¹ The concept was developed by Keohane, Robert O. and Nye, Joseph S., in Power and Interdependence (Boston: Little, Brown, 1977), 24–29Google Scholar. Some implications of the condition for international collaborative action are explored in Haas, “Turbulent Fields and the Theory of Regional Integration,” International Organization, xxx (Spring 1976), 173–214.

² The literature on regimes tends to be focused on specific regimes, not the type, such as discussions of the Bretton Woods regime. For an excellent structural analysis of that regime, see Hirsch, Fred, Doyle, Michael, and Morse, Edward L., Alternatives to Monetary Disorder (New York: McGraw-Hill, 1977)Google Scholar. Oran Young's essay in this issue, on the other hand, treats as a “regime” any international agreement which focuses and structures convergent expectations, rights, and obligations of actors with respect to a substantive matter. Although I am in broad agreement with his definitions of regime attributes, my purpose is narrower. I seek to reserve the concept for the particular kind of agreement that features “issue linkage”—a subset of Young's universe, but a broader notion than is implied in the literature on the decay of the Bretton Woods regime. I thoroughly agree with Young's restriction stressing agreement instead of following the practice of some writers who equate “regime” with the more diffuse notion of “world order.”

³ Important works of this type are: Brown, Seyom, Cornell, Nina W., Fabian, Larry L., and Weiss, Edith Brown, Regimes for the Ocean, Outer Space and Weather (Washington, D.C.: The Brookings Institution, 1977)Google Scholar; Falk, Richard A., This Endangered Planet (New York: Random House, 1971)Google Scholar; and Falk, , A Study of Future Worlds (New York: Free Press, 1975)Google Scholar. For efforts to subsume world-order issues under the larger question of natural and human evolution, see Boulding, Kenneth E., Ecody-namics: A New Theory of Societal Evolution (Beverly Hills, Calif.: Sage, 1978)Google Scholar, and Wilson, Edward O., On Human Nature (Cambridge, Mass.: Harvard University Press, 1978)Google ScholarPubMed. For some political implications of the holistic approach, see Haas, “On Systems and International Regimes,” World Politics, XXVII (January 1975), 147–74, and Haas, , “Is There a Hole in the Whole?” International Organization, XXIX (Summer 1975), 827–76CrossRefGoogle Scholar.

⁴ Liberalism and neomercantilism, politicization and depoliticization are discussed in Hirsch and others (fn. 2). Collective goods and global commons as focuses for regimes are treated by Brown and others (fn. 3), and by Morse, Edward L., “The Global Commons,” Journal of International Affairs, III (Spring/Summer 1977), 1–21Google Scholar. I question the utility of collective-goods theory when applied to these concerns, on grounds of inherent analytical shortcomings, as described in Hart, Jeffrey and Cowhey, Peter, “Theories of Collective Goods Re-examined,” Western Political Quarterly, xxx, No. 3 (1977), 351–62Google Scholar. For the classic argument that collective-goods theory predicts noncollaboration, see Ruggie, John Gerard, “Collective Goods and Future International Collaboration,” American Political Science Review, Vol. 66 (September 1972), 874–93CrossRefGoogle Scholar.

⁵ On the monetary regime, see Monroe, W. F., International Monetary Reconstruction (Lexington, Mass.: Lexington Books, 1974)Google Scholar; Law of the Sea issues are summarized well in Miles, Edward, ed., “Restructuring Ocean Regimes,” International Organization, xxxi (Spring 1977)Google Scholar. For an account of the ongoing negotiations dealing with technology transfers, see Sauvant, Karl P. and Hasenpflug, Hajo, eds., The New International Economic Order (Boulder, Colo.: Westview Press, 1977)Google Scholar, Parts 5 and 6. For North-South issues, I have relied heavily on Rothstein, Robert L., Global Bargaining (Princeton: Princeton University Press, 1979)Google Scholar, and Hansen, Roger D., Beyond the North-South Stalemate (New York: McGraw-Hill, 1979)Google Scholar. For mining, see Bulkley, George W., Who Gains From Deep-Ocean Mining} (Berkeley: Institute of International Studies, University of California, 1979)Google Scholar; for fishing, see Papers of the Program of International Studies of Fishery Arrangements (Washington, D.C.: Resources for the Future, 1974).

⁶ My use of the term “issue-area” differs from Dahl's and from the version adapted by James Rosenau for use in the comparative study of foreign policy in “Pre-Theories and Theories of Foreign Policy,” in Farrell, R. Barry, ed., Approaches to Comparative and International Politics (Evanston, 111.: Northwestern University Press, 1966)Google Scholar. In Rosenau's version, each issue-area is thought to possess unique attributes, apparent to the analyst but not necessarily to the actor, making boundary maintenance a part of the definition. Whether boundaries are or are not maintained is an important empirical question. Rosenau seeks to make issue-areas “objectively real,” while my concern is to abstract only what can be experienced by the actors.

⁷ My argument obviously sidesteps a number of competing views on this matter. Some—notably those associated with the Pugwash movement—argue for the existence of an “internationale” of scientists whose knowledge can reform the world. Others—for instance Jean-Jacques Salomon—argue that science is always used in the service of power and of the state that finances science. Neo-Marxists—for instance Jürgen Haber-mas—consider science an instrument of bourgeois hegemony within the state, if not a means of oppression. Some classical economists and political realists treat scientific knowledge as an aspect of investment, and its diffusion as a result of how competition is structured internationally. All but the Pugwash view have some plausibility, though each is in fact disputed. But each also exaggerates the case by imputing a certain monolithic quality to the core actors: classes, the state, the scientific professions, entrepreneurs. My own empirical bias forces me to reject these as meaningful categories of actors on the stage of history, at least in the last quarter of this century.

⁸ But, it is said, experts always disagree with each other; how can we speak of a consensus? In the early stages of any shift in paradigm, they certainly do disagree, and there is no consensus. Even later, experts will continue to disagree on certain aspects of their field, but not on all. In the early stages of discussion regarding the creation of regimes, the use of knowledge will be no more consensual than is the current U.S. debate among the experts on the future of nuclear energy. Experts disagree with each other on the facts and on the social goals to be realized. Moreover, they do tailor the factual discussion to whatever goal they espouse. But that is not the case during the entire lifecycle of a technology or a theory. At times there is consensus (as there was with respect to nuclear energy in the 1950s and 1960s because the dissidents were few in number and outside the policy-making process); it is permanent consensus that is inconsistent with the process of scientific investigation itself. See Margolis, Howard, Technical Advice on Policy Issues, Sage Professional Paper No. 03–009 (Beverly Hills, Calif.: Sage, 1973)Google Scholar.

⁹ Tactical and substantive linkage are discussed by Oye, Kenneth A. in “The Domain of Choice: International Constraints and the Carter Administration Foreign Policy,” in Oye, Rothchild, Donald, and Lieber, Robert J., eds., Eagle Entangled (New York: Longman, 1979)Google Scholar. Fragmented linkage is discussed in Haas, Ernst B., Williams, Mary Pat, and Babai, Don, Scientists and World Order (Berkeley: University of California Press, 1977), 121–23Google Scholar.

¹⁰ I rely greatly on Rothstein (fn. 5), in this section, even though he considers the inter-issue linkage in the Integrated Commodity Program to be essentially tactical, used solely to hold the heterogeneous Group of 77 together. Rothstein also shows that, while the parties agreed on many specific economic arguments and demonstrations, they continued to disagree on whether to act in accordance with them in the creation of the UNCTAD-designed Common Fund. What mattered was whether they thought in terms of “best-case” or “worst-case” scenarios. The issue that divided them was whether, in the process of commodity price stabilization, equity (the goal of the G-77) or efficiency (the goal of Group B) should be maximized. The Common Fund that they eventually agreed to differs from the proposal UNCTAD had used to link the issues.

¹¹ Rothstein (fn. 5), 150–51.

¹² Schelling, , The Strategy of Conflict (New York: Oxford University Press, 1963), 4Google Scholar.

¹³ See Snyder, Jack L., “Rationality at the Brink: The Role of Cognitive Processes in Failures of Deterrence,” World Politics, xxx (April 1978), 347–53Google Scholar; Jervis, Robert, Perception and Misperception in International Politics (Princeton: Princeton University Press, 1976), chaps. 4 and 11Google Scholar.

¹⁴ This distinction originates with Toulmin, Stephen, “The Complexity of Scientific Choice: A Stocktaking,” in Shils, Edward, ed., Criteria for Scientific Development, Public Policy, and National Goals (Cambridge, Mass.: M.I.T. Press, 1968), 63–79Google Scholar. The full rationale underlying this discussion and the matrix is contained in Haas, Williams, and Babai (fn. 9), chaps. 3, 5, and 6. For illustrations of the distinction in the early history of nuclear energy, see Weart, Spencer R., Scientists in Power (Cambridge, Mass.: Harvard University Press, 1979)Google Scholar.

¹⁵ This discussion runs parallel with and is indebted to Lindblom's, Charles E.Politics and Markets (New York: Basic Books, 1977)Google Scholar.

At this point we must face the question of “who decides”: a given phenomenon may be coded as an “issue” by the analyst and seen as an “issue-area” by the participating actors; two analysts may well code the phenomenon differently. How reliable, then, are our attributions?

It is useful to distinguish between three ways of assigning attributes. (1) The simplest way is to code phenomena on the basis of the actors' self-perception; if they characterize their style as “eclectic,” that is it. (2) A more complex procedure is to code the actors' behavior, not only their self-perception, on the basis of categories defined by the analyst. This procedure calls for inferring self-perceptions and matching them with actual behavior for consistency. (3) The third procedure is to neglect the actors' perceptions and behavior altogether and make attributions solely on the basis of categories set up by the analyst. The first two methods stress cognitive concepts; the third way relies on structural considerations. Making attributions on the basis of structural concepts is the procedure favored by analysts working in the Marxist, state-centric, and Parsonian traditions. It is illustrated by Rosenau's use of the notion of issue-area.

In opting for the second of these procedures, we code by asking the following questions of the negotiations being studied:

1. How did the actors define the substantive items about which they are negotiating before the session under examination?

2. In the session being studied, is there evidence that these items are being redefined and regrouped under larger headings corresponding to more ambitious goals not previously on the agenda?

3. In the session being studied, is there evidence that the redefinition and regrouping of items under larger headings is tending toward the acceptance by all actors of some core concept (or doctrine) which represents the ambitious social goals?

If we were to compare the LOS, monetary, and NIEO negotiations before and after 1967, 1971, and 1977, respectively, we would arrive at the following answers:

The price of this procedure is lack of uniformity and of full comparability among issue-areas and regimes. Uniform empirical attributions that correspond to the actors' self-perceptions are sacrificed to a scheme that allows assessments of change, but the assessment is a function of the analyst's view of the world, which may or may not be shared by the actors. Comparison of different regimes and their evolution is possible, but only if we neglect the salience of these regimes to the various actors. Since the actors' self-assessments are subordinated to the analyst's way of ordering the world, we have no way of being sure that the Law of the Sea and monetary management are of equal importance in terms of the allocations of time, energy, and attempts to plan which the governments have allocated.

¹⁶ Various modes of collaboration for reducing uncertainty, involving issues similar to technology transfers, are explored systematically and searchingly by Cowhey, Peter in The Problems of Plenty (Berkeley: University of California Press, forthcoming)Google Scholar. Cowhey shows how such collaboration can be carried out in the absence of a full-fledged regime, and at what cost. Both structural and cognitive explanations are used and fully integrated, with “structure” being identified as an unchanging and consensual knowledge base.

¹⁷ Thomas, , The Lives of a Cell (New York: Viking, 1974), 57Google Scholar.

¹⁸ One cannot speak of “a theory” of negotiation—only of competing “theories” that differ from, each other in what they attempt to predict and in how they view the process. See the special issue of the journal of Conflict Resolution, xxi (December 1977). A specification of givens that characterize negotiations relating to regime construction includes: (1) lack of a superordinate legal framework; (2) tacit acceptance of a procedure that is unstable because one or more of the parties may seek to upset it during the negotiations; (3) conviction that any resulting agreement will be sufficiently in the common interest to be implemented; (4) initial lack of accurate information in minds of negotiators with respect to the ranking of utilities in the minds of their opponents; (5) all negotiators are the instructed agents of governments and are rational actors in Schelling's sense; (6) no party has a veto power; (7) voting and adjudication cannot be used to break stalemate; (8) all parties have mixed motives because the utilities attaching to the good(s) being bargained over are not homogeneously distributed; (9) there are no deadlines; (10) mediators are active; their role is facilitated by the heterogeneous distribution of utilities; (11) non-agreement is a possibility accepted by all. Obviously, such characteristics differentiate these negotiations sharply from most game-theoretic models and all real situations that abstract from industrial relations. See also Zartman, , ed., The Negotiation Process (Beverly Hills, Calif.: Sage, 1978), 70–71Google Scholar.

The diffusion of new knowledge through various mechanisms is analyzed in Haas, Williams, and Babai (fn. 9), Part III; Touscoz, Jean, La Cooperation Scientifique Inter-nationale (Paris: Editions Techniques et Economiques, 1973)Google Scholar. The literature on this subject is summarized and reviewed by Schroder-Gudehus, Brigitte, in Price, Derek de Sola and Spiegel-Rosing, Ina, eds., Science, Technology and Society (Beverly Hills, Calif.: Sage, 1977)Google Scholar.

¹⁹ Zartman (fn. 18, 1978), 75ff., argues that the “convergence/concession” model of learning through negotiations is less powerful than the alternative “formula/detail” model. The latter, while still true to the constraints summarized in fn. 18, is said to capture more accurately the manner in which conflicts are gradually resolved; the hard negotiations are over finding a “formula” to “capture” the zone of joint gains while the details can be deduced less conflictually afterward. I would counter that this argument applies to the kinds of intra-regime negotiations carried on in adapting and maintaining the rules by incrementally tinkering with the norms, but not to the creation of a new regime where the new norm is essential.

²⁰ See John G. Cross, “Negotiation as a Learning Process,” in Zartman (fn. 18, 1978), 29–54, for a formal model of this process. In the NIEO negotiations, neither the convergence of knowledge nor the agreement on mutual utilities has occurred. But there is a literature which, if more widely accepted, would contribute to both. On the knowledge side, see the studies in Cline, William R., ed., Policy Alternatives for a New International Economic Order (New York: Praeger, 1979)Google Scholar. On the adjustment of utilities, see the ingenious argument for reconciling the “graduation,” “global reform,” and “basic needs” scenarios embedded in the NIEO controversy, in Hansen, Roger D., Beyond the 'North-South Stalemate (New York: McGraw-Hill, 1979)Google Scholar. For an interesting effort to define a consensus on both dimensions, see Freeman, Christopher and Jahoda, Marie, eds., World Futures (New York: Universe Books, 1978)Google Scholar.

²¹ The political reordering is described in Rothstein (fn. 5), chap. 7. It presupposes the kind of intellectual reordering found in Hansen (fn. 20). Rothstein argues that the parties must recognize that any single NIEO negotiating episode is merely a step in a continuing process that incorporates learning. There is evidence that the European countries and the United States do recognize this. However, the developed countries must also accept that their concern with “efficiency” cannot be squared with the emphasis placed by the G-77S on “equity,” and that the issue must be fudged or reinterpreted in short-term versus long-term considerations. This, in turn, cannot be accomplished without re-linking issues that are now treated separately; whatever issue-specific agreements are now concluded must be seen in the context of overall welfare choices that will involve substantive issue-linkage later. The separate deals must be made to interlock so as to provide side-payments to countries (especially the poorest) that perceive only costs in extensive issue-packaging. But Rothstein sees a danger in a global regime that includes all the parts; he prefers separate regimes, all of which are committed to the norm of “dependability” in meeting the G-77's aspirations, and are thus coordinated normatively rather than in terms of rules and procedures. Thus, the G-77's reliance on the maximum common denominator can continue to hold the coalition together symbolically without infecting every item on the agenda.

²² For evidence that there is some support for these hypotheses in the history of various multilateral scientific/technological programs (rather than regimes), see Haas, Williams, and Babai (fn. 9), chap. 10. The interaction of eclectics and rationalists is too farfetched to be credible. It resembles an encounter between representatives of basically different cultures. We also ignore the interaction between skeptics and rationalists, but for a different reason. Such an encounter prevails now if we take world-order modelers as rationalists and the governments of developing countries as skeptics; however, in terms of global negotiations they are on the same side! Moreover, they cannot meaningfully work together unless they agree on an integration of development strategies (means) with the goals on which both claim to be in agreement. And such integration depends on consensual knowledge of causation. The development of such knowledge is the link postulated in the skeptic-pragmatic interaction.

²³ The lawyers' simpler meaning of the word “regime” is captured in the exhaustive with health, food, and the weather. He isolates the rules in each field and analyzes the procedures for implementing them. He pays little attention to underlying norms, and none to the changing expectations of the parties, which might tell us something about the linking of old issues at a more complex level of understanding. Whether the activities discussed constitute a “regime” in my sense is not apparent from this study. On the other hand, Rangarajan, L. N., in Commodity Conflict (Ithaca, N.Y.: Cornell University Press, 1978)Google Scholar, analyzes the conditions under which multi-commodity management is feasible; in doing so, he describes the preconditions for a regime in precisely the sense here advocated, even though he does not use the term.

²⁴ The typology and the rationale on which it is based are adapted from Ruggie, John Gerard, “International Responses to Technology: Concepts and Trends,” International Organization, XXIX (Summer 1975), 570–74Google Scholar. For various ways of matching the typology with cognitive modes in explaining institutional variation, see Haas, Williams, and Babai (fn. 9), chap. 6.