Abstract
Individuals and institutions sometimes have morally stringent reasons to not do a given action. For example, an oil company might have morally stringent reasons to refrain from providing revenue to a genocidal regime, or an engineer might have morally stringent reasons to refrain from providing her expertise in the development of weapons of mass destruction. But in some cases, if the agent does not do the action, another actor will do it with much worse consequences. For example, the oil company might know their assets will be bought by a company with worse environmental and labor practices. Or the engineer might know her position will be filled by a more ambitious and amoral engineer. I call this the moral problem of worse actors (MPWA). MPWA gives reason, at least some of the time, to consider otherwise morally impermissible actions permissible or even obligatory. On my account, doing the action in the circumstances of MPWA remains morally objectionable even if permissible or obligatory, and this brings additional moral responsibilities and obligations to the actor. Similarly, not doing the action in the circumstances of MPWA may also bring additional (but different) moral responsibilities and obligations. Acknowledging MPWA creates considerable challenges, as many bad actors may appeal to it to justify morally objectionable action. In this paper, I develop a set of strategies for individuals and institutions to handle MPWA. This includes appeals to integrity and the proper attribution of expressive responsibility, regulatory responsibility, and compensatory responsibility. I also address a set of related concerns, including worries about incentivizing would-be bad actors, concerns about epistemic uncertainty, and the problem of mala in se exceptions.
In circumstances of significant injustice, imperfect and incomplete institutional arrangements, and non-compliance with basic moral norms, individuals and institutions sometimes have morally stringent reasons not to do a given action. For example, an oil company might have morally stringent reasons to refrain from providing revenue to a genocidal regime, or an engineer might have morally stringent reasons to refrain from providing her expertise in the development of weapons of mass destruction. But in some cases, if the agent does not do the action, another actor will do it with much worse consequences. For example, the oil company might know their assets will be bought by a company with worse environmental and labor practices. Or the engineer might know her position will be filled by an immoral, ambitious peer. I call this the moral problem of worse actors (MPWA). On my view, MPWA gives reason, at least some of the time, to consider otherwise morally impermissible actions permissible or even obligatory. However, doing the action in the stead of a worse actor remains morally objectionable even if permissible or obligatory, and this brings additional moral responsibilities and obligations to the actor. Similarly, not doing the action and allowing a worse actor to do it may also bring additional (but different) moral responsibilities and obligations.
Acknowledging MPWA creates considerable challenges, as many bad actors may appeal to it to justify morally objectionable action. In this paper, I develop a set of moral strategies for individuals and institutions to handle MPWA. These include appeals to integrity (both individual and institutional) and the proper attribution of expressive responsibility, regulatory responsibility, and compensatory responsibility. I also address a related set of worries, including incentives problems, concerns about epistemic uncertainty, and the problem of mala in se exceptions.
I: THE MORAL PROBLEM OF WORSE ACTORS
Agent A is an average or above average moral actor. A is considering whether or not to do an action ϕ which will have consequences C. Everything else being equal, A ought not ϕ. However, Agent B is a below average moral actor. B intends to ϕ if A does not. If B does the action in A's stead, consequences D will result. Consequences D are much worse than consequences C. A can easily, reliably foresee B's intention to ϕ if A does not ϕ. A can also easily, reliably foresee the badness of D in comparison to C. Only one actor may ϕ at any given time.
This is the moral problem of worse actors (MPWA). Everything else being equal, A ought not do the action. However, in light of the foreseeable actions of potential worse actors who would take A's place, A may need to revise her normative assessment, even if the action might standardly be a very morally bad thing to do. In my view, as the foreseeable moral difference between consequences C and D grows, it is increasingly morally permissible for A to do the bad action. If the moral difference between consequences C and D grows large enough, then it might be morally obligatory for A to do the bad action, even if it is a very bad thing to do. The otherwise morally impermissible may become, all things considered, morally permissible, if still objectionable, when facing MPWA. In cases in which the consequences of a worse actor performing the action are sufficiently large, the moral actions which are otherwise impermissible become obligatory, if still objectionable.
For the purposes of this paper, I will define a morally bad action as one that, in most circumstances, violates some important normative constraint. These normative constraints are based on moral reasons. In the context of MPWA, the reason or reasons generating the normative constraint must be compared against the foreseeable harm that other agents might cause. However, the moral weight of the reasons opposing undertaking the action cannot be entirely overridden or abandoned. For example, if an innocent person is violently attacked by a culpable aggressor and as a matter of necessity must use violence to defend herself, the standard moral prohibitions against the use of force are abandoned, simply because the culpable aggressor has forfeited her rights against the use of force. Morally objectionable actions, in the specific sense used in this paper, are those that may be permissible, but still act against some morally important reason. This objectionableness will ground the additional duties of an actor who finds herself facing MPWA, regardless of what she does.
MPWA arises in various areas of philosophical debate. Versions of the problem can be found in standard debates over the intend/foresee distinction, the doing/allowing distinction, broader debates between consequentialists and non-consequentialists, and perhaps most commonly in debates over the ‘dirty hands’ problem in political philosophy. Echoes can be found elsewhere, for example in political theory between realists and their opponents. Despite the frequent appearance of MPWA, there is no systematic treatment of the problem. Instead, there are two general tendencies in the literature—either MPWA is used to score points in one of the debates discussed above (rather than be suitably be addressed as a problem in its own right) or MPWA is restricted to a limited space, such as to politics, or to supreme emergencies. This is striking, as MPWA is common in moral life, for both individual and institutional actors.Footnote1 My intention here is not to resolve these other familiar debates, but rather to give an account of how to handle MPWA, assuming that both absolutist deontological and absolutist consequentialist approaches to the problem commit one to other difficult to defend views. I say the problem has to be handled because it cannot be easily solved—no matter what one says about these cases, serious tensions and difficulties remain.Footnote2 The approach I recommend not only attempts to provide the best way through these difficult cases, but also aims to reduce the frequency with which MPWA arises and to mitigate the harms that result from MPWA. It will behoove us to start with two individual examples and two institutional examples.
Captain Responsible commands a unit of soldiers and is given an order to raid a distant village that he believes is mistaken, given the lack of a grave threat coming from the village, a high probability of causalities on both sides, and a high probability of causalities among the civilian population. Captain Responsible knows that if he refuses the orders, the raid will be led by Captain Reckless. Captain Responsible calculates that if he carries out the raid, only a handful of individuals will die, but if Captain Reckless carries out the raid, hundreds will die.
Bernard Williams usefully provides our second example. George is a recent graduate with a PhD in chemistry, in bad health, and needing to provide for a family. He is offered a good job at a laboratory that pursues research into chemical and biological weapons, and has limited other prospects. George objects to these weapons, but is told that if he does not take the job, the position will be filled by someone without such moral scruples, who would likely pursue the development of harmful weapons with great zeal.Footnote3
Institutional moral agents also face the MPWA.Footnote4 Consider the following two cases. In the late 1990s and early 2000s, Talisman Oil was a minority partner in the Greater Nile Petroleum Operating Company (GNPOC) in Sudan. Human rights campaigners heavily criticized Talisman Oil for its operations in Sudan. GNPOC had poor security, environmental, and labor practices and poor community relations. More importantly, GNPOC provided substantial revenue to the Government of Sudan, who at the time was prosecuting a war against various rebel groups in South Sudan that took a heavy toll on civilians. Several million people had died in Africa's longest running civil war, and more than 5 million had been displaced. Once oil began flowing, revenues were directly used to continue the prosecution of the war.Footnote5 Under heavy pressure, Talisman sold its stake in GNPOC, but the assets were purchased by Oil and Natural Gas Company of India (ONGC). ONGC was less susceptible to pressure from human rights organizations, less likely to ensure GNPOC met best practices in security, environment, and labor, and was less likely to undertake philanthropic efforts in the community. The Canadian government, who regularly pressured the Sudanese government over human rights abuses, also subsequently lost influence, while the Indian government, which was unlikely to raise human rights issues with the Sudanese government, gained influence.
The United States is the largest arms dealer in the world.Footnote6 Many of these agreements have been signed with governments that are known human rights violators. However, suppose the United States considers suspending an arms agreement with a known rights violator. In the absence of an effective global arms control regime or rigorously enforced UN sanctions, the United States can reasonably foresee that Russia (or the UK, France, China, or any number of other arms exporters) will sell to these rights abusers. The US may therefore believe that other governments that authorize arms deals will be less likely to use their resulting influence to improve human rights, and therefore a decision to not sell arms would be worse for the citizens in the arms importing country.Footnote7
MPWA in a multi-polar world
For institutional moral agents, MPWA is exacerbated in a multi-polar world.Footnote8 In the 1980s, when activists were campaigning against the apartheid government in South Africa, they could reasonably hope that sanctions and divestment campaigns would reach a large proportion of the countries, and thus companies, that did business (or potentially would do business) in South Africa.Footnote9 But such a feat would be difficult to repeat today. In contrast, when Talisman left Sudan, their assets were quickly purchased by ONGC of India. And had an Indian company not bought their assets, any number of other potential companies would be available. Given the large number of extractive companies whose headquarters, and thus much of their regulatory framework, is in countries where human rights advocacy, especially regarding foreign policy, is highly circumscribed, it is not likely that successful divestment campaigns could be run in the many dozens of countries that could potentially invest. This makes the anti-apartheid model infeasible and inadvisable for today's campaigners. If activists could successfully impose sanctions from all of the countries that would potentially invest in a bad actor, then they need not worry about the moral problem of worse actors. For example, if Sudan divestment campaigners could actively prevent investment flowing to Sudan from China, India, Malaysia, Qatar, Kuwait, and so on, then they would be able to prevent worse actors from entering, and thus not face MPWA. But if campaigners can reasonably foresee that there is only a negligibly small possibility of successfully persuading the governments and/or companies of China, India, Malaysia, Qatar, Kuwait and so on to refrain from investing in Sudan's oil sector, then they face MPWA. In the late 1980s and early 1990s, especially following the collapse of the Soviet Union, campaigners could reasonably hope to succeed in preventing investment from most of the major centers of capital that would potentially invest in South Africa. But today campaigners must face MPWA head on: if American and European companies leave Burma, the Democratic Republic of Congo, or Equatorial Guinea, they will easily be replaced by companies from South and East Asia, the Middle East, Latin America, etc.
Setting aside relevant cases
There are several similar cases to MPWA that I hope to set aside for the purposes of this paper. However, I recognize that the problems are so closely related that whatever I have to say in relation to MPWA must be able to survive challenges presented by related problems.
The moral problem of equal actors also involves an actor A intending to ϕ, but who can foresee others will do the same action if she does not. But in this version, the other actors will not ϕ worse, they will do it with the same consequences. In the moral problem of not much better actors, A intends to ϕ, which she ought not do, but she can foresee a variety of other actors will ϕ if she does not, only they will ϕ with slightly better consequences. In all of these cases, the moral problem exists in circumstances in which there is only one action or set of actions to be done.
There is another set of cases that we must set aside for this paper. In some cases, harms aggregate. For example, in a war zone, many potential arms traders can sell arms to combatants. If it is the case that each additional sale of arms will lead to a greater proportion of the population armed, and thus more suffering and death, then this is not a case of MPWA. We also will not address cases where harm is significantly over-determined, in the sense that a very large number of actors may undertake a given action, and only one action is sufficient to bring about the expected harm. For example, if hundreds of companies are considering dumping toxic waste in a river, and only one would be sufficient to spoil it, this does not count as MPWA. MPWA only arises when the harmful actions of others can be blocked by doing the same action. It is my hope that focusing on what I take to be both a difficult and clear case, with a single actor pondering a single action but foreseeing her potential replacements and the worse consequences that would flow from their actions, I may be able to say something useful and coherent which will inform what should be said about the other cases.
Two absolutist strategies
At this point, the nature of the problem should be reasonably clear and it should be distinguished from nearby moral problems. We can now turn to two initial responses that attempt to solve the problem. MPWA is no problem whatsoever if you embrace one of two absolutist positions. An absolute deontologist may simply deny that the foreseeable consequences of others’ actions are relevant for normative assessment of one's own actions. They simply hold that leading an ill-fated military raid or developing chemical weapons or funding genocidal regimes or selling arms to dictators and autocrats is morally wrong, full stop, and one ought not do that. Furthermore, the absolute deontologist holds that the consequences of the foreseeable actions of others that may result from one's own refusal to act are not and cannot be relevant for guiding one's own action.
Alternatively, an absolute consequentialist may simply deny that there is a moral problem in the first place. They may assert that leading an ill-fated military raid or developing chemical weapons or funding genocidal regimes or selling arms to ruthless dictators are simply morally unobjectionable just in case doing the action produces the best overall consequences, as compared to other possible alternatives.
I don't claim to be able to give arguments to persuade the person who firmly holds either absolutist position. But these views, stated clearly, are subject to a reductio ad absurdum sufficient to give rise to doubts that they adequately handle the case. To make the reductio objection stick against the absolute deontologist, we simply need to make the consequences of allowing the worse actor to proceed large enough that no moral theory could plausibly deny that those consequences should bear some weight in decision making. We can stipulate, for example in the case of George and chemical weapons job, that his failure to take the job results in the development of a new chemical weapon which is then used on a large scale and results in the deaths of many hundreds of millions of people. If this is an easily, reasonably foreseeable consequence, it is a considerable mark against a moral theory that it simply deny that these foreseeable consequences should be taken into account in making a decision.Footnote10
To make the reductio objection stick against the absolute consequentialist view, simply adjust the cases in the other direction. We raise the moral badness of the act under consideration, and diminish the difference in consequences between actor A and actor B choosing to do the act. Suppose Captain Responsible is mandated to undertake a genocidal attack against unarmed, innocent villagers. Suppose a commander is considering whether or not to carry out the attack. He knows that 1,000 people, all innocent, will die if he leads the attack. But he knows that if he does not lead the attack, his second in command will lead the attack, and 1,000 people will be killed, and one other person will have his finger broken. It is absurd for the absolute consequentialist to claim that the commander is morally obligated to carry out the genocidal attack, because failure to do so will bring about worse consequences. The mere statement of the view, ‘A is morally required to commit genocide so as to prevent an additional broken finger’, should be sufficient to refute it.Footnote11
The moderate pluralist response
If the reader is persuaded that we ought to abandon both absolutist positions, then she must endorse a moderate pluralist position that assigns some weight to the morally stringent reasons for not doing a given action, and some weight to consequences that will follow if another actor is permitted to cause great harm. The moderate pluralist must hold that, at least some of the time morally reflective agents should do actions that it would otherwise be impermissible to do and go against moral reasons that retain weight in guiding normative action and evaluation. These actions are permissible if the consequences of preventing others from causing greater harm are significant enough. Doing the bad action may in fact become obligatory if the bad consequences of not acting are very great.
I propose that a distinct class of moral actions is captured in these cases—on my view, morally objectionable actions are those that one is permitted and/or obliged to do to prevent others from causing greater harm, but that retain bad-making features which arise out of moral reasons that they otherwise ought not do. The language of moral objectionableness retains two key points: first, that the action violates important normative constraints based on weighty moral reasons, even if one is permitted or obliged to do it, and second, that the actor bears additional moral responsibilities for doing this objectionable action.
The recognition of MPWA, and the granting of moral permission to sometimes do objectionable actions, creates considerable challenges. Most notably, it seems to permit would-be bad actors to justify their willingness, or even eagerness, to do bad things. In what remains of this paper, I develop several strategies for responding to MPWA.
II: HANDLING THE MORAL PROBLEM OF WORSE ACTORS
In Section II, I explore four responses that should be used to handle MPWA. These responses merely ‘handle’ the problem because they do not make it go away, and any given strategy for responding to MPWA will be unsatisfactory. I explicate the moral significance of appeals to integrity, expressive responsibility, compensatory responsibility, and regulatory responsibility, with reference to individual and institutional cases. However, I do not provide any precise formula for applying these four components of the strategy and I recognize that any suitable response to MPWA will be context-sensitive. It will matter whether the agent benefits a great deal, not at all, or suffers as a result of undertaking the action. It will matter whether she has an enduring relationship with institutions such that she can easily discharge regulatory responsibilities. It will matter whether the individuals who may be harmed by the agent are similar or different to those who would be harmed by the next in line. Relevant considerations will include the degree to which the actor receives or forgoes benefit from the bad action, the actor's intention in undertaking the bad action (that is, doing the action in order to prevent the worse actor from causing more harm, versus doing the action merely because the worse actor is present),Footnote12 and the actor's moral history with regards to similar cases. The particular recommendations for the engineer, solider, oil company, or arms trader will vary depending on the case, the actor involved, her circumstances, and so on. The strategy described below, if taken on board, is intended to accomplish three related tasks: (1) to provide normative guidance to agents facing MPWA; (2) to mitigate the moral badness of MPWA when it arises; and (3) to make MPWA occur with less frequency over time.
Integrity
In the second example above, Bernard Williams famously argues that the integrity of the worker who built weapons of mass destruction would be compromised if he followed the consequentialist's prescriptions. Williams relies on an understanding of integrity as ‘identity conferring’, a faithful allegiance to those things that most deeply identify who you are and what you are about. But it is not clear why this understanding of integrity would lead George to reject the job. On Williams’ implied conception, George's integrity is insensitive to the associational and role-based commitments George has, including to feed his family. George's integrity is also insensitive to the consequences of not abandoning his purist commitment to certain deontological beliefs about the production of WMDs. If George fails to take the job, his family suffers greatly as a result, and much worse weapons of mass destruction are developed, are we inclined to name George as a paragon of integrity? George's integrity, in my view, is better understood as his faithful commitment to all of his normatively significant commitments, and some reliability in taking those commitments into account when considering how he should act.
An appeal to integrity can help to maintain recognition of the morally stringent reasons to not do certain actions, even if these reasons are outweighed by other considerations. If an agent undertakes a bad action in circumstances of MPWA, we may still call them a person of integrity, if they do so begrudgingly, acknowledging the morally stringent reasons that would otherwise prohibit the action. To take seriously this actor's integrity, in at least some of the cases in which they face MPWA, they must not do the bad action.Footnote13 For example, if George consistently takes promotions at the company making WMD's, getting ever larger bonuses for his productivity, switches to a company that is an even worse producer of WMD's, and so on, we can say with certainty that he is not a person of integrity and not acting with any serious recognition of the moral wrongness of mass killing. Similarly, if an oil company begrudgingly continues to do business with one or two pariah regimes because their presence may benefit local populations (and appropriately discharges their expressive, compensatory, and regulatory responsibility), we may properly say that the company has institutional integrity. But if the company aggressively signs contracts with a host of rights abusers, always proclaiming that it is permissible because some unnamed competitor would be worse for the country, we can say that this company lacks institutional integrity and does not genuinely recognize the morally stringent reasons for not providing revenue to genocidal regimes. If an institution is engaged in morally undesirable conduct justified by the worse actors problem, the institution should have public, transparent procedures for identifying their most objectionable actions, and explaining how these will either be reformed or compensated in the future. Prospectively, individuals or institutions weighing whether or not to do a given action must reflect on the frequency with which they undertake such actions, and the ways in which they will avoid objectionable entanglements in the future. As I will discuss in the next section, institutional integrity also requires that exit be accompanied by appropriate condemnation of the relevant action.
Expressive responsibility
There is a strong expressive component of our moral conduct that is often overlooked. I accept (without further argument here) that individuals are morally responsible for the content of their beliefs and attitudes, and that they are responsible for expressing condemnation of moral wrongs when committed and for expressing praise of moral goods when appropriate. As Brennan and Lomasky write,
there exist norms of expression. Through one's spoken utterances and other expressive activity, one aligns oneself with certain values and opposes others. To the extent that it is inherently and not simply instrumentally good to identify—and to identify oneself with—that which is good (and evil to align oneself with that which is evil), an ethics of expression is not reducible to even a sophisticated consequentialism.Footnote14
This expressive responsibility neatly connects back to the issue of integrity. The person who reluctantly does the bad action to prevent worse consequences may publicly name the action as bad and state a preference for an alternative state of affairs. By doing so, she establishes a public standard by which her future conduct should be evaluated. The person of integrity will maintain allegiance to all of her moral commitments. After George reluctantly takes the job and tells his friends that he would rather not develop chemical weapons but needs to in order to preserve his family's welfare, they (and he) are then in a position to assess his future conduct. Does he in fact, as he said he would, take good care of his family, make slow and halting progress at work, and independently support efforts to constrain or eliminate the production and use of chemical weapons? Prospectively, agents that may do harm in MPWA cases must determine how they will voice moral recognition of the morally stringent reasons that weigh against them undertaking the action.
Regulatory responsibility
Agents facing MPWA must regulate conduct such that MPWA problems are less common and the harms that follow from their actions are mitigated. Moral agents have both internal and external regulatory responsibilities.
Internal regulatory responsibility refers to an individual's responsibility for ensuring that her own (otherwise impermissible) conduct is carried out according to the highest standard so as to achieve the best possible consequences, and that the conduct of her peers meets similarly high standards. For example, Captain Responsible must conduct the raid in such a manner so as to inflict minimal possible damage, and, in so far as possible, ensure that soldiers under his command conduct themselves in the most morally desirable manner. In professional contexts, internal regulatory responsibility may require of an individual that they take steps to deliberately undermine the harmful action. External regulatory responsibility refers to an individual or institution's responsibility to see to it that a regulatory structure is put in place such that the frequency with which MPWA occurs decreases or is eliminated.
At the individual level, if the state outlaws the production of weapons of mass destruction, then morally conscious job applicants need not consider the speed at which their competitors would produce them. At the institutional level, if the United Nations implements rule-based, effective targeted sanctions against rights abusers, oil companies need not worry about the entrance of worse actors into the oil markets of rights abusers. Absent widespread agreement among states on multilateral regulation, or international institutions with powers of enforcement over noncompliant states, smaller groups may ‘impose virtue’ on non-compliant actors. One model, recently advocated by Leif Wenar in the context of natural resource reform, is the ‘club and sanction’ model.Footnote16 On this model, a group of like-minded countries that seek to reform the international trade in natural resources can agree to a set of policy reforms, and sanction those countries that do not comply with the recommended reforms, as a way of both (1) incentivizing those countries to comply and (2) compensating harmed individuals for the harms they suffer as a result of non-compliance.Footnote17
Another alternative is to seek unilateral solutions that ‘impose virtue’ on others. For example, a single state could pass regulations that prescribe conduct not only for domestic actors but for foreign nationals who seek to interact with that state. In the United States the Security and Exchange Commission can impose a variety of regulations on companies that plan to list with the SEC. Given the size and significance of US capital markets, most foreign companies would prefer to comply with imposed regulations and remain listed with the SEC than leave US capital markets to list on a foreign exchange. This would be especially true if other large capital markets, including the UK or Japan, joined the sanctioning club.
While government regulation is often the strongest, most reliable, and most legitimate form of regulation, it is not the only game in town. Industries may come to regulate themselves through activist pressure, out of fear of impending government regulation, or out of a sincere desire to bring about reform. Here the ‘economy of esteem’Footnote18 has a role to play, as not all morally bad conduct can or should be regulated by the government. Therefore, the establishment and promotion of justified social norms, and the ‘soft’ enforcement of those norms through the economy of esteem, is a promising strategy to minimize the moral problem of worse actors.
Individuals, and especially institutions, may thus connect their expressive responsibility with regulatory responsibility for the establishment and promotion of norms and clear standards for conduct. It is very easy to adopt the arms-dealer morality if there are no standards against which to judge institutional conduct. These clear standards allow for a fair and public evaluation of whether institutional actors are in fact better than the next in line. They also chart a course for improving bad conduct. Though critics often disregard corporate efforts to establish non-enforceable norms, I find these efforts an important component of the establishment of norms which may over time become widely accepted, significantly governing conduct, and evolve into enforceable norms.Footnote19
Both the actors doing the action, and others who potentially could, bear regulatory responsibility. Of course regulatory responsibility may fall on other stakeholders as well. But the actual or potential participants in the system which may produce the harmful action bear the greatest regulatory responsibility.Footnote20
Compensatory responsibility
The final component of the strategy to handle MPWA is the proper attribution of compensatory responsibility. In circumstances of MPWA, if an actor undertakes an action that violates some stringent moral reason, they must accept a considerable share of the burden for compensating victims harmed by the action. An oil company operating in Sudan or Burma, even if morally better than the next in line, must accept a larger share of the burden for compensating the direct and indirect victims of their activities.Footnote21 George may be required to donate a portion of his income to NGOs working with the survivors of war. Captain Responsible must see to it that he or his military compensates victims of the attack.
Compensatory responsibility is much harder to attribute to the actor who forgoes doing the bad action and permits a worse actor to enter. Two standard considerations in attributing compensatory responsibility are whether a person caused harm and whether they benefited from it. Here we must distinguish between two cases. In the first, an actor is currently doing the bad action or is expected to do it. In the second, the actor has no existing or expected relationship to the bad action. In the first case, it seems plausible to attribute some (lesser) compensatory responsibility to the actor who had a determinate relationship with those who are harmed. In the second case, it is not plausible to attribute compensatory responsibility to the actor. Consider an oil company's operations in Sudan. Suppose that the active company departs for moral (not business) reasons, and is replaced by a worse actor. It is plausible to argue that the company owes some compensation to those Sudanese citizens who were harmed by their departure. Intuitively, all other oil companies, which could have entered the market and prevented a worse actor from entering, cannot plausibly be said to owe compensation to the Sudanese for their failure to enter. This is best explained by the fact that in the former case, the company has a determinate relationship with the future victims and a clear causal relationship to the next actor in line, whereas in the latter case the company is merely one of thousands that could enter the market, and has no determinate relationship with either the future victims or the next actor.Footnote22
III: OUTSTANDING CONCERNS
I hope to have persuaded the reader that MPWA is a real problem, that it is not easily dismissed, and that the four-pronged strategy sketched above is the best that can be done in difficult moral circumstances. In what remains of this paper, I address several outstanding concerns that readers might have, which either attempt to reject that MPWA is in fact a genuine problem, or reject my pluralist response.
Are there really binary choices?
One might object that MPWA, as currently stated, implies that the agent is always facing a binary choice. But, the objection holds, agents rarely if ever face such options. They can choose from a range of possible actions. The army commander leading a risky and ill-advised attack can form a coalition of like-minded soldiers to oppose the action, send word ahead to warn the villagers, arrest his second in command, and so on.
Nonetheless, sometimes agents do face binary choices, and one morally relevant consideration is the foreseeable actions of others. Either the oil company leaves Sudan or they stay. Of course there are different ways of staying (improve security, labor, and environmental practices, engage the government on human rights, expand humanitarian and community programs, and so on) and different ways of exiting (place human rights conditions on the selection of buyers of your assets, voice condemnation of government and name human rights deficits as reason for departure, call on governments to strengthen sanctions, and so on). But these options do not change the fact that there remains a simple binary choice—either they provide revenue to a genocidal government, or they do not.
The complex interactions of the many roles of moral theory
There are a number of roles that morality should play in our practical decision making and public policy formulation. Morality prescribes action. This is arguably the most important and most common function of moral reasoning: to recommend to moral agents how they ought to proceed. Morality is used to establish common standards of conduct. Moral reasoning thus has an explicit social and political role, to establish justifiable norms of conduct that govern human relationships. Morality attributes responsibility to moral agents, and thus is important in justifying punishment, heaping praise or blame on moral agents, and justifying compensatory (or cost sharing) claims.
There is a serious tension, but no clear contradiction, in holding that we may prescribe one action, for example that an oil company ought to continue to do business in Sudan (in the most responsible manner possible), while also holding that we (1) seek to establish common standards in which doing business with genocidal regimes is seriously morally objectionable, (2) can blame corporations who do business with genocidal regimes for some harms that result, (3) punish those whose operate irresponsibly (4) while also making claims for some remedial compensation against those who have or are profiting from a business relationship with a genocidal regime.
Epistemic uncertainty
At this point one might note that predicting the actions of would-be worse actors, over whom one does not have direct control, is an uncertain business. One cannot know with absolute certainty whether other bad actors will enter, or whether those alternative actors will in fact bring about considerably worse consequences. For this reason, one ought not factor such information into one's moral decision making, and thus should disregard MPWA.
It is true that we have different degrees of credence in whether others will do the same action, and whether they will do so with better or worse consequences. However, I deny that any uncertainty whatsoever about the consequences of our future action implies that no weight at all should be placed on those foreseeable consequences. We are always uncertain about the consequences of our actions. Perhaps we are a bit more uncertain about what others will do who are considering taking our place than we are about the other consequences of our actions. But such uncertainty is not paralyzing in daily decision making and ought not paralyze our moral reasoning. Our uncertainty about the future does not imply that we should make present judgments based on what things may come. It simply implies that we should consider the variety of possible states of affairs that would result from our actions, and the degree of credence we have in each of those states of affairs coming about.
Epistemic uncertainty is best handled by placing a discount rate on the significance of MPWA as our uncertainty rises. The degree to which we should discount MPWA in our moral considerations is a function of both the degree of certainty in MPWA being present, and the size of the difference in consequences that would result. Suppose the United States provides financial and military assistance to a rights abusing regime. And suppose the United States determines they ought to halt this support. With a low degree of credence (say, 40%), suppose that the US foresees a different backer will provide the same government the arms and money it desires. However, this new backer will not support the peace negotiations which are underway, which look promising at the moment. This subsequent collapse in peace negotiations may lead to many years, or even decades, of continued fighting. In this case, we may still give some weight to MPWA, because the consequences are so great, even if our credence is rather low. The lower our degree of confidence and the smaller the differential consequences between present and future actors, the less weight we should give to concerns about a worse actor entering.
The incentives problem
The mere recognition of the moral problem of worse actors might create perverse incentives for the would-be second worst actor to create the conditions such that worse actors have reason to enter a given ‘market for action’. This would be problematic if arms traders like the US, for example, can simply excuse their behavior by encouraging the presence of other market participants. It is not clear that the incentives problem is particularly severe. First, actors are already excusing their action by reference to what others would do in their stead. Recognizing this as a real problem doesn't increase the incentives for people to use the problem as an excuse. Second, by laying out a strategy for handling the problem in concert with acknowledging MPWA, we have in place a mechanism for reducing its prevalence and severity over time. This is surely worth any potential increased incentives for appealing to the excuse of worse actors.
Publicity and moral slippage
In addition to creating a market for bad actors, the public recognition of MPWA and the endorsement of the strategy outlined above may result in otherwise decent people getting a taste for bad action, weakening their important deontological convictions. If a moral system permits good people to do bad things, perhaps in the long run this would result in far more moral harm as a result of significant moral slippage than would ignoring MPWA.
First, MPWA is already given as a public reason by a wide variety of actors. This is unsurprising, as it is a very simple moral point to make. Clearly identifying MPWA and giving it systematic treatment is presumably unlikely to exacerbate the frequency with which the problem occurs. Rather, it should be the case that a clear and proper understanding of MPWA should lead to its mitigation, if people take on expressive responsibility, compensatory responsibility, and regulatory responsibility.
Second, we need not endorse the view, best known from Sidgwick, that it might be best to hide from citizens moral principles that they would like to be misuse or abuse. Transparency, dialogue, and democratic dispute are, in the long run, likely to bring us closer to moral truth and produce better overall outcomes than a dichotomous morality for the elites and the masses. However, if the public promotion of this approach to MPWA leads many people to become accustomed to doing otherwise very bad things, and over time this makes them more likely to do these bad things even when the circumstances do not call for it, or the incentives problem discussed above is realized in large numbers, then MPWA ought not be promoted. As Thomas Nagel (1972)Footnote23 writes, it is ‘particularly important not to lose confidence in our absolutist intuitions, for they are often the only barrier before the abyss of utilitarian apologetics for large-scale murder’ (176). My overall goal is to reduce the frequency with which the problem is faced and mitigate its attendant harms. If the recommended strategy fails by its own lights, then it should be abandoned. Third, in most of the cases where MPWA exists, it is the nature of the action that its possibility is already public knowledge. Acknowledging MPWA does not make it likely that the knowledge of the problem is any more likely.
A mala in se exception? Or objection?
If I am correct that sometimes we ought to do what is highly morally objectionable because the consequences of allowing others to do the same action would be worse, then this might mean that even the most horrendous sorts of actions should be carried out by morally decent people.
Here we need to carve out an exception. No serious morality can require that a decent person commit certain kinds of acts. If this is correct, then it is the case that at most some very bad action becomes morally permissible, but not obligatory, when facing MPWA. This mala in se exception then permits individual actors to avoid being morally required to do very bad things.
However, one might argue that this is an objection to the entire argument of the paper. If it cannot be obligatory to do certain morally reprehensible actions, then it must be the case that the MPWA justification for doing bad actions doesn't hold. If it doesn't hold, then we just must abandon my approach to MPWA, and find some explanation for why the foreseeable bad consequences of another doing the same action are not a weighty moral reason. I hope I can resist this move. Just as the consequential weight of others’ actions can be so great that it requires consideration in guiding one's actions, so too can the deontological weight of a bad action become so great so as to permit that one not be obliged to do it.
CONCLUSION
I feel some obligation to note that MPWA, or more accurately some variation of it, is often used to justify moral misconduct.Footnote24 It is therefore with some hesitation that I attempt to clearly identify MPWA, and give it legitimacy as a reason to persist in otherwise impermissible action. However, I hope doing so accomplishes two things. First, it helps to separate out genuine cases of MPWA from cases that appear similar but are mere excuses. Second, it helps to identify a set of actions and institutional arrangements that can make the moral best of it for agents facing MPWA—minimizing the frequency with which MPWA is genuinely faced and mitigating the harms that result when it cannot be avoided. Importantly, this shows that the actor engaged in bad action, justified by MPWA, must bear significant additional moral costs.
I have left quite a bit unresolved. I have said that sometimes moral agents should do the bad action but sometimes they should not. I have said nothing about the frequency with which this should occur, or the decision procedure that should be used when thinking about specific cases. I argued that the consequential differences between the good actor and the worse actor undertaking the action must be sufficiently large to invoke MPWA—but I have not said anything about how large these thresholds should be. Nor have I said how much compensation, regulation, or expression is sufficient in cases of MPWA. It may be that it is difficult to say anything precise about these cases. Hopefully others will be able to advance this discussion further than I have been able to take it.
ACKNOWLEDGEMENTS
Thanks to Holly Lawford-Smith, Eamon Aloyo, Luara Ferracioli, and several anonymous reviewers for very helpful written comments. Audiences at Oregon State University and Australian National University provided excellent feedback as well. All mistakes remain my own.
Notes
1 I concur with Michael Stocker who argues that the moral structure of dirty hands problems, in which conflicts between what one normally ought not do and what would bring about much better consequences, are a regular feature of moral conflict. See Michael Stocker, ‘Dirty Hands and Ordinary Life’, in Cruelty and Deception: The Controversy over Dirty Hands in Politics, ed. Paul Rynard and David P. Shugarman (Peterborough: Broadview Press, 2000), 32–3.
2 As a Holly Lawford-Smith suggested to me, dirty hands problems do not have clean solutions. If they did, they would not be dirty hands problems at all.
3 Bernard Williams, ‘Integrity’, in Utilitarianism: For and Against, ed. J. J. C. Smart and Bernard Williams (New York: Cambridge, 1973), 97–8. It is of course striking in this passage how deeply unjust gender norms informs Williams design of the case (that the mother working places great strain on the family, that the father should be the primary breadwinner, and so on). Nonetheless, the general structure of the case, considering undertaking harmful actions that would be done with greater fervor by another, remains useful for our purposes.
4 I believe it makes sense to talk of institutional moral agents, and that the duties these agents bear are not reducible to the duties of their individual members, and that this position is important for conducting foreign policy. However, nothing in this article relies on this belief. For a plausible defense of this view, see Toni Erskine, ‘Kicking Bodies and Damning Souls: The Danger of Harming “Innocent” Individuals While Punishing “Delinquent” States’. Ethics & International Affairs 24, no. 3 (2010): 261–85. For a dissenting view, see Seumas Miller, The Moral Foundations of Social Institutions: A Philosophical Study (Cambridge: Cambridge University Press, 2010).
5 And before revenues were coming in, one expects Sudan's creditors provided finance in anticipation of future payment from oil revenues. On the role of oil in Sudan, and the changes from western to non-western oil companies, see Luke Patey, ‘State Rules: Oil Companies and Armed Conflict in Sudan’. Third World Quarterly, 28 (2007): 997–1016.
6 In 2010, 44% of all arms trade agreements in developing countries were signed with the United States, followed by 24% with Russia, and 5% with China. The following year, the US increased its sales to an astonishing 79% of all agreements with developing countries. Thom Shanker, ‘US Arms Sales Make up Most of Global Market’, New York Times, August 26, 2012, http://www.nytimes.com/interactive/2012/08/26/world/middleeast/united-states-arms-sales-rise.html?ref=middleeast (accessed May 15, 2014).
7 This example is perhaps the weakest of the four. It is difficult to assess the degree to which arms sellers have influence over rights abusers. Arguably, because they are in an intimate and influential relationship, such influence exists. But in many cases it may not. Furthermore, even if the influence exists, it is not clear how much harm can be prevented by the arms seller in preventing human rights abuses. Nonetheless, most commentators believe that the sale of arms confers political influence. Consider, for example, the relationships between Russia and Syria, or China and Sudan. In both cases analysts suspect that Russia and China provide international protection to pariah regimes, in part, to maintain favorable economic ties including arms sales.
8 To a lesser extent the multi-polarity of potential actors is also present in some individual cases. For example, if George is only one of five potential people with the skills to work at the chemical manufacturer, and only he and one other person are highly skilled such that they could advance the company's projects, then he might reasonably hope to dissuade at least one more, or more optimistically four more, potential employees from working at the plant. But if there are many dozens of equally qualified applicants, and many are hard to contact or unlikely to be persuaded by George's moral argument, then he faces MPWA.
9 At the end of the campaign, nearly 200 American companies had left South Africa and over $1 billion in capital had left the country.
10 I concur with Altman and Wellman, who argue, ‘If one accepts the benefits of seriously unjust social arrangements, then one is required to contribute to efforts to eliminate (or, at least, ameliorate) those arrangements and, if feasible, one's contribution must be such that it does more to weaken the arrangements than one's acceptance of the benefits does to perpetuate them … One does have a duty to neither support nor profit from institutions that wrongly harm others, but it is a duty that does not demand moral purity. Rather, it demands a course of action that, on balance and over time, is reasonably calculated to do more to undermine the injustices with which one is complicit than to perpetuate them’. Andrew Altman and Christopher Heath Wellman, A Liberal Theory of International Justice (Oxford: Oxford University Press, 2009), 145.
11 If one finds this statement unfair, because the consequentialist might not think that broken fingers meet the threshold of counting in the consequential calculation, then just adjust the example to some case that exceeds this threshold. For example, ‘A is morally required to commit genocide to prevent one additional death’.
12 I am thankful to Gerhard Overland for raising this issue with me.
13 It is a conceptual possibility that a good person always does the bad action to prevent worse actors from doing it. Throughout their life as a moral agent, they consistently prevent bad actors from entering, but express remorse, condemn the action, psychologically regret that they must do it, internally and externally regulate conduct, and make compensation. But in the real world where people are constituted as most normal people are, there is good reason that, from the outside, we should doubt that people who consistently do the bad action are in fact giving due weight to deontological reasons and, from the inside, are constituted in such a way that they are not morally corrupted by their persistent involvement in these bad actions.
14 Loren Lomasky and Geoffrey Brennan, ‘Is there a Duty to Vote?’ Social Philosophy and Policy 17, no. 1 (2000): 80.
15 Robert Goodin discusses what I have termed expressive responsibility in the context of norm violation. He argues that, ‘When honouring social norms in the breach, therefore, you must engage in the more florid and rhetorical displays associated with civil disobedience. You acknowledge the rule that you are breaking, and … you openly acknowledge that you are indeed breaking it. You engage in lots of hand wringing … You promise faithfully to comply with the norm under other circumstances in the future, and you entreat others to do likewise. You emphasize that your action should not be taken as precedent by others’. Robert Goodin, ‘Norms Honoured in the Breach’, in Norms and Values, ed. Michael Baurmann, Geoffrey Brennan, Robert E. Goodin and Nicholas Southwood (Baden-Baden: Nomos Verlagsgesellschaft, 2010), 296.
16 This is my term, not his, for the general model in which a group of countries adopt a policy and undertake coercive measures to see that the policy is adopted in non-compliant countries. Another prominent ‘club and sanction’ proposal is Joseph Stiglitz's model for climate change reform. Stiglitz argues that countries can agree to a set of emissions targets, and countries which fail to meet those targets can be sanctioned in proportion to the degree to which they fail to meet their obligations. Such sanctions in effect impose a carbon tax on the non-compliant country (though the full tax is at least initially borne by exporting industries), thus incentivizing countries to comply, and compensating both industries who are compliant (by ensuring they are competing on a more level playing field, at least with regards to climate emissions) and generating needed finance for climate adaptation and compensation. See Joseph Stiglitz, Making Globalization Work (New York: Penguin Books, 2007), 161–86.
17 See Leif Wenar, ‘Property Rights and the Resource Curse’, Philosophy and Public Affairs 36, no. 1 (2008): 2–32.
18 Geoffrey Brennan and Philip Petit, The Economy of Esteem: An Essay on Civil and Political Society (Oxford: Oxford University Press, 2005).
19 OECD Guidelines for multinational companies, the United Nations Global Compact, company specific codes of ethics, and other efforts may be representative of this approach to norm development.
20 We need some principled reason to determine who counts as in the group of potential actors that bears regulatory responsibility. For example, if an oil company has no track record of operating in conflict affected countries, do they bear regulatory responsibility for oil extraction in those countries? I am inclined to follow Iris Marion Young's ‘social connection’ model of responsibility. For Young, even if I do not directly cause a harm, if I participate in a system in which harm is foreseeable, and share some social connection to the harmed, I bear responsibility for adjusting that system or compensating the harmed. In this case, even though the oil company does not work in conflict-affected countries, it nonetheless participates in and benefits from the overall system of international trade in oil, of which trade from bad countries is a central component. See Iris Marion Young, Responsibility for Justice (Oxford: Oxford University Press, 2011), 95–123.
21 Companies appear to implicitly accept this point, as they are at pains to highlight community development and outreach projects in countries that are plagued by human rights abuses.
22 On the relevance of the determinacy of an agent to those whom might claim compensation or assistance, see Violetta Igneski ‘Perfect and Imperfect Duties to Aid’ Social Theory and Practice 32, no. 3 (2006): 439–66.
23 Thomas Nagel, ‘War and Massacre’, Philosophy and Public Affairs 1, no. 2 (1972): 123–44.
24 Egregiously, opponents of the abolition of the slave trade in Britain argued that such a halt in the trade by the British would result in other nations increasing their trade. Perhaps interestingly, part of the response by then Prime Minister William Pitt to this argument was to note that other countries would not in fact replace the British trade, as the French were busy putting down rebellion in St. Dominique and other countries did not have the capital to increase their presence in the trade. But, this was quickly followed by this exclamation: ‘How, Sir! Is this enormous evil ever to be eradicated if every nation is thus to prudentially to wait till the concurrence of all the world shall have been obtained?’ Quoted in Adam Hochschild, Bury the Chains: Prophets and Rebels in Fight to Free and Empire's Slaves (New York: First Mariner Books, 2006), 233.