Abstract Traditional articulations of the conception of dirty hands, as the doing of wrong in order to do right, invite construals of the issues raised thereby as mired in conceptual confusions and inconsistencies, and moreover as generating unproductive discussions of the scope of the proposed notion itself. The status of the concept of dirty hands is thus precarious, in spite of its provenance in the work of political thinkers such as Machiavelli. This essay articulates one nonparadoxical conception of dirty hands, as the uncomfortable phenomenology of agents authorized to act on behalf of others, when they are called upon to do things that, while morally correct (or at least putatively so), are also personally or morally distasteful. The key point is that the dirt of action remains on these agents’ hands, even if moral responsibility for the relevant actions (when the agents are indeed legitimately authorized to perform them) should be conceived as falling on the persons in whose name they take the said action. This articulation opens up for investigation questions about the phenomenology of authorized agency, as well as the moral labor involved in leadership and in other contexts of authorized distasteful action. These are questions of social ontology. Here there is important philosophical work that remains to be done. 1. Introduction: The Puzzle Suppose that you are a democratically elected, or otherwise legitimate political leader, charged with (among other things) the security of a population. You receive a credible bomb threat from an organization with a reputation for following through on such threats. Your forces, it turns out, have managed to capture a member of this organization, and they now hold him prisoner. There are good reasons to believe that this prisoner knows the whereabouts of the bomb. You employ ordinary interrogation techniques, but he refuses to give up the information. So the only way you can get it is via some form of severe coercion. Suppose it is true that torturing this prisoner, or even better his innocent child, would produce the information you need—an expert in relevant scientific fields pertinent to this question tells you so. Do you proceed with torture? This is the classic dirty-hands situation—the so-called ticking time bomb case. I will refer to cases with a similar structure as the cardinal cases of dirty hands.1 Michael Walzer coins the term “dirty hands” to refer to cases of, as he puts it, “supreme emergency”2 in which an otherwise execrable violation of a moral law or principle is required to avert a great evil that would eventuate absent that violation. In his book Just and Unjust Wars Walzer argues that a situation of supreme emergency justifies the Allies’ terror bombing of German cities in the early stages of World War II (1977, 267–68). The deliberate massacre of thousands of German noncombatants, he argues, was required by supreme emergency, despite the profound immorality involved, because failure to do so risked a Nazi victory, and the consequences of that victory were so dire that the price of committing a severe transgression to prevent it was worth paying. Back to you and your prisoner: why is yours a case of dirty hands? It is to do with the fact that in doing the things that leaders might be called upon to do in supreme emergencies—murder, torture, rape, slavery, to name only the more despicable of such things—they get hands dirty by performing what appear on their face to be morally execrable and at the very least personally repellent things, things people in private life are rarely if ever called upon to do as a matter of duty. There is of course the issue of what qualifies as a supreme emergency. In your situation, it might not be an entire people or way of life at stake. Still, many lives are in the balance. Is it a case of supreme emergency? It is a recurring problem in your line of work—an occupational hazard, if you will—to determine whether you are in a case of supreme emergency. Perhaps it would be an overstatement to say that supreme emergencies are a routine matter for you.3 But if it’s a “must” for you to do some of these things in some cases, however rare, how is it that you are doing wrong when you do them? That’s the puzzle generated by the classic conception of dirty hands as “doing wrong in order to do right.” The deed in question is a wrong (again according to the classical conception) that, seen in its larger context, constitutes a right. A wrong that is also a right. But how could one and the same act be, genuinely and literally, both, at one and the same time, in one and the same context? If doing the dirty is nonetheless the correct thing to do, for a leader in a situation of emergency (supreme or otherwise), in which lives and lifeways hang in the balance, then how can your hands really be dirty when you do the right thing? After all, you had to do it as the means to something else—to the preservation of multitudes, for instance—that’s what makes for its rightness. But if it is indeed the right thing to do, how can it also involve doing something that is simultaneously morally wrong? This leaves many philosophers perplexed, convinced that those who speak of dirty hands have fallen under a spell of confusion. So if it is correct that you, as the person authorized to act in supreme emergency, are genuinely choosing the lesser of two evils, and choosing the lesser evil is the morally right thing to do in the circumstances, then you have done right and there is the end of the matter. Your hands cannot also be dirty. One natural response to this criticism, for the believer in the dirty-hands concept, is, with Machiavelli, to insist that we are dealing with two categories of imperative: the moral and the political. When you, as the person authorized to act in supreme emergency, heed the political imperative therein, you have in so doing failed to heed the moral imperative and so your hands are dirty. So you are doing the politically right thing by doing the morally wrong thing. The trouble with this answer lies in its simplistic assumptions regarding the relevant imperatives. It supposes that moral imperatives are indeed permissibly violated (at least in supreme political emergencies), albeit not without getting oneself morally dirty. But this position presupposes very unsophisticated moral theories. Here are three simple sophistications that will dissolve the issue of dirty hands. (1) An absolutism in moral theory, that simply refuses to acknowledge any higher duty than the moral imperative, would simply not admit permissible violations of them even in supreme emergencies. Alternatively (2) a position that acknowledges other imperatives in addition to moral imperatives, would not acknowledge the violation of moral imperatives as resulting in dirty hands. Finally (3) a moral theory that acknowledges, within its own scope, a variety of prima facie duties (e.g., Ross 1930, 1939), might acknowledge a diverse array of imperatives, including not torturing prisoners and preventing harm to innocents, but proceed to proclaim that, in any decision context, supreme emergency or otherwise, at most one of the prima facie imperatives survives at the end of the day as the true moral imperative. As long as one heeds the resulting actual duty/imperative, there will be no dirty hands. With any of these three sophistications in one’s account of the relevant imperatives, there are no dirty hands, no strange paradoxes. So where is the false step in the analysis of dirty hands as cases of “doing wrong in order to do right”? It’s tempting to say, for example with Nielson (2000), that the very notion of dirty hands is wrong-headed altogether. The categories of right and wrong cannot be overlapping, hence there are no cases of dirty hands—there are only cases where it is hard to discern the right from the wrong, and/or it is also hard to carry through with doing the right thing. But, as I will argue here, it is a mistake to discard the notion of dirty hands. What we require is an analysis of the concept that better illuminates the difficulties experienced in the paradigm cases. (Quite likely, we will require multiple such analyses if, as we multiply the cases within the scope of the concept, the difficulties to be illuminated cluster around multiple points of commonality.) My aim in this paper is to illuminate commonalities shared by at least some of the cardinal cases of dirty hands. These revolve around the demands of certain social roles of responsibility that result from what I shall refer to as authorization. It will be a task for another time to determine the scope of this analysis, and whether different analyses are a better fit to cases (also deserving the label of “dirty hands”) that fall outside this scope but still within of the scope of the philosophical problem going by the name of “moral dilemmas.” By way of presaging, my strategy will be to articulate a distinction between “dirt” and “blood.” The latter will denote the familiar conception of moral responsibility—the moral stain—whereas the former will denote certain features of the phenomenology (the experiential aspects) of executing one or more acts, as an agent, that may result in an outcome that—at least taken on its own—is bad or evil. The ultimate analysis will be that in certain supreme emergencies, agents are called upon to get dirt on their hands via performing acts that lead to outcomes with undesirable features, even though they are doing the right thing—that is, their hands get dirty but they are nonetheless morally unstained, and hence blood-free. The point is that, while it may well be that in the preponderance of real-life cases blood and dirt are acquired bonded together, blood and dirt can also (in other cases) come apart. And this coming apart, at least in the cases I am seeking to characterize, is a result of the deployment of a certain social technology—the technology of authorization. II. Phenomenology of Hard Choices People sometimes feel that the choices they are confronting are hard, not only in the moment of choosing, but in moments subsequent to choice, when they reconcile themselves to the choices they have made. Certainly the cardinal cases of dirty hands belong to this category of choices. So perhaps it might be said that those who speak of dirty hands have taken the phenomenology of being in these circumstances too seriously. They take the regret that lingers afterwards too seriously. They confuse—or so it might be said—the phenomenology of difficult decision-making, and the disquiet or remorse that remains when once the decision is carried out, with a residual moral stain. For example, we might find it difficult to discipline or deprive our children of certain goods. And regrets or remorse associated with such decisions can linger. But for all that, we do not and should not view our hands as morally stained thereby. True, the dirty-hands theorist can respond. And indeed we do not view such things as the disciplining of children as cases of dirty hands. It is, rather, because we are inclined to do so in supreme emergency cases that the language of dirty hands is appropriate to them. The phenomenology there is quite different. In disciplining our children, we might (if indeed we do) regret for a brief moment the action or its short-term consequences, not the very situation or the broader consequences of our actions. We know that disciplining children is part of parenthood, and we will ultimately affirm doing it as right because of the larger consequences; it is because a parent must always be attuned to those wider consequences that we view disciplining children as part of the job description. We do not, however, view supreme emergencies as inevitabilities of leadership or constitutive of it. Choosing between great evils is not an inevitability, not part of the very job description. Those who complain about the exaggeration of the scope of dirty hands will have to confess that it is because we do not see supreme emergencies as a routine part of political life that we are prepared to speak of sometime-dirty hands.4 Philosophical theories of choice are often held responsible not only for the “data points”—for generating the acknowledged correct solutions to dilemmas—but also for explaining why some choices are easy while others are hard. Because the latter are a certain kind of “data” as well. I will not be presuming any specific philosophical theories of the moral domain for our discussion here. But it is worthwhile noting that different moral theories have an easier time than others explaining data points involving how hard a choice is. For instance, when an agent’s options each involves the breaking of what is sometimes referred to, at least colloquially, as a moral “rule,” then a deontological theory will have a much easier time explaining the difficulties an agent might encounter in making such a choice than a consequentialist theory will—generally speaking. This is, to a first approximation, because consequentialist theories have much fewer degrees of freedom—they have only consequences to play with, and they have to produce a complex balance that pits good against bad in a way that produces the correct final reckoning. By contrast, deontological theories can speak of numerous duties of numerous kinds that may or may not be overridden by others, and so are in a position to explain more data points easily with these riches at the same time, including the data of choice difficulty, and moral-stain phenomenology. The point here is really that the feeling rather than the reality of doing wrong is especially important in understanding the phenomenology of dirty hands. A philosophical account of such cases that carries the implication that the sense of doing wrong is simply illusory or misplaced should be regarded as failing to capture the true moral contours of the case. Relatedly, a philosophical account that carries the implication that the sense of doing wrong attaches to making ordinary hard choices too (breaking a promise, for instance, in order not to do a great but not supreme harm) will also be lacking. III. Separating Dirt from Blood I want to maintain that supreme emergencies, and perhaps other occasions too, can call for doing execrable things, at least when enough is at stake. I don’t want to advance here a view on what would constitute “enough.” That requires independent attention. Suffice it to say that even if the scope of dirty hands is limited, it is nonetheless significant. (How can supreme emergencies fail to be significant?) My account is aimed exclusively at illuminating a taxon of such circumstances, with the potential for extension to nonemergency contexts too. The taxon involves the technology of authorization, a technology that develops in human societies in which there is differentiation of authority by social roles. We will require some distinctions to set out this idea. Let’s say that a person P is the target of an imperative, whenever an imperative of some kind is directed at P. Now, a person can be the target of an imperative by virtue of any number of things. Some imperatives are targeted at all human beings—or, more precisely, at all human beings as persons. Other imperatives are targeted at human beings by virtue of certain roles they occupy in their community or society. So we have already been speaking in our ongoing example of you as being the target of imperatives to do repellent things by virtue of being a person in a position of political authority, someone who commands certain resources of a community and empowered to act on behalf of others in times of supreme emergency. Now, what happens when someone is a target of an imperative by virtue of being authorized to act, within the confines of a role, on behalf of others? This is characteristic of the cardinal cases of dirty-hands cases I am here taking on. Roles, because they concentrate and confer power, create perhaps as many problems as they solve—they create the potential for corruption. But let us set aside cases of the exercise of power for corrupt purposes. Those are not cases of dirty hands, but simply of bad behavior. An important caution: the exercise of legitimate power for corrupt purposes, against clear guidelines for its use, cannot be thought of as authorized. What is authorized must fall within the parameters of the authorization itself. What is authorized is inherently bounded by the nature of the authorization. Still, authorization, as such, is not conditional upon doing the right thing; there seems to be no logical inconsistency in contemplating the authorization of wrongdoing. So it cannot be a simple conceptual point, even if true, that one cannot be authorized to do morally unacceptable things. Again, this is not the place to discuss what may or may not be authorized, and specifically whether heinous acts can be authorized in an emergency. I will only stipulate that there may be cases (however rare) of supreme emergency where the authorization can extend to distasteful things, like murder and torture, when enough is at stake. The aim of this paper will not be to proclaim that the notion of authorization, in certain circumstances, renders certain behaviors right that would otherwise be wrong. The point is rather to identify the difference in phenomenology between those who come to complain of dirty hands and those who do not, whatever might be the moral facts. While it may not be a conceptual point, it is often taken as an axiom that one cannot be authorized to perform illegal acts. The question regards the limits of authorization. It is not my aim here to discuss all the limits of my stipulation. Suffice it here to remark that where authorization fails, such as in the case of illegal acts, the attempt to authorize becomes something else, as though by a form of alchemy. Suppose citizen Abel seeks to commission citizen Cain to commit murder which let’s presume qualifies as an illegal act. When Cain accepts the commission, the moral/legal responsibility for the deed would—if authorization of illegal acts were permitted—fall upon Abel. Legal systems will no doubt disagree about how to deal with such cases, but in many countries the authorization is not permitted to go through unproblematically: Cain retains some responsibility for the act. But neither does Abel’s attempt at authorization fail entirely: Abel bears some responsibility too. American law, like criminal law in many countries, introduces the notion of conspiracy to handle such cases, to ensure that neither party to the deal can be considered blameless. Consequently, while responsibility for murder cannot be entirely deflected via authorization, responsibility for conspiracy can be deflected since it is only via deflection of something that Abel comes to be guilty, jointly with Cain, of something that Abel literally does not do. So I hereby stipulate, for purposes of the analysis, that when someone is a duly authorized agent of another person or collectivity, acting within the intended scope of that authorization, then the responsibility for the action performed by the person empowered to act falls on the authorizer and not on the authorized. To reiterate: it does not matter if we don’t (as we really cannot) consider this to be true universally. In fact, we shall consider that it indeed does not hold in all circumstances—that authorization can fail, or simply fail of the intended deflection. What matters at this point is that we view deflection of responsibility as one of the central functions of authorization, even if we must acknowledge it cannot perform this function in all circumstances.5 I ask only that we view authorization as a device for bringing it about (even if it does not universally succeed) that when an agent does something on behalf of a second party, then if that second party has authorized the original to do it, then ownership of the responsibilities appertaining thereto, moral and otherwise, fall upon that first party. I view this stipulation as offering a working notion of the function of authorization, a fundamental principle in the ontology of responsibility, a corollary to the familiar notions of political authority/legitimacy at least in places where democracy prevails. My aim here will be to argue that there are residual effects of acting as an agent on behalf of someone else, consequent upon being the instrument of doing so, that are not (or even cannot be) transferred to the person on whose behalf one acts, and that “dirt” is among these. I claim then that in at least the classic cases of dirty hands, whether the dirty deed is morally right or morally wrong, the phenomenology of dirt—the residues of action when once all forms of responsibility have been subtracted—falls on the authorized individual even though the responsibility (moral and otherwise) falls on the citizenry whose authorization puts the authorized in the position to deploy the resources necessary to act. The authorized person will have dirt on their hands, wherever dirt is pursuant upon action, because authorized persons are literally the instruments used to perform the actions in question, and not because the authorized bear the moral responsibility for it. Dirty hands is the phenomenology of guilt, and not the reality of it. We often also speak of blood on our hands to denote guilt. And my point is simply that dirty hands and bloody hands are contrary points being made. The problem of dirty hands is, quite simply, the dirt without the blood. So when you tortured that child to save the millions, you got dirt on your hands. You got dirt on your hands because you were the instrument (possibly even an indirect one, since the torture might actually have been carried out by a third party) that brought about the torture. Here are some of the many reasons that you might have found the action repellent: (1) it was simply personally distasteful to you, (2) you would never have done such a thing as a private citizen, and (3) you believed it was your prima facie duty, as a human being, not to torture innocents. (Note that  is only one way of describing the moral situation from your perspective; another might be that it was because normally the prohibition against torture has the best consequences.) But you, as the authorized agent, do not bear the moral responsibility for the deed to which you committed persons and/or resources. More precisely, you do not bear that responsibility in your role as authorized agent: you may bear a proportion of it, but only in your capacity as a member of the collectivity that you serve, not in your capacity as leader of it. Because the moral responsibility is literally deflected from the authorized agent to the authorizing party. The responsibility for the action is deflected, but not the dirt. Because if you’re doing your job right, the citizenry will quite typically not even know the things for which they bear responsibility. Part of your job, as their authorized agent, in many cases, is to keep your constituents from having to know about the things done in their name. It keeps their hands clean. It is one of the services you perform and for which you are remunerated: to keep the dirt to yourself.6 Under certain circumstances certain citizens, and certainly those appointed to oversee the actions of authorized individuals, will uncover the things done in the name of the citizens. If the authorized person has done their job well, when light is shone on their deeds, then the citizens will feel the weight of that for which they are collectively responsible (right or wrong), but they will not feel the grit of it. That is the difference between the phenomenology of the agent and the phenomenology of the authorizing party. You as agent feel so much more connected to the harms you inevitably caused by your direct actions than they do in whose name you caused them. For you, and because of the dirt now on your hands, it is hard to weigh the good against the bad of your action and have the feeling follow the calculation faithfully. The conception I am advancing operates also in cases outside of extreme emergencies, as it ought to do, since the notions of responsibility and the deflection of it, which I am invoking, are perfectly general. Suppose that a judge sentences a prisoner to death. The judge does not carry out the sentence. The sentence is carried out by an executioner (or team of such). Wherever the responsibility for the judgment falls—whether your preferred legal theory says it falls on the judge or on those who appointed the judge—the dirt of the execution will be on the executioner’s hands alone. Do executioners experience the phenomenology we spoke of in connection with supreme emergencies? It is an empirical question. My guess is that they do. Thus they would have incentives to ensure that the means of execution meet a certain standard—not a standard associated with whether the prisoner deserves the sentence, but rather with whether the means are appropriate. Nothing that the executioner does, so long as it’s in service of the sentence, will count as a perversion of justice; at the very worst, it will amount to a botching of the sentence. IV. The Ontology and Justification of Authorization: What Democratic Authorization Can Do Authority is a matter of legitimacy to act, within specified scope and express constraints. Moreover, the ontology of authority involves, at least quite frequently, the deflection of responsibility. The authority to act is a matter of being empowered to act. And the empowerment of leaders, at least in democratic contexts broadly construed, is the primordial form of agent-principal relationship. The agent-principal relationship involves construing one agent as an instrument of or for another, not in a collaborative vein as people might work together to achieve a shared end,7 but in what may better be thought of as a representative vein: the one party is construed as standing in for the other by virtue of the latter’s articulated wish that it be so. The act of articulating such a wish is an act of empowering or authorizing someone else to act on one’s behalf. Humans everywhere recognize the empowerment of agents to act on behalf of other persons. The power to empower is one instrument in the toolkit of human social powers of organization and society-building. Construed in this way, the notion of authorization is both morally neutral and independent of the judgment as to whether the action in question, which consequently becomes duly commissioned and authorized, is morally or otherwise either permissible or (further to permissible) compulsory. The power to empower allows us to diversify and divide labor widely. It enhances the ability of persons to act broadly and at larger scales. Perhaps more importantly, it enhances human society’s ability to deliver justice on behalf of persons at diminished capacity or no longer able to act, or whose action in their own behalf would be morally risky. It enlarges the tools in the human toolkit, for good or for ill. Suppose that Abel kills Baker’s innocent child for (let’s assume) personal gain. Suppose it is right to demand justice on behalf of the wronged—the child, foremost, and Baker secondarily. But who can serve that justice upon A? The child is dead. It is not conceptually impossible that Baker could deliver justice. But his position of having also been harmed puts him at risk of not merely delivering (some) justice, but of escalating the cycle of violence and injustice. It would be better if there were someone to act on behalf of the wronged, or even better, someone who can act on behalf of justice. While it might be the case that our human moral sentiments in the heat of the moment are drawn to vengeance and find its delivery satisfying, our better judgment says otherwise: sober judgment is better satisfied when justice is delivered impartially, when the hand of fate or the hand of God delivers a measured punishment that fits the crime. Human sentiments thus acknowledge a role for neutral parties in the delivery of justice. And everywhere our institutions of justice, evolved over many generations, are testimony to this fact. Since we are not satisfied with the hand of fate or the delinquent hand of God, we have erected everywhere institutions for delivery of justice, peopled by agents empowered (even required) to act not exactly as surrogates for, but rather in the name of parties wronged. The existence of such institutions is just one more instance of the profound division of labor that humans have devised to increase their influence all over their environment and thereby to increase the scope of individual lives. Samuel Scheffler (2005) speaks fittingly of a moral division of labor in this connection. Not only do these institutions provide for better and more reliable delivery of justice, but they also help deter wrong-doing. They do it not only in promising punishment for crime, but also in promising neutral agents for its enforcement—agents whose scruples (we can only hope) are beyond all doubt. (This is not to say that unjust legal systems, for example ones that attach death sentences to the smallest infractions, could not also achieve deterrence, and achieve it better—though I think they would not as a matter of fact achieve it better. This is only to say that a level of deterrence can be achieved via such a system. And it remains an open question whether that level of deterrence is optimal.) We recognize the value of this range of institutions when they are not functioning within spec. In times and places where law enforcement is partial, haphazard, undersourced, idiosyncratic, unreliable or simply unable to enforce certain laws (either the ones on the books or the ones that should be), a space opens up for vigilante justice. Suppose Batman metes out justice to the Riddler by stealing back something the latter has purloined in breach of existing property law. Justice surely demands return of the stolen object. But Batman’s actions in breaking and entering the Riddler’s place of residence, and purloining the previously purloined, multiplies the violations of justice in the situation. Vigilante justice multiplies wrongdoing, rather than delivering justice for wrongs committed. And while there might be room for honest error in a suitably authorized justice system, there is none in vigilante justice: if Batman is in error as to the identity of the purloiner, he has no defense of any kind. Vigilante justice operates without a margin of error—because, unlike the agent of a system (however actually just), who is held harmless when acting in the name of justice, the vigilante acts as himself (that’s what makes his case one of vigilantism). If the justice system convicts the wrong person for larceny while acting in good faith, it is not—and rightly so—subject to criminal prosecution. It is immune from such prosecution when operating within the parameters of its role. The same goes for Good Samaritan institutions and those who in certain happenstances are viewed by the law as operating within such parameters. To end the cycle of vengeance and prevent the space for vigilante justice opening up, the institutions of justice require agencies of reliable, impartial, and capable enforcement of good laws. Good laws, furthermore, hold people harmless when they act in good faith as agents of justice within the scope of their role. That is part of the very logic of institutions of justice. It is because authorized agents of the state, like agents of the justice system, cannot be seen as acting “as themselves” that they must be held harmless as themselves when so acting. This is the origin of the phenomenology of that cluster of dirty-hands cases on which I am seeking to throw some light and to which we turn soon enough. Before we proceed, however, there is an important objection to this idea to which we must turn now. The principle I am boldly proclaiming here insists that agents of the state, when performing even distasteful tasks in the line of duty, must be held harmless as themselves (as mere citizens of the state) when acting within the scope of authorization and in good faith—no more than they must be held harmless as themselves when they perform un-distasteful duties. This principle of deflection (as I’ll refer to it) states that it is the state, and thus in proportion the citizenry, on whose shoulders responsibility falls. This is the logic of due authorization. It is the fundamental difference between a democracy (or, more generally, a representative government) and other forms of government such as monarchies or dictatorships. The principle is entirely benign when applied to a state whose laws and operations are within the bounds of decency. In such a state, the well-informed citizenry is presumably content (by the very definition of the case) to accept moral responsibility for actions taken in their name, responsibility deflected onto them from their duly commissioned agents. This principle is voided when duly commissioned agents go rogue or when the state’s very machinery is somehow corrupted, resulting in laws that violate common decency or constitute crimes against humanity, which are presumably such that the citizenry would readily disavow them en masse (were they in a position to do so freely). Under such conditions, the principle cannot be said to hold for the simple reason that in such contexts the conditions of due authorization themselves do not hold. We routinely say of such regimes that the reins of power have been unlawfully or undemocratically seized. And this is enough to block the deflection of responsibility, as follows. Suppose that the machinery of a certain state is seized by a Party that enacts laws abhorrent to a preponderance of citizens (and, if we like, also to the wider world), and suppose that official O orders a subordinate S, who is not a member of the Party, to perform an act A in line with that Party’s mission but constitutes (let’s suppose) a crime against humanity. (This means, among other things, that there is no military necessity in the performance of A—nothing required by supreme emergency, as we’ve been using the term.) If S proceeds to perform A in the role of a subordinate, the responsibility for A is not deflected to the commanding officer O. The logic of authorization does not wash the responsibility for A from S’s hands, simply because the logic of democratic authorization is not applicable in the present case. Democracy has been usurped, so the logic of democratic authorization does not apply. This is why, to quote the liberal activist Wendell Phillips speaking on January 28, 1852 to members of the Massachusetts Anti-Slavery Society: “Eternal vigilance is the price of liberty; power is ever stealing from the many to the few.” One of the risks we run in a democracy is that some of us, as citizens in positions of some power over others, will (God forfend!) find ourselves no longer being held harmless when we execute the duties associated with our positions, because our democracy has been undermined or compromised. Still, the international community should say something more about the relationship between O’s dictates and S’s deeds. We should say that, while O does not bear full moral responsibility, O is not innocent either; similarly, we should say that, while S is not entirely free of moral blame, S cannot be entirely guilty either. Both points are down to the inequality of power enjoyed by S and O respectively—to the asymmetry in their social or other roles vis-à-vis the deed in question. But the precise extent of O’s guilt, and the precise extent of S’s innocence require a great deal of further inquiry, which demands also attention to the specifics in each case. These are the difficulties encountered in the Nuremberg Trials. The Nuremberg Principle VII would seem to agree with some of our results here when it proclaims: “Complicity in the commission of a crime against peace, a war crime, or a crime against humanity … is [itself] a crime under international law.” It agrees with our verdicts here insofar as it is interpreted to say that the logic of authorization does nothing to absolve someone acting under orders of at least some moral responsibility for the deeds they perform when the machinery of the state has been seized by anti-democratic forces. Unlike the difficulties encountered by the Nuremberg Principles I–VI, which deal with articulating defensible notions of crimes against humanity, as well as with the relations between laws of a sovereign state and those of the international community, the difficulties of applying the Nuremberg Principle VII lie entirely with operationalization of the notion of complicity, and how it differs from pure coercion—a problem that arises also, as we have already seen, even when no international conventions or laws are at issue. (When I commission you to kill my nemesis, you are not absolved of responsibility for the crime; and this is down to the fact that the act in question is a crime; the extent of your moral responsibility will depend on the extent to which you are complicit rather than merely coerced.) To be sure our principle regarding the deflection of responsibility in cases where democracy is in question, or where coercion may be at play, has to be complicated. But the fact that we recognize that complications are in order in such cases cannot constitute a defeat of that principle. It must instead be viewed as supportive of its appropriateness in the less problematic cases—in the cases where the technology of authorization is able to fulfill its intended purposes. V. What Democratic Authorization Cannot Do To repeat: agents of the state, when performing a distasteful task in the line of duty, must be held harmless as themselves (as mere citizens of the state) when acting within the scope of authorization and in good faith. As we’ve said, this is in the logic of authorization. But that does not mean that the phenomenology of so acting will line up with the facts of responsibility. And this is the rub. When there is misalignment, dirty hands result. Once more: I do not claim that this is the only route to dirty hands; there may be others. This route to dirty hands results from the features of human social life associated with social roles and some of the responsibilities that come with them. There are multiple social roles that individuals play simultaneously—for example, one can be a child and a parent at one and the same time. The cases of dirty hands I am illuminating result from being authorized to act for a community as well as being a morally sensitive person in that authorizing community at the same time. As a morally sensitive person, the agent sees their own hand reaching out to torture the innocent child (as in your case), hears the child cry out with their own ears, watches as the child’s parent suffers the secondary torture of their child’s primary torture. You as the authorized agent cannot distance yourself from the experience in the way that an arbitrary member of your community might. As someone who knows herself as an individual person too and, as such, the target of moral imperatives of not harming innocents, your sensitivities clash with your responsibilities as the authorized agent. As the authorized agent, you are held harmless; but you nonetheless experience yourself to be the agent of harm. Just as the digging instrument suffers the dirt that results from the action but not the moral responsibility for it, the authorized agent of the state gets hands dirty when acting on behalf of the state, and cannot be put at arm’s length from it. That experience is sometimes felt as a kind of mockery of their good will as a human being, even though the logic of their action is such as to prohibit the guilt from falling on them. The logic of authorization cannot shield the authorized agent from experiencing themselves as the instrument of an action that they indeed would be forbidden from taking as an arbitrary member of the authorizing body. Dirty hands is about being an instrument of harm. The dirt sits uncomfortably with an authorized agent’s sense as a moral individual who is, as such, a target also of certain moral imperatives. Authorization cannot remove the dirt, even as it deflects the moral responsibility. VI. Hard Cases of Authorization I’ve illustrated the points above with what may be referred to as easy cases, where the authorized agent acts squarely within a clear and precisely prescribed scope of authorization. But of course there are hard cases of authorized or quasiauthorized agency. The easy cases share the feature that the authorized agent follows the letter as well as the spirit of the authorizing agent’s expressed instructions and intentions—and indeed where the intentions can be expressed publicly. The hard cases do not have that feature. They are cases where the authorizing agent’s intentions or instructions are less clear, where the authorizing agency is itself an uneasy alliance of agents who are disagreed about what is appropriate to the role of the authorized agent, or where the authorization cannot be made public. Here are a few such hard cases, for the purposes of illustration. I send an agent to the grocery store with instructions to purchase fixings for lunch. My agent returns with foodstuffs that I would not myself have chosen. Did I really authorize the purchase of those foodstuffs? A political party nominates a candidate that represents a compromise of the various factions within the party. Once elected, that person votes for or enacts a totality of policies that no one faction endorses in entirety. Whose will do those policies represent? Undercover agents are sometimes tasked, covertly, with actions that the authorizing agent cannot acknowledge publicly for a variety of reasons, including that they involve or require violations of national or international laws in their execution. It seems as though there is no agent who can or should take responsibility for these actions, even if public opinion might demand someone be held accountable. In case (A), there is such a thing as the will of the authorizing agent, which as it happens is not entirely satisfied, in case (B) the idea of such a will is controversial, and in case (C) the will of the authorizing agent (insofar as it can be said to exist) must not be acknowledged. Especially in the last case, the experiences of the authorized agents is especially instructive vis-à-vis the sort of dirty-hands problems we are treating here: they illuminate even more deeply the agent’s experience of ambiguity. They can represent an extreme of the ticking time bomb case because, at least in some cases, the moral entity who is the target of ordinary moral imperatives is so very attenuated. The conception of dirty hands can shed some light on their experience as well—more precisely, on the price they pay—in dirty hands—in order to serve their country in their particular social roles. Understanding their experience better will pay dividends in terms of shedding light on the hidden moral costs paid by authorized agents who labor in secret. It will illuminate better the nature of moral labor they perform on our behalf and the costs they pay so that we don’t have to. And this understanding will in turn open out onto a larger conversation on the division of moral labor. This is a topic to which too little philosophical attention has been paid. Spies and other persons serving in a variety of clandestine operations within military agencies are our hands and feet where the light of day does not shine. The tasks they perform are possible only if they are allowed to conduct their work in secret, permanently outside of public scrutiny, and possibly also permanently unnoticed. They might be asked to steal, maim, torture, and kill so that we can remain secure and in addition blissfully ignorant of what they are doing. While they bear only a fraction of the full moral responsibility of their actions (they bear exactly the same amount that any arbitrary member of the citizenry bears, which may not be inconsiderable), they carry one hundred percent of the dirt. How can it not take a toll? But what exactly is that toll? Of course there are questions as to whether the range of actions actually taken in the course of clandestine operations are morally permissible. People of good will can disagree. That is not the point here. The point is merely that whichever actions are ultimately deemed permissible and authorized by legitimate political actors, the dirt will always be borne by the men and women on the ground. Indeed they might be the only ones with full knowledge of what is actually being done in the name of the countries they serve. That too takes a toll. People who perform clandestine tasks often have to give up their identities and the possibility of normal lives of blissful ignorance entirely. They have dis-integrated lives. And such personal lives as they do manage to have are much more attenuated. This renders the experience of dirty hands even more stark, since it is not offset by a richer moral life. The services they perform make our simpler (un-dis-integrated) lives easier to live and more phenomenologically palatable, at the cost of theirs being especially unpalatable. They perform moral labor on our behalf, at great cost to themselves. In that regard they are just like soldiers, who might also maim, torture, and kill not in their own name but in the name of a larger cause. The only difference between spies and soldiers is that the latter are given credit for their service while the former are not—they cannot trade on the social capital of public service, such as it might be. One might also inquire about public service of a more mundane sort—accountants who work for federal agencies, lawyers who work for departments of justice, or police officers who work for local precincts. These are persons who, in their social roles, are potentially exposed to the dirtying of hands. They might be tasked with separating people from their homes or property, breaking up families, putting people on trial, and subjecting people to a variety of penalties. (We’ve already mentioned the executioner.) And while many of these individuals are suited to their work by personal temperament (and come to it by choice), others might nonetheless (and quite understandably, even predictably) find elements of that work unpleasant, distasteful, or downright repugnant, even as they might also find it good, right, or necessary. The difference here, when compared with the work of clandestine operators, is primarily that this work is routinely subjected to public scrutiny and (at least on balance) surviving that scrutiny. These persons are, after all, upholding published laws, ostensibly approved and presumed acceptable to the citizenry. Their service is very visible and arguably also morally required. Only infrequently is there any sense of moral wrongdoing associated with carrying out the demands of these roles. Role occupants might not even enjoy a sense of choosing in executing those demands. Still, my analysis of dirty hands admits of such cases of dirty hands in virtue of the fact that the experiences of performing these tasks might conflict with a person’s role-independent sense of what they owe others as individual persons—and not as instruments of a larger entity tasked with carrying out the expressed of the citizenry. VII. The Acquisition of Dirt Rendered More Complex through Social Ontology The analysis I have been advancing relies on the idea that one and the same experience can be refracted through different lenses for an initial parsing by its subject, and then recombined into a rich phenomenology, again for its subject, on the now-familiar model of parallel processing with multiple cognitive systems. One might even conceive of social roles as niches whose occupants are trained (by persons occupying other social niches, no doubt) for specialized cognitive processing of experience, so as to facilitate execution of the duties of these roles. Thus, social niches set the stage for specialization, and even hyperspecialization—a process described eloquently by Millgram (2015). Dirty hands, as my analysis here has it, results from dual processing, first through the lens of being an ordinary citizen, and then through the lens of being an authorized agent trained to perform a specialized task or social role. The analysis suggests that it might also be the case that idiopathic routes to specialized cognitive processing—meaning, independent of official training protocols associated with well-defined social roles—might produce a related effect. Take the example of a scientist who makes a discovery that generates the capability of a certain weapon of mass destruction. When that weapon is deployed (by others, let’s stipulate, absent consultation with the original scientist), will the original scientist experience dirty hands? Some have. Others have instead offered the following reasoning: what I did was simply to make the relevant knowledge available; others were responsible, independently of myself, for its deployment in action contexts; therefore my hands are clean. I have been arguing, of course, that clean is ambiguous, that there are really two questions, and that the question of dirt-freedom is separate from and further to (if not necessarily independent of) the question of stain-freedom. Those scientists who experience their work as instrumental to the abhorrent outcome might well feel their hands dirty, while knowing full well that the moral stain is much more widely shared. What is interesting in this case is the (stipulated) fact that the scientist in question does not participate in the relevant decision making that leads to deployment of the weapon. Due to this fact, my analysis of dirty hands does not comprehend this as a case of dirty hands—not if we’re allowed to take for granted that no one can be an agent who does not participate either in carrying out the deed or in the deliberation process. But it might be made to comprehend the case by a more sophisticated account of agency that admits of key innovations (such as scientific discoveries that are instrumental in the development of relevant technologies) as forms of agency in the effects that they render possible—and actual. But it will then become a challenge why there are those who do not experience agency in such circumstances. I propose that this might be addressed with the innovation of another philosophical technology: the concept of vacancy. Consider that many of the social roles occupied by people are subject to what I will call a social vacancy condition: if the role is not currently occupied, there will arise a certain appreciation that someone must be found to fill the role via a prescribed appointment process. Leadership roles fall in this category, as do many other roles such as cooks in a restaurant or teachers in a school. The point I’m seeking to make here is that not all currently occupied or occupiable roles are thought of as obligatory by some conception of the projects that they contribute to fulfilling. Some roles fall outside this category. Consider now a role that falls in the category of socially nonobligatory (socially discretionary) role—perhaps the role of a scientist or researcher. (Of course there might be disagreement in any given community as to whether such a role is obligatory, but let’s assume for purposes of this example that our community does not maintain obligatory roles of scientists—that the existence among us of such scientists as happen to be gainfully employed as scientists is instead socially discretionary), and similarly for the job or role of scholar. I propose that the social non-obligatoriness of a role can contribute to a greater sense of agency for its occupant when some performance in it produces a technology that is easily adapted to the bringing about of harm. So when scientists or scholars who experience their roles as socially nonobligatory—when they feel themselves as having come to the role in an especially elective way—they will also feel more responsible agentically (though perhaps not more responsible morally) for its downstream effects, however remote. VIII. Concluding Remarks on Social Ontology Elements of the analysis on offer here of certain cases of dirty hands depend crucially on social ontology. The concept of dirt is dependent on the concept of authorization. And the concept of social vacancy is important for distinguishing among differently experienced cases of dirt. This is a fitting result: moral conceptions (for instance, the conception of moral stain) should be integrally related to social ontology, and there should be mutual conditionalization of each upon the other.8 Notes 1 Some believe we should not employ cases such as these because they help license abuse of power in a promiscuous way (Andy Blunden, personal correspondence). But the point of toy examples like this one, in the hands of philosophers, is really to explore whether our considered judgments regarding situations of a certain kind help to build a case for the idea that use of expedients like torture amounts to abuse of power, when our duly authorized leaders employ it in extreme contexts. If considered judgment fails to support the justification of these expedients, there is no worry about toy examples serving as fig leaves. And if, contrariwise, considered judgment supports their use, at least when the duly authorized leader does due diligence, then once again there is no abuse of power when the expedients are deployed responsibly (since that’s what considered judgment supports), and furthermore responsibility for the deed falls on the entirety of the authorizing body (namely, us the citizenry). Withholding permission to employ the philosophers’ toy examples amounts to withholding permission to have the conversation about abuse of power, which means that the judgment call will be made (since it will indeed have to be made, in cases sufficiently like the toy examples) without the benefit of philosophical inquiry. Moreover the discussion of toy examples helps us raise, rather than obstruct the discussion of, important questions about the scope of permissible use of expedients. 2 The term “dirty hands” is an allusion to a book of that name by Jean-Paul Sartre (1949), whose protagonist has to decide whether to assassinate the leader of his political party in the service of larger political ideals of the party. 3 Neil Levy (2007, 52–53, note 26) writes, “Politicians must make deals, compromise with interests they abhor, distribute favours and neglect relationships,” suggesting that these cases result in dirty hands. But this statement, I would contend, exaggerates the scope of dirty hands beyond recognition. 4 Coady (2014) writes, “The idea that dirty hands (in contrast to sheer bad behaviour and corrupt activity) are commonplace in politics is highly dubious,” but proceeds to complain about using the language of supreme emergency anyway, however restricted. 5 And no doubt different moral theories will circumscribe, conditionalize, or otherwise modify this basic precept of authorization differently, as do legal systems generally. For example, moral theories might hold the authorized agent as partly responsible for or at least complicit in the deed. (If a person X takes out a contract on the life of person Z, and person Y takes on that contract, person Y will not be held morally or legally blameless for the harm that comes to Z in consequence of Y’s actions.) The contours of this precept will thus vary in various accounts, but the core function of the precept can nonetheless be honored by all. 6 I don’t believe this is a cynical view. I don’t think our public servants are there to shield us from moral ugliness. I think that in fact many of the things that get their hands dirty are morally required. So no moral stain. It is just that some of these things leave hands dirty in the sense I am propounding here. That there are cases of dirty hands is a tragedy, not a moral failure. 7 Thalos (2008) articulates two different conceptions to describe two different ways in which people can work together in this way. 8 I am grateful to a number of people for candid reactions—to Andy Blunden, Teresa Blankmeyer Burke, Robert Richardson, and Cindy Stark. I am most grateful to Leo Zaibert for pointed comments on a host of specific points. These comments have forced me to sharpen and clarify my views substantially. Still, Leo is not to blame for any murkiness or lack of attention to detail that still remains—for these things I have only myself to thank. REFERENCES Archard David 2013. “Dirty Hands and the Complicity of the Democratic Public,” Ethical Theory and Moral Practice 16: 777– 90. Google Scholar CrossRef Search ADS Coady C.A.J. 2014. “The Problem of Dirty Hands,” The Stanford Encyclopedia of Philosophy ( Spring 2014 Edition), Zalta Edward N., ed., https://plato.stanford.edu/archives/spr2014/entries/dirty-hands. Levy Neil 2007. “Punishing the Dirty,” in Primoratz Igor, ed., Politics and Morality , New York: Palgrave Macmillan, 38– 53. Google Scholar CrossRef Search ADS Millgram E. 2015. The Great Endarkenment , Oxford: Oxford University Press. Nielson K. 2000. “There Is No Dilemma of Dirty Hands,” in Rynard Paul, Shugarman David P., eds., Cruelty and Deception: The Controversy over Dirty Hands in Politics , Peterborough, Ontario: Broadview Press; Australia: Pluto Press, 139– 55. Ross W.D. 1930. The Right and the Good , Oxford: Oxford University Press. Ross W.D. 1939. Foundations of Ethics , Oxford: Oxford University Press. Sartre J-P. 1949. “Dirty Hands,” in Three Plays , New York: Knopf. Scheffler S. 2005. “The Division of Moral Labour,” Proceedings of the Aristotelian Society 79: 229– 53. Google Scholar CrossRef Search ADS Thalos M. 2008. “Two Conceptions of Collectivity,” Journal of the Philosophy of History 2: 83– 104. Google Scholar CrossRef Search ADS Walzer M. 1977. Just and Unjust Wars: A Moral Argument with Historical Illustrations , New York: Basic Books. © The Author(s), 2018. Published by Oxford University Press. All rights reserved. For Permissions, please email: firstname.lastname@example.org
The Monist – Oxford University Press
Published: Apr 1, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 12 million articles from more than
10,000 peer-reviewed journals.
All for just $49/month
Read as many articles as you need. Full articles with original layout, charts and figures. Read online, from anywhere.
Keep up with your field with Personalized Recommendations and Follow Journals to get automatic updates.
It’s easy to organize your research with our built-in tools.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera