I concur (see below)
which I'm not at all sure of
> (warning!), but it may be that pre-literate societies,
> dependent on memory for transmitting ideas and information,
> tend to be very slow to change their world-views, etc.,
> because one can easily get confused between what one
> remembered one's elders said and the new ideas one thought of oneself.
That is a very idealistic and mentalist interpretation of history - asserting in essence that social change originates in human mind. I think that whatever evidence we have about pre-historical societies is that they were very diverse culturally precisely for the reason of the absence of written record. But as far as memory per se is concerned - these allgedly had brillinat memory cf. the !Kung bushmen who memorized th elocation of individual plants in the desert, whose roots they could retrieve even after the upper part was long gone. Werner Herzog in his film _Where the Green Ants Dream_ put that quit eemphtaically that the native herdsmen may not know how to count but they will immdediately notice absence of as single anaimal from their large herds precisely because of their excellent memory.
As far as the materialist explanation for religion is concerned, thebest I've seen is that proposed by Emile Durkheim in his _Elementary Forms of the Religious Life_ where is argues that religious is an a cognitive script (Knatial apo priori form - to which he makes specific reference) of social organization. From that perspective, religion is a language that defines meaning of certain sigifiers (hierophanies) - but unlike in the "secular" language where the relation between the signifier and the signified are pretty arbitrary, there is a certain level of isomorphism in hierophanies i.e. an object becomes a hierophany because its physical from resembles ir incarnates a symbolic concept it represents (e.g. a tree become a hieropahny of etrnal life because it undergoes seasonal changes, an dgrows bigger and stronger).
Wojtek