument is two fold. First of all, he argues that almost everyone is culpable for the gross injustice of the current world order. The way that people act, and the way they benefit from unjust institutions, make them culpable. From this culpability, he argues, there derive duties of reparation or compensation.Footnote 6 In Pogge's view, global poverty is an injustice, caused and perpetuated by those better off. Those made worse off by their actions are entitled to remedy.

Campbell argued against Pogge that it is not those who are culpable who should be targeted as a means to ending global poverty.Footnote 7 It is, rather, those who have the means to solve the problem.Footnote 8 Pogge is lucky, because those who are culpable and those who can provide aid coincide, namely, the citizens in better-off countries. But that coincidence is contingent, rather than necessary. If those in poor countries were fully responsible for their own situation, then there would be no duty to end global poverty on the justice story.Footnote 9 Campbell argues that Pogge's theory fails to account for poverty as a result of natural disasters, or in fact any poverty that is not the result of the culpable conduct of others. A better ground, Campbell contends, is the moral principle of humanity, which holds that there is a straightforward correlation between the existence of suffering and our moral duty to end or minimize it.Footnote 10 Everywhere that there is suffering, such as that caused by the existence of poverty, there will be a duty to minimize it – and that is likely to cover more cases overall than appeal to justice and the culpable conduct of others can do.Footnote 11

From the existence of suffering in the world, the principle of humanity derives a duty to provide aid. A failure to provide such aid in spite of suffering is, then, a failure of required action; we omit to do something that we have a duty to do – or we omitted in the past to do something that we ought to have done. Inversely, in Pogge's formulation the suffering caused by global poverty continues to exist because of our culpable conduct. We have a duty to stop doing whatever actions of ours are conceivably causing or contributing to the unjust world order that perpetuates such suffering. Failure to stop doing so is a failure of the duty to refrain from harming. We act in a way we ought not to. The basic claim I am making is that the injustice argued to be implicit in global poverty involves action, while the inhumanity argued to be implicit in global poverty involves omission.Footnote 12 The reason for making this claim is simply that many experiments have shown subjects to have a general bias towards omissions causing harms, which is to say, subjects usually judge actions that cause harms to be much worse, morally speaking, than omissions that cause equivalent harms. If this bias is robust, we can expect people to judge the harm claimed by cosmopolitan arguments from justice to be worse than the equivalent harm claimed by cosmopolitan arguments from humanity. And if, as I shall soon discuss, people are more motivated to compensate for morally worse harms, then we can expect arguments from justice to be more motivating overall, and thereby more feasible as a means for alleviating the effects of global poverty.

But first, what exactly is an omission? There are several available formulations.Footnote 13 Jari Talja gives the following definition:

1. (i) it is not the case that S performs a,

2. (ii) S has the ability of performing a,

3. (iii) S has an opportunity to perform a,

4. (iv) S has a reason to perform a.Footnote 14

Which is to say, when an agent has the ability to perform a particular action, the opportunity to perform it, a reason to perform it, and yet does not perform it, she has omitted to perform the particular action. This suffices to distinguish omissions from non-actions. The non-actions of an agent are potentially infinite; there are many, many things that I have no opportunity to do, no ability to do and no reason to do (and so of course, do not do). But non-actions are not our concern here.

Of course, there are difficult cases where it is hard to say whether something is an action or an omission (the standard case is killing versus letting die,Footnote 15 but there are many). Talja uses the example of the Queen's Guard, whose job is to remain perfectly still. Is ‘remaining still’ an action, or the omission of moving? If a guard moves, that is an action, but it involves an omission, omitting to stay still.Footnote 16 Despite the difficult cases, there are many where it is quite clear what counts as an action and what counts as an omission. When I give a homeless person €10 I am acting, and when I choose not to contribute to his well-being in this way I am omitting to do something I could (easily) do. When I jump into the lake to save one of Singer's ubiquitous drowning babies I am acting, and when I decide to walk on by (perhaps I believe I have done my fair share of saving drowning babies this year) I am omitting to do something I could do. And so on.

I take duties of justice (to refrain from causing harm) and duties of humanity (to provide aid to those who are suffering) to be cases that are often clear in the same way. The injustice argued to be implicit in global poverty comes from people's harming others, or making them worse-off, in spite of negative duties to refrain from doing so. The inhumanity argued to be implicit in global poverty comes from people's failure to provide aid to others, or make them better-off, despite positive duties to do so. In the next section I want to turn to some empirical literature on the actions/omissions distinction, in order to determine whether there is a presumptive case in favour of either inhumanity or injustice as better motivators in a cosmopolitan argument. Which of actions and omissions are taken more seriously by the general population? A failure in which category, if internalized by agents as a genuine failure, would be more likely to motivate them to take the required kind of compensatory action, or support their governments in taking it?

I have already mentioned, in outlining the distinction between actions and omissions and mapping them to the justice and humanity positions, respectively, that people in general have an omission bias, in that they think omissions causing harms are less bad than actions causing equivalent harms.Footnote 17 And if it is true that people in general have a bias towards harm by omission, then it should be true that they will be easier to motivate (pending establishment of a link between guilt and remedy) if they can be brought to accept that they themselves, by acting, are harming the global poor, rather than merely omitting to help the poor when they could. I am asking about how individual citizens are likely to respond to alleged harms by action, in comparison to alleged harms by omission.

Jonathan Baron and Ilana Ritov show that subjects have a strong omission bias in the case of vaccinations.Footnote 18 They suggest that the preference for harm by omission over harm by action actually tracks the more fundamental distinction between indirect and direct action. They conjecture that ‘avoid direct harm’ is a kind of cognitive heuristic easily instilled in early human development.Footnote 19 The authors note that the omission bias is stronger and weaker across different subjects, and that there is some evidence of its being counteracted by ‘action bias’, a newer cognitive heuristic that attempts to mitigate against ‘doing nothing’ (they give the example of the 1960s’ slogan ‘if you aren't part of the solution, you're part of the problem!)’Footnote 20 Evidence of a robust omission bias shows up in many other places.Footnote 21

One of the moral psychology research laboratories in the United States has run a large-scale online study directed at testing the availability of three different principles to conscious moral reasoning.Footnote 22 The principle that actions causing harm are morally worse than omissions causing equivalent harm was one of the three principles tested. Scenarios directed at testing that principle were coupled together, the only thing that made the difference being whether the scenario featured harm as a result of an act, or an omission. For example, one set of scenarios was the classic trolley problem. In the first scenario of the pair, respondents were asked to consider a person's throwing a fat man onto some railway tracks in order to stop the death of three people; in the second scenario of the pair, respondents were asked to consider a person's failing to pull a lever that would stop a platform from opening up and dropping a fat man onto the tracks below, preventing the death of three people.Footnote 23 Notice that in these cases the harms are not overdetermined; the harm (the fat man's death) is counterfactually dependent on the person's action (throwing the fat man) or omission (failing to pull the lever).

The researchers found that ‘across scenarios with different content, subjects judged action as worse than omission’. Subjects ‘readily provided sufficient justifications for the action principle’.Footnote 24 This last point is significant. Another principle being tested by the research laboratory was the ‘contact principle’, the principle that harms involving direct contact between individuals are morally worse than harms not involving direct contact (think fistfight v. gunfight). Usually, once participants realized they were appealing to this principle, they noticed that it was fallacious and reconsidered their previous judgements. Not so with the ‘action principle’ – 95 per cent of all sufficient justifications cited the target principle, i.e. the principle that actions causing harm are morally worse than omissions with equivalent consequences.Footnote 25

Vindicating the omission bias?

We have just seen that subjects show an omission bias. That makes it at least a prima facie constraint upon what we can achieve in a society. But now we have to ask a further feasibility question. When we ask what the activists can achieve, we usually take the rest of society (whom they have to convince) as pretty much fixed. But should we simply take for granted that part of what characterizes citizens is a bias towards harms by omission? The alternative is to focus on what is diachronically feasible. We could try to change the omission bias, which might make the humanity argument a more feasible means later on. Perhaps we could try to convince people that the omission bias is an irrational bias to have. But is it?

Talja distinguishes omissions from refrainings. In his view, refrainings require a further condition, namely that they are intended. So all cases of intended omissions are refrainings, but there can be, he thinks, omissions that are not intended. Johanna Cordes-de Waal gives evidence to support the idea that the reason many subjects show an omission bias is that in many cases they perceive omissions as non-decisions.Footnote 26 This supports Talja's distinction between unintended omissions and intended omissions (refrainings).Footnote 27 One explanation of a failure to act in spite of an ability, an opportunity and a reason, is that the agent deliberates about whether to act right up until the time that she no longer has the opportunity. For example, a subject seeing a drowning child might think for so long about the benefits to the child of being rescued, versus the costs to herself of doing the rescuing, that she fails to decide whether to act or not, and the child drowns. It is widely believed that intended harms are worse than unintended harms, and that people are more responsible for the former than the latter. Most legal systems distinguish between the two, and treat intended harms more harshly, as evidenced by the distinction between murder and manslaughter. If all omissions were non-decisions, were unintended, that would seem to vindicate the omission bias.

Cordes-de Waal argues that the omission bias is underwritten by judgements of causality, with intending as an intermediary state. Where causation is not supported by intending – for example, when an agent causes an outcome accidentally or negligently – people do not show an omission bias, and neither when there is intending without causation, such as when an agent intends an outcome but for some reason fails to bring it about.Footnote 28 People in general seem to believe that if an agent wants a certain outcome to obtain, he can make it happen, i.e. that there is a reliable correlation between intending and causing. Of course, this is not true as a conceptual matter. We can accidentally bring about outcomes that we did not intend, so causality is not a foolproof indicator of intending. As mentioned above, people generally agree that if an act is not intended then a person should not be held responsible for it, at least not to the same degree as they should be for an act that is intended, and perhaps not at all. If it is true that the situation of the global poor is not intended in the humanity formulation, and yet intended in the justice formulation, then, given the omission bias, people should be thought less responsible (or not responsible at all) in the former and more responsible (or wholly responsible) in the latter.

If there is some irrationality involved in the omission bias, it might be that harms by omission are not generally unintended, compared to harms by action.Footnote 29 Carolina Sartorio argues, to vindicate the omission bias in at least a subset of cases, that when an outcome would have happened anyway (for example, in cases where there is a back-up actor in place ready to do what you might fail to do), omissions producing harm really are less morally bad than actions producing equivalent harm.Footnote 30 She gives the case of an assassin firing at a victim when there is a bulletproof screen in the way. A spectator has the option to remove the screen and thus make it the case that the assassin's bullet succeeds in hitting the victim. But if the spectator does not remove the screen, the assassin's assistant will remove the screen instead. Given that the removal of the screen is going to happen anyway, the question is whether failing to remove the screen is just as bad as removing the screen. In the former case the spectator stands by as the assassin's assistant removes the screen and the assassin shoots the victim; in the latter case the spectator removes the screen and aids the assassin in killing the victim.

Sartorio's argument is that harm by omission in this kind of case is importantly different from harm by action, because harm by omission does not create a new threat to the victim, or facilitate an old threat. When the spectator fails to remove the screen she is simply standing by as an old threat (of the assassin and his assistant) are realized, but when she removes the screen she creates a new threat (that the screen is removed by her. Imagine that the shot missed and then she closed the screen; it might be that the assistant removes it and the assassin tries again). This difference between creating a new threat to a victim or not is what underwrites the acceptable bias between actions and omissions, on Sartorio's account. It is a genuine causal asymmetry, which tracks through to a genuine moral asymmetry.Footnote 31 The problem with this account is that although it vindicates some cases of omission bias, it does not vindicate them all. Cases of harms caused by omission where the harm would have occurred anyway are rare, and only a small subset of those harms caused by omission in general. The omission bias persists in cases where the outcome occurs only as a result of a particular agent's omission, for example, where the outcome is counterfactually dependent on the agent.

I have surveyed several possible vindications of the omission bias. These include its being justified in cases where the harm in question does not create a new threat to the relevant agent(s), where the harms are not intended, where the harms result from a non-decision (although, of course, we should not accept all cases of non-decision as justifications of an omission bias, because for some non-decisions an agent might know full well that the time for a decision has passed, practically guaranteeing the harm). The question, then, is whether the omission involved in failing to assist the global poor is one that we would be justified in considering to be less morally bad, compared against the actions involved in continued contribution to the plight of those same poor. It is hard to see how it could be. Whether or not aid goes to the poor depends counterfactually on me; I ‘create a new threat’ to the relevant agents by failing to provide aid when I could.Footnote 32 And there is enough public awareness about the situation of the world's poor that it would be hard to argue that a failure to do anything to address it is not intended.

It is plausible, however, to view the omissions involved in the humanity arguments as non-decisions, where the decision process begins but reaches no conclusion, so that the omissions are not strictly intended. Figuring out one's charitable duties is hugely cognitively demanding (what would they be if everybody did their share? What are they given that not everyone will do their share? What would be enough to make a difference? Should I give to mainstream charities, or smaller charities that seem to be struggling? And so on). Robert Goodin argues that because figuring out one's charitable duties is so cognitively demanding, and because the demandingness can create an impasse where people do nothing because they cannot come to a final decision, charitable donations of all stripes should be institutionalized as part of the general tax citizens pay.Footnote 33 (A further complication is a case in which omissions are caused by a genuine failure to even consider that one might have an obligation to act, so there is not only no decision but no decision process. Talja would categorize these as ‘non-actions’ rather than omissions, because they do not meet the necessary conditions of omissions,Footnote 34 which would put them outside the scope of this article). So, to the extent that the omissions involved in the humanity arguments are non-decisions, it is possible to argue that the omissions bias is justified. To the extent that they are not plausibly viewed as non-decisions, the omissions bias cannot be shown to be justified.

Where the bias is justified, it would be difficult to talk people out of it, and undesirable in any case (because intellectually dishonest). Where it is not justified, we might try to talk people out of it. But at this point a practical question arises: if what we care about is alleviating global poverty, why bother with a first step of changing people's minds about a bias they happen to have in some cases, and then using the humanity argument to convince them? Why not accept the bias and jump right in with immediately using the justice argument? It seems more practical to work around this psychological constraint, i.e. to accept that people in general demonstrate an omission bias, which may or may not be justified depending on the particular case, and to use the argument most likely to be efficacious in light of that constraint.

But is that not intellectually dishonest? We accept a bias that is in all likelihood irrational in some cases, and capitalize on it in order to convince people to do something we want them to do. In response, I think there is not enough morally at stake in the fact of people's having a sometimes-unjustified bias towards harms by omission for the interim step of dissolving that bias to be worth our effort. In pursuing the good, truth is a pro tanto constraint, but it can be trumped by practical considerations, as I think it is in this case. It just does not matter much if people think that harms by omission are worse than equivalent harms by action in the subset of cases in which they really are not. So we should reject in this instance the opportunity to expand the set of feasible alternatives by engaging the first of a series of steps intended to change the background conditions with which we must contend. On balance, we can achieve more by accepting the omission bias and using the justice argument than we can by attempting to undermine the omissions bias (especially considering the fine distinctions involved in showing exactly where it is justified and where it is not) and using the humanity argument.

To wrap up this section, Campbell's argument to end global poverty claims that people are causing harm by omitting to fulfil their duties to alleviate suffering. Pogge's argument to end global poverty claims that people are causing harm by acting in ways that violate their duties to refrain from contributing to an unjust world order. Experiments have shown that people generally take actions causing harm to be morally worse than omissions causing equivalent harm. Therefore, if people accept the premises of both Pogge's and Campbell's arguments, they are likely to see Pogge's as implicating them in the more grievous transgression.Footnote 35

The question of this article is to what extent it is feasible for a group of dedicated activists to use either the justice or the humanity arguments to gather support for policy reform related to the global poverty problem. We have seen that if they were to use the humanity argument, people would probably be less responsive, because of their bias towards harms by omission as being less morally egregious. This gives us a prima facie reason to go with the alternative, Pogge's justice argument. The question that remains to be answered is whether people who take themselves to be implicated in a more grievous transgression are more likely to be motivated to provide remedial or compensatory action, and to stop in whatever actions they are currently doing that add to the alleged harm. In the next section, I shall explore the connection between guilt and shame, and remedial action. (There is in fact a further connection, that between accepting that one has transgressed and feeling guilt and shame, but I will assume that in normal people evolved pro-social emotions take care of that step in a more or less automated way).

Moral guilt and its alleviation through action

Much research has shown that guilt, under the right conditions, is conducive to remedial action. Franklin Boster and his students, for example, created an experiment in which subjects, instructed to complete a questionnaire, were left in a room with a student about to sit an examination. After instructing both the subject and the student of their respective tasks, the experimenter leaves the room, commenting to the student that she supposes she can trust him given that there is another person in the room (the subject). In one condition the student cheats, and is caught by the experimenter, who reprimands the subject for not having come to tell her. Later, the subject is offered the chance to participate in a further experiment, and told explicitly that there would not be extra course-credit, but that it would be a chance to feel good about helping the experimenters. The authors found that subjects made to feel guilty about their complicity in the student's cheating were much more likely to comply with the request to participate in the further experiment, when it was made explicit that they would feel good about themselves by doing so. Their idea was that people are motivated to expel or lessen feelings of guilt, and they seize the opportunity to do so when it is presented.Footnote 36 Many studies show that the more guilty a person feels the more likely compliance is,Footnote 37 and that people experiencing guilt usually desire to make reparations, and to avoid future actions that might induce further feelings of guilt.Footnote 38 But does guilt always lead to increased compliance or reparation? If not, what can go wrong?

R. S Lazarus argues that guilt conduces to remedial action only when people believe that they have control over the given situation, and believe that there is an action available to them that will make their feelings of guilt lessen or disappear.Footnote 39 Imagine that a male student has stolen a valuable book from the office of one of his university professors. If he believes he has control over the outcome (it is not the case that he has already been caught and is awaiting the disciplinary committee), and that there is at least one action he could undertake to alleviate his guilt (buy the professor a new book, return the book in secret, simply apologize and explain what motivated the theft in the hope that the professor will understand), then he is likely to engage in one of the actions likely to reduce his feelings of guilt. Under the right conditions, guilty-feeling people are likely to engage in guilt-reducing behaviours.

But things can go wrong such that one or more of these conditions is not met. For instance, it might be that a situation is out of the guilty person's control. If the student awaits the disciplinary committee, then his control over the situation is lost. Likewise, it might be that there are no available remedial actions, or at least none that would suffice to alleviate his feelings of guilt. It might be that there is simply nothing relevant he could do. Imagine the professor has noticed the stolen book and no longer leaves students alone in his office, even though he has no idea which of his students is the culprit. Or it might be that there is something he could do, e.g. persuade the professor's secretary to smuggle the book back into the office, but it would not make him feel any less guilty for having stolen the book in the first place.

Furthermore, things can go wrong even when these conditions are met. Lisa Lindsey argues that both actual guilt and merely anticipated guilt trigger the intention to remedy, which in turn triggers remedial action, but only when the guilt comes from a source perceived as honest. Her experiments take bone-marrow donations as their subject matter. She found that subjects in general perceived the information they were given about blood-borne cancers and the bone-marrow treatments that could cure them as honest, in which case the triggers from guilt or anticipated guilt to intention and from intention to action (registering to be a donor, or seeking more information about it) worked as predicted. But she warns caution, saying that if a guilt-producing message is perceived to be manipulative or dishonest, it can create what psychologists call ‘reactive attitudes’, in which case subjects are likely to act counter to the recommendations of the message.Footnote 40 One interesting feature of Lindsey's experiments is that she uses a ‘naturalistic’ condition, in which subjects are given the straight facts about the number of people dying from diseases curable with bone-marrow transplants, and a high anticipated guilt condition, in which subjects are given the same facts as in the naturalistic condition with the addition of several stories about diseased individuals who either died, failing to secure the necessary transplant, or lived, having secured the necessary transplant. Although Lindsey's intention was to compare ‘normal’ guilt-inducing information with information explicitly designed to produce more guilt (and, therefore, hopefully more action), what she actually does is remind us of the fact that people respond better to human stories than they do to abstract statistics.Footnote 41

Gian Vittoria Caprara and his colleagues posit a model in which negative affect, including feelings of guilt, produce the desire for reparation only in some people. In others, it produces instead a fear of punishment. The authors argue that if a subject focuses on fear of punishment he can be more prone to anger and anxiety, and ultimately aggression, while if a subject focuses on need for reparation he is more prone to pro-social behaviour. Their idea is that different people are disposed to respond to negative affect in different ways. The authors ran a cross-cultural experiment on pre-adolescents from Italy, Hungary and the Czech Republic, and found, among other things, that female subjects were more likely to be pushed by negative affect into a need for reparation, while male subjects were more likely to be pushed by negative affect towards a fear of punishment and subsequent aggression. The authors’ main aim is to stress that while much recent literature has concentrated on the positive effects of guilt in leading to remedial action, there is a darker side. They remind us that guilt plays a large role in a variety of psychological disorders, especially in depression, and that without an effective means to offset-guilt persons will probably not try to offset it at all.Footnote 42

This last point is very important for the global poverty discussion. It is all very well to say that a view like Campbell's, combined with the omissions bias, alleges less harm and therefore stands to create less guilt than a view like Pogge's, which combined with the omissions bias alleges greater harm and, therefore, stands to create more guilt. What the empirical data discussed above show is that if the humanity argument gave a simple means of making reparation for the alleged harms, while the justice argument gave only very difficult means of reparation, we might do better overall to use the humanity argument. Greater guilt might mean greater remedial action under the right conditions, but part of creating the right conditions is enabling people to act in ways that will offset their guilt. So a crucial condition of an argument wanting to impact on global poverty is to give people a way to alleviate their guilt, such as by recommending remedial duties that agents will see as within their capacity to fulfil. Fortunately, both Campbell and Pogge make recommendations about remedial duties, mostly centring around donations to some variety of a global relief fund.Footnote 43 The lesson to take from the literature on guilt and motivation, then, is that when it comes to marketing the arguments alleging guilt in global poverty, more guilt is better, but we must be careful to give people ways to alleviate the guilt. We must be sensitive to the fact that making people feel guilty might result in fear and aggression as much as in reparation, depending on the disposition of the subject in question. We must also be careful that we have the right set of arguments – that they will produce guilt and not some other emotion. Guilt inspires remedial action, even if the path from the one to the other can be disrupted under certain conditions.Footnote 44

What I have tried to establish is that were a group of activists to begin trying to gather the support necessary to institutional political reform (let us say that the final objective is to have enough support such that a certain non-trivial percentage of the country's gross domestic product would go to foreign aid specifically for poverty relief), they would do better by engaging the justice argument rather than the humanity argument. This is because, I have argued, the justice argument alleges a harm by action, which people have been demonstrated to find more morally egregious than the alternative argument's allegation of a harm by omission, and because people's accepting implication in a more morally serious harm is likely to inspire the motivation to remedy or compensate that harm, so long as such a remedial action is available.Footnote 45 Using the justice argument makes global poverty relief more feasible than using the humanity argument would; in a pairwise comparative feasibility assessment, the justice argument comes out on top. In the next section I will consider a few objections to this argument.

Objections

One objection to the argument presented above is that dealing with the debate over justice and humanity by means of the distinction between actions and omissions is too indirect. Why bother with actions and omissions? Why not just find out what people think about the moral gravity of injustice as compared with the moral gravity of inhumanity? If people think injustice is morally worse than inhumanity, then the same conclusions follow, but we get to them directly. The response is that it is probably only moral philosophers who have clear ideas about the distinction between injustice and inhumanity, and maybe not even them. And moral philosophers are only a small proportion of the relevant population – the big question is what is likely to motivate everyone, or at least a majority, into the kind of support for global poverty relief necessary to mobilize governments. Furthermore, even if people did have clear ideas about injustice compared with inhumanity, there might be scope worries. Even if they think injustice is worse, they might not think that justice extends beyond geographical borders (for example, they cannot be responsible for injustices in other countries). Then intuitions would become blurred. These problems are avoided by approaching the issue via actions and omissions. People's intuitions about harms caused by actions and omissions respectively are fairly robust, and caused harms do not face the same scope worries.

Another objection might be that there is little point in discussing which of the arguments for global poverty relief to prefer if we do not care about which is true. If we are only instrumentalists about the arguments, then we might as well just use both. Under certain conditions it will be true that we should use both. It might turn out that the omission bias is not robust across cultures, so that in some places the humanity argument actually carries more sway. The investigation here has simply been to see which of the two arguments is more likely to succeed in bringing about the desired outcome, especially in developed, democratic countries (because that is where most of the empirical experimentation has been done). I have argued that people have a bias towards harms by omission, in which case, given that more guilt usually equals more compliance or reparation, we should allege harms by action for maximum effect. If there is a society likely to respond to a two-pronged attack (it's unjust and it is anti-humanitarian) then we should allege harms by action and by omission. The point was to take the academic debate between justice and humanity as bases of the duty to relieve global poverty, and weigh in on it from a different angle, by asking which is likely to be seen as a more serious harm by people in general and, therefore, which is likely to be more efficacious in inspiring remedial action.Footnote 46

In any case, it is not irrelevant which of the arguments is true. One way to think about this issue is in terms of comparing the epistemic weight of the arguments with their rhetorical weight.Footnote 47 These come apart, so we can ask about each. Notice that Pogge's argument is much more controversial than Campell's argument, or in fact any of the humanitarian positions. That there are assistance-based reasons to provide aid seems undeniable (although we can argue about their strength, and about the appropriate level of demandingness, and about an individual's fair share in a context where others are defecting). But Pogge's arguments have been attacked from multiple angles, such as being insufficiently empirically informed;Footnote 48 as claiming a strong conclusion from weak premises when in fact getting only weak from weak or strong from strong;Footnote 49 as mis labelling duties of assistance as duties of contribution;Footnote 50 as mistakenly extending a claim about our duties to our own victims as compared with innocent bystanders, to a claim about our duties to our own victims compared with other victims in urgent circumstances.Footnote 51 More generally we can ask about whether the kind of contribution Pogge alleges is really sufficient to a justice-based argument – there is no straightforward causal connection, but rather something much less tangible, such as failing to protest (or engage in other symbolic acts) against a government, which in its turn upholds certain international rules and institutions that themselves harm the poor. The argument faces a problem of indirectness (I cause harm via my role in a democracy, in that I have some power to influence the actions of my government) and of dilution (I am just one among many millions of other citizens to whose preferences the government is accountable). This difference in controversiality might lead us to judge that Campbell's argument is epistemically more weighty (which is to say, more likely true) but rhetorically less weighty (because people are more motivated to remedy contribution-based harms), while Pogge's argument is epistemically less weighty (less likely to be true, because faced with more problems of a seemingly serious nature) but rhetorically more weighty (because people take contribution-based harms more seriously). If epistemic and rhetorical considerations have equal worth, then it is just a matter of looking at which is more weighty overall – maybe they are equal and then we can just choose between them arbitrarily.Footnote 52 Notice, however, that Pogge is being discussed here as a representative of the justice-based position; even if the challenges to his own position are too great, there are numerous other justice-based cosmopolitan positions that are not plagued by the same difficulties.Footnote 53

A further objection is that the correlation I try to establish here between more guilt and more compensation is sensitive to a threshold worry. The fact that harms by action are seen as morally worse than harms by omission, and therefore poverty-as-injustice is seen as more morally grave than poverty-as-inhumanity, does not entail that the guilt involved in harm by omission is not sufficient to discharge the duties necessary to impact upon global poverty. If there is some threshold that justice-guilt reaches that humanity-guilt does not, then we are justified in preferring justice-guilt. But if both reach the required threshold, then there is no reason to prefer one over the other. One reply to this objection is to point out that people are generally aware of suffering in developing nations, and they generally fail to do anything about it.Footnote 54 So we might hope that justice-guilt stands a better chance of getting through to. But that is contingent, as mentioned already, on people accepting the premises of Pogge's argument.

That last point is the essence of the final objection, and it is related to the worry about epistemic v. rhetorical weight. We want to know about what is feasible with respect to motivating ordinary people. So we should not be asking which of the justice and humanity arguments stands a better chance of inspiring action, assuming that people accept the premises of them both. Rather we should be asking which of those arguments stands the best chance of having its premises accepted. If the arguments are good, we might think, they will motivate on their own (at least an anti-Humean, or a judgement internalist about moral motivation, will think so).Footnote 55 The big question is not which of Campbell's or Pogge's arguments would be more successful in practice, but how resistant to the alleged harms people will be in either case. The fact that Pogge's argument implicates people in a harm that is morally more serious from their own perspective might be a reason to think people will not accept its premises and, therefore, not be forced to any action that might be recommended by the conclusion. If people are not aware of the argument, or will resist it quite strongly, then the practical considerations change. For example, the activists might be faced with a choice between spending scarce resources on an effort to get people to accept the premises of contribution-type arguments, or spending them getting people to donate towards humanitarian projects. It may be that people would give more, and be more motivated, if they accepted contribution-based arguments, but that on balance, we could do more for the alleviation of global poverty by proceeding on humanitarian grounds (this is just to suggest that when we subtract the cost of making the antecedent of that conditional true from the amount raised once it is true, we would have less than if we merely raised funds as things stand, using humanity-based arguments). Similarly, in the unlikely event that people's acting on justice-based considerations would actually do negligible good compared to their acting on humanity-based considerationsFootnote 56, we might consider sticking with the latter. In that sense it might turn out that the humanity argument is a more feasible means of realizing the desired cosmopolitan outcome, because it will be much harder for people to resist the idea that when someone is suffering and they have the capacity to help, they should, at least in some situations, do so.

In response, we have some reason to be optimistic about people accepting their role in contributing to global harms. For example, the fair trade, eggs from free-range hens, free-range meat and anti-sweatshop movements are all gaining in momentum. People accept that their consumer decisions can implicate them in harms in their own, and other, countries. Thus, it is not unreasonable to think that a similar conclusion might be reached about their implication in international trade arrangements. The latter case is more complicated, because the alternatives are more ephemeral. When fair trade coffee is sold next to non-fair trade coffee, or non-sweatshop clothes are sold in a store next to sweatshop clothes, it is easy enough for informed persons to make a choice to avoid causing (and endorsing) harm. But there are no alternative international trade arrangements set up and waiting for endorsement from people and, even if there were, they are not something that one person alone could choose. Nonetheless, the long-term idea is the same; when enough people insist on fair trade, companies will be driven to provide fair trade, and when enough people insist on their governments implementing fair international trade arrangements, governments will be forced to implement those arrangements.Footnote 57 The fact that people are not resistant to allegations of causal implication in harms in general shows that we have no reason to expect resistance to Pogge's arguments (or other justice-based arguments). What remains to be seen is whether, given exposure to the premises of his argument, they will accept the particular claim that they are causally involved in their government's relation with other governments, and what they owe, in particular, as a result. If all we ask of them is to support greater international aid packages in the short term and reform of international trade arrangements in the long term, there is no reason to expect resistance on a level that would make accepting the conclusion of injustice arguments unlikely.

Note, however, that the costs of persuasion are not fully explained by the omission bias. People will sometimes not accept the conclusions of arguments for ideological reasons. We do not normally expect economic libertarians to be sensistive to arguments for the substantial redistribution of wealth based on welfare considerations (even though we might hope for it). The omission bias is just one part of the story about what motivates ordinary people.