Frame Game

The Subjective Earthling

This is a reply to Robert Wright’s recent “Earthling” column. 

Here’s a way to understand this debate or conversation or dialectic or whatever it is. It began with a one-dimensional game-theory model in which we, the United States, were supposed to figure out how to appease terrorists so they wouldn’t attack us again. I made the model two-dimensional by pointing out that while terrorists were trying to dictate our behavior through negative feedback (killing a bunch of us), we could try to dictate their behavior by administering similar negative feedback to them (killing them, destroying their infrastructure, confiscating their assets, etc.).

Bob Wright has improved the model again by adding a third dimension: other players. He asks whether I agree with these three propositions:

a) It makes sense to distinguish between these two different groups—terrorism’s elites on the one hand and the world’s millions of discontented Muslims on the other.b) Our relationship with the former is essentially zero-sum and with the latter is essentially non-zero-sum.c) The object of the game is to hurt the former and help the latter while minimizing the extent to which pursuing each of these goals compromises the other goal.

My answer to all three questions is yes. And since Bob has done me the favor of adding a third dimension to my model, I’m going to return the favor by adding a fourth to his. In his footnote, he observes that many “leaders of terrorism-sponsoring states … have the realistic option of changing policies in a way that would pivot that relationship [between them and the United States] toward the non-zero-sum.” I don’t see how we can stipulate up front exactly which players can be pivoted in this way. Can Iran be pivoted? Syria? Fatah? Hamas? Who exactly within these governments or organizations can be coaxed or prodded into a non-zero-sum relationship with us?

All other things being equal, every player we shift into the non-zero-sum category improves the outcome of the game. But all other things aren’t equal. As Bob notes, many variables are in play. Our efforts to shift players into the non-zero-sum category, like our efforts to give positive reinforcement to players who are already in that category, must “minimize the positive reinforcement” given to players in the zero-sum category.

So here’s the problem. Given Bob’s three-point framework, we can’t maximize the outcome without knowing who can be coaxed into a non-zero-sum relationship with us, and at what cost.

The same mystery variable applies to people in the non-zero-sum category. Which of them are at risk of sliding into the zero-sum category? What’s the cost (in terms of positive reinforcement for our zero-sum enemies) of giving these at-risk individuals or entities enough positive reinforcement to prevent them from becoming, in Bob’s words, “more terrorist foot soldiers”?

My suspicion is that at this point game theory can’t help us. In the context of cultural conflict, game theory can factor in a person’s decision to abandon a zero-sum posture, but it can’t predict that decision or what it requires. Some people are, as Bob says, “implacably devoted to ending global capitalism and American civilization.” Others who seem similarly devoted may turn out not to be implacable.

This leads back to the paradox I was getting at with my crude model of mutually administered consequences. There’s a subjective way of looking at the world, and an objective way. From the objective perspective, people are predictable. From the subjective perspective, they aren’t. The point of my piece, to put it in the language of game theory, is that when you’re in a zero-sum conflict, the best psychological strategy is to convince your enemy that your violent retaliatory behavior is objective (i.e., that you always strike back and can’t be deterred) while refusing to believe the same of your enemy.

It seems to me that by definition, game theory can’t handle this variable, since the subjectivity or objectivity of anyone’s behavior is a metaphysical rather than an empirically answerable question. And yet we have to answer it in the case of each player—Is this regime implacably devoted to our destruction? How much American stubbornness will drive this discontented Muslim to terrorism?—in order to apply the four-dimensional model. I don’t see how a mere mortal can solve this puzzle, so I’ll defer to a higher authority. God grant me serenity to accept the things I cannot change, courage to change the things I can, and wisdom to know the difference.