Kin Selection, True Altruism, and Cooperation
Reciprocal Altruism and the Prisoner's Dilemma
Tit for Tat and the Iterated Prisoner's Dilemma
Cases of Tit for Tat in Nature
Neighbor-Stranger Recognition and Tit for Tat in Birds
The Evolution of Sociality and Eusociality
We have already dealt with many of the issues underlying the evolution of cooperation among closely related kin. The example we have spent most of our time on is the evolution of social insects. In the case of social hymenoptera, kin selection is accentuated by the generally higher degree of relatedness among sisters compared to other groups. The kind of sacrifices individuals are willing to make in the case of kin are easily understood in terms of Hamilton's equation for inclusive fitness. An individual may opt to aid kin because such an act would enhance that individuals total fitness if we consider the total number of shared genes that the individual has with the kin. As we will see below, not all social or eusocial animals are haplodiploid. Explaining the existence of societies in such groups may require other mechanisms. The evolution of cooperation among non-kin may be a step towards understanding the evolution of social groups in a wider spectrum of animals. "Cooperation" is an act that occurs between two unrelated individuals that has a net mutally beneficial result in the long-term, but the act may entail fitness costs in the short-term.
True altruism is defined as some kind of act that is detrimental to an individual's fitness and enhances another individuals fitness, but, the two individuals do not share a genetic relationship. In human societies, individuals appear to aid non-related individuals and in some cases it is clear that there can be no relationship between the giver and recipient of the altruist act.
Bob Trivers (1971) considered the case of true altruism and the conditions for the spread of altruistic genes. If an altruist distributed their acts randomly in the population, the altruist gene would ultimately disappear as cheaters could easily take advantage of the altruist, but not give back in return. The altruist must recieve some benefit from its actions and cheaters would give nothing in return. Cheaters would spread in the population.
However, Trivers also realized that altruists might spread if altruist did not distribute their good deeds randomly, but distributed them to other individuals that showed evidence of reciprocity. Under such conditions, a reciprocal altruism might evolve.
We will explore the issue of reciprocal altruism with a simple model of cooperation -- the Prisoner's dilemma.
Axelrod and Hamilton devised a game theoretic treatment of cooperation that illustrates conditions necessary for the evolution of cooperation. Consider two prisoners in a jail and that have been caught with stolen property.
The detective interviews the prisoner's separately.
The prisoners know that if they both keep quiet, there is not enough evidence to be convicted of theft, they just get one year for possession of stolen goods.
If both confess to theft they get 9 years in jail.
However, if one rats on the other who keeps quiet, then the stool pigeon goes free and the other prisoner gets 10 years (one extra year for not helping the police).
Let us draw up a pay-off matrix that describes the various consequences of confessing or keeping quiet for player A (the same matrix applies for B):
Player B | |||
| Player A | Cooperation | Defection | |
| Cooperation | -1 | -10 | |
| Defect | 0 | -9 | |
Let us consider all of the payoffs, where player A Defects given cooperation by B is termed D|C:
D|C > C|C > D|D > C|D
Player A should defect as D|C is the best payoff. However, player B should do the same. When both defect they get the worst payoff. In the long run then it pays to Cooperate as it provides the second highest payoff. Therein lies the dilemma. The entire justice system is set up in this fashion.
Two students end up with identical tests.
They know that if they both keep quiet, there is not enough evidence to be tossed out of class for cheating, they get 50 on the midterm, which is 50% of the final grade so they loose 25%. But given that they were good cheaters (they scored 100) they still pass.The proctor sets them up in separate rooms and offers them the following deal:
If both confess, they fail and get 0 on the test (i.e., they loose 50% of their grade).
However, if one rats on the other (who keeps quiet), then the stool pigeon gets his full score, and the other cheater gets 0 for the class (for not helping the proctor they loose 100% of their grade).
Let us draw up a pay-off matrix that describes the various consequences of confessing or keeping quiet for player A (the same matrix applies for B):
Student B | |||
| Student A | Cooperation | Defection | |
| Cooperation | -25 | -100 | |
| Defection | 0 | -50 | |
Let us consider the payoffs, where player A Defects given cooperation by B (short D|C)
D|C > C|C > D|D > C|D
Player A should defect as D|C is the best payoff. However, player B should do the same. When both defect they get the worst payoff. In the long run then it pays to Cooperate as it provides the second highest payoff. Therein lies the dilemma. They should cooperate.
Axelrod and Hamilton considered the following game in which players do not just go through one cycle of the dilemma but they go through repeated cycles of the dilemma.
Let us consider our test cheating game in which the player's know one another (of course because they cheated to begin with). And they take courses with one another throughout their tenure as students.
One of the solutions that did very well in this game was the Tit-for-Tat strategy.
Consider our students, they might get caught and loose grades, but they still pass the course. So if their partner cooperates, they should too. They keep up this strategy throughout school. If their partner defects then they defect, and end the association.
There are in fact three play movements that underlie the Tit-for-Tat strategy:
Nice in which both players cooperate on the first move of the game,
Retaliatory in which a player defects if an individual defected on the prior move, and
Forgiving in which a player cooperates with a past defector that now has chosen to cooperate.
In the long run, the nice tit for tat will be the best strategy, thus it pays to cooperate, providing that the individual recognition and descrimination exists to exclude the cheaters.
The evolution of cooperation via a tit for tat strategy makes it possible for unrelated individuals to participate in some kind of interaction. This act would form the kernel for many kinds of social interactions. Behaviors that underlie such reciprocal altruism could arise and spread in a population if and only if, altruists can distinguish cheaters. Indeed, in human societies this particular formalism is bundled up in many human transactions. Credit card companies and the credit card holder have tit for tat relations. If a credit card company does not loan money it does not gain from loaning. While a delinquent debtor does get a short term payoff, the bad credit rating makes it impossible to get money in the future. Thus, the debtor pays of debts and the credit card company takes in interest. The debtor gains the buying power for goods (e.g., a house) that they might not have otherwise. How about the animal kingdom? Are there examples of tit for tat?
We will look at the familiar Neighbor Tit for Tat in territorial birds. The evolution of tit for tat is contingent on a number of other behaviors such as individual recognition of partner in the cooperation. Many animals show interesting cooperations when dealing with predators. Predator inspection behaviors could reflect a tit for tit. An alternative for such behaviors is the selfish herd hypothesis which we will also explore. Finally, simultaneous hermaphrodites and reciprocal fertilization presents a unique example of tit for tat, in which the fitness costs and benefits of cooperation are tightly related to fitness.
In the case of territorial birds with a dear enemy relationship, each male receives a mutual benefit for such "cooperation". Recall that Rene Godard also showed that when one bird detects its neighbor in an inappropriate place (via a playback manipulation), the focal bird carries out a tit for tat song fest and escalates hostilities at the border with the intruder. This would be the retaliator move. The bird does settle down again after apparent "cooperation" from the neighbor in a forgive-and-forget move (though the bewildered neighbor must wonder why his neighbor got so upset to begin with).
Let us consider the pay off matrix:
Neighbor | |||
| Focal Male | Cooperation | Defection | |
| Cooperation | Cooperation-save energy by avoiding confrontations | Sucker's payoff looses territorial resources or females | |
| Defection | Temptation to gains resources and females and keeps its own | Mutual defection -- come up even | |
Again the payoff matrix satisfies the conditions for the prisoner's dilemma in that:
Temptation> Cooperation > Mutual Defection > Sucker's Payoff
Lee Dugatkin has explored the iterated prisoner's dilemma game in a curious behavior found in fish -- predator inspection. The scenario is as follows:
The risks
in of defection and being the recipient of the sucker's payoff are extreme.
In this photo the bass attacked the inspectors. Thus the sucker's payoff
could be death.
Dugatkin tested a number of alternative hypotheses in addition to tit for tat:
The dilution effect would be a tendency in prey to group together in large numbers and overwhelm the feeding capacity of the predator. At least some prey would surive.
The selfish herd is another idea we can attribute to W.D.Hamilton. Hamilton modeled the behavior of prey using the simple rule:
Dugatkin ruled out these competing hypotheses and used mirror experiments to test tit for tat of an individual against itself as well as with live conspecifics. Guppies are capable of recognizing and remembering their partner's behavior and employ tit for tat-like strategies over the course of many inspections.
An interesting example of reciprocity is to be found in fishes, the only known class of vertebrates to exhibit simultaneous hermaphrodites. Sea basses have been investigated by Fisher 1988. Clutches of eggs are portioned out by the pairs of sea basses. One partner deposits a few eggs, and the partner fertilizes them, they reciprocate in the roles.
Neighbor | |||
| Egg initiator | Cooperation | Defection | |
| Cooperation | Cooperation-all eggs fertilized | Sucker's payoff gets eggs fertilized but does not get to fert. | |
| Defection | Temptation just produce sperm | Mutual defection -- nobody parcels eggs | |
This game may be a prisoner's dilemma if:
Temptation> Cooperation > Mutual Defection > Sucker's Payoff
Produce Sperm> Fert Eggs > Nobody Parcels Eggs > Looses Sperm
Species of bass do reciprocate against cheaters. Mates wait significantly longer to parcel eggs to a partner that failed to reciprocate than to partners who had reciprocated in the immediate past.
The evolution of eusociality is the most extreme form of animal societies in that members of the colony sacrifice reproductive opportunities. Eusociality is found in systems that do not have haplodiploidy such as naked mole rats.
We will explore the evolution of animal societies in greater detail. We will see that harsh conditions in the environment which establish new breeding pairs or groups, will favor the evolution of sociality with regards to cooperative breeding. Relatedness may contribute to the evolution of sociality.
Consider the actual measures of relatedness in bees relative to the theoretical maximum of 0.75. Many species show estimates of relatedness far less than 0.75 and in some cases it is less than 0.50!
Thus, haplodiploidy does not invariably lead to high degrees of relatedness.
Consider termites which form colonies very similar to ants in their extreme worker specializations. Termites have warrior castes with spray nozzles similar to ants, and there is a single bloated, egg-laying queen in the termite colony that produces all the eggs.
Such a system could indeed have much higher relatedness if individuals breeding in the colony practised brother-sister mating or mother-son matings. Recall that such oedipal inbreeding can generate higher levels of relatedness which can rival the 0.75 of hymenoptera. Similar sib-sib inbreeding can produce progeny with 0.75 relatedness. In both cases sibs normally share 1/2 of their genes, but because of the probability of identity by descent, which is 0.25, the probability of shared genetic material is 0.75. If inbreeding is not a problem then inbreeding can readily promote the evolution of eusociality. In a species that has been inbred for a long time, delerious recessive mutations would have been purged from the population long ago, so inbred species may be prone to evolve sociality
It turns out that naked mole rats have a high degree of relatedness within a colony, and great differentiation among colonies. Such inbreeding might contribute to eusociality.
Spec sheet on the Florida Scrub Jay
In the case of Florida Scrub Jays, the helpers at the nest are invariably offspring of the pair that they help. The added fitness of helpers is a tremendous boost to inclusive fitness (see table in Alcock, p 516). The dense territories of the jays makes it next to impossible for a subordinate or young bird to establish territories.
Not all cases of sociality involve inbreeding. In acorn woodpeckers, they definitely avoid inbreeding.
Stacey, P. B., and W. D. Koenig (1984). Cooperative breeding in the Acorn Woodpecker. Scientific American 251(2):114-121.
