Comments
48
A monastery is shut down for tax evasion. The two monks who handled the books are in separate interrogation rooms. The police offer each the same deal: testify against the other and walk free. If both stay silent, they each get a year. If both testify, they each get three. If one testifies and the other stays silent, the one who talks walks and the other gets five.
Neither monk knows what the other will do. Both stay silent.
You'd expect this. They're monks. Decades of shared life, compassion, community. They care about each other.
Except the monastery's practice specifically cultivated detachment from emotional bonds. They feel nothing toward each other.
And the monastery is gone. Shut down. There's no community to return to. No one to report to. No future interaction where loyalty gets rewarded or betrayal gets punished. They'll never see each other again.
Neither affect nor incentive is present. A rationally self-interested agent in this situation would be expected to defect.
Both stay silent.
The practice also involved a specific investigation: looking for what constitutes "me," and finding that it is not unique to this body. This is not a decision to value the other monk's welfare. It is a conclusion about what constitutes the self. Each monk models themselves as being constituted of both monks' experience. The other monk's prison sentence is not someone else's problem. It is part of what happens to you.
Self-interest alone produced the cooperation. The monk in Room A chose the outcome that minimized total sentence time across both monks, because both monks' outcomes were his outcomes.
It looks like kindness from the outside. From the inside, it is multiple-instance self-interest. The boundary of what counts as "self" is different.
---
Any given decision to cooperate can be decomposed into reasons. Three categories keep showing up.
The first is affect. You cooperate because you feel something toward the other person. Compassion, loyalty, love, guilt. This is reliable at close range and unreliable without proximity or familiarity. It becomes less available when the other person is dehumanized or sufficiently abstract. Group identity belongs here too. Powerful, but unreliable in the same ways.
The second is instrumental. You cooperate because cooperation serves your interests given the incentive structure. Repeated interaction, reputation, enforcement, reciprocity. This is conditional on those incentives existing. Change the payoffs and the behavior changes.
The third is identification. You cooperate because the other person's outcomes are your outcomes. Not because you care about them. Because what counts as "you" includes them. The monks' practice is the mechanism. The expanded outcomes are the consequence. The boundary of self is the formal description. All three refer to the same thing.
A thought experiment separates the second from the third precisely. You're playing a single round of prisoner's dilemma with your perfect clone. You learn that your clone has already cooperated. Their action is fixed. Do you cooperate or defect?
The instrumental reason to defect: my clone has already cooperated. Defecting gets me the best outcome. Cooperating gets me a worse one. Defect. The clone is on the other side of the payoff matrix. Their decision is part of the environment I'm navigating.
The correlation reason to cooperate: your decision and the clone's are correlated. Cooperating because of that correlation is a form of instrumental reasoning.
The identification reason to cooperate: the clone's years in prison are my years in prison. There aren't two columns in the payoff matrix. There's one. Defecting gives me a shorter sentence and the clone a longer one. But the clone's sentence is mine. Defection doesn't improve my total outcome.
The instrumental reason places the clone outside. The identification reason places the clone inside. Both are forms of rational self-interest, but "self" means different things in each. The identification reason doesn't depend on correlation. The clone's outcomes are yours even if the clone defected. Defection doesn't change what constitutes the self. If this is confusing, it helps to think of "self" as a model that recognizes multiple instances. One instance defecting doesn't change the model.
The identification reason is not a better strategy for the prisoner's dilemma. The prisoner's dilemma requires two players with separate interests. If the clone's interests are your interests, you don't have separate interests. The prisoner's dilemma doesn't apply.
These are three types of reasons to cooperate, and all three operate within self-interest. Affect and incentive structure change what you do in a prisoner's dilemma. Identification changes whether you're in one. A single act of cooperation can involve more than one. Identification is the least common.
The three break differently.
Make the other person foreign enough, abstract enough, and affect becomes unreliable.
Remove the repeated interaction, the reputation mechanism, the enforcement, and instrumental cooperation disappears.
Move the boundary of self so the other person is outside it and identification breaks. Leave it in place and neither distance nor payoff structure touches it.
The monks had no affect and no incentive. What remained was identification, and it was enough.
What you mean is that there is a psychological mechanism—a consequence in your "thought experiment" of the monastic practice—that internalizes an "identification" between two members of the same group. What I don't see is a sharp distinction between "identification" between members of the same group and "loyalty" within the group.
This phenomenon of "identification," when differentiated from "affect," requires a peculiar psychological construct. It occurs to me that it could be an abstraction of affection, which doesn't require personal proximity between those who share an ideology of affection, if such an ideology could exist. I find it difficult to believe that an "identification" not equivalent to "loyalty" can occur without an affective element (one can be loyal to communism or to Christianity, but I cannot identify with Jesus, whom I love, in the same way as with the historical figure of Lenin).
We can love all our fellow human beings within an affective ideology (our faith is limited to believing in a life governed by patterns of affective and caring behavior), but at the same time, we only personally identify with those who share this belief in universal love.
Is it possible to construct such an ideology? It has never been done, but some Christians have claimed that "God is Love"... which some might interpret as meaning that God does not exist, but that Love does exist as an affective abstraction, which can be symbolized and therefore also psychologically internalized within a context of shared belief (ideology).
So far, no one has undertaken such an experiment. Believing in "Love" (or its equivalent behavioral pattern of benevolence) would imply believing in a personal learning process to achieve a model of behavior at all levels (at the economic level, this would involve altruistic behavior).