¹ Cf. Kaplan, Abraham, The Conduct of Inquiry (San Francisco 1964)Google Scholar, Part IX, “Explanation.”

² An excellent statement of what may be meant by “rational discussion” of this sort is to be found in Flathman, Richard E., The Public Interest (New York 1966)Google Scholar, Part II.

³ Compare David Easton's well-known remarks in The Political System (New York 1963)Google Scholar, chap. 9.

⁴ Smith's phrasing of the questions he wished to study may be found at p. 26.

⁵ The literature of the pluralist approach to American politics is immense and continues to proliferate. The best recent work in the genre is Rose, Arnold, The Power Structure (New York 1967)Google Scholar. Other well-known examples of this approach are Dahl, Robert, A Preface to Democratic Theory (Chicago 1956)Google Scholar and Who Governs? (New Haven 1961)Google Scholar; Polsby, Nelson W., Community Power and Political Theory (New Haven 1963)Google Scholar; Bell, Daniel, “Is There a Ruling Class in America? The Power Elite Reconsidered,” in his The End of Ideology, rev. ed. (New York 1965)Google Scholar; Riesman, David, Glazer, Nathan, and Denney, Reuel, The Lonely Crowd (New Haven 1950)Google Scholar, Part II; Lipset, Seymour Martin, The First New Nation (New York 1963), 318–48Google Scholar; Parsons, Talcott, “The Distribution of Power in American Society,” World Politics x (October 1957), 123–43CrossRefGoogle Scholar; Galbraith, John K., American Capitalism (Boston 1952)Google Scholar; and Truman, David, The Governmental Process (New York 1953)Google Scholar. For an interesting and somewhat different viewpoint, see Kornhauser, William, “Power Elite or ‘Veto Groups,’” in Lipset, Seymour Martin and Lowenthal, Leo, eds., Culture and Social Character (Glencoe 1961)Google Scholar.

⁶ Among the many works calling into question various aspects of the pluralist approach are the following: Lowi, Theodore, “The Public Philosophy: Interest-Group Liberalism,” American Political Science Review lxi (March 1967), 5–24CrossRefGoogle Scholar, and “American Business, Public Policy, Case-Studies, and Political Theory,” World Politics xvi (July 1964), 677–715Google Scholar; Bachrach, Peter, The Theory of Democratic Elitism. (Boston 1967)Google Scholar and, with Baratz, Morton S., “Two Faces of Power,” American Political Science Review LVI (December 1962), 947–52Google Scholar; Kariel, Henry, The Promise of Politics (Englewood Cliffs 1966)Google Scholar, esp. chap. 4; Lynd, Robert S., “Power in American Society as Resource and Problem,” in Kornhauser, Arthur W., ed., Problems of Power in American Democracy (Detroit 1957)Google Scholar; Ono, S., “The Limits of Bourgeois Pluralism,” Studies on the Left v (Summer 1965), 46–72Google Scholar; and McConnell, Grant, Private Power and American Democracy (New York 1966)Google Scholar.

⁷ Intensity of preference, which also affects influence on outcomes, is also distributed unequally—at least with regard to the ability to express it in relevant ways. On the prerequisites and problems of participation, see Dahl, Preface, 71 ff.

⁸ For example, the cases of COMSAT or the 1962 tax cut or the continuing oil depletion allowance are evidence of this. It should be observed that even where access to decision-making processes is distributed equally, compromise outcomes may be highly unequal because the prevailing social context favors one group or another in bargaining. Some demands may simply not be made because a group sees no hope in making them or diinks of tliem as “illegitimate”; some demands may not have to be made because the desiderata contained in them are automatically granted by custom or long-settled law. On this point, see Bachrach and Baratz.

⁹ And even when presidential or party leadership is most centralizing, voter control through representatives or parties is not always technically possible. Some of the technical problems of majority voting and voter control are discussed in Dahl's Preface, chap. 4. And see also McCloskey, Herbert and others, “Issue Conflict and Consensus Among Party Leaders and Followers,” American Political Science Review LIV (March 1960), 405–27Google Scholar, on the divergence of interests and goals between political leaders and political followers.

¹⁰ McConnell, Private Power, and Lowi, “The Public Philosophy,” contain especially good accounts and critiques of this process.

¹¹ P. 488.

¹² , Mills, The Power Elite (New York 1957)Google Scholar, chap. 11.

¹³ For an exhaustive critique of those assumptions and of deterrence policy itself, see Green, Philip, Deadly Logic: The Theory of Nuclear Deterrence (Columbus 1966)Google Scholar; and Etzioni, Amitai, The Hard Way to Peace (New York 1962)Google Scholar.

¹⁴ It should be said that this is sometimes his argument. Especially throughout chap. 4, we can find examples of both viewpoints; on the whole, however, I think that celebration of die techniques themselves is Smidi's dominant attitude. See, for example, his concluding remarks (pp. 315ff.).

¹⁵ Examples of Smith's approach to the methodological question may be found on pp. IO-II, 24–25, 64–65, and 316, among many other instances. The language he uses in describing the adoption of innovative techniques is invariably the language of necessity, e.g., “The 'analytic' style in defense decision making is probably an inescapable concomitant of the growing complexity of defense issues” (p. 316).

¹⁶ See Green, Deadly Logic, esp. chaps. 2–4 and 7, and “Method and Substance in the Arms Debate,” World Politics xvi (July 1964), 642–67Google Scholar. Smith expresses one caveat on the question of value-neutrality: “There have doubtless been occasions when analytic studies have made unrealistic assumptions, drawn false or misleading inferences from empirical data, and failed to take into consideration vital aspects of a broad problem. But this is not an argument against the advisory function in principle; this is an argument for more and better analysis and an increased professionalism in the advisory ranks” (pp. 315–16). Unfortunately, the notion that professionals are especially adept at avoiding unrealism, falsity, and narrowness is one for which little evidence has ever been offered, at least with regard to the social sciences. Indeed, in the complicated area of national security analysis only Laplace's Demon will ever know what is “real” and “true” before the fact. What is needed is not truer assumptions but an analytic method that generates varying assumptions about the political world and calls forth the full range of its practitioners' political wisdom and moral judgment. The RAND technique of systems analysis does none of these things.

¹⁷ The more we look back at the development of deterrence theory, the harder it becomes to tell whether these propositions were ever really empirically separable. On this point, see my comments in Deadly Logic, 106ff.

¹⁸ This point was made first by Sol Stern in his “Who Thinks in a Think Tank?” New York Times Magazine (April 16, 1967), 28. The quotations cited here are taken from p. 122 of Stern's article.

¹⁹ These are, of course, the titles of RAND-sponsored works by Leites and Selznick on Soviet ideology and politics.

²⁰ Apparently RAND personnel are able to live with the following passage from an internal Air Force study, written about 1952, and quoted by Smith: ““The lawyer-client relationship of RAND to the Air Force places upon RAND certain restrictions. It is inevitable that the three Departments of the National Military Establishment will compete for budgets, facilities, and military responsibilities. As a result, it is inappropriate for RAND to “represent” more than one of the services'” (p. 83).

²¹ Smith's list of RAN D “contributions” begins at p. 109.

²² The quotation is from Stern's article, p. 119, and is apparently based on an interview with Amron Katz of RAND. Stern also mentions other RAN D studies related to the war effort, at least one of which has generated useful scholarly data that are being kept secret for political reasons.

²³ I have remarked on this point at more length in Deadly Logic, chap. 7.

²⁴ Arnold Rose takes a qualifiedly hopeful view of the growing military role in American politics (pp. 134–52), but at one point (p. 141) he writes: 'There is evidence to support the view that the United States is moving in the direction of the garrison state—the already mentioned use of military 'expertise,' placement of military men into civilian posts, and the increasing trend toward secrecy in government. There is also some evidence to support the thesis that science, technology, labor, and industry are becoming increasingly dependent upon the military.” Later (p. 142) Rose says: “If the garrison state should develop in the United States the trends suggest that the political elite would dominate it here also, rather than an economic or military elite.” It is not clear if Rose thinks diat such an eventuality would be preferable to the alternatives, or, if so, why.

²⁵ At p. 143 Smith offers a long list of research institutes in the field of national security that have been inspired by RAND's success.

²⁶ It is important to note that Smith never questions whether the words “academic,” “university,” and such, are done justice by the close ties between such scientific research centers as M.I.T.—which has a retired Air Force general as its vice-president—and the government. One must at least ask whether the presence on the RAN D board of such men as the presidents of Rice University and Case Institute, the provost of M.I.T., the director of the Lawrence Radiation Laboratory, or the vice-president of the Brookings Institution actually adds to the Board any viewpoint unique to academia.

²⁷ The Politics of Oil (New York 1961)Google Scholar, chap. 12.

²⁸ At p. 97, Smith does emphasize RAND's attempt “to avoid identification with partisan causes.” Th e apparent assumption that “nonpartisan” causes, whatever they may be, are somehow free of the taint of politics is an assumption that can be supported only by the most narrow, Bentleyan view of what constitutes political life and political action. On this point, see below, p. 322.

²⁹ Thus the invocation of the institutions referred to in n. 25 above is rhetorical rather than analytic, since the nature of their work, support, and power is left totally unexplored by Smith.

³⁰ The concept of “countervailing power” seems to me to be at the core of the pluralist approach; see Galbraith for the original formulation of this notion.

³¹ For an expression of this viewpoint, see Snyder, Glenn H., “The Politics of National Defense: A Review of Samuel P. Huntington, The Common Defense” Journal of Conflict Resolution vi (December 1962), 368–73CrossRefGoogle Scholar. A cogent reply to Snyder, and indeed die most cogent handling of the “national interest” question that I know of, is Cohen, Bernard C., “Military Policy Analysis and the Art of the Possible,” Journal of Conflict Resolution vi (June 1962), 154–59CrossRefGoogle Scholar. As the dates suggest, Cohen's article is not actually a reply to Snyder's; rather, it is a review of Snyder's book Deterrence and Defense (Princeton 1961)Google Scholar.

³² Compare the remark of Walter Stein, “It is common, indeed usual, to be uncertain of a boundary but quite certain of what lies well to the east or west of it...” (in , Stein, ed., Morals and Missiles: Catholic Essays on the Problem of War Today [London 1959], 25Google Scholar).

³³ Some discussions of democratic theory have seemed to suggest that what is meant by “pluralism” is contention over particular policies, but consensus on basic beliefs; when one attempts to act outside that consensus, one can only expect to be ignored or thwarted. (See, for example, Dahl's Who Governs?, chap. 28.) It may be that Smith, Glenn Snyder, and others have some such view in mind with regard to the “foreign policy consensus,” as Gabriel Almond calls it in his The American People and Foreign Policy (New York 1950)Google Scholar. However, I do not think that point is applicable to this discussion. Descriptively, no doubt, the proposition about consensus is unexceptionable and even tautological—depending, however, on what one means by “basic beliefs.” My own suggestion is that such a phrase becomes highly normative if it is taken to mean anything more than the words and symbols that are considered legitimate as descriptions of the goals one seeks or the actions one undertakes. The “democratic creed” thus refers to no particular (i.e., operational) ends—except for those for whom the word “un-American” has an intimate meaning.

³⁴ Perhaps they have been beyond debate at RAND, for as I mentioned earlier, RAND has published no works calling any of those premises into question. And it has certainly sponsored no contributions to the growing literature of “cold-war revisionism,” as represented by, for instance, Alperovitz's, GarAtomic Diplomacy From Potsdam to Hiroshima (New York 1965)Google Scholar.

⁸⁵ The ultimate significance of the RAND bases study only became obvious to outsiders with the publication of Wohlstetter's, Albert“The Delicate Balance of Terror,” Foreign Affairs xxxvii (January 1959), 211–34CrossRefGoogle Scholar, which made quite clear that RAND theorists envisaged a secure deterrent as being both immense and multipurposed.

³⁸ Writers as diverse as James Warburg, Frederick L. Schuman, Walter Lippmann, and George F. Kennan disagreed with American European policy in whole or in part and questioned the premises on which it was based. On “minimum deterrence,” see Singer, J. David, Deterrence, Arms Control, and Disarmament (Columbus 1962)Google Scholar. It must be remembered, incidentally, that Smith has chosen as his example of RAND's contribution what is certainly the best major RAN D study; if he had chosen instead the work of Herman Kahn and his colleagues on civil defense—work that quite obviously took its logical cues from Wohlstetter's study—he would have had to defend the indefensible rather than merely ignore the controversial. (For an extended critique of die civil defense study see my Deadly Logic, chap. 2.)

⁸⁷ See Almond, 138.

³⁸ This is true even though one notes RAND's successes in die completely noncontroversial areas I have mentioned earlier. From a critical standpoint these successes would be viewed as residual benefits that do not balance the general costliness to American pluralism of RAND's operations; furthermore, there is no reason to believe that RAN D or any similar organization was the only potential source of all the innovations tfiat Smith attributes to it.

³⁹ See n. 28 above.

⁴⁰ I have not emphasized this point, but in a way Smith's refusal to deal with it is astonishing. Are scholarship and “creative research” at all compatible with secrecy? Are we to consider classified work a contribution to knowledge or simply, until it is made public, a fiction? Can a scholar or scientist accept the military's rationalizations for secrecy and remain true to his vocation? Perhaps there is a responsible argument to be made for mixing military secrecy with civilian research (though I have never seen one that was not merely self-serving on the part of the military), but surely the whole question deserves a chapter in a book such as Smith's, rather than relegation to a footnote.

⁴¹ For example, a recent major interdisciplinary study at RAND, published as The Urban Transportation Problem, by J. R. Meyer, J. F. Kain, and M. Wohl (Cambridge, Mass., 1965), was financed chiefly by a grant from the Ford Foundation. That study itself happens to be another excellent example of the limitations that seem built into the sophisticated methods RAND has developed. Meyer et al. compare auto and rail transportation and find the latter wanting on the whole; but they do not consider urban air transportation, which is especially important in the northeastern urban corridor. Thus they take no notice of the tremendous social and economic costs that air travel is causing; and they therefore fail to make the obvious observation that the auto can never compete with the airplane as a mode of interurban transportation, but the high-speed railroad can, and may be much cheaper. And as with RAND's strategic studies, I think the fault lies not so much with the practitioners of systems analysis but with the spurious value-neutrality of the method itself. One would have to feel the destruction caused by the auto and the airplane in combination—one would have to have that specific bias—in order to be led to that observation.