Israel’s assassination of a top Hamas leader in Tehran in July, Ukraine’s incursion over the summer into Russia, and a recent series of increasingly assertive Chinese air and maritime interceptions in the South China Sea have fanned fears that long-simmering conflicts could escalate into broader wars. In the wake of these provocations, analysts fret about the heightened risk of military accidents and strategic misperceptions. They worry that incidents of this sort could ratchet up tensions until policymakers lose control and stumble into wars they do not intend to fight. As U.S. Secretary of State Antony Blinken said in August, attacks in the Middle East “raise the risk of dangerous outcomes that no one can predict and no one can fully control.”

Although provocative incidents can push crises up the escalation ladder, truly inadvertent wars are rare. History provides few examples of conflicts that have erupted without policymakers’ authorization, and leaders frequently exercise restraint to avoid combat, especially in high-stakes situations. During the 1962 Cuban missile crisis, for example, U.S. policymakers held off on retaliating after Soviet troops shot down an American spy plane, stepping back from the brink of war. When faced with the risk of a spiraling conflict, rivals often find off-ramps to de-escalate crises. This brinkmanship requires careful choreography: states must learn how to pressure their adversaries just enough to shape their behavior without crossing thresholds that could trigger a significant response.

Even then, crossing redlines does not make conflict inevitable. The death of three U.S. soldiers in an Iranian-backed drone strike in January did not start a war between Washington and Tehran. In April, the massive drone and missile attack that Iran launched against Israel did not ignite a full-scale conflict between those two states. But to avoid war, leaders on both sides must restrain themselves at moments of crisis without losing face or showing weakness. To do so, they must carefully consider their actions—how, when, and where to pressure rivals in ways that avoid triggering escalatory retaliation. They must also establish direct or indirect communication with adversaries, facilitating arrangements that permit both sides to claim success in their coercive actions while reducing the potential for misinterpretation. Understanding how to navigate the interplay of pressure and restraint empowers leaders to step back from the edge of war.

THE SECRET HISTORY

The fear of inadvertent escalation is not new to international relations. Political scientists have spent decades arguing about whether military mobilization plans caused European states to “sleepwalk” into World War I. During the Cold War, policymakers worried that weapon malfunctions, false alarms from early warning systems, and unauthorized actions by military officers could spark a nuclear conflagration. Some academics have explored how unintended wars could unfold from technical failures in military systems. Others have suggested that states stumble into conflicts when military actions create momentum that makes it impossible for political leaders to back away from the brink. Still others have argued that leaders might respond with major military strikes if they mistakenly perceive a rival’s limited actions as an existential threat.

Although scholars describe different pathways to inadvertent war, their frameworks have a common trait: the assumption that policymakers have limited control over escalation. According to these researchers, states end up in wars they did not choose to fight because of chance or chain reactions in the military. But this does not accord with reality. Even during the tensest moments of the Cold War, the United States and the Soviet Union never accidentally fell into conflict. Instead, leaders always found a way out.

The Cuban missile crisis was a close call. Soviet air defenses had shot down a U.S. spy plane over Cuba without approval from Moscow, and the United States considered responding with retaliatory airstrikes that could have led to war. U.S. President John F. Kennedy and his joint chiefs, however, refrained from retaliating for fear that airstrikes could start a nuclear exchange. During another exceptionally tense moment, in 1983, the Soviet Union mobilized forces after mistakenly assessing a NATO military exercise as Western preparation for a nuclear first strike. But senior U.S. commanders again held off on responding. In each of these cases, policymakers ultimately decided to step back from war, recognizing the potentially catastrophic implications of escalation.

BLURRED LINES

Rival countries routinely engage in brinkmanship during crises, taking risky actions that heighten the prospects of war. The reason why is obvious: doing so can push a rival to change its behavior. Even if that doesn’t work, by ratcheting up tensions, leaders demonstrate that they are committed to achieving their objectives. Provocations such as raids, aerial interceptions, and ground incursions signal leaders’ willingness to act against adversaries and suggest that additional actions will follow if rivals don’t accede to their demands.

But provocations are inherently dangerous. Chinese fighter jets, for example, frequently use risky maneuvers when intercepting U.S. reconnaissance planes, making collisions more likely. The unpredictable nature of these actions heightens the risk of accidents, miscommunication, or misjudgments that could lead a minor incident to spiral into a broader conflict.

What makes crises so unpredictable is that the thresholds, or redlines, that might trigger a war are often not publicly known. They also do not fall into neat categories. They can be geographic: attacks in certain locations, for example, will trigger escalation, whereas strikes elsewhere may be ignored. But they can be based on the type of target, as well. Attacks on military contractors may fall below the threshold for retaliation, for instance, but attacks that kill military members may trigger a sharp response. The intensity of a rival’s actions can also help determine thresholds. A large-scale attack may spark more significant retaliation than a single precision strike.

Policymakers often deliberately keep these limits vague to strengthen their hands. Although officials sometimes announce explicit thresholds, too much clarity can weaken deterrence by enabling rivals to know just how far they can go. In contrast, ambiguity can enhance deterrence by forcing opponents to exercise restraint, lest they cross an escalation threshold.

Consider the Philippines’ calculations for responding to Chinese provocations in the water around its territory. It is unclear what would drive Manila to use force in response to aggressive Chinese moves against Philippine ships. It is equally unclear how Beijing would respond to Manila’s actions and whether such a crisis would lead the Philippines to invoke the U.S.-Philippine Mutual Defense Treaty, which commits Washington to defend the country and could thus draw U.S. forces into the dispute. The uncertainty surrounding these interactions may make Beijing more cautious than it otherwise would be.

But uncertainty also raises the prospect that provocations will lead to a crisis that could spiral out of a leader’s control. The tension between using brinkmanship to pressure a rival and the desire to limit escalation forces leaders to navigate crises cautiously, probing how far they can go while keeping a situation under control.

ON THE LEDGE

Policymakers have to carefully calibrate their actions. They must show enough capability and resolve to advance their objectives while providing rival leaders with the space to back down. They do so, in large part, by avoiding significant affronts to a rival’s honor and by anticipating and then not crossing a rival’s redlines.

States often control escalation by limiting the physical effects of their coercive actions. Avoiding casualties or major infrastructure damage makes it easier for targeted states to refrain from serious retaliation. Russia and Iran have downed U.S. military drones to convey their displeasure with Washington’s reconnaissance missions, for example, but they have avoided the escalation risks of downing a manned aircraft. Likewise, Israel responded to Iran’s April attack by striking a single radar at a critical Iranian air defense site rather than launching a larger and more destructive operation. Although the attack caused little physical damage, it demonstrated Israel’s ability to target advanced systems deep within Iran. Since the strike generated limited harm, Tehran could downplay the attack at home and avoid launching a significant retaliation.

In addition to selecting targets and using precision munitions, states can minimize harm by forewarning of their actions, allowing the targeted states to strengthen defenses and otherwise prevent damage. Before retaliating for Israel’s attack on the Iranian embassy in Damascus this April, for instance, Tehran telegraphed its response plan. Iranian officials publicly threatened imminent strikes, privately warned regional governments of the impending attack, and messaged to Israel and the rest of the world that they did not seek a full-scale war. By the time Iran launched its missile and drone barrage nearly two weeks later, Israel and its partners were prepared to shoot most of them out of the sky, ensuring minimal physical damage and casualties.

But limiting destruction and the loss of lives is only part of the story. The location, timing, and method of attack can be just as important for managing escalation, even if the physical outcomes are the same. Iranian officials would undoubtedly have seen Israel’s killing of Hamas’s political chief, Ismail Haniyeh, as far less provocative if it had taken place in Gaza instead of in Tehran. Similarly, Moscow would likely view a Ukrainian ground force assault on a Russian military base as more escalatory than a drone strike on the same facility.

As a result, decision-makers frequently avoid actions that directly challenge a rival’s territory. For example, Washington seeks to deter Iranian-sponsored attacks on U.S. forces by targeting Iranian Revolutionary Guard Corps facilities and Iranian-affiliated militias in Iraq and Syria instead of carrying out direct strikes in Iran. In doing so, the United States tacitly acknowledges that attacking Iranian territory would cross an escalation threshold.

Policymakers can also use coercive tools that are more deniable or otherwise less visible to the public. In the 1950s, Soviet and American pilots waged a covert air war over the Korean Peninsula that Washington and Moscow both kept hidden from the public. Today, Ukraine often refuses to take responsibility for its drone strikes on Russia. States also increasingly use “gray zone” tactics such as cyberwarfare or rely on proxies such as Russia’s Wagner paramilitary company to further their aims in a plausibly deniable manner. The political scientist Austin Carson has argued that these “backstage” activities allow governments to secretly apply pressure while avoiding demands for escalation from the public, which often grows more hawkish after visible confrontations.

Once a state carries out coercive actions, policymakers can announce their intent to avoid further escalation. After Iran’s January 2020 missile attack on a U.S. military base in Iraq, Tehran issued a public statement to the UN secretary-general stating that it “took and concluded” military operations to retaliate for Washington’s assassination of Major General Qasem Soleimani, a top Iranian commander, noting that it “does not seek escalation or war.” Iranian Foreign Minister Javad Zarif tweeted a similar message, stressing that Iran’s military actions were done. The United States launched no military response to the attack, opting instead to impose additional economic sanctions on Iranian firms and officials.

But even after an attacker has achieved its goals and suggested it wants to go no further, its rival must find some way to reestablish deterrence. Officials must rewrite their rules of engagement, creating new thresholds that make clear that future aggression will be met with resistance. Iran’s response to Israel’s embassy attack signaled a new normal by demonstrating a willingness to directly attack Israeli territory, a task Tehran had previously delegated to its regional proxies. Such new, unwritten rules and thresholds push already fraught relations up the escalation ladder, creating uncertainty and fear that should cause leaders to exercise greater restraint.

Controlling escalation, however, comes with tradeoffs. Actions that are too restrained may not shape a rival’s behavior. U.S. strikes on Yemeni drone and missile sites, for example, have failed to halt Houthi attacks on ships in the Red Sea. Although this is partly a tactical issue—the Houthis are adept at hiding and moving launchers—Washington has also failed because its actions have not imposed costs large enough to force the Houthis and their Iranian patrons to back down. More aggressive U.S. actions might deter the Houthis more effectively, but they are also more likely to provoke Iranian escalation. Getting the Houthis to back down through force, then, might come at the cost of a general escalation in the region, ultimately a more counterproductive (not to mention dangerous) outcome for all involved.

IT TAKES TWO

Even the best efforts to avoid escalation can fail. Decision-makers may misjudge their rivals’ thresholds, taking actions that opponents perceive as more provocative than intended, as Israel did when it attacked Iran’s embassy in Syria. Israeli officials expected a minor retaliation, not an onslaught of hundreds of missiles and drones.

If tensions do rise, states can try to de-escalate. But that can be challenging since policymakers face pressure to ramp up during crises. Leaders understandably fear that looking weak will harm them politically. Constituents may punish leaders at the polls for failing to act. Other rivals closely observe a state’s crisis behavior to assess capability and resolve, and appearing weak in one crisis can weaken a state’s bargaining position in future confrontations. Such concerns are particularly severe when backing down involves reneging on a commitment, such as an agreement to defend another country or a public pledge to stand firm in a crisis. In September, for instance, Gilberto Teodoro, the Philippine defense secretary, announced that he expected American intervention in the event of a Chinese attack on Philippine military outposts. Similarly, U.S. President Joe Biden has repeatedly described Washington’s defense commitment to the Philippines as “ironclad.” As a result, it will be difficult for the United States to back down from its treaty obligations without being labeled as untrustworthy.

To complicate matters, new technologies make it harder to avoid escalation pressures. Commercial imagery satellites, cellphones, and other smart devices create a world with fewer secrets. This increased transparency makes it difficult to hide the covert and gray-zone actions that leaders often use to engage in less escalatory backstage confrontations. Meanwhile, social media provides a platform for inflammatory content that can stoke escalation.

Still, as the tit for tat between Iran and Israel showed, war is almost never inevitable. The road to conflict is an action-reaction process. Leaders decide whether and how to respond to a rival’s moves, and they often search for ways to lower the temperature. Escalation to war, after all, is not always in a state’s interest. Victory is not guaranteed, and the costs of fighting might outweigh the gains. As a result, states are often better off coming to a settlement that advances their strategic objectives without going to battle—even if a leader suffers political or reputational consequences because of it.

CLIMBING DOWN THE LADDER

To avoid playing what the international relations scholar James Fearon has called the “costly lottery” of war, leaders find ways to back down from intense crisis escalation while preserving their reputation and ensuring deterrence. To do so, policymakers must craft arrangements in which all parties can claim success or find face-saving off-ramps. In Iran and Israel’s exchange last spring, for instance, Tehran was able to project strength to both domestic and international audiences simply by showcasing its ability to launch large-scale strikes on Israel, even though the attack caused minimal damage. For their part, Israeli leaders emphasized that they can safeguard the country from even a mass attack.

Rival leaders can also tacitly collude with each other to avoid war. This frequently involves mutually deciding to keep each other’s actions hidden from the public. In the 1950s, to avoid stoking pressures to escalate, neither Moscow nor Washington disclosed their air war over Korea. Beyond such unspoken coordination, communication between rivals—either directly or through intermediaries (such as Qatar, in the case of Israel and Hamas)—can help leaders step back from war. Officials can clarify intentions and thresholds and diffuse tension after accidents, avoiding miscalculations and further escalation. There is significant precedent for this type of coordination. The close calls of the Cuban missile crisis spurred Washington and Moscow to set up a crisis hotline in 1963, and the United States established a similar connection with Beijing in 2007. Other rivals might benefit from emulating this approach.

As crises become more common and intense, the role leaders play in pulling states away from the precipice of war becomes increasingly important. When tensions push states to the brink, decision-makers must play a high-stakes bargaining game and identify ways to pursue their aims and deter future harm while avoiding war. But they need not panic about inadvertent war. The tools of restraint lie in their hands.