"Surely what makes religion religion isn’t going to be the presence of certain empty words. But then why won’t birthday parties be religious ceremonies? Why not high school graduations? If people wish to study rituals and ceremonies independent of whether they occur in a religious context, that seems a good idea. But it’s hard to believe that religion extends as far as ceremony does."

That's the question. So maybe what I'm proposing is a different methodology: starting from the other end so to speak--not asking for conditions under which beliefs about some grand SR that grounds value, "world view," wisdom literature or ethical system counts as religious but rather asking for conditions under which ceremonies, fetishes, and cultic practices count as religious.

But more substantively I'm puzzled by the suggestion that "human religiousity tends to evolve" in the direction of satisfying your definition.

(1) First, there's the question of identity conditions for religions. Even if a few members of the population, say Plotinus, believe in a supramundane reality that grounds value, etc. most of those Greeks are still worshiping stocks and stones. OK you can say: their practices count as religious because the beliefs and practices of some who practice the same religion or belong to the same cultural community or whatever satisfy your conditions for paradigmatic religion. But then of course there's the question of what counts as same religion or same cultural community for your purposes. The religion of the masses hasn't evolved. Can you make the case that it counts as a primitive or degenerate form of the same religion or that they belong to the same community in the requisite sense, without begging the question?

(2) Secondly, as a matter of empirical fact, one could make the case that religions tend to evolve in the direction of ceremonies and traditions without any connection to any supramundane reality, whether grand or trivial. Greco-Roman religion among the elite evolved in this direction: the gods became allegorical figures, representations of various virtues, and religious ceremonies became civic rituals. Or look at contemporary Judaism. Only 48% of Americans who give Jewish as their "religious preference" profess to believe in God. Of the other 52% I grant many are just giving an ethic identity. But others do engage in various ceremonies and communal practices comparable to what the elite Romans were doing: affirming communal identity, engaging in rites of passage, etc.

Again you can say that this represents the demise of religion rather than its evolution. But why pick the point at which, for an elite, ceremonies become associated with "world view," wisdom, ethics and the vision of a grand supramundane reality as the terminus of religious evolution rather than the point at which the practices of the community become civic rituals or affirmations of communal identity without commitment to any supramundane reality, whether grand or trivial--other than an interest in taking the snapshot of religious evolution at its height in the place where it satisfies your theory?