Necessity Knows No Law: On Extreme Cases and Uncodifiable Necessities

Alon Harel* &Assaf Sharon**

‘NECESSITY KNOWS NO LAW’: ON EXTREME

CASES AND UNCODIFIABLE NECESSITIES†

This article analyses the category of extreme cases – cases involving catastrophic con-sequences the avoiding of which requires severe measures (e.g. torture, shooting aplane in 9/11 situations, etc). We first reject two traditional solutions to extremecases: deontology/threshold deontology (as traditionally understood) and consequen-tialist solutions. Our proposal maintains that what is most pernicious is not the viola-tion of moral rules as such but their principled or rule-governed violation.Maintaining a normative distinction between acts performed under the direction ofprinciples/rules, on the one hand, and unprincipled, context-generated acts, actsperformed under the force of circumstances, on the other, allows for accommodatingthe necessity of infringements in extreme cases within a (non-conventional) deontolo-gical framework. Agents who perform acts under extreme cases ought not to be guidedby rules or principles. Instead, they ought to make particular judgments not governedby rules. We also establish that this solution follows from the Kantian conception ofhuman dignity.

Keywords: rules/principles/consequentialism/deontology/extreme cases/emergencies/dignity/Kant

i Introduction

In a recent decision, the German Federal Constitutional Court declared aprominent provision of a new German anti-terrorism law unconstitutionaland void. Section 14 of the Civil Aviation Security Act1 enacted in theBundestag in June 2004 authorized the Minister of Defence to orderthat a passenger airplane be shot down if it could be assumed that theaircraft was being used against the lives of others and if downing it wasthe only means of preventing this present danger. This section was

* Phillip P Mizock and Estelle Mizock Chair in Administrative and Criminal Law, HebrewUniversity Law Faculty, Professor of law, Boston University Law School

** Department of Philosophy, Stanford University† We are grateful to Amir Engel, Yuval Eylon, Michael Otsuka, Ariel Porat, Ram Rivlin,

Jesper Ryberg, and Levi Spectre for their important comments on previous drafts.This paper was presented at the conference on Constitutionalism and the CriminalLaw at the Faculty of Law, University of Toronto, Oxford University legal theoryworkshop, University of Frankfurt in a workshop on rights, the Hebrew Universityconference on criminal law in times of emergency, the Hebrew University politicaltheory workshop, and Roskilde University philosophy department. We are grateful tocomments of participants in these forums.

1 11 January 2005, BGBI 1 at 78.

(2011), 61 UNIVERSITY OF TORONTO LAW JOURNAL DOI: 10.3138/utlj.61.4.845

explicitly drafted with the attacks of 9/11 in mind.2 The GermanConstitutional Court declared the provision unconstitutional on thegrounds that it violates the constitutional right to dignity (article 1 of theGerman Constitution) and the constitutional right to life (article 2 of theGerman Constitution).3 In particular, the Court emphasized that the pro-vision treated the innocent passengers aboard such a plane as objectswithout providing them due legal protection.4 Indeed, something seemsamiss in permitting the killing of innocent civilians in these circumstances.Yet the idea that no action ought to be taken to protect the lives of others –most likely many more in number – seems equally disturbing.

This is but one instance of a large class of policies and measures thathave been vigorously debated anew in the wake of the attacks of 11September 2001. Does the threat of mass suffering and death justifytorture, pre-emptive strikes, invasion of privacy, rendition, and other vio-lations of fundamental rights? How are we to adjudicate such questionsfrom a theoretical moral perspective? Specifically, does the recognitionof the imminence of such threats and the urgency of averting them,even at significant costs, entail that our reasoning concerning suchissues ought to be based solely (as consequentialists argue) or partially(as threshold deontologists suggest) on the consequences? Or can non-consequentialist concerns, such as concerns for fundamental rights, liber-ties, or human dignity, exceed (wholly or partially) the consequentialistcalculations and require us to act in a way that cannot be justified exclu-sively on consequentialist grounds?

These doubts are new expressions of old disputes. The object ofdispute, although probably of broader scope, is often treated in relationto a family of legal and moral phenomena we shall call extreme cases. Theseare cases characterized by radically irregular circumstances, typicallyinvolving catastrophic consequences the avoiding of which requiressevere measures. The leading examples all involve situations in whichactions that are normally regarded illegal and immoral are necessary toprevent great harms, usually the death of many people. Extreme casesevoke a familiar dilemma. On the one hand, there are strong deontolo-gical sentiments – sentiments that suggest that we ought not to torture

2 For a useful summary of the case, see Oliver Lepsius, ‘Human Dignity and the Downingof Aircraft: The German Federal Constitutional Court Strikes Down a Prominent Anti-terrorism Provision in the New Air-transport Security Act’ (2006) 7 German LawJournal 761.

3 BVerfG, 1 BvR 357/05, 59 NJW 2006, 751 [1 BvR 357/05].4 Ibid; an English translation is available at ‘Judgment of the First Senate of 15 February 2006

on the basis of the oral hearing of 9 November 2005: BvR 357/05,’ Entscheidungen,online: Das Bundesverfassungsgericht ,http://www.bundesverfassungsgericht.de/entscheidungen/rs20060215_1bvr035705en.html?Suchbegriff=luftsicherheitsgesetz . .

846 UNIVERSITY OF TORONTO LAW JOURNAL

http://www.bundesverfassungsgericht.de/entscheidungen/rs20060215_1bvr035705en.html?Suchbegriff=luftsicherheitsgesetz



or kill innocent individuals even if this is likely to save lives.5 The prohibi-tions on killing and torturing are grounded in the sacredness of life (orlife’s intrinsic value) that is used sometimes to justify the claim that thelife of a person ought not to be sacrificed even for the sake of savingthe lives of others.6 On the other hand, there is an equally powerful con-viction that when the threat is grave, when consequences might be cata-strophic, lives may have to be sacrificed and rights may have to beinfringed upon. This intuition is often triggered by the invocation ofextreme scenarios involving dirty bombs in urban centres and uncoopera-tive terrorist prisoners. The case of the rogue plane brought before theGerman court triggers the same dilemma. The killing of innocent civi-lians by the state seems morally and constitutionally intolerable. Yet, atthe same time, allowing a plane to crash into a densely populated areais clearly an unacceptable alternative.

Two positions reject the existence of the dilemma entirely, therebyindicating the outer limits of the debate; another traditional view rep-resents an intermediate stance. The first rejects the deontological intui-tion and claims that consequences – and consequences alone – are todetermine what ought to be done in all cases, extreme cases included.The other extreme is occupied by absolutist deontology. This positionrejects the intuition that deontological rules must sometimes be infringedto prevent catastrophic outcomes. According to the absolutist deontolo-gist one is never allowed to kill the innocent or to violate rights, comewhat may. A notable intermediate position between consequentialismand absolutist deontology is threshold deontology. Threshold deontolo-gists maintain that deontological constraints apply so long as the negativeconsequences remain under a certain threshold. Once the threshold isreached, consequentialist considerations should apply.

Part II of this article rejects each of these positions, showing thatneither of the first two recognizes the seriousness of the dilemma ofextreme cases, while the third fails to provide a principled solution forit. Instead, we propose in Part III a new position reconciling the conflict-ing sentiments. Unlike threshold deontology, this solution is not merely acompromise that lies between the extremes of consequentialism andabsolute deontology. Our proposal turns on the idea that, correctly

5 The deontological intuition can also be phrased in terms of rights – agents ought notto violate rights even if doing so is likely to lead to more rights being respected. See e.g.Robert Nozick, Anarchy, State and Utopia (Oxford: Basil Blackwell, 1974) [Nozick]. For ageneral analysis of this feature of rights, see Alon Harel, ‘Theories of Rights’ in MartinP Golding & William A Edmundson, eds, The Blackwell Guide to the Philosophy of Law andLegal Theory (Malden, MA: Blackwell, 2005) 191 at 197–201.

6 Note that we do not maintain that intrinsic value always implies prohibitions of such atype. It is not therefore intrinsic value as such that generates the prohibitions, but thedistinctive characteristics of the intrinsic value of human life.

EXTREME CASES AND UNCODIFIABLE NECESSITIES 847

understood, what is most pernicious for deontology is not merely the viola-tion of moral rules as such but their principled or rule-governed violation.Analysing this distinction, we explicate a conception of deontology onwhich this proposal rests. The fundamental idea of our proposal is therejection of the identification of deontology with exceptionless rules.Deontological constraints, we argue, are not incompatible with exceptions;in fact, exceptions can be required by deontology. Under this understand-ing, deontology is based on the idea, not of absolute, exceptionless moralrules, but rather of unconditional duties and prohibitions. Maintaining anormative distinction between acts performed under the direction of prin-ciples or rules and unprincipled, context-generated acts – acts performednot under the guidance of rules but under the force of circumstances –allows for accommodating the necessity of infringements in extremecases within a deontological framework. The actions required in extremecases ought to be performed strictly as acts of necessity. The agents whoperform them ought not to be guided by general rules or principles.Instead, they ought to make particular judgments not governed by rules.We apply these ideas to the case of rogue planes and the GermanCourt’s decision in Part IV, where we also suggest how the category ofexceptional cases and the distinctive modes of reasoning with which theyought to be treated can be shown to follow from the Kantian conceptionof human dignity. Part V draws some further consequences and con-clusions elaborating on and extending these arguments.

ii Deontology: Categorical and absolute impermissibility

What shall we do when the only way to prevent calamity involves doingwhat would otherwise be clearly and quintessentially wrong? Are we per-mitted, or even required, for example, to shoot down a civilian planethreatening to crash into an urban centre? Within a deontological frame-work, the answer, it seems, is simple. If the prohibition on killing theinnocent is a deontological prohibition, it must always be wrong toshoot down a plane containing innocent civilians. In its ruling, theGerman Constitutional Court expresses this thought:

Such a treatment ignores the status of the persons affected as subjects endowedwith dignity and inalienable rights. By their killing being used as a means to saveothers, they are treated as objects and at the same time deprived of their rights;with their lives being disposed of unilaterally by the state, the persons on boardthe aircraft, who, as victims, are themselves in need of protection, are denied thevalue which is due to a human being for his or her own sake.7

7 See 1 BvR 357/05, supra note 3 at para 122. Note that the deaths of the plane’spassengers are foreseen but unintended effects of the act of preventing the plane


One natural way of understanding this paragraph is as an affirmation of anabsolute deontological prohibition forbidding the downing of a plane underthese circumstances.8 Is it, however, impermissible to down a plane under cir-cumstances in which it threatens the lives of innocent victims? Similar ques-tions have of course been debated by contemporary moral theorists,particularly in the context of deontology.9 It is sometimes thought that con-sequentialists would be more favourably disposed towards the view that it ispermissible to take such action under extreme circumstances, while(some) deontologists will be committed to the position that it is never per-missible to do so. Yet most committed deontologists would oppose theview that it is never permissible to take actions of this sort.

Even orthodox advocates of deontological prohibitions concede thatcertain significant risks warrant the infringement of rights – includingthe right to life.10 A leading strategy for dealing with the dilemma ofextreme cases is what has come to be known as threshold deontology.Thus, most contemporary deontologists agree that deontological

from being used to kill. Yet, the mere fact that this is a foreseen rather than anintentional killing does not imply that there are no serious moral concerns aboutshooting the plane. The case of the plane is similar to the case known in theliterature as the grenade case – a case in which an agent triggers a grenade to stopa trolley, foreseeing that that triggering will kill innocent bystanders. In such a casewe cannot say that the bystanders are used as a means (as their presence, like thepresence of the passengers in the plane, is not necessary for the saving of lives); andyet, it is commonly assumed by deontologists that triggering the grenade isimpermissible; see e.g. Frances Kamm, Morality Mortality, vol 2 (Oxford: OxfordUniversity Press; New York: Oxford University Press, 1996) at 173 [Kamm].

8 Before proceeding let us mention only one important qualification to thisdeontological prohibition. Many deontologists have argued that it is permissible insuch circumstances to kill people who are destined to die in any case for the sake ofsaving others: See Kamm, ibid at 198–200. The Court was not oblivious to thisconcern; see 1 BvR 357/05, supra note 3 at para 130. Yet we are not confident thatthe treatment of the Court is satisfactory. This so-called principle of secondarypermissibility suggests that shooting the plane down may be permissible as thepassengers of the plane are destined to die in any case. To overcome this, we simplymodify the case such that the principle of secondary permissibility is inapplicable.One can assume, for instance, that shooting the plane down will kill a fewunfortunate persons on the ground who would not die otherwise. Alternatively, onecan assume that some of the passengers on the plane would not die unless theplane was shot down or, at least, that the passengers’ chances of survival are higherif the plane is not shot down. Under such circumstances, it is evident that manydeontologists would maintain that shooting the plane is impermissible.

9 See e.g. Shelly Kagan, The Limits of Morality (New York: Oxford University Press, 1989);Samuel Scheffler, ‘Introduction’ in Samuel Scheffler, ed, Consequentialism and Its Critics(New York, Oxford University Press, 1988) 1.

10 The most well-known deontologist who makes this concession is Robert Nozick in hisfamous footnote; see Nozick, supra note 5, at 30, n*; see also Thomas Nagel, MortalQuestions (Cambridge, UK: Cambridge University Press, 1979) at 56.


injunctions can be overridden under certain circumstances. Even if oneconcedes that shooting down a plane carrying fifty passengers in orderto save fifty victims is not justified, the numbers can surely be fiddledwith until an acceptable ratio is achieved. What about shooting downfifty passengers to save 1 000 victims? What about 10 000? And whatabout shooting down two to save 50? The issue surely must not hingeon playing with the numbers. As a matter of principle, there must besome ratio of victims to potential victims that would indeed justify thedowning of the plane.11 This is not merely an abstract observation ofmoral philosophers. The duties to protect are an established componentof many constitutions, including the German Federal Constitution. Thisduty entails a duty to protect the potential victims of a terrorist attack,and such a duty, enshrined in the Constitution, may under certain cir-cumstances require infringing some people’s rights.

Some of the difficulties faced by threshold deontology are familiar andneed not be rehearsed here.12 Let us, however, point out one difficultywhich, to our mind, has not received due emphasis. Threshold deontol-ogy we argue is not faithful to the underlying values and commitments ofdeontology of at least one central brand – Kantian deontology.

The basic challenge faced by threshold deontology is to address thefollowing question: if the life of one person cannot be sacrificed forthe sake of saving one other person, or two, or even one hundred, whycan it be sacrificed to save one thousand, or ten thousand, or onemillion? The natural answer of a Kantian threshold deontologist is tomaintain that, while sacrificing one person to save two or three violatesthe victims’ dignity, the sacrifice of one to save a thousand does notviolate the victims’ dignity. In other words, there is a threshold abovewhich the sacrifice of life does not constitute a violation of dignity.

This answer, however, fails because Kantian deontology opposes notmerely the crude consequentialist metric of exchange; it abhors thevery process of exchanging lives as such. The concept of dignity is atthe foundation of the deontological theories examined here. Yetdignity ‘admits of no equivalent’ (as Kant phrased it)13 and thereforethe willingness to sacrifice one for the sake of two or one for the sakeof ten lives or a thousand lives raises the same concern. Allowing

11 The literature concerning threshold deontology is vast. For a useful review of theliterature, see Eyal Zamir & Barak Medina, ‘Law, Morality and Economics:Integrating Moral Constraints with Economic Analysis of Law’ (2008) 96 Cal L Rev323 at 343–7.

12 See e.g. Eyal Zamir & Barak Medina, Law, Economics and Morality (New York: OxfordUniversity Press, 2010) at 51–6; Larry Alexander, ‘Deontology at the Threshold’(2000) 37 San Diego L Rev 893.

13 Immanuel Kant, Groundwork of the Metaphysics of Morals, ed and translated by Mary JGregor (Cambridge, UK: Cambridge University Press, 1998) at 42 [Groundwork].


consequences to determine whether dignity was violated or not seemsinconsistent with the notion of dignity as understood by Kant. Asdignity ‘admits of no equivalent,’ sacrificing the lives of one in exchangefor a thousand lives is not more respectful of dignity than sacrificing thelives of one in exchange for a hundred. It is the very sacrificing of one lifein exchange for other lives (and not the number of lives sacrificed) thatcompromises dignity.

It follows, therefore, that if threshold deontology is to be justified, it isnot because sacrificing a life of one person for the sake of saving a milliondoes not violate dignity but because the compromising of the dignity ofthe victim is sometimes justified. The threshold deontologist may main-tain that there are two principles at play here: a dignity-based deontolo-gical principle forbidding the killing of innocent people and anotherprinciple mandating the saving of innocent people and the latter prin-ciple may at times override the former. So when the first principle is out-weighed by consequences, it is replaced by a different principle with adifferent rationale; that is, one that is not deontotological and is notdignity-based. This is a possible answer but it remains somewhat ad hocunless it is explained how it is that consequences come to matter. Inother words, if, as deontology seems to assume, consequences do notdetermine rightness and wrongness of actions, why does this changewhen their weight increases? Incorporating consequences in this way cer-tainly seems to deviate from the Kantian framework for which, as dignityforms a foundational value, it must override any other non-foundationalprinciples.

This difficulty seems to us to be mirrored in the literature on thresholddeontology. Threshold deontologists commonly profess that they cannotprovide us with a clear statement of the threshold; they cannot provide us,that is, a rule determining when a threshold has been crossed. Moreover,we are not even provided with a clear list of criteria for making this deter-mination. This renders the idea of a threshold an unprincipled compro-mise between the two sides of the dilemma, since it remains unclear whyconsequences start to matter all of a sudden. We believe that the difficultyof specifying a threshold or even specifying the criteria for what thisthreshold may be is no accidental theoretical shortcoming. It suggeststhat such a determination cannot be made because there is no specifiablereason supporting it and hence no criteria for determining the threshold.In any case, such reason rationalizing the idea of a threshold is not pro-vided by threshold deontologists and conflicts with the most influentialview of deontology – the Kantian tradition.

If setting a threshold for deontological considerations is implausible,how else can deontology accommodate legitimate violations? Or, alterna-tively, how can setting a threshold on what are basically deontologicalconstraints be explained?


iii Necessity knows no law

In his discussion of law in the Summa Theologica, Aquinas addresses casesof emergency in which ‘the observance of [a] law would be hurtful to thegeneral welfare’ and says that in such cases the law ‘should not beobserved.’14 He argues that if ‘the peril be so sudden as not to allow ofthe delay involved by referring the matter to authority, the mere necessitybrings with it a dispensation, since necessity knows no law.’15 What does itmean that ‘necessity knows no law’ and how does this claim justify actingcontrary to law in cases of necessity? Aquinas’s dictum, we propose, canbe interpreted in two ways: an instrumental (pragmatic) way and a prin-cipled way. Under the first, the law is simply not rich enough to capturethe complexity and diversity of circumstances and consequentialistreasoning requires admitting exceptions to the law. As Aquinas says, ‘Alllaw is ordered to the common well-being of men and gains the force oflaw precisely from this fact.’16 On the other hand, the principled, non-consequentialist interpretation requires differentiating between rule-based and non-rule-based behaviour independently of the desirabilityof the actions that result from such reasoning. There are circumstancesin which it is simply wrong to use rule-based reasoning. Aquinas mayhave acknowledged this when he observed, ‘He who in a case ofnecessity acts besides the letter of the law, does not judge the law; butof a particular case in which he sees that the letter of the law is not tobe observed.’17

This ambivalence between pragmatic and principled justifications forresisting rule-based reasoning can also be identified among jurists. Manycontemporary jurists oppose the incorporation of exceptions into legalrules. Often this sentiment is based on the pragmatic concern that bycodifying exceptions one provides overly strong incentives to abuse thecodified exception.18 Yet, in other contexts, the opposition to codificationis grounded in a more principled position, which is expressed in some-what vague or metaphorical terms. Incorporating exceptions into thelaw, it is claimed, is against the ‘genius and spirit of our law’;19 it is

14 St Thomas Aquinas, Summa Theologica, translated by the Fathers of the EnglishDominican Province (Westminster, MD: Christian Classics, 1981) vol 2 q 96, art 6.

15 Ibid.16 Ibid, q 96, art 4.17 Ibid, q 96, art 4.18 See e.g. Richard Posner, Not a Suicide Pact: The Constitution in a Time of National

Emergency (New York: Oxford University Press, 2006) at 85–6; Oren Gross, ‘TheProhibition on Torture and the Limits of Law’ in Sanford Levinson, ed, Torture: ACollection (New York: Oxford University Press, 2004) at 229, 240.

19 See e.g. Jeremy Waldron, ‘Torture and Positive Law: Jurisprudence for the WhiteHouse’ (2005) 105 Colum L Rev 1681 at 1719.


alien to any system of law;20 it threatens to stretch or deforms the law.21

Incorporating provisions designed to address exceptional situations isdestructive to the very spirit of the law and its underlying values. What,for instance, seems particularly offensive to the opponents of legalizingexceptions to the prohibition against torture is the fact that

an effort is also being made to see whether something like torture can be accom-modated within the very legal framework that purports to prohibit it. TheAmerican executive seems to be interested in the prospects for a regime ofcruel and painful interrogation that is legally authorized or at least not categori-cally and unconditionally prohibited . . . An effort is being made to see whetherthe law can be stretched or deformed to actually authorize this sort of thing. Theadministration does not just take the prisoners to the waterboards; it wants todrag the law – our law – along with them.22

Metaphors such as ‘drag[ging] the law . . . to the waterboards’ seem tocapture an important sentiment. Yet these are metaphors, and to providethe basis for a normative argument, metaphors must be unpacked.

To explicate these metaphors, we suggest, it is not enough to invokepragmatic considerations such as the concern that incorporating excep-tions invites abuse and manipulation. Instead, principled considerationsought to be appealed to. Such considerations focus, for example, onthe permissibility and impermissibility of certain forms of reasoning. Inextreme cases, we suggest, the agent not only ought to act differentlythan in non-extreme cases but she also ought to reason differently.

Extreme cases of the type we discuss here constitute a distinctive cat-egory,23 characterized by circumstances that command not only actionswhich are normally impermissible but also forms of reasoning whichare not sanctioned in non-extreme cases. Before we establish this, twocomplementary observations ought to be noted here. First, if a threatarises to the lives of a great number of people, then if one has thepower to prevent it, one ought to do so. The imperative to save life(when this involves no significant risk to oneself) is itself a deontologicalimperative.24 When catastrophe threatens, it does not matter what the lawdictates – whoever can prevent it ought to do so. Second, the actionscharacteristic of extreme cases are such that, when unjustly performed,their being permitted or even ordered by the legal system cannot serveto exculpate the agent. Thus, if a plane does not constitute a genuine

20 Ibid.21 Ibid at 1741.22 Ibid.23 This, of course, is not to deny that there are vague cases.24 If I see a person drowning and can save her (without risking my own life), I ought to do

so, even if I know she will crash her car and kill ten people in the future.


threat to many innocents, if, so to speak, there is no ticking bomb,25 thenthe fact that someone was ordered to down the plane or torture the pris-oner does not justify his actions. Legislation or legal authorization of suchacts plays no justifying role in extreme cases – neither in exculpating theagent when the act is wrongly performed nor in justifying her inactionwhen action is necessary. In extreme cases, therefore, legal directivesshould not affect the agent’s reasoning. An agent capable of acting toprevent a calamity must do so regardless of legal direction and authoriz-ation. And, when no such threat exists or it can be reasonably avoided,the agent must refrain from the action regardless of legislation andauthorization.

This, then, is one sense in which ‘necessity knows no law.’ In general,the law ought to guide the behaviour of individuals by shaping their prac-tical deliberations. When conducting themselves, individuals ought toincorporate the law into their practical reasoning and follow the dictatesthat apply to them. Legislation of extreme cases cannot or ought not beincorporated into one’s reasoning in this way and is therefore inert.Moreover, the codification of extreme cases can also infect reasoning innon-extreme cases. ‘Dragging the law to the waterboards’ is an apt meta-phor for describing the effects of incorporating exceptions into the legalrules because it suggests the spreading of the poisonous effects from theextreme cases to an entire array of cases. What it seems to imply is notonly that it is epistemically impossible to have perfect laws that issuethe correct directives in each and every case. It also suggests that it iswrong to amend our laws in the light of extreme cases (by introducingexceptions), even when this can improve the performance of agents –that is, generate better decisions on their part – since agents ought notto be guided or instructed by such legal rules or, for that matter (as weshow below), by any rules.

These conceptions regarding legislation, we suggest, can be groundedin and explained by moral considerations. Put crudely, the claim is thatthe incorporation of the exception into the law is wrong because it ‘nor-malizes’ the exception and corrupts the law. Codification of extreme casesputs them on a par with other legal directives. The law prohibiting thekilling of innocents and the law permitting it are on a par, separatedmerely by variations of circumstance. But this undermines a crucial differ-ence – only the prohibition on killing innocents is an acceptable prin-ciple. Only the forbidding of torture is an acceptable rule, not itssanction. Such actions are, in principle, wrong. Although particularinstances of these types may constitute unavoidable exceptions, rulesallowing them cannot be accepted as norms. By legislating them,

25 Or, more precisely, if one is not warranted in believing with overwhelming confidencethat such threats exist.


legislators undermine their status as exceptional, as cases that exceed thereach of principled, rule-like directives. And the legal system becomesone that incorporates and places on equal footing directives forbiddingmistreatment of individuals and directives allowing it.

These ideas might have a basis in Kant. In his Lectures on Ethics and hisessay ‘On the Supposed Right to Lie from Philanthropy,’ Kant expressestwo opposite positions regarding the apparently similar cases of therobber at the door and the murderer at the door. In the Lectures onEthics, Kant presents the case of a ‘necessary lie.’ This is a case inwhich a robber appears at one’s door asking whether one’s money is inthe house. In this case, Kant says, ‘the lie is a weapon of defence’ and,hence, allowed.26 In his polemical article ‘On the Supposed Right toLie from Philanthropy,’ he presents the case of a murderer askingabout the whereabouts of an innocent person for the purpose ofkilling him. In such a case, Kant surprisingly says, it is forbidden tolie.27 These apparently irreconcileable positions have triggered awealth of literature proposing solutions, emendations, or criticisms ofKant’s views.28 It is our aim here neither to survey these works nor totake issue with them but merely to present one way of thinking aboutthe issue that brings Kant’s positions into greater accord with oneanother.

In his essay, Kant’s main concern is to reject the idea advanced byBenjamin Constant against his theory of morality, according to whichduties are secondary to rights. That is to say, A has a duty to B to w

only if B has a right to be w’ed by A. Thus, to one who has forfeitedhis right to be told the truth (by threatening to harm another), thereis no duty of truthfulness.29 Countering this view, Kant urges that moralduties are unconditional; thus they cannot depend on there being a cor-responding right. Duties, says Kant, precede rights, for otherwise theywould not be unconditional (as they must if they are to constitutemoral duties at all). The question, then, is what makes a duty conditional.Without going into the intricacies of Kant exegesis, we propose that aproper understanding of this notion requires distinguishing it from an

26 Immanuel Kant, Lectures on Ethics, translated by Peter Heath, ed by Peter Heath & JBSchneewind (New York: Cambridge University Press, 1997) at 204.

27 Immanuel Kant, ‘On a Supposed Right to Lie from Philanthropy,’ in Mary J Gregor, ed,Practical Philosophy, translated by Mary J Gregor (New York: Cambridge University Press,1996) 605 at 611–5 [Kant, ‘Right to Lie’].

28 For some recent accounts see Tamar Schapiro, ‘Kantian Rigorism and MitigatingCircumstances’ (2006) 117 Ethics 32; Allen W Wood, Kantian Ethics (Cambridge, UK:Cambridge University Press, 2007) at 240–58.

29 Benjamin Constant, Des reactions politiques; cited in Kant, ‘Right to Lie,’ supra note 27 at425.


absolute duty. An absolute duty is one to which there can be no excep-tion. An unconditional duty, on the other hand, is a duty that is not con-ditioned by anything contingent such as the agent’s desires or theexpected consequences. The duty to tell the truth irrespective of howmuch the lie may serve the social good is unconditional because itincludes no clauses conditioning it on consequences. The duty not totorture so long as it serves the greater good of society is conditionalbecause it includes such a conditioning clause. Moral duties, Kantclaims, must be categorical and hence unconditional; that is, not con-ditioned on contingencies.

In the Lectures, on the other hand, Kant is concerned not with ques-tions of moral principles and the relation between rights and dutiesbut rather with the substantive moral question of what is to be done incases of necessity. In this context, what is at issue is a practical matter:if one is put in a position where acting according to the dictates ofmoral principles will cause or enable unavoidable harm, can one strayfrom the principles? The question is not about which principles oneought to adopt in the first instance, and Kant nowhere in this discussionsuggests that the duty not to lie should be amended so as not to apply tocases of necessity. The question is whether, given existing moral prin-ciples, there is room for deviation from them. As far as moral principlesgo, duties are unconditional and hence the only acceptable principleregarding lies is that they are forbidden. When it comes to practice,there can arise cases in which this duty ought not to be upheld. Butwhile this can warrant deviation from the rule, it does not warrant itsreplacement by another rule or its being amended.30 This, then, is asecond sense in which ‘necessity knows no law.’

For the Kantian-inspired position just presented, the focus of the dis-cussion is the agent’s reasoning. By incorporating lying as an exception tothe rule that dictates truthfulness – that is, by making the duty of truth-fulness conditional – an agent who considers whether to follow the ruleof truthfulness has to consider the possibility that lying is permissible byour norms. Even when lying is eventually rejected on the grounds that thecircumstances do not call for it, the idea that lying is permissible has beenadmitted into one’s system of moral laws. But under the present proposal,the moral duty dictates not only not to lie but not to consider the possi-bility that lying is permissible by our rules. In other words, the prohibition

30 To be sure, this does not yet alleviate the tension between Kant’s texts, since he (atleast) seems to maintain in the essay that one is forbidden to lie to the murderer atthe door. But as our aim here is not to defend or interpret Kant, we merely suggestthat this may be one way of bringing the two theoretical positions he expressescloser together.


on lying is characterized by non-conditionality. Incorporating rule-gov-erned exceptions into deontological rules, as Kant says, underminestheir unconditional status.31

It follows that there is a compelling reason not to incorporate exceptionsto unconditional duties. But does it also follow that it is never permissible toviolate such duties? If killing innocents is always prohibited by the rules, howcould it ever be permissible to actually do so? How could the duty remainunconditional when it is sometimes permissible to violate it?

iv By force of circumstance

Extreme cases, we argue, require action that is done out of necessity andnot under the direction of law. Under the circumstances of the particularcase, agents ought to act on the basis of the concrete factual necessity, thecondition of things, the force of specific circumstances, the necessity ofthe moment, and similar situation-specific necessitating considerations.These cases resist rule-governed normativity; to put it crudely, beinginherently irregular, they cannot be regulated by general rules. It is theforce of the circumstances – circumstances in light of which (some of)our principles collapse – that necessitate the action, not generalprinciples.32

What makes a case extreme is precisely the definite and common con-viction that it is one where following the laws is unacceptable. The ques-tion then becomes: what is to be done when faced with such a case? Oneoption is to jettison or, at least, amend the law so as to incorporate thedeviant case. But given the Kantian thought about deontological rulesas unconditional, this cannot be the answer. The alternative is to thinkin terms of rules and exceptions. If something is a genuine exception

31 This view contradicts, of course, the views of many contemporary philosophers whodeny that the reasoning of an agent (and, in particular, the agents’ intentions) canever affect the rightness of the act. See e.g. Judith Jarvis Thomson, ‘Self-Defense’(1991) 20 Philosophy and Public Affairs 283. Yet although this view has gainedpopularity in recent years, it is by no means self evident or universally endorsed; seee.g. Jeff McMahan, ‘Intention, Permissibility, Terrorism, and War’ (2009) 23Philosophical Perspectives 345–72.

32 In his discussion of torture, Christopher Kutz, ‘Necessity, Torture and ExistentialPolitics’ (2007) 95 Cal L Rev 235 at 264, develops an idea which bears somesimilarity to this analysis and argues that ‘in the hypothetical, ideal case, the tickingbomb example does confront us with real necessity, not in our imaginations.Confronted by real, existential necessity, we find that our principles yield. But wemust be precise what this means. The image of ourselves torturing, or authorizingtorture, is not a deduction from ethical principles. It is rather a recognition that ourprinciples could imaginably be unable to withstand the pressure from concreteopposing values.’ Kutz concludes that necessity in such a case is not necessity asjustification but necessity as fact.


to a rule, it is not just that the rule doesn’t apply to it but that some other,amended rule does. An exception is an instance of the type to which therule applies but which does not fall under the rule.

To treat an instance as a genuine exception is not to change or replacethe rule. Rather, it requires holding fast to our laws while admitting thatthere are cases when one ought to act beside them. This should not besurprising. As our rules are constrained by the concepts and categorieswe have – which, it is only natural to assume, are in various ways imper-fect – there will always be cases where our laws issue incorrect judgments.Assuming that their shortcomings are unavoidable, changing the laws willnot free us from these imperfections. Given our imperfect conceptualapparatus and ordering of the (factual and normative) world, the bestrules we can have are optimal, not perfect. When encountering a particu-lar case in which our rules seem to issue an incorrect judgment, we some-times change our view of the case and sometimes amend our rules.Sometimes, however, we can do neither. When discrepancies of theextreme sort arise, this does not mean that a different rule is preferableor that our judgment about the case is wrong but rather that we haveexceeded the reach of our rules and must treat the case as an exception.

This idea might seem odd or even irrational, but it seems to us unavoid-able, given deontological assumptions. If deontology is the idea that moralityconsists of law-like (that is unconditional and hence exceptionless) norma-tive judgments and we assume that the imperfection of our cognitive andconceptual abilities entails that perfect normative judgments exceed ourreach, then even if we reach the best laws attainable by us, there will stillremain cases in which the judgments resulting from them will not becorrect. If this is true, then the proper response when such a case arisesis not to change our laws – which, per hypothesis, were the best we canhave – but rather to deviate from them in that instance.33

We may conclude that, from a deontological point of view, there is adistinction to be made between performing an unlawful or impermissibleact and doing so under a rule or a principle permitting such acts. On thebasis of this claim, it is possible to articulate a conception of action inextreme circumstances that is respectful of deontological tenets yet atten-tive to the gravity of the consequences. So long as it is performed withoutguidance of a permitting principle, an action aimed at saving people’slives, for example, may be in violation of a valid law or moral requirementwhile leaving its deontological status intact. To get a clearer view of this,consider again the reasoning of the agent. When downing a rogue plane,for example, the agent may reason from the normative premise that thereis a principle directing him to down a plane threatening many lives and

33 We in no way suggest that this is Kant’s argument or even that this is consistent with hisconception of morality, specifically in the Grundlegung (Groundwork, supra note 13).


from the factual premise that a particular plane does threaten thatnumber of lives to arrive at the conclusion that he ought to shoot itdown. Such reasoning clearly involves a principle according to whichshooting at civilians is permissible in cases in which the ratio of thosekilled and those saved is appropriate. If such principled exceptions,however, are excluded, the agent will reason in another way. He willstart with the premise that saving innocent lives (at least insofar as he isin a position to do so) is his duty and, recognizing the factual truththat the only means to achieve this is by downing the plane, concludethat this is what he ought to do. Notice that in this mode of reasoning,no principle permitting the harming of innocents is employed.34

On the face of it, this may seem like mere sophism, an empty analyticdistinction carrying no genuine moral weight. But this appearance, webelieve, is misleading. Within a proper deontological conception, thereasoning of the agent is a key factor in the determination of themoral status of his actions. This, we take it, is what Kant means whenhe says ‘an action from duty . . . does not depend upon the realizationof the object of the action but merely upon the principle of volition inaccordance with which the action is done.’35 Thus, an action’s moralstatus can differ depending on whether it is done as a permissible actunder the rules or as an exceptional one. Specifically, there is moral sig-nificance in the difference between harming innocents as a principled,permissible act and doing so as an exceptional, unavoidable measure inorder to save lives.

This analysis, we propose, is grounded in a key concern shaping deon-tology. Kant famously voices this concern in the Groundwork of theMetaphysics of Morals when he claims that ‘everything has either a priceor a dignity. What has a price can be replaced by something else as itsequivalent; what on the other hand is raised above all price and thereforeadmits of no equivalent has a dignity.’36 This conception of dignity asincommensurable need not entail its absolute inviolability. A morereasonable interpretation of Kant’s thought, we suggest, is that dignityis not to be considered as a currency to be traded, as a replaceableresource to be weighed against equivalents. Deontological morality, weurge, consists in an aversion to the treatment of human life as having a

34 For Kant, this should not be described in terms of means–ends reasoning but asrendering the action itself one of saving lives. For, as Kant says, a categoricalimperative ‘represents an action as objectively necessary and makes it necessary notindirectly, through the representation of some end that can be attained by theaction, but through the mere representation of this action itself (its form), andhence directly’; Immanuel Kant, The Metaphysics of Morals, ed and translated by MaryJ Gregor (Cambridge, UK: Cambridge University Press, 1996) at 15.

35 Kant, Groundwork, supra note 13.36 Ibid at 42.


‘market value,’ as Kant puts it, as capital to be weighed and traded.Human life and (perhaps) other comparably fundamental values mustnot be treated as goods to be maximized, as assets that can be quantitat-ively measured and traded against other goods. The weighing of livesagainst lives is what, under this conception, the deontologist primarilyaims to avoid. Fundamental values and rights must shape the way wedeliberate by setting constraints on the ends we may morally endorse.37

Harming the innocent can never be justified, although it may be an una-voidable consequence of a necessary action.38

In extreme cases, the urgent need to avoid catastrophic consequencesclashes with the stringency of deontological constraints. Absolute deon-tology sacrifices the former element to uphold the latter.Consequentialism and threshold deontology, on the other hand, bothtry to resolve the conflict by administering a procedure of comparisonbetween the competing considerations in each case. But if they can beweighed against consequences, deontological constraints are no longerunconditional and the values underlying them are mere instruments,goods of ‘market value’ not ‘dignity.’ Thus, by entering the weighinggame, both positions (albeit in different ways) sacrifice the gist of deon-tology – the idea of treating human beings as ends in themselves and ofconstraints on deliberation and action stemming from their intrinsicvalue or ‘dignity’ as unconditional. The alternative sketched here, onthe other hand, allows for addressing the urgency of avoiding cata-strophic consequences while preserving these deontological sensitivities.Extreme cases might require practical infringements of deontologicalnorms, but do not warrant (nor require) their normative perversion.Extreme cases truly are exceptional.

Let us explore one possible objection to our proposal. According tothe deontological conception proposed here, there are circumstancesunder which planes ought to be downed. And, of course, agents arecalled upon to shoot the planes under such circumstances and not toshoot them under different circumstances: circumstances in which therule prohibiting the downing of planes applies. It follows that agentsshould identify those circumstances and judge whether the exceptionalcircumstances prevail. To do so, it seems that that the agent ought to con-struct a rule identifying the exceptional circumstances.39 But constructing

37 Similar aversion to reasoning by weighing is expressed by Nozick, supra note 5, andechoed by Rawls in John Rawls, A Theory of Justice, revised ed (New York: OxfordUniversity Press, 1999) at 3.

38 Note that this essay focuses exclusively on Kantian dignity, which as Kant himselfexplicitly maintains cannot be traded; it is not necessarily a general feature of otherintrinsic goods.

39 Jesper Ryberg (see acknowledgements at note † supra) has raised the objection thateven if ex ante the agent ought not to use rules, courts ex post ought to use such


such a rule and acting on its basis are precisely what we argue isimpermissible.

Our argument denies the second premise of this argument.Exceptional cases ought to be identified by the agent but their identifi-cation does not have to be based on rules. In fact, our claim has beenthat there cannot be a rule specifying what is to count as an exception,for then it wouldn’t be an exception. To claim that the distinctionbetween standard cases, to which the rules apply, and exceptionalcases, to which they don’t, must itself be based on rules falls into familiardifficulties. One need not rehearse the complex arguments that ledWittgenstein to the conclusion that ‘when I obey a rule I do notchoose. I obey the rule blindly.’40 It suffices to note that, if determiningwhether and how a rule applies to a case requires the application of arule, we are faced with a vicious regress of rules. This evident difficultyled Kant to conclude that ‘general logic contains and can contain norules for judgment . . . If it sought to give general instructions how weare to subsume under these rules, that is to distinguish whether some-thing does or does not come under them, that could only be by meansof another rule. This in turn, for the very reason that it is a rule, againdemands guidance from judgment.’41 In fact, even when rules can bearticulated for our judgments, we often reason on the basis of particularjudgments, not rule-like generalizations. Just as the mathematical calcu-lation of the ballistics of a thrown ball are a redescription of ourmental process when catching a ball, not properly of what we do, sothe description of reasoning and judgment in terms of rules is sometimesa redescription. In short, the claim that the identification of extremecases requires rule-governed judgments is, at best, in need of argument,an argument we do not see forthcoming. Thus, we see no difficulty inclaiming that a judgment that some rule is inapplicable is not one thatis guided by rules; it is simply made without the guidance of rules.42

rules. The rules determined by judges will inevitably guide the agents’ actions in thefuture. We deny however that courts must use rules in such cases. Their reasoning inour view ought to resemble that of the agent although, of course, their conclusionmay differ.

40 Ludwig Wittgenstein, Philosophical Investigations, translated by GEM Anscombe (Oxford:Basil Blackwell, 1953) at 219.

41 Immanuel Kant, Critique of Pure Reason, translated by Kemp Smith (Houndmills, UK:Macmillan, 1995) at B172.

42 This claim does not presuppose that the judgment whether the rule applies or not iseasy or that the correct answer is evident or that mistaken decisions cannot occur. It isperhaps somewhat consoling for those who are concerned about moral errors to pointout that an error may be excused at times when the judgment, although mistaken, is areasonable one.


v Rogue planes

We are ready now to apply these observations to the case of the plane.Downing a passenger plane is the sort of action that is to be done onlywhen the specific circumstances of the case are such as to make it practi-cally necessary. An agent may only properly perform such actions whenshe acts merely on the basis of the brute necessity of the circumstances.If the performance is based on the dictates of law or on the decision ofa superior hierarchy authorized by the state, it is no longer performedstrictly on the basis of the necessity of the moment; it becomes part ofregular institutional practice. The legal system recognizes necessity as ajustification; it exempts individuals from legal responsibility when theyact under the circumstance of necessity. But acting on the basis of neces-sity is dictated by circumstances, not by authorization or specific direc-tives. If it is the right thing to do, downing a passenger plane is to bedone regardless of directives, statutes, or rules.

Once legislation authorizing the downing of the plane is reinstated, itpermeates the actions of agents acting under its direction even in cases inwhich these agents judge (or ought to judge) that the plane ought not tobe downed. Downing the plane presents itself as a legitimate option thatsometimes ought to be chosen and, at other times, rejected. Regardless ofwhether or not the act of shooting down a passenger plane underextreme circumstances constitutes a violation of the passengers’ dignity,the endorsement of such a norm treats the duty not to hurt innocentsas conditional and thereby can be said to fail to respect the passengers’dignity. When, on the other hand, the act of downing a plane in excep-tional circumstances is dictated by the force of the circumstances ratherthan governed by a rule, one does not act under the directive of law orunder authorization and thus one is not incorporating a principle accord-ing to which the lives of the passengers are dispensable or exchangeable.Clearly, downing a plane can be permitted only in circumstances of gravenecessity. But – and this is the main point – it can be permitted only asan act of necessity, as an act performed strictly from the necessity of thecircumstances and not under the direction of any rules orauthorizations.43

43 Does this analysis justify the decision of the German Court? Arguably, it could bemaintained, the commander who orders the shooting could be making the one-off,non-rule governed choice which he then transmits to the soldier ordering him todown the plane. In such a case, it could be argued, the passengers’ dignity is notviolated as the shooting is not guided by rules. The decision maker who ordered theshooting was reasoning in the very same way required by our analysis, and it seemstherefore that the shooting is permissible and therefore that the Court ought not tostrike down the law. In our view, this idea is misleading as the soldier who shoots theplane is following orders; i.e., he is operating under a rule requiring him to follow


This analysis is grounded in an understanding of deontology as consist-ing of constraints on moral deliberation; namely, that human lives oughtnot to be weighed against each other. Under normal circumstances, thereis no weighing of lives against lives because downing the plane is not alegitimate option. In exceptional circumstances, there is also no weighingof lives against lives because the shooting is grounded in the urgent needto save lives, which is not guided by any metrics of weighing one lifeagainst another. The rule prohibiting the downing of a plane knows noexceptions, but there are exceptional cases beyond the rule.

vi Consequences and conclusions

We have argued that extreme cases constitute a distinct normative cat-egory. This distinctness, stemming from the exceptionality of the circum-stances, figures into the forms of reasoning appropriate for them andrestricts the types of considerations that may enter into such reasoning.Our leading example thus far has been the case of rogue planes. Butthe analysis applies to numerous current debates including torture, tar-geted assassinations, and other similar acts.

Torture is strictly forbidden. For some deontologists, no principles per-mitting torture can be acceptable. And yet when a ticking-nuclear-bombscenario is invoked, even a deontologist (at least of the brand we havebeen defending) will recognize that torture may be unavoidable.Extreme cases require extreme measures. Violations of our most funda-mental norms may be unavoidable. But – and this is the crucial point –this does not entail a rejection or modification of our basic rules. Whatallows the torture is solely the necessity of avoiding catastrophe, not adifferent law allowing torture under some conditions. Other consider-ations ought not to be incorporated into the determination of what isto be done in such cases. In particular, long-term considerations suchas deterrence and punishment or institutional considerations such aslegal authorization should carry no weight in extreme cases.44 When

such orders. The case of a soldier following orders made by a commander who is notguided by rules is therefore a case in which the shooting itself is guided by rules and istherefore impermissible. We thank Michael Otsuka for raising this objection.

44 Ironically, it is the leading classical theorist of English constitutional law whounderstood that in exceptional circumstances, no relevant distinctions betweenauthorized and unauthorized individuals ought to be made. In his discussion ofmartial law, Albert Venn Dicey, Introduction to the Study of the Law of the Constitution(London: Elibron Classics, 2005) at 506, argues that officials can use brutal means toprotect the peace, including the infliction of instant punishment and, ‘if need be,put to death persons aiding and abetting the enemy or refusing such aid to theEnglish army’; Dicey however continues and says, ‘Let it too be noted that what istrue of a general holds good of every loyal subject according to his situation and the


torture, for instance, is necessitated by extreme circumstances, it is to beperformed as an act of self-preservation, not as an act of punitive justiceor as the implementation of state policy.45

Doubtlessly, articulating rough-and-ready identity conditions for extremecases is difficult, maybe even impossible. Nevertheless, although the needto determine whether a certain situation counts as extreme or not may arise(when an actor’s decision is contested in court, for example), and acknowled-ging that hard cases are unavoidable, we venture to speculate that these will bethe exception, not the rule. Given the nature of extreme cases – they are, afterall, extreme – the common cases will be clear-cut. For, if a situation is notclearly extreme, it will normally not constitute an exception. For the mostpart, at least, extreme cases fit the label, ‘you know it when you see it.’46 Inany case, we should not let this epistemic difficulty undermine the moral cat-egory of extreme cases. Clearly, no general detailed criteria can be given forwhat constitutes such cases and vague and borderline cases are undoubtedlypossible. But these shortcomings are unavoidable in matters such as these andneed not detract from the plausibility of the present account.

One significant upshot of this is that extreme cases have little to teachus about the norms we should embrace and policies we should institute.Contemporary debates regarding the legal norms that are to guide the‘war on terror’ (most notably the question of torture) have been greatlyshaped by examples falling under the category of extreme cases (suchas ticking-bomb scenarios). The proposal on offer, however, entails thatdeliberations about these issues – important and pressing as they mayor may not be – should not be swayed by the intuitions invoked byextreme cases. This is not merely due to their rarity or mismatch withreality but because their inherent exceptionality renders them irrelevantfor questions of institutional norms and general instruction.47

authority which he derives from it, e.g., subordinate officer, of a magistrate or even of aprivate citizen who is helping to resist an invader’; ibid.

45 For more on the irrelevance of the moral status of the potential victim of torture(particularly his or her culpability in creating the threat) and the illegitimacy ofconsiderations of punitive justice which, we suspect, often creep into the argumentsand deliberations regarding torture, see Alon Harel & Assaf Sharon, ‘What’s ReallyWrong with Torture?’ (2008) 6 Journal of International Criminal Justice 241 at 256.

46 This phrase was famously used by the court with respect to pornography; See Jacobellis vOhio, 378 US 184 (1964) at 197 (Stewart J, concurring): ‘I shall not today attemptfurther to define the kinds of material I understand to be embraced within thatshorthand description [‘hard-core pornography’]; and perhaps I could neversucceed in intelligibly doing so. But I know it when I see it, and the motion pictureinvolved in this case is not that.’

47 David Luban, ‘Liberalism, Torture, and the Ticking Bomb’ (2005) 91 Va Law Rev 1425at 1445–52, argues persuasively that the conditions assumed in ticking bomb scenariosare hardly, if ever, met in reality.


To conclude, we have tried to re-examine the limits of rule-governedbehaviour and argued that recognizing these limits may help in maintain-ing a non-fundamentalist deontological morality attentive to the moralimport of emergencies. More specifically, we argued that while traditionaldeontologists characterize deontology in terms of the special strength oreven absoluteness of moral directives, a closer examination of the intui-tions giving rise to deontological convictions suggests that what actuallycharacterizes deontology is the unconditional nature of moral rulesrather than their special or absolute strength. This conception of deon-tology shapes a moral stance with regard to extreme cases allowing forthe resolution of the dilemma they trigger. Following this reasoning toits further ramification, we explicated the sense in which extreme casesconstitute a unique category.


Necessity Knows No Law: On Extreme Cases and Uncodifiable Necessities

Documents

Transcript of Necessity Knows No Law: On Extreme Cases and Uncodifiable Necessities