It is a long-standing argument among religious believers that religiosity were necessary for morality. In a recent Trends in Cognitive Sciences article (requires subscription), Pyysiäinen and Hauser argue that morality can arise and indeed can be found without and before any religious education and thus religion is a by-product of pre-existing cognitive properties of the brain. Indeed, religion is not ubiquitous, as for instance the Hadza's religion has been described as 'minimal', and yet, cooperation and morality are - as in all human cultures - thriving. In fact, there is a clear negative correlation between socioeconomic status and supernatural beliefs, further arguing that religiosity is not really all that important for morality to evolve or to persist. Pyysiäinen and Hauser cite a series of studies in moral psychology showing that moral judgments for unfamiliar moral dilemmas are unaffected in individuals without any religious background. In their press release, the authors conclude: This leaves open some other, less social cognitive factors contributing to the origin of religiosity, to which to authors allude towards the middle of their article: "[...] the concept of God is based on extending to non-embodied agents the standard capacity of attributing beliefs and desires to embodied agents. According to this view, religious beliefs are a by-product of evolved cognitive mechanisms." The authors are referring to 'theory of mind'. Besides this, still social capacity, there are several other factors contributing to the origins of religion. One such factor is of course our concept of causality and our hunt for last causes. However, the factor that is, of course, closest to my own field of research is that religion works as an operant behavior. This means that religion, for instance, can provide us with a feeling of control where, ultimately, there is none (think rain dance). This is not counter-intuitive and so I'm not the only person who has realized that this may be an important contributing cognitive factor. There is even prior evidence that when experiencing or remembering an experience of lack of control, these cognitive capacities for imagining control and order are enhanced.

These insights leave us with a set of pre-existing cognitive abilities providing a fertile ground on which the evolution of religion could occur as a by-product: Our capacity to detect agency (so helpful in our social interactions that we see it even in non-living objects), together with the concept of causality imply that everything happens for a reason and that this reason is the intention of someone. This someone can be controlled using certain rituals as evidenced, for instance, by the rain occurring after a rain-dance. This someone obviously punishes you if you do not perform these rituals, so of course this someone will also punish you if you do not cooperate or otherwise violate the rules of the in-group. In this way, religion provides you with a sense of order and controllability in an uncontrollable world which, in turn, keeps you sane, your society functioning and thus competitive and alive. As one of the commenters on the press release noted, 'competitive' may be the key word here, with religion providing a further tool for promoting self-sacrifice and suicidal fighting which might have provided some particularly religious groups with a competitive advantage.
Methinks it's about time for someone to develop a computer model for the evolution of religion, the data are starting to provide enough parameters for such a project.

Also in reply to one of the comments on the authors' press release, a very pertinent video via Pharyngula:

---

Ilkka Pyysiäinen, & Marc Hauser (2010). The origins of religion: evolved adaptation or by-product? Trends in Cognitive Sciences : doi:10.1016/j.tics.2009.12.007