1. See Hoffmann, Stanley, Contemporary Theory in International Relations (New Jersey, 1960), pp. 29–53Google Scholar; Bull, Hedley, ‘International Theory: The Case for A Classical Approach’ in Knorr, Klauss and Rosenau, James (eds.), Contending Approaches to International Politics, hereafter Contending Approaches (Princeton, 1969), pp. 20–38Google Scholar; Young, Oran, ‘Professor Russett: Industrious Tailor to A Naked Emperor’, World Politics, xx (1969), pp. 486–511.CrossRefGoogle Scholar

2. Alker, Hayward, ‘The Long Road to International Relations Theory: Problems of Statistical Nonaddivity’ in Kaplan, Morton (ed.), New Approaches to International Relations (New York, 1968), pp. 137–69Google Scholar; Singer, J. David, ‘The Incompleat Theorist: Insight Without Evidence’ in Contending Approaches, pp. 62–86Google Scholar; Small, Melvin and Singer, J. David, Resort to Arms: International and Civil Wars, 1816–1980 (Beverly Hills, Calif., 1982), pp. 14–77Google Scholar.

3. A notable exception is Robert Jervis’ systematic critique of the Stanford Group's content analysis. See Jervis, Robert, ‘The Cost of Quantitative Study of International Relations’ in Contending Approaches, pp. 177–217.Google Scholar

4. See Kaplan, Morton, ‘The New Debate: Traditionalism vs. Science in International Relations’ in Contending Approaches, pp. 46–49Google Scholar, esp. footnote 9. See also Kaplan's vitriolic reply to one of his more recent critics (Kenneth Waltz) in his Professionalism in International Theory: Macrosystem Analysis (New York, 1979), pp. 1–92Google Scholar.

5. See Russett, Bruce, ‘The Young Science of International Polities’, World Politics, xxii (1969), esp. pp. 90–92.Google Scholar

6. Oran Young, op. cit. p. 509.

7. A summary of The War Trap can be found in Bueno de Mesquita, Bruce, ‘An Expected Utility Theory of International Conflict’, American Political Science Review, Ixxiv (1980), pp. 917–931.CrossRefGoogle Scholar

8. Cited in Political Sciencefrom Yale, Yale University Press Booklet, 1981, p. 2Google Scholar.

9. Zagore, Frank, American Political Science Review, Ixxvi (1982), p. 739Google Scholar.

10. See Campbell, Donald and Stanley, Julian, Experimental and Quasi-Experimental Design for Research (Chicago, 1963), p. 5Google Scholar.

11. Singer, J. David (ed.), Quantitative International Politics (New York, 1968), pp. 51–158Google Scholar. Dina Zinnes has written that the primary purpose of these works is hypothesis testing, to ‘draw conclusions about what is (or was) ’. See Zinnes, Dina, ‘A Consumer's Guide to Texts on Mathematical Modeling’, World Politics, xxxi (1979), p. 435Google Scholar. In Kenneth Waltz's terminology this is equivalent to establishing laws. Hypothesis testing is to be distinguished from mathematical modeling, which is Mesquita's strategy. As Zinnes correctly observes, mathematical modeling's ‘avowed goal is to explain…answers to the why's of what is observed’. Ibid. cf. the recent issue of International Political Science Review, iii (1982)Google Scholar which is devoted to ‘Mathematical Approaches to International Relations’.

12. J. David Singer (ed.), op. cit., Quantitative International Politics, pp. 246–86.

13. Ibid. p. 283.

14. von Neumann, John and Morgenstern, Oskar, Theory of Games and Economic Behavior (Princeton, 1944), p. 16Google Scholar.

15. Ibid. pp. 1–33.

16. Cf. Ibid. pp. 16–20 with Mesquita, pp. 29–32.

17. Mesquita uses the definition of war provided by Singer, J. David and Small, Melvin, The Wages of War (New York, 1972), p. 381Google Scholar. For definitions of intervention and threats, see Mesquita, pp.. 98–9.

18. Kenneth Waltz makes a nice distinction between laws and theories. He maintains that ‘laws establish relations between variables’ whereas ‘rather than being a collection of laws, theories are statements that explain them’. See Waltz, Kenneth, Theory of International Relations (Reading, Mass., 1979), pp. 1Google Scholar and 5. Mesquita does not make such a distinction, he uses ‘law like statements’ and ‘theory’ interchangeably. His notion of theory is more like that of Marion Levy, who sees scientific theories as ‘highly generalized propositions containing variables with powerful deductive interdependencies…’.See Marion Levy, ‘Does It Matter If He's Naked? Bawled the Child’ in Contending Approaches, p. 92.

19. See New York Times, The Pentagon Papers (New York, 1971), pp. 382–458.Google Scholar

20. This important conflict is not included in Mesquita's set of twentieth century conflicts.

21. For a similar example, see Downs, Anthony, An Economic Theory of Democracy (New York, 1957), esp. pp. 21–35Google Scholar. In a recent article Dunleavy and Ward have criticized Downs’ (and by implication Mesquita's) ‘Friedmanite’ way of handling assumptions as ‘naive verification-ism’. See Dunleavy, Patrick and Ward, Hugh, ‘Exogenous Voter Preferences and Parties With State Power: Some Internation Problems of Economic Theories of Party Competition’, British Journal of Political Science, xi (1981), p. 351CrossRefGoogle Scholar.

22. Cf. von Neumann and Morgenstern, op. cit. pp. 16–20.

23. For a statement on the pitfalls of theories and approaches which exclude from their purview analysis of the rationality of ends, see Hoffmann, op. cit. pp. 34–5.

24. Waltz, op. cit. p. 28, makes a similar point.

25. See Przeworski, Adam and Teune, Henry, The Logic of Comparative Social Inquiry (New York, 1970), pp. 4Google Scholar, 17 and 20–3.

26. Victim or defender, as used by Mesquita, merely means the state that was first attacked. It says nothing about whether the victim wins or loses the conflict.

27. The equations are presented in the form of words to make for easier reading. Readers interested in the precise mathematical equations should consult Mesquita, pp. 56–64.

28. Stoessinger, John, Why Nations Go To War (New York, 1978), pp. 231–3.Google Scholar

29. Jervis, Robert, Perception and Misperception in International Politics (Princeton, 1976), p. 378Google Scholar.

30. Ibid.

31. Ibid. p. 379.

32. See Singer, J. David (ed.), Correlates of War: I (New York, 1979)Google Scholar and Correlates of War: II (New York, 1980)Google Scholar. See also Part 1 of J. David Singer and Associates, Explaining War (Beverly Hills, Calif., 1979).Google Scholar

33. All calculations in this paper are based on Mesquita's data as presented in Table 2. Figures are rounded for convenience.

34. This figure is computed in the following way: 14 + 24 + 32/3 = 23.

35. Hildebrand, David, Laing, James and Rosenthal, Howard, ‘Prediction Analysis in Political Research’ American Political Science Review, Ixx (1976), p. 510Google Scholar.

36. Ibid.

37. Mesquita establishes ‘levels of significance’ for some of his calculations (p. 126) but these computations appear to bear out his theory well enough to make it unnecessary for him to introduce any qualifications.

38. I would like to thank my anonymous referee for pointing this out.

39. See Popper, Karl, The Logic of Scientific Discovery (London, 1959)Google Scholar. Cf. Lakatos, Imre, ‘Falsification and the Methodology of Scientific Research Programmes’ in Lakatos, Imre and Musgrave, Alan (eds.), Criticism and the Growth of Knowledge (Cambridge, 1970)CrossRefGoogle Scholar and Lijphart, Arend, ‘Comparative Politics and the Comparative Method’, American Political Science Review, Ixv (1971)Google Scholar, esp. 692. In addition to suggesting the notion of ‘theory infirming’, Lijphart also suggests presenting theories in probabilistic terms. Ibid. pp. 686–7. Lakatos, however, is skeptical of this approach. He writes: ‘Replacing proof by probability…turned out to be insufficient. It was soon shown by Popper…that under very general conditions all theories have zero probability, whatever the evidence; all theories are not only equally unprovable but also equally improbable. ’ See Lakatos, op. cit. p. 95.

40. See Singer, J. David and Small, Melvin, ‘Alliance Aggregation and the Onset of War, 1815–1945’ in Singer, J. D. (ed.), Quantitative International Politics; Henry Kissinger, The Necessity of Choice (New York, 1960)Google Scholar; and Hoffmann, Stanley, ‘International Systems and International Law’ in Knorr, Klauss and Verba, Sidney (eds.), The International System (Princeton, 1961).Google Scholar

41. Oran Young has argued persuasively against an excessive concern for theories with policy-relevance because of intellectual, ethical and moral difficulties. See Young, Oran, ‘The Perils of Odysseus: On Constructing Theories of International Relations’ in Tanter, Raymond and Ulman, Richard (eds.) Theory and Policy in International Relations (Princeton, 1972), pp. 199–203Google Scholar.

42. To clarify matters, an example of my recomputation is necessary. The figures for column two, row one (wars from 1816–1899) are 30, 4, 3, 29 and they are translated to 0.88, 0.12, 0.09 and 0.91 respectively in Table 3. 0.88, for instance, is obtained by 30/34: this means that the figure indicates the proportion of positive utility states which initiated conflicts. This is slightly different from another possible way to compute: 30/33, which gives the proportion of conflicts initiated by positive utility states.

43. Coding this way increases the number of positive is, thereby lending more support, in terms of absolute numbers, to Mesquita's generalization. However, this also undermines Mesquita's assertion that there is no difference between the two centuries. It shows that like war and interventions, the proportion of negative utility states initiating threats is increasing. Although there are a few cases like the United Kingdom and China elsewhere (interventions), no readjustment is necessary in these cases because they do not affect the main direction of the graphs. Note that the broken lines in the graphs represent the new values obtained by our coding procedure.

44. Brodie, Bernard, ‘War in the Atomic Age’ in Brodie, Bernard (ed.), The Absolute Weapon (New York, 1946), p. 76Google Scholar.

45. Cited in Osgood, Robert and Tucker, Robert, Force, Order and Justice (Baltimore, 1967), p. 15Google Scholar.

46. On this point see especially Hoffmann, Stanley, The State of War (New York, 1965), pp. 233–6Google Scholar; Robert Osgood and Robert Tucker, op. cit. pp. 123–8; Mandelbaum, Michael, ‘International Stability and Nuclear Order: The First Nuclear Regime’ in Gompert, David et al., Nuclear Weapons and World Politics (New York, 1977)Google Scholar, and his The Nuclear Revolution: International Politics Before and After Hiroshima (Cambridge, 1981), pp. 8–9Google Scholar. Note that for our purposes, nuclear weapons need not have changed the international system, it is sufficient that they introduce a significant amount of restraint in decision making.

47. Mesquita infers the decision making pattern from the decisions made. We do the same here.

48. Bueno de Mesquita, op. cit. p. 929.

49. Waltz, op. cit. pp. 116–23.

50. Kissinger, Henry, White House Years (Boston, 1979), p. 195Google Scholar.

51. Kissinger, Henry, A World Restored (New York, 1964)Google Scholar.

52. Moore, Barrington, Social Origins of Dictatorship and Democracy (Boston, 1966), pp. xi–xviiGoogle Scholar.

53. See Feis, Herbert, The Road to PearlHarbor (Princeton, 1950), pp. 282–6Google Scholar, 327–30.

54. Ibid. See Togo, Shigenori, The Cause of Japan (New York, 1956)Google Scholar for the assessment of Japan's war ability against the United States.

55. See Halperin, Morton, Bureaucratic Politics and Foreign Policy (Washington, DC, 1974), pp. 166–8.Google Scholar

56. Ibid.

57. See Levy, op. cit.

58. Ibid.

59. Feyerabend, Paul, Science in a Free Society (Norfolk, 1978), p. 111Google Scholar.