Printer Friendly

Rational custom.


Customary international law is understood to require that state practices be followed from a sense of legal obligation, though international lawyers have long puzzled over how those obligations come into being. Recent work applying rational choice theory suggests, unsettlingly, that the entire inquiry is misconceived: practices commonly attributed to obligations are merely behavioral regularities that arise from intersecting state interests, and the role of legal obligations is minimal at best. This Article attempts to explain how the rational choice critique and traditional doctrine may be reconciled. Rational choice theory, it is argued, responds to a genuine problem in the existing understanding of customary international law; if properly leavened, the approach provides a useful and insightful tool for analyzing the relevance of law to state practices--albeit one confirming suspicions that custom is not so widespread as may have been supposed, and may prove less resilient too. But expanding the range of games used to analyze state behavior, and understanding the interdependence of those games, also shows how the rational choice and traditional doctrine are not so very far apart--and how legal obligations may have an instrumental role to play in each.
  I. The Irrationality of Customary International Law
 II. Rationality without Law: The Rational Choice Critique
     and Its Limits
     A. The Rational Choice Critique
        1. Coincidence
        2. Coercion
        3. Cooperation/Prisoner's Dilemma
        4. Coordination
        5. The Games' Implications
     B. Conventional Objections
        1. The Empirical Claim
        2. The Categorical Claim
III. Resurrecting Custom's Rationality
     A. Toward Rational Custom
        1. Coordination and the Case of the Territorial Sea
        2. Assurance and the Case of the Continental Shelf
        3. Specifying Opinio Juris
     B. Gaming Reputation
 IV. Implications for the Future of Custom
     A. Custom's Modest Potential
     B. Reforming the Customary Rules of Custom


It was only a matter of time before international law limped into the wide-angled sights of rational choice theory. (1) Rational choice and game theory are popular tools for analyzing all manner of domestic legal problems, (2) and international lawyers are increasingly willing to draw on other disciplines to try to make some sense of theirs. (3) Rational choice theory may be considered alien to international law's norm-laden nature, (4) but perhaps the critical perspective is needed. (5) The fit is also better than one might expect. For example, because international law typically presumes unitary and sovereign states, it indulges the rational choice conceit that states single-mindedly address their important interests (6)--including, for example, in deciding whether to enter into (and abide by) international agreements. (7)

Customary international law (custom, (8) for short) is equally ripe for the picking. Custom is credited with establishing foundational rules like diplomatic immunity and state responsibility, generating new human rights principles, supplementing treaty regimes, and even creating the obligation to adhere to treaties in the first place. (9) Custom's potential significance was recently evidenced in intergovernmental claims that, whatever the status under the Geneva Conventions of al Qaeda members held prisoner by the United States, they would need to be treated "humanely and in accordance with customary international law." (10) Custom's importance to international law, in short, is beyond question. (11)

Even custom's most ardent supporters, however, have difficulty explaining how it arises, and more particularly, why customary practices should be considered binding on states. Critics increasingly cite this as a mortal failing, (12) and the very aspiration to a theory of obligation is like a mating call for rational choice theorists. Positive theory is usually measured by its capacity to predict future behavior, something that rational choice--which offers no insight into which goal(s) a rational actor will pursue--has found to be a sore point. (13) But custom's assertion that states sometimes act out of a sense of legal obligation, rather than some baser interest, surely invites alternative explanations of their behavior.

Unsurprisingly, the first generation of scholarship applying rational choice methods to customary international law--and, in particular, a series of works by Professors Jack Goldsmith and Eric Posner--has been deeply skeptical about that law's legitimacy, and proffers a devastating conclusion: what is commonly regarded as law is simply behavioral regularity dictated by state interests. States may act predictably, it is conceded, and even cooperate with one another (though such instances are less common, and less stable over time, than is typically claimed). Rational choice theory can even help explain why, and under what circumstances, such conduct may arise. But even when real cooperation emerges, it is argued, law has nothing whatsoever to do with it. At best, law-talk is just one of many fungible means by which a state can signal its intentions to keep on doing what it has done before. (14)

International lawyers have been slow to respond, (15) leaving the impression that custom may be rational, or it may be law, but not both. This impression is misplaced. This Article attempts the challenging task of resuscitating custom--restoring it, at the least, to its conventional position among the walking wounded--while at the same time respecting important insights from rational choice theory. That theory, I explain, helps clarify the circumstances under which customary international law is likely to arise, and further commends some important qualifications to prevailing legal doctrine. Rational choice theory, from this perspective, is no more hostile to customary international law than law and economics was to the common law; in both cases, the critique was as much justificatory as destructive, and in the long run may revive institutions otherwise in danger of collapse.

Part I briefly explains some of the signal failings of the traditional theory of customary international law. These have been well described elsewhere, and it is not my purpose to recite the litany of objections to existing doctrine, nor to add one more attempt at a jurisprudential fix. To the contrary, I suggest, most reform attempts share with the traditional theory a central flaw: the failure to pay sufficient heed to the interests of states, which are widely regarded as the originators and the subjects of customary international law.

This mutual failing opens the door, in effect, for rational choice. Part II summarizes the first generation critique, noting both its insights and some of its failings. This scholarship has made an enormous contribution to the literature, both in terms of theoretical rigor and by demonstrating the contestability of paradigmatic customary rules. At the same time, it illustrates the path dependence of attempts to sort complex interactions into a few tried and true games, tends to suppose an inordinately narrow set of potential interests, and evinces an unduly restrictive understanding of the standards for customary international law.

While such objections might be the basis for departing from a rational choice analysis altogether, Part III demonstrates that rational choice theory is in fact consistent with the instrumental use of custom as a legal construct. Using as examples rules involving the delimitation of territorial seas and the right to exploit the continental shelf, I suggest that considering two additional, highly conventional games--and, equally important, recalling that games may be interdependent--demonstrates the potential value of a customary regime even within the confines of game theory.

This second generation exercise is deliberately schematic, and does not consider exhaustively either the case examples or the theory behind the modeled games, in part because doing so would render both subjects needlessly insular and inaccessible. Nonetheless, it has important implications both for rational choice scholarship on international law and for more conventional doctrinal analyses. First, it posits circumstances under which customary international law might be expected to arise, as opposed to circumstances in which patterns of behavioral regularity--or misleading assertions of international obligations--merely generate false positives. This helps separate the wheat from the chaff, without paving the entire field. Second, identifying rational circumstances for custom, rather than merely asserting that states recognize and desire to obey law, suggests minor but important modifications to the traditional doctrines of customary international law. On balance, I suggest, rational choice theory is broadly consistent with the standard claims made for customary international law, and it is far more constructive in its application than may initially appear. On this view, rational choice comes to disinter, rather than to bury, custom.

Nothing in this analysis suggests that rational choice theory offers a comprehensive or exclusive picture of customary international law, and indeed part of my claim is that the theory's sights have been set too high. As critics of rational choice point out in other contexts, actors may rationally pursue ends that defy simple modeling, like a desire to behave appropriately, or otherwise act in ways that seem inconsistent with any reductionist vision of rationality. (16) And as international lawyers often argue, norms may have a substantial role to play. States may, for example, obey custom because they respect international legal processes or the particular principles it has produced, (17) and potential customary rules could be classified in terms of their superior moral content. (18) I do not mean to ignore these possibilities, or to suggest that the rational choice method ought to wholly supplant them. (19) Instead, I argue only that the gulf between rational choice and the traditional approach to custom has been exaggerated, and that by indulging in an appropriately leavened rational choice analysis, we may better understand and defend the integrity of the law we appear to have.


According to Article 38 of the Statute of the International Court of Justice (ICJ), international custom amounts to "a general practice accepted as law." (20) The surrounding language is confusing, (21) but the Article's gist is that both practice and a sense of obligation (commonly described as opinio juris) are indispensable. This view is consistent with mainstream commentary, (22) and has considerable intuitive appeal. An articulated sense of obligation, without implementing us age, is nothing more than rhetoric. (23) Conversely, state practice, without opinio juris, is just habit. (24)

Assuming that states have a coherent and discernible intent when they act is obviously problematic, as is trying to evidence that intent. (25) The problem with opinio juris that has most occupied international lawyers, however, involves articulating an internally coherent definition of the understanding that states are supposed to manifest. (26) The standard version seemingly requires that states perceive themselves to be acting in accord with a preexisting obligation. This works well enough for established customary international law: for example, confronted with claims that custom has changed, a state could examine pronouncements by other states to see if they continue to regard the rule as obligatory. But requiring a preexisting obligation makes it impossible to create new custom, unless the states pioneering the rules routinely mistake themselves to be followers--an unlikely and perverse foundation indeed. Because all rules have to begin somewhere, this chronological problem seems to afflict all of custom.

International lawyers have described this conundrum in loving detail, and have proposed various tweaks to the standard version in order to address it. (27) They have suggested, for example, that pioneering states should be treated differently, (28) that opinio juris may be inferred from persistent practices, (29) or that opinio juris may be deemphasized when the evidence of state practice is particularly compelling. (30) More radically, "instant" custom, shorn of prior practice, might be drawn from United Nations (U.N.) resolutions and the like, notwithstanding their lack of independent legal authority. (31)

Whatever their conceptual merits, these and kindred solutions share with the standard version an underappreciated flaw: namely, a disregard for state authority over the generation of customary international law. That states are in charge of determining what is or is not law remains the fundamental tenet of international law, (32) but the notion that those creating new custom have mistaken it to be obligatory--or that the accuracy of their belief is irrelevant (33)--imagines a regime in which pioneering states may be lead astray by a false premise (or, if they themselves have asserted an obligation knowing it to be false, can legitimately lead others astray). In practice, moreover, the standard version virtually guarantees heavy reliance on nonofficial sources. State claims to be obligated, as previously explained, will frequently appear premature and counterfactual. In their stead, reckonings from state practice or from nonbinding declarations are almost inexorably used to fill the gap. (34) The result erodes the privileged role that states have in determining whether new law exists.

Proposed revisions to the standard version encounter something of the same problem. If one relaxes the prerequisite that states act out of a perceived obligation (in favor, say, of an increased emphasis on the quantity or quality of state practice), it becomes even less clear how those states (let alone their more passive peers) ever become obligated--how, in other words, their habits became binding. What is more, states agreed to the approach taken in Article 38, whatever its failings, and might also be said to have acquiesced in the standard version of opinio juris that has been teased from it. (35) If so, revising this standard version to any substantial degree is not only an unauthorized attempt to bind states, but also one directly contrary to their preferences.

The dilemma this poses for reform was recently illustrated in the final report issued by the International Law Association's Committee on Formation of Customary (General) International Law. (36) The Committee expressly acknowledged the primacy of states in determining the ground rules for custom. (37) At the same time, it recognized that states rarely speak to "the principles of customary law-formation in the abstract," (38) and wound up fashioning its own eclectic approach. The Final Report rejected, for example, any general requirement of state consent, but opined that the consent of a particular state is sufficient to bind it "to a corresponding rule of customary international law" (39)--"usually" even if the rule is merely "alleged" by that state. (40) At the same time, while a general belief by states that a practice is legally obligatory or permissible "is sufficient to prove the existence of a rule of customary international law," the Committee cautioned that it was not "necessary to the formation of such a rule to demonstrate that such a belief exists, either generally or on the part of any particular State." (41) The upshot is that opinio juris may be proven by any of a variety of methods, but definitively falsified by none; (42) opinio juris, indeed, may be rendered largely irrelevant by sufficient practice (and vice versa). (43) The diversity of method almost seems demand-centered--motivated, in other words, by a desire to accommodate the customary international law already said to exist, potentially at the expense of a more intellectually rigorous approach. (44) As such, it is of a piece with a good chunk of existing scholarship. (45)

The result raises fresh concerns regarding custom's integrity. If opinio juris may be established by a variety of means, and even ignored, state participants will find it increasingly difficult to apprehend how they can avoid creating legal obligations for themselves. Generating new methods for establishing customary international law is arguably illegitimate, in any event, unless states consent. Even if scholars have been put in charge by delegation or default, their theories should be plausibly consistent with state interests--and, at the barest minimum, sufficiently clear so that states can confirm or contraindicate it. Small wonder that some urge instead that custom be radically truncated. (46)


If, as I have argued, prevailing attempts to elaborate a theory of opinio juris neglect state interests, rational choice theory seems like the perfect antidote. It assumes, in keeping with customary international law's premises, that states are unitary actors, (47) but eschews any assumption that its legal norms explain how states behave; instead, it supposes that states act to promote their interests, (48) and seeks to explain what they do in those terms. (49) The usual tool for this positive enterprise is game theory. The parties, their strategies, and their payoffs are all variables in these games, but certain other assumptions are more fundamental. The central one, unsurprisingly, is rationality. Actors are assumed to favor higher payoffs over lower ones; to choose strictly dominant strategies (that is, those that prove to be the best choice regardless of what the other party does) and, conversely, to avoid strictly dominated ones; and to have adopted the assumption that other actors behave likewise. It follows that actors are inclined toward what are called Nash equilibria, positions where neither actor can improve on its strategy given the other's strategy. (50)

A second, discretionary assumption involves the independence of the game. By a confusing convention, the literature distinguishes between "cooperative" games, in which the actors are able to make binding agreements before and during the game, and "noncooperative" games, in which the actors can enforce commitments solely within the game itself--such as through shared interests and credible threats and promises. (51) Problems of international law are better conceived as noncooperative, or self-enforcing, games, given the conspicuous lack of external enforcement mechanisms; such an assumption is even more appropriate to customary international law, where the issue of obligation is very much in play. (52)

A final set of assumptions concerns the number of actors, their options, and their state of knowledge. For simplicity's sake, classical game theory typically employs two-by-two bimatrices, assuming--often counterfactually--that only two actors, facing the same two choices, are involved. While this may seem particularly inapposite to the international setting, it is at least useful as a starting point. (Moreover, as explained below, simpler games may be defended as presenting the optimal circumstances for custom's formation: if law is difficult to establish under such conditions, it may be prohibitively difficult under more realistic assumptions. (53)) Within this simpler universe, each actor knows both its payoffs and those of the other party (or at least those payoffs' ordinal rankings). The actors are also assumed to know the strategies each has available, but not to know the strategy actually chosen by the other actor--whether because they are acting simultaneously, or because circumstances require that the actor decide before it gains credible information. (54)

A. The Rational Choice Critique

In the most extensive and compelling application of rational choice theory to customary international law, Professors Goldsmith and Posner explain core customary international law doctrines as resulting from baser calculations of self-interest. More particularly, they differentiate between four "strategic positions" in which states might exhibit behavioral regularities that traditional theory might perceive as customary international law.

1. Coincidence. First, states may have coincident interests that motivate parallel, but wholly independent, behavior. To illustrate, Professors Goldsmith and Posner provide the example of two belligerents that refrain from attacking each other's fishing boats because each concludes it would be an inefficient use of its navy. (55) The payoffs in this uncontroversial game are illustrated in Game 1.


State J knows that its best strategy if State I "attacks" will be to "ignore" (because, looking at the left-most values in the left column, the payoff of 2 in the lower row exceeds the payoff of -2 in the upper row), and it also knows that if State I "ignores," it should also play "ignore" (because for the left-most values in the right column, the payoff of 3 in the lower row exceeds the payoff of -1 in the upper row). State I, symmetrically, will prefer to play "ignore" regardless of State J's strategy.

In this game, accordingly, each party is motivated to move to the lower right cell, irrespective of the other's behavior. The result is one that could be anticipated even without a formal model: states that have coincident interests will behave similarly, and continue to do so for so long as their interests coincide.

2. Coercion. Alternatively, a stronger state might coerce a weaker state into a particular course of conduct. This is a special kind of bullying, however, in which the stronger state desires to coerce the weaker state into the same conduct that the stronger state would independently elect for itself, with the result that the two states will engage in conforming behaviors. Unlike a coincidence of interest, however, the weaker state would not independently elect that course, but instead must be forced to adopt it.

Modeling this game is more complicated, as an adaptation of Game 1 reveals. The initial payoffs, reflected in Game 2A, show that the stronger State I's dominant strategy remains to ignore. A comparison of its payoffs (right-most in each cell) in the left and right columns indicates that it prefers ignoring to attacking regardless of State J's decision. But the weaker State J, by contrast, would prefer to attack irrespective of State I's decision.


Stronger State I can only achieve its ends by altering State J's payoffs, likely by some form of reprisal. Professors Goldsmith and Posner assume, in their narrative account of this game, that the "cost of punishing the small state is trivial," and that a threat to do so will be credible. (58) But the cost of attacking State J's fishing vessels is probably not zero--otherwise, the dominant strategy for State I would more likely be to attack them--and it seems equally likely that the cost of some other punishment, one sufficient to deter State J, would also not be trivial. (59) If State J knows this, State I's threat either to attack State J's fishing vessels or to pursue another form of punishment may not be credible (absent introduction of some form of doomsday device, or some other means of surrendering control). Maintaining a credible threat may require State I to maintain a link between this norm and the resolution of some other issue involving State I's interests, perhaps within the context of a multilateral setting. (60) Even simple coercion, apparently, may be better effectuated within a more elaborate institutional context.

Ignoring these complications, and assuming that State I can impose without cost a penalty of 2 on State J for attacking, the resulting game might look like Game 2B--with the resulting norm that neither state would attack.


3. Cooperation/Prisoner's Dilemma. The third game depicted by Professors Goldsmith and Posner is the standard bilateral repeat prisoner's dilemma, undoubtedly the best known of these games. (61) Unlike the first game, in which the identical behavior of ignoring fishing vessels was encouraged by the small advantage gained and the more substantial cost imposed by the diversion of naval resources, each state in this game is encouraged to attack so long as the other refrains from doing likewise. While both would profit were they each to refrain from attacking, neither can be confident of the other's decision, and so attacking will be the dominant strategy for both. The payoffs from a one-stage game are illustrated in Game 3:


If the game is repeated, it turns out, the parties may wind up cooperating--so long as they care sufficiently about the future, expect their interaction to continue indefinitely, and have sufficiently low relative payoffs for defection. (63) In that case, apparent behavioral regularities will emerge.

4. Coordination. Fourth, and finally, states might be coordinating their conduct. In the simplest form, that emphasized by Professors Goldsmith and Posner, the states are indifferent to which of two desirable equilibria results, so long as they happen upon the same solution. Once their actions are coordinated, there is no incentive to deviate. The difficulty lies in the fact that neither option dominates for either party, so arriving at coordination is far from automatic. Again, behavior extrinsic to the game is required.


5. The Games' Implications. It seems plausible that some international problems resemble, at a high level of abstraction, the assumptions and payoffs presupposed by these games. Where that is the case, analyzing those problems in terms of the games is surely instructive. Among other things, the games help explain how behavioral regularities may emerge, as well as the pressures that may make them difficult to sustain.

Professors Goldsmith and Posner, however, make two more ambitious claims. The first is essentially empirical: behavioral regularities conventionally regarded as establishing customary international law may be described by one of the specified games. They substantiate this claim through trenchant descriptions of four reputedly robust examples of custom:

(1) the rule that property on a neutral party's ship (other than contraband) is immune from seizure, which they describe as reflecting either coercion or coincidences of interest; (65)

(2) diplomatic immunity from criminal prosecution, which they describe as "an amalgam of independent, bilateral repeat prisoner's dilemmas"; (66)

(3) the three-mile rule establishing national jurisdiction over territorial seas, which they consider to have depended on each of the profiled games under divergent sets of circumstances; (67)

(4) the Paquete Habana rule exempting enemy coastal fishing vessels from prize rules permitting belligerents to capture enemy ships and goods at sea, which they consider to be explained largely by coincidences of interest, but also to some degree as coercion, and possibly as cooperation arising from prisoner's dilemmas. (68)

Drawing considerable succor from these examples, they appear to suggest that all of the regularities presently regarded as customary international law may be described in similar terms. (69) The argument clearly depends on whether their examples are convincing, and whether they are truly representative. Both aspects are potentially problematic, (70) and Part III briefly discusses one of their examples below, and an additional rule, in order to suggest some limits to the analysis.

The "payoff" for these narrative accounts and their generalization, in any case, lies in a second, categorical claim: to the extent behavioral regularities may be described in these game-theoretic terms, they cannot be considered customary international law. For some of the specified games, this claim should be uncontroversial. Coincidences of interest describe essentially independent behaviors that happen to be similar, but are undertaken without respect to any party other than the acting state itself. Such acts have nothing to do with any sense of legal obligation. Similarly, coercion makes smaller states act out of fear, not legal obligation. (71)

The other games are closer to the line. Cooperation arising out of bilateral prisoner's dilemmas may resemble acts taken out of a sense of legal obligation, but even there, Professors Goldsmith and Posner argue, "[n]ations do not act in accordance with a norm that they feel obliged to follow; they act because it is in their interest to do so." (72) The point is not merely semantic. State behaviors will ebb and flow with changes in the underlying payoffs; such changes, it is argued, "are the sole determinants of whether states engage in the behavioral regularities that are labeled norms of [customary international law]," suggesting that legal norms have no influence whatsoever on state behavior. (73) Bilateral prisoner's dilemmas are further incompatible with the traditional account because they cannot easily be extrapolated to multilateral cooperation. (74) Many of the same points are made regarding coordination games: they are highly unlikely to evolve multilaterally, and even if they did, would reflect state interests rather than perceived legal obligations. (75)

B. Conventional Objections

As made clear in Parts I and III, the rational choice approach to custom has pronounced virtues relative to its alternatives, and many are apparent in the insightful and creative critique provided by Professors Goldsmith and Posner. The limitations to their approach are also evident. Before undertaking a second-generation attempt to reconcile rational choice with custom, it is important to identify potential limits to the existing approach.

1. The Empirical Claim. The assertion that many, or all, of custom's rules may be described in terms of a few specific games raises several obvious concerns. Some have to do with basic game-theoretic assumptions. More than two states are ordinarily involved in the formation of customary rules, and their options usually involve more than two choices. It may seem unrealistic, moreover, to suppose that states know one another's payoffs, but remain wholly ignorant of the strategies their peers are pursuing. One might expect, for example, that states will sometimes lack information regarding the depth or intensity of one another's interests, especially on matters of substantial international concern. Absent full information, a state may mistakenly assume that its own interests and payoffs are universal--and perhaps even conclude, in observing discrepant behavior by other states, that they behave in an irrational, even strictly dominated, fashion. (76)

While important, I think, such objections may easily be overstated, and their implications are far from obvious. Relaxing these assumptions may, as Professors Goldsmith and Posner suggest, simply worsen the prospects for achieving custom. Introducing additional parties, for example, makes it still more difficult to explain the emergence of stable behavioral regularities. (77) Similarly, if states sometimes act contrary to their own interests, or lack sufficient information concerning their counterparts' payoffs and strategies, they may engage in unproductive dissembling and bargaining that make collective undertakings like custom still more difficult. (78)

A second set of concerns emerges from within the rational choice model. Game theory has generated a broad range of paradigms with meaningfully different implications for cooperative behavior, and identifying which best fits the particular context may be quite difficult. (79) Describing a situation in terms of a particular game, moreover, tends to shape the behavioral account. (80) It is easy, in consequence, to overapply certain iconic games and neglect other explanations both from within and without game theory. Focusing exclusively on prisoner's dilemmas, for example, supposes stringent conditions under which custom would have an excessively difficult time arising. Debatable diagnoses of simple coordination games, on the other hand, would too readily exclude explanations of how regularities may require a helping hand, and so inadvertently exclude any role for legal rules. (81)

Professors Goldsmith and Posner are not preoccupied with any single game, of course, and to their considerable credit, attempt to test their approach against actual examples of customary norms. (82) Nonetheless, their accounts are substantially, and perhaps inevitably, arbitrary in particularizing state interests. Assuming, that is, that they have identified the most important interests, the weighting of those interests and the resulting payoffs seem inescapably artificial.

Take, for example, their discussion of diplomatic immunity from criminal prosecution. (83) As they explain, according immunity to a foreign state's diplomats, while potentially risking national security and public discord, also benefits the host state in several regards--principally, by facilitating communication with the foreign state, and by reducing the risk that the foreign state will retaliate against diplomats the host state has itself stationed abroad. (84) One could cite other advantages to immunity (such as avoiding the need for costly, potentially compromising prosecutions) and disadvantages (such as enduring smuggling), (85) or challenge their weighting of legal integrity, (86) national security, (87) or even communication. (88) In the end, however, it is unclear how it would affect their analysis. They essentially assume the conditions of a prisoner's dilemma, and further assume that two-state cooperation will nonetheless be promoted because the long-term benefits of facilitating communication outweigh any countervailing interests . (89)

It is natural to assume procooperative payoffs, of course, because cooperation does predominate in this area. The particular structure of their payoffs, however, largely dictates their conclusion that the cooperative strategy of bilateral, reciprocal immunity is both the best solution and, ultimately, an unstable one. (90) Professors Goldsmith and Posner acknowledge in passing that a different game, the Battle of the Sexes, might more accurately portray the relevant payoffs, (91) but they do not adequately address the degree to which that changes their analysis. As I show below, modeling diplomatic immunity in terms of that game might lead to a very different conviction regarding custom's instrumental value. (92)

The tendency of rational choice analysis to indulge in Just So Stories about real-world phenomena is well understood. (93) Any attempt to shoehorn complicated political and legal problems into a matrix encounters similar problems, which might be regarded as the inevitable price of attempting analytical clarity. But because the work of Professors Goldsmith and Posner is uncommonly ambitious, such problems have a particular gravity. First, and most obviously, if different games may be appropriate even for the examples they discuss, surely we should be cautious in supposing that other customary rules necessarily fall into one or the other of these games.

Second, any arbitrariness in analyzing these illustrative rules makes it more difficult for rational choice to claim a decisive edge over traditional modes of analyzing custom. Those applying traditional versions of custom, for example, have assumed that functional reciprocity is the foundation for according diplomatic immunity, just as do Professors Goldsmith and Posner. (94) The latter's claim to better explain violations of custom, too, seems exaggerated. Such violations may be "inexplicable" within the terms of the immunity rule itself (though in other cases, elaborate explications are found equally unsatisfactory), (95) but it would surprise no one to learn that rogue states (being, well, roguish) are more likely to breach immunity norms, or that customary rules of this kind may collapse when stakes are high. (96)

In point of fact, many of Professors Goldsmith and Posner's incisive criticisms of purported customs--such as the lack of clarity in a custom's terms, widespread deviation, or the absence of any nontransient commitment by states (97)--are equally cognizable within the standard version as deficiencies either in state practice or in opinio juris. If these examples are really the best customary international law has to offer, and Professor Goldsmith and Posner's criticisms are accurate, then there is very little custom by custom's own lights. This is an important conclusion. It is less clear, however, whether the rational choice perspective offers any distinctive advantages in assessing what is or is not custom, or whether it has genuinely exhausted the latent potential of the standard version.

2. The Categorical Claim. The argument that behavioral regularities of the kind potentially emerging from coincidence, coercion, coordination, or cooperation games cannot be described as custom depends heavily on the peculiar standards of customary international law. Haphazardly coincidental behavior, everyone would agree, does not establish a legal obligation. On the other hand, routinely coincidental behavior is not ordinarily considered inconsistent with legal obligation simply because it is motivated by self-interest. Professors Goldsmith and Posner demean the legal potential of such behaviors, for example, by noting that regularly refraining, by coincidence, from attacking the shipping vessels of other states is no different in character than "the behavioral regularity of nations not sinking their own ships." (98) Assuming the analogy holds, the fact remains that the self-interest in avoiding injury to one's own self is not usually seen as contraindicating the existence of a binding legal norm against such acts. The common law criminalization of suicide, (99) for example, did not lose its legal character simply because most compliance is consistent with rational self-interest.

But customary international law is of course special. Opinio juris arguably requires that a state act with reference to legal obligation, not out of "mere" self-interest. Professor Goldsmith and Professor Posner's game-theoretic analysis may be defended, accordingly, as simply taking custom seriously. (100) Their blunt conclusion--that customary international law is not, by its own terms, law--also happily avoids the tiresome task of tinkering with the standard version. By their account, instances of cooperation may resemble previous descriptions of customary practices, but calling them law adds little. Indeed, doing so sows confusion by suggesting that the norms will survive changing state interests, when in fact state practices are considerably more elastic.

At the same time, the fact that this account so comprehensively denies custom of its legal character gives pause. Although cooperation arising from a prisoner's dilemma may in theory be stable, (101) it would appear never to be sufficiently stable to be regarded as the basis for a legal norm. The implication, it would seem, is that once one stipulates that international relations amount to a noncooperative game--leaving states to enforce commitments against one another without benefit of external compulsion--one has essentially assumed away the prospect for international law.

Customary international law, however, is not a run-of-the-mill null hypothesis, but instead comprises an identifiable--if problem-plagued--institution. One might wonder, for example, why states would continue to speak of international legal obligations of this sort. Anticipating this objection, Professors Goldsmith and Posner suggest that states may be papering over real conflicts of interest in order to avoid jeopardizing other cooperative relations. Alternatively, states may be seeking to explain how their behaviors are in keeping with a previously established cooperative understanding. As they note, even the latter is consistent with the hypothesis that states are engaged in the rational pursuit of self-interest, since they have an interest in explaining past acts insofar as they may provide clues about future performance. (102)

Much of this analysis rings true, for reasons explored previously. Litigants, commentators, and even the ICJ often claim that custom has been established when that appears unlikely. (103) Part of this, certainly, is "cheap talk" of the kind described by the rational choice critique--in particular, claims by pioneering states that they are acting in accord with prior obligations, when in fact their statements and acts are better regarded as innovative.

At the same time, these explanations understate the degree to which law-talk is genuinely constitutive. Professors Goldsmith and Posner recognize that legal rationalization of behavior may be important in shoring up a state's reputation as a party that is generally trustworthy, though they are highly skeptical of state representations. (104) The more important point, however, is that states are signaling how far they may be trusted with respect to legal obligations--employing a particular language for denominating cooperative and coordinative understandings, and in doing so alluding to a tangible regime. States consented, for example, to the statutes of both the Permanent Court of International Justice (PCIJ) and the ICJ, each with concrete (if somewhat opaque) language enabling those courts to apply customary international law. (105) States also participate, less directly, in reform efforts like those conducted by the International Law Commission. (106) To be sure, states have in a number of respects resisted the decisive articulation and application of custom, such as by placing few disputes before the ICJ, and to that extent have undermined the regime's potential. But while that resistance is partly justified on the ground that such adjudication is a costly irrelevancy, the dominant consideration has been the risk of losing--proving, indirectly, the significance of the institution. (107)

Professors Goldsmith and Posner downplay the persistence of a customary regime, regarding legal commitments as largely fungible with moral and other normative commitments. (108) But their inattention to the distinction between legal and nonlegal commitments retards the potential of their approach. It is nonlegal norms, for one, that are characteristically disassociated from self-interest; (109) as I demonstrate in the next Part, state self-interest is not strictly inconsistent even with the standard version of opinio juris. More important, states may act out of self-interest while still making significant instrumental use of the customary international law regime, and that behavior may in fact be modeled in rational choice terms.


Neither the standard version of opinio juris, nor its rational choice critique, it would appear, successfully account for the role of states. The standard version does a poor job of establishing coherent ground rules to which states might be said to have consented--or, for that matter, that might plausibly be consistent with state interests. The rational choice critique, on the other hand, places the interests of states front and center, but makes little headway in explaining state adherence to existing legal institutions, other than to suggest that they comprise some form of "organized hypocrisy." (110) Each approach, in short, finds itself limited by its failure to accommodate the other.

Nothing about this conflict is inevitable. Though rational choice theorists are often associated with "thick" interest analysis, in which actors are presupposed to pursue certain elementary and autonomously maintained preferences, the theory is equally open to the supposition that states instrumentally pursue other objectives, such as legal compliance. (111) The theory is also open to incorporating consideration of regimes like customary international law, at least where the account is sensitive to the possibility that states may choose such regimes, rather than simply being subject to them. (112)

What, then, stops rational choice theory from taking such factors into account? For one, taking normative or relational interests into account arguably disserves the theory's parsimony and leaves it underspecified. (113) But in scrutinizing the pretensions of customary international law, rational choice theory has far less of a burden than it otherwise might: its value in this endeavor lies less in predicting the outcome of state interactions than in analyzing whether baser explanations suffice. Put simply, its application here is more destructive than constructive, which affords the luxury of being less parsimonious.

Second, even if rational choice does not require ignoring customary international law, custom's prerequisites may require looking past state interests. Here, too, the divide is less pronounced than it may initially appear. International lawyers hesitate to say that states disregard legal obligations whenever their other interests dictate, but they often concede that such obligations are largely created and followed because it is in the states' interests to do so. (114) Nor is the opinio juris requirement inherently hostile to state interests. As explained earlier, many international lawyers acknowledge the impossibility of requiring that pioneering states develop norms only out of some sense of prior obligation, and suggest other means by which the subjective element may be evidenced. (115) In the particular case of permissive customary rules--those allowing states to engage in particular conduct, as opposed to placing duties upon them--it is especially clear that no prior obligation need be asserted. (116)

This same indulgent approach has been reflected in judgments rendered by the ICJ. The ICJ has sometimes stated the rule of opinio juris in the textbook fashion, as in the North Sea Continental Shelf cases. (117) In that same judgment, however, the ICJ also indicated that the practices "should ... have occurred in such a way as to show a general recognition that a rule of law or legal obligation is involved," language that is far less demanding. (118) Five years later, in the Fisheries Jurisdiction case, the ICJ evidenced considerable uncertainty as to the impact of aspirational statements leading up to the Third Law of the Sea Conference, but its divided opinions made clear that prior commitment to a principle was not a threshold requirement. (119) And in the Nicaragua judgment, the ICJ allowed that "[r]eliance by a State on a novel right or an unprecedented exception to [al principle might, if shared in principle by other States, tend towards a modification of customary international law." (120)

The ICJ's position, if it can be generalized, seems to be predominately instrumentalist: rather than dwelling on the legal characterization of a particular act, the opinio juris requirement is intended to exclude evidence of practices engaged in without regard to legal obligation, and to identify rules that in the end have "become binding on the other Party." (121) Much the same approach is evinced in parts of the International Law Association's Final Report. Opinio juris, at a high level of generality, requires practice "in circumstances which give rise to a legitimate expectation of similar conduct in the future." (122) Conversely, "a practice that is generally followed but which states feel legally free to disregard does not contribute to customary law." (123)

Especially if the standard version is to be so relaxed, however, it is necessary to stay connected to the foundational interests of states. (124) The task, more specifically, is to identify circumstances in which behavioral regularities are (1) plausible in rational choice terms, while maintaining meaningful instrumental use of customary international law, and (2) compatible with the legal ground rules of custom, in particular its subjective element. To be clear, I do not propose to specify all such circumstances, or to classify even a fraction of custom along these lines. Nor would I claim that all conventionally recognized custom--to the extent any such consensus could be reached--might be sorted into games of this or kindred natures. But it is useful to explore, however tentatively, whether the limits of rational choice theory have been exhausted, and whether that theory and the principles of customary international law are inextricably at odds.

A. Toward Rational Custom

As explained in Part I, traditional theory has had to confront the embarrassing absence of an explanation as to how custom might arise--arise, that is, in a fashion consistent with a sense of legal obligation, so that rules of law could be distinguished from mere habits. If it is to appeal to rational choice theorists, moreover, such a theory must also render custom a comprehensible vehicle for promoting state interests, not merely one for generating law in keeping with some notion of the public good. These are demanding conditions, but not nearly so out of reach as may be supposed.

1. Coordination and the Case of the Territorial Sea. As previously explained, legal regimes are unlikely to be relevant in a coordination situation of the kind modeled in Game 4. Signaling can be helpful in initiating play, but once a state has moved it is relatively unlikely to change its position, as it has no incentive to defect. Absent a collective mistake, moreover, other states will follow the first move, as they are, by hypothesis, indifferent among any agreed-upon alternatives. In this scenario, law does very little work.

A different situation is suggested by the well-known Battle of the Sexes game, an important variant of the coordination game. The politically incorrect namesake story assumes, for better or for worse, that a man and a woman each favor a different form of entertainment (often boxing and ballet, respectively, though the examples used depend on the writer's nationality). The highest payoff for each would be to persuade the other to acquiesce in the first's preferred entertainment, but the second-best would be to accompany the other to the entertainment he or she prefers, with solo expeditions falling in a predictable order. Companionship, in other words, trumps personal preference.

The same sorts of payoffs can be produced in a game of international relations. One might reimagine the fishing hypothetical originally constructed by Professors Goldsmith and Posner, for example, to suppose that the issue of how best to preserve fishing fleets during time of war (and prevent the wasteful expenditure of military resources) admits of more than one solution: for example, refraining from attacking fishing vessels, or removing any potentially offensive weaponry (e.g., explosive harpoons) from such vessels. (125) Here, too, there are two Nash equilibria, each favored by a different state; though not indifferent between the two results, each state would still prefer either equilibrium to discrepant actions. Once ensconced in one of those positions, neither will wish to deviate unless assured that the other state will follow.


Coordination games are generally regarded as presenting little need for conventional legal constraints: because there is no incentive to deviate, there is no need for an enforcement mechanism. Much the same may be said for the Battle of the Sexes variant. But the trick in a coordination game is getting to that "sticky" outcome, and that is especially difficult in the Battle of the Sexes, because states are not indifferent among the collegial equilibria. The consequences become apparent as soon as the number of players is expanded and the game iterated. In the early stages of an international practice, states might conceivably migrate from one solution to the other in search of a critical mass, (126) particularly if the distributional gains to states benefiting may be shared with defectors. (127) It is not hard to imagine the transition from a two-state game to an n-state game resulting in states clustering around more than one solution, with diminished payoffs for each. (128) The attendant uncertainties, and the prospect of prestabilization "defection," risk the ability to achieve one or the other equilibrium. Attaining an equilibrium, too, is likely to be retarded by transitional undercommitment as some states wait to see what happens. (129)

Even without formal institutions capable of enforcing a particular solution, (130) it may be possible to fashion a multilateral solution by focusing instead on the development of the standard. It is at these early stages where credible commitments, backed by reputational investment in the customary international law regime, may usefully diminish uncertainty and allow coordination to be attained more rapidly and with less friction. The existence of customary international law, in this variant, permits a state to commit to one of the equilibria and to have its representation regarded as binding. Customary international law, then, facilitates a choice not just between norms and anarchy, but also between norms in circumstances where states may be well disposed toward binding obligations.

One potential example of such a rule is the same three-mile limit that Professors Goldsmith and Posner critically examine. That rule is not an uncontroversial example of customary international law, which arguably impairs its value as an illustration for what is wrong with the traditional approach. Perhaps "conventional wisdom" regarded the three-mile limit as a customary rule "during most of the nineteenth and the first half of the twentieth century," (131) but many contemporary commentators disputed that position, and even favorable commentary has questioned whether the rule lasted so long. (132)

Still, few customary rules are uncontested, and a case can be made that the three-mile limit was a rule during the period 1875-1930, (133) after which point it began to unravel. (134) While the question of when it ceased having any pretense to being a rule (and why) is of some interest, the more relevant question for instant purposes is how it came to be considered a rule in the first place. The jurisprudential origins--in particular, the relationship between the three-mile rule and the position that the territorial sea extended as far as a nation's ability to defend it, later expressed as the cannon-shot rule--have been debated at length, but two points deserve emphasis. First, many standards, not limited to the cannon-shot measure and the three-mile limit, were available. (135) Consistent with that understanding, the early champions of the three-mile limit, principally the United States and England, did not genuinely contend that their choice was fore-ordained, and their earliest discussions were quite tentative. (136)

Second, although the initial exposition of the three-mile limit was provisional, its adherents showed considerable resilience during the relevant period. (137) Numerous inconsistencies (or, perhaps, subtleties) may be cited, but their significance is debatable, and it is difficult to say conclusively that they abnegated an otherwise viable rule. (138) For example, although different limits were generated for different purposes, some of the most important--such as the additional range for enforcing customs violations--were evident in the early going. (139) The fact that additional distinctions were introduced later on is common to many kinds of law, not a distinctive flaw of customary international law. And while Professors Goldsmith and Posner appropriately note the discretionary enforcement of the rule by the United States and England, (140) they do not lavish attention on the attempted rationalizations of those occasions, nor provide any means of assessing their significance as against evidence that the rules were observed. (141)

Instead, these categorical and individual exceptions are highlighted in order to suggest that state interests--which may be hypothesized for each of the exceptions--better explain the sum of state behavior than would any rule of law. This is a powerful argument but ultimately, I believe, overstated. First, as suggested above, it relies on a false opposition between customary international law and state interests: the opinio juris requirement does not ineluctably require that custom's content be foreordained and imposed upon pioneering states, nor does it insist that rules require behavior inconsistent with state interests. (142)

Second, while exogenous factors like maritime power and fishing patterns surely influenced state interests and behavior, state conduct seems in fact to have been relatively stable and predictable, (143) in keeping with the pattern indicated by a Battle of the Sexes. States clearly had the latitude to choose different means of designating their territorial seas, and they were not indifferent among the possibilities. But they seem also to have appreciated that some coordination was preferable to none--in the immediate case, so as to create rules of the road for coastal states and the maritime powers alike.

The idea that a behavioral regularity might emerge in keeping with a Battle of the Sexes game is not alien to Professors Goldsmith and Posner, who expressly entertain the possibility with respect to the particular problem of fishing rights. (144) But they fail to draw the connection between such circumstances and an acknowledged embarrassment for their approach: why would powerful states like the United States and England, mindful of their varied interests, agree to (and maintain) a three-mile territorial sea in the first place? Their answer is that neither state was sufficiently powerful simultaneously to protect fishing vessels at home and abroad, and so was inclined to coordinate--with a mile-measured band being a reasonable focal point, and the cannon-shot standard, being roughly three miles at the time, close at hand. (145)

This is not implausible, though one might find fault with lumping the United States and England together, across time and across issue areas (146)--a criticism very much in the spirit of Professor Goldsmith and Professor Posner's own skepticism about custom's generalizing tendencies. But it is also difficult to confirm this particular calculus of national interests in the historical record. At bottom, it simply assumes, in the fashion of nonempirical rational choice discussions, that one of many potential foci will be chosen--and, for that matter, that any focus will emerge (147)--and ignores the instrumental role of custom in doing so. As previously explained, settling on the three-mile limit would have been aided by the ability of states to commit to that standard, and the institution of customary international law enabled them to do so. An indispensable part of that function involved law's obligation, and the sense that deviating from a legally framed commitment would impair the ability to make similar commitments credibly in the future. Custom worked for a period, in short, because it provided nations interested in establishing a stable standard with a mechanism for coordination--a distinctive kind of oath, as it were, that permitted states to focus on a rule that happened to serve their mutual interests. (148)

2. Assurance and the Case of the Continental Shelf. A second game meriting further consideration is popularly known as the Stag Hunt. In the basic story, two hunters must each decide whether to hunt stag or hunt hare; they must work together to bag a stag, but either one could snare a hare; and the half share of the stag to which each might conceivably be entitled has a greater payoff than a single hare. Like the Battle of the Sexes, the Stag Hunt has two pure strategy equilibria: if one hunter is hunting hare, the other one should as well, for the time spent pursuing a stag will be for naught; but if one is hunting a stag, the other should as well. Acting strategically requires anticipating the other's move. (149)

In its international form, the game suggests a common problem confronting those desiring to cooperate. Sticking with the fishing hypothetical, one might suppose that each of two potential belligerents might weigh whether to place military observers adjacent to the other party's fishing vessels--the plan being to deter the occasional attack by creating the potential for one of the attacker's own to be at risk. For the scheme to succeed, clearly, both states must be involved. The alternative is to use the military personnel in some unrelated capacity, with a lesser payoff (say, by marching them in parade formation).


Like the Battle of the Sexes, participants in a Stag Hunt have the incentive to continue in any mutual cooperation they achieve, and would be inclined to defect only in the event that others cease cooperating. Unlike the Battle of the Sexes, states would not be distracted by another cooperation point. Cast in this light, legal institutions should have little role to play. (151)

The problem arises, though, when one relaxes any information assumptions. If State I is unsure as to whether State J will place observers, it will also be unsure as to which strategy it should pursue. It then becomes appropriate to consider the mixed strategy equilibrium--basically, the point at which State I would be indifferent to State J's particular choice, which depends on State J's probability of placing observers and the relative payoffs. (152) Technicalities aside, it becomes valuable for the parties to be able to communicate their likely moves. In the pure case, such communications have credibility without stringent enforcement mechanisms, because the parties have no incentive to mislead. (153) But one state does not, unfortunately, have perfect information about another's payoffs, and so may pursue sub-optimal strategies from the beginning--or, for that matter, constantly fear that the other state will defect, thus making an apparently stable equilibrium highly unstable in practice. (154)

Customary law, under these circumstances, may provide an effective and low-cost means of communicating. A nation taking a position concerning what international law requires, or should require, is surely indicating where its own interests lie. Its communication also signals that it is less likely to defect than would otherwise be the case, lest its reputation as a law-abiding nation be squandered. Finally, but not insignificantly, it communicates how it perceives the situation in question: because collaboration problems are generally difficult to resolve and hard to enforce under the best of conditions, states may use custom to signal their understanding that the instant "game" is more tractable.

It is tempting to describe all customary international laws as emerging in this kind of game, because each potential rule requires just such a choice between collaborative gain and independent pursuits. A more specific example may lie in the initial formation of the continental shelf regime. The origins of that doctrine are almost universally traced to a single source, the Truman Proclamation of 1945, which announced a succinct but reasonably comprehensive position: (1) coastal states have exclusive rights to their contiguous continental shelves; (2) where the shelf extends to the shore of another state, or is shared with an adjacent state, the boundary is to be determined by those states according to equitable principles; and (3) rights to the shelf do not impinge in any way on the character of the high seas above or associated rights to navigation. (155)

The Truman Proclamation justified this approach in terms of interests, adverting to the need to find new sources of mineral sources (and the emerging technological ability to exploit those residing in the shelf), the need for orderly exploitation (and the superiority of contiguous states in ensuring that), and the security interests that contiguous states have in regulating offshore activities. (156) At the same time, the Proclamation clearly and self-consciously sounded in international law. The president carefully justified the position in terms of common interests, not solely those relevant to the United States, and sought to establish a universal policy that would be "reasonable and just" for all interested states. (157) The Proclamation also stated the objective of obtaining "recognized jurisdiction" over the shelf--recognized, presumably, by other states--and appeared to acknowledge the rights of other states to jurisdiction of identical scope. (158) As an accompanying press release explained, the proclamation was "concerned solely with establishing the jurisdiction of the United States from an international standpoint." (159)

The legal adviser to the State Department later endorsed the objective of trying to "fill[]the gap in international law on this subject" with a principle more faithful to the freedom of the seas than its alternatives. (160) But there was no attempt in the Proclamation or afterwards to argue that the U.S. position was dictated by international law, and virtually no one thought that it was. (161) There was likewise little support for the proposition that the Proclamation had by itself established customary international law; the overwhelming consensus, however, was that it contributed to, and even enabled, its creation. (162) The position rapidly attracted the agreement of other states, permitting the ICJ to describe the proclamation as the "starting point of the positive law on the subject," (163) and to cite the ensuing 1958 Continental Shelf Convention as "an example of a legal theory derived from a particular source that has secured a general following." (164)

In this broad outline, at least, the emergence of a pre-Convention customary international law of the continental shelf reflects some characteristics typical of a Stag Hunt. At the time of the Truman Proclamation, states desiring eventually to exploit the seabed, but uncertain about the eventual legal regime (if any), faced difficult questions concerning their appropriate strategy. If the continental shelf would come to be viewed as res nullius--owned by no one, but available for occupation--self-help would clearly be the order of the day, and states might engage in a literal race to the bottom. If the shelf were instead regarded as res communis, to which all had a right regardless of the ability to exercise it, exclusive national occupation would at least in theory be impermissible in the absence of international license (though in practice states would likely be able to obtain certain property interests, subject to dispossession, and to protect the exploitative activities of their nationals). (165) In either case, those in a position to exploit the shelf could do so without regard to the shelf's location; much the same result would transpire if states were to propose inconsistent rules of customary international law, with none prevailing. (166)

Under such circumstances, coastal states might be forced to choose between premature and costly exploitation of their shelves, on the one hand, and waiting for the emergence of a cooperative norm, on the other--and in the latter case, risking not only that such a norm would not emerge, but also that their resources would be exploited by another state, and their security threatened. (167) Like a hunter uncertain whether to pursue a hare, or pursue instead a stag in the hope of cooperation, states had to elect between a cooperative scheme with potentially greater return, independent action, or a strategy based on their imperfect understanding of the strategies of other states. By dint of the Truman Proclamation, the United States successfully sought to create a more orderly, cooperative regime, initiating that process by staking out a legal position and articulating its view of the regime best reflecting equilibrium interests. (168) Not insignificantly, the proposed principle was generally recognized as the best cooperative outcome. (169) International law permitted, under those circumstances, the opportunity to commit credibly to a relatively uncontroversial outcome. (170)

3. Specifying Opinio Juris. If successful, the above descriptions delimit circumstances in which custom may be in accord with the state interests privileged by rational choice theory. But it remains to detail an approach to opinio juris that more coherently distinguishes between practices establishing and failing to establish law. I attempt to do so in this Section, in part to lay bare some substantial objections to my account.

The conditions for generating customary international law in a Battle of the Sexes game might be generalized in the following way:
   Principle 1: Customary international law can be created when a
   sufficient number of states credibly commit to a norm of conduct
   that offers the opportunity for mutual gain through coordination.

Several important objections pertain equally to the second principle elaborated below, and so I briefly postpone addressing them. But this first principle may be peculiarly incompatible with the standard description of the subjective element. At custom's formative stage, at least, states belonging to each "sex" within the international community might simultaneously claim to be engaged in propagating customary international law. The resulting conflict between incompatible legal assertions is all too reminiscent of the embarrassing position pioneering states encountered in the standard account, and inconsistent to boot with even the most primitive commitment to right answers in legal inquiries.

Three answers seem at least partly satisfactory. First, precisely because customary international law permits, by hypothesis, the easier coordination of state interests, it diminishes the likelihood that competing factions of states will make equally legitimate claims regarding their legal obligations. Second, nothing in the theory of customary international law suggests that claims about it are universal or immutable. It is accepted that two or more groups of states may be governed by "special" custom (also known as a local, or particular, custom) governing their particular relations, (171) and it is fundamental that custom may be eliminated by a sea change in state practice. (172) The fact that one of two competing rules may be left standing does not mean that neither one was a legal norm until that occurs. Third, and finally, it is not unprecedented for international lawyers to distinguish between established customary international law--which later-subscribing states, at least, should describe as already binding them--and budding rules, for which more basic means of distinguishing practices from happenstance may suffice. (173) While this appears to be an entirely ad hoc deviation within the standard version, the Battle of the Sexes scenario provides a clear explanation as to why these positions need to be distinguished.

A second condition under which custom may be created, based on the Stag Hunt, might be crystallized as the following:
   Principle 2: Customary international law can be created when
   one or more states propose a norm of conduct that they credibly
   indicate will bind their conduct if and only if a sufficient
   number of other states agree likewise, and a sufficient number
   of states in fact do so.

States attempting to create international law under these circumstances act less out of the perception that their behavior is obligated, even on a transitional basis, than out of a willingness to be obligated. In some regards, this may appear more akin to a legal obligation than does the first principle. The subjective character of the enterprise is much like the opening of a multilateral treaty for signature--with the difference being, of course, that the terms, the number of parties participating, and the means of enforcement are unspecified. (174)

Conditionality alone is not inconsistent with the traditional understanding of the subjective element. States acting out of a perceived legal obligation are not supposed to have perceived that the law is immutable. Doing so would tend to conflate customary international law with peremptory norms, such as the prohibition on genocide, from which derogation by ordinary means is impossible; although it is unclear whether such norms are a species of customary international law, strictly speaking, (175) no one contends that they are the only kind of custom.

The problem, again, lies with conditional assertions by pioneering states. Such states could not be deemed to be acting out of a sense of legal obligation identical with the norm they seek, any more than a party extending an offer can be deemed to be acting pursuant to contract. That said, offerors are by that very act assuming legal obligations, (176) and pioneering states could be understood as undertaking a similarly conditional obligation--consistent, in my view, with opinio juris as it is conventionally applied to pioneers, (177) and surely consistent with more revisionist views. (178) Equally important, to the extent that it presses the limits of the standard version, it is not susceptible to criticisms of prior reform proposals. (179) Rational choice theory allows us to specify circumstances under which custom may arise that are plausibly consistent with state interests, including the interests reflected in the rules of recognition developed for custom. More importantly, though, such an exception would obtain only under conditions that presumptively reflect state interests.

B. Gaming Reputation

Assuming that these additional games and derivative principles may be reconciled with the standard version of custom's requirements, the more serious tension may be with the sound application of rational choice theory. Customary international law is largely predicated on the notion that states care about establishing and upholding their reputations for acting consistently with their legal obligations. But as rational choice theorists have observed, states may have more pressing concerns, and can promote their interests in ways not contingent on their credibility for keeping international commitments. Some of those means may actually be in tension with having a reputation for being law-abiding, such as where a nation can better achieve its ends by appearing to be tough, unpredictable, or heedless of public opinion. (180)

It would be a mistake, accordingly, to assume blithely that all states want to be regarded as law-abiding, or that that those states generally considered law-abiding wish to maximize that reputation. But it is equally mistaken to assume that the existence of these caveats is fundamentally inconsistent with customary international law. To maintain custom, it is only necessary that a group of nations invest their reputations in a nontrivial number of rules; (181) those nations are also free to withhold their legal reputations from any number of other contexts in which it would disserve them, such as where they want to maintain credibility as the zealous defenders of paramount national interests. States peculiarly concerned with developing reputations antithetical to one for compliance might also absent themselves altogether from customary international law (or, better still, subscribe briefly to a norm but then breach it flagrantly). This kind of considered disobedience is actually quite rare. (182) In any event, its potential appeal scarcely means that states privileging a more conventional "good" reputation lack that opportunity.

A more complex model of reputation, indeed, may make custom's assumptions more, rather than less, plausible. For those states valuing it, reputation becomes costly to jeopardize on any particular occasion; putting the reputation in jeopardy might, in fact, alter the relevant payoffs. (183) But this risk has substantial virtues for cooperation. Once states do place their reputations on the line, it can improve their strategic situation, which in turn may embolden them to undertake the initial risk. (184)

Such use of reputation, indeed, can easily be modeled as part of an embedded game. The discussion so far has centered on simple, normal-form games in isolation from one another--an entirely defensible heuristic, given the difficulty of anticipating and evaluating the potential interactions between problem areas. But states do not, in fact, interact solely with respect to one rule or the other, and it is also possible to understand their interaction with respect both to an individual rule and to the system of customary international law.

Take, for example, the game most recently modeled, the Stag Hunt, and for simplicity's sake employ the same payoff scheme as used previously: in this case, with the greatest payoffs derived from establishing the continental shelf regime described by the Truman Proclamation, and the less productive path involving the (submarine) land-grab encouraged by a res nullius principle (which might for simplicity's sake be equated to the absence of a customary rule). In Game 6, recall, there are three equilibria: the top left cell, the bottom right cell, and a mixed strategy. The difficulty is to ensure that the states involved settle on the equilibrium with the greater payoff, but they cannot know for certain what strategy the other party will in fact adopt; they also lack any external agency for ensuring that promises are kept.

Superimposing another game, involving reputation, can resolve the situation, as illustrated in Game 7. (185) Assume now that State J can obtain a payoff of 4 by abstaining from the Stag Hunt (equivalent, for these purposes, of the value to it of refraining from placing its reputation for obeying customary international law on the line). Nonetheless, State J moves "right" to participate in the game. State I knows this much, but does not know which move, "up" or "down" row-wise, that State J has made or will make--in this scenario, because it understands State J has invoked custom to bind itself, but does not know in fact whether State J will live up to its word.

But State I does suppose that State J would never move down (assert res nullius, with a payoff of 3) after moving right, since that payoff is strictly dominated by a strategy of staying left (and not invoking the language of custom at all, with a payoff of 4). Because State I will always expect State J to play up (adhere to the Truman Proclamation) after it has moved right, State I will always adopt the same policy (that is, by playing "left" column-wise). State J, for its part, knows (1) that State I is rational, and (2) that State I knows that State J will not play dominated strategies, so it can predict with confidence that State I will move left. Aware of that, State J will adopt the strategy, as initially envisioned, of moving right into the bimatrix, leaving both with the optimal payoff of 5 each.

Under the appropriate circumstances, then, states may perceive a sufficient reward--the gap, in a simple game, between their best and next-best options--to justify invoking customary international law and their reputation for abiding by it. (186) Changing conditions and payoffs, to be sure, may overwhelm the interest states have in sustaining their reputations, and the prospect of that happening may equally deter a state's interest in risking its reputation. But both the Battle of the Sexes and the Stag Hunt scenarios suggest that there will be relatively little interest otherwise in undermining any equilibrium once achieved. As the practice of legal discourse indicates, moreover, a state's decision to discount the prospect of changing payoffs--and to qualify purposefully for opinio juris--may be entirely rational.


Suggesting that customary international law may be consistent with rational choice theory does not, of course, wholly redeem the status quo, but only suggests that its claims may be descriptively accurate for a certain class of rules. This is an important finding, particularly in light of the searching criticisms advanced by Professors Goldsmith and Posner. But to the extent this leaves custom in its initial posture, it remains unsatisfactory, given widespread concerns about custom's future potential and its intellectual integrity. In this Part, I describe some implications for custom's prospects and for the reform of its framework principles.

A. Custom's Modest Potential

One lingering impression left by rational choice analysis, certainly, is that states are likely to breach their international obligations when it is rational to do so. It is conceivable that states might want to look past their short-term interests and lash themselves to the mast in order to avoid longer-term harms. But the circumstances in which they would want to be bound unalterably are probably few, (187) and even the cost of committing themselves through treaties or other formal agreements may be excessive in light of the possibility that state interests will change and require breach. (188) Customary international law, like informal agreements, may allow states a relatively moderate means of credibly committing themselves, while at the same time permitting less costly escape when the circumstances demand it. (189)

A further insight, arguably, is that the broad-based, multilateral arrangements ordinarily contemplated by custom are untenable because they are unlikely to arise among a sufficiently large number of states. (190) But this does not exhaust the possibilities of custom. As I have argued elsewhere, the notion of local, special or regional customary international law--custom forged between a small number of relatively homogenous states, binding among them only--is increasingly prevailing over its critics, and represents an incremental strategy that helps resolve both realist and liberal criticisms of custom. (191) As a consequence, groups no larger than those participating in a bilateral game may bind themselves to observe principles of customary international law amongst themselves. Rational choice theory does not discredit this doctrinal byway, but instead helps explain and legitimate its emergence as an alternative to the more problematic alternative of widespread rules.

Both of these implications imply, however, that broad-based custom will arise infrequently, and that is less likely to arise with respect to areas of particular importance to the states involved--where reputation may be a secondary consideration--or in circumstances involving a significant likelihood of deviation. This prediction seems entirely defensible from within the rational choice framework, as does the further implication that customary international law is least likely to exist where it would be most helpful. The greater the potential benefits of cooperation, the greater the incentive of individual states to defect, and the greater the need for draconian enforcement mechanisms (192)--mechanisms absent in the case of custom.

A corollary is that there is also "an inverse relationship between the maximum number of countries that can sustain full cooperation and the total gains to cooperation." (193) This suggests a troubling paradox for norms like those concerning human rights. One of the key claims respecting human rights regards their universality (194)--the notion, among other things, that such rights are not culturally relative. (195) The price for securing this rhetorical advantage, however, may be prohibitive. Intuitively, "selling" norms to a broader and more heterogeneous group will likely water them down and ambiguate them. Further, to the extent that pioneering states are concerned not only with their reputation for conforming to a fully flowered norm (lex lata), but also with regard to the norm they are pioneering (lex ferenda), the prospect that the norm will be altered or breached by late-comers may deter pioneers from investing sufficient reputation in it from the beginning. That possibility, surely, warrants careful attention in any modification to the standard version.

B. Reforming the Customary Rules of Custom

For reasons described in Part I, there is reason to be wary of any free-form efforts to rationalize--for want of a better word--the rules respecting custom's formation. But rational choice analysis, and the redemption of custom via concerns for state reputation, does commend some specific cautions for any revisionist efforts. One concerns the treatment of estoppel, a principle the ICJ has generally been reluctant to invoke against state parties. (196) As noted previously, the International Law Association, in its recently issued Final Report, argued for the more aggressive use of that principle, (197) arguing that the Court's reluctance was "inconsistent with normal judicial attitudes to concessions by the parties, and also with theory." (198)

Whatever theory the Association had in mind, it likely was not rational choice theory, which instead suggests that any estoppel principle should continue to be applied judiciously. In particular, the Stag Hunt scenario--or, less abstractly, the experience with the formation of continental shelf doctrine--suggests that applying estoppel against pioneering states risks fundamentally misunderstanding the nature of their representations, and unnecessarily undermines the mutually advantageous formation of custom. Particularly to the extent that a pioneering state may misconstrue a rule's potential appeal, it would be inappropriate to bind that state to its legal rhetoric in the event that its move failed to galvanize support by other states.

Second, and more important, the sovereign equality of states normally presumed by international law may require qualification. (199) Courts and commentators have sometimes asserted that when assessing state practice, and even when looking for proof of opinio juris, extra weight must be given the states most keenly interested in the problem area at hand. (200) The precise reason has always been unclear. Interested states may generate the most practice and opinion across topics, or create the most defined and stable position on a particular topic, but neither seems to warrant giving their views any special weight. (201) Perhaps such states are more likely to learn of the relevant practices, and thus may be more fairly charged with acquiescence if they fail to object, though that requires increasingly unrealistic assumptions regarding the availability of legal information. (202)

More plausibly, such states, if disregarded, are perhaps more likely to persistently object and destabilize rules with which they disagree. (203) Finally, the fact that interested states will be particularly affected by any customary rule--quite apart from whether that makes them more demonstrative about it--may be seen as warranting giving them special consideration. (204) Either of these rationales, however, would worsen custom's already antidemocratic nature, at least if democratic values are understood to require treating individual states as equal participants in formulating custom. (205) Moreover, while privileging "specially affected" states may be relatively straightforward when it turns on the distinction between coastal and landlocked states, (206) extending any of these rationales to other, less rigidly defined, areas--such as the greater interest nations with space programs have, for the foreseeable future, in rules relating to its exploitation--is obviously more problematic. (207)

Rational choice analysis suggests a different, if not necessarily more manageable, form of discrimination. As the discussion in Part III makes clear, a state's investment in its legal reputation is of paramount importance in considering its potential contribution to the customary law formation: where a state invokes the language of obligation, but in fact has a weak or poor record of adhering to established norms, it is less likely to be making a genuine commitment, and it is correspondingly more likely that any states simulating its behavior will be doing so simply as a matter of coincidence. Distinguishing credible from noncredible states may be particularly important in cases where they are "specially affected," but for very different reasons. If the matter's relative importance among states correlates with the matter's relative importance to the state in question, (208) a state specially affected by a putative custom may place greater value on getting the rule right, but it will also be more inclined to renege on any commitments it has made.

For these reasons, the occasional privileging of specially affected states should be displaced by, or at least married with, a more controversial focus on whether the state in question has established or meaningfully compromised its interest in a law-abiding reputation. (209) Other indicia, like a state's willingness to subject itself to the compulsory jurisdiction of the ICJ, may also be probative. (210) Finally, the awkward relationship between customary international law and treaties may take on a new dimension. As it stands, courts and commentators have noted that treaties "can codify the existing law," "cause the law to crystallize," or "initiate the progressive development of new law," (211) though the position varies by treaty and commentator. (212) The argument here suggests instead that treaties are valuable to the extent they evidence state commitment to a given rule, perhaps by signaling high switching costs or, at a minimum, dissonance in the event positions were to differ. Treaties, in this view, are less valuable as a means of extrapolating obligations from signatories to nonsignatories than as a means of evidencing the credibility of extrinsic commitments made by the parties themselves.


The first generation of rational choice criticism is right, on balance, to be skeptical of claims that customary international law is widespread, or that a particular rule binds a particular state in a matter of keen interest to it. But given persistent recognition of custom as a legal institution, one might be equally skeptical of claims that custom cannot exist, and slow to assume that its principles require legal rules and individual state interests to be antinomies. Reconsidering the application of rational choice theory, it would appear, suggests that customary international law may be a valuable vehicle for simultaneously advancing individual and aggregate state interests, even putting aside the possibility of more altruistic or normative ambitions. Maximizing cooperation's potential requires both improving the cooperation between international legal doctrine and rational choice analysis and generating criteria that better allow the states themselves to decide.

(1.) Viewed broadly, rational choice theory explains behavior as the product of individualistic attempts to optimize outcome preferences, usually reduced to self-interest--though the importance of that assumption, and the relationship between rational choice and kindred methods (such as positive political theory, political economy, and public choice) is the subject of debate. See Daniel A. Farber & Philip P. Frickey, Foreword: Positive Political Theory in the Nineties, 80 GEO. L.J. 457, 458-62 (1992) (describing the sociology of this and related terms). Nothing here turns on terminology. As will become apparent, though, this Article follows the first wave of rational choice criticism in assuming that states are the relevant actors, rather than disaggregating them into other self-interested institutions or actors--a move typically associated with public choice approaches. See, e.g., John K. Setear, Treaties, Custom, Rational Choice, and Public Choice, 94 AM. SOC'Y INT'L L. PROC. 187, 187 (2000) (distinguishing between rational choice and public choice approaches to treatymaking).

(2.) For wonderful illustrations, see DOUGLAS G. BAIRD ET AL., GAME THEORY AND THE LAW (1994).

(3.) Some of the seminal works include Kenneth W. Abbott, Modern International Relations Theory: A Prospectus for International Lawyers, 14 YALE J. INT'L L. 335 (1989); William J. Aceves, The Economic Analysis of International Law: Transaction Cost Economics and the Concept of State Practice, 17 U. PA. J. INT'L ECON. L. 995 (1996); Anne-Marie Slaughter Burley, International Law and International Relations Theory: A Dual Agenda, 87 AM. J. INT'L L. 205 (1993); Jeffrey L. Dunoff & Joel P. Trachtman, Economic Analysis of International Law, 24 YALE J. INT'L L. 1 (1999); Anne-Marie Slaughter et al., International Law and International Relations Theory: A New Generation of Interdisciplinary Scholarship, 92 AM. J. INT'L L. 367 (1998). For a recent survey focused on the issue of compliance, see Kal Raustiala & Anne-Marie Slaughter, International Law, International Relations and Compliance, in HANDBOOK OF INTERNATIONAL RELATIONS 538, 538-58 (Walter Carlsnaes et al. eds., 2002).

(4.) Professors Dunoff and Trachtman suggest that this is part of why law and economics analysis in general has been resisted, though they argue that the objection is misconceived. Dunoff & Trachtman, supra note 3, at 9-12.

(5.) See, e.g., Jack L. Goldsmith & Eric A. Posner, Understanding the Resemblance Between Modern and Traditional Customary International Law. 40 VA. J. INT'L L. 639, 663 (2000) [hereinafter Goldsmith & Posner III (contending that "It]he main problem with international law scholarship ... is that it is too normative"); Anthea Elizabeth Roberts, Traditional and Modern Approaches to Customary International Law: A Reconciliation, 95 AM. J. INT'L L. 757, 769 (2001) ("The greatest criticism of modern custom is that it is descriptively inaccurate because it reflects ideal, rather than actual, standards of conduct."). But see infra notes 49, 111 and accompanying text (noting the normative element of rational choice analysis).

(6.) This simplifying assumption can be faulted, however, for failing to take into account significant subnational and supranational actors that operate as domestic and international constraints. The assumption is not inherent to rational choice analysis, though relaxing it complicates matters. See Andrew Kydd & Duncan Snidal, Progress in Game-Theoretical Analysis of International Regimes, in REGIME THEORY AND INTERNATIONAL RELATIONS 112, 127-34 (Volker Rittberger ed., 1993) (discussing "two level" rational choice games involving domestic and international fronts); Duncan Snidal, Rational Choice and International Relations, in HANDBOOK OF INTERNATIONAL RELATIONS, supra note 3, at 84-85 (discussing the incorporation of additional actors and changed preferences); cf. Edward Rubin, Rational States?, 83 VA. L. REV. 1433, 1439-51 (1997) (discussing the legitimacy of viewing governments in federal systems as rational actors).

(7.) Early examples of work applying rational choice analysis to treaties include JON HOVI, GAMES, THREATS AND TREATIES: UNDERSTANDING COMMITMENTS IN INTERNATIONAL RELATIONS 1 (1998) (discussing "commitment problems in international politics"); John K. Setear, An Iterative Perspective on Treaties: A Synthesis of International Relations Theory and International Law, 37 HARV. INT'L L.J. 139, 139-42 (1996) (applying institutionalist theory to the law of treaties); John K. Setear, Law in the Service of Politics: Moving Neo-Liberal Institutionalism from Metaphor to Theory by Using the International Treaty Process to Define "Iteration," 37 VA. J. INT'L L. 641,642 (1997) (arguing "that the international legal rules governing ... treaty obligations can provide some crucial definition to ['neo-liberal Institutionalism']").

(8.) For the avoidance of doubt, I refrain from using that shorthand to describe state practices not necessarily rising to the level of customary international law, though that would be an equally valid usage.

(9.) RESTATEMENT (THIRD) OF FOREIGN RELATIONS LAW pt. 1, ch. 1, introductory note at 18-19 (1987) [hereinafter RESTATEMENT (THIRD)]. The conventional wisdom that custom undergirds the obligation to abide by treaties simply reinforces, of course, the importance of explaining why states are any more obligated by custom--without moving much closer to an answer. G.G. Fitzmaurice, The Foundations of the Authority of International Law and the Problem of Enforcement, 19 MOD. L. REV. 1, 9 (1956).

(10.) Rosemary Bennett, Straw Warns U.S. on Issue of Terror Suspects, FIN. TIMES (London), Jan. 21, 2002, at 1 (quoting Jack Straw, U.K. foreign secretary).

(11.) See, e.g., Brigitte Stern, Custom at the Heart of International Law, 11 DUKE J. COMP. & INT'L L. 89, 89 (2001) ("It is not contested ... that custom enjoys privileged status in the international order: 'custom is even more central than the treaty.'" (quoting and translating PAUL REUTER, INTRODUCTION AU DROIT DES TRAITES 38 (1972))).

(12.) See, e.g., N.C.H. Dunbar, The Myth of Customary International Law, 8 AUSTL. Y.B. INT'L L. 1, 2 (1983) (arguing that customary international law is "merely a fiction or myth"); J. Patrick Kelly, The Twilight of Customary International Law, 40 VA. J. INT'L L. 449, 481-82 (2000) (stating that international legal rules, such as jurisdictional principles, are often ignored); Roberts, supra note 5, at 757 (noting that "[t]he demise of custom as a source of international law has been widely forecasted"); Phillip R. Trimble, A Revisionist View of Customary International Law, 33 UCLA L. REV. 665, 665 (1986) (noting the "popular view [of] international law" is that it is "a charade").

(13.) Compare DONALD P. GREEN & IAN SHAPIRO, PATHOLOGIES OF RATIONAL CHOICE THEORY 33-46 (1994) (describing methodological and empirical limits to rational choice analysis), with Snidal, supra note 6, at 79-82 (noting criticisms of rational choice theory, but discussing some of the "significant and growing body" of empirical rational choice research).

(14.) Jack L. Goldsmith & Eric A. Posner, A Theory of Customary International Law, 66 U. CHI. L. REV. 1113, 1136 (1999) [hereinafter Goldsmith & Posner I]; Goldsmith & Posner II, supra note 5, at 654-60; Jack L. Goldsmith & Eric A. Posner, Further Thoughts on Customary International Law, 23 MICH. J. INT'L L. 191, 192 (2001) [hereinafter Goldsmith & Posner III[; Jack L. Goldsmith & Eric A. Posner, Moral and Legal Rhetoric in International Relations: A Rational Choice Perspective, 31 J. LEGAL STUD. S115, S124 (2002) [hereinafter Goldsmith & Posner IV]; Setear, supra note 1, at 188. For an earlier discussion, see FERNANDO R. TESON, A PHILOSOPHY OF INTERNATIONAL LAW 73-103 (1998).

(15.) After this Article was completed and accepted for publication, two articles emerged that grappled with the application of game theory to customary international law: Mark A. Chinen, Game Theory and Customary International Law: A Response to Professors Goldsmith and Posner, 23 MICH. J. INT'L L. 143 (2001) [hereinafter Chinen II: Peirre-Hugues Verdier, Cooperative States: International Relations, State Responsibility and the Problem of Custom, 42 VA. J. INT'L L. 839 (2002). See also Mark A. Chinen, Afterword, 23 MICH. J. INT'L L. 201,201 (2001) (responding to a reply by Professors Goldsmith and Posner) [hereinafter Chinen II]; cf Michael Byers, Taking the Law Out of International Law: A Critique of the "Iterative Perspective". 38 HARV. INT'L L.J. 201, 201-05 (1997) (criticizing an earlier application to law of treaties). Mr. Verdier's analysis appears largely to accept the critical stance taken by Professors Goldsmith and Posner with respect to substantive customary rules, but would differentiate "secondary rules" like those regarding state responsibility. Verdier, supra, at 850. Professor Chinen's analysis applies more readily to substantive rules, though it equally emphasizes the function of treaties in coordinating cooperation. Chinen I, supra, at 145. The analysis here differs in important regards, particularly as concerns the fundamental challenge that rational choice analysis poses for the opinio juris requirement.

A third, contemporaneous paper is closest in approach to the present Article, though it too came to my attention only belatedly: Andrew T. Guzman, International Law: A Compliance Based Theory, 90 CAL. L. REV. 1823 (2002). Professor Guzman's focus in this exceptionally interesting paper is on the broader question of compliance with international law, but he takes a similar view of the effect of reputation on the possibility of compliance with customary international law--that is, extending the relevant models beyond one-shot games. Compare id. at 1847-87, with infra Part III.B. He does not address, however, the potential compatibility of custom with the particular games discussed here (or the chosen illustrations), and unlike this Article, comes to the conclusion that customary international law must be wholly reconceived in order to withstand the rational choice analysis. Guzman, supra, at 1874-78. But see infra note 178.

(16.) See, e.g., MARTHA FINNEMORE, NATIONAL INTERESTS IN INTERNATIONAL SOCIETY 28-29, 139-43 (1996) (advancing a constructivist account of custom).

(17.) See generally Harold Hongju Koh, Why Do Nations Obey International Law?, 106 YALE L.J. 2599 (1997) (reviewing ABRAM CHAYES & ANTONIA HANDLER CHAYES, THE NEW SOVEREIGNTY: COMPLIANCE WITH INTERNATIONAL REGULATORY AGREEMENTS (1995) and THOMAS M. FRANCK, FAIRNESS IN INTERNATIONAL LAW AND INSTITUTIONS (1995), and discussing competing theories of compliance). But see infra note 182 (noting emerging empirical work on compliance).

(18.) See, e.g., Frederic L. Kirgis, Jr., Custom on a Sliding Scale, 81 AM. J. INT'L L. 146, 149 (1987) (explaining that, in practice, "[t]he more destabilizing or morally distasteful the activity ... the more readily international decision makers will substitute one element [opinio juris or state practice] for the other," and vice versa); Roberts, supra note 5, at 774-84, 788-91 (advocating an approach to custom incorporating sensitivity to procedural and substantive normativity). For normative criticisms of employing game theory, see TESON, supra note 14, at 84, 88-92; compare Chinen I, supra note 15, at 156-59 (suggesting that a gulf exists between game theory critiques and normative evaluation of law).

(19.) Cf. Saul Levmore, From Cynicism to Positive Theory in Public Choice, 87 CORNELL L. REV. 375, 376 (2002) (arguing that "[p]ublic choice does not, or at least need not, claim to be the only game in town; it simply follows the common pattern found in the social sciences and sciences of making simplifying assumptions--beginning in this case with rational political actors").

(20.) Statute of the International Court of Justice, June 26, 1945, art. 38, 59 Stat. 1055, 1060, 3 Bevans 1153, 1187 (directing courts to "apply ... international custom, as evidence of a general practice accepted as law"); e.g., Continental Shelf (Libya v. Malta), 1985 I.C.J. 13, 29-30 [paragraph] 27 (June 3) (looking to state practice and opinio juris as primary evidence of customary international law). As one chamber of the ICJ recognized, "a chamber of the Court in its reasoning on the matter, must obviously begin by referring to Article 38, paragraph 1, of the Statute of the Court." Delimitation of the Maritime Boundary in the Gulf of Maine Area (Can. v. U.S.), 1984 I.C.J. 245, 290 (Oct. 12); see also KAROL WOLFKE, CUSTOM IN PRESENT INTERNATIONAL LAW 1-5 (2d rev. ed. 1993) (describing the authority and genesis of Article 38).

(21.) As is often observed, the Statute's literal direction that courts are to apply "international custom, as evidence of a general practice accepted as law," Statute of the International Court of Justice, supra note 20, art. 38, 59 Stat. at 1060, 3 Bevans at 1187, appears to have the test exactly backwards--that is, a general practice accepted as law is properly regarded as evidence of international custom, not vice versa. See WOLFKE, supra note 20, at 5-9 (describing criticism of Article 38's terms). But this nonliteral version is the one actually applied, even by the ICJ. ROSALYN HIGGINS, PROBLEMS & PROCESS: INTERNATIONAL LAW AND HOW WE USE IT 18-19 (1994).

(22.) The terms, of course, have varied. The best-known alternative formulation, developed by Judge Manley O. Hudson for the International Law Commission, would require:

[1] concordant practice by a number of States with reference to a type of situation falling within the domain of international relations;

[2] continuation or repetition of the practice over a considerable period of time;

[3] conception that the practice is required by, or consistent with, prevailing international law; ...

[4] general acquiescence in the practice by other States[; and]

[5] ... [establishment of] the presence of each of these elements ... by a competent international authority.

Manley O. Hudson, Article 24 of the Statute of the International Law Commission, [1950] 2 Y.B. Int'l L. Comm'n 24, 26, U.N. Doc. A/CN.4/16; see also RESTATEMENT (THIRD), supra note 9, [section] 102(2) ("Customary international law results from a general and consistent practice of states followed by them from a sense of legal obligation."). For reviews of varying formulations, see IAN BROWNLIE, PRINCIPLES OF PUBLIC INTERNATIONAL LAW 4-11 (5th ed. 1998); Jo Lynn Slama, Note, Opinio Juris in Customary International Law, 15 OKLA. CITY U. L. REV. 603 (1990).

(23.) This practice requirement immediately qualifies the normative character of custom. See, e.g., Daniel Bodansky, Customary (And Not So Customary) International Environmental Law, 3 IND. J. GLOBAL LEGAL STUD. 105, 109 (1995) (describing the traditional approach to practice as an attempt "to describe the existing norms that govern the relations among states" without "advocat[ing] or prescrib[ing] new norms," and which "draws a clear distinction between lex lata (law as it is) and lex ferenda (law as it should be)"). It also distinguishes custom from other forms of international law. With custom, it is clear, deviance by a pivotal number of states could decisively repudiate a rule. Other international legal principles are understood to have a binding character notwithstanding state deviance. See, e.g., CHAYES & CHAYES, supra note 17, at 113-14 (explaining that "departure from a norm, even frequent or persistent departure, does not necessarily invalidate it"). How many states would have to deviate from custom in order to change it, or conform in order to establish custom in the first instance, are among the many difficulties with the practice element. See, e.g., Kelly, supra note 12, at 503-07 (describing flaws in the theory and practice of state practice).

(24.) See, e.g., North Sea Continental Shelf (F.R.G.v. Den.: F.R.G.v. Neth.), 1969 I.C.J. 3, 44 (Feb. 20):
   The States concerned must therefore feel that they are conforming to
   what amounts to a legal obligation. The frequency, or even habitual
   character of the acts is not in itself enough. There are many
   international acts, e.g., in the field of ceremonial and protocol,
   which are performed almost invariably, but which are motivated only
   by considerations of courtesy, convenience or tradition, and not by
   any sense of legal duty.

S.S. Lotus (Fr. v. Turk.), 1927 P.C.I.J. (ser. A) No. 9, at 28 (Sept. 7) (concluding that abstention from exercising criminal jurisdiction over acts committed aboard vessels on the high seas could be considered an international custom "only if such abstention were based on their being conscious of having a duty to abstain"); MALCOLM N. SHAW, INTERNATIONAL LAW 58 (4th ed. t997) (discussing the relative importance of overt action and opinio juris); Michael Akehurst, Custom as a Source of International Law, 1974-75 BRIT. Y.B. INT'L L. 1, 33 (explaining that "opinio juris is ... needed in order to distinguish legal obligations from non-legal obligations, such as obligations derived from considerations of morality, courtesy or comity").

States might agree to be so obliged, of course, whether by treaty or otherwise. But most agree that consent is an inadequate explanation for much of what is called customary international law. E.g., MICHAEL BYERS, CUSTOM, POWER, AND THE POWER OF RULES 142-46 (1999) (discussing the validity of inferred consent and problems arising from it). But see, e.g., WOLFKE, supra note 20, at 162 (concluding that states may be bound "exclusively by their own sovereign will"). Even consent-oriented theories typically allow for consent by tacit or passive means, though any presumption favoring consent seems counterfactual. Maurice Mendelson, The Subjective Element in Customary International Law, 1995 BRIT. Y.B. INT'L L. 177, 185-94.

(25.) See, e.g., ANTHONY A. D'AMATO, THE CONCEPT OF CUSTOM IN INTERNATIONAL LAW 82-84 (1971) (illustrating the difficulties of finding the subjective intent of a state and its decisionmakers).

(26.) As one commentator put it,
   The precise definition of ... the psychological element in the
   formation of custom, the philosophers' stone which transmutes the
   inert mass of accumulated usage into the gold of binding legal rules,
   has probably caused more academic controversy than all the actual
   contested claims made by States on the basis of alleged custom, put


(27.) For illustrative descriptions of the chronological problem, and the range of purported solutions--including the "mistake" argument--see BYERS, supra note 24, at 130-33 (describing the "chronological paradox" and various responses and circumventions); D'AMATO, supra note 25, at 47-56, 66-68 (describing the "error" hypothesis in which states believe they are following a legal norm when none exists); O.A. ELIAS & C.L. LIM, THE PARADOX OF CONSENSUALISM IN INTERNATIONAL LAW 1-28 (1998) (discussing workable theories and solutions); Olufemi Elias, The Nature of the Subjective Element in Customary International Law, 44 INT'L & COMP. L.Q. 501,502-08 (1995) (arguing that opinio juris, although problematic, is still necessary).

(28.) One might, for example, relieve initiating states of any need to (mistakenly) perceive their actions as obligatory. See Elias, supra note 27, at 508 n.33 (explaining that, under theories that postpone rather than deny the operation of opinio juris in the creation of law, change may actually supplement the legal force of a new practice because it can be declared and, thereby, recognized). But that approach places great weight on distinguishing between the initiating states and latecomers, without offering any justification for treating the second and third (or forty-first and forty-second) states differently.

(29.) RESTATEMENT (THIRD), supra note 9, [section] 102 cmt. c (stating that "[e]xplicit evidence of a sense of legal obligation (e.g., by official statements) is not necessary; opinio juris may be inferred from acts or omissions").

(30.) E.g., Kirgis, supra note 18, at 148-51 (arguing that the ICJ does deemphasize opinio juris when state practice is clear and consistent); Mendelson, supra note 24, at 202-08 (arguing that international and state courts rely on well-established practice when it exists, without the need for opinio juris).

(31.) For the seminal discussion, see Bin Cheng, United Nations Resolutions on Outer Space: "Instant" International Customary Law?, 5 INDIAN J. INT'L L. 23, 35-40, 45-48 (1965).

(32.) As the Permanent Court of International Justice stated in the Lotus judgment,
   International law governs relations between independent States. The
   rules of law binding upon States therefore emanate from their own
   free will as expressed in conventions or by usages generally accepted
   as expressing principles of law and established in order to regulate
   the relations between these co-existing independent communities or
   with a view to the achievement of common aims.

S.S. Lotus (Fr. v. Turk.), 1927 P.C.I.J. (ser. A) No. 9, at 18 (Sept. 7); accord Military and Paramilitary Activities in and Against Nicaragua (Nicar. v. U.S.), 1986 I.C.J. 14, 135 (June 27) ("[I]n international law there are no rules, other than such rules as may be accepted by the State concerned, by treaty or otherwise.... "). Nonstate actors play a substantial, and increasing, role in generating international norms, but virtually everyone regards states as central and indispensable participants. E.g., BYERS, supra note 24, at 13 (using the "working assumption" that states "are the principal actors").

(33.) Professor Akehurst insisted on this distinction. Akehurst, supra note 24, at 36-37 (agreeing with Professor D'Amato that, in regard to the obligatory character of a particular custom "what matters is statements, not beliefs").

(34.) For an unapologetic description of how states and their legal advisers rely on the academy, see Louis B. Sohn, Sources of International Law, 25 GA. J. INT'L & COMP. L. 399, 401 (1995-96) (asserting, with approval, that "international law is made, not by states, but by 'silly' professors writing books, and by knowing where there is a good book on the subject"); id. at 399 (submitting that "states really never make international law on the subject of human rights," but rather "lilt is made by the people that care," namely professors and other authors). For one criticism of legal solipsism, see Daniel M. Bodansky, The Concept of Customary International Law, 16 MICH. J. INT'L L. 667, 677-78 (1995) (reviewing WOLFKE, supra note 20, and concluding that, while Wolfke's book is a "good example" of traditional legal scholarship, its exegesis of legal texts does little to teach us "how custom actually operates").

(35.) See infra notes 105-07 and accompanying text.

(36.) COMM. ON FORMATION OF CUSTOMARY (GEN.) INT'L LAW, INT'L LAW ASS'N, FINAL REPORT: STATEMENT OF PRINCIPLES APPLICABLE TO THE FORMATION OF GENERAL CUSTOMARY INTERNATIONAL LAW (2000) [hereinafter FINAL REPORT], The Final Report contains both formal principles and accompanying commentary; for clarity's sake, citations to the former are by section.

(37.) E.g., id. introduction at 4 (describing the report's inductive approach to custom, by which "the rules about the sources of international law, and specifically [customary international law], are to be found in the practice of States ... not [in] an a priori method of reasoning").

(38.) Id. introduction at 3.

(39.) Id. [section] 18; see also id. [section] 14(ii) (providing that "for a specific State to be bound by a rule of general customary international law it is not necessary to prove that it participated actively in the practice or deliberately acquiesced in it"); id. [section] 14 cmts. a-e (discussing further the parameters under which a state may become bound by custom).

(40.) Id. [section] 18 cmt. d.

(41.) Id. [section] 16; see id. [section] 1(iii) (stating that consent need not be shown before a state will be bound by rules of general customary international law). The principles allow, however, that state claims may sometimes contraindicate a conclusion that particular conduct gives rise to a customary rule, where: (1) the practices, by their nature, are not of the kind regarded by states as giving rise to legal obligations; (2) practices otherwise capable of creating legal obligations are undermined by the contrary understanding of states as a whole; or (3) the conduct is too ambiguous to be regarded as precedent-setting, absent proof "that the State or States concerned intended, understood or accepted that a customary rule could result from, or lay behind, the conduct in question." Id. [subsection] 17(i), (ii), (iii), & (iv).

(42.) E.g., id. [section] 1(iii) ("Where a rule of general customary international law exists, for any particular State to be bound by that rule it is not necessary to prove either that State's consent to it or its belief in the rule's obligatory or (as the case may be) permissive character.").

(43.) Id. [section] 19 (asserting that strong evidence of opinio juris "may compensate for a relative lack of practice, and vice versa"); id. [section] 19 cmt. a, at 40 (noting the prior submission that "the subjective element is not in fact usually a necessary ingredient in the formation of customary international law--certainly on the part of any given State which is allegedly bound by the putative customary rule"); id. [section] 19 cmt. a, at 41 ("[W]hatever the theory, the result is the same: the more practice, the less the need for the subjective element.").

(44.) It is telling, for example, that while the Final Report asserts that strong evidence of opinio juris "may compensate for a relative lack of practice, and vice versa," id. [section] 19, it does not address the possibility that a marginal sufficient showing of practice may make necessary an especially strong showing of opinio juris (or vice versa).

(45.) Some scholars have eloquently argued, for example, that because the standard version of custom's elements retards its ability to keep pace with global needs, relaxation of those elements-or the development of new international legal forms--is necessary. E.g., Jonathan I. Charney, Universal International Law, 87 AM. J. INT'L L. 529, 529-31 (1993) (arguing that in order to resolve pressing global problems, "it may be necessary"--contrary to international legal traditions--"to establish new rules that are binding on all subjects of international law regardless of the attitude of any particular state"); Krista Singleton-Cambage, International Legal Sources and Global Environmental Crises: The Inadequacy of Principles. Treaties, and Custom, 2 ILSA J. INT'L & COMP. L. 171, 171-73 (1995) (summarizing the claim that "international law is largely inadequate in the face of current environmental crises" to accommodate human needs and state interests). Others, however, would be content with letting international agreements pick up the slack. E.g., W. Michael Reisman, The Cult of Custom in the Late 20th Century, 17 CAL. W. INT'L L.J. 133, 133-35 (1987) (acknowledging custom's limitations in coping with emerging global problems, but suggesting that its real deficiencies run deeper).

(46.) See supra note 12 (citing authorities).

(47.) See supra note 6 and accompanying text (noting the potential relevance of subnational and supranational actors that operate as domestic and international constraints).

(48.) See supra note 1 (noting the relevance of self-interest to defining rational choice theories); infra note 111 and accompanying text (noting the "thin" and "thick" conceptions of interests).

(49.) See Goldsmith & Posner II, supra note 5, at 662 ("The rational choice account seeks to explain accurately the behaviors associated with CIL [customary international law]. Whether CIL is or is not the law is beyond its concern."); supra note 5 and accompanying text (noting complaint about custom's normativity). Some would argue, though, that the rational choice approach is itself normative in character. See, e.g., GEORGE TSEBELIS, NESTED GAMES: RATIONAL CHOICE IN COMPARATIVE POLITICS 30-31 (1990) (arguing that assumptions regarding rationality are normative in character); Jon Lister, Introduction to RATIONAL CHOICE 1, 1 (Jon Elster ed., 1986) (claiming that rational choice is a normative theory before it is anything else): Snidal, supra note 6, at 85-86 (emphasizing the normative dimensions of rational choice).

(50.) For a description of Nash equilibria, see BAIRD ET AL., supra note 2, at 21-23.

(51.) See JAMES D. MORROW, GAME THEORY FOR POLITICAL SCIENTISTS 75-76 (1994) (illustrating parlance).

(52.) To be clear, however, regarding a game as noncooperative in nature does not mean that the actors cannot cooperate, nor does it prejudge the question of whether they can be expected to live up to their commitments. HOVI, supra note 7, at 4 & n.4.

(53.) See infra notes 74-75 and accompanying text.

(54.) See BAIRD ET AL., supra note 2, at 9-10 (describing routine assumptions in normal form games).

(55.) Goldsmith & Posner I, supra note 14, at 1122-23.

(56.) This game is drawn directly from id. at 1122. This assumes a baseline payoff of 3 for the ignore/ignore cell; a loss of 4 for attacking the other state; and a loss of 1 in the event a state is itself attacked.

(57.) For purposes of illustrating this game, I assume the same costs and benefits specified by Professors Goldsmith and Posner--that is, a benefit of 3 for ignoring the other state's fishing vessels and sparing one's own navy, a price of 4 for attacking the other state's boats, and an additional price of 1 for being attacked oneself--save that State J achieves a benefit of 5 by destroying State I's fishing fleet, perhaps because it wards off overfishing or some other environmental harm. One can also imagine payoffs that would encourage State J to attack only when State I did not.

(58.) Goldsmith & Posner I, supra note 14, at 1123-24.

(59.) State I might, in the alternative, seek only to minimize the advantage accruing to State J from attacking, but that too is unlikely to be costless.

(60.) Cf. Lisa L. Martin, The Rational State Choice of Multilateralism, in MULTILATERALISM MATTERS: THE THEORY AND PRAXIS OF AN INSTITUTIONAL FORM 103-06 (John Gerard Ruggie ed., 1993) (describing suasion problems).

(61.) This game is directly derived from Goldsmith & Posner I, supra note 14, at 1124.

(62.) Id. at 1124-25.

(63.) See ROBERT AXELROD, THE EVOLUTION OF COOPERATION 169-91 (1984) (describing experimental and anecdotal evidence of cooperative strategies).

(64.) This game is derived directly from Goldsmith & Posner I, supra note 14, at 1127.

(65.) Id. at 1148-49.

(66.) Id. at 1154, 1151-58.

(67.) Id. at 1167.

(68.) Goldsmith & Posner II, supra note 5, at 654-60.

(69.) Goldsmith & Posner I, supra note 14, at 1139 (claiming that behavioral regularities reflecting the profiled games, "rather than the notion of universal state practices followed from a sense of legal obligation[,] account for the CIL identified by courts and scholars"); Goldsmith & Posner II, supra note 5, at 655 (introducing "a rational choice account of the behaviors associated with CIL compliance" corresponding to three of the four games, and noting the potential relevance of coordination to other problems).

(70.) Using rational choice theory to explain particular results is generally fraught with peril. GREEN & SHAPIRO, supra note 13, at 20-23. So too are attempts to provide universalist explanations for phenomena, even within discrete areas like customary international law. Id. at 26-30.

(71.) Goldsmith & Posner I, supra note 14, at 1132.

(72.) Id.

(73.) I may misunderstand them on this point. They explicitly "deny the claim that CIL is an exogenous influence on states' behavior," leaving the possibility that CIL is an endogenous influence--that is, that it affects the payoffs in some way. Id. I would have sympathy with any such account. See infra text accompanying note 108. But they also specifically contrast themselves with those claiming that "the sense of legal obligation puts some drag" on deviations from customary norms, Goldsmith & Posner I, supra note 14, at 1132, and it would be inconsistent with the spirit of their argument were norms to come in through the back door. See also Goldsmith & Posner II, supra note 5, at 672 (claiming that "[n]ations would act no differently if CIL were not a formally recognized source of law").

(74.) Id. at 1132.

(75.) Id. at 1132-33.

(76.) States may also lack information even about one another's previously revealed preferences. See, e.g., Kydd & Snidal, supra note 6, at 117 (noting that "in reality, states often have remarkably vague knowledge of the other's past behaviour").

(77.) Goldsmith & Posner I, supra note 14, at 1128-31, 1132-33.

(78.) Cf. JON ELSTER, THE CEMENT OF SOCIETY 50-96, 152-185 (1989) (describing inefficiencies and collective action problems associated with bargaining under conditions of uncertainty).

(79.) BAIRD ET AL., supra note 2, at 188.

(80.) Id. at 45; Kydd & Snidal, supra note 6, at 114 (noting "[t]here is a significant danger that models will dominate the substantive analysis in ways that distort rather than enhance our understanding of international regimes").

(81.) Thus, for example, Professors Goldsmith and Posner mention several times, but do not explore at length, the coordination variant popularly known as the Battle of the Sexes. Goldsmith & Posner I, supra note 14, at 1128 n.39, 1154 n.111, 1166 n.168. I describe this game a little more in Part III.A.1, but for immediate purposes would note that it contains a mixed message for the potential contribution of legal norms. The availability of more than one equilibrium presumably makes it less likely that states would settle in on a behavioral regularity of the kind that might be mistaken for legal rule--the distributional consequences, presumably, will be a distraction and irritant. This makes it an inferior hypothesis to consider routinely. At the same time, if the paradigm is more descriptively accurate, it may suggest that something else--potentially a rule, or some other kind of link among games--is going on, and that the simple and relatively autonomous self-interest is less likely to be a complete explanation. Cf. Martin, supra note 60, at 101 (noting, in contemplating a battle of the sexes, that "[c]oordination games can have major distributional implications, which sometimes make cooperative solutions difficult to achieve," but that stability may be achieved "once an equilibrium has been established either by convention or by agreement"). The immediate point, put simply, is that modeling the world of custom through the four games selected may have the unintended effect of suppressing richer explanations of particular moment to those taking on the traditional account of custom.

(82.) They clearly regard this empirical dimension as central to their analysis. See Goldsmith & Posner III, supra note 14, at 192 (replying to Professor Chinen).

(83.) Such immunity was widely understood to exist well prior to the Vienna Convention, which largely codified customary international law. EILEEN DENZA, DIPLOMATIC LAW: A COMMENTARY ON THE VIENNA CONVENTION ON DIPLOMATIC RELATIONS 229-30 (2d ed. 1998); Rosalyn Higgins, The Abuse of Diplomatic Privileges and Immunities: Recent United Kingdom Experience, 79 AM. J. INT'L L. 641, 641-42 (1985); see also Vienna Convention on Diplomatic Relations and Optional Protocol on Disputes, Apr. 18, 1961, art. 31(1), 23 U.S.T. 3227, 3240-41,500 U.N.T.S. 96, 112 ("A diplomatic agent shall enjoy immunity from the criminal jurisdiction of the receiving State."); id. art. 41, 23 U.S.T. at 3247, 500 U.N.T.S. at 120 (stating that the duty of diplomats to observe the laws of the receiving State is "without prejudice to their privileges and immunities").

(84.) Goldsmith & Posner I, supra note 14, at 1152-53. Note that there are two kinds of communication interests at play here: a state's interest in fostering "closed" communication with its own diplomatic agents stationed abroad (involving, for example, the state's tendering of diplomatic instructions, and its receipt from the agent of espionage), and a state's use both of its own agents and foreign diplomats stationed on its territory for bilateral communications. The latter complicates the analysis of otherwise rival interests, since a host state's threat to take retaliatory action against foreign diplomats stationed on its territory may seem to be cutting off its nose to spite its face. The suggestion that "each state also benefits if it can successfully harass or harm the other state's ambassador, thereby preventing the ambassador from engaging in espionage, without causing a breakdown of communication with the ambassador's home state," id. at 1154, thus delimits what may be a very fine line, and it is unclear how a typical state would evaluate the relevant risks.

(85.) See LINDA S. FREY & MARSHAL. FREY, THE HISTORY OF DIPLOMATIC IMMUNITY 494-96 (1999) (describing diplomatic smuggling as a "common," and seemingly intractable, abuse of privilege).

(86.) As Professors Goldsmith and Posner note, prosecution of diplomats for criminal offenses would help "to preserve the integrity of the criminal law," Goldsmith & Posner I, supra note 14, at 1153, presumably by showing that no one is above the law. But alternative sanctions are available that might partially satisfy that end, such as prosecution by the offending diplomat's own state, and host states may seek--and sometimes do obtain--a waiver of the immunity. See FREY & FREY, supra note 85, at 496-99 (citing examples of when countries waived a diplomat's immunity); CLIFTON E. WILSON, DIPLOMATIC PRIVILEGES AND IMMUNITIES 81-82 (1967) (citing examples of when countries waived a diplomat's immunity and when a diplomat's own state initiated proceedings). Host states also have the unilateral recourse of expulsion. See Vienna Convention on Diplomatic Relations and Optional Protocol on Disputes, supra note 83, art. 9(1), 23 U.S.T. at 3231-32, 500 U.N.T.S. at 102 (codifying the principle that a host state "may at any time and without having to explain its decision," declare any member of a diplomatic mission persona non grata, or simply not "acceptable," obligating the sending state to "either recall the person concerned or terminate his functions with the mission"); FREY & FREY, supra note 85, at 496-99 (citing examples).

(87.) Professors Goldsmith and Posner acknowledge that preserving a communication route to the foreign state may exact a cost in terms of espionage. See Goldsmith & Posner I, supra note 14, at 1153 (recognizing that open communication routes enable diplomats to engage in espionage). But national security concerns are not prominent when the prospects for two-state cooperation are being assessed, id. at 1153-54, though one could well imagine them playing a prominent role in any rational choice explanation as to why a customary norm had not emerged. But see WILSON, supra note 86, at 83-86 (describing instances in which national security concerns have trumped respect for diplomatic immunity).

(88.) There are abundant means for communicating with a foreign state. See HIGGINS, supra note 21, at 86 (identifying permanent diplomatic missions, visits by heads of governments, special missions, and official representations at ad hoc or regular conferences). If communication is paramount, moreover, alternative sanctions like expulsion, see supra note 86, would seem to pose similar problems, at least in the short term.

(89.) See Goldsmith & Posner I, supra note 14, at 1153-54 (asserting that each country will want to retain a diplomatic presence in the other country). Their commitment to the stability of this cooperation is, however, equivocal. For example, the claim that a bilateral immunity rule is made feasible because "[t]his cooperative strategy (immunity) has a clear all-or-nothing quality that is relatively easy to monitor," id. at 1154, is marginally inconsistent with their recognition elsewhere that violations of that rule "will not necessarily be overt," id. at 1156 n.117.

(90.) The claim that the terms of diplomatic immunity are in fact determined bilaterally rather than multilaterally is difficult to evaluate. Id. at 1154-58. Nothing in their description is strictly inconsistent with the idea of a multilateral custom subject to reservation by bilateral, local customs, and it is hard to assess the implicit claim that the differences across bilateral relationships are more marked than the similarities. It also slights the significance of multilateral negotiations and international organizations, such as the United Nations, where discrimination among foreign diplomats is relatively difficult. See Convention on Privileges and Immunities of the United Nations, Feb. 13, 1946, art. 4(11), 21 U.S.T. 1418, 1426-30, 1 U.N.T.S. 15, 20-23 (extending the same privileges and immunities accorded other diplomats to "Representatives of Members to the principal and subsidiary organs of the United Nations and to conferences convened by the United Nations"); United Nations Headquarters Agreement, June 26, 1945, [section] 15, 61 Stat. 3416, 3428, 11 U.N.T.S. 11, 26 (requiring the United States to accord to representatives of United Nations member states "the same privileges and immunities ... as it accords to diplomatic envoys accredited to it"). Finally, the terms of diplomatic immunity have in an important sense been socially constructed by international discussions, as the punctuated evolution of theories of diplomatic immunity indicates. See WILSON, supra note 86, at 1-25 (examining traditional theories justifying diplomatic privileges and immunities). While the role of such theorizing should not be exaggerated, it seems plausible that it influences the way states conceive of their responsibilities, which in turn may have an effect on patterns of conduct and customary rules.

(91.) Goldsmith & Posner I, supra note 14, at 1154 n.111; see also supra note 81 (describing and noting references to the Battle of the Sexes game).

(92.) See infra Part III.A.1.

(93.) See, e.g., Kydd & Snidal, supra note 6, at 115 (acknowledging that "the rational model is very flexible in terms of, for example, attribution of preferences and information to different actors [so] a clever scholar can reconstruct any single empirical case from a rational actor perspective").

(94.) See FREY & FREY, supra note 85, at 373-74 (noting the ascendance of functional necessity and reciprocity as a rationale for immunity); WILSON, supra note 86, at 17-25 (same). It may be questioned, though, whether any rule directly premised on functional necessity would have sufficient integrity to be characterized as a legal rule. See BROWNLIE, supra note 22, at 351 (describing the functional theory as "fashionable but somewhat question-begging"); WILSON, supra note 86, at 22 (noting that "some" might characterize such a rule as "disturbingly vague").

(95.) Goldsmith & Posner I, supra note 14, at 1152; cf. infra note 139 and accompanying text (discussing the claim that customs concerning the territorial sea lose their integrity if they are rife with exceptions).

(96.) Id. at 1156-57. Indeed, these considerations helped motivate the move to supplant custom with the Vienna Convention, as well as subsequent initiatives to strengthen those norms. FREY & FREY, supra note 85, at 497-538.

(97.) See Goldsmith & Posner III, supra note 14, at 192 (providing a nonexhaustive list of findings from empirical studies).

(98.) Goldsmith & Posner I, supra note 14, at 1131-32.

(99.) See Cruzan v. Dir., Mo. Dep't of Health, 497 U.S. 261, 294 (1990) (Scalia, J., concurring) (describing common law criminal liability for suicide in evaluating the claim that limiting assisted suicide violated substantive due process).

(100.) To this extent, their reply to Professor Chinen seems well taken. Compare Chinen I, supra note 15, at 178 ("As long as there is a practice [in which states refrain from attacking other states' coastal vessels], it should not matter why it is followed ... and nothing in the standard account of customary international law requires otherwise."), with Goldsmith & Posner III, supra note 14, at 194 ("But of course the standard account does require otherwise. It requires that the behavioral regularities be followed from a sense of obligation.").

(101.) See AXELROD, supra note 63, at 55-79 (specifying conditions for stability).

(102.) Goldsmith & Posner II, supra note 5, at 663-65. This enhances and qualifies somewhat the explanation provided in their initial work. See Goldsmith & Posner I, supra note 14, at 113-538 (proposing two models--the rogue state model and the coordination model--to explain how states care about their reputation and do not admit to violations of customary international law, but instead justify their actions by reinterpreting their own prior statements).

(103.) Sir Robert Jennings, President Emeritus of the International Court of Justice, once observed that "most of what we perversely persist in calling customary international law is not only not customary law: it does not even faintly resemble a customary law." Robert Y. Jennings, The Identification of International Law, in INTERNATIONAL LAW: TEACHING AND PRACTICE 3, 5 (Bin Cheng ed., 1982); see also Bodansky, supra note 23, at 111 (describing an emerging recognition that "there is a divergence between the traditional theory of customary law, which emphasizes consistent and uniform state practice, and the norms generally espoused as 'customary'").

(104.) They back away, for example, from characterizing all such representations as "casuistry," but then describe it as "cheap talk," Goldsmith & Posner II, supra note 5, at 663-64, meaning that the cost of making the representation is zero. Goldsmith & Posner I, supra note 14, at 1136 n.50.

(105.) The articulation of that standard in preparing the statute for the Permanent Court was surely conspicuous. See WOLFKE, supra note 20, at 1-5 (discussing the ambiguity behind the international custom standard). For a description of the process by which the international community created the ICJ, see generally Manley O. Hudson, The Succession of the International Court of Justice to the Permanent Court of International Justice, 51 AM. J. INT'L L. 569 (1951). Its transparency in any particular application, concededly, may have left something to be desired. U.S. Secretary of State John Foster Dulles, for example, once described the international law of satellites as "quite obscure," a Russia-like "mystery wrapped in an enigma," and suggested that he would be unable to state definitively the law in the absence of codification. John F. Dulles, U.S. Secretary of State, News Conference (Feb. 7, 1956), in 34 DEP'T ST. BULL. 280-81 (1956).

(106.) The Commission is a United Nations entity, like the ICJ, in which members function in their individual capacity rather than as representatives of their national governments. Nonetheless, its statutes ensure that "[g]overnments have an important role at every stage" of the Commission's work. UNITED NATIONS, THE WORK OF THE INTERNATIONAL LAW COMMISSION at 21, U.N. Sales No. E.95.V.6 (5th ed. 1996).

(107.) For a thorough examination of this point, framed in terms highly amenable to rational choice analysis, see Arthur W. Rovine, The National Interest and the World Court, in 1 THE FUTURE OF THE INTERNATIONAL COURT OF JUSTICE 313, 317 (Leo Gross ed., 1976).

(108.) This tendency is especially evident in their most recent work. Goldsmith & Posner IV, supra note 14, passim (adverting to the function of "moral and legal rhetoric" in international relations); id. at S133-36 (addressing why nations employ moral and legal rhetoric). The exception is their explanation as to why states might advert to customary international law, rather than merely "custom": "The word 'law' makes explicit the obligatory content of the practice. Appeal to the law is a way of saying that past actions provide evidence of future intentions, that we have done well by acting consistently with them, so you should not deviate." Id. at S137. It is unclear why the same function could not be explained by calling the custom "moral," or appending the words "important" or "obligatory." But as explained in Part III, this aspect of their account--more so than the more detailed and critical version presented in their earlier work--is, if properly elaborated, essentially consistent with the opportunities presented by customary international law as it stands. Their resistance to this claim appears based in large part on an arguably exaggerated regard for the standard version of the opinio juris requirement. See id. at S137 ("The notion that law necessarily implies that the parties have submitted to an outside authority, real or metaphysical or moral, is a modern confusion."). Compare id., with infra notes 114-23 and accompanying text (noting latitude in standard version and practice), and infra Part III.A.3 (positing alternative principles). They also suggest that any complex interactions must be left to diplomacy or treaties, and may be read to imply that custom is limited to bilateral relations or their near equivalent, see Goldsmith & Posner IV, supra note 14, at S137-38 (asserting that diplomacy and treaties facilitate coordination better than custom in complex situations because of the discussion and codification process), which would continue to distinguish their position. A final point of difference would be their skeptical position regarding the potential for investing value in custom-specific reputation. See infra Part III.B.

(109.) See ELSTER, supra note 78, at 101 (suggesting a distinction between social and legal norms in part on the ground that "obedience to the law is often rational on purely outcome-oriented grounds," whereas social norms are by definition not outcome-oriented); H.L.A. HART, THE CONCEPT OF LAW 225-26 (1961) (distinguishing international law from moral rules due, in part, to the role of self-interest in the former). This does not. of course, exhaust the differences between legal and moral claims. See HART, supra, at 222-25 (stressing the difference in state practices between moral and legal reproaches, the moral indifference of many legal rules, and the amenability of international law to change by common consent).

(110.) See STEPHEN D. KRASNER, SOVEREIGNTY: ORGANIZED HYPOCRISY 14-20, 228-38 (1999) (arguing that international legal sovereignty is a construct subject to violation and manipulation).

(111.) See Russell B. Korobkin & Thomas S. Ulen, Law and Behavioral Science: Removing the Rationality Assumption from Law and Economics, 88 CAL. L. REV. 1051, 1060-75 (2000) (describing critically the "thick" and "thin" versions of rational choice analysis as applied to individuals, and noting implications of each for validating theory). Compare GREEN & SHAPIRO, supra note 13, at 17-19 (noting discrepant versions, but recognizing that much rational choice literature depends on a "thick" conception), and Edward L. Rubin, Public Choice, Phenomenology, and the Meaning of the Modern State: Keep the Bathwater, but Throw out That Baby, 87 CORNELL L. REV. 309, 318-19 (2002) (making a similar point with respect to legal scholarship on the public choice theory), with HOVI, supra note 7, at 4-6 (noting that rational choice models do not require that actors be motivated exclusively by self-interest), and TSEBELIS, supra note 49, at 21 (noting that goals amenable to rational choice analysis "may be egoistic or altruistic, idealistic or materialistic"), and id. at 26 n. 11 (recognizing that, in theory, rationality does not require the pursuit of maximum utility).

(112.) For an example, see FRITZ W. SCHARPF, GAMES REAL ACTORS PLAY: ACTOR-CENTERED INSTITUTIONALISM IN POLICY RESEARCH (1997); compare Lisa L. Martin & Beth A. Simmons, Theories and Empirical Studies of International Institutions, 52 INT'L ORG. 729, 742-43 (1998) (anticipating theories that involve states electing, and then abiding by, institutions). Of course, at some point, difficult to define, taking account of institutions becomes institutionalism (which is essentially compatible with rational choice approaches) and then constructivism (which is, generally speaking, incompatible, because it challenges the individualism assumed by rational choice).

(113.) See Goldsmith & Posner III, supra note 14, at 200 (arguing that the "additional complexity [of particular evolutionary game theory models] does not, in our view, make international behavior any easier to understand, and thus parsimony counsels against it").

(114.) E.g., RESTATEMENT (THIRD), supra note 9, pt. 7 introductory note at 149 (contending that "international law generally is largely observed because violations directly affect the interests of states, which are alert to deter, prevent, or respond to violations"); Phillip R. Trimble, International Law, World Order, and Critical Legal Studies, 42 STAN. L. REV. 811, 833 (1990) (asserting that a state "may decide to forgo the short-term advantages derived from violating those rules because it has an overriding interest in maintaining the overall system"). Both examples are cited in Robert O. Keohane, International Relations and International Law: Two Optics, 38 HARV. INT'L L.J. 487, 490 (1997). See also MARK E. VILLIGER, CUSTOMARY INTERNATIONAL LAW AND TREATIES: A MANUAL ON THE THEORY AND PRACTICE OF THE INTERRELATION OF SOURCES 285 (2d ed. 1997) (defending compatibility of interest with custom); Byers, supra note 15, at 202 (observing that "[m]ost international lawyers have long accepted that states are not only the subjects but also the creators of international law," that international law results from "coordinated or at least (in large part) common behavior, and that rules of international law therefore reflect the long term interests of most, if not all, states").
      As one commentator elaborated:

   States' acts in international relations are in all cases rather well
   calculated and conscious (this does not mean that states do not
   miscalculate or that their acts always lead to desired results). And
   in most cases states are also conscious that their behaviour may
   contribute to the formation of customary rules which will become
   binding on them (otherwise they would not pay so much attention to
   disclaimers and protests). Often they will that a pattern of their
   behaviour become a general rule.... States seldom have as the primary
   motive of their behaviour in international relations the formation of
   rules of customary international law.

Rein Mullerson, The Interplay of Objective and Subjective Elements in Customary Law, in INTERNATIONAL LAW: THEORY AND PRACTICE 161, 164 (Karel Wellens ed., 1998).

(115.) E.g., RESTATEMENT (THIRD), supra note 9, [section] 103(2)(d) (citing as evidence of custom "pronouncements by states that undertake to state a rule of international law"); PETER MALANCZUK, AKEHURST'S MODERN INTRODUCTION TO INTERNATIONAL LAW 45 (7th rev. ed. 1997) (suggesting that while "[o]pinio juris is sometimes interpreted to mean that states must believe that something is already law before it can become law," that "is probably not true, since "what matters is not what states believe, but what they say"; "[i]f some states claim that something is law and other states do not challenge that claim, a new rule will come into being, even though all the states concerned may realize that it is a departure from pre-existing rules"); Hugh Thirlway, The Law and Procedure of the International Court of Justice 1960-1989 (Part Two), 1990 BRIT. Y.B. INT'L L. 1, 43 (arguing that "at the initial stage of the development of the custom, it is sufficient that the States concerned regard the practice as what the Court ... referred to as 'potentially norm-creating', as conforming to a rule which either already exists or is a useful and desirable rule which should exist" (quoting North Sea Continental Shelf (F.R.G.v. Den.; F.R.G.v. Neth.), 1969 I.C.J. 3, 42 (Feb. 20))); see supra notes 27-31 and accompanying text.

(116.) See, for example, the classic formulation by Judge Hudson requiring the "conception that the practice is required by, or consistent with, prevailing international law." Hudson, supra note 22, at 26 (emphasis added). To be sure, it is not always clear whether a rule is permissive or negative in character, and sometimes one involves the other. See Akehurst, supra note 24, at 37-38 (insisting on the distinction, but noting nuances).

(117.) See supra note 24 (quoting North Sea Continental Shelf, 1969 I.C.J. at 43); see also Military and Paramilitary Activities in and Against Nicaragua (Nicar. v. U.S.), 1986 I.C.J. 14, 108-09 (June 27) (citing North Sea Continental Shelf).

(118.) North Sea Continental Shelf, 1969 I.C.J. at 43 (emphasis added).

(119.) Compare Fisheries Jurisdiction (U.K. v. Ice.), 1974 I.C.J. 1, 23 (July 25) (describing statements in preparatory work as "manifestations of the views and opinions of individual States and as vehicles of their aspirations, rather than as expressing principles of existing law"), with id. at 48 (opinion of Forster, Bengzon, de Arechaga, Singh, and Ruda, JJ.) (concluding that, in light of the uncertain prior state of customary international law and contradictory, imprecise practices, "such declarations and statements and the written proposals submitted by representatives of States are of significance to determine the views of those States as to the law on fisheries jurisdiction and the opinio juris on a subject regulated by customary law"). The majority's opinion, moreover, elsewhere regarded state approval of proposed legislation as providing "overwhelming support" for an emerging priority for coastal state fishing rights, notwithstanding the impossibility of claiming that such approval was legally compelled. Id. at 26; see generally 1 D.P. O'CONNELL, THE INTERNATIONAL LAW OF THE SEA 33-34 (1982) (discussing Fisheries Jurisdiction).

(120.) Military and Paramilitary Activities, 1986 I.C.J. at 109.

(121.) Asylum Case (Colom. v. Peru), 1950 I.C.J. 266, 276 (Nov. 20). That case was particularly concerned with establishing a "regional or local custom peculiar to Latin-American States," rather than a universal custom, but its inquiry was premised on a more general understanding of Article 38 of the Court's Statute. Id.

(122.) See FINAL REPORT, supra note 36, [section] 1(i) (providing a "working definition" of custom). As noted previously, though, the Final Report goes on to suggest that opinio juris may wholly be disregarded under certain circumstances. See supra notes 41-44 and accompanying text.

(123.) RESTATEMENT (THIRD), supra note 9, [section] 102 cmt. c.

(124.) One such attempt, as previously noted, would be to relate the Court's articulation of this requirement to a principle of consent. See WOLFKE, supra note 20, at 13-29 (describing cases). But see supra note 24 (noting contrary views).

(125.) It would be still easier to reimagine the patrol example they use to illustrate the coordination game, where bordering states are indifferent between two means used to operationally divide their territory--either permitting patrolling up to a river between them, or up to a road that divides them. Goldsmith & Posner I, supra note 14, at 1128. Each state might in fact prefer a different effective border (say, because one is better girded against amphibious attack), while still preferring any mutual solution to different approaches.

(126.) Cf. George W. Downs, Enforcement and the Evolution of Cooperation, 19 MICH. J. INT'L L. 319, 322 n.8 (1998):
   [In contrast to standard theory,] [o]ne could argue that enforcement
   might still play a role in some coordination games in order to lend
   greater stability to a particular pattern of distribution. If the
   relative power of one or more States increased, the party that gained
   in power might violate the agreement in order to renegotiate its
   terms and thus claim a larger distributional benefit.

(127.) In the hypothetical, for example, a state with vessels that profit particularly from whaling--putting to one side any international law on that subject--might have an active interest in persuading others away from the "disarm" norm.

(128.) This is a familiar problem in the literature concerning standards and networks. For U.S. consumers, the obvious example involves the mobile telephony sector, where several systems have emerged without achieving the unity of the Global System for Mobile Communications (GSM) services in Europe. See, e.g., Andy Dornan, Waiting for Wireless in the United States, NETWORK MAG., Dec. 1, 2001, at 54.

(129.) This will not, presumably, plague those states which have the clearest distributional advantages at stake, and which are the most likely to commit early as a consequence.

(130.) Cf. Martin, supra note 60, at 101-03 (discussing multilateral solutions to coordination games).

(131.) Goldsmith & Posner I, supra note 14, at 1158 n.134 (citing authorities).

(132.) Professors Goldsmith and Posner recognize, of course, this division of authority. See Goldsmith & Posner I, supra note 14, at 1161 n.148.

(133.) See, e.g., RESEARCH IN INT'L LAW, HARVARD LAW SCH., DRAFT CONVENTION ON TERRITORIAL WATERS, art. 2, in 23 AM. J. INT'L L. 242, 243 (Supp. 1929) (proposing terms for an international convention reflecting, in proposed Article 2, the three-mile limit); SAYRE A. SWARZTRAUBER, THE THREE-MILE LIMIT OF TERRITORIAL SEAS 130 (1972) (suggesting that "the greatest years for the three-mile limit were those from 1876 to 1926.... If domestic legislation, international instruments, court decisions, and the writings of publicists are a fair measure, then by 1926, the three-mile limit was in every sense a rule of international law"); Thomas Baty, The Three-Mile Limit, 22 AM. J. INT'L L. 503, 504 (1928) ("The three-mile limit rule has ... in modern times been as little infringed in practice and as little contradicted by practical diplomatists as any rule of law."); Bernard G. Heinzen, The Three-Mile Limit: Preserving the Freedom of the Seas, 11 STAN. L. REV. 597, 634 (1959) ("Since the three-mile or one-league limit had the consent of virtually all coastal states claiming a territorial sea, it may therefore be said that it was, at the turn of the [twentieth] century, generally accepted as a customary rule of international law.").

I should stress that a number of commentators disagreed. See, e.g., CHRISTOPHER B.V. MEYER, THE EXTENT OF JURISDICTION IN COASTAL WATERS 512-20 (1937) (summarizing state practices permitting more extensive Norwegian claims); STEFAN A. RIESENFELD, PROTECTION OF COASTAL FISHERIES UNDER INTERNATIONAL LAW 29-98 (1942) (cataloging conflicting views of commentators from 1800-1942); id. at 279-80 (illustrating the wide range of scholars' opinions regarding the territorial sea when polled between 1800 and 1899 and again after 1900); Manley O. Hudson, The First Conference for the Codification of International Law, 24 AM. J. INT'L L. 447, 457 (1930) ("The history of the last century has failed to invest the 'three-mile limit' with any particular sanctity, and recent conquests of distance make it seem in many respects archaic."). As O'Connell observes:
   [D]uring the critical period from 1876 to 1914, thirty-three
   jurists believed that the territorial sea expanded with the evolving
   range of artillery; twenty-six believed that State practice had
   established it at three miles; five proposed other fixed limits; five
   argued for different limits for different purposes; eight ambiguously
   referred to both the three-mile limit and the cannon-shot; and seven
   thought that there was no consensus on the matter.... The principal
   conclusion that emerges from this review is that, on the outbreak of
   World War I, which was to demonstrate the effect of developing
   technology upon the criterion for delimiting belligerent and neutral
   waters, the three-mile rule was a minority opinion among jurists.

1 O'CONNELL, supra note 119, at 153-54 (citations omitted).

(134.) Most conclude that the three-mile limit's own limits were reached by the 1930 Hague Conference for the Codification of International Law, which discussed but failed to adopt the three-mile margin, and later events surely confirmed that the margin had expanded. See, e.g., Henry M. Arrudo, Comment, The Extension of the United States Territorial Sea: Reasons and Effects, 4 CONN. J. INT'L L. 697, 701-02 (1989) (concluding that, in addition to the Hague Conference of 1930 and U.S. implementation of "different-purpose jurisdictional limits beyond a three-mile limit," the failure of the first and second United Nations Conferences on the Law of the Sea (UNCLOS I and II) and the codification of the twelve-mile limit at the third U.N. Conference (UNCLOS III) "put an end to the three-mile limit as a rule of international law"); Robert Jay Wilder, The Three-Mile Territorial Sea: Its Origins and Implications for Contemporary Offshore Federalism, 32 VA. J. INT'L L. 681,685 (1992) (noting that "[i]n 1945, a total of forty-six nations"--"a full 77% of the total number of coastal or island nations"--claimed a three-mile limit, but "[b]y the early part of 1989, only ten nations (or 14%)" did (citing Annotated Supplement to the Commander's Handbook on the Law of Naval Operations, NWP 9 (Rev. A)/FMFM 1-10, at tbl. ST1-6 (October 1989))).

Some further assert that the failure of the Hague Conference, among other international fora, proved that the three-mile limit had never existed. See, e.g., BYERS, supra note 24, at 11420 (describing the subsequent defection of scholars from support for the three-mile limit, and apparently regarding that as conclusive); RIESENFELD, supra note 133, at 280 (concluding on the basis of state practice that "it is clear that, as the Hague Codification Conference made obvious to all the world, there is no such thing as a universally recognized three-mile rule"); accord Joseph Walter Bingham, The Continental Shelf and the Marginal Belt, 40 AM. J. INT'L L. 173, 174 (1946); Goldsmith & Posner I, supra note 14, at 1160-61 n. 147. But that reasoning seems dubious. The failure to ratify the rule during attempts at codification, in which states may feel freer to opt for a new norm or to seek exemptions for themselves, does little to contraindicate the existence of a background rule of customary international law. See Heinzen, supra note 133, at 639 (arguing that the 1930 Hague Conference did not necessarily belie the three-mile limit); Jesse S. Reeves, The Codification of the Law of Territorial Waters, 24 AM. J. INT'L L. 486 487-99 (1930) (describing features of the conference that frustrated agreement). Indeed, had the Hague Conference been successful, it might have been read as supplanting customary international law altogether. North Sea Continental Shelf (F.R.G.v. Den.; F.R.G.v. Neth.), 1969 I.C.J. 3, 43 (Feb. 20) (rejecting compliance examples on the ground that "over half the States concerned, whether acting unilaterally or conjointly, were or shortly became parties to the Geneva Convention, and were therefore presumably, so far as they were concerned, acting actually or potentially in the application of the Convention," so that "[f]rom their action no inference could legitimately be drawn as to the existence of a rule of customary international law in favour of the equidistance principle").

(135.) See, e.g., 1 O'CONNELL, supra note 119, at 124-25 (noting proposed limits of sixty miles, two days' sailing distance, the range of vision, and one hundred miles); H.S.K. Kent, The Historical Origins of the Three-Mile Limit, 48 AM. J. INT'L L. 537, 537 (1954) (describing the origin of the three-mile limit)

(136.) Secretary of State Thomas Jefferson's seminal letter to his British counterpart, in which he blithely equated the cannon-shot and three-mile limits, is worth quoting at length:
   [B]efore it shall be finally decided to what distance from our
   seashores the territorial protection of the United States shall be
   exercised, it will be proper to enter into friendly conferences and
   explanations with the powers chiefly interested in the navigation
   of the seas on our coasts, and relying that convenient occasions may
   be taken for these hereafter, finds it necessary in the meantime to
   fix provisionally on some distance for the present government of
   these questions. You are sensible that very different opinions and
   claims have been heretofore advanced on this subject. The greatest
   distance to which any respectable assent among nations has been at
   any time given, has been the extent of the human sight, estimated at
   upwards of twenty miles, and the smallest distance, I believe,
   claimed by any nation whatever is the utmost range of a cannon ball,
   usually stated at one sea league. Some intermediate distances have
   also been insisted on, and that of three sea leagues has some
   authority in its favor. The character of our coast, remarkable in
   considerable parts of it for admitting no vessels of size to pass
   near the shores, would entitle us, in reason, to as broad a margin of
   protected navigation as any nation whatever. Reserving, however, the
   ultimate extent of this for future deliberation, the President gives
   instructions to the officers acting under his authority to consider
   those heretofore given them as restrained for the present to the
   distance of one sea league or three geographical miles from the
   seashores. This distance can admit of no opposition, as it is
   recognized by treaties between some of the powers with whom we are
   connected in commerce and navigation, and is as little, or less,
   than is claimed by any of them on their own coasts.

Letter from Thomas Jefferson, U.S. Secretary of State, to George Hammond, British Minister (Nov. 8, 1793), reprinted in 1 JOHN BASSET MOORE, A DIGEST OF INTERNATIONAL LAW 702-03 (1906); see also RIESENFELD, supra note 133, at 137 n.33 (noting the provisional nature of U.S. assertions).

(137.) For a thorough, if partisan, review of the evidence, see Baty, supra note 133, at 517-32; C.J. Colombos, Territorial Waters, 9 GROTIUS SOC'Y 89, 96-97 (1924) (asserting that demands for the extension of maritime zones beyond three miles "have been always resisted").

(138.) See, e.g., Baty, supra note 133, at 509-15, 532-37 (reviewing purported discrepancies). It should be mentioned that those nations regarded as favoring broader territorial waters were at least as inconsistent. Professors Goldsmith and Posner cite the Scandinavian countries, Spain, Portugal, and Russia as leading examples, see Goldsmith & Posner I, supra note 14, at 1159-60, but those nations' opposition to the three-mile limit was not unblemished. Russia is their best example: it appears to have been relatively aggressive in its assertion of broader limits, and though it frequently capitulated, it did so under circumstances of duress arguably consistent with Professor Goldsmith and Professor Posner's theory. See PHILIP C. JESSUP, THE LAW OF TERRITORIAL WATERS AND MARITIME JURISDICTION 26-31 (1927) ("The striking thing about the Russian case is the fact that whenever she has endeavored to deny the three-mile limit as a rule of international law, she has been forced to withdraw from her position."); RIESENFELD, supra note 133, at 194-203 (describing settlements between Russia and various nations over territorial seas limits); Baty, supra note 133, at 513, 520-21,523, 535-36 (describing instances in which Russia abandoned claims or acquiesced in foreign rights up to three miles from shore). But while Scandinavian exceptionalism was prominent in the eighteenth century, see Heinzen, supra note 133, at 605-12, and maintained by Sweden and Norway well afterward, see RIESENFELD, supra note 133, at 188-94; Baty, supra note 133, at 511, it was betrayed by Denmark, does not appear to have had practical application against foreign powers, and was premised in special local historical circumstances rather than as any more general objection to the three-mile limit. Baty, supra note 133, at 511; see also JESSUP, supra, at 31-41, 63 (explaining that even were Scandinavian assertions uncontested and consistent, they "could be admitted without destroying the three-mile rule as a principle of international law, since they are set up, not in opposition, but as exceptions to that rule"); Kaare Bangert, Denmark and the Law of the Sea, in THE LAW OF THE SEA: THE EUROPEAN UNION AND ITS MEMBER STATES 97, 102 (Tulio Treves ed., 1997) (describing contemporary Danish practice). Spain and Portugal likewise acquiesced in foreign challenges to their assertion of a six-mile limit. See JESSUP, supra, at 41-43 ("Spain must be grouped with those other powers like Portugal and Russia, which are desirious of extending the three-mile zone but are unable to do so."); Baty, supra note 133, at 511 ("The Spanish and Portuguese claim to a six-mile limit was successfully challenged by France and Great Britain, and the Peninsular states withdrew their pretensions."). But see RIESENFELD, supra note 133, at 175-80 (regarding Spanish and Portuguese assertions as more resolute).

(139.) See Goldsmith & Posner I, supra note 14, at 1161 (noting that "throughout the period many nations enforced antismuggling and related security laws outside the three-mile band").

(140.) Id. at 1161-65.

(141.) See JESSUP, supra note 138, at 10-18, 49-60 (recounting the histories of British and U.S. territorial waters claims); Baty, supra note 133, passim (same). For a valuable collection of British and U.S. official documents articulating (and attempting to rationalize) their respective positions concerning the extent of territorial waters, see THE EXTENT OF THE MARGINAL SEA 53395 (Henry G. Crocker ed., 1919) (British documents); id. at 630-95 (U.S. documents). There were certainly occasions on which the limit was acknowledged at a cost. In the late nineteenth century, for example, the United States seized several British ships engaged in sealing in the Bering Sea at a distance of more than three miles from the Alaska Territory. The issue was sufficiently serious that the parties agreed, by treaty, to submit the controversy to arbitration (though not so serious, as was conceded in the arbitration, as to risk war). William Williams, Reminiscences of the Bering Sea Arbitration, 37 AM. J. INT'L L. 562, 582 (1943); see also Editorial Comment, The Fur Seal Question, 1 AM. J. INT'L L. 742, 742-48 (1907) (noting subsequent, related diplomatic controversies). Though it would have been to the U.S. advantage in that controversy to take an aggressive stance in favor of expanded territorial limits, it instead acquiesced in posing the question to the arbitrators of whether it had a right to protect the seals (or an ownership right in them) "when such seals are found outside the ordinary three-mile limit"--a submission that hamstrung their prospects for success. JESSUP, supra note 138, at 54-55; RIESENFELD, supra note 133, at 268-71. But cf. MEYER, supra note 133. at 307-08 (quoting U.S. arguments for a right to self-defense beyond the three-mile limit).

(142.) See supra notes 114-23 and accompanying text. Professors Goldsmith and Posner also suggest that enforcement was more consistent with power relationships than with the rule of law: some nations asserted territorial jurisdiction beyond three miles, but they also tempered those assertions in "in the face of threats of retaliation" by countries like England and the United States. Goldsmith & Posner I, supra note 14, at 1160 n.145. Though there were cases in which maritime powers threatened the use of force--for example, to protect English shipping outside Spain's three-mile limit--and other cases in which threats must have been implicit, it was common for nations simply to protest excessive assertions of territory, in manners that appear from secondary accounts to be more like legally premised diplomacy than coercion. See id. at 1160 n.145 (citing RIESENFELD, supra note 133, at 144-46, and Heinzen, supra note 133, at 630); Baty, supra note 133, at 511,513, 522-23, 525-26, 528 (citing assertions of broader territorial seas and successful, apparently noncoercive, protests).

(143.) See supra notes 133-34, 137-38 and accompanying text.

(144.) See Goldsmith & Posner I, supra note 14, at 1166 n.168.

(145.) Id. at 1165-67. In their view, nations with such strong navies, and interested in maximizing areas in which their nationals could fish, should have desired "the narrowest possible territorial sea," and should have been able to succeed in obtaining it. Id. at 1165; see also MARK W. JANIS, SEA POWER AND THE LAW OF THE SEA 39 (1976) ("Large expanses of high seas served British naval missions well. These missions were to protect Britain's far-flung colonies and trade routes and to project British power ashore throughout Europe and the world.").

(146.) Some discussion tends to focus on fishing rights, for example, at the expense of other germinal issues like neutrality. See, e.g., Heinzen, supra note 133, at 614-19 (emphasizing neutrality issues as influencing development of the three-mile limit). And neutrality naturally meant one thing to an infant America, eager to avoid being entangled in the conflict among the European powers, e.g., JESSUP, supra note 138, at 49-60, and something different to a maritime giant like England, see, e.g., JANIS, supra note 145, at 39 ("Traditional law of the sea was largely the 19th century creation of British sea power," which was "[u]nrivalled for most of the period 1815 to 1914"). Both potential distinctions were illustrated in British consideration (and, ultimately, rejection) of U.S. proposals to extend the three-mile limit to five miles, following British protests of an episode in which U.S. ships stationed outside British neutral waters fired on Confederate ships within those waters. British internal deliberations noted the varied applications and purposes of the three-mile limit, with Lord Palmerston explaining, in opposition to the U.S. proposal, that: "The whole discussion turns upon the assumption that England is generally to be Neutral, but England is more likely to be Belligerent than Neutral and it is evident that any extension of territorial limits from Coasts would be more unfavorable than advantageous to England as a Belligerent." RIESENFELD, supra note 133, at 148 (quoting memorandum of Nov. 28, 1864): see also D.P. O'CONNELL, THE INFLUENCE OF LAW ON SEA POWER 24-25 (1975) (noting the evolution of British interests and policy). Nothing in this account, of course, is inconsistent with a realist's skeptical view of the three-mile "rule," but makes it difficult to conclude that the observed practice can be explained as a mere coincidence of interests among powerful states.

(147.) It seems equally plausible that powerful states would assert a very large territorial sea for themselves, and the narrowest possible such sea for others--dispensing, essentially, with any common understanding, perhaps on the premise that every coastline is different. Cf. Philip Marshall Brown, The Marginal Sea, 17 AM. J. INT'L L. 89, 90-91 (1923) (noting the substantial variation in the physical circumstances of littoral states as an obstacle to any three-mile limit). The question, in other words, is not merely why a narrow-three mile band was elected (as opposed to an even narrower margin), but why powerful nations presupposed that any common margin was necessary. Cf. Edwin Borchard, Resources of the Continental Shelf, 40 AM. J. INT'L L. 53, 61 (1946) (asserting that since the United States "is free to adopt any rule that suits its interests best, and it is doubted whether foreigners could or would raise serious complaint," the United States should not continue to adhere to a three-mile limit that disserves its interests).

(148.) See infra Part III.A.3 (discussing the consistency of coordination games with the traditional requirement of opinio juris); infra Part III.B (discussing reputation issues in greater depth).

(149.) See BAIRD ET AL., supra note 2, at 35-37 (discussing the Stag Hunt game).

(150.) I structure the payoffs here so as to assume no competition in pursuing the inferior course (here, parading). If that assumption were incorrect--if there were, for example, a shortage of floats or uniforms, so that the decision by a second state to parade diminished the payoff for the first--the payoffs in the lower right quadrant would be reduced.

(151.) Martin, supra note 60, at 106-07 (citing Arthur A. Stein, Coordination and Collaboration: Regimes in an Anarchic World, in INTERNATIONAL REGIMES 115, 119 (Stephen D. Krasner ed., 1983)).

(152.) Specifically, if P is the probability that State J will place observers, and the payoffs reflect the values to State I of conducting a parade ([Payoff.sub.parade]), placing observers when State J does as well ([]), and placing observers when State J does not do so ([Payoff.sub.solo]), where [Payoff.sub.parade]=([P.sub.o] * [Payoff.sub.solo]) + ((1-[P.sub.o])* [Payoff.sub.o]). Under the specified payoffs, the probability would be 3/5: that is, if State J had a 3/5 probability of pursuing "observers," State I will be indifferent among its options; a greater probability will tip State I toward a preferred strategy of following "observers" itself, and a lesser probability will lead it to "parade." For a clearer explanation, see BAIRD ET AL., supra note 2, at 37-39 (discussing examples of mixed strategy).

(153.) Id. at 39-41 (illustrating and defining focal points).

(154.) See EDNA ULLMANN-MARGALIT, THE EMERGENCE OF NORMS 123 (1977) (explaining that in the Stag Hunt case, "joint co-operation is stable pending not just anyone's actual desertion but pending anyone's belief, or even suspicion that someone will desert").

(155.) Proclamation No. 2667, 3 C.F.R. 67, 67-68 (1943-1948) reprinted in 59 Stat. 884, 884-85 (1945). Some questioned whether the disassociation of the continental shelf from the law governing the high seas was purely formal. LAWRENCE JUDA, INTERNATIONAL LAW AND OCEAN USE MANAGEMENT 99 (1997) (describing debates within the International Law Commission).

(156.) Proclamation No. 266T 3 C.F.R. at 67-68, reprinted in 59 Stat. at 884-85; see also Herman Phleger, Legal Adviser, U.S. Department of State, Address Before the American Branch of the International Law Association (May 13, 1955), in 4 DIGEST OF INTERNATIONAL LAW 761 (Marjorie M. Whiteman ed., 1965) (explaining why the exercise of jurisdiction over the natural resources and seabed of the continental shelf is reasonable and just).

(157.) Proclamation No. 2667, 3 C.F.R. at 67-68, reprinted in 59 Stat. at 884-85.

(158.) Id. The generalizable nature of the U.S. claim--presented as one for all coastal states, most with shelves, regardless of their immediate abilities to exploit them--distinguished its law-forming character and played a significant role in attracting support. BYERS, supra note 24, at 91-92; ZDENEK J. SLOUKA, INTERNATIONAL CUSTOM AND THE CONTINENTAL SHELF 74-76 (1968). But see. e.g., C.H.M. Waldock, The Legal Basis of Claims to the Continental Shelf, 36 GROTIUS SOC'Y 115, 138-39 (1950) (arguing that contrasts between the shelf proclamation and the fishery proclamation of the same date suggest that the former was less in the form of international legislation, and that "the United States did not consider recognition by other States as in any way necessary to give legal propriety to its claim to the resources of the shelf").

(159.) See Press Release, White House (Sept. 28, 1945), reprinted in 4 DIGEST OF INTERNATIONAL LAW, supra note 156, at 758 (elaborating on the policy proclaimed by the president concerning jurisdiction over the continental shelf).

(160.) See JUDA, supra note 155, at 95-96 (describing memoranda authored by President Roosevelt and Interior Secretary Harold Ickes indicating an intent to form new principles of international law); M.L. Jewett, The Evolution of the Legal Regime of the Continental Shelf, 22 CAN. Y.B. INT'L L. 153, 158 (1984) (construing State Department deliberations as indicating that "the [Truman] Proclamation was designed to achieve a new acquisition rather than to assert a pre-existing right"); Phleger, supra note 156, at 764 (explaining that the United States' claims are not intended to modify the freedom of the superadjacent waters).

(161.) See 1 O'CONNELL, supra note 119, at 467 (claiming that "the continental shelf doctrine when it was enunciated in the late 1940s was a novelty without juristic antecedents"); id. at 467-75 (describing rationalizations in commentary). As Professor O'Connell also emphasizes, though, the theory was not one unilaterally imposed by the United States, which instead sought to ensure its acceptance through "extensive diplomatic overtures." Id. at 31-32; see also Richard Young, Recent Developments with Respect to the Continental Shelf, 42 AM. J. INT'L L. 849, 851 (1948) (noting that the United States first consulted with the states most likely to be directly affected by the U.S. claim, and that they did not appear to have objected).

(162.) Sir Francis Vallat, a legal adviser in the United Kingdom's Foreign Office, explained that "while the unilateral declaration of the United States cannot in itself create any new rights or any new rules of international law, it may be regarded as providing the seed from which such rights and rules may grow." Francis A. Vallat, The Continental Shelf, 1946 BRIT. Y.B. INT'L L. 333,337; see BYERS, supra note 24, at 92 (describing the Truman Proclamation as "a classic example of a conscious, successful effort to develop a new customary rule"); SLOUKA, supra note 158, at 20-22, 25, 27, 43, 74-83 (describing the growing acceptance of Truman's Presidential Proclamation as the basis for the rise of the doctrine of the continental shelf); Josef L. Kunz, Continental Shelf and International Law: Confusion and Abuse, 50 AM. J. INT'L L. 828, 829-30 (1956) (noting controversy over whether theory of continental shelf "inaugurated" by President Truman had already become custom, and stating his personal view that "a new norm of international law has not yet come into existence, although we are witnessing the formation of such new norm"); Arvid Pardo, The Future of the Sea, in THE FUTURE OF THE LAW OF THE SEA 1, 2 (L.J. Bouchez & L. Kaijen eds., 1973) (describing the Truman Proclamation as "the starting point of contemporary developments in the law of the sea").

(163.) North Sea Continental Shelf (F.R.G. v. Den.; F.R.G. v. Neth.), 1969 I.C.J. 3, 32-33 (Feb. 20).

(164.) Id. at 53; see also id. at 39 (explaining that some of the Convention's terms "were then regarded as reflecting, or as crystallizing, received or at least emergent rules of customary international law"); Ocean Drilling & Exploration Co. v. Oda, 27 JAPANESE ANN. OF INT'L L. 148, 164 (D. Tokyo, 1982) (describing the Proclamation as having "awakened the states to the possibility of a fight over the resources beneath the seabed of the high seas and initiated the formulation of international law with regard to the continental shelf"); JUDA, supra note 155, at 97 (noting that while there were substantial variations among postproclamation claims, "It]here was complete agreement that the continental shelf, however defined, did not constitute an area res nullius .... [n]or ... res communis," but instead was controlled by the contiguous coastal state with or without effective occupation); Young, supra note 161, at 849 (describing subsequent practices of states following the U.S. lead, albeit with variations, as demonstrating that the Proclamation "proves to have offered a marketable concept in the marts of international law").

(165.) See 1 THEODORE O. KRONMILLER, THE LAWFULNESS OF DEEP SEABED MINING 3-5, 107 (1980) (describing and evaluating the competing principles of res nullius and res communis under traditional customary international law); Jewett, supra note 160, at 170-76 (same).

(166.) See SLOUKA, supra note 158, at 78-79 (illustrating the application of alternative principles to the continental shelf). But see Waldock, supra note 158, at 139 (characterizing the U.S. position as a mere variant of res nullius).

(167.) As Professor Slouka explains, these were not hypothetical considerations, though there is room to dispute how realistic the beggar-thy-neighbor and security concerns were. SLOUKA, supra note 158, at 32-35, 74-83. A related concern, obviously, was that private exploitation of the continental shelf required greater stability than these regimes might afford. See JUDA, supra note 155, at 96 ("An obvious and perhaps implicit motivation [for the Truman Proclamation] is the need for a legal regime which provided the legal stability and assurances that would be necessary to attract investors to undertake offshore operations.").

(168.) These interests were, it bears reiterating, synonymous with its own. See JUDA, supra note 155, at 96 ("The Truman Proclamation on the Continental 'Shelf may be seen as a prime example of a carefully tailored and self-serving legal claim. The history of its evolution demonstrates how a legal doctrine is shaped to serve the underlying political interests of the state putting it forward.").

(169.) E.g., Bingham, supra note 134, at 177-78 (explaining that there was no international established practice contrary to the proclaimed policy). Were this more akin to a Battle of the Sexes situation, recall, one would anticipate the emergence of a rival focal point. With respect to the continental shelf issue, the far more expansive claims of certain South American nations were regarded as marginal, see SHAW, supra note 24, at 433 (describing rival assertions by Argentina, E1 Salvador, and Chile to far broader shelves, and including waters as well), and attempts by Great Britain to establish occupational rights through bilateral agreements apparently stopped well short of establishing any continuing, universal alternative to the regime advocated by the Truman Proclamation. 1 O'CONNELL, supra note 119, at 471-75; SLOUKA, supra note 158, at 71-73, 85 n.39.

(170.) In their later work, Professors Goldsmith and Posner suggest that the Truman Proclamation arguably exemplifies a situation in which "It]he rhetoric of CIL can convey meaningful information about focal points when nations face a coordination problem," but that this does not "depend on the normative gravitational pull of CIL." Goldsmith & Posner III, supra note 14, at 193; see also Goldsmith & Posner IV, supra note 14, at S118 (citing Truman's declaration as an example of nondeceitful "international talk"); id. at S129 (explaining that Truman's declaration, by clarifying the area "over which [the United States] plans to exert control," warned away other states and reduced the potential for conflict). It is by no means clear why, in their account, such an announcement "is credible and influences the behavior of foreign nations." Id. at S129. In any event, a concession that custom-talk may identify focal points does not, in my view, adequately convey custom's function in facilitating the creation of those points in the first place. The U.S. claim appealed in part because of its inherently reciprocal nature, and was credible because of the investment of U.S. legal reputation, rather than because of any more implicit threat of conflict. See infra Part III.B. Nothing in this requires, moreover, that a particular custom (or even the system of customary international law) have a mysterious "normative gravitational pull," but only that its legal "rhetoric" convey something distinctly of value.

(171.) See Concerning Right of Passage Over Indian Territory (Port. v. India), 1960 I.C.J. 6, 39 (Apr. 12) (stating, in response to objection that "no local custom could be established between only two States," that "[t]he Court sees no reason why long continued practice between two States accepted by them as regulating their relations should not form the basis of mutual rights and obligations between the two States"); RESTATEMENT (THIRD), supra note 9, [section] 102 cmts. b, e (recognizing the existence of special customary law); D'AMATO, supra note 25, at 233-34 (arguing that scholars have "failed to draw a basic distinction between special (or 'local' or 'particular') customary international law and general international law"); WOLFKE, supra note 20, at 88-89 (contending that the distinction between particular and general international law has been recognized by the ICJ).

(172.) See, e.g., Akehurst, supra note 24, at 39 (describing potential for protesting states to change customary rules).

(173.) See, e.g., ELIAS & LIM, supra note 27, at 12 (describing a "stages" approach to formation of customary international law); FINAL REPORT, supra note 36, introduction at 7 (explaining that because "[p]art of the confusion may be caused by a failure to distinguish between different stages in the life of a customary rule," "[o]nce a customary rule has become established, States will naturally have a belief in its existence: but this does not necessarily prove that the subjective element needs to be present during the formation of the rule").

(174.) This need not, however, be the case. In each of the two principles suggested here, references to a "sufficient" number of states may be construed either in traditional (and underspecified) customary international law terms, or may be specified by the pioneering state or states, much as with special or local custom. See supra note 171 (noting doctrine of special or local custom); infra note 191 and accompanying text (indicating the salience of local custom to rational choice critique).

(175.) See, e.g., Roberts, supra note 5, at 783 (discussing the relative normativity of customs).

(176.) See RESTATEMENT (SECOND) OF CONTRACTS [section] 35(1) ("An offer gives to the offeree a continuing power to complete the manifestation of mutual assent by acceptance of the offer.").

(177.) E.g., Military and Paramilitary Activities in and Against Nicaragua (Nicar. v. U.S.), 1986 I.C.J. 14, 109 (June 27) ("Reliance by a State on a novel right or an unprecedented exception to the principle might, if shared in principle by other States, tend towards a modification of customary international law."); see supra notes 114-23 and accompanying text (describing the theory and practice of opinio juris). Two popular treatises elaborate the argument, albeit without regard to state interests or rational choice. Professor Harris, for example, queries:
   Would it be correct to say that in the early days of the formation
   of a new rule the state or states adopting the practice either do
   not think about whether it is binding or, if they are thinking
   about its significance in the development of international law,
   put it forward more as an "offer," which other states can accept
   or reject, rather than as something which they are convinced is
   already binding? On this view, the feeling of obligation, if it
   arises at all, arises only later when there has been general
   adoption or acceptance of the practice or "offer."

D.J. HARRIS, CASES AND MATERIALS ON INTERNATIONAL LAW 41 n.4 (5th ed. 1998). Professor Shaw is more assertive:
   [O]ne has to treat the matter in terms of a process whereby states
   behave in a certain way in the belief that such behaviour is law
   or is becoming law. It will then depend
   upon how other states react as to whether this process of
   legislation is accepted or rejected. It follows that rigid
   definitions as to legality have to be modified to see whether
   the legitimating stamp of state activity can be provided or not.
   If a state proclaims a 12 mile limit to its territorial sea in
   the belief that although the 3 mile limit has been accepted law,
   the circumstances are so altering that a 12 mile limit might
   now be treated as becoming law, it is vindicated if other states
   follow suit and a new rule of customary law is established. If
   other states reject the proposition, then the projected rule
   withers away and the original rule stands, reinforced by state
   practice and commonly accepted.

SHAW, supra note 24, at 69.

(178.) The principle would, for example, fit squarely within Professor D'Amato's description of behavior constituting the "acceptance" of a practice as law as required by Article 38, and of rule-articulation by a pioneering state as evidence of opinio juris. D'AMATO, supra note 25, at 73-87. His theory seems to have been conceived of as a revision to the traditional approach to opinio juris, see id. at 73-74, but its departure was most obvious in respects not strictly relevant here, such as in the variety of actors who might "articulate" the rule and the interplay between that element and an arguably relatively restrictive approach to relevant state practices. See, e.g., Akehurst, supra note 24, at 1-3 (criticizing Professor D'Amato's approach to practice, partly on the ground that it departs from the academic consensus); id. at 35-37 (quibbling with Professor's D'Amato's approach to opinio juris, but agreeing in principle); Mendelson, supra note 24, at 195 n.70 (describing Professor D'Amato's approach as "unconventional," including as to opinio juris, but principally as to the sources of articulation and the ease by which assertions could be confirmed as law). Any discrepancy is perhaps less clear now than it once was. See supra notes 114-23 and accompanying text (describing contemporary approaches to opinio juris).

In claiming consistency with the standard version of customary international law and its incremental variants, this Article differs from Professor Guzman's argument: while he reaches the conclusion that states may have an incentive to preserve their reputations for legal compliance in multistage games, he would appear to agree with Professors Goldsmith and Posner that any such function is inconsistent with the function of opinio juris in the standard version. See Guzman, supra note 15, at 1887 ("The theoretical problems with [customary international law] stem from the commitment of legal scholars to the traditional definition that emphasizes ... opinio juris."). Not insignificantly, his argument that international law should be reconceived as consisting of those norms the violation of which would harm a country's reputation as law abiding, while compatible in many respects with that presented here, would include "soft law" that under traditional approaches has a murkier standing. Without additional rules of recognition, this would pose serious difficulties for those pioneering states seeking to determine the status of their conduct.

(179.) See supra notes 27-46 and accompanying text.

(180.) See Goldsmith & Posner I, supra note 14, at 1135 ("[A] reputation for compliance with international law is not necessarily the best means--and certainly not the only means--for accomplishing foreign policy objectives." (citing Keohane, supra note 114, at 496-99)); Jack Goldsmith, Sovereignty, International Relations Theory, and International Law, 52 STAN. L. REV. 959, 985 (2000) (arguing, similarly, that maintaining an outstanding reputation for compliance with international agreements is not the preeminent diplomatic tool).

(181.) As the number of occasions for relying on one's reputation dwindles, so does its value; as value diminishes, so does investment in that reputation, and credibility for conforming to it.

(182.) See LOUIS HENKIN, HOW NATIONS BEHAVE: LAW AND FOREIGN POLICY 47 (2d ed. 1979) (asserting, famously, that "almost all nations observe almost all principles of international law and almost all of their obligations almost all of the time"); Koh, supra note 17, at 2599 n.2 (citing empirical work supporting Professor Henkin's claim). As Professor Hathaway and other participants in a burgeoning compliance literature have shown, compliance depends upon the kind of international law at stake--in terms of its content and form--as well as the kind of nation at issue. See Oona A. Hathaway, Do Human Rights Treaties Make a Difference?, 111 YALE L.J. 1935, 2020-25 (2002) (concluding that although ratification of human rights treaties may sometimes be associated with worsening human rights practices, their ratification by democracies appears to be more closely associated with better practices). For a compelling argument that genuinely "rogue" states are relatively few and far between, see Goldsmith & Posner IV, supra note 14, at S115-16 (noting the exceptional nature of Athenian candor and contrasting it with Hitler's behavior). One reason, perhaps, is because the reputation is costly to achieve and maintain. Cf. AXELROD, supra note 63, at 152-53 (describing bullying). More commonly, states will simply lack reputations on which they could rely, or they will regard it as too costly to establish them. Charles Lipson, Why Are Some International Agreements Informal?, 45 INT'L ORG. 495, 509-10 (1991).

(183.) This recalls Professor Fearon's more general observation that lengthening the shadow of the future, which is thought to improve the ease of enforcing international conventions, may simultaneously make negotiating them much more difficult. James D. Fearon, Bargaining, Enforcement, and International Cooperation, 52 INT'L ORG. 269, 286-93 (1998).

(184.) Cf. Lipson, supra note 182, at 508 (explaining that treaty pledges purposefully raise the political costs of noncompliance and positing that informal agreements, though generally less reliable than treaties, involve less of a reputational stake).

(185.) This discussion is derived from BAIRD ET AL., supra note 2, at 191-93. In their similar example of an embedded game, they suppose that "staying left" in the initial game involves a symmetrical payoff for each participant, but that is not indispensable to the game's operation.

(186.) Professor Chinen makes a similar argument in his recent article, but he premises it in research relating to communities of individuals. Chinen I, supra note 15, at 168-70. But see Goldsmith & Posner III, supra note 14, at 197 (suggesting "many reasons why individuals would be less self-regarding than nations").

(187.) Cf. JON ELSTER, ULYSSES UNBOUND 88-174 (2000) (discussing skeptically precommitment in politics, especially as concerns constitutions).

(188.) In tentatively extending their analysis to treaties, Professors Goldsmith and Posner remark on the superiority of treaties in communicating shared understandings that help resolve cooperation or coordination problems. See generally Goldsmith & Posner I, supra note 14, at 1170-72. But the choice to employ treaties is also conventionally understood (even by political scientists) to reflect a relatively public and serious investment of reputation, such that states would hesitate to enter into them if they strongly suspected that they would likely breach--unless, presumably, the gains from such a strategy were exceptional. See Lipson, supra note 182, at 508 (noting that "It]he effect of treaties ... is to raise the political costs of noncompliance.... not only for others but also for oneself," so that "It]he more formal and public the agreement, the higher the reputational costs of noncompliance"); id. at 508-09 (concluding that "one crucial element of treaties is that they visibly stake the parties' reputations to their pledges"); id. at 511 (noting that, despite limitations to the reputational argument, "treaties are a conventional way of raising the credibility of promises by staking national reputation on adherence").

(189.) For an analysis of these and other virtues of informal agreements, see Lipson, supra note 182, at 500-01,514-23.

(190.) See supra notes 74-75 and accompanying text (discussing the difficulties in achieving multinational cooperation).

(191.) Edward T. Swaine, The Local Law of Global Antitrust, 43 WM. & MARY L. REV. 627, 706-25 (2001).

(192.) George W. Downs et al., Is the Good News About Compliance Good News About Cooperation?, 50 INT'L ORG. 379, 380, 384-87 (1996).

(193.) Scott Barrett, A Theory of Full International Cooperation, 11 J. THEORETICAL POL. 519, 523 (1999).

(194.) See, e.g., Douglas Cassel, Does International Human Rights Law Make a Difference?, 2 CHI. J. INT'L L. 121,127 (2001) (arguing that the concept of universality demonstrates the fundamental nature of human rights and increases their legal and political relevance).

(195.) Indeed, a recent article argued that not only were the normative implications of cultural relativism overstated, but that the virtue of universal human rights law consisted in large part of its "maximally inclusive" nature, which "accommodates the greatest diversity of alternative cultural conceptions of human dignity." Robert D. Sloane, Outrelativizing Relativism: A Liberal Defense of the Universality of International Human Rights, 34 VAND. J. TRANSNAT'L L. 527,594 (2001).

(196.) See lan Sinclair, Estoppel and Acquiescence, in FIFTY YEARS OF THE INTERNATIONAL COURT OF JUSTICE 120 (Vaughan Lowe & Malgosia Fitzmaurice eds., 1996) (concluding that "the Court will be reluctant to penalize a state unduly for inconsistency of conduct, and in particular to find that the conduct relied on has created an estoppel in the strict sense"); id. at 111-16 (citing cases). The Court appears more amenable, however, to closely related principles of acquiescence. Id. at 106, 116. Its underlying ambivalence is well illustrated by the Nicaragua proceedings. In the jurisdictional phase, it rejected on the facts the U.S. claim that Nicaragua was estopped from invoking the Court's compulsory jurisdiction. Concerning Military and Paramilitary Activities in and Against Nicaragua (Nicar. v. U.S.), 1984 I.C.J. 392, 414-15 (Nov. 26). At the merits phase, too, it observed that "It]he mere fact that States declare their recognition of certain rules is not sufficient for the Court to consider these as being part of customary international law, and as applicable as such to those States." Concerning Military and Paramilitary Activities in and Against Nicaragua (Nicar. v. U.S.), 1986 I.C.J. 4, 87 (June 27). But it also noted the U.S. acceptance of a principle reflected in the final act adopted at a 1975 Helsinki conference, concluding that "[a]cceptance of a text in these terms confirms the existence of an opinio juris of the participating States prohibiting the use of force in international relations," id. at 90, and ultimately concluded that "it can be inferred that the text testifies to the existence, and the acceptance by the United States, of a customary principle which has universal application." Id. at 97.

(197.) See supra notes 39-40 and accompanying text (discussing FINAL REPORT, supra note 36).

(198.) FINAL REPORT, supra note 36, [section] 18 cmt. d, at 40 n.96.

(199.) See BYERS, supra note 24, at 36 (acknowledging and critiquing this principle).

(200.) For example, in the North Sea Continental Shelf case, the ICJ observed that "it might be that, even without the passage of any considerable period of time, a very widespread and representative participation in the convention might suffice" to establish a general international principle, "provided [the participation] included that of States whose interests were specially affected." North Sea Continental Shelf (F.R.G. v. Den.; F.R.G.v. Neth.), 1969 I.C.J. 3, 42 (Feb. 20); accord id. at 43 (indicating that "an indispensable requirement" in order to establish international law would be for the relevant practice to "includ[e] that of States whose interests are specially affected"); Fisheries Jurisdiction (U.K. v. Ice.), 1974 I.C.J. 3, 90 (July 25) (citing North Sea Continental Shelf cases); FINAL REPORT, supra note 36, at 26 (asserting that "it is not simply a question of how many States participate in the practice, but which States").

(201.) Thus Professor Akehurst, for example, explained the reliance on specially affected states as a matter of fact--due simply to the likelihood of their contribution--rather than in any normative terms. See Akehurst, supra note 24, at 22-23. Rational choice theory does support, however, giving greater heed to those states that have in fact articulated views, as opposed to the silent majority who nonetheless are bound under traditional theory--while reputation-oriented variants of the kind articulated here have a harder time explaining compliance by silent states, given the relatively weak investment indicated by their acquiescence.

(202.) Cf. Mendelson, supra note 24, at 186 (noting that dissemination of information is reducing potential disparities, but that information overload may pose identical difficulties).

(203.) See BYERS, supra note 24, at 37 (noting the relative influence of interested states) (citations omitted).

(204.) One of these latter possibilities appears to have motivated Professor Baxter's position. See Richard R. Baxter, Treaties and Custom, 129 RECUEIL DES COURS DE L'ACADE DE DROIT INTERNATIONAL 25, 66 (1970) (suggesting that special weight be accorded to "the size of the State, the volume of its international relations, and ... the contribution it makes to the development of international law").

(205.) See, e.g., VILLIGER, supra note 114, at 32-33 (rejecting the argument that special weight should be attributed to the size and importance of a state when assessing state practice); Kelly, supra note 12, at 518-26. The Final Report recognizes this criticism, but responds that "leaving aside the question what is meant by 'democratic' in this context, it should be noted that customary systems are rarely completely democratic: the more important participants play a particularly significant role in the process." FINAL REPORT, supra note 36, [section] 14 cmt. e.

(206.) But see MARTIN IRA GLASSNER, ACCESS TO THE SEA FOR DEVELOPING LANDLOCKED STATES 16-38 (1970) (discussing the interest of landlocked states in, among other things, rights of transit).

(207.) See Mendelson, supra note 24, at 186-87 (noting that dissemination of information is reducing potential disparities, but that information overload may pose identical difficulties); see, e.g., Olivier M. Ribbelink, Technological Development and the Development of the Law of Outer Space, 10 HAGUE Y.B. INT'L L. 3, 8 (1997) (noting the challenge posed by the New International Economic Order and the debate between those states favoring free access to materials in outer space and those states favoring regulated access).

(208.) The two are not necessarily correlated. One can imagine, for example, that a state might be among the few routinely having to cope with the recovery of valuable antiquities off its coast, but that such matters were relatively unimportant to its vision of the national interest.

(209.) See, e.g., Hersch Lauterpacht, Sovereignty over Submarine Areas, 1956 BRIT. Y.B. INT'L L. 392, 395 (invoking "[u]nilateral declarations by traditionally law-abiding states, within a province which is particularly their own").

(210.) See Statute of the International Court of Justice, supra note 20, art. 36(2), 59 Stat. 1055, 1060, 3 Bevans 1153, 1187 (declaring that states parties, given the consent of any other state, can recognize the jurisdiction of the Court as compulsory). This would suggest, accordingly, that arguments for the nullification of compulsory jurisdiction on grounds of nonuse are unwise. See J. Patrick Kelly, The International Court of Justice: Crisis and Reformation, 12 YALE J. INT'L L. 342, 362-69 (1987) (discussing the ICJ and the ineffectiveness of compulsory jurisdiction). It may be appropriate, however, to qualify acceptance of that jurisdiction in order to avoid strategic exploitation by other states. See, e.g., Anthony A. D'Amato, Modifying U.S. Acceptance of the Compulsory Jurisdiction of the World Court, 79 AM. J. INT'L L. 385, 387-94 (1985) (explaining strategic problems and describing solutions). The latitude of individual states to do so, and the advisability of so doing, are complex subjects, and beyond the scope of this Article. For a collection of essays on compulsory jurisdiction and related issues, see generally THE INTERNATIONAL COURT OF JUSTICE AT A CROSSROADS (Lori F. Damrosch ed., 1987).

(211.) Jonathan I. Charney, International Agreements and the Development of Customary International Law, 61 WASH. L. REV. 971, 971 (1986).

(212.) See FINAL REPORT, supra note 36, pt. IV, at 42-54 (suggesting that treaties establish no presumption in favor of any potential effect on customary international law, but may be probative, depending on circumstance).

EDWARD T. SWAINE, Assistant Professor, The Wharton School, University of Pennsylvania. A.B. Harvard, 1985; J.D. Yale Law School, 1989. I received helpful comments on an earlier draft from participants at the University of Chicago Law School's International Law Workshop in November 2001 and at University of Florida's Hurst Seminar in January 2002, and from Jeffrey Dunoff, Eric Orts, Eric Posner and Alan Strudler. Errors remain my own.
COPYRIGHT 2002 Duke University, School of Law
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2002 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Customary International Law
Author:Swaine, Edward T.
Publication:Duke Law Journal
Date:Dec 1, 2002
Previous Article:The juvenile death penalty and international law.
Next Article:A public trust argument for public access to private conservation land.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters