Elephants in the Room

President Trump and the Risks of Nuclear War

How command and control works when the military wakes up the president vs. when the president wakes up the military.

A deactivated Titan II  nuclear missile in Green Valley, Arizona on May 12, 2015. (Brendan Smialowski/AFP/Getty Images)
A deactivated Titan II nuclear missile in Green Valley, Arizona on May 12, 2015. (Brendan Smialowski/AFP/Getty Images)

On Tuesday, the Senate Foreign Relations Committee held a hearing — the first of its kind in over 40 years — exploring the president’s authority to order the use of nuclear weapons. I was one of three witnesses, the others being former STRATCOM commander Gen. Bob Kehler and former Acting Undersecretary of Defense for Policy Brian McKeon. You can read my prepared testimony here.

Senator Bob Corker (R-Tenn.) convened the hearings in response to concerns raised by many experts and political leaders about President Donald Trump’s fitness to wield the most fearsome power of the presidency: the ability to initiate a nuclear war.

Of course, this has been a contentious issue from the very beginning of the nuclear age. My first major book, Guarding the Guardians, explored it in some detail, focusing especially on the early period (which happened to be the only period with relatively abundant archival material when I was doing the research in the late-1980s). My research showed that nuclear command and control involved balancing between two competing desiderata — on the one hand, reinforcing deterrence by assuring we could always present adversaries with a credible capability of launching a nuclear strike even in the most dire of scenarios; and, on the other hand, reinforcing stability by assuring we would never have an accidental, unauthorized, or inadvertent use of nuclear weapons. Civilian and military leaders did not always agree on how to strike the balance between those objectives, and where the system ended up at any given point in time changed with changes in technology, the evolving Soviet threat, and other factors.

In my testimony, I argued that the time was ripe for another look at this question, even as I cautioned against hastily adopting any particular legislative fix. I do think Congress and outside experts should be debating the various proposals, and the pros and cons of the reforms should be viewed against the backdrop of a new, top-to-bottom review of nuclear command and control. In the past, such reviews turned up important anomalies that warranted policy or technical changes and I would not be surprised to learn that a new review likewise found areas needing improvements.

At a minimum, the rapidly evolving cyber threat has raised new concerns of a type that was considered more manageable during the Cold War: the prospect that an adversary could be operating within the command and control system. So much of system is decades old, and likely long-overdue for modernization and upgrades. These two facts alone would justify a fresh look.

That said, many of the legislative proposals floating around raise hoary constitutional issues about war powers or are based on policies that have been carefully considered and rejected by previous generations of strategic leaders. It is possible that looking at old problems with new perspectives will lead us to change our mind about the advisability of these reforms. But we should be wary about second- and third-order consequences and so should scrutinize proposals with as jaundiced an eye as we scrutinize claims by nuclear operators that suggest “all is well, nothing to see here…”

During the question and answer period, two other major concerns emerged that are worth clarifying.

The first concerns confusion about how streamlined the decision-making process is and how great are the risks associated with that process. It is commonplace to hear that no one can stop the president from ordering the use of nuclear weapons, that he alone can make the decision and needs no one else to second it, and that he could do so in only minutes. This concern gives the impression that the president could take the country from peaceful stability to a nuclear war with about the same effort and carelessness with which he can fire off a tweet.

This misunderstanding of nuclear command and control arises out of a conflation of two very different contexts: situations in which the military wakes up the president vs. situations in which the president wakes up the military.

When the military wakes up the president

In the former context, it is the military that has been monitoring global events and has detected an urgent threat that requires a presidential decision — for instance, an actual missile attack or an imminent launch from a rogue state like North Korea. In this launch-under-attack context, the decision-making process is indeed quite truncated and the system would be primed to implement the president’s orders. While the process also does everything technically possible to ensure that the president would have the benefit of counsel from his national security team, it is technically possible that the president would have to make a decision before all of his advisors had had the chance to weigh in. And, pursuant to Article II of the Constitution, the president would have the final say and would not need concurrence from subordinates for the order to be legal.

Of course, the president would not truly be acting alone. After all, the military would already be providing a certain degree of de facto concurrence since it was they who made the decision that the threat required a presidential decision, and they who presented the range of options to the president for the decision. But the electorate on Nov. 8, 2016, did decide that it would be Donald Trump who would make this decision in this crisis situation and, should it come to it, I believe the nuclear command and control system would recognize the president’s decision as legal — and carry out that order.

It is hard to identify any legislative fix that would: (a) prevent the president from making that decision in that extreme scenario that would also (b) pass Constitutional muster and that would (c) not undermine deterrence.

When the president wakes up the military

It is a different matter in the other context: when it is the president who wakes up the military and tries to get them to go from peacetime to war, i.e. to launch a preventive nuclear attack. In the preventive case, it is not reasonable to believe that the streamlined procedures of an emergency response would operate without anyone raising objections. The steps the president would have to take in order to pass a nuclear order to someone who could physically launch the missiles would simultaneously alert the rest of his national security team. Efforts to bypass the senior leadership would themselves further alarm subordinates, increasing the likelihood that they would draw in the rest of the national security team, even if ordered not to.

The military is trained to reject illegal orders and the president trying to order the military to go from peacetime to nuclear war without consulting with his national security advisors would set off alarms up and down the system about whether the orders were legal.

The president does not need anyone else to help him fire off a tweet, but he does need many others to help him fire off a nuclear intercontinental ballistic missile. If he were trying to do so it would take an enormous effort of persuasion that would involve many more people than are involved in the streamlined, launch-under-attack scenario.

What would happen in this second scenario? That question led to the second major topic of concern in the Q&A period: how robust is the military training to resist illegal orders and how confident could we be that the Pentagon would view an order to “launch a preventive nuclear war without notifying my national security team” to be illegal?

In testimony, General Kehler repeatedly emphasized that the military does not follow orders blindly and the ubiquity of lawyers at multiple layers of command gives us high confidence the legal questions would be asked (and need to be answered) before a nuclear strike actually happened.

This is true, though it is also true that the military are trained that authenticated orders from the national command authority have a presumption of legality. (Note: the presumption is even stronger in launch-under-attack scenarios, because the United States has long embraced the legal concept of anticipatory self-defense, which could result in a decision to strike under circumstances where the United States has not yet suffered an attack, but one is deemed to be imminent or even underway.)

Nevertheless, I am inclined to share General Kehler’s confidence that a rogue president would find it exceedingly hard to persuade the military to act in preventive war scenarios as rapidly as they are trained to act in launch-under-attack scenarios. Part of this comes from my understanding of the civil-military context of national security. Presidents already find it challenging to persuade the military to embrace policies that the military object to — and which are far less consequential than preventive nuclear war. Another of my books, Armed Servants, explores in some detail the push and shove of civil-military relations. And still a third (co-authored with Chris Gelpi), Choosing Your Battles, shows that the military are hardly chomping at the bit to initiate the use of force. To be sure, I also found, in Guarding the Guardians, that the military did favor a system inclined to the always rather than the never side of the always/never dilemma. This is in part why U.S. political leaders insisted that there be civilian control of the arsenal. Yet all of these policies were the result of a lengthy bureaucratic struggle that involved many more people than just the president and the few nuclear operators required to launch a missile.

The longer timeline of a preventive war scenario gives the opportunity for all these actors to weigh in on the president’s decision. Yes, the president could still carry the day, as President George W. Bush did in 2003 when he ordered the invasion of Iraq in a similar preventive scenario. But Bush’s team spent over a year debating the decision. The military weighed in repeatedly. And, crucially, Congress voted to give the president the authority to do what he did.

Critics of the Iraq War decision can say this is cold comfort, since Bush ended up making a decision they deplored. But it would not be accurate to say that the president alone made the decision without anyone else involved in the process.

There is abundant murky middle ground that is worth exploring further.

  • What about a slowly escalating crisis — something in between launch-under-attack and cold-start preventive war?
  • What about the peculiarities of the Korean situation, where the original armed conflict of the Korean War is still technically operative, suspended only by a cease-fire?
  • And what about the risks of a “Gulf of Tonkin” scenario? For instance, myriad Chapter VII resolutions authorize the United States (and others) to take steps short of the use of force that might nevertheless produce an escalatory spiral that would result in North Korea taking military action that would provide legal cover for U.S. military action.
  • Are there other ways to strengthen nuclear command and control that would extend the decision time afforded to the president in any scenario, maximizing opportunities for the president to get the benefit of counsel from his advisory team and from Congress?

I encouraged the senators to continue to review nuclear command and control, seeking answers to these and other questions. Even if the most alarmist concerns are over-stated, the stakes are high enough to warrant a fresh look at this issue.

Peter D. Feaver is a professor of political science and public policy at Duke University, where he directs the Program in American Grand Strategy.

Trending Now Sponsored Links by Taboola

By Taboola

More from Foreign Policy

By Taboola