The game theory is clear: it can be in your interest to make threats that it would not be in your interest to carry out. So you can gain from committing to carrying out such threats. But only if you do it right. Your commitment plan must be simple and clear enough for your audience to see when it applies to them, how it is their interest to go along with it, and that people who look like you to them have in fact been consistently following such a plan.

So, for example, it probably won’t work to just lash out at whomever happens to be near you whenever the universe disappoints you somehow. The universe may reorganize to avoid your lashings, but probably not by catering to your every whim. More likely, others will avoid you, or crush you. That’s a bad commitment plan.

Here’s a good commitment plan. A well-run legal system can usefully deter crime via committing to consistently punish law violations. Such a system clearly defines violations, and shows potential violators an enforcement system wherein a substantial fraction of violations will be detected, prosecuted, and punished. Those under the jurisdiction of this law can see this fact, and understand which acts lead to which punishments. Such acts can thus be deterred.

Here’s another pretty good commitment plan. The main nations with nuclear weapons seem to have created a mutual expectation of “mutually assured destruction.” Each nation is committed to responding to a nuclear attack with a devastating symmetric attack. So devastating as to deter attack even if there is a substantial chance that such a response wouldn’t happen. This commitment plan is simple, easy to understand, clearly communicated, and quite focused on particular scenarios. So far, it seems to have worked.

Humans are often willing to suffer large costs to punish those who violate their moral rules. In fact, we probably evolved such moral indignation in part as a way to commit to punishing violations of our local moral norms. In small bands, with norms that were stable across many generations, members could plausibly achieve sufficient clarity and certainty about norm enforcement to deter violations via such threats. So such commitments might have had good plans in that context.

But this does not imply that things would typically go well for us if we freely indulged our moral indignation inclinations in our complex modern world. For example, imagine that we encouraged, instead of discouraged, mob justice. That is, if we encouraged people to gossip to convince their friends to share their moral outrange, building off of each until they chased down and “lynched” any who offended them.

This sort of mob justice can go badly for a great many reasons. We don’t actually share norms as closely as we think, mob members are often more eager to show loyalty to each other than to verify accusation accuracy, and some are willing to make misleading accusations to take down rivals. More fundamentally, we might say that mob justice goes bad because it is not based on a good commitment plan. Observers just can’t predict mob justice outcomes well enough for it to usefully encourage good behavior, at least compared to a formal legal system.

Now consider the subject of making peace deals to end wars. Such as the current war between Russia and Ukraine. An awful lot of people, probably a majority, of the Ukrainian supporters I’ve heard from seem to be morally offended by the idea of such a peace deal in this case. Even though the usual game theory analyses of war say that there are usually peace deals that both sides would prefer at the time to continued war. (Such deals could focus on immediately verifiable terms; they needn’t focus on unverifiable promises of future actions. In April 2022 Russia and Ukraine apparently had a tentative deal, scuttled due to pressure from Ukrainian allies.)

Many of these peace deal opponents are willing to justify this stance in consequentialist terms: they say that we should commit to not making such deals. Which, as they are eager to point out, is a logically coherent stance due to the usual game theory analysis. We should thus “hold firm”, “teach them a lesson”, “don’t let them get away with it”, etc. All justified by game theory, they say.

The problem is, I haven’t seen anyone outline anything close to a good commitment plan here. Nothing remotely as clear and simple as we have with criminal law, or with mutually assured destruction. They don’t clearly specify the set of situations where the commitment is to apply, the ways observers are to tell when they are in such situations, the behavior that has been committed to there, or the dataset of international events that shows that people that look like us have in fact consistently behaved in this way. Peace deal opponents (sometimes called “war mongers”) instead mainly just seem to point to their mob-inflamed feelings of moral outrage.

For example, some talk as if we should just ignore the fact that Russia has nuclear weapons in this war, as if we have somehow committed to doing that in order to prevent anyone from using nuclear weapons as a negotiating leverage. The claim that nations have been acting according to such a commitment doesn’t seem to me at all a good summary of the history of nuclear powers. And if the claim is that we should start now to create such a commitment by just acting as if it had always existed, that seems even crazier.

If we have not actually found and clearly implemented a good commitment plan, then it seems to me that we should proceed as if we have not made such a commitment. So we must act in accord with the usual game theory analysis. Which says to compromise and make peace if possible. Especially as a way to reduce the risk of a large nuclear war.

The possibility of a global nuclear war seems a very big deal. Yes, war seems sacred and that inclines us toward relying on our intuitions instead of conscious calculations. It inclines us toward mob war strategy. But this issue seems plenty important enough to justify our resisting that inclination. Yes, a careful analysis may well identify some good commitment plans, after which we could think about how to move toward making commitments according to those plans.

But following the vague war strategy inclinations of our mob-inflamed moral outrage seems a poor substitute for such a good plan. If we have not yet actually found and implemented a good plan, we should deal with a world where we have not made useful commitments. And so make peace, to avoid risking the destructions of war.