War-Proneness, War-Weariness, and Regime Type: 1816-1980

Abstract

Previous empirical research has demonstrated that war-weariness is not a universal characteristic of na tion-state behaviour. But a scholarly consensus agrees, for reasons originally advanced by Kant, that war-weariness is more applicable to democratic than to nondemocratic states. This is true despite evi dence that democracies are not more peaceful than autocracies — with the notable exception that war fare virtually never occurs between democracies. Empirical analyses of war (interstate and extrasy stemic) and dispute data from the Correlates of War Project reveal no statistically significant (p < .05) relationship between war-weariness and the conflict behaviour of the three principal democracies (France, United Kingdom, and United States) during the nineteenth and twentieth centuries. Some pat terns of behavior, especially the severity of subsequent wars, do appear to be in the predicted direction. However, the coefficients are so small that even if the number of cases were large enough to make the results statistically significant, only a minute fraction of the variance in international violence would be explained. These results, and those of other researchers, refute war-weariness as a general hypothesis. But they do not disprove the impact of war-weariness in particular cases such as France in 1940 and the United States in 1954.