ribbonsofnight wrote:If this scenario perfectly satisfies the conditions of the prisoners dilemma something is wrong with the relationship.
No kidding: the PD requires that the two parties know nothing about each other except that they are "human". The Iterated PD, which is not the PD, allows knowledge of past behaviour.
The analysis of the PD turns on what is meant by "human". The traditional analysis in terms of strict economic rationality turns on a contradiction: it pretends to ask "What would a strictly 'rational' being do in this situation" and then behaves as if BOTH players are not strictly rational beings. As if one player was completely different in every respect from the other. This is weird, because it is then no longer an human game. The other player could be a random number generator for all you care.
This is important because if all you know is that the other player is economically rational (ie that "human" == "homo economicus") you know that they will behave identically to you. The off-diagonal elements of the pay-off matrix simple vanish in a cloud of contradiction, and the while "dilemma" becomes a trivial choice.
It is only when you treat one player are a rational being and the other player as a magic unicorn or equivalent that there is any dilemma to be found. On this view the PD highlights the randomness and irrationality of human behaviour rather than demonstrating any particular issue for economically rational individuals living in an economically rational population: in such a population everyone would make the same choices faced with the same PD-type situation, because beings of the same kind behave the same way (that is, causality holds.)