About a week ago, Graham Harman posted a couple of severe disagreements with my previous post – ones which I consider serious, enticing, and complimentary in their severity. I’d like to respond, but I’m putting it off since my attention has been elsewhere lately and a good response would require some significant effort on my part.
Instead, what I have been working on recently is reading through Danièle Hervieu-Léger’s Religion as a Chain of Memory. In it, she argues that a good sociological working definition of “religion” is that it is a way of believing that appeals and/or is committed to the authority of a tradition, around the (continually reconstructed) memory of which meaning is generated for the individual and community. There are a number of aspects of her approach that I find very helpful, not the least of which is her insistence that her definition of religion (or the religious) is provisional and entirely dependent on concrete sociological analyses (to which she often refers). More interesting to me right now, though, is a related point she makes drawing on Jean Séguy’s reading of Weber. In following a sociological impetus to eschew substantial or ontological definitions of religion, she argues that a more appropriate goal should be
to comprehend changes in the sphere of religion, considered by way of its tangible socio-historical manifestations. … Hence attention needs to be drawn to the process of change itself. Definition (if the term can still be used unambiguously) is a dynamic concept, whose aim is not that of fixing the subject matter but of pointing to the lines of transformation around which it reconstructs itself. 
Hervieu-Léger underplays, I think, the potential of this “dynamic” approach to understanding religions. She articulates it in the context of an examination of the problem of finding the place of religion in “modern” societies, as this is indeed the overall object of her book. And this is one of the things I find wanting in this work (there are a couple more, which I might bring up in a later post) – viz., an ongoing assumption of a phenomenon or complex of phenomena called “modernity” whose distinction from the “pre-modern” goes more or less unquestioned (even when she aims to show how both the religious and what is not religious, and their relation, have not undergone any essential transformation with the advent of modernity). I would argue that understanding religion primarily in terms of change and dynamism would be productive in general, and not just within modern Western milieux. This would be so not because what we usually call religions are, with regard to their content or to the practices associated with them, committed in any special way to social or historical change (it should be obvious that this is not always – even usually not – the case). It is because religions do in fact operate as ways of believing, of orienting individuals and groups in the world, with reference specifically to the past and to the future in a particular way for which tradition is a good enough word.
So, primarily there’s a diachronic element to this approach (I want to call it a “differential definition of religion”), consisting of identifying a religion as such according to the way it relates to its past and to its future, as well as the way it changes over time. But there’s also a synchronic element that would have to do with a religion’s present, yet it would be no less differential. This comes out repeatedly, in fact, in Hervieu-Léger’s account (though not all that explicitly) as she deals with the relationship between religions or the religious (and here she includes “traditional religions” – Christianity, Islam, etc. – as well as newer religious movements or sects) and other spheres of modern society. A provisional, differential definition of religion in its synchronic aspect would take into account the network of internal and external relationships by way of which a religion is continually reconstituting itself in the present among an irreducible religious plurality. Such a re-constitutions way be affirmative or negative, productive or destructive, but they are inevitable. At no particular time or place, though – and this is the payoff of this approach – can any one manifestation, understanding, abstraction, or negotiation become the definition of what it is to be a religion, because religions are as such constantly being renegotiated.