¹ This is the basis of my disagreement with several of the other contributors to this volume. Donald J. Puchala and Raymond F. Hopkins, for example, treat international regimes as coextensive with international politics. Similarly, although Oran Young does not formally equate international politics with regimes, his definitions, both of regimes and of international relations, suggest such an equivalence. Also see his International Regimes: Problems of Concept Formation,” World Politics 32 (04 1980): 331–56, andGoogle ScholarCompliance and Public Authority: A Theory with International Applications (Baltimore: Johns Hopkins University Press, 1979). My concern is to develop a conceptualization of regimes that delineates a subset of international politicsGoogle Scholar.

² Waltz, Kenneth N., Theory of International Politics (Reading, Mass.: Addison-Wesley, 1979), p. 96Google Scholar.

³ On the importance of property rights, see Thomas, M. Carroll, , Ciscil, David H., and Chisholm, Roger K., “The Market as a Commons: An Unconventional View of Property Rights,” Journal of Economic Issues 13 (06 1979): 605–627. On the constrained sense of economic competition, seeGoogle ScholarHirshleifer, J., “Competition, Cooperation, and Conflict in Economics and Biology,” American Economic Review 68 (05 1978): 238–43; andGoogle ScholarHirshleifer, J., “Economics From a Biological Viewpoint,” Journal of Law and Economics 20 (04 1977): 1–52Google Scholar.

⁴ The term “complex interdependence” is most fully presented in Keohane, Robert O. and Nye, Joseph S., Power and Interdependence: World Politics in Transition (Boston: Little, Brown, 1977); yet it remains unclear, for example, if the use of force remains an option in the relations between advanced industrial societies but is dominated by other choices. Alternatively, it may be that nations sometimes prefer to threaten the use of force on a contingent basis, but recognize that the outcome resulting from the mutual use of force is the least preferred outcome for all actorsGoogle Scholar.

⁵ The A, B, outcome is also a coordination equilibrium, which David K. Lewis defines as an outcome from which neither actor can shift and make anyone better off; see Convention: A Philosophical Study (Cambridge: Harvard University Press, 1969), p. 14Google Scholar.

⁶ Individual accessibility is discussed by Elster, Jon in Ulysses and the Sirens: Studies in Rationality and Irrationality (Cambridge: Cambridge University Press, 1979), p. 21Google Scholar.

⁷ Only A, B, however, is a coordination equilibrium. The other equilibrium outcome, A2B2, is not a coordination equilibrium because each actor can shift from it and make the other better off by doing so. For Lewis this does not pose a coordination problem, which requires the existence of two or more coordination equilibria; see Convention, p. 24.

⁸ For Elster, this case is individually inaccessible. Nonetheless, he expects convergence because the outcome is individually stable. I consider this case to be individually accessible precisely because there are convergent expectations. Note that if regimes are understood to include any devices that help actor expectations to converge, then regimes might arise even in this case, although solely to provide information. The profferred information would provide each actor with assurance about the others' preferences, as would be necessary for expectations to converge on the one of the two equilibria that all prefer. I find a problem with Robert O. Keohane's treatment of the role of information in his article in this volume. He argues that, given a demand for international agreements, the more costly the information the greater the actual demand for international regimes (one of whose functions is to improve the information available to actors). It is unclear whether he means to suggest that all mechanisms which provide information are examples of regimes even when the actors' interests are harmonious, or that they are not regimes because there is no demand for agreements in such cases. Since he presents the demand for agreements as a given in his formulation, we do not know if the demand for information can be a basis for a demand for agreements or simply a basis for a demand for regimes which assumes a demand for agreements. His formulation is too imprecise to adduce the standing of assurance mechanisms and whether they do or do not constitute regimes.

₉ In some cases, actors may require mechanisms for assurance, which extradition treaties xemplify. These treaties might thus be seen as “assurance regimes,” regimes that arise when each actor's knowledge of others’ preferences is enough to allow the actors' autonomous decisions to bring them to the outcome they all most prefer.

¹⁰ The conceptualization of regimes presented here, that they arise to deal with the dilemmas of common interests and common aversions, is not, therefore, based on any inherent notion of “principles.” Indeed, it is easy to conceive of unprincipled regimes, such as OPEC. Regimes may, but need not, have some principle underlying them.

¹¹ The prisoners’ dilemma is the only two-actor example of a Pareto-deficient equilibrium that occurs when both actors have dominant strategies. It is for this reason that it has received so much scholarly attention.

¹² The role of game models in analyzing oligopolistic relations is described by Markham, Jesse W., “Oligopoly,” in International Encyclopedia of the Social Sciences, vol. 11 (New York: Macmillan, 1968): 283–88Google Scholar. Scherer, F. M. discusses the prisoners' dilemma, as a model for oligopolistic interaction in Industrial Market Structure and Economic Performance (Chicago: Rand McNally, 1970)Google Scholar. The same observation is made by Lester G. Telser, who redubs the prisoners' dilemma as it applies to oligopolies the “cartel's dilemma”; see Competition, Collusion, and Game Theory (Chicago: Aldine-Atherton, 1972), p. 143Google Scholar.

¹³ For a general discussion of suboptimality, see Elster, Jon, Logic and Society: Contradictions and Possible Worlds (New York: Wiley, 1978), pp. 122–34Google Scholar.

¹⁴ In one formulation, Jon Elster defines politics as “the study of ways of transcending the Prisoner's Dilemma”; see “Some Conceptual Problems in Political Theory,” in Power and Political Theory: Some European Perspectives, ed. by Barry, Brian (London: Wiley, 1976), pp. 248–49Google Scholar. Laurence S. Moss provides an assessment of modern and somewhat formal equivalents to the Hobbesian, and Lockean, views of state formation in “Optimal Jurisdictions and the Economic Theory of the State: Or, Anarchy and One-world Government Are Only Corner Solutions,” Public Choice 35 (1980): 17–26Google Scholar. See also Taylor, Michael, Anarchy and Cooperation (London: Wiley, 1976)Google Scholar. Elster, criticizes Taylor's, alternative in Logic and Society, pp. 156–57Google Scholar, and Ulysses and the Sirens, pp. 64, 143, 146.

¹⁵ Hardin, Russell, “Collective Action as an Agreeable n-Prisoners’ Dilemma,” Behavioral Science 16 (09 1971): 472–81. Note Elster's distinction between counterfinality and suboptimality in explaining the behavior of free riders; Logic and Society, pp. 122–23CrossRefGoogle Scholar.

¹⁶ The dilemmas discussed in this article refer to specific actors and not necessarily to the system as a whole. In the prisoners' dilemma, for example, only the prisoners themselves face a Pareto-deficient outcome. The rest of society finds the outcome of their dilemma to be optimal. This is precisely analogous to the situation of oligopolists, who prefer collusion to competition. The rest of society, however, would prefer that they compete rather than collude. The collective suboptimality need not necessarily exist for all actors in the system.

¹⁷ This literature was spawned by Olson, Mancur Jr. and Zeckhauser, Richard, “An Economic Theory of Alliances,” Review of Economics and Statistics 48 (08 1966): 266–79. Other essays linking collective goods and international cooperation includeGoogle ScholarRussett, Bruce M. and Sullivan, John D., “Collective Goods and International Organization,” International Organization 25 (Autumn 1971): 845–65Google Scholar; Ruggie, John Gerard, “Collective Goods and Future International Collaboration,” American Political Science Review 66 (09 1972): 874–93Google Scholar; and Sandier, Todd and Cauley, Jon, “The Design of Supranational Structures: An Economic Perspective,” International Studies Quarterly 21 (07 1977): 251–76Google Scholar. More recent work stresses different institutional arrangements for international collective goods. See Sandier, Todd M., Loehr, William, and Cauley, Jon T., “The Political Economy of Public Goods and International Cooperation,” Monograph Series in World Affairs 15 (1978)Google Scholar, and Snidal, Duncan, “Public Goods, Property Rights, and Political Organizations,” International Studies Quarterly 23 (12 1979): 532–66Google Scholar.

¹⁸ Indeed, international trade regimes have historically exemplified the subsystemic character of many regimes. Scholars often characterize the mid 19th century, for example, as the era of free trade. Yet several major states, including the United States and Russia, did not take part. Similarly, the post-1945 era is now commonly referred to as the period of American economic hegemony. Ironically, this characterization is of a postwar economic system established by and within the sphere of only one pole of a bipolar international system-a bipolarity that has typically been offered as the most important characterization of the age. In other words, we should continually be reminded that references to the postwar economic system are, in fact, to a subsystem that excludes the Soviet bloc. See Arthur A. Stein, “The Hegemon's Dilemma: Great Britain, the United States, and the International Economic Order,” paper presented at the annual meeting of the American Political Science Association, New York, 4 September 1981.

¹⁹ A similar argument can sometimes be made about the decision to devalue a currency or maintain par value in a fixed exchange rate system when devaluation, although every nation's dominant strategy, results in the suboptimal outcome of mutual devaluation. Richard N. Cooper uses simple games in his discussion of the choice of an international monetary regime; see “Prolegomena to the Choice of an International Monetary System, International Organization 29 (Winter 1975): 63–97CrossRefGoogle Scholar.

²⁰ In the dilemma of common interests, actors are averse to the suboptimal equilibrium outcome and resolution involves their arriving at the outcome they prefer to the equilibrium one. The dilemma is their inability individually to arrive at the outcome they prefer to the equilibrium one. In the dilemma of common aversions, on the other hand, the actors do have a'common interest in avoiding a particular outcome but their dilemma is the possibility that they might arrive at a mutual aversion without some coordination. Beyond their desire to avoid that aversion, however, they disagree about which of the multiple equilibria they prefer.

²¹ Both equilibria are also coordination equilibria. In this case, there is no minimax solution.

²² If each of the actors chooses its minimax option, the A, B, outcome results. This outcome is not their mutual aversion, but it is a Pareto-deficient nonequilibrium outcome because both prefer it less than either equilibrium.

²³ Precommitment has been variously described as the power to bind, as imperfect rationality, and as egonomics; see Schelling, Thomas C.. , The Strategy of Conflict (Cambridge: Harvard University Press, 1960), pp. 22–28Google Scholar; Elster, , Ulysses and the Sirens, pp. 36–111Google Scholar; and Schelling, T. C., “Egonomics, or the Art of Self-Management,” American Economic Review 68 (06 1978): 290–94Google Scholar. Such a formulation of prior agreement on principles does not require John Rawls's veil of ignorance; see A Theory of Justice (Cambridge: Harvard University Press, 1971)Google Scholar. Thinking ahead without agreement in strategic interaction, however, is no solution; see Schick, Frederic, “Some Notes on Thinking Ahead,” Social Research 44 (Winter 1977): 786–800Google Scholar.

²⁴ The prisoners’ dilemma is the only situation with a Pareto-deficient equilibrium in which all the actors have dominant strategies. There are other cases of Pareto-deficient equilibria in which some have dominant strategies and some contingent strategies. These too are dilemmas of common interests and require regimes for solution; in these cases, however, only those actors with dominant strategies must eschew independent decision making. Thus, the regime formed to insure collaboration in this case is likely to have stipulations and requirements that apply asymmetrically to those who must eschew independent decision making to achieve optimality and to those who must be assured that the others have actually done so and will continue to do so.

Some argue that the cooperative nonequilibrium outcome of the prisoners’ dilemma can emerge spontaneously-without collaborative agreement. Social psychologists have done extensive experiments on the emergence of cooperation in repeated plays of the prisoners’ dilemma game; the most recent review is by Pruitt, Dean G. and Kimmel, Melvin J., “Twenty Years of Experimental Gaming: Critique, Synthesis, and Suggestions for the Future,” Annual Review of Psychology 28 (1977): 163–92Google Scholar. See also Rapoport, Anatol, Guyer, Melvin J., and Gordon, David G., The 2x2 Game (Ann Arbor: University of Michigan Press, 1976). For a mathematician's deductive assessment of the prospects for the emergence of such cooperation, seeGoogle ScholarSmale, Steve, “The Prisoner's Dilemma and Dynamical Systems Associated to Non- Cooperative Games,” Econometrica 48 (11 1980): 1617–1634. See alsoGoogle ScholarAxelrod, Robert, “The Emergence of Cooperation Among Egoists,” American Political Science Review 75 (06 1981): 306–318. The conditions for this are rarely met in international politics, however. The first such requirement is that play be repeated indefinitely. Because states can disappear, and because they are therefore concerned with their own survival, international politics must be seen as a finite game by the actors. Moreover, the stakes in international politics are typically so high that fear of exploitation will insure that states follow their dominant strategy, to defect, in the absence of a collaborative agreementGoogle Scholar.

²⁵ Fisher, Bart S., The International Coffee Agreement: A Study in Coffee Diplomacy (New York: Praeger, 1972), andGoogle ScholarBilder, Richard B., “The International Coffee Agreement: A Case History in Negotiation,” Law and Contemporary Problems 28 (Spring 1963): 328–91. The latter appeared in a special issue devoted to “International Commodity Agreements.”Google Scholar

²⁶ Hardin, Garrett, “The Tragedy of the Commons,” Science 162 (13 12 1968): 1243–1248Google Scholar; Thomas, C. Schelling characterizes the commons as a prisoners’ dilemma in Micromotives and Macrobehavior (New York: W. W. Norton, 1978), pp. 110–15Google Scholar.

²⁷ The following authors all discuss coordination, although they do not agree fully on a definition: Schelling, Strategy of Conflict; Lewis, Conventions; Heymann, Philip B., “The Problem of Coordination: Bargaining and Rules,” Harvard Law Review 86 (03 1973): 797–877; andGoogle ScholarGoodin, Robert E., The Politics of Rational Man (London: Wiley, 1976), pp. 26–46Google Scholar. The distinction between collaboration and coordination made here can be compared to distinctions between negative and positive coordination and between negative and positive cooperation made by the following: Marina, v. Whitman, N., “Coordination and Management of the International Economy: A Search for Organizing Principles,” in Contemporary Economic Problems 1977, ed. by Fellner, William (Washington, D.C.: American Enterprise Institute for Public Policy Research, 1977), p. 321Google Scholar; and Pelkmans, Jacques, “Economic Cooperation Among Western Countries,” in Challenges to Interdependent Economies: The Industrial West in the Coming Decade, ed. by Gordon, Robert J. and Pelkmans, Jacques (New York: McGraw-Hill, 1979), pp. 97–123Google Scholar.

²⁸ This notion of self-enforcement differs from that developed by Telser, L. G., “A Theory of Self-enforcing Agreements,” Journal of Business 53 (01 1980): 27–44CrossRefGoogle Scholar. For Telser, an arrangement is self-enforcing if the actor calculates that defection may bring future costs. Thus, even if cheating brings immediate rewards, an actor will not cheat if others ‘responses cause it to bear a net loss. For me, regimes are self-enforcing only if the cost that an actor bears for defecting is immediate rather than potential and is brought about by its own defection rather than by the response of others to that defection.

²⁹ Standardization may reflect harmonious interests rather than coordination solutions to dilemmas of common aversions. This may, for example, explain the adoption of a common calendar.

³⁰ Schelling, provides an interesting discussion of the traffic light as a self-enforcing convention in Micromotives and Macrobehavior, pp. 119–21Google Scholar.

³¹ The organization is the governing body for almost all international civil air traffic.

³² There does exist a dilemma of common aversions that can be solved by coordination or by collaboration. Like other situations characterized as dilemmas of common aversions, the actors in the game of chicken have contingent strategies, do not agree on a most-preferred outcome, but do share a mutual aversion. In this case, the actors diverge in their assessment of the two equilibria. Unlike those of other dilemmas of common aversions, the two equilibria in chicken are not coordination equilibria. In chicken, the nonequilibrium minimax outcome is the second choice of both actors and is not Pareto-deficient. Thus, the situation is not merely one of deadlock avoidance, but one that can be solved either by coordination to arrive at one of the two equilibria or by collaboration to accept second-best. Here, too, the collaboration is not selfenforcing and requires mutual assurances about defection from a particular outcome. No-fault insurance agreements are one example of a collaboration regime to resolve a dilemma of common aversions. Note that Lewis would not consider chicken to be a coordination problem because the two equilibria in chicken are not coordination equilibria. I believe that it is a coordination problem, but one that collaboration can also solve.

³³ For background and analysis of the most recent World Administrative Radio Conference of 1979, see the articles in Foreign Policy no. 34 (Spring 1979): 139–64Google Scholar; and those in Journal of Communication 29 (Winter 1979): 143–207Google Scholar. See also “Scramble for the Waves,” Economist, 1 09 1979, p. 37Google Scholar; “The Struggle Over the World ‘s Radio Waves Will Continue,” Economist, 8 December 1979, p. 83; and “Policing the Radio,” New Statesman, 14 December 1979, p. 924.

³⁴ The absence of regimes does not mean, however, that the actors are independent of one another.

³⁵ The conditions in which misperception matters, and the ways in which it matters, are delineated in Stein, Arthur A., “When Misperception Matters,” World Politics 34 (07 1982)CrossRefGoogle Scholar.

³⁶ Goodin, , in Politics of Rational Man, p. 26Google Scholar, puts it this way: “Joint decision making is said to occur when all actors participate in determining the decisions of each actor. It implies that there was interaction between all the actors prior to the decisions and that this interaction shaped the decision of each actor.” It is not surprising, then, that two recent formulations both stress the importance of agreement as part of their definition of regimes: see Young, “International Regimes”; and Haas, Ernst B., “Why Collaborate? Issue-Linkage and International Regimes,” World Politics 32 (03 1980), p. 358Google Scholar. For interesting delineations of the range of decision-making procedures, see Knut Midgaard, “Co-operative Negotiations and Bargaining: Some Notes on Power and Powerlessness,” in Barry, Power and Political Theory; and Zartman, I. William, “Negotiations as a Joint Decision-Making Process,” Journal of Conflict Resolution 21 (12 1977): 620–23Google Scholar. Both of these authors, however, heavily emphasize the bargaining process. Various forms of international cooperation can also be seen as forms of decision making; see Tinbergen, Jan, “Alternative Forms of International Co-operation: Comparing Their Efficiency,” International Social Science Journal 30 (1978): 224–25Google Scholar.

³⁷ Alker, Hayward R. Jr, “A Methodology for Design Research on Interdependence Alternatives,” International Organization 31 (Winter 1977), pp. 37–38CrossRefGoogle Scholar.

³⁸ Although I do not define regimes by reference to their degree of institutionalization, it is the case that collaboration regimes are more likely to be institutionalized than coordination regimes, because of the requirements of enforcement.

³⁹ Difference maximization is discussed by Mc Clintock, Charles G., “Game Behavior and Social Motivation in Interpersonal Settings,” in Experimental Social Psychology, ed. by Mc Clintock, Charles Graham (New York: Holt, Rinehart and Winston, 1972), pp. 271–92Google Scholar. Taylor calls them pure difference games and designates them a subtype of games of difference generally; see Anarchy and Cooperation, pp. 73–74. See also Shubik, Martin, “Games of Status,” Behavioral Science 16 (03 1971): 117–29Google Scholar.

⁴⁰ Those who argue that world politics constitutes a zero-sum game cannot, of course, sustain their position at the extremes. After all, it is impossible for all dyadic relationships to be zerosum or constant-sum in a world of more than two actors. Thus, even if some relationships in international politics are zero- or constant-sum, there must also exist some subset of relationships that are nonconstant-sum and therefore provide a basis for regime formation among this subset of nations. Yet Robert Gilpin still claims “that in power terms, international relations is a zero-sum game.” See U. S. Power and the Multinational Corporation: The Political Economy of Foreign Direct Investment (New York: Basic Books, 1975)Google Scholar.

⁴¹ Recent exponents of the predominance model of stability, as opposed to the classical balance-of-power model of stability, include Organski, A. F. K., World Politics, 2nd ed. (New York: Knopf, 1968), pp. 338–76Google Scholar, and Modelski, George, “The Long Cycle of Global Politics and the Nation-State,” Comparative Studies in Society and History 20 (03 1978): 214–35Google Scholar. The international political economy variant of the argument is provided by Krasner, Stephen D., “State Power and the Structure of International Trade,” World Politics 28 (03 1976): 317–47Google Scholar; see also Stein, “The Hegemon's Dilemma.”

⁴² Note that this is precisely the way in which Krasner develops his argument in “State Power.”

⁴³ Burns, Tom and Buckley, Walter, “The Prisoners’ Dilemma Game as a System of Social Domination,” Journal of Peace Research 11 (1974): 221–28CrossRefGoogle Scholar.

⁴⁴ Jervis, Robert, “Cooperation Under the Security Dilemma,” World Politics 30 (01 1978): 167–214CrossRefGoogle Scholar.

⁴⁵ The quotation and the substantive discussion are from Pannenberg, Charles O., A New International Health Order: An Inquiry into the International Relations of World Health and Medical Care (Germantown, Maryland: Sijthoff and Noordhoff, 1979), pp. 179–80Google Scholar.

⁴⁶ Regulations founded on health reasons and scientifically based can still become the basis of political disagreement, as the Japanese response to the California medfly spraying in 1981 demonstrates.

⁴⁷ See, for example, Gourevitch, Peter Alexis, “International Trade, Domestic Coalitions, and Liberty: The Crisis of 1873–1896,” Journal of Interdisciplinary History 8 (Autumn 1977): 281–313Google Scholar; and Kurth, James R., “The Creation and Destruction of International Regimes: The Impact of the World Market,” paper delivered at the American Political Science Association Meeting, Washington, D.C., 08 1980Google Scholar.

⁴⁸ North, Robert C., “Toward a Framework for the Analysis of Scarcity and Conflict,” International Studies Quarterly 21 (12 1977): 569–91CrossRefGoogle Scholar.

⁴⁹ Note that this clearly distinguishes domestic sectoral from international structural approaches. Although both approaches can be seen as delineating the determinants of actor preferences, the international structural perspective can be claimed to determine the constellation of all actors' preferences. Thus, the existence of offensive weapons creates a prisoners' dilemma situation for any pair of nations. On the other hand, the sectoral approach explains one actor's preferences at a time, and thus must be linked with an analysis of the interaction between actors to explain outcome. This is, of course, why the analysis of foreign policy is not equivalent to the analysis of international relations. Thus, the works of Allison, Gourevitch, Katzenstein, and Kurth, among others, which explain foreign policy by reference to domestic economic or bureaucratic interests, remain incomplete precisely because they do not incorporate relations between nations.

⁵⁰ The recognition of the multiple determination of actor interests also makes possible an issue approach to international politics that is not necessarily issue-structural.

⁵¹ One can, of course, expect there to be lags between changes in interests and actor behavior; see Nicholson, Michael, Oligopoly and Conflict: A Dynamic Approach (Liverpool: Liverpool University Press, 1972). Schick distinguishes realization lags from adaptation lags in “Some Notes on Thinking Ahead,” p. 790Google Scholar.

⁵² One can argue that regimes actually change actor preferences. The property rights argument about dealing with externalities through changes in liability rules is an example of a situation in which prearranged agreements are specifically devised in order to change utilities in subsequent interaction; see Conybeare, John A. C., “International Organization and the Theory of Property Rights,” International Organization 34 (Summer 1980): 307–334CrossRefGoogle Scholar.

⁵³ Akerlof, George A., “A Theory of Social Custom, of Which Unemployment May Be One Consequence,” Quarterly Journal of Economics 94 (06 1980): 749-75CrossRefGoogle Scholar.