That would be a different argument, though. As I say in the article, evolutionary functions can be abused or overextended, in which case pragmatic value may not longer be joined with the trait’s naturalness.

You can argue that religion is socially useful, but that utilitarian calculation would likely be incomplete. How do you weigh religion’s benefits against its drawbacks?

Also, do you still benefit psychologically from religion if you know you’re being pragmatic about it? This is a problem with Pascal’s Wager too. If you must fool yourself somehow into believing you’re not just being pragmatic, and that the religion’s creed is true, this self-deception will have its downside as well as its upside. For instance, the credulity will transfer over to politics, as we see with Trumpism. So the badness of Trumpism would be relevant to a pragmatic assessment of religion.