¹ Most of the discussion in this essay is relevant, mutatis mutandis, to the whole field of political science. Since my principal concern here is international relations, however, I will not emphasize this point explicitly in my argument.

² For a helpful discussion of some of the meanings associated with the term “theory” see Rapoport, Anatol, “Various Meanings of ‘Theory’,” American Political Science Review, LII (1958), 972–88.CrossRefGoogle Scholar And for an instructive discussion that includes comments on some of the meanings attached to the term “theory” by European writers see Runciman, W. G., Social Science and Political Theory (Cambridge 1965).Google Scholar

³ For good introductory discussions of the formal implications of this definition of theory consult, inter alia, Brodbeck, May, “Explanation, Prediction, and ‘Imperfect’ Knowledge,” in Brodbeck, May, ed., Readings in the Philosophy of the Social Sciences (New York 1968), 363–97Google Scholar, and Hempel, Carl G., Philosophy of Natural Science (Englewood Cliffs 1966).Google Scholar

⁴ Popper, Karl R., Conjectures and Refutations: The Growth of Scientific Knowledge (New York 1963), esp. 33–65.Google Scholar

⁵ For an account that stresses some of these subjective elements in the construction of theories see Polanyi, Michael, Science, Faith, and Society (Chicago 1964).Google Scholar

⁶ Any well-known text in philosophy of science will serve as an introduction to these rules, even though there are various significant issues on which the leading scholars in this field disagree. Consult, inter alia, Hempel, Carl G., Aspects of Scientific Explanation (New York 1965)Google Scholar; Nagel, Ernest, The Structure of Science (New York 1961)Google Scholar, and Popper, Karl R., The Logic of Scientific Discovery (New York 1959).Google Scholar

⁷ This point is explicated in a particularly suggestive fashion by Kuhn. See Kuhn, Thomas S., The Structure of Scientific Revolutions (Chicago 1962).Google Scholar

⁸ For an introductory discussion of modern logic see Kemeny, John G. and others, Introduction to Finite Mathematics (2nd ed., Englewood Cliffs 1966), 1–57.Google Scholar Suggestions for additional readings in this area can be found on p. 57.

⁹ Whether or not the logical model associated with a theory actually mirrors reality in some objective sense is a subject of considerable controversy, much of it metaphysical in nature, among philosophers. But this issue is not relevant to the argument set forth in this essay.

¹⁰ Parsimony has two distinguishable manifestations. First, in the case of two theories designed to explain the same phenomenon, the one that requires the smallest number of assumptions and/or premises is superior in terms of parsimony. Second, all theories gain in parsimony to the extent that they are able to explain the maximum number of different phenomena in terms of the minimum number of assumptions and/or premises. In both cases, however, parsimony has to do with minimizing the formal apparatus of one's model while expanding explanatory leverage to the fullest extent possible.

¹¹ See, for example, Carl G. Hempel, “The Theoretician's Dilemma,” in Hempel (fn. 6), 173–226.

¹² For an argument along these lines that is particularly relevant to the social sciences see Rapoport, Anatol, Two-Person Game Theory (Ann Arbor 1966), pp. 186–214.Google Scholar

¹³ For clear statements of this view consult ibid., chap. 12, and Rapoport (fn. 2).

¹⁴ This is a key point of divergence between non-empirical disciplines such as logic and mathematics and all of the empirical disciplines.

¹⁵ In some parts of physics, for example, theorists come up with sophisticated mathematical models on a day-to-day basis, but the great bulk of these models are discarded almost as quickly as they are formulated.

¹⁶ There is some debate among scientists about the exact point at which a theory should be discarded for lack of predictive accuracy. It is sometimes argued, for example, that if a theory is sufficiently elegant and aesthetically pleasing, it should be pursued, at least for a while, even in the face of negative empirical evidence. And the example of the ultimate vindication of the Copernican system of astronomy is often cited in support of this position. In the context of the present discussion, however, this is a nuance that should not be allowed to obscure the point that models which fail the test of predictive accuracy should not be retained indefinitely.

¹⁷ In accordance with the general orientation of my argument, this paragraph stresses the potential functions of theory for policymakers. There are, of course, other functions of theory which have little bearing on the problems of policymaking. The very act of constructing theories, for example, is regarded by many theorists as a rewarding process in itself. Moreover, many investigators pursue the development of theories from an interest in explanation rather than prediction, even though the logical status of explanation and prediction is treated as identical by most philosophers of science.

¹⁸ Note, however, that the theory of the dynamics of the solar system has considerable practical value because it allows man to manipulate his own behavior in the light of predictions concerning the behavior of the solar system.

¹⁹ Not only is the intellectual tradition in this field heavily nontheoretical, but also those who have provided financial support for research on international questions have seldom displayed a serious interest in the development of viable theories.

²⁰ Richardson, Lewis F., Arms and Insecurity (Pittsburgh and Chicago 1960).Google Scholar

²¹ Rapoport, Anatol, “Lewis F. Richardson's Mathematical Theory of War,” Journal of Conflict Resolution, 1 (September 1957), 249–99CrossRefGoogle Scholar, and Rapoport, Anatol, Fights, Games, and Debates (Ann Arbor 1960), Part I.Google Scholar

²² Peter Busch, “Mathematical Models of Arms Races,” pp. 193–233 in Russett, Bruce M., What Price Vigilance? (New Haven 1970).Google Scholar

²³ McGuire, Martin, Secrecy and the Arms Race (Cambridge, Mass. 1966).Google Scholar

²⁴ Briefly, rationality can be defined in terms of the following conditions: 1) the actor evaluates alternatives in his environment on the basis of his preferences among them, 2) his preference ordering is consistent and transitive, and 3) he always chooses the preferred alternative.

²⁵ There is now a large literature on the theory of games. A lucid review that is still highly relevant is Luce, R. Duncan and Raiffa, Howard, Games and Decisions (New York 1957).Google Scholar For a more recent overview that is quite sensitive to the difficulties of applying game theoretic models to problems in the field of international relations see Rapoport (fn. 12), together with his N-Person Game Theory (Ann Arbor 1970).Google Scholar

²⁶ Strategic behavior is the behavior of an individual member of a group which involves a choice of action contingent upon that individual's estimates of the actions (or choices) of others in the group, where the actions of each of the relevant others are based upon a similar estimate of the behavior of group members other than himself. Strategic interaction is the set of behavior patterns manifested by individuals whose choices are interdependent in this fashion.

²⁷ In fact, game theoretic solutions do not even correspond well to empirical outcomes in highly restricted experimental situations. For a helpful discussion of this subject see Rapoport, Anatol and Orwant, Carol, “Experimental Games: A Review,” Behavioral Science, VII (January 1962), 1–37.Google Scholar

²⁸ For discussions of manipulative bargaining see Schelling, Thomas C., The Strategy of Conflict (Cambridge, Mass, 1960)Google Scholar, and Young, Oran R., Bargaining (Princeton Center of International Studies 1970), mimeo, Part IV.Google Scholar

²⁹ See especially Rapoport, Anatol, Strategy and Conscience (New York 1964), Parts I and II.Google Scholar

³⁰ Boulding, Kenneth, Conflict and Defense (New York 1962).Google Scholar

³¹ Riker, William, The Theory of Political Coalitions (New Haven 1962).Google Scholar

³² A collective good is any good that cannot be withheld from any member of a specified group once it is supplied to some member of the group. For an interesting, if somewhat preliminary, attempt to construct a theory of alliances based on the concept of collective goods see Olson, Mancur Jr. and Zeckhauser, Richard, “An Economic Theory of Alliances,” Review of Economics and Statistics, XLVIII (August 1966), 266–79CrossRefGoogle Scholar, and Olson, and Zeckhauser, , “Collective Goods, Comparative Advantage, and Alliance Efficiency,” in McKean, Roland N., ed., Issues in Defense Economics (New York 1967), 25–63.Google Scholar For another analysis of collective goods that could be applied to issues in the field of international relations see Frohlich, NormanOppenheimer, Joe A., and Young, Oran R., Political Leadership and Collective Goods (Princeton 1971).Google Scholar

³³ For a discussion of these models consult Young (fn. 28), esp. Part II.

³⁴ For some interesting efforts to construct models of this kind within the general conceptual framework of game theory see Rapoport, Anatol and Chammah, Albert, Prisoner's Dilemma (Ann Arbor 1965).CrossRefGoogle ScholarPubMed Hayward Alker, Jr. is also interested in the development of models of this kind. See, for example, Hayward R. Alker, Jr. and Cheryl Christensen, “From Causal Modelling to Artificial Intelligence: The Evolution of a U.N. Peace-Making Simulation,” to be published in Jean LaPonce, ed., Experimentation and Simulation in Political Science (forthcoming).

³⁵ For a survey of relevant material on complex organizations see March, James G. and Simon, Herbert A., Organizations (New York 1958).Google Scholar For preliminary efforts to move toward logical models of bureaucratic behavior consult Tullock, Gordon, The Politics of Bureaucracy (Washington 1965)Google Scholar and Downs, Anthony, Inside Bureaucracy (Boston 1967).CrossRefGoogle Scholar

³⁶ That is, there may be excellent reasons to treat the problems of international relations as special cases of more general theoretical constructs, rather than develop entirely new theories of international relations.

³⁷ On transnationalism see Robert O. Keohane and Joseph S. Nye, Jr., eds., Transnational Processes and International Organization, a special issue of International Organization, XXV (No. 3, Summer 1971).Google Scholar On bureaucratic politics see Allison, Graham T., “Conceptual Models and the Cuban Missile Crisis,” American Political Science Review, LXIII (September 1969), pp. 689–718CrossRefGoogle Scholar, and Neustadt, Richard, Alliance Politics (New York 1970).Google Scholar

³⁸ At a minimum, a conceptual framework, or an approach to analysis, is an interrelated set of concepts (with their definitions), variables, and assumptions or premises. For an elaboration of this concept with reference to the field of political science see Young, Oran R., Systems of Political Science (Englewood Cliffs 1968).Google Scholar

³⁹ For a good introductory discussion of these approaches consult Herz, John H., Political Realism and Political Idealism (Chicago 1951).Google Scholar

⁴⁰ The practical value of projections derived from constant conjunctions will generally increase as a direct function of the lack of predictions based on viable theories.

⁴¹ That is, some may wish to regard these techniques as being based on theories since it is possible to achieve logical closure through the simple procedure of assuming that the future will be like the past and that intervening factors are irrelevant. Even if this argument (which is not very persuasive) is granted, however, there is still the problem that predictions derived from such models do not fare very well in most areas within the social sciences.

⁴² For a variety of examples along these lines consult Kahn, Herman and Wiener, Anthony J., The Year 2000 (New York 1967).Google Scholar

⁴³ For clear definitions of the key concepts associated with oligopoly and imperfect competition in general in the field of economics see Samuelson, Paul A., Economics (5th ed., New York 1961)Google Scholar, chap. 25.

⁴⁴ Thus, in economic relationships, oligopoly covers the ground between monopoly on the one hand and perfect competition on the other.

⁴⁵ For a good introduction to the study of oligopolistic markets see Fellner, William J., Competition Among the Few (New York 1949).Google Scholar

⁴⁶ For an extended discussion of strategic interaction and the analytic problems it raises see Young (fn. 28), General Introduction.

⁴⁷ For a study of international relations by an economist who emphasizes this link see Schelling (fn. 28).

⁴⁸ For an introductory discussion of the application of this assumption to the decision-making problems faced by typical firms see Samuelson (fn. 43), chap. 24.

⁴⁹ For a discussion of many of the issues raised by this perspective on the behavior of the firm see Cyert, Richard M. and March, James G., A Behavioral Theory of the Firm (Englewood Cliffs 1963).Google Scholar

⁵⁰ See, inter alia, Allison, (fn. 37).

⁵¹ To see this compare such works as Cyert and March (fn. 49), Gordon Tullock (fn. 35), and Anthony Downs (fn. 35), with recent discussions of “bureaucratic politics” such as Allison (fn. 37) and Neustadt (fn. 37).

⁵² All of the problems referred to in this paragraph are discussed at greater length in Young (fn. 28), General Conclusion. Accordingly, I have deliberately restricted the present discussion to a rudimentary statement of the issues.

⁵³ Note that this analysis deals only with the development of viable theories and not with other intellectual activities in the field of international relations.

⁵⁴ For an excellent statement of this view see Rapoport (fn. 12), 186–214.

⁵⁵ For a discussion that suggests this conclusion see Graham T. Allison and Morton Halperin, “Bureaucratic Politics: A Paradigm and some Policy Implications,” in this issue.

⁵⁶ Premises are empirical generalizations, but they play the same analytic role that assumptions play in logical models.

⁵⁷ For a convenient collection that illustrates a wide variety of efforts along this line consult Singer, J. David, ed., Quantitative International Politics (New York 1968).Google Scholar

⁵⁸ In fact, there is good reason to believe that those who do not derive pleasure from playing with logical models will rarely become important theorists (as that concept is employed in this essay) in any discipline.

⁵⁹ Note, however, that a viable theory of the behavior of individual actors developed by others might be plugged into the models of those concerned with interactions among actors with excellent results. In economics, for example, this has been the role of the traditional simple theory of the firm for students of market interactions.

⁶⁰ This argument need not hold of course if one's primary objective is not the development of viable theories.

⁶¹ For arguments to this effect with specific reference to international relations see Busch (fn. 22), and Shubik, Martin, “On the Study of Disarmament and Escalation,” Journal of Conflict Resolution, XII (March 1968), 83–101.CrossRefGoogle Scholar

⁶² For some illustrations of these possibilities see, inter alia, Frohlich, Oppenheimer, and Young (fn. 32), and Cross, John G., The Economics of Bargaining (New York 1969).Google Scholar

⁶³ There is widespread confusion among social scientists concerning these issues. Whereas physical scientists regard the failure of their initial model as perfectly normal, many social scientists (other than economists) tend quite erroneously to interpret initial failures as evidence that viable theories cannot be constructed about the substantive issues at hand.

⁶⁴ On the contrary, as I indicated earlier in this essay, some theorists regard simplicity as so important that they are willing to hang onto a simple and elegant model for a time even in the face of initially negative empirical evidence.

⁶⁵ One striking example among many in the field of international relations is Snyder, Richard C. and others, eds., Foreign Policy Decision Maying (New York 1962).Google Scholar

⁶⁶ For a more detailed statement of my views on this subject see Young, Oran R., “Professor Russett: Industrious Tailor to a Naked Emperor,” World Politics, XXI (April 1969), 486–511.CrossRefGoogle Scholar

⁶⁷ Though heavily inductive work is not logically at odds with efforts to develop viable theories, some of the inductive projects focusing on international relations have probably served to set back the cause of theory in the field. This happens when people unwittingly mistake the inductive studies for the sum total of the theoretical enterprise. Under these circumstances, when the highly limited results of the inductivists breed disillusionment (as is now occurring in the field of political science in general), the whole idea of developing viable theories tends to suffer from its mistaken identification with the heavily inductive work.

⁶⁸ This conclusion is relevant primarily to those academics and intellectuals who regard themselves as theorists. A moment's reflection is sufficient to affirm that numerous academics and intellectuals working on essentially nontheoretical projects have exercised substantial influence on policymaking in the international Held in recent years.

⁶⁹ Note, in this connection, the deference often accorded to the views of well-known nuclear physicists on numerous issues in the field of national security which are largely political in nature.

⁷⁰ Among other things, this suggests that those who wish to obtain influence as theorists of international relations should pay more attention to emphasizing the accuracy and utility of their predictions than to instructing unreceptive policymakers in the intricacies of their theoretical models.

⁷¹ With respect to the field of physics, such well-known figures as Vannevar Bush, Arthur Compton, and James Conant are cases in point.

⁷² I recognize that this argument raises questions about the definition of the term “intellectual.” One interesting definition of an intellectual, for example, is that he is an individual who regards the exploration and manipulation of ideas as highly valued ends in their own right. Such an individual would not commit the “sin” I am discussing here since he would not, for the most part, be interested in the practical impact of his work. In the text, I employ the term “intellectual” in its more mundane sense to refer to the class of academics, writers, and so forth.

⁷³ In fact, it is probably possible to construct an argument to the effect that policymakers have recently displayed too much sensitivity to the advice of those who purport to offer theories of international relations.

⁷⁴ This is a well-known proposition concerning the bases of scientific creativity. But the stimulus of practical problems must be clearly distinguished from an excessive impatience to achieve results or find answers directly applicable to real-world problems. The resultant tension constitutes yet another reason why the processes of scientific creativity rest on a set of delicate balances.

⁷⁵ Contrast this with the enormous sums of money that are currently required to support advanced theoretical work in fields such as high energy particle physics.

Unfortunately, the current policy of the National Science Foundation and the major private foundations with respect to research in the field of international relations does not appear to be conducive to basic theoretical work. In short, the present guidelines of the NSF and the large private foundations place a great deal of weight on shortterm relevance to current issues of national concern as a criterion in the selection of projects to support despite the fact that these organizations have, at least nominally, a greater commitment to basic research about international phenomena than any other major funding agencies within the United States government and American society in general.