I'm all for rationalism, for being guided, so far as possible, by reason and (good) reasons. But when rationalism leads someone to underestimate the force and the importance of beliefs which they see as beyond reason, that's not so clever. Here's Jonathan Miller in conversation with Laurie Taylor. Miller:
And he [a questioner] said, "No, seriously, are you an atheist?" And I said, when I say I'm not an atheist, it's not that I'm an agnostic and I'm hedging my bets. It's just that I cannot think it is worth having a word to describe it. It's so trivial, my not believing in God in the present day. I don't believe in witches, but I don't call myself an ahexist.
Yes indeed, and I don't believe there's a zhnarp hiding in my desk drawer or anywhere else about town, yet I haven't adopted the label 'azhnarpist'. Religion, however, is different from belief in zhnarps - believe me it is - in being an all but universal feature of human societies and having attracted the passions of generations of human beings, to say nothing of having inspired architecture, art, music, literature, works of charity and some less benign things than that. It's kind of useful to have a word for not believing in its most central propositions.
Then there's this. Taylor suggests that the revival of religion in fundamentalist forms has confounded the expectations of sociologists, and Jonathan Miller replies:
But we have to remember that in the case of the Islamic fundamentalism it's associated with an objection to what they see as the military tyranny inflicted upon their people by the United States and by NATO. It's a politically motivated thing.
There you go - what comes back Taylor's way is sociology of the playpen. A religious phenomenon as widespread as this Miller deals with as an expression, essentially, of politics. Leave aside the Comment-is-Free level of the remark, but you might suppose from it that religion could have no independent power over people's thinking.