¹ Until I complete my argument that it is more useful to focus on the preventive motivatioas a contributory cause of war than on preventive war as a type of war, I will continue to refeto the concept of “preventive war” in my review of the literature.

² Thucydides, , The Peloponnesian War, trans, by Warner, Rex (New York: Penguin, 1954), 1/23Google Scholar; Howard, Michael, The Causes of Wars (Cambridge: Harvard University Press, 1983), 18Google Scholar; Morgenthau, Hans J., Politics Among Nations, 4th ed. (New York: Knopf, 1967), 202–3Google Scholar; Gilpin, Robert, War and Change in World Politics (Cambridge: Cambridge University Press, 1981), 94.CrossRefGoogle Scholar

³ Organski, A.F.K., World Politics, 2d ed. (New York: Knopf, 1968), chap. 12.Google Scholar For other theories of hegemonic transitions, see Gilpin (fn. 2); Modelski, George, “The Long Cycle of Global Politics and the Nation-State”, Comparative Studies of Society and History 20 (April 1978), 214–35CrossRefGoogle Scholar; Thompson, William R., ed., Contending Approaches to World System Analysis (Beverly Hills, CA: Sage, 1983)Google Scholar; Charles F. Doran, “Power Cycle Theory and the Contemporary State System”, ibid., 165–82; Väyrynen, Raimo, “Economic Cycles, Power Transitions, Political Management and Wars between Major Powers”, International Studies Quarterly 27 (December 1983), 389–418CrossRefGoogle Scholar; Chase-Dunn, Christopher, “Interstate System and Capitalist World-Economy: One Logic or Two?” International Studies Quarterly 25 (March 1981), 19–42.CrossRefGoogle Scholar

⁴ Gilpin (fn. 2), 231–44; Keohane, Robert O., “Theory of World Politics: Structural Realism and Beyond”, in Finifter, Ada W., ed., Political Science: The State of the Discipline (Washington, DC: American Political Science Association, 1983), chap. 16Google Scholar; Modelski, George and Morgan, Patrick M., “Understanding Global War”, Journal of Conflict Resolution 29 (September 1958), 391–417CrossRefGoogle Scholar

⁵ Organski, A.F.K. and Kugler, Jacek, The War Ledger (Chicago: University of Chicago Press, 1980), chaps. 1 and 3Google Scholar; Organski (fn. 3), 371. Note that the identification of leader and challenger is critically dependent upon the conception of the geographical scope of the system and the basis of power in the system. See Levy, Jack S., “Theories of General War”, World Politics 37 (April 1985), 344–74.CrossRefGoogle Scholar

⁶ Gilpin (fn. 2), 191, 201.

⁷ Schroeder, Paul W., “World War I as Galloping Gertie: A Reply to Joachim Remak”, Journal of Modern History 44 (September 1972), 319–45CrossRefGoogle Scholar; Taylor, A.J.P., The Struggle for the Mastery of Europe, 1848–1918 (Oxford: Oxford University Press, 1954), 166Google Scholar; Anderson, M. S., 18th century Europe, 1713–1789 (New York: Oxford University Press, 1966), 34.Google Scholar For other cases, see Levy, Jack S. and Collis, Rick, “Power-Cycle Theory and the Preventive Motivation: A Preliminary Empirical Investigation”, paper presented to the American Political Science Association, New Orleans, August 1985.Google Scholar

⁸ Fischer, Fritz, Germany's Aims in the First World War (New York: Norton, 1961)Google Scholar, and Fischer, , War of Illusions (New York: Norton, 1975)Google Scholar; Albertini, Luigi, The Origins of the War of 1914, 3 vols., trans, by Massey, I. M. (London: Oxford University Press, 1952–57)Google Scholar; Ritter, Gerhard, The Sword and the Scepter, trans, bNorden, H. (Coral Gables, FL: University of Miami Press, 1970), Vol. 2Google Scholar; Geiss, Immanuel, ed., July 1914 (New York: Charles Scribner's Sons, 1967)Google Scholar

⁹ Lebow, Richard Ned, Between Peace and War (Baltimore: The Johns Hopkins University Press, 1981), 254–63Google Scholar; Evera, Stephen Van, “The Cult of the Offensive and the Origins of World War I” International Security 9(Summer 1984), 58–107CrossRefGoogle Scholar; Snyder, Jack L., “Perceptions of the Security Dilemma in 1914”, in Jervis, Robert, Lebow, Richard Ned, and Stein, Janice Gross, Psychology and Deterrence (Baltimore: The Johns Hopkins University Press, 1985), 153–79Google Scholar. Lebow's more recent position can be found in his “Windows of Opportunity: Do States Jump Through Them?” International Security 9 (Summer 1984), 147–86CrossRefGoogle Scholar. Lebow's comparative analysis is based more on opportunities arising from a favorable dyadic balance of military power than on constraints and fears arising from an impending negative shift in power and potential, and may be more relevant for static power models than for a power transition hypothesis. For quantitative empirical studies bearing on this question, see Maoz, Zeev,Paths to Conflict (Boulder, CO: Westview, 1982)Google Scholar, chaps. 4–6; Wayman, Frank Whelon, “Power Transitions, Rivalries, and War, 1816–1970”, mimeo (University of Michigan, 1983)Google Scholar. For formal theoretical treatments, see Morrow, James F., “A Twist of Truth, A Reexamination of the Effects of Arms Races in the Occurrences of War”, mimeo (University of Michigan, 1984)Google Scholar; Niou, Emerson M. S. and Ordeshook, Peter C., “Preventive War and the Balance of Power: A Game-Theoretic Approach”, Journal of Conflict Resolution 31 (September 1987).CrossRefGoogle Scholar

¹⁰ Organski (fn. 3), 376; Evera, Stephen Van, “The Causes of War”, Ph.D. diss. (University of California, Berkeley, 1984), 72–76Google Scholar; Snyder (fn. 9), 160–61.

¹¹ Van Evera (fn. 10), 71.

¹² Note that the emphasis is on the decline in relative strength, consistent with the conception of power in relational and zero-sum terms. See Gilpin, Robert, U.S. Power and the Multinational Corporation (New York: Basic Books, 1975), chap. 1CrossRefGoogle Scholar. A state whose capabilities are increasing in absolute terms may have an incentive for preventive action if its adversary is growing even faster. Note also that the emphasis is on the perception of changing power differentials by state decision makers.

¹³ Preventive military action is not, of course, the only possible “solution” to the problem of an impending decline in relative military potential. Future security might also be provided by alliances. Although the conditions under which states resort to alliances rather than to preventive military action is an important research question, my working hypothesis is that great powers are hesitant to rely on others to satisfy their long-term security needs. Alliances tend to be transient in nature and excessively affected by domestic politics and personalities; consequently, they are unreliable over the long haul. Moreover, alliances deal with the symptoms rather than with the causes of a future threat, whereas preventive military action sometimes provides the hope of dealing more directly with the source of the threat (though the true sources are often economic).

Another alternative to preventive military action, for certain states under certain conditions, would be an attempt to reverse one's decline through internal means, such as a policy of industrial revitalization. Although such efforts are more viable over the long term than the short term, this may be the preferred alternative if the immediate military threat is not too great. At this stage, the preventive motivation for war can best be analyzed by isolating it from alternative policy instruments, but ultimately the role of alliances and of internal change will have to be included in an integrated model. (The possible importance of industrial revitalization has been emphasized to me by David Lake.)

¹⁴ My assertion that states tend to give greater weight to losses than to gains is consistent with recent findings from experimental psychology. If A is preferred to B, an individual at A would be willing to pay more to avoid dropping to B than the same individual at B would be willing to pay to move up to A. See Kahneman, Daniel and Tversky, Amos, “Prospect Theory: An Analysis of Decision Under Risk”, Econometrica 47 (March 1979); 263–91CrossRefGoogle Scholar; Kahneman, and Tversky, , “Choices, Values, and Frames”, American Psychologist 39 (April 1984), 341–50CrossRefGoogle Scholar; Knetsch, Jack L. and Sinden, John A., “Willingness to Pay and Compensation Demanded: Experimental Evidence of an Unexpected Disparity in Measures of Value”, Quarterly Journal of Economics 99 (August 1984), 507–21CrossRefGoogle Scholar. This phenomenon is related to the concept of framing and to risk orientation, which are discussed later. See fn. 45.

¹⁵ For other definitions of “preventive war”, see Vagts, Alfred, Defense and Diplomacy (New York: King's Crown, 1956), 263Google Scholar; Lider, Julian, On the Nature of War (Farnborough, England: Saxon House, 1977), 63Google Scholar; Lebow (fn. 9, 1984), 154; Van Evera (fn. 11), 60–61; Organski (fn. 3), 371; Brodie, Bernard, War and Politics (New York: Macmillan, 1973), 25Google Scholar; Gilpin (fn. 2), 191; Fischer (fn. 8, 1975), 468. Fischer's definition, which requires not just the perception of inevitability but also the actual intention by the target state to launch a war against the preventer within a few years, is very restrictive. It is also self-serving, for it facilitates his thesis of German war guilt and the argument that the German decision for war was not a preventive action deriving from perceived military necessity, but instead an unprovoked war of aggression. Because the concept of the preventive motivation refers to the motivation of the preventer, it is proper to focus on his perceptions alone, without including the intentions of the target.

¹⁶ For example, German Chancellor Bethmann-Hollweg insisted to the military that they must allow Russia to mobilize first. He was concerned about the reaction not only of England, but also of the Social Democrats in Germany who, he believed, would approve war credits only for a defensive war against Russia. Similarly, German Admiral von Miiller argued that Germany should “present Russia or France or both with an ultimatum which would unleash the war with right on our side.” Moltke later agreed that “the attack must come from the Slavs.” See Fischer (fn. 8, 1975), 162; (fn. 8, 1961), 33; Fischer, , World Power or Decline? (New York: Norton, 1974), 30Google Scholar. See also Vagts (fn. 15), 290–91.

¹⁷ The Japanese attack on Pearl Harbor may be one example. It should be emphasized that preventive military action can be taken by any state in decline relative to a particular adver-sary, though it may be more common for the leading state in an international system or at least in a regional subsystem.

¹⁸ This discussion builds on some earlier distinctions made by Van Evera (fn. 9), 64n; (fn. 11), chaps. 1–2. See also Snyder (fn. 9), 160; Harkavy, Robert E., “Preemption and Two-Front Conventional Warfare” (Jerusalem Papers on Peace Problems, No. 23, Hebrew University, 1977), 6–7Google Scholar. Note that the distinction between prevention and preemption applies across the entire continuum of military action, from the surgical use of force to all-out war. The Israeli attack on the Iraqi nuclear reactor in 1982 was preventive, whereas the Israeli attack against Egyptian airfields in 1967 was preemptive.

¹⁹ The less time for these corrective measures, the more similar the situation is to preemption and the greater the incentive for military action. The preemptive motivation is generally stronger than the preventive one; the presence of both motivations is particularly destabilizing.

²⁰ Van Evera (fn. 9), 63–65.

²¹ In a crisis in which both states consider preemption, the stronger state may have an incentive to allow the adversary to strike first, as illustrated by Israel in 1973. The weaker state rarely has the luxury of waiting. For this reason, preemption is generally perceived as more legitimate for weaker states. This legitimacy reduces, to a certain extent, the potential diplomatic and domestic political costs of preemption, and thus makes such preemption more likely.

²² Van Evera (fn. 9, p. 64) and Snyder (fn, 9, p. 160 n.) argue that preemption involves the simultaneous incentive by both parties to strike first, whereas with prevention, the incentive is one-way. My analysis suggests that they are wrong on both counts.

⁹ There are some secondary differences between prevention and preemption. Whereas the preventer basically “wants war” in the sense that it would demand enormous concessions by the adversary and make few (if any) of its own in order to avoid war, that is not necessarily true for the preemptor, who may prefer to avoid war but feel compelled by circumstances to preempt. Finally, even when prevention involves (as it usually does) the perception of a high probability of a future war, the preemptor perceives the probability of an immediate war to be even greater.

²⁴ “Windows” generated by the anticipated formation of a hostile military coalition present some analytical problems, for they blur the distinction between preemption and prevention. The greater the certainty that such a coalition will form—and particularly if it has already formed—the greater the expected likelihood that war will follow, and consequently the greater the incentive to strike first. The situation then begins to resemble preemption more than prevention, and our hypotheses on the preventive motivation are less likely to be valid. An example of the preemption of a hostile coalition before it is formed is Frederick the Great's action initiating the Seven Years' War (see fn. 37); an example of the preemption of a coalition after it has formed is France's initiation of the Revolutionary Wars. Betts's argument that throughout most of the nuclear age the likelihood of preventive war has been “close to zero” while the likelihood of preemption has been considerably higher, is, if correct, but one example of the utility of this distinction. See Betts, Richard K., “Surprise Attack and Preemption”, in Allison, Graham T., Carnesale, Albert, and Nye, Joseph S. Jr., eds., Hawks, Doves, and Owls (New York: Norton, 1985), 54–79.Google Scholar

²⁵ This situation raises a difficult methodological problem. If B preempts, how do we determine whether this is exclusively a preemptive action by B or whether it also involves the preventive motivation by A? These situations cannot be distinguished on behavioral grounds alone, for observed state behavior may be identical in the two situations. It is therefore necessary to examine the actual motivations of As decision makers in order to determine if they wanted a war to block the rise of B, yet preferred that B initiate the war. This problem is complicated by the possibility that a state may launch a purely aggressive war for nonpreventive reasons but, for diplomatic and internal political purposes, will try to justify it as preventive action that is necessary in the face of a rapidly rising adversary and the inevitability of a future war.

²⁶ Mutual preventive motivations could occur because of a faulty assessment, by at least one state, of existing military capabilities and their trends in the near future. One example might be the Franco-Prussian War, with France fearing a rising Prussia and Prussia being concerned about the impact of French army reforms. Another might be the Russo-Japanese War, where the rising Japanese may have feared the completion of the Russian railroad in Manchuria. See Howard, Michael, The Franco-Prussian War (London: Methuen, 1981), 54Google Scholar; Langer, William L., “The Origin of the Russo-Japanese War,” in Langer, , Explorations in Crises (Cambridge: Harvard University Press, 1969), 37, 40CrossRefGoogle Scholar. Mutual preventive motivations could also occur if two states evaluate military power from two different analytical frameworks. A global power may perceive the erosion of its global dominance by a rising regional power; the latter may simultaneously perceive the erosion of its regional dominance by an expanding global power. For example, the attack on Pearl Harbor may have been in part a preventive action by the Japanese to consolidate their regional superiority before the opportunity vanished; at the same time, the United States may have undertaken strong coercive measures in part to prevent the erosion of its global power position by an expansionist Japan.

²⁷ Levy, Jack S and Morgan, T. Clifton, “The Frequency and Seriousness of War: An Inverse Relationship? Journal of Conflict Resolution 28 (December 1984), 731–49CrossRefGoogle Scholar.

²⁸ In fact, we might expect to find the preventive and scapegoat motivations to occur together frequently, for they may be the product of the same underlying processes. Economic decline often generates social problems and political insecurity for the elite at the same time it undercuts the military power and potential of the state. It is often argued, for example, that German and Austro-Hungarian leaders were driven to war in 1914 by the combined fear of the decline of German military power abroad and the rise of social democracy and disorder at home. V. R. Berghahn concludes that the ruling elites in Germany “were increasingly haunted by the nightmare of impending internal chaos and external defeat so that an offensive war appeared to be the only way out of the general deadlock.” Berghahn, , Germany and the Approach of War in 1914 (New York: St. Martin's, 1973), 213Google Scholar. See also Fischer (fn. 8, 1975), 398; Ritter (fn. 8), 227–39. For more general analyses of the use of force externally for internal political purposes, see Mayer, Arno J., “Internal Causes and Purposes of War in Europe, 1870–1956: A Research Assignment,” Journal of Modem History 41 (September 1969), 291–303CrossRefGoogle Scholar; Levy, Jack S., “Domestic Politics and War,” Journal of Interdisciplinary History 18 (Spring 1988)CrossRefGoogle Scholar.

²⁹ Howard (fn. 2), 16; Morgenthau (fn. 2). As Richard W. Mansbach and John A. Vasquez argue, traditional realist theory emphasizes the “issue of power” and minimizes the “power of issues.” See their critique of realism in In Search of Theory (New York: Columbia University Press, 1981)Google Scholar. See also Keohane, Robert O. and Nye, Joseph S., Power and Interdependence (Boston: Little, Brown, 1977)Google Scholar.

³⁰ Recall that the preventive motivation is only one of several possible responses to declining military power and potential. A complete analysis would have to evaluate the relative costs and benefits of alternative policy options, particularly alliances. See fn. 13; also, Most, Benjamin A. and Starr, Harvey, “International Relations Theory, Foreign Policy Substitutability, and ‘Nice’ Laws,” World Politics 36 (April 1984), 383–406CrossRefGoogle Scholar.

³¹ For an argument reaching similar conclusions, see Russett, Bruce M., “Pearl Harbor: Deterrence Theory and Decision Theory,” Journal of Peace Research 4 (1967), 89–105CrossRefGoogle Scholar. Also see Betts. (fn. 24), 62.

³² In cases where the preventer is not fearful of being overtaken in the near future, certain critical thresholds of adversary strength may still be important (for example, Israel's concern 1982 about Iraq's acquisition of a nuclear capability).

³³ Thompson, , “Succession Crises in the Global Political System: A Test of the Transition Model,” in Bergesen, Albert L., ed., Crises in the World System (Beverly Hills, CA: Sage, 1983)Google Scholar, chap. 5,93–116, at 96; also, Organski (fn. 3), 376.

³⁴ Organski (fn. 3) and Thompson (fn. 33) suggest that a rapid rise may increase the challenger's incentives for war, but this is less significant than the effect of a narrowing power differential on the stronger state.

³⁵ On the importance of systemic interests for leading states, see Waltz, Kenneth N., Theory of International Politics (Reading, MA: Addison-Wesley, 1979)Google Scholar, chap. 8; Snyder, Glenn H. and Diesing, Paul, Conflict Among Nations (Princeton: Princeton University Press, 1977)Google Scholar, chap. 6; Gilpin (fn. 2). The incentives for prevention by a leading or “hegemonic” state will be particularly great since it has a disproportionate influence on the distribution of benefits and influence in the system. See ibid..

³⁶ Fischer (fn. 8 1975), 461. Though statesmen frequently use the term “inevitability,” they presumably mean a very high probability of war. For one of the few discussions of this important variable, see Lebow (fn. 9, 1981), 254–63. Note that a limited preventive attack involves fewer risks and may be more common in the absence of such fatalism.

³⁷ A major reason why the preventive motivation was important in the Seven Years' War in Europe was that the war was perceived as inevitable by Frederick. See Dorn, Walter L., Competition for Empire, 1740–1763 (New York: Harper & Row, 1963), 312–14Google Scholar; Vagts (fn. 15), 277; Mowat, R. B., A History of European Diplomacy, 1740–1763 (Hamden, CT: Archon, 1971), 244–46.Google Scholar The elder Moltke's belief in the inevitability of war undoubtedly contributed to his advocacy of preventive action against France in 1867 and against Russia in 1887 (Howard, fn. 26, p. 41). In the period prior to World War I, the perception by German and Austrian decision makers that war was inevitable (against France or Russia or both) was critical. German Chief of Staff Moltke, for example, said in 1912, “I believe a war to be unavoidable and: the sooner the better.” This fatalism was shared by political leaders and the general public as well as by the military. See Fischer (fn. 8, 196:), 29–38; Fischer (fn. 8, 1975), 162; Albertini (fn. 8), vol. II, p. 610; vol. III, p. 441; Ritter (fn. 8), 107; Geiss (fn. 8), 41; Fay, Sidney B., The Origins of the World War (New York: Macmillan, 1928), Vol. I, p. 43Google Scholar; Lebow (fn. 9, 1981), 255–57.

³⁸ In actuality, of course, victory and defeat constitute two ends of a continuous range of possible outcomes, each associated with an expected probability of occurrence and net cost or benefit. Note that the positive benefits from a victorious war may be a powerful incentive for war, but this would exist independently of declining power, and hence independently of the preventive motivation.

³⁹ This hypothesis explains why weaker states rarely succumb to the preventive motivation. If we control for the fear of the future or degree of anticipated decline, this hypothesis is valid for only a restricted range of probabilities. If the expected probability of victory is too large, reflecting an enormous military advantage for the preventer over a weak but growing opponent, there is no immediate threat to one's position and fewer incentives for preventive action.

⁴⁰ On the importance of expectations of intervention by third states, see Blainey, Geoffrey, The Causes of War (New York: Free Press, 1973)CrossRefGoogle Scholar, chap. 4; de Mesquita, Bruce Bueno, The War Trap (New Haven: Yale University Press, 1980).Google Scholar

⁴¹ Vagts (fn. 15, p. 290) argues that “preventive wars cannot easily be undertaken by an alliance or a coalition. Preventive wars are essentially wars of a single autocratic power acting alone or in absolute dominance within a coalition.” On the tendency for great powers to align against the leading state, to balance rather than bandwagon, see Waltz (fn. 35), 126–27; Walt, Stephen M., “Alliance Formation and the Balance of World Power,” International Security 9 (Spring 1985), 3–43.CrossRefGoogle Scholar The main exception is a situation in which the rising challenger expounds a revolutionary ideology which is perceived by other great powers as a threat to their own internal stability as well as to their military superiority (for example, revolutionary France in 1792). See also fn. 13.

⁴² On the concept of the offensive/defensive balance, see Levy, Jack S., “The Offensive/Defensive Balance of Military Technology and the Incidence of War: A Theoretical and Empirical Analysis,” International Studies Quarterly 28 (June 1984), 219–38CrossRefGoogle Scholar; Jervis, Robert, “Cooperation under the Security Dilemma,” World Politics 30 (January 1978), 167–214.CrossRefGoogle Scholar On the link between offensive doctrines and preventive war, see Van Evera (fn. 9), 64, and (fn. 11), 74–76; Posen, Barry, The Sources of Military Doctrine (Ithaca, NY: Cornell University Press, 1984), 69–70.Google Scholar The importance of this factor for preventive war should not be exaggerated, however, for the marginal impact of the offensive/defensive balance of military technology is greatest when the difference in military capabilities is small, but the preventive motivation is weakest under those conditions.

⁴³ On the consequences of misperception, see Levy, Jack. S., “Misperception and the Causes of War: Theoretical Linkages and Analytical Problems,” World Politics 36 (October 1983), 76–99CrossRefGoogle Scholar; Blainey (fn. 40), chap. 3; Lebow (fn. 9, 1981), 242–47. On the sources of misperception, see Jervis, Robert, Perception and Misperception in International Politics (Princeton: Princeton University Press, 1976)Google Scholar, and “Perceiving and Coping with Threat,” in Jervis, Lebow, and Stein (fn. 9), 13–33; Nisbett, Richard and Ross, Lee, Human Inference (Englewood Cliffs, NJ: Prentice-Hall, 1980).Google Scholar

⁴⁴ Risk refers to a situation in which the probabilities of different outcomes are known, whereas in a situation of uncertainty the probabilities of various outcomes are not known. Because of the inherent unpredictability of international behavior, uncertainty is more important than risk. I shall use risk orientation loosely to refer to attitudes toward uncertainty as well as risk. Technically, the analysis of the expected utility of preventive war and delay is inseparable from an analysis of orientation toward risk.

⁴⁵ Given a choice between a certain gain x and a lottery involving an expected value y > x (in typical experiments x and y differ by 20–30%), individuals generally choose x; but given a choice between a certain loss x and a lottery involving an expected value y < x, they choose to gamble and choosey y. These findings are robust and contrary to the assumptions of expectedutility theory. They are generally explained by the tendency to give disproportionately high weight to nearly certain outcomes (the “certainty effect”), by the asymmetry of losses and gains and the steeper slope of loss curves, and by the tendency to frame the choice in terms of the status quo rather than of an absolute or arbitrary standard. See fn. 14. Also see Hershey, John C. and Schoemaker, Paul, “Risk Taking and Problem Context in the Domain of Losses; An Expected Utility Analysis,” Journal of Ris% and Insurance 47 (March 1980), 111–32CrossRefGoogle Scholar; Jervis, Robert, “War and Misperception,” Journal of Interdisciplinary History 18 (Spring 1988).CrossRefGoogle Scholar

⁴⁶ This same experimental finding suggests that the rising state facing a near-certain future gain should be risk-adverse and prefer to avoid gambling on a risky war—contrary to Organski's argument (fn. 3) that the rising challenger will initiate the war.

⁴⁷ Brodie (fn. 15), 26.

⁴⁸ Fischer (fn. 8, 1975), 377,461; Vagts (fn. 15), 290–91. Bismarck's interpretation of this case appears to refer more to preemption than to prevention.

⁴⁹ Austrian Chief of Staff Conrad perceived that the nationality problem was dividing the army, and feared that in a country in which two-thirds of the population were Slavs, the army would soon no longer fight for Germans and Magyars. He questioned “whether we should wait until France and Russia were ready to attack us together or whether it was not more desirable to settle the inevitable conflict earlier.” Fischer (fn. 8, 1975), 398. See also Vagts (fn. 15), 304; Ritter (fn. 8), 229–31.

⁵⁰ Fischer (fn. 8, 1975), 371, 399; Turner, L.C.F., The Origins of the First World War (New York: Norton, 1970), 74Google Scholar. Domestic politics may affect the outbreak of war, of course, quite independently of the fear of one's deteriorating international position. See fn. 28.

⁵¹ Huntington, Samuel P., The Soldier and the State (New York: Vintage, 1957) chap. 3Google Scholar; Betts, Richard K., Soldiers, Statesmen and Cold War Crises (Cambridge: Harvard University Press, 1977)Google Scholar; Posen (fn. 42), chap. 2; Vagts (fn. 15), 263.

⁵² Edinger, Lewis J., “Military Leaders and Foreign Policymaking,” American Political Science Review 57 (June 1963), 392–405CrossRefGoogle Scholar; Betts (fn. 51); Epstein, Klaus, “Gerhard Ritter and the First World War,” Journal of Contemporary History 1 (July 1966), 193–210CrossRefGoogle Scholar; Goerlitz, Walter, History of the German General Staff, trans, by Battershaw, Brian (Westport, CT: Greenwood, 1975), 101Google Scholar.

On several occasions, Bismarck blocked military demands for preventive war. After he departed as chancellor in 1890, the German military gradually gained the upper hand. Although military pressures for preventive war were rejected in the crises of 1905, 1909, and 1912, these pressures had become much stronger in 1914. The political influence of the military was also greater in part because of the ignorance of military issues and planning by political leaders, and the resulting tendency to defer to the military on security issues. See Kennan, George F., The Decline of Bismarck's European Order (Princeton: Princeton University Press, 1979), 364–66Google Scholar; Craig, Gordon A., The Politics of the Prussian Army, 1640–1945 (Oxford: Oxford University Press, 1955)Google Scholar, chap. 7; Fischer (fn. 8, 1961, 1975); Geiss (fn. 8), 38–48; Berghahn (fn. 28), chap. 9: Levy, Jack S., “Organizational Routines and the Causes ofWar,” International Studies Quarterly 30 (June 1986), 193–222CrossRefGoogle Scholar.

⁵³ Divisions over preventive war may also follow generational lines, with the younger military officers being the most eager. Vagts (fn. 15, pp. 291, 306–8), argues that this was true for both Germany and Austria in the decades prior to World War I. See also Ritter (fn. 8), 308.

⁵⁴ This variable is potentially important but conceptually difficult, and is generally neglected. See Lebow (fn. 9, 1981), 247–54. Although the question whether democratic states are less war-prone than nondemocratic states has yet to be answered conclusively (though democracies almost never fight other democracies), it is not unreasonable to hypothesize that the preventive motivation for war is less intense in democratic states. See Chan, Steve, “Mirror, Mirror on the Wall…: Are the Freer Countries More Pacific?” Journal of Conflict Resolution 28 (December 1984), 617–48CrossRefGoogle Scholar.

⁵⁵ The war and capability data for the post–1815 period are from J. David Singer's Correlates of War project, available at the Inter-University Consortium for Political and Social Research at the University of Michigan. For war data spanning the last five centuries and a justification of the temporal extension of the system, see Levy, Jack S., War in the Modern Great Power System, 1495–1975 (Lexington: University Press of Kentucky, 1983)Google Scholar.

⁵⁶ Such an analysis would bear on the question of the sufficient conditions for war; a complete analysis would also have to consider necessary conditions and to include cases of unchanging power differentials.