Posts Tagged ‘axelrod’

Cooperation (2): The road to hell is paved with good intentions

February 24, 2008

To paraphrase French writer Gide, bad cooperation is done with beautiful feeling. More exactly, cooperation is not in the realm of sentiment. The example of the software tournaments organized by Axelrod proves it well enough: programmed robots with no emotions can demonstrate cooperative behavior. We will even see in a future paper that passion can thwart cooperation. Nevertheless, man is made of passion as much if not more than of reason. Whether we are delighted or must resign ourselves to it, it is that way. Thus the emotive dimension has necessarily an impact on cooperation. If cooperation is not about intentions, it is about communication of intentions. It is not only a matter of cooperating by acts but also of making known one’s intention to cooperate.

The plot thickens when we consider that cooperating with someone consists in carrying out an action favorable to his/her interests and that, conversely, a hostile behavior must be punished immediately by a hostile behavior of the same intensity. The problem is that everything is a matter of judgment:

  • how can I know the interests of my partner?
  • am I even always aware of my own interests?
  • if I punish a behavior which I consider hostile whereas it was neutral in my partner’s mind (error of judgment), my punishment is likely to be misunderstood
  • does there exist a scale which makes it possible to compare the intensity of the cooperative or unfavorable acts?
  • When we try to function in a cooperative mode with our partners, we should act in such a way that our actions are as understandable as possible. That requires certain competences in psychology and communication, to convey the right message: I acted in such way with you and I explain why to you.

    Cooperation (1) : tit-for-tat

    February 16, 2008

    This paper is the first of a series of three articles originally published in French on cooperation and its issues in everyday life. In the present article, we will reconsider cooperation theory as brilliantly studied by
    Robert Axelrod . Among other achievements, Robert Axelrod is the author of the book The Evolution of Cooperation.

    The book reviews various strategies of interaction between members of a group. Among these strategies, there is the model called “tit-for-tat”. The principle of tit-for-tat cooperation consists in interacting in a cooperative mode with a partner as long as he/she does the same. At the first defection, the model suggests reacting immediately and with the same intensity; then, at the next turn, re-proposing cooperation. In any event, with this model, one never attacks first, one reacts to the first aggression and then starts again in a cooperative mode.

    In his book, Robert Axelrod showed that this strategy was effective and could be essential, even in absence of friendship or love, on the sole strength of interests well understood by both parties. This model brings a long-term value and lasting benefit to human groups.

    Axelrod describes and analyzes tournaments organized in the 1980s between artificial agents (small software), each having its own strategy. Some always cooperated, whatever the other’s behavior. To some extent, even when they were hit, they turned the other cheek; others always cooperated to a certain point where they deserted, “betraying” their partners; still others had a random behavior, etc. In the end, the model known as tit-for-tat proved to be one of the most robust.

    As human beings, we can choose this strategy consciously, without being programmed. But is it so simple? Of course not. Next week, we will start surveying limits which hinder us in our willingness to cooperate.