There has been no lack of theories on the cause of war. But we do lack theories that hold up when tested against the facts of history. This deficiency of all existing theories has led a group of scholars to try to reverse the typical way of arriving at an explanation for war. Instead of coming up with a theory and then looking for the evidence, they have decided to look first at the evidence. Their first undertaking was to collect the most precise information possible about wars, their length, destructiveness, and participants. But before they could do even this they needed careful definitions of terms, so it would be clear which events belonged in the category of 'war', when a state could be considered 'participating in a war', what in fact a 'state' was, and so on. Like all definitions, theirs were somewhat arbitrary, but they carefully justified their choices and, more important, they drew up their definitions first, before arriving at their conclusions so that they could not be accused of defining events in a way that would prove their presuppositions. After agreeing on definitions, they set out to collect data. Even though they confined themselves to wars fought in the last 150 years, they encountered difficulties in getting precise information on items such as the number of casualties. Nevertheless, they argue, their results are better than any that preceded them. These basic facts about wars were published in a handbook, The Wages of War 1865—1965, edited by two leaders of the project, J. David Singer and Melvin Small. Even though this is only the beginning of the project, it already provides some answers to questions about wars. You might hear a street corner preacher tell you that the end of the world is at hand, because the number of wars is increasing just as the Bible prophesies. If you want to check the validity of such an assertion, you could turn to The Wages of War and answer the question using the best available data. The next step in the project is to identify conditions or events that seem to be associated with wars. They are not looking for explanations, but just for correlations, that is, items that usually accompany each other. It is for this reason that they have named their project 'The Correlates of War'. Starting with their collection of data on wars, they could examine the hypothesis of Woodrow Wilson that autocracies are the cause of wars. If this were true, then autocracies would fight other autocracies and democracies might fight autocracies in defense, but democracies would never fight democracies. After defining 'democracy' in a way that could be measured (for example, the frequency with which officeholders change office) they would see if any of the wars they had identified in the last 150 years had been fought between two countries clearly identifiable as democracies. If they could find no such wars, they could say there was a correlation between democracy and peace. It would not yet be a proof that autocracies cause war. There could be other explanations—the world might contain only one or two democracies. But a correlation would be an important first step. The Correlates of War project is just entering this second stage. It will be some time before a full theory appears. Even when the project does produce a theory of war (if it finds evidence to warrant such a theory), it may not provide the final word on the subject. Any such project must make decisions early in the research, such as what counts as a war and what does not. These decisions can crucially affect the outcome, even though it might not be evident for a long time that they will. Here is an example of this problem. The Correlates of War project counts the wars fought by Prussia under Bismarck as three separate wars because each stopped before the next one started. On the other hand, Hitler's belligerent moves against neighboring countries in 1939 and 1940 (Poland, Denmark, Belgium, France, Norway)