The success of religion may be the fault of non-believers (or, if you look at it the other way around, thank god for the atheists!) 

At least that is one interpretation of a recent individual-based simulation study of social evolution conducted by James Dow at Oakland University in Rochester, Michigan, and published in a recent issue of the Journal of Artificial Societies and Social Simulation (vol. 11, no. 2 2).

Dow built a simulation program (appropriately called evogod) that explored the question of how religion — i.e., a system based on passing along false or unverifiable information about the world — can spread in a society. There are, of course, several theories out there about the evolution of religion, falling into two broad categories: either religion is somehow advantageous and is therefore the result of natural selection, or it is a byproduct of other characteristics of the human brain and social organization.

The first possibility comes in two main flavors: the advantage could accrue to religious individuals (standard individual-level natural selection) or groups (invoking the more controversial mechanism of group selection). Dow’s study explores the possibility that religious belief spread because of an individual advantage of some sort.

The first interesting result from the simulations is that under most tested scenarios religion actually does not survive!   This is presumably because there is an obvious cost (in terms of sheer Darwinian fitness) to buying into fanciful notions about how the world works. How is it possible, then, that practically every human society has gotten the religious virus?   The most surprising result of Dow’s study is that religion spreads only if non-religious people help it by supporting the religious!

How is this possible?

The simulation’s structure was not designed to address the question of what mechanism could induce non-religious people to help religious ones, but some possibilities have been suggested nonetheless.

According to Dow, “if a person is willing to sacrifice for an abstract god then people feel like they are willing to sacrifice for the community” (the so-called “greenbeard” effect). This is a social version of a well-established evolutionary idea known as the “handicap principle,” where males who can parade useless and costly attributes (be they peacock's feathers or Ferrari sports cars) are more likely to attract females because they are sending the indirect signal that their genes are so good that they can waste energy and resources just to please the female. It attempts to induce the female to imagine what sort of offspring they might be able to produce if only the female would consent to ...

As bizarre and irrational as this sort of scenario may seem, there is independent empirical evidence, for instance from studies of Israeli kibbutzim, that religious people do tend to receive more assistance than less religious ones from the rest of the community, again perhaps because they inspire trust.

Ironically, of course, this trust originates not because the religious provide more truthful information about the world, but precisely because they display a high degree of commitment to delivering non-verifiable information!

Humans, you’ve got to love them.