Mike over at 2 blowhards has a post up where he muses upon religion, religious feeling, and traditional vs. "quasi-religion" (economism, scientism, etc.). I have read a lot of books on religion in the context of psychology, sociology and biology in the past year. Here are the general viewpoints that I found:

• Religion as a byproduct of the interaction of specialized cognitive domains. This is the sort of idea expressed in Scott Atran's In Gods We Trust and Pascal Boyer's Religion Explained. There are other other prominent thinkers, like Steven Pinker, who also support this position. This explanation tends to highlight the importance of universals, that is, the common religious sentiments of all human populations, and builds up from first principles, and can explain much of the "mystical" and "emotional" impact of religion.
• Religion as a functionally oriented social adaptation. This is the argument behind Darwin's Cathedral, by David Sloan Wilson. Wilson's ideas owe a lot to early 20th century thinkers like Emille Durkheim who emphasized the utility of religion to society, and even wondered if religion was not an expression of the society itself. This sort of thinking emphasizes the importance of social constraints and internal regulators in shaping religious life, through ritual & dogma. Because Wilson's ideas are often set in the context of "group selection," he can explain the differences between cultures (as religions are "selected"), but tends to look at each human as a "black box," and so neglects internal dynamics in a society caused by variation from person to person.
• Religion as a set set of supernatural goods & services. This is Rodney Stark's position, elaborated in A Theory of Religion, and suffusing his numerous popular books. Stark uses "rational choice theory," and posits the existence of religious "firms" who compete for "clients" (believers) by offering supernatural goods and services (afterlife, absolution from sin, etc.). There are various types of firms, for instance, those who aim toward individual transactions, like many magical religions that emphasize the priest-believer interaction, or those who focus on "collective action," such as congregational faiths. Stark's thesis explains the dynamics within a society as well as between societies. Stark uses an individualist paradigm, and though he does not elaborate on the psychological factors in detail like Atran, he takes for granted similar motives and underpinning factors, though ultimately, like Wilson, he might view religious feeling as a "black box" where only the consequences matter. Stark also implicitly integrates the idea of a social organism like Wilson when talking of faiths that eliminate the "free rider" problem through social constraints and tit-for-tat behavior.

Of course, there are many other theories, though not as fully fleshed out in my opinion. Jared Diamond for example seems to express the opinion that organized religion is a property of hierarchical societies where the elite uses faith to mainpulate the masses. Some believe religion is a psychopathology, or at least, an unconventional psychological state that humans need to explain or systematize. And so on. Another common idea is that religion is a pre-science, a way of explaining the world.

As time passes, I am moving toward a synthetic view. Atran & co. have convinced me that religion is "normal," that is it an emergent property of various cognitive domains, so there is no "God module," nor is there a particular gene that causes the "abnormal" mental states that might characterize religion (I believe mystical ecstatic experience is probably atypical, no matter the charisma these individuals accrue to themselves). Stark's thesis works in the context of places like the United States where horizontal change, that is, conversion, is rife, and people are relatively free of social constraints or stigma when switching religions. Wilson's ideas are more appropriate to organically integrated societies like India, or with minorities that are insulated by social distance from the majority, like Jews. Diamond's idea that religion is used by the elite to manipulate the masses is obviously true, but it seems likely religious feeling pre-dated complex societies, and I see no evidence that more stratified societies are "more religious" than hunter-gather peoples, unless you define religion has having to do with temples, priests and complex theological dogmas. There is some truth to all the above viewpoints, and I think one thing that they indicate is that some things that might be considered "religion" or fillling the "religion-shaped-hole" don't really fall under that umbrella.

For example, I have often wondered if Communism could be thought of as a "religion." Diamond and Wilson might not object, in that it basically seems to be a system that the elite uses to manipulate the masses. There is collective action and sacralization of founding fathers. It does have "religion like" tendencies. But, it seems that after the fall of Communism as the dominant political ideology in places like the former Sovient Union, it had few adherents. Where Stark would say that Communism could not provide "supernatural goods and services," Atran would say it does not satisfy the need for "collective identity" solidified by a "counter-intuitive supernatural agent." I think Stark and Atran are basically right, if we look at humans as "black boxes," rather than their psychological motives and needs, Communism seems like a religion, but it mimics certain aspects of large organized religions, rather than being rooted in religious instincts.

The "theories of religion" might be less important for a positive evaluation of what is a religion, and how religions interact, than in eliminating pretenders to the throne, and excluding certain theoretical models.