Full Text

Modern law's response to mass atrocities vacillates equivocally in how it understands the dramatis personae to these expansive tragedies, at once extraordinary and ubiquitous. Is there any principled order to this? If not, should we care?
We sometimes treat the parties as individual "natural persons"-say, Slobodan Milošević, as defendant before the International Criminal Tribunal for the former Yugoslavia. At other times, the characters emerge as social groups of various sorts, which law treats as irreducible to the individuals composing them. The law so re- gards, for instance, the victims of genocide, persecution, aggression, apartheid, and certain war crimes.
This would not be worthy of note, were it not often perfectly possible to con- ceptualize the same facts in either fashion, individualizing now what was, a mo- ment ago, taken to be collective in character, and vice versa. In the historical episode just mentioned, for instance, the party liable for most Balkan war crimes and genocide would then be "The State Union of Serbia and Montenegro." And the victims would be seen as the people targeted by the war crime of "attacks against individual civilian persons not taking part in hostilities."
In atrocity's aftermath, the law thus often authorizes us to move in either direc- tion, or in both at once, as on the facts just described. The puzzle is more complex still. The law sometimes instructs us that we must individuate, not collectivize. And sometimes, it tells us the very opposite, on what would appear to be identical
facts. It is not clear whether any principled rationale or regularity exists in the way we manage such matters. Even if there is indeed an intelligible pattern in the law's workings, does it correspond to how mass atrocities actually occur, that is, with factual regularities in the evidentiary record of these catastrophic conflagrations? An affirmative response would be reassuring. For it means that reality is out there, that it can be determined through available methods of discovery, and that our law closely reflects it. There is order to the universe.
Or do the commonalities in our legal responses, such as they are, arise simply from the consistent but unthinking application of longstanding rules bearing no genuine relation to the events themselves, rules derived from law's encounters with wrongs very different in scale, motivation, and the organization of their co- mmission? After all, our handling of these unusual cases depends, to an embar- rassing extent, on the most tenuous and counterintuitive of "legal fictions." All is flux.
We might ask much the same questions in a normative key. Must the logical order we may desire in the world's legal response to these colossal horrors metic- ulously track their true character, as historians would describe and social scientists explain it? Or is it enough that the law's conceptual structure, as a fluid intellectual creation, allows us to contest the meaning of these events, whatever we take their "true facts" to be, in juridically cognizable terms, even ones devised for the very occasion? We must then acknowledge that the law is simply-forgive the clunky cliché-a social construction, which we may put up or tear down at will. All of the law is, after all, created to serve our purposes, which vary widely from one histor- ical circumstance to another. The law is not intended to naively "mirror" the em- pirical world, even to the limited extent we could hope to discern it from the widely conflicting accounts of these infinitely complex episodes.
I here defend a middle ground. On one hand, any satisfactory legal response to mass atrocity must correspond to essential facts, often startlingly novel, about what actually transpired, even as legal actors may contest what these are. On the other hand, any acceptable atrocity-response must also cohere with preexisting rules for understanding the parties and processes involved. It must nonetheless also enable opposing counsel to offer competing interpretations of the events' true legal import. We need not understand law's coherence to preclude all lawyerly contrivance.
The curious wavering discernible in our atrocity-responses-individuating here, collectivizing there-seems to reflect an underlying uncertainty within the law about who "really" occupies any of the three recurrent categories: perpetrators, victims, and beneficiaries. The law is unwilling to commit to any particular answer to the question, at least before carefully assessing evidence about the details of a particular episode. We are at least somewhat uncomfortable with this uncertainty, insofar as we let ourselves acknowledge it. For we intuit that there must surely be right and wrong answers to such morally momentous questions. And we know that how we identify each of the three characters-who falls within, who is left out-has weighty implications for law's legacy in the ashes of these sad affairs.
Our ambivalence plays out across a wide legal landscape, in areas not generally examined in conjunction. It arises even when designing the public monuments by which we memorialize the victims of mass violence. The larger work from which this essay draws shows how the recurrent dissonance between our conflicting in- clinations-to isolate and to aggregate, to split or lump-finds expression, and re- ceives variable resolution, in the law of international crime, of civil compensation for atrocity victims, of counterterrorism, and of state responsibility for atrocious human rights abuse. These are the four fields to which the global community today chiefly turns in confronting the consequences of these cataclysms. I here il- lustrate the issues with examples drawn from several such episodes throughout the world.
It is conceivable that no coherent ordering (however understood) yet exists nor will prove possible, whatever the analytical firepower brought to bear. Call this the "null hypothesis." It suggests that mass atrocity must elude understanding and conceptual clarification, thereby resisting satisfactory alleviation through law. Even when stretched to its limits, our familiar analytical apparatus for handling simpler criminality, between single perpetrator and single victim, may not satisfac- torily accommodate very large-scale wrongs implicating many people in both ca- pacities, sometimes the same people-child soldiers, for instance.
We may conclude that the seeming incoherence of the world's legal response to mass atrocity merely mirrors the infinite variation evident within and across the events themselves, requiring that for such episodes we incorporate some unique amalgam of individualizing and collectivizing strategies. To insist upon greater uniformity in either our social understanding of mass atrocity or our methods for its legal redress would then be not so much hopefully premature as fundamen- tally misguided.
The Parties to Mass Atrocity
Wrongdoers
Let us start with the wrongdoers. Who are they, really and truly? For instance, who "perpetrated" the Gulag, exactly, in both the layman's and lawyer's sense? What party or parties would be responsible, according to today's international law? More precisely still, which individuals and institutions could be held criminally or civilly liable for what portion of this vast complex of violent human rights violations and crimes against humanity, committed by thousands of people over several decades?
There are several plausible answers. We might choose to define the Gulag's "actual" architect and responsible party as Joseph Stalin alone or, alternatively, as the sovereign state he governed, the Union of Soviet Socialist Republics. The first approach would today urge Stalin's prosecution before the International Criminal Court (which I'll hereafter call "the Court"). The second would urge a finding of "state responsibility" (a technical term here) by the International Court of Justice (ICJ). Or perhaps we might prefer something between the two, enlarging the scope of liability beyond any single individual, yet not so wide as an entire nation-state. How about the top leaders of the Committee for State Security (that is, the KGB)? We might credibly classify them all as forming a "joint criminal enterprise," through which each such leader became liable for the wrongs of every other. Alternatively, why not the KGB itself, much as today it is sometimes possible to convict a company in its own right, apart from any fur- ther liability attaching to particular managers or directors? And why should we not hold KGB leaders "jointly and severally liable" in tort for "wrongful death" and "false imprisonment" of their millions of victims?
But why stop there? We might also point the accusatory finger at low echelons within this infamous "archipelago" of slave labor camps spreading across a thou- sand miles of Siberian tundra. A particular commandant governed each camp, several commandants over the course of its history. The agile legal mind could quickly spin out still other scenarios, all readily cognizable within existing juridical terms.
The questions just posed concern the range of parties legally responsible for mass atrocity. Similar questions arise regarding the range of beneficiaries from these wrongs. These people are not always identical to the perpetrators.

Beneficiaries
We may wonder, for instance, precisely who was "unjustly enriched" by the Shoah-what persons, which organizations? From them, well-settled law could re- quire restitution of illegally acquired assets to successors of their true owners. We might ask much the same of those many loyal Communist Party members who eagerly took possession of the most valuable private dwellings within the most de- sirable neighborhoods of Central and Eastern European metropolises, after these properties were seized without compensation following Soviet conquest. Do the rightful owners of these homes have a legal claim against only the state, which en- joys certain immunities in any event? Or may such owners also vindicate their claims against the private parties who took possession of the property?
If this is so, then why should Argentina's courts not employ essentially the same legal rules of restitution to order the return of those adolescents abducted in infancy from parents soon "disappeared" to surviving biological relatives they have never met? Hundreds of infertile couples among the country's ruling armed forces joyously accepted such children into their families through surreptitious, ir- regular forms of "adoption." These military elites knowingly benefited from multiple acts of kidnapping.
This example reveals that those who importantly benefit from the wrong need not necessarily do so in pecuniary terms, yet a second difference between our cases. Nor are these beneficiaries always individuals. Those profiting, for instance, from "conscripting or enlisting children under the age of fifteen," which is a war crime, clearly include the "armed forces or groups . . . using them to participate actively in hostilities."
Victims
The same questions I have posed thus far about wrongdoers and beneficiaries also arise with respect to victims and the legal remedies available to them. The Court's Statute defines victims to include both natural persons and "organizations or in- stitutions" of certain sorts. It requires the Court to allow direct participation by victims of the alleged atrocities. The judges construe this as a right to participate at almost every stage of the proceedings. Since there are often thousands of imme- diate victims, the novel procedures to which this right gives rise are already prov- ing highly cumbersome, to say the least.
The Court's Statute also authorizes "reparations . . . for the benefit of victims of crimes within the jurisdiction of the Court, and of the families of such victims."
The preferred source of these reparations is to be "money and other property col- lected through fines or forfeiture," obtained by "order directly against a convicted person." Further reparations will issue from the Court's Victims' Trust Fund, to which states and others, such as private parties, companies, and NGOs, voluntarily contribute.
All these provisions, procedural and remedial, depend on some notion of "the victim," a concept the Statute nowhere defines. One might plausibly construe the notion broadly or narrowly. The broader one's construction, the greater the number of people entitled to participate and to monetary recovery. Though the question has not been definitively resolved, the Trust Fund's administrators and many of the judges themselves today understand the notion of victim very broadly indeed, ex- tending it to all those harmed not merely by the particular crime the Court formally finds but to the larger "situation" of which it comprised a part. The Statute expressly enshrines that key term-situation-as well, and leaves it similarly undefined.
This definitional ambiguity allows the prosecutor, the Court, and its Fund ad- ministrators considerable discretion over how to determine the "true" temporal and geographical contours of a given "situation" of atrocity-occurrence. The inde- terminacy in both statutory concepts-victim and situation-combine to make it difficult to draw principled, consistent boundaries around the class of those enti- tled to reparations. If we spaciously interpret either concept, these persons could easily number in the hundreds of thousands.
Even if we narrowly construe the class of victims from a given episode, there re- mains the prior question of why any victims, conceived as individuals, should oc- cupy so central a role, to the extent that many voices (in The Hague and beyond) clearly demand they do. For most of modern history, Western law axiomatically has regarded the true victim of all crime as "society at large"-a species of collec- tivity-not the individual persons more immediately injured, however severe their suffering. Thus the question of who is a "genuine" victim of mass atrocity often becomes surprisingly, disconcertingly difficult. This is so whether we wish to con- sider the matter from the perspective of legal discourse or of common morality.
Defining the Parties to Mass Atrocity
Groups or Individuals?
In notable respects, the questions posed thus far will be unfamiliar, dizzyingly so perhaps, to all but a small number of readers. How does the law answer them all? How should it? Do the ghastly depravities here examined stand out in ways unique not only to world history but to the law as well? We are probably inclined to view the implicated parties, even the wrongs themselves, differently than in more famil- iar forms of illegality, even the most violent. In garden-variety crimes, individual people are always at the center of the action and the law's attention, as wrongdo- ers, victims, and beneficiaries. And the wrongdoers are also the beneficiaries of the wrong, so there are really only two classes of dramatis personae, not three or more (for example, active collaborators, influential bystanders).
With mass atrocities, though, various kinds of collectivities-from stiffly bureaucratic organizations to more casual human congeries-also prominently occupy the stage. Whether or not they can sue, be sued, or be prosecuted, they will often appear the "real" parties in interest to these events, so far-reaching in scope and scale. We could not begin to understand what transpired-even at the most intimate, interpersonal level-by first reducing these events to what any given individual did to any other, then somehow tallying all this up, whatever that might mean. To understand mass atrocity as a series of discrete acts of mano-a-mano violence would clearly miss many characteristics essential to our as- sessment of those engaged in it, victimized by it, and benefiting from it. This is presumably what people mean when they casually say of such events that "the whole is greater than the sum of the parts."
To move the law very far toward the opposing pole, to wholly collectivize its characters, presents difficulties of its own. How much can the law convincingly take on board the "structural" features of mass atrocity? At what point do we end up succumbing to suspicions that no one is responsible because almost every- one is responsible, and that no one deserves to be singled out for remedial treat- ment as a victim because society at large is the "real" victim of massive violence?
Nothing would be lost, to be sure, in making collectivities central to the legis- lative reform of such large institutions as the military and police, whose members perpetrated the crimes. Civil suits alleging grave human rights abuse, moreover, always target sovereign states (one form of collectivity) and often seek remedies with a collective dimension, such as structural injunctions. Yet often it is not the state, but some other form of collectivity (for example, a nonstate military or- ganization, like the Lord's Resistance Army) that is the relevant wrongdoer. In international law, at least, the nonstate organization can be neither prosecuted nor civilly sued for atrocity damages. The law of "state responsibility" therefore becomes irrelevant. Leaders and their associates at all levels of the organization may be criminally prosecuted, to be sure, usually via the concepts of "control over an organizational apparatus" or "joint criminal enterprise." But the latter doctrine, at least-so prominent in international atrocity prosecutions today-suffers the same problems long recognized in dubious nineteenth century notions of "group mind" and "crowd psychology."
There is no single "right answer," much less an obvious one, to any of these ca- pacious queries, nor even to many of the more specific questions preceding and generating them, questions emerging from the more concrete controversies I have briefly mentioned. This would be true even if we could gather the full histor- ical archive for any such episode-which, of course, we cannot. We law professors like to summarily deflect all such questions through our potted, perennial rejoin- der: "Why do you ask?" Yet, in these situations, why one asks does not wholly determine how one may legitimately answer. There are several plausible responses to any of the preceding questions, each compatible with a number of reasons for posing them. Also, we sometimes have more than one reason for asking. For in atrocity's aftermath, people will be simultaneously concerned with deterrence, moral desert, incapacitation, and rehabilitation (of the lower echelons, at least). The "justices" sought may be at once retributive, restorative, expressive, transitional, corrective (compensatory or restitutive), even distributive. Each of these justice- forms will have differing implications for who the law should regard as the genuine perpetrators, victims, and beneficiaries. Some of the competing justices must inev- itably assume priority over others, and their respective imperatives will run mutually afoul.
In its evolving responses to mass atrocity, as noted, the law today oscillates un- certainly between what we might loosely call individualism and collectivism, or (more crassly) between the retail and wholesale levels of analysis. This is partly because for any given historical episode it is always possible to understand each of the central characters-the perpetrators, their victims, relevant bystanders, and the beneficiaries (of both the wrong and its later remedy)-in either way, if rarely with equal persuasiveness. All these personages have fluid contours. None is a natural kind. This simple fact opens a panorama of possibilities for how the law might conceptualize the wrong, assign responsibility for it, and seek its partial alleviation.
Each of these understandings makes certain assumptions. Let us call these sociological, though they are perhaps philosophical as well. Every such mode of apprehending atrocity invites not merely discrete factual inquiries into what transpired in a given place and time; it also implicates theoretical questions about more general processes likely at work within all these historical experiences, each bearing some family resemblance to the rest. To employ any of these understand- ings in our legal response to a given episode is to construct our case on the foun- dations of an implicit sociology, the merits of which will often be intelligently contestable.
We may favor a given collectivity over the next smaller subset because we think its boundaries more accurately correspond to the set of individuals most truly re- sponsible for the relevant wrongs. Or we may favor the bigger entity only because blaming it will better ensure that victims receive adequate compensation. If it has a broader revenue base, in particular, a given collectivity can more effectively spread the costs of compensation among a larger numbers of payers. Thus, all citizens of a state held liable for waging aggressive war may be taxed to pay reparations to the state victimized by the unlawful attack. If the reparations are onerous (as they were for Germany following the First World War), this will make many people angry, however; those who did no wrong later discover that they must pay the price of their compatriots' misdeeds. This may be a prudential reason to demur from such a course, though the law permits it.
To similar effect, in designing particular prosecutions, the law might treat the true victim of mass terror as collective in nature: the Ukrainian kulaks, the European Jews, the Cambodian upper-middle classes. Whatever the rationale for making this move, it makes a particular statement to all concerned. The legal system may tack this way because lawmakers understand the victims as inherently a collectivity-at some essential, irreducible level. Or the law may take this direc- tion merely because of a contingent determination that it is impracticable to do otherwise. Even when many individual victims survive, the "information costs" re- quired to identify and extensively communicate with them may be excessive. The harm suffered by the true, individual victims sometimes cannot be significantly alleviated, if at all. Thus, for instance, during the distribution of Jewish assets from Swiss banks, those who first established these accounts were obviously in no position to demand their life savings, with the decades of accumulated interest. Only some portion of their descendants, for that matter, came forth to reclaim this substantial legacy.
With these primary, preferred options clearly inadequate, a second natural al- ternative-invited by standard thinking in the law of remedies-would have been to seek to reestablish Jewish institutions and associated cultural life across Eastern Europe, where most victims had lived. For Nazi leadership had specifically intend- ed to destroy the communities that had generated and continued to transmit this shared cultural heritage and activity, not merely to kill individual Jews engaged in it. The desired victim was to be, and in fact proved to be, a distinct social collec- tivity, which the perpetrators understood in precisely these terms. That social en- tity would be European Jewry "as such," in the language of the Genocide Convention. Nazi leaders themselves accurately conceptualized the target of their exterminating fury as a flourishing population of many millions, rich with thousands of vibrant communal institutions across a pan-continental geography. Why should not the law then treat that very same social entity, victimized in this collectivizing fashion, as a remedial beneficiary, if only when more immediate vic- tims could no longer pursue their individuated claims?
Yet by the 1990s this too was unimaginable, notwithstanding the remarkable upsurge of philo-semitism in some such lands. There may have been some nostal- gic appeal to reviving "the world of our fathers." By this point in history, however, not even the billions of dollars suddenly available to the task could re-create the thousands of thriving shtetl communities populating much of the region as late as 1943. Collectivizing legal relief in this fashion, however desirable in theory perhaps, was simply impossible.
Such experiences reveal that where the law allows several ways to understand "the victim" of mass atrocity, and where substantial resources are available to those who would successfully claim that status, the question will almost certainly find its way into the courts. Yet disagreements of this sort need not entirely in- volve a grubby scramble for material resources once belonging to atrocity's genu- ine victims. They sometimes arise in large part from reasonable difference of opinion over the meaning of what truly happened.
Thus, some legal actors may sincerely understand the actual victim in collective terms, whereas other such actors-in a position to resolve the matter-sincerely think it simply more historically accurate to conceptualize the victims (now plu- ral) solely in terms of identifiable individuals (for example, Anne Frank, Aleksandr Solzhenitsyn, Osip Mandelstam). There may be little doubt, after all, about the essentially similar nature of their suffering-extreme misery and/or death through lengthy "camp" incarceration-when seen through a wide-angle lens, from the wholesale perspective. The law may nonetheless conclude that particular victims differ in ways normatively relevant to remediation, as in the measure of their compensatory recoveries. (Even in a unitary class action, within U.S. law, it is rare that every plaintiff within a large class of claimants recovers exactly the same amount in damages; here, virtually within the same breath, we both collectivize and individualize.)
In short, the legal decision-makers, acting fully within the law, could often plau- sibly choose to comprehend both the victims and perpetrators of mass atrocity in a variety of ways, as isolates or aggregates, and as aggregates of very different sizes or shapes. However they do so, courts must then construct a form of redress de- signed, again, either to particularize or generalize. They may adopt some mix of remedies falling variously at any number of points along this spectrum. A judge has to delineate, in terms broad or narrow, the proper boundaries around those she will regard as legitimate claimants to relief. These may turn out to range from a few isolated individuals to entire institutions, sometimes very large ones. The positive law itself, even when relatively well settled, often autho- rizes considerable judicial discretion, and so in practice offers little determinate guidance.
At times, judges will categorize the beneficiary as a still looser human constel- lation, largely of their own devise. This amorphous configuration may come into view only in the very course of the lawsuit itself. It becomes clearer over time, as facts emerge at trial, that many people or institutions not before the court bear some close, morally significant resemblance to actual victims who cannot be found or whose harms cannot be alleviated. In these situations, the equitable doc- trine of cy pres offers a common, well-established means for enlarging the scope of those who will benefit from the court's order.
At times, even clear victims of mass atrocity, though not usually the most im- mediate, may seek collectivized modes of redress. This approach may be used only partly as a fallback "Plan B" for when more individuated remedies fail. There may also be some genuine recognition that the group itself has suffered loss, in ways irremediable, even in the best of circumstances, through its analytic "reduction" to a number of particular members. These are perhaps the most poignant cases, because persons who are clearly individual victims choose (under some constraint, to be sure) to reinterpret their personal suffering as part of a wrong much larger than their own, a wrong best remediable in ways responsive to its inherently col- lective character. Hence, for instance,
The African American community does not press for reparations so that every African American will receive a reparations check in the mail. What is sought are better schools, better housing, and a pool of capital that will allow Black-owned businesses to flourish. The demands of the Herero People against Germany focus on the equivalent of a new Marshall Plan geared toward revitalizing a collective that was nearly wiped-out in the early part of the th century. Suits relating to Shell Oil's treatment of the Ogoni People seek broad collective remedies aimed at restoring the land of the Ogoni region of Nigeria and the Ogoni People's self-sufficiency. Among those advancing or support- ing suits against corporations that allegedly propped up the Apartheid regime are many who oppose individual remedies and favor instead that funds be spent on communal needs.
In any courtroom controversy (if such controversies ever get that far), counsel for the litigants will often seek to scramble around the variables I have been de- scribing in a number of ways, redraw the boundaries around the characters to match the evidence-and, of course, to serve their side to the dispute. Thus, for a particular episode of mass atrocity, judges may end up pigeonholing its perpe- trator as a collective actor while conceptualizing its victim as distinct individuals; the judges may then construct a remedial schema entirely collectivist in character. There are dozens of other readily cognizable permutations.
Constructing the Relation Between Past, Present, and Future in the Law's Atrocity-Response
Patterns of the Past-A Preliminary Sociology
We must first ask whether there exist discernible patterns in how the law tacks in one direction or the other-individuating here, collectivizing there-when grap- pling with episodes of mass atrocity, from early modern times through the present day. We may plausibly hypothesize that the law's responses are not random, as the interacting characters assume first collective then individuated incarnation. If em- pirical patterns become apparent, what accounts for them, sociologically speaking? Here, there are three variables to the equation. They interact in ways that make every national experience somewhat unique, of course. Yet the way they combine also discloses certain regularities, amenable to fruitful investigation, but one that has as yet not been seriously undertaken.
First, the applicable law will make several doctrinal options available to legal actors. On a given legal question, the choice-set may be wide or narrow. So the question becomes: How much may court or counsel, by adopting this or that cog- nizable formula, enlarge or contract at will the scope of legal parties to perpetra- tion, victimization, and remediation? The range of doctrinal possibilities sets the terms of legal discussion, limits the scope of disagreement over how to "define the situation," as sociologists say, and determines the parties to it.
Second, there are the larger normative goals of legal policymakers. For instance, do they chiefly valorize retribution against individual wrongdoers? If so, they will pursue criminal prosecutions, maybe many of them, against perpetrators at all lev- els of a miscreant institution. Or perhaps, instead, they will decide to prioritize collective reconciliation. Prominent criminal trials stir up old intergroup antago- nisms, and so imperil this objective. Policymakers therefore concentrate their ef- forts on establishing the legal machinery for a "truth commission" and related forms of "restorative" justice, as in Rwanda's gacaca.
Third, all pertinent legal actors-prosecutors, defense counsel, judges, and Justice Department policymakers-will obviously face situational constraints and opportunities, specific to time and place. These sometimes make it more cost- ly to individuate than to aggregate, sometimes the reverse. Among an array of potential defendants, for instance, individuals may be more powerful than the in- stitutions they controlled, or vice versa. Among potential beneficiaries, the same will sometimes be true, as in the Swiss bank litigation discussed above, where Jewish organizations lost out to a class of Jewish individuals. The incentives of prosecutors and counsel to civil plaintiffs alter in light of these disparities.
These three variables and the values attached to them in a given situation deter- mine how much flexibility exists, both within the law itself and in the circum- stances determining its use. Laymen may be surprised to learn just how much discretion legal actors enjoy in these situations and how much these extralegal contingencies determine its exercise.
Normative Goals for the Future
What might justify the decision to maneuver the law in one direction or the other, as it struggles to properly construe the wrongdoer, injured party, and beneficia- ries? Decisions about whether to individuate or collectivize the characters partly depend on what type of justice one seeks in the aftermath of dictatorship or mass violence. Here, we encounter only with "elective affinities," not logical entailments.
We may anticipate, for instance, that legal measures designed chiefly to prevent renewed atrocities-measures often loosely grouped under the label "transitional justice"-would gravitate toward more collectivistic responses. These would focus on reforming the institutions-usually the military and police-chiefly responsible for implementing evil policies, whereas contemporaneous demands for "corrective" justice, focused on redressing past harm, will often (not always) invite more individuated forms of judicial resolution. And where "retributive" jus- tice springs most saliently to our moral sentiments, the claims of criminal law will assert themselves with special urgency; and criminal law almost always conceives the perpetrator, as well as the wrong's victim and beneficiary, in highly individu- ated terms. The relative strength of these distinct moral sentiments, within a given society emerging from mass atrocity, will influence the direction of its legal response.
A second normative concern arises from the common moral intuition that those who benefit from legal redress, after atrocity or in any other circumstance, ought to be those who suffered from the wrong. In mass atrocity cases, more than in almost any others, the alignment between legal redress and those who suffered is often off, in ways that many find ethically troubling. This tends to happen where the law, sometimes unavoidably, assumes an individuating approach in defining one of the characters but a more aggregative angle on another. For instance, the law concludes that, though only "natural persons" have been the "true" victims of a given atrocity, a collective remedy should nonetheless exclusively issue, from which individual victims may benefit only indirectly, if at all. The question arises whether the remedy draws its resources through litigation or by other means.
Conversely, we should predict dissatisfactions as well where courts regard the "real" victim as a collective actor-a tribal "people," for instance-yet legal redress must target only individual members, many of whom may have suffered from no readily recognizable, legally cognizable harm. Examples of both scenarios regularly arise among the legal cases; in fact, these tensions often lie close to the surface of the social and political conflicts underlying and precipitating the litigation. That is notably the case with respect to "affirmative action" in the United States, initially devised as a sincere if deeply unsatisfactory remedy for centuries of slavery and Jim Crow.
Pragmatics of the Present
Whatever the law and whatever their goals, legal actors of course confront situa- tional constraints and opportunities. Seeing opportunity where others see con- straint is often the mark of political and legal intelligence. Circumstances that at first seemed to require individuated atrocity-response later yield-in the hands of savvy leaders-to the logic of collectivization instead. Thus, for instance, upon agreeing to accept Holocaust reparations from the Federal Republic of Germany, Israeli President David Ben-Gurion promptly pocketed it all ($89 bil- lion thus far) for the State's treasury. Holocaust victims themselves recovered nothing, though the Prime Minister had promised to use the funds to help resettle individual survivors. Though such state-to-state reparations were long (and largely remain) the only kind contemplated by public international law, today only highly "statist" societies, like socialist Israel of the day, would seriously attempt such fla- grant expropriation of their citizens' individual rights in tort and property.
Such political calculations are omnipresent when choosing a legal policy of atrocity-response. In identifying the "real" wrongdoer, the decision over whether to individuate or aggregate depends both on where the practical problem most seems to lie and on who (within its general vicinity) appears most politically vulnerable to legal assault. Thus, when the chief obstacle to post-atrocity justice resides in the person of a single individual, his criminal prosecution is provision- ally indicated-that is, unless he remains (like Gen. Augusto Pinochet, for a dec- ade after leaving Chile's presidency) politically untouchable by his country's courts.
Somewhat similar is the situation where the primary problem is located among a small to moderate number of individuals. Here, lawmakers perform the vetting of long-entrenched personnel, a process known as "lustration." This is the proce- dure first employed in Czechoslovakia, then widely across Central and Eastern Europe, to formally discharge high-level civil servants directly implicated in large- scale human rights abuse during communist rule. New democratic reformers in these countries concluded that the central problem was not deeply "structural." It did not reside in the civil service as such, requiring its abolition or fundamental reorganization. Rather, the threat to the democratic transition lay in the enduring ideological and personal loyalties of certain individuals still occupying the bureaucracy's top ranks.
By contrast, where the problem (as perceived) stems from the nature of an in- stitution, not its ephemeral personnel and policies, more collectivist modes of policy-response come prominently to mind. They incline us toward legislation, an executive order, or tort litigation against the responsible state. All three devices seek to create or reorganize the workings of collective entities, public or private. Change in the behavior of individuals laboring within them, regardless of their prior commitments and activities, will follow in train. In their responses to mass atrocity, certain countries move decidedly in one direction along this legal spectrum, others in the opposite, for reasons which will remain obscure until someone seriously investigates the question.
The Quixotic Quest for "Coherence"
It may be that no compelling answers will present themselves to any of the ques- tions here posed-not the empirical and descriptive, the explanatory, the norma- tively philosophical, nor the pragmatic and policy-driven. There would nonetheless be no small value in establishing the inconsistent contours of our am- bivalent juridical rejoinder to these deeply disconcerting episodes. For no one re- ally believes that the law can-in its apparent conceptual confusion-remain indifferent to events of such moral magnitude.
A clear sense of what a properly coherent response to these events might entail could at least give us criteria for evaluating particular instances of response, na- tional and international, judicial or otherwise juridical. Quite frequently, of course, a given body of law encompasses two opposing moral principles-usually, utilitar- ian and deontological-and "reconciles" them only insofar as it demarcates par- ticular pockets of rules within which one principle will prevail over the other. The result is something of a "patchwork quilt," in the language of critical legal studies.
Within atrocity law, the conflicting impulses, as I call them, involve something different. For these are not normative principles at all, but rather implicit socio- logical premises-almost ontological assumptions-about what mass atrocity is, about who the parties to it really were. Where legal actors differ over their implicit sociology of these events, they will differ as well over how the law should respond to them. If we should seek coherence in this response, it must be of a proper sort. The improper sort is clear enough: the law should not follow either path on all occasions, without exception, individualizing or collectivizing at every point, em- bracing one endpoint along this continuum, yielding nothing to its antithesis.