[Homepage] [Current Issue] [Past Issues]


Dirty Hands, Cosmopolitan Value and State Evil:
Reflections on Torture

 Richard Matthews


Mt. Allison University




The debate about state use of torture has surged in the last 5 years. Oren Gross, Alan Dershowitz, Michael Gross, and Fritz Alhoff are a sample of those arguing either for the legitimation of, or excusing, the use of torture by state officials in prosecuting struggles against other states or non-state actors. Although the arguments for torture vary, I intend to explore, evaluate, and reject one specific argument first made by Michael Walzer in his 1970’s paper “Political Action: The Problem of Dirty Hands”.


I will first locate this within the general ethical division we remark in the history of western philosophy. Of the deontological, utilitarian, and virtue-ethical traditions only the first seems to be unequivocally opposed to the state use of torture. Not one of the defences of torture that I have encountered relies on the deontological principle, and I think not surprisingly, for deontology offers the strongest foundation for unconditional prohibition of torture. It takes its cue from the Kantian categorical imperative, in particular the end in itself formulation. This specifies that one may never use an individual merely as a means to prudential purposes, no matter how pressing. The categorical imperative stresses the absolute worth and dignity of the individual will. Since state torture essentially requires the state agent to break the will of individuals and groups through the infliction of intense physical and psychological suffering, it is a direct attack on the most fundamental value of deontological theory.


The utilitarian and virtue ethical traditions are divided on the question of torture. Utilitarians like Fritz Alhoff unequivocally support the use of torture provided that the amount of torture employed is minimal, that there are no alternatives, and that there is a reasonable prospect of a greater good being achieved through the employment of torture then would otherwise happen. However, W.E. Twining and William Casebeer both note that careful consideration of the consequences of the use of torture in terms of unintended harms, destruction of political and social institutions, generation of an aggrieved and increasingly hostile population, and the like, could be sufficient for some utilitarians to support an absolute practical ban on torture, even if they might be willing to consider its use in principle.


I want to emphasize the following point here: on Alhoff’s analysis, there is a prior right not to be tortured, but that this may conflict with the rights others have not to be killed by acts of violence. If the balance of goods and harms shows that torture is required, then the order to torture can be given and the commander and the torturer both should then operate with a clear conscience because they have done nothing wrong. Not only have they done nothing wrong, but they have done something morally obligatory and so they should have a clean conscience having made such difficult choices. On such an analysis, torture is neither good nor evil except in terms of its consequences. But if it is good, then one is morally obligated to torture and torturing another human being would then be the right thing to do. (Alhoff 2005)


Recognizing the role of conscience is important because it is one distinguishing feature of the problem of dirty hands.  The problem of dirty hands, or the problem of tragic choices, is primarily an issue within virtue ethics. Virtue ethics is concerned with character, with the kind of person that one ought or ought not to be. Good characters have the dispositions to act well and to do good things. They choose the mean between vices and, in virtue of their practical wisdom, do not choose the bad.


At first sight one might conclude that this means that the good person could never willingly choose evil, but the issue is not so clear. According to Stephen de Wijze, the dirty hands problem is a difficulty specifically for the moral person, not for the immoral person. Dirty hands will arise only when a good person, through unfortunate circumstances, is compelled to deliberately and knowingly choose to act in an evil manner for the sake of some greater good.


So what precisely is the problem of dirty hands? 


Alan Dershowitz, while praising Amnesty International for doing its job and taking the high road, says that such a high road is not open to the government official. State officials have to make hard judgments about choices between evils and are often not in the position simply to refuse to act in an evil fashion. (Dershowitz 2004) Michael Walzer argues that it is impossible to govern innocently, and that a given action might be a moral wrong and yet considered in utilitarian terms still be the right thing to do. (Walzer 2004) To act with dirty hands you must do something wrong in order to achieve some higher good. 


More precisely, Michael Stocker defines a dirty hands situation in the following way:


An act is one of dirty hands if (i) it is right, even obligatory, (ii) but is, nonetheless somehow wrong, shameful, and the like.


And de Wijze affirms that


The political world is one characterized by what Stuart Hampshire calls ‘Experience’ (as opposed to ‘Innocence’), the ‘idea of guilty knowledge and the expectation of unavoidable squalor and imperfection, of necessary disappointments and mixed results, of half success and half failure. A person of experience has come to expect that her usual choice will be the lesser of two or more evils.


 Dirty hands reasoning involves balancing evils one against another. One acquires dirty hands by knowingly choosing an act that is evil for the sake of some desired greater good. What distinguishes dirty hands from standard utilitarian justifications of deliberate infliction of suffering is that the dirty hands act remains evil no matter the consequences. It is evil if it fails, and it is evil if it succeeds. The individual with dirty hands is culpable and guilty even though he may have acted rightly in the circumstances.


Dirty hands situations involve what De Wijze and Stocker call ‘impossible oughts’. Normally we recognize the Kantian maxim that ought implies can, that to claim that we ought to do something presupposes that it is possible for us to carry out that act and that it is possible for us to act otherwise. But the problem with dirty hands situations is that we are put into a position where we have conflicting obligations, conflicting oughts where satisfying the one obligation entails violating the other. 


The classic Kantian dilemma is worth mentioning here. Suppose you live in a racist state in which laws have been enacted requiring the arrest and imprisonment of all members of the hated race in a concentration camp. Suppose further that you are helping members of that group by hiding them in your house. A police officer comes by and asks you whether you know the location of any of that ethnic minority. How do you act? You have an unconditional obligation to obey the law. You have an unconditional obligation not to tell lies. But you also have an unconditional obligation to protect innocent individuals from harm.  You ought to do all of those things. But meeting one obligation entails violation of the others. In a situation like this, dirty hands theorists argue that there are correct ways to act, but that these correct actions require you to violate at least some of your obligations.


An appropriate description of the problem is the following: in a dirty hands situation, the official morally ought to do a and morally ought to do b, while he cannot do a as well as b.” (de Haan 2001; Morscher 2002)  There is no logical incoherence because there is no propositional conflict. The problem is rather that that there are two different cherished principles in the premise set, the satisfaction of one of which requires the violation of the other. It is a problem of incompatibility of values in specific circumstances. 


Furthermore, Stocker and de Wijze ensure the compatibility by making the principles contingent on desires. They are absolute, but they are not absolutely action-guiding. They are only action-guiding contingent on wise choice in the given circumstances. This has the unfortunate consequence of making the norms only formally absolute. That is, according to logical form they are universally quantified obligations and prohibitions. But to be action-guiding they have to be relativized to the circumstance of choice.[1]  Here the appropriateness of choice is dependent upon practical wisdom and desire.


De Wijze introduces the following analogy to illuminate the point.


Consider the desire to eat a large jam doughnut and the mutually exclusive desire to lose weight. Deciding to eat the doughnut does not abolish the desire to lose       weight. It simply overrides it and leaves one feeling guilty. There may be good reasons to eat the doughnut and in a certain set of circumstances the right thing to do, politeness to a host, for example. But because deciding to eat the doughnut is the right thing to do the desire to diet does not disappear. (De Wijze 1994)


Substitute any pair of valued moral principles, and suppose that in some given set of circumstances they become mutually exclusive, and we see the point. In that situation neither principle can be action-guiding by itself. Desires for the realization of cherished principles can conflict, and in some cases can do so in a manner where our desire for the one principle excludes the possibility of realization of the other. We think that we ought to do both, but we cannot achieve both. Hence, in that given situation, it is impossible for the norms to be action-guiding. Their action-guiding nature is contingent on our adoption of the relevant desire, here guided by circumstantial considerations and cost-benefit calculation.


The way Stocker puts the issue is that these impossible ‘oughts’ are taken into account both in judging that an action ought to be done and in evaluating the morality of the action itself. Hence they are double counted, for although they yield the conclusion that we ought to do the immoral thing for the sake of some perceived greater good, nonetheless, they are still counted as evil. This yields the consequence that the prohibited act remains evil, but that nonetheless it is an evil that we are morally obligated to commit under certain circumstances. 


In virtue of its being evil, the official who orders it, as well as any subordinate officials who carry it out, should feel ashamed at having used such means. It is appropriate that they feel sullied by having engaged in such actions. Indeed, for the dirty hands theorists, the fact that they are ashamed and sullied is the mark that they are moral. What gives us hope that they will not inappropriately choose evil is precisely their sense of their own shame at having committed evil.


What constraints obtain for dirty hands choices? To distinguish it from flatly immoral actions it can only be done for moral reasons. We incur dirty hands when we justifiably betray a person, value, or moral principle. One cannot betray the principle for selfish reasons or because it gives pleasure. One can only do the evil in order to achieve a higher good. The costs of the action have to be reasonably calculable and there has to be a reasonable chance that the choice of the immoral action will prevent the greater evil. (De Wijze 1994) Furthermore, as no general account of reasonableness can be given, the cost-benefit calculations depend on the particular circumstances of the case.


The Dirty Hands Argument As Used To Defend Torture.


How do dirty hands considerations generate an argument in favour of torture? Well, in international law torture is absolutely forbidden. Article 2 of the CAT forbids torture unconditionally. This prohibition is non-derogable even under conditions of state emergency.  Dirty hands theorists agree with the convention’s prohibition against torture. However they disagree about its non-derogability. The state official can act rightly in violating the convention against torture, provided that it is done for the sake of a higher good and has a reasonable prospect of success under the circumstances.


Suppose then you have reasonable information of a substantial terrorist threat against your state. Assume further that you have captured an individual whom you have reasonable grounds to believe has vital information about this threat. The individual is recalcitrant and refuses to divulge the information voluntarily. Time is short and normal interrogation techniques have already been tried and failed. The dilemma then is the choice between torturing the individual and failing to meet the obligation to look after public security. Meeting one obligation compels you to violate the other.


There are many good reasons for treating the above ‘ticking bomb’ dilemma as a priori incoherent, but I will not explore them here. It does at least allow the modelling of a moral dilemma.  This in turn sheds light on how the tragic choice theorist reasons in cases of deliberate evil-doing. Dirty hands theorists do not conclude from the thought experiment that all choices are equal. They agree that the choice to torture the suspect will be the right action. Consequentialist considerations are the tie-breaker here, as the suffering experienced by the torture victim will sometimes be outweighed by the consequences that will be suffered by the state the public official is obligated to defend. The dirty hands theorist does not believe that the wrong of torture is sublated by the good consequences hopefully attained through its use. It remains a wrong no matter its good consequences. Nonetheless, in the relevant circumstances it is the right thing to do and the public official is morally obligated to torture the suspect.


Why Dirty Hands Dilemmas Do Not Morally Entail Torture


I have a number of worries. The first is that there does not seem to be any place for the concept of an indefensibly wicked action. Dershowitz offers cautious support for an exceptionally wide range of actions provided they are subject to accountability and judicial control. He explicitly states his cautious heuristic support for the use of rape as a form of torture. (Dershowitz 2004) De Wijze lays out a set of 10 possible examples of dirty hands situations, and then excludes the atomic bomb attacks on Hiroshima and Nagasaki as well as the terror bombing campaigns in Europe as flatly immoral acts even if they save lives and minimize troop casualties. I think he is right to do this, but he offers no argument as to why he can infer this. Further he neglects to explore how the correctness of this argument generates an absolute opposition to torture and severely circumscribes the kinds of evils that may be deliberately done. What is there to rule out the idea that such things might be the right thing to do assuming that the situation is sufficiently grim for a given state? (or rather, assuming that the state officials believe the situation to be sufficiently grim)


Perhaps dirty hands theorists have an argument they can offer in support of the simply wicked act, but if they do I have not seen them offer it. If they do, then there will be no possible circumstances in which they can commit such acts. Further they will have to circumscribe the number of cases in which we can choose evil. We will then have to show why it is that those cases are importantly demarcated from those acts in which we can knowingly will that which is evil. The fact that such demarcations do not take place in any systematic way in discussions of tragic choice reasoning is worrying, because it raises the possibility that the dirty hands situation is far too permissive. If the conflict of values really exists as they suggest, and if the judgments are entirely based on circumstance and cost-benefit analysis, then I see no a priori reason as to why the statesman might be morally obligated to choose any evil act, in principle, should that be dictated by circumstance and cost-benefit analysis.


How do these reflections bear on torture? Well, I have raised the question about what kinds of actions are permitted. There is enormous variation in torture techniques, some more violent and destructive than others. Some of these techniques involve the torture of more than one individual. So suppose then that the only way that will work to get the suspect to reveal this information he holds will be to force the brutal rape, torture, and murder of his or her loved ones, including children. This has been a tactic employed by some interrogators in the past, and so is not a fantastic option. If you are prepared to torture the suspect to avert the ticking bomb threat, and if there are no principled reasons for rejecting torture above all else, then why would you not conclude that we should so torture the suspects’ loved ones?


Although I will not argue the point here, I would like to suggest that such an action is flatly and indefensibly evil. It is the kind of action which happens for the usual kinds of consequential state reasons that interrogators employ. But what is the difference really between that and employing the same tactics on the suspect him or herself. Part of what disgusts people about such cases is the innocence of the third parties. In cases of state torture of non-state actors, the actors are suspects and the probability is far higher than in a court situation that they are likely to be non-complicit.  In all of the cases under consideration here, the interrogators have some reason to believe the individuals have information, but then in the conditions under which such interrogations occur, the information is only ever of limited reliability. As dirty hands theorists concede, public officials have to act under conditions of often horribly inadequate information.


            Dirty hands public figures, unlike those who are simply immoral, will never choose an evil means unless they believe that there is no alternative. But having made that choice, dirty hands theorists offer nothing to clearly rule out any of the available possible evil actions except for consequentialist reasoning and the specific historical circumstances. The dirty hands public official will order that evil which he or she believes to have the greatest chance of producing the desired good consequences. In principle, he or she might then order any torture whatsoever, provided that it offers some reasonable prospect of success.


A second problem concerns the reasonableness requirement for choosing evil. Dirty Hands theorists insist that you have to maximize the chance that the choice of evil will bring about the desired good. Otherwise you are not acting morally. In the case of torture, what will be required to maximize the chances that the torture is successful in alleviating the perceived threat? Well, if you are going to torture well, then you have to be prepared in advance to do it. There is no point giving a suspect to untrained interrogators, because they will use the wrong torture methods, ask the wrong questions, and increase the risk of inadvertently either killing the prisoner or perhaps driving him insane and thereby rendering him useless for intelligence purposes. Either way, this has little reasonable chance of success.  To do evil well, one has to practice and research it.


But consider the institutional conditions that the statesman now has to establish.  You will have to conduct research on effective torture methods to maximize efficiency and ensure that counter-productive techniques are not employed. Gary Jones explicitly advocates this (Jones 1980). This means that you will have to have psychological and medical personnel performing and evaluating the research. Hence you will have to suspend or alter a crucial range of medical ethical research norms and principles. Trained medical personnel will also be necessary to evaluate torture victims to ensure that they can survive the torture, or to temporarily repair any damage caused so that the interrogations can continue.  Furthermore judicial and police ethical norms and principles will have to be altered to allow for the proper functioning of torture interrogations. Not to mention that you will have to establish torture training institutions and provide any necessary logistical backup required. If you are talking about a military or an intelligence agency, then these are far reaching provisions indeed. Not only that; since news of the torture is bound to leak out eventually, the population will have to be taught or at least encouraged to think that the use of torture is occasionally legitimate. 


If you care about the corrosive effects of torture on public institutions and civil liberties, these reflections are merely a starting point for concern and not even remotely exhaustive. Part of the reason absolutist opponents of torture are disturbed by arguments in favour of torture goes well beyond compassion for the victims of torture. It extends to recognizing the inevitable impact of torture on the character of public life itself.


But there are further consequences that should also lead us to absolutely forbid the practice. In reading the dirty hands scenarios, one gets the impression that the threat to character is incurred by the public figure who issues the order. However the evils are far worse than that. Those who institutionalize atrocities damage not only their individual character, but also those of any individuals unfortunate enough to carry out the orders. In instituting the policy, a whole range of individuals have to participate and risk becoming accustomed to these procedures. Some of them will have strong character, some of those will be simply immoral, but most are likely to be somewhere in between, the impressionable subjects who, like the subjects of Milgram’s prison experiments, will simply carry out orders and gradually become acclimatized to inflicting violence on others. In issuing orders to torture, public figures do not just damage their own souls; they also harm that of their states. What is not recognized is that the harms are inevitable. This is not a matter of identifying possible risks and taking steps to minimize them. Certain evil-doing policies inevitably wreck individual and social character. Torture is simply a paradigmatic class of such practices.


Furthermore, the use of torture historically has had a radicalizing effect on the populations which identify with the tortured individuals The French use of torture in Algeria was fuel to the fire on the Algerian War for Independence. The British use of torture in Northern Ireland in 1972-73 was a key contributor to the increase in violence during the troubles. Israeli use of ‘moderate physical pressure’ is a continuing radicalizing irritant for Palestinians that has produced no clear reduction in violence levels, and the use of torture at Guantanamo Bay, Abu Ghraib and other prisons known and hinted at has contributed to a massive increase in hostility towards the United States. What this suggests is that torture is a violence intensifier.


            A further problem is hinted at by Stocker. He remarks “I here will follow Aristotle in not discussing the possibility that one might lose one’s virtue, perhaps even become base, as a result of doing or even having to do morally required base acts.” (Stocker 1986)  This is the possibility that choice of evil means itself might destroy the character of the person making the choices. The idea is that the taking of such choices, especially if repetitive, might transform a just individual into an evil person. The problem is, once you start getting your hands dirty, how do you stop? Further if you are incapable of stopping because of the evil times in which you live, then how can you any longer say that the person has dirty hands, rather than being flatly immoral?


For those who use dirty hands considerations to defend torture, this is a crucial issue and one that cannot responsibly be avoided. For if we deliberately order torture, an act that we know to be unconditionally evil, and since state torture requires institutional policy, training, and some form of broad state support, it is not clear that public officials can preserve their good character as the tragic choice theorists envision. In an interview with Mark Danner, author of Torture and Truth: American, Abu Ghraib, and the War on Terror, Tom Engelhardt asks Danner the following question: “How does the US get up to its elbows in blood so regularly? (Englehardt 2006)” Danner’s reply is that if you persistently deceive and tell lies, then such filth is inevitable.


The problem of dirty hands generates a further difficult question: once you start getting your hands significantly dirty, and once you institutionalize the dirt, then how can you ever stop? In the context of torture, for example, tragic choices commit state officials not just to a single act of torture but to its institutionalization in public life. Not only that, it contains its own logical imperative always to torture in those situations where the circumstances dictate. Logically it is not just about one act in a unique circumstance, but a universal set of acts in possible circumstances. The combination of the institutionalization along with the generalization means that torture is neither conceptually nor empirically going to be a single act committed once. It must be repeated, and its evils must be far more extensive than the standard over-simplified cost-benefit examples suggest.  It requires utilization of torture in every possible situation in which the circumstances dictate. Additionally preparation for possible situations involves advance research and training, and regularization to maximize the likelihood of political success. If it is not a matter of the choice of a single act employed on discrete occasions, but demands policy decisions that change the shape of public institutions across the entire society, then how do you avoid the conclusion that the society gradually but consciously becomes evil.


Cosmopolitan Value Commitments As A Possible Solution


As mentioned above, De Wijze is rightfully and laudably concerned to limit the scale of violence that dirty hands reasoning can unleash, but apart from simply denying that great evils may be committed, he does not explore why these can never be integrated into a lesser evils calculus and he does not explore the criteria or values which might prohibit them. Yet if they are flatly evil, as he seems to suggest, then he must tacitly be invoking something more than contextual moral reasoning. Supplementing the dirty hands reasoning with cosmopolitan values offers one possible solution.


As its theorists typically envision it, the dirty hands position stresses the messy, murky, and confusing nature of politics. Mistakes and foul-ups are effectively inevitable. For Max Weber, for example, the region of human political affairs is deeply paradoxical and irrational. In this context, the pursuit of an ethic of love or any unconditional ethic is a catastrophe. (Weber 2004) For Weber, as well as for those like Richard Posner who follow him, a public official who pursues such an ethic is a disaster waiting to happen. Posner goes so far as to say that you should never allow such an altruist into a position of responsibility. (Posner 2004)


            Here a remark of Martha Nussbaum is valuable.  She discusses the way in which various thinkers, including Nietzsche and Bernard Williams, assert the importance of constructing politics on the recognition that the world is horrible and unintelligible. It lacks any intelligible rational structure and provides no reason to hope for progress. (Nussbaum 1997) For tragic choice theorists, to make decisions with altruistic and cosmopolitan considerations in mind is utopian and a recipe for disaster.


A cosmopolitan, for Nussbaum, is a world citizen. To be a cosmopolitan is to believe that one’s allegiances are owed to humanity as a whole (not to mention, perhaps, to other species and the environment). This means that moral consideration can never extend merely to my own children or to the members of specific communities. For a cosmopolitan there is no special prior obligation to the state; other human beings and ethnic groups count as well, and certain moral obligations to them are not trumped by considerations of community interest or national security. Ethnocentric political considerations should never trump the obligations to humanity as a whole.  The consequences of neglecting this are appalling, yet the dirty hands position is anti-cosmopolitan.


            We can see the antithesis to cosmopolitanism in the dirty hands commitment that the obligation of political figures is owed to the state. They lie under no obligations to maximize the happiness of the greatest number of sentient beings. Nor are they morally compelled to respect the autonomy of all rational beings. They are not committed to do what they can to realize a kingdom of ends or to otherwise make the world better.  These can be subordinate hopes of a public official, but should they become primary, then the public official will make decisions that are against the interests of his own state – such as refusing to institute torture policies, drop atomic bombs, or make other decisions that they might believe to be necessary to resolve a state of emergency.


Since the political world is irrational and unintelligible, there is no better world to attempt to make. Still worse, the attempt to create a better world, if carried out by politicians under conditions of emergency, will create a catastrophic state for their communities. Hence their obligations lie against cosmopolitan values. They are resolutely anti-utopian and also anti-egalitarian. The only possible egalitarianism is internal to the specific state that responsible politicians represent. There, the egalitarianism is itself an entirely contingent matter. Public officials as conceived by tragic choice theorists lie under no unconditional obligation to try to make their states recognize the humanity in its own citizens, for example. Their obligation is to the public good as they and their communities conceive it.


The absence of a categorical imperative or other suitably cosmopolitan principle is terrible, especially combined with the resulting moral parochialism. The preservation of the well-being of one’s own community is perfectly compatible with the disproportionate infliction of suffering on another community. Walzer’s defense of the use of terror bombing by the English in the earlier stages of World War II is one example. (Walzer 1977) Perhaps, although this is deeply disputed, the terror bombings were consequentially necessary for the defense of England. But in tragic choice reasoning the well-being of one’s own citizens are taken to be morally prior to that of any others. Consequently it is perfectly possible to treat the welfare of the others as less or effectively non-existent. Terror bombing and other horrific means can thereby be justified provided the state of emergency is believed to exist and provided there are no perceived alternatives for mitigating the threat. Dirty hands reasoning has also been used to justify Hiroshima and Nagasaki as well as widespread torture and murder during Argentina’s dirty war. A dirty hands argument can be and is formulated for pretty much every act of state violence.


Unlike sophisticated utilitarian or Kantian theories, and unlike cosmopolitan virtue ethical theories, the dirty hands theory excludes from consideration the costs to the populations of other states. Or, rather, it is not that they are necessarily excluded from consideration, but they are always trumped by the priority ascribed to the obligations public officials have to their states. So, in the case of a conflict, and provided the state of emergency is judged to be serious enough, almost any level of causalities and suffering can be inflicted upon another state or population.


This also helps to show how it is that public officials can use dirty hands reasoning to justify the deaths of half a million children due to sanctions, or launch illegal invasions that kill hundreds of thousands or millions of people and destabilize entire countries, while believing that they are acting morally in doing so. In the absence of a cosmopolitan imperative, they have no clear moral restrictions. Although they agree that the killing of over half a million people is wrong, and a wrong not made right by any ‘success’ that might be had (if it is had), nonetheless, some dirty hands theorists are prepared to accept that level of suffering as justifiable under the circumstances.


In a discussion of the causes of the Bengal famine in 1943 and 1944 in India Amartya Sen observes that although it was not the sole contributing cause to the starvation of over three million people, the British government refused to divert shipping from the war effort to help feed the starving people, because the shipping was needed for the war. Furthermore, they also diverted a section of the available grain to the troops at the front, again tolerating the sacrifice of huge amounts of people and consequent enormous suffering – all for the sake of a perceived national security good.


            I mention India because the absence of cosmopolitanism allows the ugly faces of racism and sexism to rear their heads.  The British government was not fighting the war for the sake of the Indian population. It was fighting it for the sake of the preservation of the British empire, of which the home Islands were the primary population to be defended.  Given a food shortage and a conflict between the feeding of the Indian population, and a similar food scarcity in England, to where would English political loyalties lie? On dirty hands grounds, an English public figure will differentially evaluate populations and sacrifice the ‘less important’ where a choice has to be made.


            A way of thinking about the problem is that the dirty hands claim that the evils are real is irrelevant. It is irrelevant because the evils do not really count for anything. The only thing that happens is that the dirty hands individual feels guilty afterwards – by which time it is too late. The only thing you can do, possibly, is to punish the public figure who has acted ‘correctly’, and perhaps provide some reparations to those who have suffered. Quite how this is possible given the scale of many of the evils is beyond me.  The evils are held to be real, but it makes no difference. It is not merely that tragic choice thinking fails to prevent them; it actively supports their occurrence. Tragic choice theorists still infer an obligation to do them when a lesser evils calculation requires it


The Good Of The State Is The Good


When we speak of public decision-making, we have to recognize that there is an epistemic gap between those actions which are genuinely good for a given community, and the political judgments of specific public figures. In debates about deliberate evil doing, the problem has to do with specific judgments that may or may not help the community to thrive. These are fallible and made in the context of confusing and morally unintelligible contexts. Hence, in the absence of prior non-derogable cosmopolitan norms, it is perfectly possible to systematically oppress women, even though one might believe it to be a wrong. All that is needed are the convictions that such oppression is necessary for the good of the community, that the sexism and other oppression are a lesser evil in comparison to the alternatives, and that the situation is sufficiently exigent to warrant it.


In such a context, questions of racism, sexism, and political and economic justice will inevitably be suppressed in favour of whatever values the public official believes his state represents.  And, in the use of the tortures and other atrocities to save England, Argentina, the US, and other countries from various perceived threats, as a matter of fact racist and sexist strategies are deliberately chosen. Why? Because the sacrifice of those concerns is held to be a priori less significant to the relevant public officials than are the obligations they owe to their states.


Work on feminism and other issues of justice and equality are cosmopolitan issues. If dirty hands theorists prove unable to integrate cosmopolitan concerns into their theory, then they will inevitably justify situations of deliberate oppression of women and other cultures, provided that in the judgment of the relevant public officials the oppression is a lesser evil in comparison to the perceived alternatives.


            It is here that I think we should keep in mind an important remark of Thomas Nagel’s:


What I shall offer, therefore, is a somewhat qualified defense of absolutism. I     believe it underlies a valid and fundamental type of moral judgment – which cannot be reduced to or overridden by other principles. And while there may be other principles just as fundamental, 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. (Nagel 1979)


Nagel speaks of absolute values, where I prefer to speak of cosmopolitan ones. But I believe the intuition is the same. We have to concede that there is at least a small set of non-derogable principles and act-classes. If we prove incapable of doing so, then tragic choice considerations will and do provide the justification for staggering levels of deliberately inflicted suffering and murder.


We can outline how the tragic choice dilemma might be restricted by considering the invalidity of the inference that all universal principles hold only prima facie and can be over-ridden whenever the needs of the situation demand it. One can infer from the situational character of moral decision-making that some moral principles hold prima facie, but not that all can be so treated. That is not a valid virtue-ethical inference. There are good virtue ethical reasons for thinking that some classes of acts are pernicious a priori.


We can identify this class as that set of actions that are inevitably vice creating. If there is such a set of actions, a set of actions that inevitably makes individuals and communities bad, and if the point of moral education and practice is to make communities good, then there is no plausible circumstance in which such actions could count as either justifiable or excusable. If there are such classes of actions, then virtue ethicists can support absolute principles. They can just as much support an absolute prohibition against terror bombing or torture as can the utilitarian or the deontologist. But they can only do this if they allow a carefully chosen and defined set of cosmopolitan prohibitions to have force. If they remain parochial and ethnocentric, then the dirty hands problem permits just about everything. The only constraint is that the circumstances have to demand the atrocity.


The preceding argument is not intended to show that dirty hands dilemmas are impossible. What I have argued is that careful reflection on the preconditions for state torture of suspected state and non-state actors has consequences which entail that torture should be absolutely forbidden. It suggests that virtue ethics supports and needs cosmopolitan principles and values.  I doubt that any exercise of practical wisdom could prevent the corruption of public institutions that the efficient deployment of torture entails unless certain cosmopolitan moral principles are taken as absolute limits even in states of emergency.





Alhoff, Fritz. 2005. Terrorism and Torture. In Philosophy 9/11: Thinking About the War on Terrorism, edited by T. Shanahan. Peru, Illinois: Open Court Publishing.


de Haan, J. 2001. The Definition of Moral Dilemmas: A Logical Problem. Ethical Theory and Moral Practice: An International Forum Vol. 4, No.3.


De Wijze, Stephen. 1994. Dirty Hands - Doing Wrong to Do Right. South African Journal of Philosophy Vol. 13, no.1 (February 1994):27-33.


Dershowitz, Alan. 2004. Tortured Reasoning. In Torture: A Collection edited by S. Levinson. Oxford: Oxford University Press.


Englehardt, Tom. You can do anything with a bayonet except sit on it. A Tomdispatch interview with Mark Danner  2006 [cited. Available from www.tomdispatch.com/index.html?pid=63903.


Jones, Gary. 1980. On the Permissability of Torture. Journal of Medical Ethics 6 (March 1980):11-13.


Morscher, Edgar. 2002. The Definition of Moral Dilemmas: A Logical Confusion and a Clarification. Ethical Theory and Moral Practice: An International Forum Vol. 5, no. 4 (December 2002):485-491.


Nagel, Thomas. 1979. War and Massacre. In Mortal Questions. Cambridge: Cambridge University Press.


Nussbaum, Martha. 1997. Kant and Cosmopolitanism. In Perpetual Peace: Essays on Kant's Cosmopolitan Ideal, edited by J. Bohman, and Lutz-Bachman, Matthias. Cambridge, Mass: The MIT Press.


Posner, Richard. 2004. Torture, Terrorism, and Interrogation. In Torture: A Collection, edited by S. Levinson. Oxford: Oxford University Press.


Stocker, Michael. 1986. Dirty Hands and Conflicts of Values and Desires in Aristotle's Ethics. Pacific Philosophical Quarterly Vol. 67 (January 1986):36-61.


Walzer, Michael. 1977. Just and Unjust Wars. New York: Basic Books.


———. 2004. Political Action: The Problem of Dirty Hands. In Torture: A Collection, edited by S. Levinson. Oxford: Oxford University Press.


Weber, Max. 2004. Politics as a Vocation. In The Vocation Lectures, edited by D. a. S. Owen, Tracy B. Indianapolis/Cambridge: Hackett Publishing Company.




[1]The use of the words ‘absolute’ and ‘unconditional’ are mine, but I think the description is faithful to the claim that these really are evils – they are not relative evils that become sublated under the circumstances. They remain evil come what may. But I am unsure of the point of treating them as unconditional evils once you make them contingent on desire and practical judgment.