上海交通大学：《法律与道德》课程教学资源（案例）洞穴奇案50周年续写，哈佛法律评论原文

I999] THE SPELUNCEAN EXPLORERS 885 $I2-A does not apply to those who kill in self-defense,even though there is no express statutory provision making self-defense a legally sufficient justification.Our conclusion to this effect is based not on literal language,but on the (literal)absurdity of not allowing self- defense to count as a justification.Those justices who purport to be "textualists"here are running afoul of well-established law;I cannot believe that they would remain "textualists"in a genuine case of self- defense. Nor is it clear that the statute would apply,for example,to a police officer(or for that matter a private citizen)who kills a terrorist to pro- tect innocent third parties-whether or not there is an explicit provi- sion for justification or excuse in those circumstances.Where the killing is willful,but necessary to prevent a wrongdoer from causing loss of innocent life,a mechanical or literal approach to this statute would make nonsense of the law.A statute of this breadth creates a risk not of ambiguity,but of excessive generality-the distinctive sort of interpretive puzzle that arises when broad terms are applied to situations for which they could not possibly have been designed and in which they make no sense. A possible response would be that to promote predictability,exces- sive generality should not be treated as a puzzle at all;we must follow the natural meaning of the words,come what may.But as I have sug- gested,our self-defense jurisprudence makes this argument unavailable in the current context.But put the precedents to one side.In ordinary parlance,people routinely counteract excessive generality,and thank goodness for that.For example,a parent may tell his child:"Do not leave the house under any circumstances!"But what if there is a fire? A judge may tell his law clerk:"Do not change a single word in this opinion!"But what if by accident,the word "not"was (not?)inserted in the last sentence?Interpreting statutes so as to avoid absurdity could not plausibly undermine predictability in any large-scale or global sense.Nor is it clear that absurdity would be corrected by the legislature before or after the fact.Whether the legislature would cor- rect the absurdity is an empirical possibility,and it is no more than that.Even the most alert people have imperfect powers of foresight, and even the most alert legislature cannot possibly anticipate all appli- cations of its terms. I conclude that when the application of general language would produce an absurd outcome,there is a genuine puzzle for interpreta- tion,and it is insufficient to invoke the words alone.The time- honored notion that criminal statutes will be construed leniently to the criminal defendant strengthens this point.I am therefore unwilling to adopt an approach that would,in all cases,commit our jurisprudence to taking statutory terms in their ordinary sense. HeinOnline--112 Harv.L.Rev.1885 1998-1999

THE SPELUNCEAN EXPLORERS § 12-A does not apply to those who kill in self-defense, even though there is no express statutory provision making self-defense a legally sufficient justification. Our conclusion to this effect is based not on literal language, but on the (literal) absurdity of not allowing self￾defense to count as a justification. Those justices who purport to be "textualists" here are running afoul of well-established law; I cannot believe that they would remain "textualists" in a genuine case of self￾defense. Nor is it clear that the statute would apply, for example, to a police officer (or for that matter a private citizen) who kills a terrorist to pro￾tect innocent third parties - whether or not there is an explicit provi￾sion for justification or excuse in those circumstances. Where the killing is willful, but necessary to prevent a wrongdoer from causing loss of innocent life, a mechanical or literal approach to this statute would make nonsense of the law. A statute of this breadth creates a risk not of ambiguity, but of excessive generality - the distinctive sort of interpretive puzzle that arises when broad terms are applied to situations for which they could not possibly have been designed and in which they make no sense. A possible response would be that to promote predictability, exces￾sive generality should not be treated as a puzzle at all; we must follow the natural meaning of the words, come what may. But as I have sug￾gested, our self-defense jurisprudence makes this argument unavailable in the current context. But put the precedents to one side. In ordinary parlance, people routinely counteract excessive generality, and thank goodness for that. For example, a parent may tell his child: "Do not leave the house under any circumstances!" But what if there is a fire? A judge may tell his law clerk: "Do not change a single word in this opinion!" But what if by accident, the word "not" was (not?) inserted in the last sentence? Interpreting statutes so as to avoid absurdity could not plausibly undermine predictability in any large-scale or global sense. Nor is it clear that absurdity would be corrected by the legislature before or after the fact. Whether the legislature would cor￾rect the absurdity is an empirical possibility, and it is no more than that. Even the most alert people have imperfect powers of foresight, and even the most alert legislature cannot possibly anticipate all appli￾cations of its terms. I conclude that when the application of general language would produce an absurd outcome, there is a genuine puzzle for interpreta￾tion, and it is insufficient to invoke the words alone. The time￾honored notion that criminal statutes will be construed leniently to the criminal defendant strengthens this point. I am therefore unwilling to adopt an approach that would, in all cases, commit our jurisprudence to taking statutory terms in their ordinary sense. 1999] 1885 HeinOnline -- 112 Harv. L. Rev. 1885 1998-1999

r886 HARVARD LAW REVIEW [Vol.1I2:1834 II As I will suggest,the key to this case lies in showing that the best argument for the defendants is unavailable,because a conviction here would not be analogous to a conviction in the most extreme or absurd applications of the statutory terms.But before discussing that point,I pause to deal with some alternative approaches.Troubled by a con- viction in this case,the defendants and several members of this Court have urged some creative alternatives.It is suggested,for example, that under the extreme circumstances of the collapse of the cave opening,the law of civil society was suspended and replaced by some kind of law of nature.See supra,at 1855 (Foster,J.).To the extent that this argument is about a choice of law problem,I do not accept it. There is no legitimate argument that the law of some other jurisdiction applies to this case,and I do not know what is meant by the idea of the "law of nature."The admittedly extreme circumstances themselves do not displace the positive law of this state.Extreme circumstances are the stuff of hard cases,and what makes for the difficulty is the ex- treme nature of the circumstances,not anything geographical.The question is what the relevant law means in such circumstances,and to say that the law does not"apply"seems to me a dodge.The view that extreme circumstances remove the law's force is a conclusion,not an argument. Nor is this a case in which a constitutional principle,or a principle with constitution-like status,justifies an aggressive construction of the statute so as to make it conform to the rest of the fabric of our law. When a statute poses a problem of excessive generality,a court may properly avoid an application that would raise serious problems under the Constitution,including,for example,the Equal Protection Clause, the First Amendment,or the Due Process Clause.If a legislature in- tends to raise those issues,it should be required to focus on them with some particularity.Though it cuts in a different direction from the "plain meaning"idea,this principle is also a close cousin of the void- for-vagueness doctrine,designed to require legislative,rather than merely judicial,deliberation on the underlying question.But there is no such question here. Several members of this Court emphasize the "purpose"of the law. See,e.g.,supra,at 1857 (Foster,J.).They claim that the defendants should not be convicted because while their actions fall within the statute's letter,they do not fall within its purpose.I have considerable sympathy for this general approach,which is not terribly far from my own,and I do not deny that purpose can be a helpful guide when statutory terms are ambiguous.Statutes should be construed reasona- bly rather than unreasonably,and when we do not know what statu- tory terms mean,it is legitimate to obtain a sense of the reasonable HeinOnline--112 Harv.L.Rev.1886 1998-1999

HARVARD LAW REVIEW II As I will suggest, the key to this case lies in showing that the best argument for the defendants is unavailable, because a conviction here would not be analogous to a conviction in the most extreme or absurd applications of the statutory terms. But before discussing that point, I pause to deal with some alternative approaches. Tr-oubled by a con￾viction in this case, the defendants and several members of this Court have urged some creative alternatives. It is suggested, for example, that under the extreme circumstances of the collapse of the cave opening, the law of civil society was suspended and replaced by some kind of law of nature. See supra, at 1855 (Foster, J.). To the extent that this argument is about a choice of law problem, I do not accept it. There is no legitimate argument that the law of some other jurisdiction applies to this case, and I do not know what is meant by the idea of the "law of nature." The admittedly extreme circumstances themselves do not displace the positive law of this state. Extreme circumstances are the stuff of hard cases, and what makes for the difficulty is the ex￾treme nature of the circumstances, not anything geographical. The question is what the relevant law means in such circumstances, and to say that the law does not "apply" seems to me a dodge. The view that extreme circumstances remove the law's force is a conclusion, not an argument. Nor is this a case in which a constitutional principle, or a principle with constitution-like status, justifies an aggressive construction of the statute so as to make it conform to the rest of the fabric of our law. When a statute poses a problem of excessive generality, a court may properly avoid an application that would raise serious problems under the Constitution, including, for example, the Equal Protection Clause, the First Amendment, or the Due Process Clause. If a legislature in￾tends to raise those issues, it should be required to focus on them with some particularity. Though it cuts in a different direction from the "plain meaning" idea, this principle is also a close cousin of the void￾for-vagueness doctrine, designed to require legislative, rather than merely judicial, deliberation on the underlying question. But there is no such question here. Several members of this Court emphasize the "purpose" of the law. See, e.g., supra, at 1857 (Foster, J.). They claim that the defendants should not be convicted because while their actions fall within the statute's letter, they do not fall within its purpose. I have considerable sympathy for this general approach, which is not terribly far from my own, and I do not deny that purpose can be a helpful guide when statutory terms are ambiguous. Statutes should be construed reasona￾bly rather than unreasonably, and when we do not know what statu￾tory terms mean, it is legitimate to obtain a sense of the reasonable 1886 [VOL. 112:1834 HeinOnline -- 112 Harv. L. Rev. 1886 1998-1999

I999] THE SPELUNCEAN EXPLORERS I887 goals that can be taken to animate them and to interpret them in this light. But there are two problems with making purpose decisive here. First,there is no ambiguity in the statutory terms;when text is clear, resort to purpose can be hazardous.Second,the purpose of any stat- ute can be defined in many different ways and at many levels of gen- erality;and at least in a case of this kind,it is most unclear which characterization to choose.Is the purpose of this statute to reach any intentional killing?Any intentional killing without sufficient justifica- tion?Any intentional killing not made necessary by the circum- stances?To reach willful killings while at the same time limiting judi- cial discretion?To make the world better on balance?Any answer to these questions will not come from the statute itself;it is a matter not of excavating something but of constructing it.Where the statute is not ambiguous,we do best to follow its terms,at least when the out- come is not absurd.It is that question to which I now turn. Ⅲ Thus far,I have urged a particular view of this case:the statute contains no linguistic ambiguity.At most,the statute raises the dis- tinctive interpretive problem created by excessive generality.We have long held that self-defense is available by way of justification.It is unclear whether -and we need not decide whether-the statute would or should be inapplicable to some other cases in which a life was taken "willfully"in order to prevent the death of innocent people. For purposes of analysis let us assume,without deciding,that the stat- ute would and should not be so applied.The question then is whether this case is sufficiently like such cases.If it is,then we will have to reach the difficult question of whether an exemption would be allowed in those extreme cases. In cases that seem to raise a problem of excessive generality,it is often useful to proceed by identifying the exemplary or prototypical cases,that is,the cases that are most clearly covered by the statute.I do not mean to suggest that a statute's reach is limited to such cases; generally it is not.But an identification of the prototypical or exem- plary cases can help in the decision whether an application is so far afield as to justify an exemption.The exemplary or prototypical cases within the purview of this statute include those of willful killing of an innocent party,motivated by anger,greed,or self-interest.It is also possible to imagine cases that are at an opposite pole but that seem covered by the statute's literal language:when a defendant has killed someone who has jeopardized the defendant's own life,we have a le- gally sufficient justification under our law,no matter what the statute literally says.And why would cases of this kind be at the opposite pole?The answer is that,in such cases,the victim of the killing is HeinOnline--112 Harv.L.Rev.1887 1998-1999

THE SPELUNCEAN EXPLORERS goals that can be taken to animate them and to interpret them in this light. But there are two problems with making purpose decisive here. First, there is no ambiguity in the statutory terms; when text is clear, resort to purpose can be hazardous. Second, the purpose of any stat￾ute can be defined in many different ways and at many levels of gen￾erality; and at least in a case of this kind, it is most unclear which characterization to choose. Is the purpose of this statute to reach any intentional killing? Any intentional killing without sufficient justifica￾tion? Any intentional killing not made necessary by the circum￾stances? To reach willful killings while at the same time limiting judi￾cial discretion? To make the world better on balance? Any answer to these questions will not come from the statute itself; it is a matter not of excavating something but of constructing it. Where the statute is not ambiguous, we do best to follow its terms, at least when the out￾come is not absurd. It is that question to which I now turn. Thus far, I have urged a particular view of this case: the statute contains no linguistic ambiguity. At most, the statute raises the dis￾tinctive interpretive problem created by excessive generality. We have long held that self-defense is available by way of justification. It is unclear whether - and we need not decide whether - the statute would or should be inapplicable to some other cases in which a life was taken "willfully" in order to prevent the death of innocent people. For purposes of analysis let us assume, without deciding, that the stat￾ute would and should not be so applied. The question then is whether this case is sufficiently like such cases. If it is, then we will have to reach the difficult question of whether an exemption would be allowed in those extreme cases. In cases that seem to raise a problem of excessive generality, it is often useful to proceed by identifying the exemplary or prototypical cases, that is, the cases that are most clearly covered by the statute. I do not mean to suggest that a statute's reach is limited to such cases; generally it is not. But an identification of the prototypical or exem￾plary cases can help in the decision whether an application is so far afield as to justify an exemption. The exemplary or prototypical cases within the purview of this statute include those of willful killing of an innocent party, motivated by anger, greed, or self-interest. It is also possible to imagine cases that are at an opposite pole but that seem covered by the statute's literal language: when a defendant has killed someone who has jeopardized the defendant's own life, we have a le￾gally sufficient justification under our law, no matter what the statute literally says. And why would cases of this kind be at the opposite pole? The answer is that, in such cases, the victim of the killing is 1999] 1887 HeinOnline -- 112 Harv. L. Rev. 1887 1998-1999

I888 HARVARD LAW REVIEW [Vol.112:1834 himself an egregious wrongdoer,one whose unlawful,life-threatening misconduct triggered the very killing in question.In such a case,ap- plication of the ban on willful killing would indeed seem absurd.It is hard to identify a sensible understanding of the criminal law that holds a defendant criminally liable in such circumstances.In fact,the law recognizes a legally sufficient justification in such circumstances,de- spite the literal language of the statute.If this case were akin to those at this pole,I have suggested that we would have an exceedingly hard question. But-and now I arrive at the crux of the matter-we have here a quite different situation.The victim was not a wrongdoer,and he did not threaten innocent persons in any way.His death was neces- sary only in the sense that it was necessary to kill an innocent person in order to permit others to live.The question is not whether we would agree,if we were legislators,to apply the statute in such situa- tions;to overcome the ordinary meaning of the statutory terms,the question is whether it would be absurd or palpably unreasonable to do so.The clear answer is that it is not. It is hardly absurd to say that there is no legal justification or ex- cuse for a willful killing in a situation like this one,even if more peo- ple on balance will live (or the killing is otherwise justified by some cost-benefit calculus).Many people who engage in killing can and do claim that particular excuse.To be sure,this case is different from the exemplary or prototypical ones in the sense that the killing was neces- sary to save lives.But there is nothing peculiar or absurd about ap- plying the law in such circumstances.People with diverse views about the criminal law should be able to accept this claim.Those who be- lieve in retribution and those who believe in deterrence should agree that the outcome,whether or not correct,is within the domain of the reasonable.Retributivists and Kantians are unwilling to condemn someone who has killed a life-threatening wrongdoer.But retributiv- ists and Kantians could certainly condemn the defendants here,who, to save their own lives,took the life of a wholly innocent person,one who withheld his consent at the crucial moment.For the retributivist, those who have killed,in these circumstances,have plausibly commit- ted a wrongful act,even if that act was necessary to save a number of lives.It is not unreasonable to say that the victim deserved to be treated as something other than a means to other people's ends.At the very least a conviction could not,for a retributivist,be deemed ab- surd. For their part,those who believe in deterrence should concede that a verdict of "innocent"could,in the circumstances of this case,confuse the signal of the criminal law and hence result in more killings.Many people who willfully kill believe that the outcome is justified on bal- ance,and we should not encourage them to indulge that belief.A judgment that N.C.S.A.(N.s.)$I2-A protects all blameless victims HeinOnline--112 Harv.L.Rev.1888 1998-1999

HARVARD LAW REVIEW himself an egregious wrongdoer, one whose unlawful, life-threatening misconduct triggered the very killing in question. In such a case, ap￾plication of the ban on willful killing would indeed seem absurd. It is hard to identify a sensible understanding of the criminal law that holds a defendant criminally liable in such circumstances. In fact, the law recognizes a legally sufficient justification in such circumstances, de￾spite the literal language of the statute. If this case were akin to those at this pole, I have suggested that we would have an exceedingly hard question. But - and now I arrive at the crux of the matter - we have here a quite different situation. The victim was not a wrongdoer, and he did not threaten innocent persons in any way. His death was neces￾sary only in the sense that it was necessary to kill an innocent person in order to permit others to live. The question is not whether we would agree, if we were legislators, to apply the statute in such situa￾tions; to overcome the ordinary meaning of the statutory terms, the question is whether it would be absurd or palpably unreasonable to do so. The clear answer is that. it is not. It is hardly absurd to say that there is no legal justification or ex￾cuse for a willful killing in a situation like this one, even if more peo￾ple on balance will live (or the killing is otherwise justified by some cost-benefit calculus). Many people who engage in killing can and do claim that particular excuse. To be sure, this case is different from the exemplary or prototypical ones in the sense that the killing was neces￾sary to save lives. But there is nothing peculiar or absurd about ap￾plying the law in such circumstances. People with diverse views about the criminal law should be able to accept this claim. Those who be￾lieve in retribution and those who believe in deterrence should agree that the outcome, whether or not correct, is within the domain of the reasonable. Retributivists and Kantians are unwilling to condemn someone who has killed a life-threatening wrongdoer. But retributiv￾ists and Kantians could certainly condemn the defendants here, who, to save their own lives, took the life of a wholly innocent person, one who withheld his consent at the crucial moment. For the retributivist, those who have killed, in these circumstances, have plausibly commit￾ted a wrongful act, even if that act was necessary to save a number of lives. It is not unreasonable to say that the victim deserved to be treated as something other than a means to other people's ends. At the very least a conviction could not, for a retributivist, be deemed ab￾surd. For their part, those who believe in deterrence should concede that a verdict of "innocent" could, in the circumstances of this case, confuse the signal of the criminal law and hence result in more killings. Many people who willfully kill believe that the outcome is justified on bal￾ance, and we should not encourage them to indulge that belief. A judgment that N. C. S. A. (N. S.) § 12-A protects all blameless victims 1888 [VOL 112:1834 HeinOnline -- 112 Harv. L. Rev. 1888 1998-1999

I999] THE SPELUNCEAN EXPLORERS 188g creates a clear deterrent signal for those whose independent judgments may not be trustworthy.From the point of view of deterrence,apply- ing the statute in this instance would,at the least,not be absurd, which is sufficient to justify my conclusion here. I would not entirely exclude the possibility that the defendants would have had a legally sufficient excuse if the unfortunate proceed- ings had been consensual at all times.It is conceivable that the ab- surdity exception would apply in that event as well.But this case is emphatically not that one,because the victim's consent was with- drawn before the dice were thrown.At that point,the victim ex- pressly said that he did not wish to participate in this method of de- ciding who would live or who would die.Where,as here,there was no consent to participate in the process that led to an unconsented-to death,the answer is clear:Those who killed acted in violation of the statute. Thus,it should be possible for those with diverse views of the pur- pose of the criminal law to agree that there is nothing absurd about following the ordinary meaning of the statutory text here.Indeed,I do not understand any of those justices who disagree with my general conclusion to disagree with this particular point.Their disagreement stems not from a judgment of absurdity,but from a willingness to dis- regard the text and to proceed in common law fashion-a willingness that would,in my view,compromise rule-of-law values.For example, Justice West urges the need for an individualized hearing,not because she thinks the conviction absurd,but in order to ensure individualized justice.See infra,at 1899 (West,J.).Justice Easterbrook thinks this case is analogous to self-defense,see infra,at 1913 (Easterbrook,J.), but he seems to take our jurisprudence to mean that courts may make particularized inquiries into the circumstances of killings.He does not suggest that a conviction would be absurd.I do not understand Jus- tice Stupidest Housemaid or Justice De Bunker to find absurdity here. And while I very much agree with Justice De Bunker's suggestion that criminal statutes should be narrowly construed,see infra,at 1902(De Bunker,J.),I would apply that suggestion only in cases of genuine tex- tual doubt. Some members of this Court plainly believe that the killing was morally excusable,because it was necessary in order to ensure that more people would live,and because the victim originally designed the plan that led to his death.See,e.g.,infra,at 1916-17(Easterbrook,J.). But that moral argument cannot be taken to override the natural meaning of the statutory terms,at least where the outcome is one that reasonable people could regard as justified.A serious underlying con- cern here is that to allow an exception on the claimed principle would be likely to undermine the statutory prohibition,either in principle or in practice.In principle,it is at least unclear that an exemption in this case could be distinguished from a claimed exemption in other cases in HeinOnline--112 Harv.L.Rev.1889 1998-1999

THE SPELUNCEAN EXPLORERS creates a clear deterrent signal for those whose independent judgments may not be trustworthy. From the point of view of deterrence, apply￾ing the statute in this instance would, at the least, not be absurd, which is sufficient to justify my conclusion here. I would not entirely exclude the possibility that the defendants would have had a legally sufficient excuse if the unfortunate proceed￾ings had been consensual at all times. It is conceivable that the ab￾surdity exception would apply in that event as well. But this case is emphatically not that one, because the victim's consent was with￾drawn before the dice were thrown. At that point, the victim ex￾pressly said that he did not wish to participate in this method of de￾ciding who would live or who would die. Where, as here, there was no consent to participate in the process that led to an unconsented-to death, the answer is clear: Those who killed acted in violation of the statute. Thus, it should be possible for those with diverse views of the pur￾pose of the criminal law to agree that there is nothing absurd about following the ordinary meaning of the statutory text here. Indeed, I do not understand any of those justices who disagree with my general conclusion to disagree with this particular point. Their disagreement stems not from a judgment of absurdity, but from a willingness to dis￾regard the text and to proceed in common law fashion - a willingness that would, in my view, compromise rule-of-law values. For example, Justice West urges the need for an individualized hearing, not because she thinks the conviction absurd, but in order to ensure individualized justice. See infra, at 1899 (West, J.). Justice Easterbrook thinks this case is analogous to self-defense, see infra, at 1913 (Easterbrook, J.), but he seems to take our jurisprudence to mean that courts may make particularized inquiries into the circumstances of killings. He does not suggest that a conviction would be absurd. I do not understand Jus￾tice Stupidest Housemaid or Justice De Bunker to find absurdity here. And while I very much agree with Justice De Bunker's suggestion that criminal statutes should be narrowly construed, see infra, at 1902 (De Bunker, J.), I would apply that suggestion only in cases of genuine tex￾tual doubt. Some members of this Court plainly believe that the killing was morally excusable, because it was necessary in order to ensure that more people would live, and because the victim originally designed the plan that led to his death. See, e.g., infra, at 1916-17 (Easterbrook, J.). But that moral argument cannot be taken to override the natural meaning of the statutory terms, at least where the outcome is one that reasonable people could regard as justified. A serious underlying con￾cern here is that to allow an exception on the claimed principle would be likely to undermine the statutory prohibition, either in principle or in practice. In principle, it is at least unclear that an exemption in this case could be distinguished from a claimed exemption in other cases in 19991 1889 HeinOnline -- 112 Harv. L. Rev. 1889 1998-1999