though perhaps only because the bourgeois French (including pied noir) society of the 1930s is more like our society than is the France of the Bourbon Restoration. The variety of human and social types is also greater in Stendhal's novel, as is the emphasis on social cla.s.s, clericalism, and love; and the author maintains a more critical perspective on his protagonist than Camus does on his.

Farcical Trials I called the trial of Meursault in The Stranger "a sinister farce"; we shall encounter other trials of that character in subsequent chapters. a.s.sociating law with injustice is a standard move in fictional depictions of law, but sometimes the dominant mood is ridicule rather than indignation. I discuss two examples of trials as outright farce: that of the Knave of Hearts in Alice's Adventures in Wonderland, and that of Joan of Arc in George Bernard Shaw's play Saint Joan.

The Knave is accused of having stolen tarts that the Queen of Hearts had made. (It is no surprise that a knave should be accused of knavish conduct.) The trial is before the King of Hearts, a dimwit. The Queen, the accuser, who sits at the front of the court beside the King though she has no official role in the trial, is incoherent with rage; it is she who famously utters, "Sentence first-verdict afterwards."76 The trial is to a jury of 12 small animals, each dimmer than the next. Surprisingly, the jurors are given slates and pencils so that they can take notes; until quite recently, jurors were forbidden to take notes. The lifting of the prohibition has been a welcome reform, but it is wasted on the jurors in the Knave's trial. Before the trial begins, Alice sees them scribbling on their slates and is told they're writing down their names "for fear they should forget them before the end of the trial" (p. 108). This elicits from Alice a scornful "Stupid things!"-and all the jurors write "Stupid things!" on their slates. "One of the jurors had a pencil that squeaked. This, of course, Alice could not stand, and she went round the court and got behind him, and very soon found an opportunity of taking it away. She did it so quickly that the poor little juror (it was Bill, the Lizard) could not make out what had be 76. Lewis Carroll, Alice's Adventures in Wonderland 122 (1865).

come of it; so, after hunting all about for it, he was obliged to write with one finger for the rest of the day; and this was of very little use, as it left no mark on the slate" (pp. 108109).

Not only are the jurors, like the judge, the witnesses, the spectators (including two guinea pigs and a dormouse who are "suppressed" for talking during the trial), and everyone else a.s.sociated with the trial, idiots, but they are terribly bullied and abject, as when the Queen hurls an inkstand at Bill the Lizard. He "had left off writing on his slate with one finger, as he found it made no mark; but he now hastily began again, using the ink, that was trickling down his face, as long as it lasted" (p. 122). The witnesses too, as when the King admonishes the first witness (the Mad Hatter): "Give your evidence . . . and don't be nervous, or I'll have you executed on the spot" (p. 111).

An ambiguous, unsigned poem, probably written by the Knave (it states that the author cannot swim, and the Knave admits that he cannot swim, as he is made of cardboard), proves to be the critical evidence, as it suggests that he returned the tarts, and in fact they are there at the trial, on the table; but this leaves unsettled whether he had stolen them. It is at this juncture, one of complete confusion, that the choleric Queen impatiently cries, "Sentence first-verdict afterwards." Alice, who has grown to a formidable height, responds, "Stuff and nonsense! . . . The idea of having the sentence first!"-and when the Queen responds, "Off with her head!" Alice ripostes, "Who cares for you? . . . You're nothing but a pack of cards!" (p. 122). The spell is broken, and she wakes from her dream.

"This is the legal process as it appears to the child in us: both frightening and ridiculous, founded on incomprehensible jargon, crazed unreason and the arbitrary power of the status quo, but capable of being overthrown by our own innocent child's-eye vision. 'You're nothing but a pack of cards' is what we all want, at some time or other, to say to those who give us doc.u.ments we can't read and subject us to legal procedures we don't understand."77 The trial of the Knave is a nonsense trial, but it intersects real trials at enough points to const.i.tute effective satire. The "Off with [his or] her head!" motif recalls the characteristic mode of execution of traitors (many of course merely political enemies) in earlier cen 77. Atwood, note 31 above, at 515.

*turies of English law. The bullying of jurors, witnesses, and spectators is an old story, as are the incompetent judge (who also has a conflict of inter-est-his wife is the crime victim), the clueless juror, the angry victim, the drive to convict an unpopular defendant, and irrelevant or fatally ambiguous evidence. "Sentence first, verdict afterwards" could be the motto of Stalin's purge trials, especially when intoned by a person of consequence, as the Queen undoubtedly is. (The English monarch was a queen when Alice in Wonderland was written.) American judges and lawyers love to quote it as a metaphor for procedural impropriety.78 Still, it is all just a dream, and too silly to be sinister. The same cannot be said for the trial of Joan of Arc in Shaw's play. It too is a farce and ridiculous, but as it culminates in the execution of an innocent young woman (moreover a historical figure rather than a fictional character), it sounds a graver note than the aborted trial of the Knave of Hearts. It is a trial in which the sentence indeed precedes the verdict. Joan is to be tried for heresy before an ad hoc ecclesiastical court in Rouen, a city controlled by the English, who are determined that she shall be executed. The Earl of Warwick tells the presiding bishop and the other ecclesiastical functionaries just before the trial begins: "I tell you now plainly that her death is a political necessity . . . I should be sorry to have to act without the blessing of the Church."79 The English clerical representative at the trial (the "chaplain") warns the court that "there are eight hundred men at the gate who will see that this abominable witch is burnt in spite of your teeth" (p. 134).

The note of the ridiculous is sounded early. The inquisitor (the prosecutor-and the original of the examining magistrate of Continental procedure) has cut the 64 original counts of heresy to 12. The chaplain is indignant. He complains that one of the most important charges against Joan is that the heavenly voices spoke to her in French (her only language); he See, for example, Southern Union Co. v. Southwest Gas Corp., 415 F.3d 1001, 1003 (9th Cir. 2005); Summers v. McLanahan, 2004 WL 14090843, at *9 (Cal. App. June 24, 2004). See also Marc Mauer, "Alice in Wonderland Goes to Criminal Court, or, How Do We Develop a More Effective Sentencing System?" 14 St. Louis University Public Law Review 259 (1994); Robert S. Pasley, "Sentence First-Verdict Afterwards: Dishonorable Discharges without Trial by Court-Martial?" 41 Cornell Law Quarterly 545 (1956). On literary references by American judges generally, see M. Todd Henderson, "Citing Fiction," 11 Green Bag (2nd series) 171 (2008).

Bernard Shaw, Saint Joan: A Chronicle Play in Six Scenes and an Epilogue 117 (1951 [1924]).

thinks they would have spoken in English. This prompts the inquisitor to reply that "it would not be very courteous to you, . . . or to the King of England, to a.s.sume that English is the devil's native language"-for both the inquisitor and the chaplain believe, or at least pretend to believe, that the voices were those of "evil spirits tempting her to d.a.m.nation" (p. 119). Another of the clerics pops up and alleges "a conspiracy here to hush up the fact that the Maid stole the Bishop of Senlis's horse"-provoking the inquisitor to reply abruptly: "This is not a police court. Are we to waste our time on such rubbish?" (p. 120).

A cleric named Ladvenu is sympathetic to Joan. He asks, "Is there any great harm in the girl's heresy? Is it not merely her simplicity?" To this the inquisitor replies with the same kind of "slippery slope" argument that we remember from Henry Wilc.o.x in Howards End: If you had seen what I have seen of heresy, you would not think it a light thing even at its most apparently harmless and even lovable and pious origins . . . Mark what I say: the woman who quarrels with her clothes, and puts on the dress of a man, is like the man who throws off his fur gown and dresses like John the Baptist: they are followed, as surely as the night follows the day, by bands of wild women and men who refuse to wear any clothes at all. When maids will neither marry nor take regular vows, and men reject marriage and exalt their l.u.s.ts into divine inspirations, then, as surely as the summer follows the spring, they begin with polygamy, and end by incest. (pp. 121122) This is nonsense, but it has the form and tone of legal argument.

The trial is short; it consists of verbal fencing between Joan and the clerics. She has the better of the exchanges, as when she responds to an invocation of custom by saying, "Do what was done last time is thy rule, eh?" (p. 128). All to no avail: "Her absolute simplicity of vision cuts raspingly through all the malign or well-intentioned errors of the world, until in its wrath the world rises up in the form of all its a.s.sembled inst.i.tutions and declares by the voice of all its a.s.sembled doctors that this girl is-as Shaw says-insufferable."80 80. Louis L. Martz, "The Saint as Tragic Hero: Saint Joan and Murder in the Cathedral,"in Tragic Themes in Western Literature 150, 160 (Cleanth Brooks ed. 1955) (footnote omitted).

*Threatened with torture, she recants. But when she learns that her sentence is merely to be commuted to life in prison she repudiates the recantation. The court then orders her excommunicated, and because the Church prides itself on its refusal to kill, "abandon[s]" her to the secular power (the English) to kill her (p. 138). (That is legalism with a vengeance.) The chaplain joins the English soldiers in rushing her to the stake. "This is irregular," one of the French clerics says: "The representative of the secular arm should be here to receive her from us." The inquisitor responds, in a further parody of legalism: "We have proceeded in perfect order. If the English choose to put themselves in the wrong, it is not our business to put them in the right" (p. 139). He acknowledges that Joan was "quite innocent. What does she know of the Church and the Law? She did not understand a word we were saying" (pp. 139140). But still he watches the execution impa.s.sively. He is inured to such sights; for, as he says, "habit is everything"-a variant of Hamlet's remark that "the hand of little employment hath the daintier sense." The chaplain had never seen a person burned at the stake, and the sight causes him to break down: "She cried to Thee [G.o.d] in the midst of it [the fire]: Jesus! Jesus! Jesus! She is in Thy bosom; and I am in h.e.l.l for evermore" (p. 141).

chapter 2.

Law's Beginnings: Revenge as Legal Prototype

and Literary Genre

liver Wendell Holmesargued, I think correctly (I'll give some evidence in this chapter), that law grows out of revenge.1 That origin has left its stamp on a number of legal doctrines and procedures, as well as on such overarching legal principles as corrective justice and retribution. Vengeful feelings play an important role in the administration of law even today. No general theory of law would be complete without attention to revenge. It is also a theme of some of the greatest monuments of the Western literary tradition.2 Literary depictions of revenge can tell us something about revenge and about the issues of law and justice that revenge adjoins or subtends, while the lawyer's and the social scientist's a.n.a.lyses of revenge can tell us something about revenge literature-can even dispel the mystery of Hamlet's delay in avenging his father's murder.

The Logic of Revenge I begin, far from the splendidly dramatic works that are my princ.i.p.al texts in this chapter, by sketching the logic of revenge. My sketch employs the Oliver Wendell Holmes Jr., The Common Law 225 (1881).

For a survey, see John Kerrigan, Revenge Tragedy: Aeschylus to Armageddon (1996).

75.

*vocabulary of rational choice (economics in a broad sense),3 but it is consistent with the historical and anthropological literature describing the practice of revenge in a variety of cultures.4 The pa.s.sion for revenge may seem the ant.i.thesis of rational, instrumental thinking. In particular, it flouts the economist's commandment to ignore sunk costs, to let bygones be bygones. When there is no possibility of legal redress to deter an aggressor, potential victims will be a.s.siduous in self-protection. But that can be extremely costly and often is futile. The alternative is retaliation against the aggressor after he has victimized you. Yet if you are "rational man" you will realize that the harm is a sunk cost. No matter how much harm you do to the aggressor in return, the harm you have suffered will not be annulled. Whatever dangers or other burdens you take on in order to retaliate will merely increase the cost to you of the initial aggression. To deter aggression in a revenge culture, the potential victim must convince potential aggressors that he will retaliate even if the expected benefits of retaliation, calculated after the aggression has occurred, are smaller than the expected costs at that time. He must in other words make a credible commitment to act in a way that may be irrational when the time to act comes. But the making of such a commitment may be rational (as conventional economic a.n.a.lysis might overlook): it may deter enough aggression to generate benefits greater than the costs of having sometimes to honor the commitment by retaliating, instead of cutting one's losses. Legal and political a.n.a.logies to the commitment to See, for example, Richard A. Posner, The Economics of Justice, ch. 8 (1981) ("Retribution and Related Concepts of Punishment"); Robert Axelrod, The Evolution of Cooperation (1984); Robert H. Frank, Pa.s.sions within Reason: The Strategic Role of the Emotions 170 pa.s.sim (1988); Maria Radinsky, "Retaliation: The Genesis of a Law and the Evolution toward International Cooperation: An Application of Game Theory to Modern International Conflicts," 2 George Mason University Law Review 53 (1994). The use of notions of rational behavior to explain or understand the behavior of characters in fiction, as I shall be doing in this chapter, is defended in Paisley Livingston, Literature and Rationality: Ideas of Agency in Theory and Fiction (1991).

Examples of this literature are Keith Otterbein, "Five Feuds: An a.n.a.lysis of Homicides in Eastern Kentucky in the Nineteenth Century," 102 American Anthropologist (new series) 231 (2000); David Cohen, Law, Violence, and Community in Cla.s.sical Athens (1995); Stephen Wilson, Feuding, Conflict and Banditry in Nineteenth-Century Corsica (1988); and Christopher Boehm, Blood Revenge: The Enactment and Management of Conflict in Montenegro and Other Tribal Societies (1984).

revenge include the self-destructive defensive measures that corporate managers sometimes take to deter hostile takeovers, such as the "poison pill," which may result in the acquirer's discovery that it owns a firm crushed by debt, or the commitment of the United States and the former Soviet Union to ma.s.sive nuclear retaliation, parodied in the "Doomsday Machine" of Dr. Strangelove.

In a society without formal legal inst.i.tutions, a legally enforceable commitment to retaliate against an aggressor is precluded; commitment has to come from instinct or culture. In human prehistory, people endowed with an instinct to retaliate would have tended to be more successful in the struggle for survival than others. The desire to take revenge for real or imagined injuries, without calculating the net benefits of revenge at the time when it is taken, may therefore have become a part of the human genetic makeup.5 The emotionality and universality of vengefulness are evidence for this speculation.

But more than instinct may be necessary to make threats to retaliate fully credible. Cultures in which revenge plays a significant role in the regulation of social interactions place great emphasis on inculcating a sense of honor.6 Shame, the reaction to being dishonored, helps overcome fear and so makes it more likely that a victim will retaliate if attacked or abused. Out of the interplay of honor, shame, and revenge grow notions of exchange, balance, reciprocity, and "keeping score"7-notions later taken up by law, initially under the rubric of corrective justice.

Another cultural overlay on the genetic impulse to revenge is the extended family. A potential victim of aggression in a revenge society needs allies; otherwise there would be no deterrence of murder. The natural place to seek them is among kin. In law's infancy murder is a purely private offense, to be dealt with by means of what today we would call a tort suit for wrongful death brought by the kin of the murder victim, implying that there is no legal sanction for murdering a kinless person, Robert L. Trivers, "The Evolution of Reciprocal Altruism," 46 Quarterly Review of Biology 35, 49 (1971); J. Hirshleifer, "Natural Economy versus Political Economy," 1 Journal of Social and Biological Structures 319, 332, 334 (1978).

6. On the relation of honor to revenge, see Frank Henderson Stewart, Honor, ch. 5 (1994).

The close connection between repayment and revenge is emphasized in Peter Singer, The Expanding Circle: Ethics and Sociobiology 39 (1981).

*though a supernatural sanction may be feared.8 Kinship increases not only the number of potential avengers but also the number of potential targets for revenge, and by doing so lowers the cost of revenge. If X kills Y, Y's family may decide to go after X's brother rather than after X himself- maybe the brother is not as well protected as X. With responsibility thus collective, people have an incentive to police their kin lest a kinsman's misbehavior lead to retaliation against them rather than against him.

A revenge culture places a premium on rhetorical skill. Even your kin may be reluctant to risk their necks on your behalf, so you have to be able both to put the offender in the wrong in their eyes and to convince them of the importance of retaliation, with all the risks that it entails, to the future security of the kin group.

This discussion shows that revenge is a system of social control, like law itself, rather than a sign of the absence of social control. But it is costly and clumsy. For one thing, it suffers from underspecialization. Instead of some people working full-time in law enforcement, leaving the rest of the community free to work full-time in other occupations, every man must spend part of his time as investigator, prosecutor, judge, sheriff, and executioner and equip himself with the skills and equipment necessary to play each role. And because not everyone has the same apt.i.tudes and resources, the duty of revenge may fall on someone unable to perform it.

An ethic of revenge makes large-scale cooperation difficult by breeding intense loyalty within the extended family; powerful loyalties within small groups r.e.t.a.r.d the formation of civic loyalties. This seems to have been Plato's objection, in the Republic, to the portrayal of Achilles in the Iliad.

In a revenge society, everyone nurses his ego, knowing that the more p.r.o.ne to retaliate against slights he is known to be, the less likely it is that anyone will dare inflict a slight on him. A common result is an overdeveloped sense of honor, such as what led to the crazy act of revenge that was the undoing of Julien Sorel. Cultures in which the sense of honor is highly developed, such as the traditional American South or America's inner cities today, are, despite the deterrent effect of a known willingness 8. On both points, see S. C. Todd, The Shape of Athenian Law 272273 (1993). But Todd believes it possible, though not certain, that under ancient Athenian law murder could be prosecuted at the behest of a person not related to the victim. Id. at 273.

to avenge slights, on balance more violent than cultures in which honor is taken less seriously.9 Acts of revenge tend to be too frequent, because the avenger, being a judge in his own cause, is likely to misjudge the balance of right and wrong in his own favor, sensing a slight when none is intended or believing that to maintain credibility he must interpret every injury to himself in the worst possible light. Aggressors become victims whose culturally honed instinct for revenge, activated by the revenge taken (perhaps rightfully) against them, is directed toward the original victim or the victim's avenger.10 Such a response is especially likely because-people being adept at rationalizing their conduct-the aggressor may not acknowledge to himself that he is a wrongdoer. The fact that vengeance may be taken against the aggressor's family rather than against the aggressor himself (if he is hiding or well protected) increases the likelihood that he will escape retribution and view himself as a victim obliged to take revenge. And it further strengthens the family at the expense of the broader community.

The possibility of evading immediate punishment may engender inter-generational feuds, as descendants of the original victim pursue the original aggressor or his descendants. Yet even that enlargement of the scope of revenge will often leave crimes between members of the same family unavenged; the son who kills his father would be the natural avenger of the crime. Beowulf contains a lament over the inadequacy of vengeance in such a situation. The plot of Oedipus Tyrannus turns on an alternative way of dealing with this situation-the "pollution" (plague) visited on Thebes as a result of Oedipus's parricide and incest.

For the threat of revenge to be an effective deterrent, moreover, the natural psychological tendencies of victims and aggressors have somehow to be reversed. Putative victims must have an unshakable commitment to retaliate violently for wrongs done to them, regardless of cost, while putative aggressors must be prudently self-restrained by rational calculation Richard E. Nisbett and Dov Cohen, Culture of Honor: The Psychology of Violence in the South (1996), esp. pp. 8891; David Hackett Fischer, Albion's Seed: Four British Folkways in America 892 (1989).

So Odysseus expresses concern that if he kills the suitors their families will avenge them, and in fact it requires a deus ex machina to prevent this. Odyssey XX.4243; XXIV.531540.

*of the costs of aggression, which include the possibility of retaliation by the victim. Justice as revenge can thus yield too little as well as too much punishment-and for the further reason that it depends on emotion. Because the average person's anger fades rapidly, revenge may not provide sufficient motivation to track down and punish-perhaps at a high cost in time and danger-an aggressor who has managed to elude swift capture. In early Roman law, as also in Anglo-Saxon and early Germanic law, a thief caught red-handed ("manifest theft") was subject to capital punishment or enslavement, whereas if he escaped and was caught later ("nonmanifest theft") the punishment was lighter.11 This pattern makes little sense from a deterrent or retributive standpoint.12 But it makes sense if criminal law tracks revenge, the desire for which tends to cool with the pa.s.sage of time.

A system of revenge thus places a premium on implacable anger, as well as on other traits, such as touchiness and unforgivingness-and, above all, the refusal to behave "rationally" in the face of slights-that further r.e.t.a.r.d the emergence of such forms of social cooperation as the market economy. Commerce depends on a willingness to compromise desires (the market translates people's subjective valuations of goods and services into finite prices) and thus to accept limitation and commensurability as norms of social interaction.

Even in a modern society, revenge has a role, as a supplement to formal law enforcement. Vengefulness in the form of righteous indignation and justified anger motivates victims of crime and their families to a.s.sist in the apprehension of the criminal-rarely is there a financial incentive to do so. James Fitzjames Stephen, the great nineteenth-century English theorist of the criminal law, explained that it is "highly desirable that criminals should be hated."13 For if one didn't hate the person who had done one (or one's family, or nation, etc.) ill, there would be little pressure to enforce the criminal law. That is why, Stephen argued, punishing affluent criminals by means of fines, though perhaps adequate from a deterrent 11. Henry Sumner Maine, Ancient Law 222223 (1917 [1861]).

But it does not make no sense: doubt of guilt is less when the criminal is caught red-handed, and that reduces the error costs of punishment.

Stephen, A History of the Criminal Law of England, vol. 2, p. 82 (1883). See also Steven Eisenstat, "Revenge, Justice and Law: Recognizing the Victim's Desire for Vengeance as a Justification for Punishment," 50 Wayne Law Review 1115, 11651168 (2004).

standpoint, would not be a satisfactory alternative to imprisonment; a fine would not quench the public's thirst for vengeance because it would not make the criminal suffer enough.

Stephen's emphasis on vengeance was natural because in his day the criminal law was privatized to a degree that we would find strange. Prosecutors, rather than being public employees, were private lawyers hired on an ad hoc basis to prosecute criminal defendants; merchants could hire lawyers to prosecute thieves; and public police were a recent innovation. Being much less professional and bureaucratized than we are accustomed to, law enforcement was closer to its roots in vengeance and more reliant on the emotions of the victims of crime.14 Emotion is an alternative to career incentives and other rational motivations for pursuing and punishing wrongdoers, though victims' emotions remain an important aid to law enforcement.

Vengeance has the further office of deterring aggressions and infringements-ranging from the mistreatment of friends and family members to breaches of faith by political leaders-that are too minor to justify activating law's c.u.mbersome and expensive machinery. But often vengeance itself has to be punished, not only when it takes the form of the revenge killings common in gangland settings but also when it takes the form of retaliation for exercising legal rights, as when an employee is fired for filing a complaint against employment discrimination.

Thus, justice as vengeance is crude from a moral standpoint once one steps outside the moral framework of the vengeance system itself. Lacking differentiated inst.i.tutions for making and applying rules of law, justice as vengeance cannot distinguish between culpable and justifiable or excusable injuries. Liability is absolute. The avenger is as "guilty" as the original aggressor. This feature of justice as vengeance makes the feud logical rather than pathological. Law provides both a superior calibration of punishment to blameworthiness and a machinery for public condemnation of wrongdoers that serves both to justify the punishment and to add a dollop of shame to it.15 See Harold J. Krent, "Executive Control over Criminal Law Enforcement: Some Lessons from History," 38 American University Law Review 275, 290295 (1989).

See Gary Jonathan Ba.s.s, Stay the Hand of Vengeance: The Politics of War Crimes Tribunals 288 (2000), and the discussion of Henry Hart's theory of punishment in chapter 7.

*A revenge system is unstable. The larger a man's kin group, the less he will be restrained by fear of retaliation.16 So there will be compet.i.tion to form protective groups large enough not to be overawed by any man or his kin. Fictive kinship, culminating in state formation with citizenship the organizing principle in place of kinship, expands the group far beyond its "natural" limits. The Icelandic sagas depict this transition.17 Iceland, previously uninhabited, was settled by Norwegians in the tenth century ad. For 300 years the Icelanders managed to get by with a rudimentary political and legal system in which revenge not only was lawful but was the basis of public order. The country's poverty made it difficult for anyone to support an entourage of feudal-style retainers who in exchange for food and shelter would place an armed force at the disposal of their liege, and lacking such an armed force he had little to offer in the way of protection to other people in the society-little, but not nothing: there were chieftains. But the small forces at their disposal-a handful of relatives, dependents, and clients taking the afternoon off from tending their flocks-limited their power to the execution of piecemeal revenge. None was able to offer a king's peace to the entire society.

A community needs an economic surplus to support specialists in coercion. Toward the end of the period of Icelandic independence the Roman Catholic Church was able at last to collect substantial taxes from the population. But it was unable to keep the bulk of the tax revenues out of the coffers of a handful of major chieftains. Six of them grew powerful enough to wage civil war, at which point the population was happy to turn to the king of Norway for protection, and Icelandic independence ended.18 Civil war had not been inevitable. Not all arms races end in war. But where the potential gains from obtaining a monopoly of political power are great and seem attainable, a miscalculation can easily lead to war.

Despite all these criticisms, I do not argue that vengeance is a bad system of justice. That depends on the alternatives. The Odyssey depicts 16. See Boehm, note 4, at 168.

17. See William Ian Miller, Bloodtaking and Peacemaking: Feud, Law, and Society in Saga Iceland (1990), discussed in my book Overcoming Law, ch. 14 (1995). See also Theodore Ziolkowski, The Mirror of Justice: Literary Reflections of Legal Crisis, ch. 3 (1997).

18. Thrainn Eggertsson, Economic Behavior and Inst.i.tutions 309310 (1990).

vengeance as thoroughly normal, and the reader thrills when Odysseus slaughters the scores of suitors and hangs the dozen maids who had slept with them. There is no other system of justice in the world of the Odyssey, just as during the Cold War world peace could be maintained only by the Soviet Union's belief that the United States would retaliate in kind to a nuclear attack even if the attack so devastated the United States as to deprive retaliation of any benefits to the few left alive.

Even before there is a nation or other organized community to take over from the victims of aggression and their families the responsibility for punishing aggressors, customs evolve to limit revenge. The principle of retribution-that is, of proportionate retaliation for a wrong-reduces the likelihood of overreactions (your life for my eye) that are likely to engender feuds.19 (So retribution is "sanitized revenge.")20 Another principle is "composition" (blood money, wergeld): the victim or his family is required, or at least encouraged, to accept payment in compensation for an injury, discharging the injurer's liability. A transfer of money or goods is less costly to society than an act of violence, which besides inflicting a net social loss rather than merely transferring wealth from one person to another may provoke further violence.

Another mitigating principle is bilateral kinship. Icelanders reckoned kinship through both the father and the mother. This not only increased the credibility of revenge as a deterrent to aggression by strengthening the family but also made it more likely that a disputant would have kin on both sides of the dispute. Caught in the middle, these kinsmen were naturals to try to make peace between the disputants. The Iliad hints at the further possibility that pity might limit the savagery of revenge.

Retribution is not feasible for all wrongs-for defamation, for instance, or where victim and aggressor are not identically circ.u.mstanced (suppose A, who has only one eye, puts out one of B's eyes). It is not an adequate deterrent if an aggressor is unlikely to be caught and punished every time, for in that case the punishment will not equal the crime when all his pun 19. This is a major theme of Axelrod, note 3 above. See, for example, pp. 121123.

20. Andrew Oldenquist, "An Explanation of Retribution," 85 Journal of Philosophy 464 (1988). See also Thomas C. Bilello, "Accomplished with What She Lacks: Law, Equity, and Portia's Con," in The Law in Shakespeare 109, 122123 (Constance Jordan and Karen Cunningham eds. 2007).

*ishments and all his crimes are summed. And a commitment to limited retaliation, or to accepting money or goods in compensation for a wrongful injury, or to be compa.s.sionate, is hard to stick by in the emotional circ.u.mstances in which revenge is provoked and administered.

As centralized inst.i.tutions of law enforcement emerge in tandem with the nation-state, vengeance, even as refined by principles of retribution, composition, and compa.s.sion, falls out of favor. The nation-state claims a monopoly of force; so taking the law into one's own hands becomes itself a crime. Practices are adopted that direct vengeful feelings into socially less disruptive channels, such as giving the victim his day in court, whether as plaintiff in a civil case or as prosecuting witness in a criminal case, in lieu of private vengeance. The damages recoverable in a private litigation, however, may no more slake the desire for revenge than the composition obtainable under customary law, or than a fine would in a case of serious crime. Revenge by victims and their relatives gives way to retribution by disinterested persons-with advantages made transparent, as we shall see, by Hamlet. Yet revenge tends to break out whenever legal remedies are blocked, as when the evildoer controls the legal machinery (Hamlet again) or is otherwise above the law, or when public law enforcement is very lax.

Intermediate stages in the evolution from revenge to public enforcement include the duel21 and, closely related, trial by battle and ritualized medieval warfare.22 Like retribution, these are devices for heading off feuds by establishing a natural end to a dispute. But feuds, duels, trials by battle or ordeal, and "rational" litigation coexisted for a long time.23 Espe For a case study, see R. S. Radford, "Going to the Island: A Legal and Economic a.n.a.lysis of the Medieval Icelandic Duel," 62 Southern California Law Review 615 (1989). See also Eric A. Posner, "Law, Economics, and Inefficient Norms," 144 University of Pennsylvania Law Review 1697, 17361740 (1996), and the references cited there. The duel between Menelaus and Paris in book 3 of the Iliad is an example of an attempt (in that instance unsuccessful) to subst.i.tute the cheaper alternative of a duel for full-scale revenge. See also Bertram Wyatt-Brown, Southern Honor: Ethics and Behavior in the Old South 352 (1982).

"The vision of the Agincourt war, as presented both by Shakespeare [in Henry V] and by contemporary sources, is that of a law suit culminating in trial by battle, which will establish justice." A. W. B. Simpson, "The Agincourt Campaign and the Law of War," 16 Michigan Journal of International Law 653, 656 (1995).

See, for example, The Settlement of Disputes in Early Medieval Europe (Wendy Davies and Paul Fouracre eds. 1986).

cially in aristocratic circles the duel persisted long after an essentially modern legal system was in place. The persistence of norms of revenge in American gangster culture is a parallel phenomenon. Aristocrats and gangsters are disinclined or unable to submit their disputes to the formal legal system, which is "owned" by a different social or political cla.s.s.

Corrective justice defines the task of the legal system to be to restore the balance between persons that a wrongful act has disturbed. It reflects the revenge-inciting indignation caused by the infringement of one's rights, but is conceived of in individual rather than group terms. Consistent with the nation-state's hostility to strong kinship groups, the family falls out of the picture. The social contract that legitimates the state and formal law, supplanting justice as revenge with corrective justice, is not a means of overcoming the selfishness and atomistic individualism of man in nature; it is the opposite. Revenge is not a selfish emotion, and a revenge ethic breeds powerful family and small-group loyalties. The state creates the conditions in which selfish behavior will not endanger social order. The state, as a collective, is hostile to competing collectives.

The relation between revenge and law is ill.u.s.trated by the rudimentary legal system of medieval Iceland. The formal inst.i.tutions of government were limited to courts and an a.s.sembly, all staffed entirely by ordinary citizens rather than by professionals. In effect there were jurors but no judges; there were no appeals either. Nor were there any sheriffs, police, soldiers, or prosecutors. The only sanctions that Icelandic courts meted out, other than for the most trivial infractions (which were punished by a fine), were outlawry, which made a man an outlaw in the literal sense- anyone could kill him with impunity-and lesser outlawry, which meant banishment from Iceland for three years. A dispute might be submitted to binding arbitration conducted by one or more men selected by the disputants. Refusal to obey an arbitral decree was punishable like other serious wrongdoing. Feuds, though lawful, were governed by norms; for example, the killing of an outlaw was not to be avenged-that is, it was not to occasion a feud.

Legal judgments were not self-executing. If the convicted defendant thumbed his nose at a judgment, the plaintiff would have to rally his kin to enforce it by force, much as if he had decided to retaliate directly against the defendant for whatever wrong had precipitated the suit. But a legal judgment might have enough suasive force to enable the plaintiff to rally

his allies and also make it easier for the defendant's potential allies to beg off, thus isolating the defendant and so vindicating the plaintiff 's decision to sue rather than to fight. Law played the same role that graphite rods play in the core of a nuclear reactor: to slow down the chain reaction. So did bilateral kinship.

Ancient Athens had a legal system not unlike that of saga Iceland.24 Prosecutions even for such serious offenses as murder and treason were inst.i.tuted by private persons ("denouncers") and tried before panels of citizens chosen at random. There were no professional judges and no appeals. There were no lawyers as such, though litigants hired rhetoricians to write speeches for them. Criminal judgments were enforced by public officers, unlike the procedure in Iceland; and the juries were larger. There were other differences. But the parallels are striking; they even include the use of banishment as a sanction, and protracted and repet.i.tive litigation by "feuding" factions.25 Even closer to saga Iceland is the society depicted in the Homeric epics, which reveal only rudimentary government inst.i.tutions and exhibit the same emphasis found in the sagas on revenge as the basic principle of social order. Other effectively stateless regimes have operated in this way, such as the mining communities that sprang up during the California Gold Rush to maintain order at a time when, and in a place where, public authority was virtually nonexistent.26 Revenge Literature Revenge is a pervasive theme in the literature of ancient Greece (as well as in other ancient literatures-think of the ferocious revenge of the Israelites for Schechem's rape of Dinah narrated in Genesis). Aeschylus's trilogy Oresteia is based on a revenge story,the legend of the House of Atreus.27 On the Athenian legal system, see Todd, note 8 above.

Cohen, note 4 above.

See, for example, Robert C. Ellickson, Order without Law: How Neighbors Settle Disputes (1991); Gary D. Libecap, Contracting for Property Rights (1989).

For a full account of the legend, see Timothy Gantz, Early Greek Myth: A Guide to Literary and Artistic Sources, ch. 15 (1993). On law, justice, and revenge in the Oresteia, see Kevin M. Crotty, Law's Interior: Legal and Literary Constructions of the Self 4273 (2001); Michael Gagarin, Aeschylean Drama, ch. 3 (1976); David Cohen, "The Theodicy of Aeschylus: Justice and Tyranny in the Oresteia," 33 Greece and Rome 129 (1986).

Thyestes had wronged his brother Atreus in retaliation for wrongs done him by Atreus. Inviting Thyestes to a banquet, ostensibly for reconciliation, Atreus kills Thyestes' sons and feeds their flesh to their unknowing father. (The banquet scene is described in hideous detail in Seneca's tragedy Thyestes and echoed in Shakespeare's t.i.tus Andronicus.) One of Thyestes' sons had not been killed, however-Aegisthus. (In some versions of the legend he is Thyestes' grandson rather than his son.) Aegisthus kills Atreus and drives away Atreus's sons, Agamemnon and Menelaus. They regain power eventually, but then go off to fight the Trojan War-a war to avenge Paris's seduction of Menelaus's wife, Helen, as well as to recover her. In their absence Aegisthus becomes the lover of Agamemnon's wife, Clytemnestra. She is nursing her own grievance: Agamemnon had sacrificed their daughter Iphigenia on the way to Troy to appease the G.o.ddess Artemis. When Agamemnon returns at long last from the Trojan War, Aegisthus and Clytemnestra kill him.

His son, Orestes, kills them. Pursued by supernatural beings, the Furies (euphemistically termed "Eumenides"-the "kindly ones")-for Orestes himself would be the natural avenger of his mother's murder were he not the murderer-he is eventually brought to trial in the Court of the Areopagus in Athens, presided over by Athena. He is acquitted of the murder of his mother, and the cycle of vengeance is broken, as there is no avenger of Aegisthus. Eumenides, the last play of the Oresteia, ascribes the founding of the Court of the Areopagus, reputed to be the first formal court in the cla.s.sical world, to a desire to end the cycle of vengeance- thus making transparent the relation between vengeance and the absence of regular inst.i.tutions of criminal justice. Until the trial it is a.s.sumed that Orestes' killing of Clytemnestra, though justifiable and indeed inescapable, must itself be avenged, just as it is a.s.sumed that Agamemnon had to be killed as punishment for sacrificing his daughter even though he had done it to get on with the war against Troy-a war ordained by Zeus. Hence the pursuit of Orestes by the Furies. And recall that although Homer presents Odysseus's revenge against the suitors as wholly justifiable, had it not been for Athena's intervention the suitors' families would have had a duty to avenge their deaths (Odyssey XXIV.531548).

Since revenge is a.s.sociated with an uncompromising form of strict liability for harms inflicted unintentionally, excusably, or even justifiably, one is not surprised that early legal systems, in which the roots of law in *revenge still show, rely on strict liability more heavily than modern legal systems do. Oedipus is guilty of parricide and incest and must be punished even though he had no reason to know that the man he had killed was his father and the woman he had married was his mother. His punishment is a metaphor for the human condition. We often suffer as a result of acts that we do in warranted ignorance of the possible consequences; in other words, we may be "punished" for acts that are "innocent" or, in the case of Agamemnon, justifiable and hence in modern legal systems excused. The implied system of liability in Oedipus Tyrannus, as in the first two plays of the Oresteia, is crude from a modern legal standpoint, yet this crudeness is bound up with its dramatic effectiveness. Absolute liability (stricter than normal strict liability, which allows for some justifications) makes law a more effective metaphor of necessity-of the external world that bears down on people, whatever their status or merit, and ruins their dreams.28 This is especially true when, as in Oedipus, "nature" and "law" are not clearly distinguished.29 But we must be wary of exaggerating the ethical gulf between the world of Oedipus and our world. Jocasta kills herself and Oedipus blinds himself, but these are effects of horror and remorse rather than punishments. The only punishment that is clearly required is that Oedipus leave Thebes; until he does, the plague will continue. We moderns would not inflict formal punishment on a modern officeholder who, without knowledge or reason to know, committed Oedipus's offenses (provided that he had killed his father in self-defense). But consider what such revelations would do to the officeholder's tenure, let alone to his and his wifemother's psychological state. Kantian morality, which judges the person rather than the act and so equates wrongfulness with blameworthiness, is not the whole of our morality. If two drivers drive with equal carelessness 28. "Through solving the riddle of the sphinx by his intelligence Oedipus was able to marry the queen, his mother; through his clever power of logic he is able to untangle the riddle of his origins and realise his tragedy . . . At each stage of the narrative, he is active in the plot, an autonomous individual, deciding to murder Laius, deciding to sleep with Jocasta. So he actively transgresses and must later feel shame. But retrospectively, when the story is considered as a whole, it is apparent that the G.o.ds and his fate have shaped his life." Jennifer Wallace, The Cambridge Introduction to Tragedy 18 (2007).

29. See Lloyd L. Weinreb, Natural Law and Justice, ch. 1 (1987).

and one has an accident and injures someone and the other does not, the first will receive greater moral as well as legal condemnation even though the conduct of the two drivers, as distinct from the consequences of that conduct, was blameworthy to the same extent. The second driver is the beneficiary of "moral luck": "in the story of one's life there is an authority exercised by what one has done, and not merely by what one has intentionally done."30 Similarly, Agamemnon made a defensible choice to sacrifice Iphigeneia but must be punished anyway.

We puzzle over this; and so does Oedipus, in Oedipus at Colonus, the sequel to Oedipus Tyrannus, written many years later. Notions of liability may have been changing, for the later play acknowledges a moral difference between strict liability and liability based on blameworthiness. Oedipus a.s.serts his moral innocence, founded on lack of premeditation or even lack of carelessness, and perhaps as a result of that innocence his death on Athenian soil becomes a blessing to Athens rather than the curse one might have expected it to be. An emergent sense of free will-a basis for distinguishing between cause and fault as grounds of liability-is also apparent in the difference between the curse that doomed Oedipus in the earlier play-a curse delivered by Oedipus against the then-unknown murderer of his father-and the curse delivered by Oedipus in the later play against his son Polynices. Oedipus could have done nothing to escape the consequences of his unwitting self-cursing, but Polynices could have escaped the consequences of Oedipus's curse simply by not attacking Thebes. (The disastrous consequences of that attack are the subject of Antigone, which I discuss in the next chapter.) Polynices rejects this escape route because he would lose face if he took it, and in particular because he would be humiliated before his younger brother, Eteocles, the defender of Thebes. He may be said to choose his fate and so to be culpable in a way that Oedipus had not been.

Yet even in the earlier play, and in the Oresteia, one senses a recoil from the punishment of a person who happens merely to be unlucky. For Oedi 30. Bernard Williams, Shame and Necessity 69 (1993). See also Williams, "Moral Luck," in his book Moral Luck: Philosophical Papers 19731980 20, 2830 (1981), and Robert B. Pippin, "Morality as Psychology, Psychology as Morality: Nietzsche, Eros, and Clumsy Lovers," in Pippin, Idealism as Modernism: Hegelian Variations 351, 367368 (1997).

*pus and Agamemnon are not depicted as wholly without fault, though their punishment is disproportionate to their blameworthiness. Oedipus is impious in trying to evade his fate. He knew it had been prophesied at his birth that he would kill his father and marry his mother; what he did not know, in his arrogant finitude, was that he was a foundling. The fury with which Agamemnon kills Iphigenia, as well as his impiety in accepting Clytemnestra's invitation to walk into the palace on a purple carpet (the original "red-carpet treatment")-as if he were too good to let the G.o.ddess earth touch him-mark him as more than an involuntary transgressor. It is apparent, moreover, that Agamemnon was not actually forced to sacrifice his daughter,31 though his refusing to do so and as a result having to abandon the rescue of Helen and the punishment of the Trojans would have been an even worse choice, given Zeus's decision that they should be punished.

The trial of Orestes dramatizes the transition from a system of absolute liability, enforced by revenge, to one of legal liability based on blameworthiness in something approaching the modern sense. Hounded by the Furies, Orestes takes refuge with Apollo, who had encouraged him to kill his mother. Orestes thinks it unfair that he should be punished for having done what a G.o.d ordained. Apollo agrees and comes up with the idea of a trial ("the first trial of bloodshed," as Athena says) to sort out blame; and the procedure at Orestes' trial, apart from the presence of divine personages, is an approximation to Athenian trial procedure in the fifth century bc. The Furies argue that if Clytemnestra's murder of her husband is punishable, which of course is the premise of Orestes' defense, so should his murder of his mother be. The reply that leaps to mind is that the killing of Clytemnestra was justifiable because it was punishment for her act; an executioner is not a murderer. But Clytemnestra, with some justification, had regarded her act as punishment for Agamemnon's killing their daughter. So Orestes would have to argue either that that killing, too, was justified-an unattractive argument of "the end justifies the means" variety-or, more subtly, that Clytemnestra's act, being not 31. He deliberated before deciding on the sacrifice. E. R. Dodds, "Morals and Politics in the Oresteia," in Aeschylus 245, 258259 (Michael Lloyd ed. 2007 [1960]).

just murder but also usurpation, had a political resonance that made it a disproportionate punishment for Agamemnon's misconduct.