Two views on political lesser evil Duas visões sobre o mal menor político

In this article I argue that doing the lesser evil in politics is always permitted and even required. I call this view “pragmatism”. I defend it against “purism”, which claims that it is never permissible to do (the lesser) evil. I reject three arguments for purism, which are based on Alan Gewirth’s principle of intervening action, on an alleged epistemic difference between doing and allowing evil, and on rule-consequentialism. I also address Terrance McConnell’s and Thomas Hill Jr.’s attempts to constrain pragmatism by claiming that doing the lesser evil is not always permitted or required. Although those constraints may apply to most spheres of action, I contend that they do not apply to political action.


Introduction
In this article I ar ue that doing the lesser evil in politics is always permitted and even required.I defend this view against the claim that it is never permissi le to do (the lesser) evil.Following Thomas Hi l Jr. 's terminology, I ca l the former view "pragmatism" and the latter "purism" (Hi l, 1983, p. 213-214).
This article is structured as fo lows.In the first section I reject three ar uments for purism: one is based on Alan Gewirth's principle of intervening a ion, another one is based on an alleged epistemic difference between doing and a lowing evil, and yet another ar ument is based on rule-consequentialism.In the second section I reject Ter ance McConne l's and Thomas Hi l Jr. 's claim that doing the lesser evil is not always permitted or required.Although this may be true of most spheres of a ion, I contend that it is not true about political a ion. 2

Purism
Moral purism claims that you should never do evil, even if it is the lesser evil.By "evil" I mean any a ion or state of affairs that has disvalue.
At the core of purism is the distinction between doing and a lowing.An agent does X when she a ively contributes to X obtaining.An agent a lows X to hap en when she forbears to prevent X. 3Purism strictly forbids doing evil, but it does not strictly forbid a lowing evil to hap en.One reason for this distinction is that a theory that strictly forbids a lowing evil is in a sense inconsistent or not fu ly a ion uiding.This is because it is not always possi le to avoid a lowing evil to hap en.It is perfectly possi le (and indeed frequent) to face a choice between letting one evil hap en or letting another evil hap en, without there being a third option in which no evil hap ens.A theory that strictly forbids a lowing evil implies that in such situations whatever you choose is (a l-things-considered) wrong.This pro lem is not present in a theory that only strictly forbids doing evil, because it is always possi le to abstain from doing evil (because I assume that it is always possi le to abstain from performing any a ion).
Purism should be distin uished from absolutism.Absolutism is the view that certain evils are strictly forbi den.This view is held by authors such as Alan Gewirth (1981) and, according to some interpretations, by Immanuel Kant (1996).My focus in this article, however, is on the more radical view according to which it is strictly forbi den to do any evil.
Notice that purists do not claim that agents are only responsi le for what they do and never for what they a low.Purism is compati le with the claim that agents have (basic or derived) duties to bring about certain desira le states of affairs or to make sure that certain undesira le states of affairs do not take place.For example, purism can be coupled with the view that agents have duties of beneficence, i.e. duties to promote other people's we l-being.Thus, purism is compati le with the claim that in some occasions omissions can be wrong, and that we are sometimes responsi le for what we a low.
What purism claims is that our duties not to do evil always trump our duties to prevent evil from hap ening.
Whenever you must choose between doing evil and letting evil hap en, you must choose the latter.This is so regar less of the amount of evil that you can prevent from hap ening.Size does not matter.There is no need to even measure.In fact, purism should be carefu ly distin uished from the view that "doing evil is always the greate evil" (i.e. that the duty to avoid doing evil is always eightie than the duty not to let evil hap en).If it were simply a matter of weight difference, then if doing evil could prevent a sufficiently high amount of evil, then the duty to avoid doing that particular evil could be outweighed.But the purist wants to deny even that possibility.
The absolute priority of not doing evil is understood in an "a l-things-considered" sense.For purism it is always a l-things-considered wrong to do evil.Purism is compatile with the claim that a lowing evil to hap en is sometimes p ima facie wrong, and sometimes even a l-things-considered wrong.But a lowing evil to hap en can never be a l-things-considered wrong when preventing it from happening requires doing evil.
The question, then, is whether purism so understood can be defended.In its face, purism is plainly wrong.Sup ose, for instance, that you can save someone from being murdered by simply te ling the murderer a sma l lie.Intuitively, the duty not to lie is over i den by the duty to protect that person's life.So purism is counterintuitively strict.It is also counterintuitively permissive.It lets you off the moral hook as soon as you are required to do some mino evil in order to avoid evil or do good (Curzer, 2006, p. 38).
Purism is such an extreme view that it is hard to find ar uments for it in the literature.Even Kant (who goes as far as saying that you must not lie to the murderer at the door) is not a purist in the sense that I am using here.Kant merely claims that some negative duties (the duties not to lie and not to coerce) are absolute (Korsgaard, 1986, p. 337), so he is just an absolutist.Moreover, it is possi le to reconstruct Kantian philosophy as a lowing for some exceptions even to those fundamental duties (Korsgaard, 1986, p. 346-349).Even though purism is not held by any author that I know of, it is important to discuss it, since many students, religious advocates, and others in the general pu lic seem to a here to it.For example, it is common to interpret Paul of Tarsus as a hering to purism, since he explicitly rejects the principle 'let us do evil that good may result' (Rom 3:8, 6:1).
There are at least three minima ly plausi le ar uments for purism, which I now turn to assess.

Alan Gewirth's principle of intervening action
Alan Gewirth's famous article on absolute rights is helpful for finding a possi le ground for the purist view (Gewirth, 1981(Gewirth, , 1982)).As mentioned, Gewirth is not purist because he does not claim that you should never do evil.He only claims that there are some evils that you should never do (violating absolute rights), so he is an absolutist.I consider his ar ument for absolutism here because it can also be used to ground purism.
Gewirth analyses the fo lowing lesser-evil choice scenario: "a gang of ter orists realistica ly threatens to explode a nuclear bomb over a large distant city unless a politica ly a ive lawyer, Abrams, pu licly tortures his mother to death" (Gewirth, 1982, p. 1).Intuitively, it seems that Abrams must torture her mother to death, since this would avoid a much greater evil.So purism must be wrong.Interestingly, Gewirth ar ues that Abrams must abstain from torturing his mother.His reason is not that agents only have responsibility for what they do, and never for what they a low.On the contrary, Gewirth believes that people's right to life grounds positive duties to protect them from homicidal projects when that can be done at no compara le cost (Gewirth, 1981, p. 11).And the cost, in this particular case, is not compara le.The death of Abrams's mother by torture is clearly less costly than the death of thousands of people due to the atomic explosion.
Gewirth's reason is that Abrams is not directly responsi le for those people's deaths; the ter orists are.He resorts to what he ca ls the "principle of intervening a ion" , according to which when there is a causal connection between some person A's performing some action (or inaction) X and some other person C's incurring a certain harm Z, A's moral responsibility for Z is removed if, between X and Z, there intervenes some other action Y of some person B who knows the relevant circumstances of his action and who intends to produce Z or who produces Z through recklessness.The reason for this removal is that B's intervening action Y is the more direct or proximate cause of Z and, unlike A's action (or inaction), Y is the sufficient condition of Z as it actually occurs (Gewirth, 1981, p. 12).
Gewirth's principle implies that if someone else is the proximate cause of an evil, my responsibility for that evil is removed.I may permissi ly a low that evil to hap en.But when I do evil, the proximate cause is me, so I a bear responsibility for it.
The principle of intervening a ion seems to provide a defense of purism.The main pro lem with purism, reca l, is that intuitively we think that there are many situations in which we should do minor (and even major) evil deeds in order to comply with our duty to avoid even greater evils.Now, if it is true that in those situations the greater evil is often brought about by an intervening agent, then according to Gewirth's principle we do not rea ly have a duty to avoid the greater evil.Therefore, it becomes more plausi le to claim that in those situations doing the lesser evil is absolutely forbi den.After a l, by avoiding doing the lesser evil we are not violating any duty at a l, not even a p ima facie duty.The intervening agent completely freed us from responsibility for the outcome.(Notice that lesser-evil choices involving intervening agents are frequent in politics.Consider for instance the case, which I mention below, of bombings against German civil population in order to stop Hitler).
But the principle of intervening a ion is implausi le.First, there are many counterexamples that intuitively show that intervening a ion does not remove responsibility.Imagine that Aurora is about to intentiona ly run over Frank with her car.Frank is hearing and sight impaired, so he is completely unaware that Aurora is about to run him over.I can avoid his death easily by pushing him away from Aurora's path.It seems clear that I have a duty to push him away, even if Aurora's a ion would clearly be the proximate cause of Frank's death, and that the primary duty to avoid Frank's death is hers.
The purist may try to save the principle of intervening a ion from this counterexample by restricting the principle' s scope.She may claim that the principle only ap lies to cases in which one has to do evil in order to avoid (a greater) evil.In the example, by warning Frank I do no evil.But why exactly does intervening a ion remove responsibility only in cases of evildoing, and not in other cases?It seems unlikely that the purist can provide an adequate answer.Moreover, even Gewirth inadvertently provides a counterexample to the claim that intervening a ion removes responsibility in cases in which evildoing is necessary: "if someone threatens to commit suicide or to ki l innocent hostages if we do not break our promise to do some relatively unimportant a ion, breaking the promise would be the obviously right course" (Gewirth, 1981, p. 10).
Lastly, even if the principle of intervening a ion were sound, it would not provide a sufficient defense of purism.That is because there are several lesser-evil choice scenarios that do not involve intervening agents.Consider, for example, a case in which the only way in which half the humanity could be saved from an unforeseea le natural catastrophe is by a ively ki ling a human being.Ki ling her would clearly be the lesser evil.Now, there is no intervening a ion to release us from the responsibility to sacrifice her.So purism needs a different explanation of why in this case we must not do the lesser evil.

Epistemic imbalance
A further ar ument for purism may be built around an important epistemic difference between doing and averting evil.When an agent chooses to do the lesser evil, she is always fu ly certain that she is doing evil, but she is less than fu ly certain that the greater evil wi l be averted.This is for the simple reason that whenever we do evil, we do evil.There is absolutely no chance that we do not do evil when we do evil, but there is always at least some chance that evil wi l not be averted when we try to avert it.For example, take Gewirth's threatened politician.He has fu l certainty that if he tortures his mother, he is torturing; but he has less than fu l certainty that if he tortures he wi l actua ly be a le to avoid the threatened explosions.He cannot even be certain that if he does not torture the greater evil wi l actua ly hap en.The ter orists may end up changing their mind, the bombs may fail, etc.
We may ca l this the "epistemic imbalance" between doing evil and averting evil.By itself the fact that there is an epistemic imbalance does not show much.Of course, purists would claim that responsi le agents must not trade certain evil occur ence for uncertain evil prevention.But this is an over ea ion to uncertainty.What a rational agent should actua ly do is simply to adjust for probabilities.She must not simply weigh the evils themselves, but she should also account for the likelihood of the evils taking place.If by doing evil there is a sufficiently high probability of averting a substantia ly greater evil, a mora ly responsi le agent should do the lesser evil (Hi l, 1983, p. 225-226). 4ut the fact of epistemic imbalance is more pro lematic than it seems once we notice the incidence of biases.Probability calculations are subject to several biases, e ecia ly in politics.Consider first the "overconfidence bias" (Kahneman, 2011).Humans tend to believe they can achieve more than what availa le evidence actua ly war ants (this is a sort of optimism bias: we tend to think that things wi l go better -or not as bad -as they statistica ly wi l).This bias may lead politicians to believe that doing the lesser evil has greater probabilities of averting the greater evil than it actua ly has.The overconfidence bias can be a gravated when combined with forms of egoistic bias: my assessment of probabilities is proba ly going to be distorted if doing the lesser evil is politica ly or economica ly advantageous to me.Egoistic biases may also lead me to believe that there is a greater evil to be avoided when in fact there is none, or when in fact it is the lesser evil (when compared with the evil that my egoistic bias pushes me to do as means).
These biases -the purist claims -only impair assessments of the evil we a low to hap en, not of the evil we do.This is because the former hap ens in the future and indirectly, while the latter hap ens now and directly.There is no need to determine probabilities when assessing the lesser evil we do.Therefore, the biases have no room to kick in.
Of course, biases can be cor ected.Once we know which biases we are vulnera le to, we can through reflection cor ect them or adjust for them.So it seems that in particular cases it is always possi le to know the actual chances of averting evil.But this is too optimistic.In real life, and e ecia ly in politics, agents do not have the time or the wi lingness to cor ect their own biases.
The epistemic imbalance ar ument is insufficient to sup ort purism.The ar ument presup oses that when assessing the evil we do we are not also subject to biases.But in fact biases may impair our assessments of whether a certain a ion is actua ly evil.
For instance, assume for the sake of the ar ument that punishing thieves is not wrong at a l because it is retributively just, and sup ose that punishing thieves also averts a great evil: high crime rates.We may i lustrate this by saying that punishing thieves would bring about 0 units of evil, since it is not evil at a l, while not punishing them and letting crime rates rise would bring about 1,000 units of evil.Now sup ose that a certain politician is a thief, and she subconsciously wants to avoid punishment.One possi le effect of her egoistic biases is that she may be led to believe that punishing thieves is evil.So sup ose she believes that punishing one thief brings about 1 unit of evil.And there are 10 thieves in her society, so punishing them would bring about 10 units of evil.Of course, she also wants to avert high crime rates, which is a great evil (even for her, since she is also at risk of being ro bed).And punishing thieves would clearly be the lesser evil (10 units of evil against 1,000).Unfortunately, according to her biased view averting the greater evil requires doing evil.And, if purism is cor ect, this is impermissi le.So the right thing for her to do (given her evidence) is to a low the greater evil.We thus end up with 1,000 units of evil, when with an unbiased reasoning we would have ended up with 0 units of evil.
Now sup ose we reject purism and embrace pragmatism.The right thing for her to do would be to punish the thieves, because that is the lesser evil.Of course, this conclusion would not be completely cor ect.The right conclusion would be that thieves should be punished because it is not evil at a l (and it would avert a great evil).But sti l the conclusion would be closer to the truth than the one she would ar ive at if she embraced pragmatism.So purism is more vulnera le than pragmatism to this kind of bias.This shows that there is no real epistemic imbalance between assessments of evil done and of evil a lowed, since both are subject to significant biases.

Rule-consequentialism
There is also a rule-consequentialist ar ument for purism.The ar ument starts from the fact that politicians often lie about lesser evil justifications.They often try to justify evils by claiming that they are necessary for averting great-er evils, when in fact there are no greater evils to be averted, or the evil deeds are neither necessary nor sufficient to avert them.What would hap en if everyone believed that purism is cor ect?Politicians would then have one less tool to deceive others.If politicians tried to justify their evil deeds by ap eal to an a leged greater evil, people would not take that justification as valid.The world would be a better place.
For rule-consequentialist reasons, then, we should act as if purism were the right rule to assess evil deeds.We would then avoid a lot of "false positives" , i.e. a lot of a leged lesser-evil justifications which are unsound.Of course, if purism were believed to be cor ect, then we would have the op osite problem of having "false negatives" , i.e. cases in which we would not let politicians do evil when in fact it would be better if they did (in order to avoid a greater evil).Sti l, the rule-consequentialist ar ument claims that the a gregate harmful effect of false negatives is far outweighed by the a gregate harmful effect that false positives would have if purism were rejected.In sum, believing in purism is the lesser evil.
Is it?It is not clear at a l that the a gregate harm from false positives actua ly outweighs the a gregate harm from false negatives.To begin with, it is hard to realistica ly assess how many a leged lesser-evil justifications are actua ly false positives.In fact, it is precisely for that reason that politicians abuse that tool.Moreover, the risk of false negatives should not be minimized.World War II would proba ly have been lost to Germany if the A lied politicians had not been a lowed by their citizens to use evil means to fight.For instance, some ar ue that bombing German cities was a necessary means to win (at least the initial bombings; but the rest were clearly unnecessary) (Walzer, 1977, p. 259-261).Moreover, purism implies that a l defensive wars are impermissi le, since even defensive wars require at least some evildoing in the form of harm to innocent bystanders.So the rule-consequentialist arument is either impossi le to assess or it implies that actua ly purism is the theory that should be rejected.

Pragmatism
Rejecting purism does not imply that pragmatism is correct.Some ar ue that political agents are not always a lowed or always required to do the lesser evil.We may ca l this view "limited pragmatism" .In this section I ar ue that limited pragmatism is wrong.
The pro tanto case for always doing the lesser evil is quite straightforward (Hi l 1983, p. 214).If an agent must choose between two (or more) evils, and one of them is the lesser evil, then it fo lows that all othe things equal she must choose the lesser evil.
Limited pragmatists ar ue that sometimes other things are not equal, and that there are reasons for setting limits on lesser-evil doing.These reasons may be of two different kinds: internal and external.
Internal reasons are reasons for claiming that what seems to be the lesser evil is actua ly the greater evil.When weighing two evils an agent may wrongly assess their relative weight, for example by failing to consider relevant reasons that make one of them a more serious evil.In such cases, another agent may point to those missing reasons in order to revise the assessment (or the agent may find out about those missing reasons herself).For example, imagine you have a friend whose presence could ruin the party you are planning for tonight at your house.You could lie to her and te l her that there is no party tonight.And you think that lying would be the lesser evil, because the suffering she wi l endure when she finds out about the lie is outweighed by the joy everyone wi l have at the party if she is not there.Now someone could try to set limits on your lesser-evil reasoning by pointing out that there is a reason you failed to adequately consider: lying is evil not only because it causes suffering but also because it is a quite serious offence to your friend's dignity, and it is actua ly the greater evil in this case.I ca l this kind of reasons "internal" because they are internal to the assessment of which of the alternatives is a greater evil.External reasons, on the other hand, are general reasons against always doing the lesser evil (the actual lesser evil).Once we have weighed two competing evils adequately, there may sti l be reasons to avoid doing the one that is the lesser evil.These reasons are "external" because they are independent from (or fa l outside of) the assessment of relative evil, and they are independent from the particular evils in play.(For examples of external reasons, see below).
In defending pragmatism what matters are external reasons.The pragmatist is indifferent to hic particular evil is the lesser evil in a given case (and why).She simply claims that whichever is the lesser evil, doing it is always permissi le and even mandatory (at least in politics).This view was held most famously by Machiave li (1950), and recently by Kai Nielsen (2000) and others.(Notice that pragmatism is compati le with non-consequentialism. Non-consequentialism is the view that what makes an a ion evil is not only its consequences, but also other non-consequentialist considerations such as the fact that the a ion breaks a moral rule.Now, what defines pragmatism is not the kinds of considerations that it a lows -consequentialist or not -but how it deals with conflicts between two or more considerations that require actions, as op osed to mere omissions).
Limited pragmatists put forward a number of external reasons for limiting lesser-evil doing.I ar ue that some of those reasons are not valid (section "Invalid reasons"), and that the ones that are valid only ap ly to some spheres of human a ion and not to political a ion (section "Valid reasons that do not ap ly to politics").
(Notice that limited pragmatism is different from absolutism.While absolutists claim that certain types of a ion are never permissi le, limited pragmatists merely say that doing the lesser evil is sometimes impermissi le.While an absolutist may claim that, for example, murdering innocent persons is always wrong, the limited pragmatist merely claims that murdering innocent persons may sometimes be wrong, even when it is the lesser evil.)

Invalid reasons
Consider first Ter ance McConne l's ar uments for limiting lesser-evil doing.He claims that doing the lesser evil is imper is i le when it ir epara ly violates a right (McConne l, 1981, p. 551). 5Notice that McConne l is not a purist.He grants that there are cases in which the lesser evil must be chosen, e ecia ly when no rights are ir epara ly violated and compensation is possi le (McConne l, 1981, p. 546-547).So he is just a limited pragmatist.
McConne l provides two grounds for his view.First, he claims that if ir epara le rights violations were a lowed (in order to do the lesser evil), then victims would be treated as means only.This is not the case, he claims, if we only a low repara le rights violations (McConne l, 1981, p. 552).McConne l may be right about this, but he is beside the point.He is only providing an internal reason for not doing (what seems to be) the lesser evil.A reason for counting ir epara le rights violations as the greater evil.So it cannot be a reason against (unlimited) pragmatism.
Another pro lem with McConne l's claim is that irrepara ly violating a right is often necessary to avoid other people's rights to be ir epara ly violated.To this objection he answers that ir epara ly violating a right myself is always a greater offence than a lowing someone else to ir epara ly violate the rights of others (McConne l, 1981, p. 552).
To back his claim, he offers the fo lowing example.Suppose that someone wi l torture A and B unless you torture C. If you do not torture C, then A and B could complain to you.However -McConne l points out -their complaint would be much softer than the complaint C would a dress to you if you tortured her in order to save A and B (McConne l, 1981, p. 554).I think McConne l is right in his analysis of this particular case, but this case alone is not enough to sup ort his claim.If the number of people I could save from torture were substantia ly higher, then thei complaint would certainly be stronger than C's complaint for torturing her.So it is not the case that violating a right myself is always a greater offence than a lowing someone else to violate others' rights.Moreover, even if McConne l were right about which would be a greater offence, he would sti l be ta king about internal reasons, not external reasons.
Second, McConne l ar ues that if rights could always be permissi ly violated in order to avert a greater evil, then "the possession of rights would be of little value" (McConne l, 1981, p. 554).This is a proper external reason.It is not a reason for regarding something as a greater or lesser evil, but a reason for not a lowing certain lesser evils.However, it is not a good reason.Having rights can sti l be of great value even if ir epara ly violating them is a lowed in some cases.Indeed, rights require that only substantially greater evils must be averted for a rights violation to be justified.Ir epara ly violating my right to life cannot be justified if just five lives could be saved (as in Judith Jarvis Thomson's transplant case [Thomson, 1985[Thomson, , p. 1396]]).Only if the number of lives that can be saved is substantia ly higher may my right be violated.So rights do have great value.
Another author that claims that there are reasons for not always doing the lesser evil is Thomas Hi l Jr.The first reason he provides is that humans have a desire to relate to others with mutual re ect, caring, and sup ort.Doing evil to others, even if it is the lesser of two evils, squelches that desire.So a way to cherish this important desire is to restrict the requirement to always do the lesser evil.Thus, Hi l claims that the lesser evil should be done only if it would prevent a substantially greater evil (Hi l, 1983, p. 226).Second, the requirement to always do the lesser evil runs counter to our desire for moral integrity.We want our acts to express the values we are interna ly committed to.When we do evil (even if it is the lesser evil) there is a lack of cohesion between our inner convictions and our outer deeds.Since the desire for integrity should also be cherished, there are reasons to restrict the requirement to always do the lesser evil (Hi l, 1983, p. 226).
Both reasons are pro lematic because they miss the way in which those two desires would be cherished by doing the lesser evil.It is true that when we do the lesser evil to someone, we are una le to relate to her with re ect, caring, and sup ort.But at the same time we are preventing a greater evil from befa ling others.This shows re ect, caring, and sup ort to the .There is no reason to think that the lack of re ect that we show towards the lesser evil's victim is higher than the lack of re ect we would show to the greater evil's victims by doing nothing to stop it.For the same reason, not doing the lesser evil would also show a lack of integrity.Important inner values would be left unexpressed, such as the value that each individual's life and suffering matter.Since the evil we are preventing is the greate evil, it seems that our integrity is even more compromised when we avoid doing the lesser evil than when we do it.

Valid reasons that do not apply to politics
Hi l offers two a ditional reasons to restrict the duty to always do the lesser evil.First, he claims that there is value in having a we l-defined set of responsibilities.Knowing in advance what our duties are is helpful in order to define our lives around them (Hi l, 1983, p. 22).This value is threatened if we have a duty to always do the lesser evil.We become perpetual policewomen and policemen of evil, always on ca l, with a high degree of unpredictability about what our duties are.
Second, Hi l ar ues that a requirement to always do the lesser evil is incompati le with having ecial relations.There is moral value in having relationships that are based on love, friendship, and mutual re ect.But in order to work, these relationships require an almost unconditional commitment.And such commitment would be undermined by the duty to always do the lesser evil.As Hi l explains, people would only be a le to make conditional commitments with each other: "I would never lie to you, beat you, ki l you, etc... unles I found that doing so would prevent a more serious offense somewhere else" (Hi l, 1983, p. 226-227).
Both reasons Hi l offers are valid for limiting the requirement to do the lesser evil.However, they only ap ly to some spheres of human a ion.Not to political a ion.Consider first the reason from autonomy.As part of their role, political actors acquire ecial o ligations.As a politician you voluntarily lay down your right not to be always on ca l.You are now doing pu lic service.It is part of your job description.Now consider ecial relations.It is true that persons are under no requirement to always do the lesser evil, because otherwise they would not be a le to have valua le ecial relations.But this claim is perfectly compati le with saying that politicians acquire ecial o ligations as part of their role.It is common knowledge that when you engage in politics, your personal relations suffer.For example, when holding office your family and friends can no longer demand preferential treatment.Favoritisms (nepotism and cronyism) are prohibited by your job description.Similarly, it could be ar ued that if you agree to become a politician, you are required to always do the lesser evil, even at the cost of losing valua le personal relations.
Notice that the view on political duties I have just presented is different from standard political realism.According to realism, when compared to other individuals politicians have wider per is ions: they are entitled to do a ions that common morality prohibits.What I hold, on the contrary, is that politicians have fewe permissions.They are not entitled, for the sake of autonomy and personal relations, to evade the requirement to do the lesser evil.
For the same reason, the view I am introducing here seems compati le with political moralism.Morality and politics are not substantia ly different.Politicians simply have more responsibilities than other persons.Notice that moralism grants that roles are legitimate grounds for ecial o ligations.Moralism only denies that role duties may conflict and trump general moral o ligations.My view is compati le with this assertion.
It may be countered that amongst politicians' a ditional responsibilities comes a strict duty to obey the law, e ecia ly in constitutional democracies.Therefore, lesser-evil a ion involving law-breaking cannot be defended by the ar ument from politicians' a ditional responsibilities. 6 In response, it must be noticed that not a l lesser-evil action involves law-breaking.Consider some examples: lying in political negotiations, betraying your political a lies or making a legislative decision that is socia ly unjust towards a group.None of these acts are i legal, but they are sti l evil in a moral sense.Therefore, even if we grant that politicians are not a lowed to break the law in order to do the lesser evil, it must be granted that legal lesser-evil acts in politics are permissi le.

Conclusion
In this article I have ar ued that in politics doing the lesser evil is always permissi le and required.The three aruments for purism that I have assessed are pro lematic.First, intervening a ions do not completely eliminate responsibility, and not a l lesser-evil choices involve intervening a ions.Second, there is no strong epistemic imbalance between doing evil and averting evil; biases not only impair assessments of evil consequences but also assessments of whether an a ion is an actual evil.Third, it is not clear that rule-consequentialism sup orts purism and it may even sup ort pragmatism.
Ar uments for limited pragmatism are also pro lematic.Some of the reasons for limiting pragmatism that I have analyzed do not rea ly eak against doing the lesser evil, but against counting certain particular evils as lesser or greater evils.These are simply internal reasons.The proper external reasons I have assessed (autonomy and ecial relations) successfu ly show that there is no general requirement to always do the lesser evil.However, I have ar ued that political agents are nevertheless under that requirement.