I have been lucky that my last two book purchases both proved interesting--I think I saw both mentioned on Arts and Letters Daily, so it's not all luck. I have mentioned Frederick Crews' Follies of the Wise. I just finished Nicholas Wade's Before the Dawn, which reports on recent techniques for using genetic analysis to take a fresh look at issues in anthropology.
On religion, Wade writes,
These sacred truths are unverifiable, and unfalsifiable, but the faithful nevertheless accept them to be unquestionable. In doing so, like assemblies of the faithful since the dawn of language, they bind themselves together for protection or common action against the unbelievers and their lies. (p. 165-166)
Think about this in the context of the matrix that I proposed here, in which some ideas are popular but not empiricist. What Wade is suggesting is that when you and I share religious beliefs, it is easier for us to trust one another, even though we are strangers. It would make sense that non-empiricist beliefs would serve as better trust cues than empiricist beliefs.
Wade says that people need protection against cheats and liars in order to deal with strangers. One sanction against cheats and liars is the threat of excommunication. By espousing religious beliefs, I signal that I care about being excommunicated, and so I signal that I am less likely to lie or cheat.
Obviously, this is a pretty tenuous hypothesis. But something to consider. I certainly believe that in business conversations and political discussions among friends, a lot of communication consists of "trust cues" that are designed to enable people to be comfortable with one another. In fact, among my liberal friends, bashing President Bush is such an important "trust cue" that it seems to be a necessary part of every conversation.
I should also point out that Wade says that "Modern states now accomplish by other means many of the early roles performed by religion" (p. 164) and that the role of religion "is now supplied by many other cultural and political institutions" (p. 168)