Abstract: 
Great powers are entering an era of deep disarray, but without the theoretical tools necessary to avert armed conflict. Foreign affairs experts still lack the ability to predict the occurrence of war with a high degree of accuracy. While realism, liberalism, and constructivism are helpful tools, they all fail to achieve the ultimate goal of any science: the power of prediction. This is due to a failure to systematically address the question of why wars start. Using rationalist explanations for war and bargaining theory as the foundation, this thesis proposes a perception square model which seeks to predict dyadic conflict. This thesis demonstrates that war and peace are functions of the perceptions of policymakers. Put plainly, wars happen primarily because leaders’ perceptions of each other are incompatible with a negotiated peace. A review of 12 historical cases buttresses the model’s explanatory power. By analyzing the perceptions of leaders of two countries and categorizing the perception situation according to the model, policymakers can predict war with a high degree of accuracy. When applied to a potential U.S.-China war over Taiwan, the model shows that relations are extremely unstable and headed towards armed conflict.