It is unlikely that George W. Bush feels constrained by international law when deciding whether to use military force abroad. Nevertheless, many of the United States' allies are reluctant to cooperate with and participate in military actions that cannot reasonably be justified under international law. And supportive allies, while perhaps not strictly necessary to the United States in its recent and foreseeable military campaigns, do make the military option easier to pursue. A war against Iraq would be difficult without access to bases and airspace in countries as diverse as Turkey, Saudi Arabia, Germany, and Canada. For this reason, at least, it would seem to be worth the president's while to adhere to international law where possible and, where this is not possible, to seek to change the rules.

Political science attends to causes and consequences of war but only fitfully welcomes study of its conduct, because few grasp how much the dynamics of combat shape politics. Bernard Brodie called for development of strategic studies on the model of the discipline of economics, because neither the military nor academia treated the subject rigorously. His call was answered in the early cold war, with mixed results. Theories about nuclear deterrence burgeoned while empirical studies of war lagged. The late—cold war impasse in nuclear strategy, rooted in NATO doctrine, shifted attention to conventional military operations and empirically grounded theory. Since the cold war, research on general theoretical questions about war and peace has been prospering, but education in military matters has been eroding. Interdisciplinary strategic studies integrate political and military elements of international conflict, but there is no recognized discipline of military science; military analysis is smuggled into political science and history departments, where it is resisted by calls to conceptualize security broadly or focus on purely theoretical work. If serious military studies are squeezed out of universities, there will be no qualified civilian analysts to provide independent expertise in policy and budget debates, and decisions on war and peace will be made irresponsibly by uninformed civilians or by the professional military alone.

For over three decades the North Atlantic Treaty Organization (NATO) has based its deterrent on the principle that the United States would retaliate with nuclear weapons if a Soviet conventional attack against Western Europe succeeded. This notion has long troubled most strategic analysts. It remained generally acceptable to political elites, however, when U.S. nuclear superiority appeared massive enough to make the doctrine credible (as in the 1950s); when the conventional military balance in Europe improved markedly (as in the 1960s); or when détente appeared to be making the credibility of deterrence a less pressing concern (as in the 1970s). None of these conditions exists in the 1980s, and anxiety over the danger of nuclear war has prompted renewed attention to the possibility of replacing NATO's Flexible Response doctrine (a mixture of nuclear and conventional deterrence) with a reliable conventional deterrence posture that might justify a nuclear no-first-use (NFU) doctrine.1

Strategic intelligence failures cannot be prevented by organizational solutions to problems of analysis and communication. Analytic certainty is precluded by ambiguity of evidence, ambivalence of judgment, and atrophy of institutional reforms designed to avert failures. Many sources of error are unresolvable paradoxes and dilemmas rather than curable pathologies. Major failures in attack warning, operational evaluation, and intelligence for strategic planning are due primarily to leaders’ psychological attributes rather than to analysts’ failures to detect relevant data. Since analysis and decision are interactive rather than sequential processes, and authorities often hear but dismiss correct estimates, intelligence failure is inseparable from policy failure. Solutions most often proposed—worst-case analysis, multiple advocacy, devil's advocacy, organizational consolidation, sanctions and incentives for analysts, and cognitive rehabilitation—are either impractical because of constraints on the leaders’ time, or they are mixed blessings because they create new problems in the course of solving old ones.