Precommitment theory and international law: Starting a conversation

Precommitment theory has already demonstrated its explanatory power in settings far beyond the interpersonal arena set forth by Jon Elster and Thomas Schelling a generation ago. Its most significant application in the legal realm has been in the area of constitutional law, though this application

This Essay offers a preliminary sketch in this direction. It attempts to ascertain whether precommitment theory offers useful tools for examining certain behavior by international actors. In particular, it explores the ways that states and international organizations seek to restrain their future conduct through treaties and other international legal undertakings. In so doing, it identifies the Elsterian precommitment as a special form of self-restraint by states, with legal consequences that formal doctrine may or may not already recognize. At the same time, the Essay also suggests that many forms of self-binding on the international plane are not bona fide precommitments. Nonetheless, the distinctions between those state actions that qualify as precommitments and those that do not can have important ramifications for the way international law and institutions should treat various commitments by states to constrain their future activity.

My Essay proceeds as follows. Part I offers a brief exploration of the connections to date between international legal scholarship and rational choice theory, the theoretical font of precommitment theory. Part II describes the workings of self-binding at the international level and the conceptual relationship between such binding and precommitment. In Part III, I offer five examples that show how precommitment theory can offer important insights for examining various international legal relationships. I conclude in Part IV with some suggestions for future research in this area.

I. International Law and Rational Choice

The systematic application of rational choice theory to international law-as opposed to international relations-is a recent phenomenon. The new insights that rational choice scholars have brought form part of a recent trend in interdisciplinary scholarship in international law. Rational choice has appeared in particular through two methodological innovations: the so-called "international relations/international law" (IR/IL) approach, and the movement of law and economics from the domestic plane to the international one.2 IR/IL, pioneered by Kenneth Abbott and Anne-Marie Slaughter, has brought to bear the full range of international relations theory to the study of legal problems. Of course international relations theory has many strands, but much of it includes rational choice concepts. IR/IL works have employed models such as collective action problems and the Prisoner's Dilemma to explain the reasons that states enter into various forms of international agreements as well as the modes by which they implement those agreements.3 Similarly, those relying upon economic approaches continue to make reference to rational choice concepts.4

At the same time, although the number of international law scholars doing interdisciplinary work is expanding, scholars applying rational choice theory remain relatively few in number.5 The situation might be compared to that in domestic law twenty years ago. On the one hand, international legal scholars-or at least good ones-are aware of the limitations of law in a world without a centralized authoritative legislature, executive, or judiciary. As all but the most rigid of positivists would recognize, political science-including political realism-must be a part of our thinking. Indeed, the greatest American contribution to the field of international law is one that brings legal realism into the international arena-the policy-oriented approach of the New Haven School.6 But that same sense that we must move beyond the formalism of European positivism has made many international lawyers, myself included, wary of theories that seek to explain all state behavior in rational choice terms. Although some models that derive from the application of rational choice theory (such as regime theory and institutionalism) have a great appeal to international lawyers, attempts to go much further by applying rational choice to other aspects of the international legal process have moved slowly.

II. Self-Binding at the International Level: Conceptual Foundations

A. A Typology of International Self-Restraints

The theory of precommitment offers a framework for understanding the reasons that, and the methods by which, individuals and communities seek to bind themselves in the sense of preventing themselves from having full liberty of action at a time in the future. Before analyzing the modalities by which states bind themselves, it is worth restating the meaning of a precommitment as elaborated by Elster. The core criterion of a precommitment is the purpose of the precommitting actor; precommitment is action done principally to restrain oneself from doing something that one would otherwise do because such restraint will itself directly improve one's future welfare. The need for restraint stems from a concern that one's passions, self-interest, preferences, or method of discounting time will cause one to do something in the future that is deleterious to one's welfare.7 This action I will call the simple precommitment.

Elster identifies one other form of self-binding as a precommitment even though the actor's purpose is somewhat different-the strategic precommitment, an idea that stems from Schelling's pathbreaking work. Here the individual binds himself, not out of fear of what he might do, but to make a credible threat to others that he will not do something. The key example is the army that burns its bridges behind it to prevent retreat and thus scare the enemy.8 The army might fear that, with the bridges intact, it would be tempted to retreat, but it might be confident that it would not retreat. Either way, the real reason to burn the bridges is to demonstrate the impossibility of retreat to the other side.

The general theory of precommitment has significant utility in explaining various forms of international commitments entered into by states and other international actors. By focusing on the temptation to violate and its relationship to self-binding, the theory invites a useful categorization of the reasons states enter into international legal obligations. I sketch these out below, with examples to follow. This Essay focuses on treaty obligations, although the insights of precommitment theory would appear to apply as well to customary international law-the uncodified form of international law that emerges gradually from the consistent practice of states coupled with a belief by them that the practice is legally required.

-First, some states sign treaties for a purpose that has little or nothing to do with the underlying subject matter of the treaty, principally to enhance their reputation or curry favor. That is, they expect a benefit from other states on a different issue in exchange for such a commitment.9 Many states, for instance, sign treaties banning certain weapon systems not because they want to prevent themselves from deploying them. They may not even have the capability, or they may have the capability but be insincere in seeking to prevent themselves from using them. Rather, they expect somehow to be rewarded by other states for opposing such weapons. If they end up carrying out their obligations, it will not be because of precommitment but because of some later decisions by themselves and others.

-Second, some states enter into treaties without particular concern that they might violate the provisions in the future, but because they want to ensure that their treaty partners observe the treaty when it is used against them. This reciprocity strategy lies at the core of many treaty relationships: state A commits to an agreement for the principal purpose of getting state B to commit to act in a certain way that benefits state A. Reciprocal trade or defense agreements are classic examples of such treaties. International relations theory has studied the advantages to states of imposing upon themselves higher costs to acting in the future instead of incurring the costs inherent in the absence of the treaty.10

-Third, and a variation of the second, some states sign treaties without concern that they will be tempted to violate the underlying norm but specifically to demonstrate their belief in a particular regime. In many instances, they are trying to get other states that would be so tempted to sign the treaty in question. Thus, many states entering into nuclear nonproliferation treaties have no intention of developing a nuclear weapon but want to encourage potential nuclear weapons states to agree not to develop the weapons.

These three potential stances of states entering into treaties seem quite prevalent. But they are not Elsterian precommitments as defined above, either the simple or the strategic form. They are not simple precommitments because states are not principally guarding against future preference shifts-trying to overcome future temptations on their part. They are not strategic precommitments because states are not engaged in unilateral self-binding for the sole purpose of gaining an advantage over other states. Rather, states are basically tying their hands solely for the purpose of, and as part (or in anticipation) of a deal in exchange for, a benefit from another actor. Their purpose for imposing those costs is not, in the end, to constrain their own behavior but to constrain or influence another actor's behavior. This purpose is similar to the idea mentioned by Elster when he quotes Jens Arup Seip to the effect that people do not try to bind themselves, but others.11

In response, it might be argued that, in the end, states enter into treaties constraining their behavior in order to benefit themselves; they are thus making a strategic precommitment. A broader argument, made by others in this Symposium, might even say that, because the states are tying their hands at Time 1 against action at Time 2, they are per se precommitting. Under either view, all treaties are precommitments, just as all contracts are.12 But these interpretations expand the meaning of precommitment to the point of uselessness or water it down to nothing more than an ordinary promise.13 There is a conceptual difference between, on the one hand, unilaterally giving up one's future choices to overcome future temptation (the simple precommitment) or to gain a strategic advantage over another (the strategic precommitment) and, on the other hand, trading in choices solely as part of a deal involving a limitation of choices by or a benefit from another. Both the simple precommitment and the strategic precommitment build on the notion from the Ulysses story of a direct connection between the self-binding and the behavior that the self-binder fears he would-or, in the case of the strategic precommitment, that he fears the adversary perceives he would-otherwise commit.14 In Arup's terms, a treaty whose purpose is solely to prevent someone else's disagreeable behavior seems different.

-Fourth, some states enter into treaties because they really do wish to tie their hands for fear of the consequences of their later choices. These states are making a true precommitment. They use the vehicle of the treaty-a commitment to others-because they see it as stronger than a wholly internal precommitment through constitutional law. Its form as a promise to other states does not per se make it any less of a precommitment. I offer examples in Part II below. Whether these true precommitments constitute a small or large proportion of international obligations will vary across subject areas. But whether the proportion of such bona fide precommitments is large or small, the factors that differentiate precommitments from other forms of commitment remain useful analytical tools.

At the same time, it should be briefly mentioned that these four categories are not hermetically separated from each other. A state, acting through its government, may have mixed reasons for joining a treaty. Thus, for instance, the government could enter into a treaty for reciprocity reasons as well as to demonstrate to other states its support for the regime; and a government precommitting against future temptation could also be hoping to demonstrate to other states its support for a regime. And, as discussed further below, different institutional actors within a government (as well as those outside it) will often have entirely distinct reasons for joining a treaty.

B. Precommitment and Sovereignty

The above categorization of state self-binding shows a variety of purposes and modalities for self-binding by states and suggests that precommitment, at least as defined by Elster, only captures one form of such behavior. Nonetheless, Elster's fundamental insight regarding the result of precommitment applies to all forms of state self-binding. That insight, that binding oneself actually furthers one's freedom in the long run,15 is one that international law understands (even if only relatively few scholars argue in explicitly rational choice terms). The typical precommitment limits the individual's ability to act as he chooses at a certain point-or many points-in the future. The international analogue of this concept of freedom to act independent of outside constraints-autonomy-is, of course, sovereignty. Sovereignty connotes a state's freedom to act as it chooses in the international arena. International law has long recognized that, when a state ties its hands through treaty commitments, it is furthering, rather than limiting, its sovereignty.

For instance, in a famous international chestnut, the S.S. Wimbledon, the Permanent Court of International Justice (P.C.I.J.), the predecessor to the International Court of Justice, considered Germany's reliance on sovereignty as a reason to escape the limitations placed on it by the 1919 Treaty of Versailles. Under that treaty, Germany accepted a permanent right of passage through the Kiel Canal for all vessels.16 During the war between Russia and Poland after World War I, however, Germany sought to protect its neutral status by denying transit to a British ship carrying arms to Poland. Britain, France, Italy, and Japan, with Poland intervening, brought proceedings against Germany in the P.C.I.J. for breach of the treaty. Germany argued that a state's ability to declare itself neutral was "an essential part of her sovereignty," and that, by signing the Versailles Treaty, Germany "neither could nor intended to renounce by anticipation" what it described as a "personal and imprescriptible right."17 The P.C.I.J. held against Germany and famously stated that "the right of entering into international engagements is an attribute of sovereignty,"18 and that, therefore, the limitations a state accepts under a treaty cannot later be renounced as impermissible infringements on that state's sovereignty.19

Modern treaty law thus rejects such excuses by states for failing to carry out treaties. Instead, treaties must be carried out in good faith.20 The law considers treaty commitments invalid only if they resulted from clearly fraudulent or coercive processes or if they violate a limited number of "peremptory norms" of international law-which clearly do not include the loose notion of sovereignty.21 Similarly, states must carry out obligations incurred under customary international law. The state may argue that the obligation is invalid or inapplicable to it only if it persistently objected to the norm while the norm was in the process of formation and, thus, implicitly never consented to the norm.22

International law's ability to see self-restraint as an attribute of-and, indeed, means of furthering-sovereignty corresponds with Elster's notion of precommitment as furthering individual autonomy. As Wimbledon makes clear (and as is obvious to international lawyers), international law enforces states' decisions to self-bind against later changes of mind (or heart) by classifying most such changed views as breaches of international law.23 International law prohibits such changes in order to allow states to make meaningful commitments in the first place; for if states could walk away from their commitments, no state would enter into them. Promoting self-restraint allows states to further their own long-term welfare as they choose.

Enforcing self-binding through rules that make breaches illegal makes international law essentially a system of proscriptions-what Calabresi and Melamed refer to as property rules.24 (International law does not work from the assumption that states may undertake unacceptable activities if they are willing to pay for them-the essence of a liability rule.) And, indeed, Calabresi and Melamed justify property rules (e.g., rules prohibiting the sale of organs) by invoking the Ulysses paradigm; they note that rules prohibiting the sale of entitlements foster what they call "self paternalism" which "allows the individual to choose what is best in the long run . . ., even though that choice entails giving up some short run freedom of choice."25 Although some have criticized international law as relying exclusively on proscriptions rather than including liability rules,26 treaties and international law's prohibition on breaches of them advance sovereignty in much the same way that self-binding and property-rule prohibitions protect individual freedom.

At the same time, although the term "sovereignty" continues to be used quite freely in international relations discourse, most international law scholars-and some international relations scholars as well-remain very put off by it.27 The reason stems from the way in which the term is invoked. Wimbledon laid to rest the acceptability of any claims by a state that sovereignty means complete freedom of action irrespective of prior legal commitments. Instead, states invoke it to argue for certain inherent limits on what international law can regulate. Thus, for instance, China routinely condemns both other states and the United Nations for criticisms of its human rights records, decrying such views as impinging on its sovereignty.28 Yet a core set of human rights norms is well established in customary international law, and the UN's authority to criticize states based on such norms is equally well accepted. Thus, sovereignty has become a subterfuge for a basic disagreement with the underlying norm. I will thus prefer the term "freedom of action," closer to that used by precommitment theorists, to the term "sovereignty."

C. Two Caveats

Before considering some examples of precommitment theory's utility for understanding certain elements of international lawmaking, it is critical to identify some obstacles that stand in the way. By obstacles, I mean not actual barriers to the process of precommitting or substantive disadvantages to precommitting, but theoretical obstacles to the application of the theory. The most fundamental obstacle, repeatedly identified by Elster in the context of constitutional law, concerns the difference between individuals, on the one hand, and societies, polities, constituent assemblies, or other groups, on the other, in terms of who is precommitting whom.29 In the international arena, this insight applies in spades. The so-called "international community," while a useful rhetorical and normative construct, is no more than that-a construct. It does not correspond to some clearly identifiable set of individuals, groups, or nations. At a minimum, the group to which those employing the term are actually referring, namely the Western powers, are not the full "international community."30 In a system as diverse as the international arena, a set of actors may in fact be seeking to tie the hands of others, rather than their own hands. Only the latter constitutes a bona fide precommitment;31 indeed, the full "international community" may rarely engage in this self-binding.

This factor does not make precommitment irrelevant to the international sphere any more than it makes precommitment irrelevant to situations involving domestic communities as opposed to individuals. But it does make it more complicated and sets up at least three possibilities for precommitment: First, sometimes a group of states is trying to constrain itself-the group-from acting a certain way. Second, at other times, the group has made the decision, but it is only trying to constrain certain of its members from acting a certain way in the future. Third, at still other times, the group is acting jointly, by treaty or otherwise, but in fact, the purpose is for each member of the group to constrain itself individually in the future.

Even within the state, different actors can have different purposes for entering into treaties. Governmental actors may not see the treaty as self-binding, but nonstate actors, and nongovernmental organizations in particular, may view the treaty this way. Or some governmental actors may have no concern about what the state might do in the future, while others might have such fears. As a result of disaggregating groups of states and groups within states, we will discover that, in fact, some precommitments are Janus-faced insofar as they can be simultaneously viewed as reciprocal bargains and self-restraints.

A second factor rendering the theory's application in the international realm more complex is the problem of enforcement, one central to international law. The law of nations has numerous mechanisms of control. These range from internal self-controls based on reciprocity and fear of adverse precedent;32 to the mobilization of shame by powerful international institutions, both intergovernmental and nongovernmental, who can impugn the reputation of states; to various forms of retaliatory sanctions of the diplomatic, economic, and (rarely) military kind. But clearly, the possibility of precommitments (made by whomever) being observed is limited by the absence of much traditional hierarchical enforcement. Ulysses may have needed others to tie him up, but, more importantly, once tied up, he was powerless to release himself without the help of his shipmates. States making such precommitments in the constitutional process will not have Ulysses' physical constraints, and those acting at the international level may be even less bound by their decisions.33

III. Self-Binding Under International Law: Some Seminal Examples

At this early stage of the endeavor to apply precommitment theory to international law, it seems best to posit a handful of important examples that suggest how the theory may help us understand the ways in which states and other participants in the international legal process make, interpret, and carry out international obligations. I offer five examples below from different subject areas of international law.

A. Precommitting to Borders: Uti Possidetis

A first example concerns a classic issue in international law-the norms for determining a state's borders with its neighbors. Under the principle of uti possidetis, states emerging from decolonization are generally obliged to accept the borders set for them by their ex-colonial power.34 This norm has been applied by newly independent states throughout Latin America, Africa, and Asia (with some notable exceptions). Thus, for example, the border between Chad and Niger-two former French colonies-corresponds to a French colonial administrative line; whereas the border between the states of Niger and Nigeria corresponds to a French-British imperial border. These borders had, in most cases, been drawn a century or more earlier with no regard for the will of the local population and little concern for pre-existing boundaries of tribal or other ethnic entities. Latin American states accepted this doctrine as a matter of practice in the early nineteenth century;35 the African states adopted it in a core early resolution (the Cairo Declaration) of the Organization for African Unity (OAU).36

The OAU accepted uti possidetis for two reasons, with two distinct audiences in mind: first, to make sure that states in Africa did not pursue territorial claims against neighbors where the latter's population included-on their side of the border-ethnic kin of those in the former; and second, to make sure that those ethnic groups within states did not try to break off and join a neighbor or become independent.37 Either prospect could have been very destabilizing for the new states, and the immediate postcolonial governments knew that there were no obvious borders to replace those that the colonial powers had left them. (The second rationale was of less concern in Latin America, though those states had an additional fear in the 1820s-that other non-Latin American powers, whether the United States or European states, might try to take over parts of their territory.)38

Adoption of uti possidetis presents an excellent example of a precommitment in that the members of the OAU were first and foremost constraining themselves, not seeking to constrain others. Given the nature of the colonial-era lines, nearly every state in Africa had a border that could have been challenged by one of its neighbors. The group was agreeing as a whole to commit all its members for the purpose of overcoming any temptation to do something in the future against the interests of Africa as a whole. Although the precommitment does not have a particularly robust enforcement mechanism, the OAU for many years made one of its central purposes the peaceful resolution of any potential border disputes.39 Even the recent wars in central Africa among the Democratic Republic of Congo, Rwanda, and other states are not really so much about changing borders as they are about changing regimes. Indeed, in the half-century since the decolonization of Africa began in earnest, no border has changed through unilateral action (although one wholly new state has emerged through secession, Eritrea).40

At the same time, the uti possidetis example highlights the need to disaggregate the precommitting entity. First, one might view each member of the OAU separately and the entire Cairo Declaration as a deal whereby each state agreed to give up something (encroaching on its neighbors' territory) in exchange for something else (a promise by its neighbors not to encroach on its territory). In that sense, it could be seen as not really (or not only) a precommitment in the Elsterian sense-the fourth category above-but a bargain along the lines of the second category above.

Second, the OAU policy was not only about a group of states restraining its behavior (or the individual states so doing), but also about Africa's new leaders sending a message to minority groups who saw ethnic kin on the other side of the border not to contemplate secession or border movement. Jeffrey Herbst has written that the core element of the OAU policy was "if an African government is in control of a capital city, then it has the legitimate right to control the nation-state."41 In that sense, the African leaders, the often handpicked heirs to colonial governors, saw the doctrine as constraining others. Yet even this constraint of others turns back to a restraint on those leaders; by telling potentially seceding groups not to attempt secession, they were also telling themselves not to accept the calls for assistance that seceding groups in neighboring states might make.

During the more recent breakups of Yugoslavia, the Soviet Union, and Czechoslovakia, the newly independent states, as well as outside states and international organizations reacting to these dissolutions, generally applied the doctrine as well. Thus, for example, the governments in the former Yugoslavia and the former Soviet Union agreed that their new borders would be the borders between their constituent republics as set many years ago by Tito and Stalin, respectively; outside actors endorsed this strategy.42 The elites of the new states made the precommitment not to adjust the interrepublican borders for essentially the same reasons as did the African states-to avoid future border wars.

But here the deconstmction of the precommitment becomes essential, for not all actors in the process participated in or ever accepted the decision to apply uti possidetis. In particular, the leaders of the Bosnian Serbs and their allies in Serbia were very much unwilling to make such a precommitment. They wanted to change the border between Bosnia and Serbia to allow Bosnian Serbs to remain in a state where Serbs were the majority (namely Serbia) rather than a minority (as they were in Bosnia). When actors in the European Union and the UN insisted on the inviolability of the interrepublican borders, the Bosnian Serb army went about conquering as much territory as they could in Bosnia to create their own state there-an activity we now call "ethnic cleansing." Croatia and ethnic Croats in Bosnia did much the same on a smaller scale. In the 1995 negotiations to end the war, outside states still formally rejected the creation of a new state, but they allowed the creation of an autonomous Republika Srpska within Bosnia (though with slightly less territory than the Serbs had actually conquered).43

The Bosnian war and its outcome demonstrate clearly the consequences of an attempted precommitment where one of the actors whose cooperation is key fails to accept the precommitment and outside actors are not prepared to enforce it. Diplomats participating in the attempts to avert war in the former Yugoslavia understood this barrier to making the precommitment effective. They criticized the European Community and UN Security Council's insistence on uti possidetis as removing an important option from their tool kit-renegotiation of borders to more closely correspond to the wishes of the local populations of the former Yugoslavia. For example, David Owen of Britain even used the language of precommitment theory in stating how this decision "put all peacemaking . . . within a straitjacket that greatly inhibited compromises . . . ."44 I have argued elsewhere that uti possidetis does not, as a normative matter, make sense for dealing with the disintegration of states today.45 But either way, precommitment theory helps us better understand the actors employing uti possidetis, their motivations, and the prospects for its success.

B. International Human Rights and Humanitarian Law

As Sanford Levinson's paper on torture seeks to demonstrate, one significant precommitment by states in international law is their agreement to refrain from using the full coercive force of the state in situations where they might be most tempted to do so.46 In human rights law, governments specifically agree that they will resist temptations to institute various inflictions on human dignity, principally against their own citizens, based on short-term considerations of the national interest. These governments have, essentially, defined their long-term goals in terms of guaranteeing basic liberties and do not wish to do otherwise based on other goals. In humanitarian law (which significantly predates human rights law), governments agree to forbear from certain actions against enemies on the battlefield (the Law of the Hague) or against those persons not taking an active part in hostilities (the Law of Geneva).47 The substantive restrictions on state power in these two bodies of law are quite different, with human rights law covering a far greater range of state conduct. But they are still both quite broad; and, moreover, they do overlap, for example, in their joint commitment to prohibiting states from torturing.48

1. The Degrees of Self-Binding.-Beyond the breadth of the precommitment in terms of the range of actions restricted, human rights law and humanitarian law provide different depths of precommitment. Both provide essentially three levels of precommitment. First, human rights law and humanitarian law both provide a total precommitment not to undertake certain grave acts against human dignity. For these actions, no threat to the state, in war or peace, however severe, is considered sufficient to permit these acts. Thus, the International Covenant on Civil and Political Rights (ICCPR), the principal global human rights treaty, recognizes certain nonderogable rights, e.g., the right to life, the ban on torture, the ban on slavery, and the ban on ex post facto criminal liability. Not even a "public emergency which threatens the life of the nation" permits abrogation of these rights.49 With respect to international humanitarian law, Article 3, common to the four Geneva Conventions of 1949, prohibits certain acts "at any time and in any place whatsoever" against persons taking "no active part in the hostilities" regardless of the nature of the conflict (interstate or internal) between the belligerents. These prohibited acts are violence to life and person, taking of hostages, outrages on personal dignity, and sentences and executions without basic judicial guarantees.50 Other serious offenses that are never justifiable are listed in each of the Geneva Conventions as "grave breaches," violations that engender individual criminal responsibility as war crimes.51

Second, human rights law recognizes other rights that may be derogated from in times of national emergency, but otherwise must be respected. This might be called an intermediate precommitment strategy. Thus, for example, the numerous procedural rights of defendants-such as the right to counsel, the right to a speedy trial, and the right to cross-examination-are afforded this status in the ICCPR.52 In humanitarian law, most of the provisions of the Geneva Conventions do not contain the strong language of Common Article 3 nor are the violations of these provisions always regarded as war crimes.53 As a formal legal matter, states cannot violate these other provisions, but their derogable (or non-war-crime) status does suggest somewhat of a looser commitment to them.

Third, both human rights law and humanitarian law recognize that certain aspects of human dignity can be constrained, but only to further a legitimate goal and in a proportionate manner. In this sort of weak precommitment, states have agreed that they will not limit individual freedom unless they can justify the government's restrictions based on specific criteria. Thus, for instance, the European Convention on Human Rights permits states to limit freedom of expression in order to further "national security, territorial integrity or public safety, for the prevention of disorder or crime, for the protection of health or morals, for the protection of the reputation or rights of others," and other reasons, as long as the limitation is "prescribed by law and necessary in a democratic society."54 The latter criterion is one of proportionality, familiar to American jurisprudence too. On top of this, the European Court of Human Rights has developed the idea of a "margin of appreciation," a doctrine that permits states significant discretion to determine, with respect to many human rights, the scope of the national interests that justify limiting those rights and the closeness of the fit between means and ends.55 In humanitarian law, states can limit many more individual liberties, but, for some, they can do so only if it helps maintain public order. Thus, for instance, the law on occupation of territory in the Fourth Geneva Convention (on protection of civilians during wartime) states that real or personal property of civilians shall not be destroyed "except where such destruction is rendered absolutely necessary by military operations" and that, "[s]ubject to imperative reasons of security, protected persons in occupied territories shall be permitted to receive the individual relief consignments sent to them."56

2. Definitional Problems.-But are these treaties precommitments after all? The four categories of state self-binding developed above from precommitment theory offer a useful typology for considering why states enter into human rights and humanitarian law commitments.

The first category above describes states that sign human rights treaties with little concern for human rights, but in order to gain advantages from other states. (Certainly, this explains the large number of states that are parties to the ICCPR but, nonetheless, have terrible human rights records.)57 This posture of states may not be undesirable overall for the protection of human rights, as outside actors can then embarrass these states into complying with their obligations; but it is not, as noted, a true precommitment strategy.58 One subcategory within this group are states entering into human rights treaties as part of a deal that provides them some other, non-human-rights benefit. Thus, states in Central and Eastern Europe joined the European Convention on Human Rights because (at least in part) the European Union made such membership a requirement for EU admission.59

The second category above-reciprocity as the purpose for commitment-applies to many states adhering to humanitarian law treaties. States in the humanitarian law area sign such treaties to help ensure that other states will respect humanitarian norms in conflicts against the former group. Thus, the United States has the strongest interest in a robust humanitarian law because it is most likely to deploy forces abroad. (Similar motivations drive states to sign treaties on diplomatic and consular relations, which protect their diplomats and citizens from mistreatment by other states.)60 The reciprocity purpose does not explain a state's decision to join human rights treaties, however, because the principal beneficiaries of these agreements are a state's own citizens, not other states.

The third category applies to those states that sign human rights and humanitarian treaties without fear they will violate them, but to show their commitment to human rights and humanitarian law and to encourage states that would be tempted to violate the norms to sign the treaties.61 The treaty commitment is merely the legal vehicle to advance the cause of human dignity, particularly regarding the practices of other states. The United States, for example, with its ever-confident attitude about its own human rights practices, eventually became party to the ICCPR for these reasons-to promote human rights abroad.62 In that sense, I disagree with Professor Levinson's characterization of the U.S. ratification of the ICCPR as a precommitment,63 although some NGOs certainly saw it that way. (Moreover, a state's prediction of its future propensity to respect the norms may prove inaccurate.) With respect to international humanitarian law, those states that rarely deploy their own military forces and thus gain no reciprocal benefit from compliance nonetheless consider the treaties important vehicles to constrain the behavior of others.64 The rationales in this category and the previous one are not mutually exclusive. States may have not only an innate belief in the international humanitarian law regime, but also a fear that violations will inure to their immediate detriment.

Fourth, some states presumably enter into human rights treaties because they really do wish to tie their hands for fear of the consequences of their later choices. The treaty, as a promise to others, increases the costs to them of changed behavior. A clear example would include states in transition from autocracy to democracy that want to lock in human rights commitments as a way to overcome temptations by future governments to renew human rights abuses.65 As noted above, outside observers identifying such state behavior as a precommitment need to take account of the possibility that not all actors within the state will have the same purpose. Some governmental officials may have little worry about the state's possible future threats to human dignity, while those inside and outside government more suspicious of the security forces might value a precommitment. Indeed, within societies that have less-than-ideal human rights records, NGOs push for accession to the treaty as a form of precommitment; they might find allies within parts of the government. One might also posit a hybrid category of states making a bona fide precommitment and also seeking to encourage other states to accept the norms.

3. Differentiating Treaties.-Precommitment theory may offer new insights into understanding a major debate in international law concerning the permissibility of reservations to human rights treaties. The debate has resulted from the tendency of many states that join such treaties (including the United States vis-a-vis the ICCPR) to issue numerous or broad reservations at the time of ratification, which keep significant parts of the treaty from applying to that state.66 Treaty law permits states to make reservations when they become parties to treaties, as long as those reservations are not prohibited by the terms of the treaty and are not incompatible with the object and purpose of the treaty.67 On one side of this debate, the UN's Human Rights Committee, an expert body established under the ICCPR to monitor states' implementation of the treaty, has stated that human rights agreements are fundamentally nonreciprocal obligations aimed at protecting private interests. As a result, it asserts that reservations to the ICCPR should be viewed with much more scrutiny than reservations to other treaties and that the normal rules for determining the validity of reservations, under which validity turns on the reactions of other states parties to the reservations, should not apply.68 The Committee has gone so far as to insist that it alone has the right to determine the validity of reservations.69 On the other side, a number of major states-such as the United States, the United Kingdom, and France, as well as the UN's International Law Commission (a body of legal experts that principally helps the UN prepare treaties)-insist that human rights treaties are just like other treaties and that reservations should be presumptively valid unless they are rejected by the parties to the treaty.70

If human rights treaties are genuine precommitments, signed by states for the express purpose of tying their own hands (the fourth category of self-binding above), then the Human Rights Committee may have a point in that the legitimacy of a state's decision to limit that precommitment through reservations ought not turn on the reaction of others to the reservations. (Of course, the Human Rights Committee wants to arrogate to itself this determination, even if it were a bona fide precommitment.) On the other hand, if human rights treaties are just a set of reciprocal obligations (the second category of self-binding above), notwithstanding that their principal beneficiaries are a state's own citizens, then the views of other states become critical. Ryan Goodman has recently examined the ways that the differing motivations of states in joining human rights treaties can affect the consequences resulting from illegal reservations to those treaties.71

C. Unilateral Statements and the No First Use Pledge

For many years, international law has accepted that states can make binding commitments without becoming parties to treaties or contributing to (or acquiescing in) the development of customary law. Most significant is the unilateral statement regarded as binding. The black-letter law on this issue was famously set forth by the International Court of Justice in the 1974 Nuclear Test cases.72 In those proceedings, Australia and New Zealand sued France over its above-ground nuclear testing in the French South Pacific. After the Court ordered France in 1973 to cease the testing pending a decision on the merits, France refused to do so, but it soon issued a variety of solemn pledges-in public from the highest levels of government-that it would cease such testing after 1974. The Court found these unilateral declarations to be legally binding on France, allowing the Court to dismiss the case as moot.73 The Court found France's intent critical, and France clearly entered into this commitment with the intent of tying its own hands.74

But before identifying these unilateral pledges as a precommitment, we should ask, as we did with the human rights treaties above, whether France's purpose was really to prevent itself from testing in the future, in which case we have a bona fide precommitment, or whether it was simply to curry favor as part of a bargain with opponents of the testing (e.g., Australia, New Zealand, and other South Pacific states). Was France concerned that it might be tempted to test nuclear weapons in the future? Though it is possible that certain elements of the French defense establishment had such a temptation that the government as a whole wanted to overcome, the government made clear in its public statements that it no longer had any need for atmospheric testing and could continue its program underground (which it immediately did).75 If France no longer had such a temptation, then its action was not a precommitment; rather, it was an attempt at a reciprocal bargain along the lines of the second category above-though a bargain that France did not announce publicly and that the ICJ refused to see.76

If France was, however, actually agreeing to constrain itself from doing something that it was otherwise tempted to do (conduct atmospheric testing), tying its own hands was nonetheless not France's ultimate motive. Rather, France had a strategy to score political points with the states of the South Pacific that were uniformly opposed to its atmospheric testing. Does that make it less of a precommitment? If France's self-restraint was inner-directed and not part of a bargain to "bind others" (in Seip's phrase),77 it would seem that one is still faced with a genuine precommitment. Some people, after all, throw out cigarettes to stop themselves from smoking with an ultimate goal of impressing friends, not improving their own health. That does not make their actions less of a precommitment.

A key contemporary example of self-binding concerns pledges not to use nuclear weapons first. In 1995, upon the renewal of the 1968 Nuclear Non-Proliferation Treaty (NPT),78 all five declared nuclear weapons states issued unilateral statements that they would not use nuclear weapons first against non-nuclear weapons states that were parties to the NPT except if they (the nuclear weapons states) or their allies were attacked by such a state in alliance with a nuclear weapons state. (China did not include the exception in its statement.)79 Thus, for instance, if Iraq were to attack Israel alone, not using nuclear weapons, the United States would be precluded from using nuclear weapons against Iraq in response. Here one might see the five states as undertaking a genuine precommitment to bind themselves in the case of an attack when they might be tempted to use nuclear weapons. The motivation is not to constrain the behavior of others-indeed, the precommitment actually increases the likelihood that Iraq would attack Israel with non-nuclear weapons by eliminating one method of deterrence-but to constrain the behavior of the five.

At the same time, another reading is possible. The five may not be concerned about their own temptations to use nuclear weapons because they generally see almost no circumstances under which they would use them.80 Rather, they are simply trying to get non-nuclear weapons states to refrain from developing nuclear weapons by guaranteeing their immunity from nuclear attack.81 In this case, their purpose seems similar to that of nuclear states signing the 1996 Comprehensive Nuclear Test Ban Treaty,82 where states seemed much more clearly motivated to bind others, accepting (or, in the case of the United States, still not accepting) self-binding as the price to pay.

Unilateral pledges such as France's in 1974 and the nuclear weapons states' in 1996 are quite rare insofar as states treat such pledges as legally binding. International legal doctrine does not now focus on the criterion of precommitment-overcoming temptation-but it should. The determination as to whether these unilateral undertakings are a precommitment paves the way for understanding, among other things, the prospects for the success of these undertakings in the long run and strategies for promoting that success. We may discover, for instance, that, if states are internally motivated to restrain themselves, outside actors can appeal to one set of domestic interests to preserve the commitment, whereas, if they have simply entered into a bargain, other forms of leverage are required. Precommitment theory thus may offer international decisionmakers a way of identifying these pledges and distinguishing them from others.

D. Amendment of the Constitutive Instruments of International Organizations

The charters of international organizations typically provide for an elaborate process of amendment, one that often requires approval of both a small group of important states as well as some high percentage of the membership as a whole. In the case of the United Nations, Article 108 of the UN Charter states that all amendments must be adopted by two-thirds of the members of the General Assembly and thereafter ratified by two-thirds of all the UN's members, including the five permanent members of the Security Council.83 This procedure has meant that, in fifty-eight years, the Charter has been amended only three times, the last of which entered into force in 1973. Each amendment concerned modest expansions to the size of the Security Council or the Economic and Social Council.

For the past decade, the UN's members have been discussing a set of amendments that would further expand the membership of the Security Council, but have been unable to agree on anything. The sword of Damocles created by the requirement of concurrence by the Permanent Five means that sensible proposals, such as eliminating permanent membership for ex-world powers like France and Britain, have no chance of passage. Moreover, many developing states oppose proposals to grant permanent (but not veto-wielding) status to a select number of such states, e.g., India, Indonesia, Brazil, and Nigeria, for fear that these states would not advance the interests of small states any better than does the existing arrangement.84

The amendment procedures are precommitments in the same sense as domestic constitutional amendment procedures. They make it hard for the member states to alter their collective constitution. Certainly, if the 1945 San Francisco conference at which the Charter was concluded is viewed as the embodiment of a continuing international community, that community was trying to prevent tinkering with a core constitutive document. But, as Elster makes clear with constitutional law, the nub of the story is about different groups within the constitutive (or constituent) assembly.85 In fact, the San Francisco conference was characterized by debates between the Big Three (or Five, if France and China are included) and other states over their relative power in the UN. The Charter's requirement of Permanent Five concurrence to amendments was a successful move by the Big Three to tie the hands of the smaller states in perpetuity. Given their power in San Francisco, the other states had no choice but to oblige.86 Thus, for the United Nations Charter at least, the amendment procedures may simultaneously be viewed as a bargain among the states.

E. The Security Council Veto

Finally, and related to the amendment procedure above, the veto held by the five permanent members of the Security Council seems like a precommitment, although it is also Janus-faced. On the one hand, the Charter makes clear that the Council cannot make any nonprocedural decision unless nine members vote in favor of the decision and China, France, Russia, the United Kingdom, and the United States vote yes or abstain.87 The positive spin on this provision of the Charter is that its drafters foresaw that any UN Security Council resolution-the only resolution of any organ of the UN that can be per se binding on members88-that did not have the support of the Permanent Five would be a useless piece of paper. For the UN to legislate in opposition to the key loci of effective power would have led to a repeat of the League of Nations' experience. In that light, one might regard the veto as a genuine precommitment by the drafters at San Francisco to avoid the temptation to make something that is law under the Charter but that is not law in the meaningful sense of having some effective power behind it.89

This view of the veto as a precommitment by all the UN's members, however, gives short shrift to the history of the Charter and the current views of the Permanent Five. In fact, as with the amendment procedure discussed above, the veto is centrally about the Permanent Five protecting themselves by constraining others. They did not want the Council to have the authority to make any binding decisions regarding international peace and security without their assent.90 Critics of the veto are correct in pointing out that the veto entrenches the power of five states, three of which no longer have the relative power they did in 1945. Although this account does not itself undercut the above view that the veto is, in the end, not merely inevitable but useful (to avoid making the UN a paper tiger), it does mean that the veto can equally be viewed as an attempt by some states to tie the hands of others.

The resistance of many states to the veto, both at San Francisco and ever since, has meant that states in the UN have devised many strategies to get around it. Once decolonization proceeded to the point that the developing states had a majority in the General Assembly, they tried to shift decisionmaking to that body and convince states that Assembly resolutions, though nominally only recommendations under the Charter, could be binding as reflections of customary international law. This practice was especially prevalent regarding foreign investment. Large majorities of the Assembly voted for legal-sounding resolutions that would permit developing states to pay no compensation to Western investors after expropriating their property as part of nationalization schemes.91 Even the veto-wielding states have sometimes seen the need to overcome their precommitment and bypass the veto in the Council. The United States, when it saw the Soviet Union use the veto to block Council action during the Korean War, endorsed a strategy whereby issues of peace and security could be shifted to the General Assembly in the event of a deadlock in the Council.92 The United States deployed the same bypass again against France and Britain in 1956, after they had vetoed Council action during the Suez War. And NATO's decision to use force against Yugoslavia regarding actions in Kosovo in 1999, without Security Council authorization, stemmed from the virtual certainty that Russia would have used its veto in the Council. The 2003 U.S. intervention in Iraq followed a similar pattern once France promised a veto of any authorization of force.

IV. Prospects

This Essay has demonstrated some of the potential and pitfalls of precommitment theory for understanding international law. It does not do justice to other areas of international relations where states tie their hands through law, but which are not quite central to international law itself. For example, under the terms of fast-track authority in international trade, the U.S. Congress precommits not to amend trade agreements negotiated by the President. It is not trying to tie anyone's hands but its own.93 The idea is to strengthen the President's hand during negotiations by giving him the ability to tell negotiating partners that any tentative agreement worked out by the negotiators cannot be amended by Congress. This process is, of course, an interesting twist on the idea of burning bridges to advance one's strategic interests. One might think that the United States would do better in trade talks if the U.S. negotiators could say their hands were tied by Congress, which might insist on changing draft texts if the negotiators did not meet congressional demands. Instead, the Executive Branch's enthusiasm for fast-track authority suggests that it believes it will do better if it denies itself this card-a case of a precommitment by Congress interacting with a precommitment by the Executive Branch.

Precommitment theory thus offers a set of tools for examining the reasons and modes for self-restraint by international actors. Its significance for international actors and scholars can be identified at three distinct levels. At a first level, it isolates important factors inherent in self-restraint and paves the way for a typology of such restraint. The theory also permits us to see that the same act of self-restraint may be viewed in completely different terms, as a precommitment or as something else, by both those participating in it and outsiders appraising it.

At a second level, the theory and the categories it sets up help international actors and scholars understand whether and how to maintain self-binding as states' temptations to violate their self-binding become strongest. For example, states entering into treaties as bona fide precommitments face certain internal motivations to comply. As a result, the parties to those treaties may be able to devise specific strategies to support one another in keeping those precommitments or responding to violations. States entering treaties solely in exchange for benefits provided by other actors may well require other forms of influence from outsiders to keep them in line. We may also be able to predict different consequences resulting from violations. As a result, the strategies individuals use to enforce their precommitments can be examined alongside mechanisms by which international law induces compliance to show which international precommitments are likely to be honored and which rejected.

In this context, Elster's recent work examining both the barriers and disadvantages to precommitment is worthy of application to international law.94 Indeed, the ex ante barriers, such as strong, standing passions or the immovable interests of the potential precommitters, may have much to say about when states will choose to precommit in international law or why some attempts at precommitment (like uti possidetis in the case of Bosnia) fail. As for the disadvantages of precommitment, certainly those who see the veto as a precommitment by the world community of 1945 would say that the changes in the composition and complexion ofthat community in the ensuing fifty-eight years makes such a precommitment completely anachronistic.

At a third and uniquely scholarly level, precommitment theory helps to fill an important gap in international law and international relations theory-namely the propensity of scholars to examine the reasons states enter into treaties and the reasons states comply with them as two almost wholly separate issues. Much recent scholarship on these issues seems to assume that, at Time 2, states are faced with a set of factors, internal and external to them, that determine whether they will comply with a treaty, but their conduct is not affected by their own reasons for accepting the commitment at Time 1. Precommitment theory properly rejects this dichotomy. It informs us that one important, perhaps even the most important, factor in a state's propensity to comply with a treaty in situations where it is tempted to violate the treaty is the original purpose for entering into the treaty.95

Future work on precommitments will need to explore the ideas in this paper in more detail. Ultimately, the value of the theory will be its ability to provide explanations and prescriptions in international law that other approaches, whether from law or international relations theory, do not. The utility of the definitions of precommitment (mine or others') and of the distinctions between genuine precommitments and mere illusions thereof will turn on the extent to which such concepts delineate interesting and important categories of legal relations among actors in the international arena. It may well turn out that very few international norms are true precommitments. But the inquiry remains beneficial if we end up scrutinizing international behavior through a new and sharper lens.