korsgaard the right to lie kant on dealing with evil

8/2/2019 Korsgaard the Right to Lie Kant on Dealing With Evil

1/26

Princeton University Press

The Right to Lie: Kant on Dealing with EvilAuthor(s): Christine M. KorsgaardReviewed work(s):Source: Philosophy & Public Affairs, Vol. 15, No. 4 (Autumn, 1986), pp. 325-349Published by: Blackwell PublishingStable URL: http://www.jstor.org/stable/2265252 .

Accessed: 26/02/2012 11:43

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of

content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms

of scholarship. For more information about JSTOR, please contact [email protected].

Princeton University Press andBlackwell Publishing are collaborating with JSTOR to digitize, preserve and

extend access to Philosophy & Public Affairs.

http://www.jstor.org
http://www.jstor.org/action/showPublisher?publisherCode=blackhttp://www.jstor.org/stable/2265252?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/2265252?origin=JSTOR-pdfhttp://www.jstor.org/action/showPublisher?publisherCode=black


2/26

CHRISTINEM. KORSGAARD The Rightto Lie: Kanton Dealing with Evil

One of the greatdifficultieswith Kant'smoralphilosophy s thatit seemsto implythat our moralobligations eave us powerlessin the face of evil.Kant'stheorysets a high ideal of conduct and tells us to live up to thatidealregardless of what otherpersons aredoing.The results may be verybad. But Kantsays that the law "remains n full force, because it com-mands categorically" G, 438-39/57).' The most well-knownexample of

This paperwas deliveredas the RandallHarrisLectureat HarvardUniversity n October,I985. Versions of the paper have been presented at the Universityof Illinois at Urbana-Champaign, the Universityof Wisconsin at Milwaukee, he Universityof Michigan, andto the Seminaron Contemporary ocial andPoliticalTheory n Chicago.I owe a greatdealto the discussions on these occasions. I want to thank the followingpeople for their com-ments: MargaretAtherton,CharlesChastain,DavidCopp,StephenDarwall,MichaelDavis,GeraldDworkin, Alan Gewirth, DavidGreenstone,John Koethe,RichardKraut,RichardStrier,and Manley Thompson. And I owe special thanks to Peter Hyltonand AndrewsReath for extensive and useful comments on the early written versions of the paper.I. Where I cite or refer to any of Kant'sworks more than once, I have inserted thereference into the text. The followingabbreviations re used:G: Foundationsof the Metaphysicsof Morals I785). The first page number is that of thePrussianAcademyEditionVolume IV;the secondis thatof the translationby Lewis WhiteBeck (Indianapolis:Bobbs-Merrill ibrary f LiberalArts,I959).C2: Critiqueof PracticalReason(1788). PrussianAcademyVolumeV;Lewis White Beck'stranslation Indianapolis:Bobbs-Merrill ibrary f LiberalArts, 1956).MMV:TheMetaphysicalPrinciplesof Virtue(I797). PrussianAcademyVolumeVI; JamesEllington's ranslation n ImmanuelKant:EthicalPhilosophy Indianapolis:Hackett,I983).MMJ:TheMetaphysicalElementsof Justice (I797). PrussianAcademyVolumeVI;JohnLadd's ranslation Indianapolis:Bobbs-Merrill ibrary f LiberalArts, I965).PP: Perpetual Peace (795). Prussian AcademyVolume VIII,translationby Lewis WhiteBeck in On History, edited by Lewis White Beck (Indianapolis:Bobbs-MerrillLibraryofLiberalArts, I963).SRL: "On a Supposed Right to Lie from Altruistic Motives" I797). PrussianAcademyVolume VIII;translationby Lewis White Beck in ImmanuelKant:Critique of Practical


3/26

326 Philosophy&PublicAffairs

this "rigorism," s it is sometimes called, concerns Kant'sviews on ourduty to tell the truth.In two passages in his ethical writings, Kant seems to endorse thefollowingpairof claims about this duty:first, one must never under anycircumstances or for any purposetell a lie; second, if one does tell a lieone is responsibleforall the consequences thatensue, even if they werecompletelyunforeseeable.One of the two passages appearsin the MetaphysicalPrinciples ofVirtue. There Kant classifies lying as a violationof a perfect duty tooneself. In one of the casuisticalquestions,a servant,underinstructions,tellsavisitor helie thathis master s not athome. Hismaster,meanwhile,sneaks off and commits a crime, which would have been preventedbythe watchman sent to arrest him. Kantsays:

Uponwhom ... does the blame fall? To be sure, alsoupon the servant,who here violateda dutytohimselfbylying, the consequence of whichwill now be imputedto him by his own conscience. (MMV, 31/93)

The other passage is the infamous one about the murderer at the doorfrom the essay, "On A SupposedRight to Lie FromAltruisticMotives."Here Kant'sclaims are more extreme, for he says that the liar may beheld legally as well as ethicallyresponsible for the consequences, andthe series of coincidences he imagines is even more fantastic:After you have honestly answered the murderer's question as towhether his intendedvictimis at home, it may be that he has slippedout so thathe does not come in the wayof the murderer,and thus thatthe murdermay not be committed. Butif you had lied and saidhe wasnot at home when he hadreallygone out withoutyourknowingit, andif the murdererhad then met him as he went awayand murderedhim,you might justly be accused as the cause of his death. For if you hadtold the truthas far as you knew it, perhapsthe murderermight have

Reason and OtherWritings in MoralPhilosophy Chicago: Universityof ChicagoPress,I949; reprint,New York:GarlandPublishingCompany,1976).LE: Lectures on Ethics (1775-I780). Edited by Paul Menzer from the notes of TheodorFriedrichBrauer,using the notes ofGottliebKutznerandChristianMrongovius;ranslatedby Louis Infield(London:Methuen&Co., Ltd., I930; reprint,New York:HarperTorch-books, I963; currentreprint,Indianapolis:Hackett PublishingCo., I980).


4/26

327 The Right to Lie: Kanton Dealingwith Evil

been apprehendedby the neighborswhile he searchedthe house andthus the deed might have been prevented. (SRL,427/348)Kant'sreaders differ about whether Kant'smoralphilosophy commitshim to the claims he makes in these passages. Unsympathetic readersare inclined to take them as evidence of the horrifying conclusions towhich Kantwas led by his notion that the necessity in duty is rational

necessity-as if Kantwere clinging toa logical point in the teeth of moraldecency. Suchreaders akethese conclusions asadefeatforKant'sethics,or for ethical rationalismgenerally;or they take Kant to have confusedprincipleswhich aremerely generalin theirapplicationand prima faciein their truthwith absolute and universal aws. Sympatheticreaders arelikely to arguethat Kanthere mistook the implicationsof his own theory,and to try to show that, by carefulconstruction and accurate testing ofthe maxim on which this liar acts, Kant'sconclusionscan be blocked byhis own procedures.Sympatheticand unsympatheticreaders alike have focused their at-tention on the implicationsof the first formulationof the categoricalimperative, he Formulaof UniversalLaw. The Foundations of the Met-aphysics of Morals contains two other sets of terms in which the cate-gorical mperative s formulated: he treatmentof humanityas an end initself, and autonomy,or legislative membershipin a Kingdomof Ends.My treatment of the issue falls into three parts. First, I want to arguethat Kant'sdefenders are rightin thinking that, when the case is treatedunder the Formulaof Universal Law, this particular ie can be shown to

be permissible. Second, I want to argue that when the case is treatedfrom the perspective provided by the Formulas of Humanity and theKingdomof Ends, it becomes clear why Kant is committedto the viewthat lying is wrongin everycase. But fromthis perspectivewe see thatKant's rigorismaboutlying is not the result of a misplacedlove of con-sistency or legalistic thinking. Instead,it comes froman attractive dealof human relationswhich is the basis of his ethical system. If Kant iswrongin his conclusion aboutlying to the murdererat the door,it is forthe interestingandimportant easonthatmoralitytself sometimes allowsor even requires us to do something that from an ideal perspective iswrong. The case does not impugn Kant'sethics as an ideal system. In-stead, it shows that we need special principlesfor dealingwith evil. My


5/26

3.28 Philosophy&PublicAffairs

thirdaimis to discuss the structurethat an ethical system must have inorderto accommodate such special principles.UNIVERSAL LAWThe Formula of Universal Law tells us never to act on a maxim that wecould not at the same time will to be a universal law. A maxim whichcannot even be conceived as a universallaw without contradiction s inviolationof a strict and perfect duty, one which assigns us a particularaction or omission. A maxim which cannot be willed as universal lawwithout contradictingthe will is in violationof a broad and imperfectduty, one which assigns us an end, but does not tell us what or howmuch we should do toward t. Maximsof lying are violations of perfectduty,and so aresupposedtobe the kind that cannotbe conceived withoutcontradictionwhen universalized.

The sense in which the universalizationof an immoral maxim is sup-posed to "contradict" tself is a matter of controversy.On my reading,which I will not defend here, the contradictionn questionis a "practical"one: the universalizedmaxim contradicts tself when the efficacy of theaction as a method of achieving its purposewould be undermined by itsuniversal practice.2 So, to use Kant'sexample, the point against falsepromising as a method of getting ready cash is that if everyone attemptedto use false promising as a method of gettingready cash, false promisingwould no longer work as a method of getting ready cash, since, as Kantsays, "no one would believe what was promisedto him but would onlylaugh at any such assertionas vain pretense"(G, 422/40).Thus the test questionwillbe:couldthis action be the universalmethodof achieving this purpose? Now when we consider lying in general, itlooks as if it could not be the universalmethod of doing anything. Forlies are usually efficaciousin achieving their purposes because they de-ceive, but if they were universallypracticedthey would not deceive. Webelieve what is said to us in a given context because most of the timepeople in that context say what they really think or intend. In contextsin which people usually say false things-for example, when telling sto-ries that are jokes-we are not deceived. If a storythat is a joke and is

2. I defend it in "Kant'sFormulaof UniversalLaw,"Pacific PhilosophicalQuarterly66,nos. I & 2 (January/April I986): 24-47.


6/26

329 TheRightto Lie:Kanton Dealing with Evil

false counts as a lie, we can say that a lie in this case in not wrong,because the universal practice of lying in the context of jokes does notinterferewith the purposeofjokes, which is to amuse and does not dependon deception. But in most cases lying falls squarely into the category ofthe sort of action Kant considers wrong: actions whose efficacy dependsuponthe factthatmost peopledo notengagein them, and which thereforecan only be performedby someone who makes an exception of himself.(G, 424/42)When we tryto applythis test to the case of the murdererat the door,however, we run into a difficulty.The difficultyderives from the fact thatthere is probablyalreadydeception in the case. If murderers standardlycame to the door and said: "Iwish to murder your friend-is he here inyour house?" then perhapsthe universalpractice of lying in orderto keepa murderer from his victim would not work. If everyone lied in thesecircumstances the murdererwould be awareof that fact and would notbe deceived by your answer. But the murderer s not likely to do this, or,in any event, this is not how I shall imagine the case. A murdererwhoexpects to conduct his bus-inessby asking questions must suppose thatyou do not know who he is and what he has in mind.3If these are thecircumstances, and we trytoascertainwhether there could be a universalpractice of lying in these circumstances, the answer appears to be yes.The lie will be efficacious even if universally practiced. But the reasonit will be efficacious is rather odd: it is because the murderer supposesyou do not know what circumstances you arein-that is, thatyou do not

3. I am relying here on the assumptionthatwhen people ask us questions, they give ussome account of themselves and of the context in which the questions are asked. Or, ifthey don't,it is because they arerelyingon a context that is assumed. If someone comesto your door looking for someone, you assume that there is a family emergency or somesuch thing. I am preparedto count such reliance as deception if the questioner knowsabout t and uses it, thinking thatwe wouldrefuse to answer his questionsif we knew thereal context to be otherwise. Sometimes peopleask me, "Suppose he murderer ust askswhether his friend s in yourhouse, withoutsaying anythingaboutwhyhe wantstoknow?"I think that, in our culture anyway, peopledo not just ask questions of each other aboutanything except the time of day and directions or getting places. Afterall, the reason whyrefusal to answer is an unsatisfactoryway of dealing with this case is that it will almostinevitably give rise to suspicion of the truth, and this is because people normallyanswersuch questions. Perhaps if we did live in a culture in which people regularly ust askedquestionsin the way suggested, refusal to answer would be commonplaceand would notgive rise to suspicion; t would not even be consideredodd or rude. Otherwisethere wouldbe no way to maintain privacy.


7/26


know you are addressing a murderer-and so does not conclude fromthe fact that people in those circumstances always lie that you will lie.The same point can be made readily using Kant'spublicitycriterion.(PP, 38i-83/129-3 1) Can we announcein advance ourintentionof lyingto murdererswithout, as Kant says, vitiatingour own purposes by pub-lishing our maxims? (PP, 383/13I) Again the answer is yes. It does notmatter if you say publicly that you will lie in such a situation, for themurderersupposes that you do not know you are in that situation.4These reflectionsmight lead us to believe, then, that Kant was wrongin thinkingthat it is never all right tolie. It is permissibletolie to deceiversin orderto counteract the intended results of their deceptions, for themaxim of lying to a deceiver is universalizable.The deceiver has, so tospeak, placed himself in a morallyunprotectedposition by his own de-ception. He has createda situationwhich universalizationcannot reach.

HUMANITYWhenwe applythe FormulabfHumanity,however,the argumentagainstlying that results applies to any lie whatever. The formularuns:

Act so that you treat humanity,whetherin yourown personorin thatof another, always as an end and never as a means only. (G, 429/47)In orderto use this formula or casuisticalpurposes,we need to specifywhat counts as treatinghumanity as an end. "Humanity"s used by Kant

specifically to refer to the capacity to determine ends through rationalchoice. (G, 437/56; MMV,392/50) Imperfectduties arise from the obli-gation to make the exercise, preservation,and development of this ca-pacityitself an end. The perfectduties-that is, the duties of justice, and,4. In fact, it will now be the case that if the murderersupposes that you suspect him,he is not going to ask you, knowing that you will answer so as to deceive him. Since wemust avoid the silly problemaboutthe murdererbeing able to deduce the truthfromhisknowledgethatyou will speakfalsely, what you announceis thatyou will say whatever snecessary in orderto conceal the truth. There is no reason to suppose that you will bemechanicalabout his. Youarenot goingto be a reliable ourceof information.Themurderer

will thereforeseek some other way to locate his victim.On the otherhand,supposethatthemurdererdoes,contrary o mysupposition,announcehis real intentions. Then the arguments that I have given do not apply.In this case, Ibelieve your onlyrecourse s refusal to answer(whetherornot the victim is in your house,or you know his whereabouts).If an answer is extortedfromyou by force you may lie,accordingto the argumentI will give laterin this article.


8/26

331 TheRight toLie:Kanton Dealingwith Evil

in the realm of ethics, the duties of respect-arise fromthe obligation tomake each human being's capacityforautonomous choice the conditionof the value of every other end.In his treatmentof the lying promisecase under the Formulaof Hu-manity, Kantmakes the followingcomments:For he whom I want to use formy own purposes by means of such apromisecannot possiblyassent to my mode of acting against him andcannotcontainthe end ofthisaction n himself.. . . he whotransgressesthe rights of men intends to make use of the persons of othersmerelyas means, withoutconsideringthatasrationalbeings, theymust alwaysbe esteemed at the same time as ends, i.e., only as beings who mustbe able to contain in themselves the end of the very same action. (G,429-30/48)

In these passages, Kantuses two expressions that are the key to under-standing the derivationof perfect duties to others from the Formula ofHumanity. One is that the other person "cannotpossibly assent to mymode of acting toward-him"and the second is that the other personcannot "contain he end of this actionin himself."These phrasesprovideus with a test forperfectduties to others:an actionis contrary o perfectduty if it is not possible for the other to assent to it or to hold its end.It is importantto see that these phrases do not mean simply that theotherperson does not or would not assent to the transactionor that shedoes not happen to have the same end I do, but strictlythat she cannotdo so: that something makes it impossible.If what we cannot assent tomeans merely what we are likely to be annoyed by, the test will besubjective and the claim that the person does not assent to being usedas a means will sometimes be false. The object you steal from me maybe the gift I intended foryou, and we may bothhave been motivatedbythe desire that you should have it. And I may care aboutyou too muchor too little to be annoyedby the theft. For all that, this must be a clearcase of your using me as a mere means.5So it must not be merely that your victim will not like the way you

5. Kanthimself takesnoticeof this sort of problem n a footnoteto thispassage in whichhe criticizesGolden-Rule ype principlesfor, among otherthings, the sortof subjectivityin question:such principlescannotestablishthe duty of beneficence,forinstance,because"manya man would gladlyconsent thatothersshouldnot benefit him, providedonly thathe might be excused fromshowingbenevolence to them"(G, 430n./48n.).


9/26


propose to act, that this is psychologicallyunlikely, but that somethingmakes it impossible for her to assent to it. Similarly, t must be arguedthat something makes it impossiblefor her to hold the end of the verysame action. Kant never spells out why it is impossible, but it is notdifficult to see what he has in mind.Peoplecannot assent to a way of actingwhen they are givenno chanceto do so. The most obvious instance of this is when coercionis used. Butit is also true of deception:the victim of the false promisecannot assentto it because he doesn't know it is what he is being offered. But evenwhen the victim of such conduct does happento know what is going on,there is a sense in which he cannot assent to it. Suppose, for example,that you come to me and ask to borrowsome money, falsely promisingto pay it back next week, and suppose that by some chance I knowperfectly well that your promise is a lie. Suppose also that I have thesame end you do, in the sense that I want you to have the money, sothat I turn the money over to you anyway. Now here I have the sameend thatyou do, and I tolerateyourattempts to deceive me to the extentthat they do not prevent my giving you the money. Even in this case Icannot really assent to the transactionyou propose.We can imagine thecase in a number of different ways. If I call your bluff openly and say"nevermind that nonsense, just take this money" then what I am doingis not accepting a false promise,but givingyou a handout,and scorningyour promise.The nature of the transaction s changed: now it is not apromisebut a handout.If I don'tcallyou on it, but keepmy own counsel,it is still the same. I am not accepting a false promise.In this case whatI am doing is pretending to accept your false promise. But there is allthe difference in the worldbetween actuallydoing something and pre-tending to do it. In neitherof these cases can I be describedas acceptinga false promise,for in both cases I fix it so that it is something else thatis happening. Myknowledgeof what is going on makes it impossibleforme to accept the deceitful promisein the ordinaryway.The question whether another can assent to your way of acting canserve as a criterionforjudging whether you are treatingher as a meremeans. We will say that knowledgeof what is going on and some poweroverthe proceedingsare the conditionsof possible assent; withoutthese,the concept of assent does not apply.This gives us anotherway to for-mulate the test for treatingsomeone as a mere means: suppose it is thecase that if the other person knows what you are trying to do and has


10/26

333 TheRight to Lie: Kanton Dealing with Evil

the power to stop you, then what you are tryingto do cannot be what isreally happening. If this is the case, the action is one that by its verynature is impossible for the other to assent to. You cannot wrest fromme what I freely give to you; and if I have the power to stop you fromwresting something from me and do not use it, I am in a sense freelygiving it to you. This is of course not intendedas a legal point: the pointis thatany action which depends for its natureand efficacy on the other'signorance or powerlessness fails this test. Lying clearly falls into thiscategoryof action: it only deceives when the other does not know that itis a lie.6A similar analysiscan be given of the possibilityof holding the end ofthe very same action.In cases of violationof perfect duty, lying included,the other person s unable to hold the end of the verysame action becausethe way that you act prevents her fromchoosing whether to contributeto the realizationof that end or not. Again, this is obviouslytrue whensomeone is forcedto contributeto an end, but it is also true in cases ofdeception.Ifyou givealying promiseto get somemoney, the otherpersonis invited to think that the end she is contributingto is your temporarypossession of the money: in fact, it is your permanentpossession of it. Itdoesn'tmatter whetherthatwouldbe allrightwith her if she knew aboutit. What matters is that she never gets a chance to choose the end, notknowing that it is to be the consequence of her action.7According o the Formulaof Humanity,coercion anddeceptionare themost fundamental forms of wrongdoingto others-the roots of all evil.Coercion and deceptionviolate the conditionsof possible assent, and allactions which depend for their nature and efficacy on their coercive ordeceptive character are ones that others cannot assent to. Coercion anddeception also make it impossible for others to choose to contribute to

6. Sometimes t is objected that someone could assent to being lied to in advanceof theactual occasion of the lie, and that in such a case the deception might still succeed. Onecan thereforeagree to be deceived. I think it depends what circumstancesareenvisioned.I can certainly agree to remain uninformedabout something,but this is not the same asagreeing to be deceived. For example, I could say to my doctor:"Don't ell me if I am fatallyill, even if I ask."But if I then do ask the doctorwhether I am fatally ll, I cannotbe certainwhether she will answer me truthfully.Perhaps what's being envisioned is that I simplyagree to be lied to, but not about anythingin particular.Will I then trust the person withwhom I have made this odd agreement?7. Asimilarconclusion aboutthe wayin which the Formulaof Humanitymakes coercionand deception wrong is reached by Onora O'Neill in "Between ConsentingAdults,"Phi-losophy&Public Affairs 14, no. 3 (Summer I985):252-77.


11/26


our ends. This in turn makes it impossible, according to Kant'svaluetheory,forthe ends of such actions to be good.For on Kant'sview "whatwe call goodmust be, in the judgment of everyreasonableman, an objectof the faculty of desire" (C2, 60/62-63). If your end is one that otherscannotchoose-not because of what they want, but because they are notin a positionto choose-it cannot,as the end of thataction,be good.Thismeans that in any cooperativeproject-whenever you need the decisionsand actions of others in orderto bring about your end-everyone who isto contributemust be in a position to chooseto contributeto the end.The sense in which a goodend is an object for everyoneis that a goodend is in effect one that everyone, in principle,and especially everyonewho contributes to it, gets to cast a vote on. This voting, or legislation,is the prerogativeof rationalbeings; and the ideal of a world in whichthis prerogative s realized is the Kingdomof Ends.THE KINGDOMOF ENDSThe Kingdomof Ends is representedby the kingdomof nature; we de-termine moral laws by considering their viabilityas natural laws. OnKant'sview, the will is a kind of causality.(G, 446/64) A person, an endin itself, is a free cause, which is to say a firstcause. Bycontrast,a thing,a means, is a merelymediatecause, a link in the chain. A first cause is,obviously,the initiatorof a causal chain, hence a realdeterminerof whatwill happen.The idea of decidingforyourselfwhether youwill contributeto a given end can be representedas a decision whether to initiate thatcausal chain which constitutes yourcontribution.Anyactionwhich pre-vents or divertsyou from makingthis initiatingdecisionis one that treatsyou as a mediate rather than a first cause; hence as a mere means, athing, a tool. Coercionand deception both do this. And deception treatsyou as a mediatecause in a specificway:it treatsyourreasonas a mediatecause. The false promiserthinks: if I tell her I will pay her back nextweek, then she will choose to giveme the money. Yourreasonis worked,like a machine: the deceivertries to determinewhat levers to pull to getthe desired results fromyou. Physical coercion treats someone's personas a tool; lying treats someone's reason as a tool.This is why Kantfindsit so horrifying; t is a directviolationof autonomy.We may say that a tool has two essential characteristics: t is there to


12/26

335 TheRight to Lie: Kanton Dealing with Evil

be used, and it does not control itself-its nature is to be directed bysomething else. To treat someone as a mere means is to treat her as ifthese things were true of her. Kant'streatmentof our duties to others inthe MetaphysicalPrinciples of Virtue is sensitive to bothcharacteristics.We are not only forbidden o use another as a mere means to our privatepurposes.We arealso forbidden o take attitudestowardherwhich involveregarding her as not in control of herself, which is to say, as not usingher reason.This latter is the basis of the duties of respect. Respect is violated bythe vices of calumny and mockery (MMV,466-68/131-33): we owe toothers not only a practical generosity towardtheir plans and projects-a dutyof aid-but also a generosityof attitude towardtheirthoughts andmotives. To treat another with respect is to treat him as if he were usinghis reason and as far as possible as if he were using it well. Even in acase where someone evidently s wrongormistaken,we ought to supposehe must have what he takes to be good reasons forwhat he believes orwhat he does. This is not because, as a matter of fact, he probablydoeshave good reasons. Rather,this attitude s somethingthatwe owe to him,somethingthatis his right.Andhe cannot forfeit t. Kant s explicitaboutthis:

Hereuponis founded a duty to respect man even in the logical use ofhis reason: not to censure someone's errorsunder the name of ab-surdity, inept judgment, and the like, but rather to suppose that insuch an inept judgment there must be something true, and to seek itout.... Thus it is also with the reproachof vice, which must neverburst out in complete contemptordeny the wrongdoerall moralworth,because on that hypothesis he could never be improvedeither-andthis latter is incompatiblewith the idea of man, who as such (as amoralbeing) can never lose all predisposition o good.(MMV, 63-64/128-29)To treat others as ends in themselves is always to address and dealwith them as rationalbeings. Everyrationalbeing gets to reasonout, forherself, what she is to think, choose, or do. So if you need someone'scontributionto your end, you must put the facts beforeher and ask forher contribution. If you think she is doing something wrong, you maytryto convince her by argumentbut you maynot resort to tricks or force.


13/26


The Kingdomof Ends is a democratic deal, and poorjudgment does notdisqualify anyone for citizenship. In the Critique of Pure Reason, Kantsays:Reason depends on this freedomfor its very existence. For reason hasno dictatorialauthority; ts verdict is always simply the agreement offree citizens, of whom each one must be permittedto express, withoutlet or hindrance, his objectionsor even his veto.8

This means that there cannot be a good reason for taking a decision outof someone else's hands. It is a rationalbeing's prerogative,as a firstcause, to have a share in determiningthe destiny of things.This shows us in anotherway why lying is for Kanta paradigmcaseof treatingsomeone as a mere means. Any attemptto controlthe actionsand reactionsof another by any means except an appealto reason treatsher as a mere means, because it attempts to reduce her to a mediatecause. This includes much more than the utterance of falsehoods.In theLectureson Ethics, Kantsays "whatevermilitates againstfranknesslow-ers the dignityof man"(LE, 231 ).9 It is an everyday emptation,even (orperhaps especially) in our dealings with those close to us, to withholdsomething, or to tidy up an anecdote,or to embellish a story,or even justto place a certain emphasis, in orderto be sure of getting the reactionwe want.'0Kantholds the Socratic view that any sortof persuasionthatis aimedat distracting ts listener's attentionfrom either the reasons thatshe ought to use or the reasons the speakerthinks she will use is wrong."

8. ImmanuelKant's"Critique f Pure Reason," ranslatedby NormanKempSmith (NewYork:St. Martin'sPress, i965) A738-39/B766-67, p. 593.9. It is perhaps also relevantthat in Kant'sdiscussionof perfectmoral friendshiptheemphasis s not on goodwill towardone another,buton completeconfidenceandopenness.See MMV, 471-72/138-39.

iO. Some evidence that Kant is concernedwith this sort of thing may be found in thefact that he identifiestwomeanings of the word "prudence"Klugheit);"Theformersensemeans the skill of a man in having an influence on othersso as to use them for his ownpurposes.The latter s the abilityto unite all these purposesto his own lastingadvantage"(G,4I6n./33n.).A similarremark s found in Anthropologyroma PragmaticPointofView(1798). See the translation by Mary J. Gregor (The Hague: Martinus Nijhoff, 1974), p.i83; Prussian AcademyEditionVolumeVII,p. 322.

i i. I call this view Socraticbecause of Socrates'concernwith the differencesbetweenreason and persuasionand, in particular,becausein the Apology,he makes a case forthecategoricaldutyof straightforwardness.ocratesandPlatoare also concernedwith a trou-blesomefeature of this moralview that Kantneglects. An argumentmust come packagedin some sort of presentation,and one may well object that it is impossible to make a


14/26

337 TheRightto Lie: Kanton Dealingwith Evil

In light of this account it is possible to explain why Kant says what hedoes about the liar'sresponsibility. n a Kantian heory our responsibilityhas definite boundaries:each person as a first cause exerts some influ-ence on what happens, and it is your part that is up to you. If you makea straightforward ppeal to the reason of another person, your respon-sibilityends there and the other's responsibilitybegins. But the liar triesto take the consequences out of the hands of others; he, and not they,will determine what form their contributionto destiny will take. By re-fusing to share with others the determinationof events, the liar takes theworldinto his own hands, and makes the events his own. The results,goodorbad,areimputable to him, at least in his own conscience. It doesnot follow from this, of course, that this is a risk one will never want totake.HUMANITY AND UNIVERSAL LAWIf the foregoingcasuistical analyses are correct,then applyingthe For-mula of UniversalLaw and the Formula of Humanity lead to differentanswers in the case of lying to the murderer at the door. The formerseems to say that this lie is permissible,but the lattersays that coercionanddeception are the most fundamental orms of wrongdoing.In a King-dom of Ends coercive and deceptivemethods can never be used.This result impugns Kant'sbeliefthatthe formulasareequivalent.Butit is not necessarytoconcludethat the formulas latlysaydifferentthings,and are unrelatedexcept for a wide rangeof coincidence in their results.For one thing, lying to the murdererat the door was not shown to bepermissible in a straightforwardmanner: the maxim did not so muchpass as evade universalization. For another, the two formulas can beshown to be expressionsof the same basictheoryofjustification. Supposethat your maxim is in violationof the Formula of Universal Law. You

straightforwardresentationof a case to someonewho is close to or admiresyou, withoutemphasis,withoutstyle, without takingsome sort of advantageof whatever t is about youthat has your listener'sattention n the first place. So how can we avoidthe nonrationalinfluence of others? I take it that most obviously n the Symposium, but also in otherdialoguesconcernedwith the relationof love and teachingsuch as the Phaedrus,Plato isat workon the questionwhether you can use your sex appeal to drawanother'sattentionto the reasons he has for believing or doing things, ratherthan as a distraction hat aidsyourcase illicitly.


15/26


are making an exception of yourself, doing something that everyone inyour circumstances could not do. What this means is that you are treatingthe reason you have for the action as if it were stronger, had morejus-tifying force, than anyone else's exactly similar reason. You are thenacting as if the fact that it was in particularyour reason, and not justthe reason of a human being, gave it special weight and force. This isan obvious violation of the idea that it is your humanity-your power ofrationalchoice-which is the condition of all value and which thereforegives your needs and desires the justifying force of reasons. Thus, anyviolationof the Formulaof UniversalLaw s also a violationof the Formulaof Humanity. This argument, of course, only goes in one direction:itdoes not show that the two formulas are equivalent. The Formula ofHumanityis stricterthan the Formulaof Universal Law-but both areexpressions of the same basic theoryof value: that your rational natureis the source of justifying powerof your reasons, and so of the goodnessof your ends.And although the Formula of Humanity gives us reason to think thatall lies are wrong, we can still give an account in the terms it providesof what vindicates lying to a liar. The liar tries to use your reason as ameans-your honesty as a tool. You do not have to passively submit tobeing used as a means. In the Lectures on Ethics, this is the line thatKant takes. He says:

If we were to be at all times punctiliously truthful we might oftenbecome victims of the wickedness of others who were readyto abuseourtruthfulness.If all men were well-intentioned t would not only bea duty not to lie, but no one would do so because there would be nopoint in it. But as men are malicious, it cannot be denied that to bepunctiliouslytruthful s often dangerous... if I cannot save myself bymaintaining silence, then my lie is a weapon of defence. (LE, 228)

The common thought that lying to a liar is a form of self-defense, thatyou can resist lies with lies as you can resist forcewith force,is accordingto this analysiscorrect.'2This should not be surprising, orwe have seen12. Of course you may also resist force with lies, if resisting t with force is notan optionfor you. This gives rise to a question about whether these options are on a footing witheach other. In many cases, lying will be the better option. This is because when you usecoercion you risk doing injury to the person you coerce. Injuring people unnecessarily swrong, a wrong that should be distinguished rom the use of coercion. When you lie you


16/26

339 TheRight to Lie:Kanton Dealing with Evil

that deception and coercion are parallel.Lying and the use of force areattempts to undercut the two conditions of possibleassent to actions andof autonomous choice of ends, namely, knowledge and power. So, al-though the Formulaof UniversalLaw andthe Formulaof Humanity giveus differentresults, this does not show thatthey simplyexpress differentmoraloutlooks.The relationbetween them is more complex than that.Two CASUISTICALPROBLEMSBefore I discuss this relation,however, I must take up two casuisticalproblems arising from the view I have presented so far. First, I haveargued that we may lie to the murdererat the door. But most peoplethink something stronger:thatwe ought to lie to the murderer-that wewill have done somethingwrongif we do not. Second, I have arguedthatit is permissible to lie to a deceiverin orderto counterthe deception.Butwhat if someone lies to you for a goodend, and, as it happens, you knowabout it? The fact that the murderer'send is evil has played no directrole in the argumentsI have given so far. We have a right to resist liarsand those who tryto use force because of theirmethods, not because oftheir purposes. In one respect this is a virtue of my argument. It doesnot license us to lie to or use violence against personsjust because wethink their purposesare bad. But it looks as if it may license us to lie toliars whose purposes are good. Here is a case: suppose someone comesto your door and pretends to be taking a survey of some sort.'3In fact,this person is a philanthropistwho wants to give his money to peoplewho meet certaincriteria,and this is his way of discoveringappropriateobjectsfor his beneficence.Asit happens, you know what is up. Bylying,

do notriskdoingthisextrawrong. ButKant hinks that ying is in itself worsethancoercion,because of the peculiarlydirect way in which it violatesautonomy.So it shouldfollow thatif you can deal with the murdererby coercion, this is a better option than lying. Othersseem to share this intuition.Cardinal ohn HenryNewman,responding o SamuelJohnson'sclaim thathe would lie to a murdererwho asked which wayhis victimhad gone, suggeststhat the appropriate hing to do is "to knock the man down, and to call out for the police"(ApologiaPro Vita Sua: Being a History of His Religious Opinions[London:Longmans,Green & Co., i88o], p. 36I. I am quoting from Sissela Bok, Lying [New York: VintageBooks, 1979], p. 42). If you can do it without seriously hurtingthe murderer, t is, so tospeak, cleanerjust to kick him off the front porch than to lie. This treats the murdererhimself more like a human being than lying to him does.

I3. I owe this example to John Koethe.


17/26

340 Philosophy &Public Affairs

you could get some money, although you do not in fact meet his criteria.The argumentthat I derivedfrom the Formula of Universal Law aboutlying to the murdererapplieshere. Universalizing the lie to the philan-thropist will not destroy its efficacy. Even if it is a universal law thateveryonewill lie in these circumstances, the philanthropist hinks youdo not know you are in these circumstances. By my argument, it ispermissibleto lie in this case. The philanthropist,ike the murderer,hasplaced himself in a morallyunprotectedposition by his own deception.Start with the first casuisticalproblem.There are two reasons to lie tothe murdererat the door.First, we have a duty of mutual aid. This is animperfect duty of virtue,since the law does not say exactlywhat or howmuch we must do along these lines. This duty gives us a reason to tellthe lie. Whetherit makes the lie imperativedepends on how one under-stands the duty of mutual aid, on how one understands the "wideness"of imperfectduties.'4It may be that on such an urgent occasion, the lieis imperative.Notice that if the lie were impermissible,this duty wouldhave no force.Imperfectduties arealwayssecondarytoperfectones. Butif the lie is permissible, this duty will providea reason, whether or notan imperativeone, to tell the lie.The second reason is one of self-respect.The murdererwants to makeyoua toolofevil; he regardsyour ntegrityas auseful sortofpredictability.He is trying to use you, and your good will, as a means to an evil end.Youowe it to humanityin yourown person not to allowyour honesty tobe used as a resource for evil. I think this would be a perfect duty ofvirtue;Kantdoes not say this specifically,but in his discussionof servility(the avoidanceof which is a perfectdutyofvirtue)he says "Donot sufferyour rights to be trampledunderfootby others with impunity" (MMV,436/99).Both of these reasons spring from duties of virtue. A person with agood characterwill tell the lie. Not to tell it is morallybad. But there isno dutyofjustice totell the lie. If we do not tellit, we cannot be punished,or, say, treated as an accessory to the murder.Kant would insist thateven if the lie ought to be told this does not mean that the punctiliouslytruthfulpersonwho does not tell it is somehowimplicated n the murder.It is the murderer,not the truthful person, who commits this crime.

14. Fora discussionof this questionsee BarbaraHerman,"MutualAid and RespectforPersons,"Ethics 94 (July i984):577-602.


18/26

341 TheRightto Lie:Kanton Dealingwith Evil

Telling the truth cannot be part of the crime. On Kant'sview, personsare not supposed to be responsiblefor managing each other'sconduct.If the lie were a duty of justice, we would be responsiblefor that.These reflections will help us to think about the second casuisticalproblem,the lie to the philanthropist. thinkit does followfromthe lineof argument I have taken that the lie cannot be shown to be impermis-sible. Althoughthe philanthropistcan hardlybe called evil, he is doingsomething trickyand underhanded,which on Kant'sview is wrong. Heshould not use this method of getting the informationhe wants. This isespecially true if the reason he does not use a more straightforwardmethodis that he assumes that if he does, peoplewill lie to him. We arenot supposed to base our actions on the assumptionthat other peoplewill behave badly.Assuming this does not occur in an institutionalcon-text, and you have not sworn that your remarkswere true, the philan-thropistwillhave no recourseto justice if youlie tohim.'5 Butthe reasonsthat favor telling the lie that exist in the first case do not exist here.According o Kant,youdo not haveadutytopromoteyourown happiness.Nor would anyone performsuch an actionout of self-respect.This is, ina very trivialway, a case of dealingwith evil. But you can best deal withit by telling the philanthropist hat you know what he is up to, perhapseven that you find it sneaky. This is because the ideal that makes hisaction a bad one is an ideal of straightforwardnessn human relations.This would also be the best way to deal with the murderer, f it were away to deal with a murderer.But of course it is not.IDEALANDNONIDEAL HEORYI now turn to the questionof what structurean ethical theorymust havein orderto accommodatethis way of thinking. In A Theoryof Justice,JohnRawlsproposesa divisionofmoralphilosophynto idealandnonidealtheory.'6In that work, the task of ideal theoryis to determine "whataperfectlyjust society would be like,"while nonideal theory deals with

15. In the Lectureson Ethics, Kant takes the positionthatyou maylie to someonewholies to or bullies you as long as you don'tsay specifically hatyourwordswill be true. Heclaimsthis is not lying,becausesuch a person shouldnot expectyou to tell the truth.(LE,227, 229)i6. John Rawls,A Theoryof Justice (Cambridge,MA:HarvardUniversityPress, 1971).Sectionand page numbersreferring o this work will appear n the text.


19/26


punishment, war, oppositionto unjust regimes, and compensatory us-tice. (Sec. 2, pp. 8-9) Since I wish to use this featureof Rawls'stheoryfor a model, I am going to sketch his strategy orwhat I will call a double-level theory.Rawlsidentifies two conceptions ofjustice, which he calls the generalconception and the specialconception.(Secs. I I, 26, 39, 46) The generalconceptiontells us that all goodsdistributedby society,includinglibertyand opportunity,are to be distributedequallyunless an unequal distri-bution is to the advantageof everyone,and especially those who fall onthe low side of the inequality.(Sec. 13) Injustice,according othe generalconception, occurs whenever there are inequalities that are not to thebenefit of everyone. (Sec. i i, p. 62) The special conceptionin its mostdevelopedform removes libertyand opportunity rom the scope of thisprincipleand says they must be distributedequally,forbidding radeoffsof these goods foreconomic gains. It alsointroducesa numberof priorityrules, forexample, the priorityoflibertyoverallotherconsiderations,andthe priorityof equal opportunity vereconomicconsiderations. Secs. i I,46, 82)Ideal theory is worked out under certain assumptions. One is strictcompliance:it is assumed that everyonewill actjustly. The other,a littleharder to specify, is that historical,economic, and naturalconditions aresuch that realization of the ideal is feasible. Our conduct towardthosewho do not comply, or in circumstances'which make the immediaterealizationof ajust state ofaffairsmpossible, s governedbytheprinciplesof nonidealtheory.Certainongoingnaturalconditionswhich mayalwaysprevent the full realizationof the ideal state of affairs also belong tononidealtheory:the problemsof dealingwith the seriously ll ormentallydisturbed, for instance, belong in this category. For purposes of con-structingideal theory,we assume that everyoneis "rationaland able tomanage their own affairs."(Sec. 39, p. 248) We also assume in idealtheorythatthere arenomassive historic njustices,such as the oppressionof blacks and women, to be corrected.The point is to work out our idealview of justice on the assumptionthat people, nature, and historywillbehave themselves so that the ideal can be realized,and then to deter-mine-in light of that ideal-what is to be done in actual circumstanceswhen they do not. The special conceptionis not appliedwithoutregardto circumstances. Specialprincipleswill be used in nonidealconditions.Nonidealconditionsexist when, or to the extent that, the'special con-


20/26

343 TheRightto Lie:Kanton Dealing with Evil

ception of justice cannot be realizedeffectively.In these circumstancesour conduct is to be determinedin the followingway: the special con-ception becomes a goal, rather than an ideal to live up to; we are to worktowardthe conditions in which it is feasible. Forinstance, suppose thereis a case like this: widespreadpoverty or ignorance due to the level ofeconomic developmentis such that the legal establishmentof equal lib-erties makes no real differenceto the lot of the disadvantagedmembersof society. It is an emptyformality.On the otherhand, some inequality,temporarilynstituted, would actually tend to fosterconditions in whichequallibertycould become a reality oreveryone.In these circumstances,Rawls's double-level theory allows for the temporary nequality. (Secs.I I, 39) The priorityrules give us guidance as to which features of thespecial conception are most urgent. These are the ones that we shouldbe striving to achieve as soon as possible. For example, if formalequalopportunity orblacks andwomen is ineffective, affirmativeaction meas-ures may be in order. If some people claim that this causes inefficiencyat first, it is neither here nor there, since equality of opportunityhaspriorityoverefficiency. The special conceptionmay also tell us which ofour nonideal options is least bad, closest to ideal conduct. Forinstance,civil disobedience is better than resortingto violence not only becauseviolenceis bad in itself,butbecause of the wayin which civildisobedienceexpresses the democraticprinciplesof the just societyit aspiresto bringabout. (Sec. 59) Finally, the general conception of justice commandscategorically. n sufficientlybadcircumstancesnone of the characteristicfeatures of the special conception may be realizable. But there is noexcuse everfor violationof the generalconception.If inequalitiesare notbenefiting those on the lower end of them in some way, they are simplyoppression.The generalconception, then, representsthe point at whichjustice becomes uncompromising.'7

17. In a nonideal case, one's actions may be guided by a more instrumental style ofreasoning than in ideal theory.Butnonideal heory s notaformof consequentialism.Thereare two reasons for this. One is that the goal set by the ideal is not just one of goodconsequences, but of a just state of affairs.If a consequentialistview is one that definesright action entirely in terms of good consequences (which arenot themselves defined interms ofconsiderations f rightnessorjustice), then nonideal heory s notconsequentialist.The second reason is thatthe ideal will also guide our choice amongnonidealalternatives,importing riteria orthis choice otherthaneffectiveness.I would ike to thankAlanGewirthforpromptingme to clarifymy thoughts on this matter, and DavidGreenstone orhelpingme to do so.


21/26


A double-level theory can be contrastedto two types of single-leveltheory, both of which in a sense fail to distinguish the way we shouldbehave in ideal and nonideal conditions, but which are at oppositeex-tremes. A consequentialist theory such as utilitarianismdoes not reallydistinguish ideal fromnonidealconditions.Of course, the utilitariancansee the difference between a state of affairsin which everyone can bemade reasonablyhappy and a state of affairs in which the utilitarianchoice must be for the "lesserof evils,"but it is still really a matter ofdegree. In principle we do not know what counts as a state in whicheveryone is "as happy as possible" absolutely. Instead, the utilitarianwants to make everyone as happy as possible relative to the circum-stances, and pursues this goal regardless of how friendly the circum-stances are to human happiness.The difference s not between ideal andnonideal states of affairs but simplybetween better and worse states ofaffairs.Kant's heoryas he understoodtrepresentstheotherextremeofsingle-level theory.The standard of conduct he sets for us is designed for anideal stateof affairs:we arealwaystoactas if we werelivingin a Kingdomof Ends, regardless of possible disastrousresults. Kantis by no meansdismissive toward the distressing problemscaused by the evil conductof other human beings and the unfriendlinessof nature to human ideals,but his solutionto these problems s different. He findsin them groundsfor a morallymotivatedreligiousfaith in God.'8Ourrationalmotive forbelief in a moralauthorof the worldderives fromour rational need forgroundsfor hope that these problemswill be resolved. Such an authorwould have designed the laws of nature so that, in ways that are notapparentto us, our moral actions and effortsdo tend to further the re-alizationof an actual Kingdomof Ends. With faith in God,we can trustthat a Kingdomof Ends will be the consequence of our actions as wellas the ideal that guides them.In his Critiqueof Utilitarianism, BernardWilliamsspells out some ofthe unfortunateconsequences of what I amcallingsingle-leveltheories.I9According oWilliams,the consequentialist'scommitmenttodoingwhat-ever is necessary to secure the best outcome may lead to violationsof

i8. See the "Dialecticof Pure PracticalReason"of the Critiqueof PracticalReason,andthe Critiqueof Teleological udgment,sec. 87.I9. BernardWilliams, n UtilitarianismForand Against,by J.J.C.Smart and BernardWilliams(Cambridge:CambridgeUniversityPress, 1973), pp. 75-150.


22/26

345 The Right to Lie: Kanton Dealing with Evil

what we would ordinarilyhinkof as integrity.Thereis no kind of actionthat is so mean or so savage that it can never lead to a better outcomethan the alternatives.Acommitment to always securingthe best outcomenever allows you to say "badconsequences or not, this is not the sort ofthing I do; I am not that sort of person." And no matter how mean orhow savage the act requiredto secure the best outcomeis, the utilitarianthinks thatyou will be irrational o regretthatyou didit, for you will havedone what is in the straightforward ense the right thing.2o A Kantianapproach,by defining a determinate deal of conductto live up to ratherthan setting a goal of action to strive for, solves the problemabout in-tegrity, but with a high price. The advantageof the Kantianapproach sthe definite sphere of responsibility.Your share of the responsibility orthe way the world is is well-definedand limited, and if you act as youought, bad outcomes are not your responsibility.The trouble is that incases such as that of the murdererat the door t seems grotesque simplyto say that I have done my part by telling the truth and the bad resultsare not my responsibility.The pointof a double-level heory s to give us both a definiteandwell-defined sphere of responsibility or everyday ife and some guidance, atleast, about when we may or must take the responsibilityof violatingideal standards. The common-sense approachto this problem uses anintuitive quantitativemeasure: we departfrom our ordinaryrules andstandardsof conduct when the consequences of followingthem wouldbe "verybad."This is unhelpful for two reasons. First; it leaves us onour own about determininghow bad. Second, the attempt to justify itleads down a familiar consequentialist slipperyslope: if very bad conse-quences justify a departure romordinarynorms,why donot slightlybadconsequences justify such a departure?A double-level heorysubstitutessomethingbetterthanthisrough quantitativemeasure.In Rawls'stheory,for example, a departurefrom equal libertycannot be justified by thefact that the consequences of libertyare "verybad" in terms of mereefficiency. This does not mean that an endless amount of inefficiencywill be tolerated,because presumablyat some pointthe inefficiency mayinterfere with the effectiveness of liberty.One might put the point this

20. Williamsalso takes this issue up in "EthicalConsistency,"originallypublishedinthe SupplementaryVolumes to the Proceedingsof the AristotelianSociety XXXIX, 965,and reprinted n his collection, Problemsof the Self (Cambridge:CambridgeUniversityPress, 1973), pp. i66-86.


23/26


way: the measure of "verybad" s not entirely intuitive but rather, badenough to interfere with the realityof liberty. Of course this is not analgorithmic criterionand cannot be applied withoutjudgment, but it isnot as inexact as a wholly intuitive quantitative measure, and, impor-tantly, does not lead to a consequentialistslippery slope.Anotheradvantage of a double-leveltheory is the explanation t offers

of the otherphenomenonWilliams s concerned about:that of regretfordoing a certain kind of action even if in the circumstances it was the"right" hing. A double-leveltheoryoffersan account of at least some ofthe occasions for this kind of regret.We will regrethavingto depart romthe ideal standardofconduct,for we identifywith this standardandthinkof our autonomyin terms of it. Regret for an action we would not dounder ideal circumstancesseems appropriate ven if we have done whatis clearly the right thing.21

KANTIANONIDEALHEORYRawls'sspecialconceptionofjustice is a stricterversion of the egalitarianidea embodied in his general conception. In the same way, it can bearguedthat the Formulaof Universal Law andthe Formulaof Humanityare expressions of the same idea-that humanity is the source of value,and of the justifying force of reason. But the Formula of Humanityisstricter, and gives implausible answers when we are dealing with themisconduct of others and the recalcitranceof nature. This comparisongives rise to the idea of using the two formulasand the relationbetweenthem to construct a Kantian double-leveltheoryof individualmorality,with the advantagesof that sort of account. The Formulas of Humanityand the Kingdomof Ends will providethe idealwhich governsour dailyconduct. When dealingwith evil circumstanceswe may departfrom thisideal. In such cases, we can say that the Formulaof Humanityis in-applicablebecause it is not designed foruse when dealingwith evil. But

21. It is importanthere to distinguish two kinds of exceptions. As Rawls points out in"TwoConceptionsof Rules" The PhilosophicalReview64 [January965]), a practicesuchas promising may have certain exceptions built into it. Everyone who has learned thepractice understands that the obligationto keep the promise is cancelled if one of theseobtains. When one breaks a promise because this sort of exception obtains, regret wouldbe inappropriate nd obsessive. And these sorts of exceptions may occur even in "ideal"circumstances.The kind of exception one makes when dealingwith evil should be distin-guishedfromexceptions built into practices.


24/26

347 The Rightto Lie:Kanton Dealingwith Evil

it can still guide our conduct. It defines the goal toward which we areworking,and if we can generatepriority uleswe will knowwhich featuresofit aremostimportant. t givesus guidanceaboutwhich of the measureswe may take is the least objectionable.Lying to deceivers is not the only case in which the Formulaof Hu-manityseems to set a more ideal standard han the Formulaof UniversalLaw. The argumentsmade aboutlying can all be made aboutthe use ofcoercion to deal with evildoers.Anothervery difficultcase in which thetwo formulas give differentresults, as I think, is suicide. Kant gives anargumentagainst suicide under the Formulaof UniversalLaw, but thatargumentdoes not work.22Yet under the Formulaof Humanitywe cangive a clear and compellingargumentagainst suicide: nothing is of anyvalue unless the human person is so, and it is a greatcrime, as well asa kind of incoherence, to act in a way that denies and eradicates thesource of all value. Thus it might be possibleto say that suicide is wrongfrom an ideal point of view, though justifiablein circumstances of verygreatnaturalor moralevil.There is also another,ratherdifferentsense of "rigorism"n which theFormulaof Humanityseems to be more rigorousthan that of UniversalLaw. It concerns the question whether Kant'stheoryallowsfor the cat-egoryof merelypermissibleends andactions,or whetherwe must alwaysbe doing something that is morallyworthy:that is, whether we shouldalways pursue the obligatoryends of our own perfectionand the hap-piness of others, when no other duty is in the case.The Formula of Universal Law clearlyallows for the categoryof thepermissible.Indeed, the first contradiction est is a test forpermissibility.But in the MetaphysicalPrinciplesof Virtue, there are passages whichhave sometimes been taken to imply that Kant holds the view that ourconduct should alwaysbe informedby morallyworthyends. (MMV,390/48) The textual evidence is not decisive. But the tendency in Kant'sthoughtis certainlythere.For completemoralworth s onlyrealizedwhen

22. Kant'sargumentdependson a teleologicalclaim: that the instinctwhose office is toimpel the improvementof life cannotuniversaliybe used to destroy ife without contradic-tion.(G,422/40) But as I understand he contradictionn conception est, teleological laimshave no real place in it. Whatmatters is not whether nature assigns a certain purpose toa certain motive or instinct, but whether everyone with the same motive or instinct couldact in the way proposedand still achieve their purpose. There is simply no argumenttoshow that everyone sufferingfrom acute misery could not commit suicideand still achievetheir purpose:ending that misery.


25/26


our actions are not merelyin accordancewith duty but fromduty, or, tosay the same thing a differentway, perfect autonomy is only realizedwhen our actions and ends are completelydeterminedby reason, andthis seems to be the case only when ourends arechosen as instantiationsof the obligatoryends.Using the Formulaof Humanityit is possible to argue for the more"rigorous"nterpretation.First, the obligatory nds can be derived morestraightforwardlyrom Humanitythan from Universal Law. Kant doesderive the obligatoryends from the Formula of Universal Law, but hedoes it by a curiouslyroundaboutprocedure n which someone is imag-ined formulatinga maxim of rejecting them and then finding it to beimpermissible.This argumentdoes not show that therewould be a moralfailing if the agent merely unthinkinglyneglected rather than rejectedthese ends. The point about the pervasivenessof these ends in the morallife is a more complicatedone, one that follows from their adoptionbythis route: among the obligatoryends is our own moralperfection.Pur-suing ends that aredeterminedbyreason,ratherthanmerelyacceptableto it, cultivatesone's moralperfection n the requiredway. (MMV,380-81/37-38; 444-47/1 o8-I i i)It is importantto point out that even if this is the correct way tounderstand Kant's ideal theory, it does not imply that Kantianethicscommandsa life of conventionalmoral"gooddeeds."The obligatory ndsare one's own perfectionand the happinessof others;to be governedbythem is to choose instantiations of these largercategories as the aim ofyour vocationand other everydayactivities.It is worth keeping in mindthat naturalperfection s a largecategory, ncludingall the activitiesthatcultivate body and mind. Kant'spoint is not to introduce a strenuousmoralismbut to find a place forthe values of perfectionism n his theory.But this perfectionismwill be a partof ideal theoryif the argumentforit is based on the Formulaof Humanityand cannot be derived rom thatof UniversalLaw.This seems to me a desirableoutcome. Peoplein stul-tifying economic or educationalconditionscannot reallybe expected todevote all their spare time to the cultivation of perfectionistvalues. Butthey can be expected not to do what is impermissible,not to violatetheFormula of UniversalLaw. Here again, the Formulaof Humanityshedslight on the situationeven if it is not directlyapplied: t tells us why itis morally as well as in other ways regrettablethat people should be insuch conditions.


26/26

349 TheRightto Lie:Kanton Dealingwith Evil

CONCLUSIONIf the account I have given is correct, the resources of a double-leveltheory may be availableto the Kantian.The Formula of Humanity andits corollary, he vision of a Kingdomof Ends, providean ideal to live uptoin daily ife as well as a long-termpoliticaland moralgoal forhumanity.Butit is not feasiblealwaysto live up to this ideal, and where the attemptto live up to it would make you a tool of evil, you should not do so. Inevil circumstances, but only then, the Kingdomof Ends can become agoal to seek rather than an ideal to live up to, and this will provideuswith some guidance. The Kantianpriorities-of justice over the pursuitof obligatoryends, and of respect over benevolence-still help us to seewhat matters most. And even in the worstcircumstances,there is alwaysthe Formula of Universal Law, telling us what we must not in any casedo. For whatever badcircumstances may drive us to do, we cannot pos-siblybe justifiedin doingsomethingwhich othersin those same circum-stances could not also do. The Formulaof UniversalLaw providesthepoint at which moralitybecomes uncompromising.Let me close with some reflections about the extent to which Kanthimself might have agreedwith this modificationof his views. Through-out this essay, I have portrayedKantas an uncompromisingdealist,andthere is much to support this view. But in the historical and politicalwritings, as well as in the Lectures on Ethics, we find a somewhat dif-ferent attitude.This seems tome tobe especially mportant:Kantbelievesthat the Kingdomof Ends on earth, the highest politicalgood, can onlybe realized in a condition of peace. (MMJ, 354-55/I27-29) But he doesnot think that this commitsa nation to a simplepacifismthatwouldmakeit the easy victim of its enemies. Instead, he draws up laws of war inwhich peace functions not as an uncompromising deal to be lived up toin the present, but as a long-range goalwhich guides our conduct evenwhen war is necessary. (PP, 343-48/85-9I; MMJ, 343-5I/I 14-25) If aKantian can hold such a view for the conduct of nations, why not forthat of individuals?If this is right, the taskof Kantianmoralphilosophyis to drawup forindividualssomething analogousto Kant's aws of war:special principlesto use when dealingwith evil.

korsgaard the right to lie kant on dealing with evil

Documents