Emotions change the outcome of The Prisoners’ Dilemma

Two prisoners sit in separate interrogation rooms without the opportunity to communicate with each other. Both are offered release if they denounce the other prisoner, who will then receive three years in prison. If no one testifies against the other, they only get one year – but if both denounce the other, the result is two years in prison each. The scenario is repeated several times and after each round the prisoners get to know what the other has chosen.

Viewed from the outside, the best option is for both prisoners to remain silent, as the total prison time will then be the shortest. But from the perspective of each prisoner, the most rational strategy is to testify against the other to avoid the maximum sentence.

Emotions make games more complex

Image

“In classical game theory, one mainly studies what happens when players act on basis of rational principles and only care about their own benefit, such as maximizing profit. But that is really a behavior one can assume that computers have. In humans, the spectrum of motivations is much richer and what has happened in the past may affect one’s future choices. So, I wanted to study how emotions affect the outcome in this type of game”, says Lina Andersson.

In her dissertation, she lets the players in the prisoners’ dilemma care about their own benefit, but also about the outcome for the other. An emotional player can be friendly, neutral, or hostile and shift in mood depending on the other player’s actions. If one player cooperates, the other may become friendly and positive to material gains for the first player, while a hostile act might get a hostile response.

“In my study I can see that if the players have full information about what the other is doing, emotions help them maintain cooperation. But if they’re unaware of what the actions of the other, they start to guess and then the outcome becomes more complex”, says Lina Andersson.

Can help us handle decision-making

With two emotional players, less patience is needed for them to cooperate, but the emotions may also complicate cooperation if they make the players reluctant to punish each other after deviations, or if the players become hostile after a punishment.

“In the traditional prisoners’ dilemma, you can expect cooperation between the players if the game is infinite, but not if it is repeated a finite number of times. If the players have emotions, they may cooperate also in finite games”, says Lina Andersson.

Situations where there is a conflict between the interest of the individual and the common good are sometimes called social dilemmas and may be looked at as merely thought experiments. But they illustrate and can help us understand decision-making situations privately, in working life and politics.

“If, for example, two colleagues are to collaborate over a long period of time, you through game theory you can see that it is advantageous if both have full information about what the other is doing. If this is difficult to achieve, you can expect more conflicts and need to create structures to handle it”, says Lina Andersson.

How well does this type of theoretical model match reality?

“The model is based on a number of basic assumptions, tested in psychology and neuroscience. But you can also examine the predictions in experiments to, for example, evaluate if people are more prone to follow the paths in the games that are favoured by emotions”.

In her dissertation, Lina Andersson also examines how fear affects financial choices, for example in the concrete example of a so-called bank run, that is when people start withdrawing their money from the bank, which leads to the bank crashing, and how emotions affect us to write online reviews of products and services.

Read the full dissertation: Emotions in Game Theory: Fear, friendliness and hostility

Text: Lars Magnusson