Adding to the general melee that is the debate about the evolution of religious belief, here's some stimulating news about a computer simulation created by evolutionary anthropologist James Dow (Oakland University in Rochester, Mi):
and the original research paper can be found here: Is Religion an Evolutionary Adaptation?
I'll take some time soon to blog more carefully on the original paper, but immediately tantalizing was the general concept: a computer simulation of interactions between "believers" and "non-believers," with the believers eventually flourishing! How cool … but of course this all depends on the assumptions built into the simulation. How were "believers" and "non-believers" defined? What type of interactions were being modeled? And how was adaptive fitness defined?
Dow used a co-evolutionary agent-based model allowing for influences of both genetically determined behavior and learned behavior. Dow explains that
The simulated agents have both inherited and learned capacities to communicate "real information" vs. "unreal information." The assumptions here are that:
So "believers" are essentially defined as communicators of unreal information, and eventually fitness is manifest in terms of the number of offspring an agent leaves when it dies. Agent fitness is affected in the simulation, though, through the complex interactions (communications) among the multiple agents. Basically: agents that receive real information increase in fitness; agents who receive unreal information decrease in fitness. But an agent's tendency to communicate either type of information is affected by the type of information being received from other agents, a form of cultural learning.
In most of the simulations, the general population increases and the non-believers dominate the believers — under these basic conditions, populations increase, but the gene frequencies for real communication rise and those for unreal communication fall.
The interesting variant occurs, prompting the article title and the pop media coverage, when Dow plays with the greenbeard parameter that determines with whom an agent communicates. [see the Green-beard effect on wikipedia]. The simulations mentioned so far used greenbeard = 0, designating that the agents with whom an agent communicates is chosen randomly from a uniform distribution of the agents (i.e. no agents are preferred for communication). When Dow incorporates the greenbeard effect, then "If one agent sees that another agent is more likely to communicate unreal information, it can select that agent as a recipient of its communication with a greater probability." Basically, Dow simulates a situation in which agents are more likely to communicate with agents that themselves are more likely to disseminate unreal information.
Under such conditions, we see the population of believer agents eventually thrive.
Now, why would one implement such a greenbeard effect in the simulation? That's not clear, and makes a nice topic for another post.
Dawkins, R. (1976) The Selfish Gene. New York: Oxford University Press.
Dow, James (2006). The Evolution of Religion: Three Anthropological Approaches. Method and Theory in the Study of Religion, 18(1), 67-91.
Dow, James (2007). A Scientific Definition of Religion. Anpere: Anthropological Perspectives on Religion <http://www.anpere.net/ccount/click.php?id=13>.
Dow, James (2008). Is religion an evolutionary adaptation? Journal of Artificial Societies and Social Simulation, 11(22) <http://jasss.soc.surrey.ac.uk/11/2/2.html>.