Sex, Love, and Anger: on the evolutionary emergence of emotionally motivated gaming strategies

Emotions obviously play an important role in human behavior. This study seeks to determine why emotions may have evolved. It is speculated that one’s genes seek to influence the individual beyond merely the construction of the individual’s body at birth. Game theoretic methods have been employed, to search for Evolutionary Stable Strategies (ESSs) that would make it more likely for an individual to pass its genes on to subsequent generations. The Genetic Algorithm (GA) was used to search for evolutionary stable strategies, by evolving a population of competing strategies. Two different gaming scenarios have been considered: aggressive/aggressive-avoidance behavior and mating behavior. Results have been found that are consistent with human behavior. For instance, males and females formed relatively long-lasting relationships that fostered the development of offspring through the age of maturity. Promiscuous relationships also emerged, though favored much more heavily by males than by females. An important (and potentially new) result was found, linking the onset of female menopause to the dependence of infant survivability on the care of its mother. Introduction and Overview It appears to be relatively difficult for a person to control his emotions. If someone unexpectedly slugs you in the face you quite rightly become angry; if you meet an attractive male or female you might very well fall in love. But why become angry; why fall in love? It is our contention that anger, love and other emotions are evolved responses and provide an adaptive advantage to an individual. Emotions are your genes’ mechanism to influence your behavior in ways that have been found, over evolutionary time, to be advantageous. This theory has been examined by running a series of games to determine if Evolutionary Stable Strategies (ESSs) emerge that are consistent with human behavior. While discovering strategies that are stable will not prove the theory (there is likely a learned component to emotions as well), it can at least point out why certain, emotionally influenced behaviors may be valuable to an individual. Two game scenarios have been considered. The first one examines aggressive behavior to determine strategies that are advantageous to an individual when confronted by the threat of aggression. For example, it may not be a wise strategy to always turn your tail and run when threatened, because if you do, a bully can easily take advantage of you. The second game examined is that of mating and subsequent reproduction. We are interested in answering questions such as, what sex should I choose for my child, how long should I stay married, and should I be promiscuous. Note that one doesn’t get to choose the sex of a child, so strictly speaking emotions can’t affect this “decision.” But the genes do choose the sex of a child and fundamentally this study is about the evolution of genes. It is traditional in game theoretic studies, for a set of strategies to be determined by an expert analyzing the system, and then a cost/benefit analysis performed to determine the evolutionary stable set of strategies. However, for anything other than the most simple of systems, no claim can be made that the original set of strategies was in any sense complete. It is a goal of this paper, especially with regards to the aggressive behavior game, to examine a more complete set of strategies to determine if initial determinations of ESSs (see, Dawkins 1976, pgs 67-76, for example) are in fact, evolutionary stable, or if they can be attacked by strategies not originally considered. For both of the games examined, a population of randomly chosen strategies has been evolved using the Genetic Algorithm (GA) (Holland, 1975). Individuals receiving higher scores in competition with other individuals have a better chance of passing on their genes to subsequent generations. Discoveries of new strategies are obtained through the use of standard mutation and crossover operations. As the population evolves over successive generations, the highest scoring strategies tend to take over the population and a set of stable strategies should emerge after many generations of the genetic algorithm. But do stable strategies in fact emerge? Perhaps there is some detail in the randomness of the genetic algorithm that inhibits the discovery of an ESS. But if ESSs can be found with the genetic algorithm, then a statement can be made regarding the possibility of evolution affecting our emotions. Since the genetic algorithm is patterned after sexual reproduction in nature, then if the genetic algorithm can discover an ESS, then it is certainly possible that natural evolution can also learn an ESS. This would imply that there might be a component of our behavior that is an ESS, learned over the eons of evolutionary time. So the next time you get angry or fall in love, think back to the genes of your distant ancestor and wonder if perhaps your behavior is controlled, in part, by something your genes learned long ago.