The idea that democracies do not go to war with other democracies, popularized by Princeton University historian Michael Doyle in 1986, has become a dangerous political cliché. It is dangerous partly because of its effect on international relations—the Clinton administration seems to believe that democracybuilding equals peace, with dubious results from Haiti to Kosovo—but mostly because, by begging important questions about the nature of democracy, it leads us onto treacherous ground.

In his thoughtful and penetrating book, Snyder, a political scientist at Columbia University, analyzes imperial Britain, revolutionary France, Germany from Bismarck to Hitler, and 19th-century Serbia. Nationalism, he concludes, is a regular and sometimes monstrous feature of young democracies. These nations suffer a wild and even vicious youth; indeed, "the process of democratization can be one of its own worst enemies." Nominal democracies without civic institutions and a sturdy middle class are especially vulnerable to nationalist demagogues. "If nationalist conflict is to be avoided," Snyder writes, "the development of civic institutions should be well underway before mass-suffrage elections are held. Likewise, it is better if a strong middle class emerges before press freedom expands and civil society groups get organized, or else these may be easily hijacked by an elite with a nationalist agenda."

At a time when support of democracy is almost reflexive, these are startling prescriptions. But Snyder makes a powerful case, one with which ancient Greeks and classical-minded Enlightenment figures such as Burke and Gibbon would have been familiar. Although the Russian election came too late for the author’s deadline, his thesis helps explain why so many Russian reformers have soft-pedaled their democratic aspirations to back Vladimir Putin’s attempt to restore a strong, centralized state.

Whereas Snyder concentrates on the problems of fledgling democracies, Deitchman considers the United States. The lonely superpower, perforce responsible for global stability, is also a media-saturated democracy that is sensitive to casualties, views the United Nations with suspicion, and expects the world to be grateful. Deitchman, formerly with the Institute for Defense Analyses, a research firm based in Alexandria, Virginia, starts with three credible scenarios. First, he asks whether the United States would really risk nuclear war with China to protect Taiwan, and he ponders the price in lost credibility if Taiwan were abandoned. Second, he asks how the United States would react to an Islamic fundamentalist rebellion against the Saudi monarchy. Finally, he wonders whether an overstretched U.S. military could still retake the Panama Canal if it fell under the control of a narco-dictatorship. He concludes that a diminished military, increasingly distant from American society as a whole, greatly complicates the effective exercise of might.