Albert Tucker, a distinguished Princeton mathematician, was invited to give a lecture in 1950 at Stanford University. There, he first posed this dilemma:
“Two men, accused of jointly breaking the law, have been confined by the police in separate rooms. Each is told that:
1- If one of them confesses guilty, but the other does not, the first will receive a reward, … and the second will be punished.
2. If both confess, both will be punished.
At the same time, each has good reason to believe that:
3.- If neither confesses, both will go free.”
Although there was earlier work by Flood and Dresher on the subject, Tucker was the first to christen it the prisoner’s dilemma. The implications of this dilemma are manifold in politics, international relations, economics and even everyday life. The story of how it came about and some of its applications are recorded in the book The Prisoner’s Dilemma by William Poundstone.
In a dilemma, it is often the case that whatever the solution, it involves a loss, a sense of remorse or regret. The prisoner’s dilemma is problematic because it defies common sense, says Poundstone. I will discuss the options and implications of the dilemma below.
There are two options for the two players: cooperate or betray. If player A confesses, he chooses to betray his partner. There are two possibilities: a) that the other player B does the same, then the final result is bad, and both are punished; b) that the other player B does not confess, chooses to cooperate, and then the result is very good for player A -he is rewarded- and bad for player B -he is punished-.
Suppose player A chooses to cooperate and does not confess. There are two possibilities: c) the other player B does the same, then the end result is good – both go free; d) the other player B chooses to betray his partner and chooses to confess, then the result is bad for player A – he is punished – and very good for player B – he is rewarded.
It is one of the most famous strategy games in which the interdependence of the players is vital. Whatever player A does, the final outcome depends on what player B does and vice versa. There is an incentive to betray the other player, but if they both follow that incentive and betray, the outcome is bad for both players. If both cooperate, the result is good, but it is challenging to obtain this outcome because of the incentive to betray and the impossibility of communicating.
Some consider that the prisoner’s dilemma has no solution. Others have pointed to the fact that it varies whether the game is played once or more than once. If played once, there is a solid incentive to betray as long as the other party does not do the same. This cannot be guaranteed and is, therefore, a dilemma.
What is most interesting is whether the prisoner’s dilemma must be played repeatedly. Then, it is how one has come to justify the need for the pact, the agreement between the players. This has been a way of justifying the existence of legal norms that guarantee that agreements will be fulfilled.
In Hobbes‘ State of Nature, there was natural liberty but great insecurity; the law of the strongest ruled. Although there could be cooperative behaviour, there was an incentive to betrayal. This leads to a social pact that guarantees peace and security, ensuring the players cooperate and not betray each other.
In a case where it is discovered that the treasurer of a political party, for many years, has 50 million euros in accounts in Switzerland, there would be a prisoner’s dilemma at some point in this case. At some point in this case, a prisoner’s dilemma scheme would emerge: both parties would have a better outcome if they did not confess and settle, but there is an incentive for betrayal.
The prisoner’s dilemma shows that pacts or agreements must promote cooperative environments. If these do not exist, there is an incentive to betray, which does not guarantee the best outcome because of the interdependence of the moves. If both betray, the outcome is terrible. If both cooperate, the result is good. If one cooperates and the other betrays, the outcome is bad for the former and very good for the latter.
There are many readings and applications of the prisoner’s dilemma. One is that cooperation, in the long run, is best guaranteed by a pact or agreement. It justifies the need for agreements that foster cooperative frameworks.