President Donald Trump is right: North Korea’s nuclear program is on a dangerous trajectory. But there is no quick fix. Nor is there an imminent threat, and it does not help to create the impression that there is one. A show of force, if carefully calibrated, can be helpful. But rhetorical excess, personal provocations directed at North Korea’s leader Kim Jong Un, and stunts like calling the Senate to a White House briefing don’t help.

Moreover, for all the bellicose posturing from Washington and Pyongyang, the most likely outcome of this latest flare-up of tensions around North Korea is more stalemate, as the Kim regime continues work on its nuclear-weapons program—unless the Trump administration considers some new, non-military,  approaches.

Where North Korea is concerned, Trump is certainly not the first U.S. president to insist that “all options” remain on the table. His four immediate predecessors couldn’t stop Pyongyang’s nuclear march, not for reasons of negligence, insufficient toughness, or poor deal-making. There are, indeed, some foreign-policy problems with no “good” solution, where good means both a strong likelihood of success and a low risk that the remedy will turn out to be worse than the threat. North Korea is one such problem.

Within three broad categories—negotiation, military action, and persuading China to force its ally to give up its nuclear program—the choices have remained largely the same over the years, as North Korea has advanced from secretly piecing together one or two warheads to boasting a usable nuclear arsenal (perhaps 10-20 warheads with lots more fissile material on the way) and openly testing short, medium, and, soon, intercontinental-range missiles to deliver them. Twenty years ago, it made sense to insist, as Secretary of State Rex Tillerson recently did, that North Korea accept denuclearization—giving up its nuclear weapons and facilities—as the precondition to negotiations. But because it fails to acknowledge all that North Korea has achieved in the interim, this position makes no sense today.

The military options have been reviewed, again and again, with the same conclusion each time: None are attractive. Seoul, a city of 10 million, lies just 35 miles from the DMZ, in range of Pyongyang’s heavy artillery. Enough of its guns could fire off a single round to inflict huge casualties on America’s ally before its planes could shut them down. While the United States could take out the North Korean nuclear or missile-launch facilities it knows of, there may be many it does not know of. And now that North Korea has mobile missiles, and solid-fueled ones that can be launched quickly, the picture looks even worse.

Under such conditions, a preemptive strike would be folly. North Korea would retaliate against Japan, South Korea, or the 28,000 U.S. troops stationed there, forcing Washington to respond. As North Korea began to lose, a conventional war would escalate into nuclear catastrophe.

Then there is the third option, increasingly prominent in recent years, of insisting that China solve this problem. This is a false hope. Short of forcing North Korea’s collapse, China cannot make it give up the very weapons it views as its only buffer against Armageddon. And China, for reasons of its own national security, will not go that far. It fears the flood of refugees that would result from the collapse of North Korea, as well as the ensuing chaos of regime change and the dangers from uncontrolled access to its neighbor’s nuclear, chemical, and biological weapons. Its biggest fear, however, is a unified, U.S.-allied Korea, with American forces directly on its border. In that case, given Washington’s treaty ties with Japan on the east, and increasingly close relations with India to the west, China would feel itself encircled.

This is not to say that China is powerless. In temporarily shutting down coal imports from North Korea, which provide the Kim regime with desperately needed revenue, it has taken a promising step. It can and should inflict more pain by shutting down leaks in UN-imposed sanctions by Chinese companies and banks. But even the extreme step of shutting down oil exports to Pyongyang would not yield Trump’s desired outcome, where Beijing “solves” North Korea if only it saw the problem as America thinks it should.

So, if the United States doesn’t like the prospect of North Korean nuclear-armed ICBMs that can reach its shores, it is time for some new thinking, forged from what it already knows about the Kim regime, but keeps forgetting.

First, threats don’t work. Indeed, they are counter-productive, because they confirm Pyongyang’s belief that it faces an unrelenting menace from the United States and its allies, and will survive only by never backing down. Another bit of wisdom that experience should have taught Washington is that Pyongyang’s protectors—first Russia, now China—have limited influence on its actions, much less than seems possible. In the 1970s, during a break in a negotiation on conventional-arms transfers, a Soviet official who had served as the USSR’s defense attaché in Pyongyang told me that in the years he lived there, he was never allowed into the North Korean defense ministry—all he had was a phone number. This was at a time when Pyongyang depended entirely on Moscow for arms and other support. The power dynamic one would expect between a superpower and a weak vassal state was inverted, with the North Koreans dictating the terms. The Chinese now find themselves in the same position the Russians once did.

If military force would be unwise, denuclearization talks a non-starter, and a Chinese solution unlikely, the most obvious alternative is negotiations to freeze, and perhaps partially roll back, North Korea’s nuclear program, rather than dismantle it, with ironclad Chinese-backed provisions to prevent cheating. This would mean de facto recognition of North Korea as a nuclear-weapons state, rewarding it for violating its Nonproliferation Treaty commitment and for spurning UN resolutions, and establishing a terrible precedent for any other state that might be tempted to follow its example. However, no other state is in a comparable position: economically isolated and therefore relatively immune to sanctions, and able to inflict very heavy casualties with its conventional forces. Another drawback is that an agreement that left in place some nuclear weapons and short-range missiles would, while largely removing the threat to America, leave its allies in danger. Such undesirable but best-available outcomes  are the price for delaying too long in acting against a proliferator.

Another possibility would be to strengthen America’s deterrence posture with continued deployments of missile defenses and an enhanced military presence in the region, making U.S. determination to act in its own and its allies’ defense more clear to North Korea, South Korea, and Japan. China would see these developments as a threat, of course—that would have to be managed.

The most extreme step on this pathway, and a last resort short of war, might be to encourage Japan and South Korea to develop their own nuclear weapons (both are now protected by the U.S. nuclear umbrella, under what is known as extended deterrence). That would cripple the global effort to halt nuclear proliferation, and could endanger, rather than safeguard, Northeast Asia. And it would have grave consequences for America’s standing in the world, reinforcing the view that Washington makes rules it demands others live by until it changes its mind and changes those rules.

A different course entirely would involve a difficult, long-term dialogue between the United States and China on whether the two countries could overcome their mutual mistrust and develop a shared vision of a neutral, united Korea. A prerequisite to such an effort would be a very clear U.S. strategy towards East Asia, including its relations with China, developed first for itself (requiring bipartisan support) and then with Tokyo and Seoul.

This list of potential approaches to North Korea is not exhaustive. Nor is it a full reckoning of their pluses and minuses. The point is that options beyond those that haven’t worked in the past do exist. There is a common thread among them: Where North Korea is concerned, neither China nor America will achieve security acting separately.

This article was originally published in the Atlantic.