REALISTS ARE hawks when they see an adversary as a deliberate, implacable, risk-prone aggressor bound to strike if not deterred or defeated by a formidable counterthreat. The perfect example of this is Nazi Germany—but only as the Western understanding of it solidified after 1938. Realists are doves when they believe the adversary has limited ambitions that can be accommodated or contained at moderate cost. This was the view of many about Germany prior to the Munich agreement. Until the eve of war, too many realists, like others who thought Winston Churchill alarmist, found excuses for Adolf Hitler’s demands and rationales for compromising with him. As the Munich crisis unfolded Sir Neville Henderson, the British ambassador in Berlin, wrote back to the Foreign Office: “One must also try to understand the German point of view. If we were in Germany’s place what would we, in the midst of all this war psychosis, be doing: exactly what I think the Germans are today doing.” It is a discomfiting irony that one of the most arresting expositions of realism ever written, E. H. Carr’s The Twenty Years’ Crisis, was an implicit brief for appeasement when it went to press shortly before the war.

Chamberlain’s approach was the path of least resistance, in a nation buffeted by memories of the Great War. Hindsight at the time suggested that the catastrophe of World War I might have been avoided if statesmen had shown more empathy for their adversaries, more willingness to take risks for peace and, above all, more military restraint as the summer crisis of 1914 worsened. Chamberlain and company avoided the fateful mistake of 1914 only to commit the opposite one.

As 1938 exemplifies the image of misreading pure aggression, 1914 exemplifies the image of the “security dilemma.” This concept, formulated by John Herz and further developed by Robert Jervis and others, is the tragic view of international conflict. It assumes that countries are usually opportunistic but averse to paying the high cost of war to make gains, yet cannot trust that their opponents’ aims are similarly limited. So both sides arm to deter or defend, but see each other’s precautionary mobilization as preparation to attack. Acting to maximize one’s own security threatens the opponent’s security. Neither may want war, but they may stumble into it out of misperception, miscalculation and fear of losing if they fail to strike first. In short, if the security dilemma is the real threat, then restraint rather than provocative military action should be the default option.

The challenge of crisis management since the first half of the twentieth century has essentially been to balance the risks in these competing images of threat. Hawks attend too much to the danger of unprovoked aggression, doves too much to the danger of misperception and inadvertent escalation between innocent parties. John F. Kennedy provided a solid example of balancing the contrasting risks in his handling of the Cuban missile crisis, coercing Nikita Khrushchev forcefully but holding back from precipitous action at crucial points and making limited concessions to salve Soviet pride. (Kennedy had recently read Barbara Tuchman’s The Guns of August, and he told his brother at the height of the crisis, “I am not going to follow a course which will allow anyone to write a comparable book about this time, The Missiles of October.”)

Deciding whether the greater danger is unprovoked aggression or accidental war in any given situation is about judging enemy motives, intentions, reactions and decisions. On dealing with enemy power, however, realists unite. Whether in a security dilemma or not, realists do not voluntarily accept vulnerability. They favor making whatever alliances are potent enough to assure deterrence or defeat of the enemy—and this is where they part company with those who condition action on moral and legal norms. Consider two examples in 1939 of faltering in power politics.

The realist solution that should have been pursued before the Munich crisis if Hitler’s unlimited ambitions had been appreciated, and at the least should have been undertaken in the months between Germany’s absorption of Czechoslovakia and August 1939, would have been an alliance of the Western democracies with the Soviet Union to outweigh Germany. That combination would have been problematic on both practical and moral grounds, and it is useful to remember that until World War II unfolded the record of Stalin’s regime was even more murderous than Hitler’s. But such an alliance might have averted the greatest man-made catastrophe in history. The three countries considered it, but when the French and British hesitated, the surprise Nazi-Soviet nonaggression pact took away that option. Only Hitler’s greatest miscalculation, attacking the USSR two years later, brought it about and enabled the defeat of Nazism—but now at an astronomically higher price.

In the interim, however, the democracies narrowly avoided a moral but suicidal urge to wage war against Moscow. When the Soviets invaded Finland shortly after they divided Poland with Germany, the British and French supported Moscow’s expulsion from the League of Nations and would have come to Finland’s aid militarily if they had not failed to get permission for their troops to pass through Norway and Sweden. Imagination reels at the thought of what would have happened had Britain and France put themselves at war with Germany and the Soviet Union at the same time. The motive was lofty, the potential consequences disastrous. Hans Morgenthau used this example in his classic Politics Among Nations to epitomize the foolishness of letting legal principles obscure the strategic imperatives imposed by the balance of power.

Who were the realists and idealists in all of this? Were the opponents of appeasement idealists? Neoconservatives today lay claim to Churchill as a model to be held up in opposition to today’s realists, whom they see as wimps on Iran and craven in their unwillingness to act against evil in Syria and Ukraine. Yet Kenneth Thompson, soulmate of Morgenthau, could write a whole book, Winston Churchill’s World View, interpreting Churchill as a consummate realist. Churchill and Chamberlain may have been equally realist in philosophical disposition, but they diagnosed the threat differently. Or were Chamberlain and company idealists in contemplating coming to the aid of Finland, and Churchill idealist in determining not to sue for peace after the fall of France, when Britain stood alone with no apparent prospects for winning at the time? Looking for lessons to apply to the present, two points stand out. First, trying to pigeonhole real-world statesmen as realist or not is pointless. Second, what can divide either realists or idealists, and cause coalitions across the theoretical divide, is their judgment about the severity of the threat in question—that is, their estimates of the consequences of opposed options.

WHEN RUSSIAN intervention stopped the advance of Ukrainian government forces against separatist rebels in 2014, many liberals in the West campaigned to give much heftier military aid to Kiev than the limited training and “nonlethal” supplies their countries had provided thus far. What could be wrong with helping victims repel aggressors? The realist counterargument begins by downplaying the question of which side deserves support according to standards of justice. Instead, it focuses on the questions of what will in fact happen if the West gives a big boost to Kiev’s military forces, and whether the main risk in the Ukraine conflict is more akin to 1914 or 1938.

If Russia threatened to conquer all of Ukraine, then there would be nothing to lose in supplying arms to Kiev. However, Moscow has refrained from conquering as much of Ukraine as it could have, and instead has established a stalemate over the limited separatist enclave. Is it likely that Moscow will allow its separatist clients to be defeated if the Ukrainian army is made stronger? Will Moscow allow Russian supporting forces to be driven out of eastern Ukraine? Will Moscow shrink from doing whatever is necessary to preserve the enclave, as it preserved the secession of South Ossetia and Abkhazia in its little war with Georgia? The answer to all of these questions is no, and in any contest of military escalation between Ukraine and Russia, Moscow is bound to prevail.

If the result of stronger military assistance to the morally right side (Kiev) is more intense combat, higher casualties and more destruction, but without victory, it will count as pointless suffering. And is it any likelier that stronger Ukrainian military action will defeat the rebels and Russians than that such action will drive the Russians to strike back harder and expand the area under rebel control, creating a “land bridge” to Crimea? In either case, the result would be worse than the episodic limited combat of 2015. (Realists also see more ambiguity about the current Kiev government’s claim to legal right, since it came to power through a coup or revolution—take your pick—against elected president Viktor Yanukovych.)

It may be too late to reverse the separation of the eastern Ukrainian enclave, as it is probably too late to reverse the loss of Crimea. If not, most realists would say that unless NATO goes to war against Russia to restore Ukraine’s integrity—an option out of the question in the West—the only hope for a settlement is a negotiated compromise that links reintegration of the enclave in Ukraine to recognition of Russian interests in the country. As even many liberals have accepted, this would involve ample autonomy for eastern Ukraine, agreement to refrain from integrating the country as a whole into the West via institutions such as NATO or the EU, and symbolic internal status such as recognition of Russian as a second national language—all of which amounts in effect to a “sphere of influence” for Russia. This would be a defeat for Kiev but a smaller one than consolidation of outright secession by the separatist region would be. The idea of a sphere of influence is repugnant to liberal moral principles but has been a recognized prerogative of great powers historically. It is also one the United States has taken for granted in its own hemisphere.