The study itself raises other questions—some wittingly, some not. For instance, all the charts and timelines contain the following footnote: "The graph does not include ethnic or other conflicts where neither warring party was a state, nor does it include cases of 'one-sided' conflict, such as genocide." If it's true, as some scholars maintain, that international politics will become increasingly dominated by actors that are not states (terrorists, ethnic enclaves, etc.), this study's methodology has little to say on the future of warfare.

As for genocide, one graph in the study, Figure 1.11, shows that the number of genocides has dramatically declined since 1990 (from 10 instances to two). But, as the authors acknowledge, it's unclear that this means. There are no good data on the number of deaths caused by genocide. (More people may have been killed in one, Rwanda—800,000, by most estimates—than in a hundred smaller genocides.) Data began to be collected in 2002. The authors trumpet the fact that fewer people were killed by genocide in 2003 than in 2002, but nothing can be inferred from two data points.

The report notes that wars are producing steadily fewer battlefield casualties than was once the case. Millions died in the two world wars and in the Vietnam War. Those were prolonged battles, involving tanks, heavy guns, and massive aerial bombing. Wars fought nowadays tend to be "low-intensity conflicts," involving fewer soldiers and more limited aims. But this trend is hardly irreversible. A war in the Middle East or a serious civil war in Iraq could cause hundreds of thousands of deaths. If a war or terrorist strike ever involves nuclear weapons, millions might be killed, depending on the targets.

The report contains some data suggesting that parts of the world are growing more militarized. For instance, from 1975-96, the ratio of security forces to population has grown by 81 percent in Sri Lanka, 71 percent in India, 65 percent in Pakistan, 63 percent in China, 42 percent in Burma, and 29 percent in Thailand. These trends don't make conflict inevitable; they don't even necessarily make it more likely; but they're not worth celebrating, either.

So, what makes this report important? First, it may well be the most comprehensive compendium of data available anywhere on the patterns and history of armed conflict. Second, it shows conclusively that systems of international control, devised to keep the peace between great powers (whether the Cold War's nuclear standoff between the United States and the Soviet Union or the Treaty of Vienna's balance of power among the five major nations of Europe in the early 19^th century), do not stave off—and, in fact, may promote—wars involving smaller powers. However, the report's data do not support the notion (however much the authors may wish otherwise) that peace flourishes in the absence or aftermath of these great-power control-systems.

Still, those data raise intriguing questions about what does cause war and peace. That graph in Figure 5.2, showing the wavelike rise and fall in the number of wars across the centuries—they seem almost mechanical, these waves, like forces of nature: war followed by peace followed by war followed by peace … But is this pattern inevitable, or can it be controlled? The report's final section notes a surge in peacekeeping activities—by the United Nations and by regional institutions—since the end of the Cold War. Some of these efforts have been disastrous (Rwanda, Somalia), but others have met success (El Salvador, Namibia, Mozambique). More than at any other time in modern history (due in part to the rapidity of global communications and transportation), the major powers have the means and methods to dampen conflicts. The question is, do they have the desire?

In short, the report can serve as the basis for two kinds of exploration. Scholars should dig into the database to examine just why the incidence of warfare has risen and fallen at such regular intervals. Diplomats should see if they can—through their own cooperative actions—keep those waves at bay.