Roberts (2005) Cooperation through interdependence

Roberts, G. (2005). Cooperation through interdependence. Animal Behaviour, 70, 901-908.

Abstract:
Cooperation is fundamental to human societies and widespread among animals, yet explaining how cooperative relationships persist without one party exploiting another remains a challenge. Although it is well established that altruism can be favoured when recipients are relatives or when they reciprocate, it is increasingly being recognized that many behaviours cannot be explained in these terms. Here, I argue that many cases of apparent altruism can be explained by interdependence. I formalize the degree of interdependence by defining an individual’s ‘stake’ in another as the dependence of its fitness on that of the other. This provides a means of valuing others, analogous to Hamilton’s valuation of relatives. It also provides a conceptual synthesis of theories of altruism, in which kinship and reciprocity become special cases. Interdependence means that cooperators can benefit as a secondary consequence of helping their recipients. Altruism can then be favoured when its costs are outweighed by the altruist’s stake in the recipient’s benefits. Whereas the exploitation problem makes reciprocal altruism inherently unstable, cooperation through interdependence can be stable because whatever others do, it is best to cooperate. I discuss the extent to which interdependence can explain examples of cooperation that cannot be explained in terms of kinship and reciprocity. 

Routes to cooperation:

• Kinship
• Relatedness does not explain much intraspecific helping in both humans and nonhumans
• E.g., cooperative breeding, social insect behavior
• Reciprocal altruism
• Empiricists have questioned its assumptions
• People are more cooperative than rationality would predict
• Human friendship goes beyond what is predicted by reciprocity.

There is a gap between the current state of theories and the evidence.
-We can develop models of reciprocal altruism with more biologically realistic assumptions.
[NOTE: This is basically what we are doing with the Sucker's Payoff model]
-There is still a need for alternative models not based on "trade" or reciprocity.

Interdependence

• An altruist might not benefit from the active return of a benefit, but as a secondary consequence (pseudoreciprocity or by-product mutualism)
• Many apparently altruistic acts have as their beneficiary an individual in whose welfare the individual has some stake.
• Altruists can be expected to benefit as a secondary consequence of their behavior.
• Therefore, cooperation could be favored without reciprocity. 

The altruist A receives some stake s of the benefit b that goes to B, minus the cost c. So cooperation we get a form of Hamilton's rule of the form sb – c > 0.

• Kin selection occurs if s = r. 
• Reciprocity if s is the probability of reciprocation.
• When reciprocity comes from beneficiary's offspring, stake also encompasses B's reproduction.  
• When A's benefits come from being observed to be altruistic, the stake encompasses all observers.
• Most interesting is the possibility that even if a recipient is unrelated and does not reciprocate, an altruist may receive secondary benefits (NOTE: this is also covered by pseudoreciprocity).

This is then given a nice game theory treatment.