So I winged out to Chicago last month for a symposium at the Pritzker Military Museum and Library, across Michigan Avenue from the justly famed Art Institute of Chicago. My panel reviewed “Global Partnerships & Treaties.” This is a critical topic. After all, the United States’ standing in the world hinges on alliances and fellowships of all types—chiefly in the rimlands and marginal seas ringing the Eurasian supercontinent. America has no strategic position in the rimlands without them.

It can accomplish little in the place that matters most.

Nor is this lost on antagonists. They grok how much the United States depends on multinational leagues. That’s why they are doing their darnedest to degrade or break existing U.S. alliances, coalitions, and partnerships and forestall new ones. At present every body of water adjacent to the Eurasian periphery is potentially or actually embattled, from the Baltic Sea to the Red Sea to the Sea of Japan and everywhere in between. As geopolitics maven Nicholas Spykman pointed out eight decades ago, the United States has no access to the rimlands without command of waters lapping against the rimlands.

Without access it cannot carry on commerce. Nor can it wield diplomacy and military might to prevent a hostile hegemon or alliance from gaining control of the rimlands and constituting a threat to the Western Hemisphere. Maritime access is precious beyond price.

Here endeth the geopolitics sermon. During the Q&A in Chicago our panel chairman raised an intriguing question, riffing on Lord Palmerston’s oft-cited maxim that nineteenth-century Great Britain had no permanent allies, only permanent interests. (Palmerston hastened to add that Albion had no perpetual enemies either.) It may be that’s a general rule. Britain founded or joined alliances when collaborating suited its interests, chiefly in shaping events on the continent of Europe. Alliances expired their shelf-lives when they no longer suited British interests. In Palmerston’s view alliances, coalitions, and partnerships are temporary arrangements. Yet those who build, maintain, or study alliances tend to put the accent on what unites them and sustains them—not how they end.

So how do alliances end?

My go-to work for puzzling through such things is an old International Security essay by Harvard professor Steven Walt, titled “Alliance Formation and the Balance of World Power” (subsequently expanded to book form as The Origins of Alliances). To oversimplify, Walt discerns three types of adhesives that bring together alliances. He is a consummate international-relations realist, so power and national interests stand above all other factors. He seems to vest little faith in lesser adhesives while conceding they do exist and could boost or degrade an alliance’s longevity.

According to the realist school, common interests—mainly in fending off common threats—apply the stickiest glue binding together an international fellowship. Next stickiest are political, cultural, and social affinities among the allies. A common language, political system, or heritage primes the partners to look at the strategic environment in similar ways and to discern similar strategies for managing it. Walt puts less stock in such binding agents.

For him the weakest adhesives are transactional. That is, a predominant ally could rent allies, bankrolling some common venture. Or the hegemon could strongarm them, compelling them to obey its bidding. The crucial point is that transactional alliances endure only as long as the transaction remains in force. Lesser allies skedaddle when a hegemonic ally no longer pays the rent, or when it can no longer hold them at gunpoint. In other words, when they no longer have any interest in material gain or need to worry about self-preservation.

So there you have three broad ways of projecting how an alliance could end, extrapolating from what brings it together in the first place and keeps it together (until it doesn’t). The allies’ interests could diverge, or a common threat that brought them together could retreat enough that countering it no longer demands concerted action. The partners would drift apart. Sociocultural affinities might loosen, if indeed they were that compelling to begin with. Or one well-resourced or domineering ally might no longer have the means or desire to pay off or coerce its junior partners.

Strategy grandmasters Sun Tzu and Carl von Clausewitz may not see eye to eye on everything, but they are confirmed alliance skeptics. Sun Tzu enjoins Chinese sovereigns of antiquity to comport themselves like the “Hegemonic King,” the ruler of a state so musclebound that antagonists dare not make common cause against it. In fact, the Chinese sage designates alliance-breaking as a strategy of choice. Doing so simplifies the military problem by keeping foes fragmented, and weak, and vulnerable to being overcome one by one should the sovereign see the need.

In the best case—if alliance-breakers utterly outmatch their foes on an individual basis—they could spare themselves from fighting to fulfill their ambitions. Lesser combatants might stand down rather than risk an unwinnable or unaffordable war. Winning without a fight is good.

Clausewitz breaks down the types of alliances, observing that there are hegemonic alliances and alliances of equals. The dynamics within the two genres are quite different, in large part because of the Golden Rule of alliances: namely that he who has the gold makes the rules. Outsized contributions to the common effort confer outsized influence within alliance deliberations, whereas diplomacy among peers is at a premium when members commit more or less the same resources to the endeavor.

There are also coercive and consensual alliances, as Walt points out.

Whatever the internal makeup of a fellowship, Clausewitz regards alliances, coalitions, and partnerships as fragile, halfhearted affairs. Politics is key. He maintains that one ally never attaches as much value to another’s cause as to its own. It’s not all in, and thus is unprepared to invest maximum resources in the cause for as long as it takes. That being the case, the ally makes a middling investment in the cause. And tepid allies look for the exit when the going gets tough.

Think chemistry. Chemists teach that there are two types of chemical bonds, ionic and covalent. Atoms sharing an ionic bond transfer electrons to one another, making this a bond that verges on unbreakable. Atoms sharing a covalent bond share electrons—making this a relatively soluble bond. Clausewitz would contend that covalent bonds represent the rule in international partnerships of all sorts.

They are separable.

With all of this in mind, why don’t we spitball on some candidate scenarios for how alliances end. First, and most obviously, an alliance might dissolve because it accomplished the goal behind which the allies rallied. Clausewitz made a career of fighting against France. He saw coalition after coalition dash itself against French armies, only for the very last coalition to prevail at Waterloo and lay the groundwork for what turned out to be a durable postwar order in Europe.

The allied powers made provision to cooperate to forestall the rise of another Napoleon. But this was no NATO, a standing alliance that stays together after the demise of the hostile power that united it.

Second, and closely related, an alliance can start to fracture as the endgame of a conflict approaches and the allies look to their postwar interests—interests that may be odds with their wartime allies’. Austrian foreign minister Klemens von Metternich, another protagonist in the Napoleonic Wars and their aftermath, noted that an alliance without a “strictly determinate aim” tends to “disintegrate” as national interests and purposes diverge. Solidarity comes easy when allies bestride “death ground,” to borrow from Sun Tzu. The prospect of being hanged concentrates the mind wonderfully. That’s less and less true once the alliance steps off death ground and danger is no longer mortal. Then self-interest reasserts itself—because it can.

And solidarity becomes fissile. Think about the Grand Alliance that fought World War II. Once it became apparent that the Allies were the likely victor over the Axis, Allied leaders—Josef Stalin and Winston Churchill foremost among them—started jockeying for geopolitical position in postwar Europe. Marriages of martial convenience tend to prove perishable.

The partners look to themselves.

Third, if alliances can fall victim to their own success, defeat almost always sounds the death knell for the vanquished. Defeat means the arrangement has glaringly failed at its mutual purpose. The Central Powers didn’t outlive World War I, nor the Axis World War II. Nor is that surprising. Losing big tends to discredit an alliance—especially if it loses so big that its enemies can impose their own terms. Namely terms meant to prevent the rise of another such hostile concord.

Fourth, allies sometimes agree to disband their commitment to one another without encountering either victory or defeat. The Anglo-Japan alliance of 1902 served British and Japanese interests well for a time, but they consented to dissolve the alliance within the framework of the Versailles Treaty that cemented an end to World War I.

Fifth, one or more alliance members could undergo regime change. Revolutionary change tends to bring about drastic change to national character, to a point where the country ceases to be an attractive, capable, or competent partner in the common weal. Four empires fell during World War I or its aftermath, three among the fallen Central Powers—Germany, Austria-Hungary, and Turkey—and one among the victors, namely Russia. Neither wartime alliance was likely to stand amid wrenching change. The 1979 Iranian Revolution spelled an end to the U.S.-Iranian alliance. Or, in effect the Soviet Union committed assisted suicide in 1991—portending ill for the Warsaw Pact, a coercive alliance from which allies, and even Soviet republics, were desperate to flee.