I am currently reading George Lindbeck, The Nature of Doctrine, in its 2009 25th anniversary edition. It is a very important book and is one of the seminal works of the "Yale School" of theology. I intend to review it properly after I have read it, but long before reading it I have encoutered and used Lindbeck's typology of theories of religion. Like all typologies its usefulness comes from the ease with which you can use it to categorise the things you come across, but also, again like all typologies, you have to be careful to realise that it is not in itself neutral but comes with its own implicit assumptions and agenda. For Lindbeck you can approach religion (or doctrine) in one of three ways (although you can also combine two or more of these ways to different extents):
1 The congnitive-propositional
This first approach assumes that religious language is concerned primarily with propositions of "fact". This tends to treat religion much like a philosophy or a science. This is the default setting, i think, for most evangelical theology if not for practice.
2 The experiential-expressivist
This approach is (a part of) the legacy of Schleiermacher who understood the essence of religion to be "religious experience." As such religion and doctrine are purely symbols which we use to make sense of and organise inner feelings. If a Buddhist and a Christian have exactly the same experience they will describe it very differently, using the symbols of their respective religions but they will still be actually having the same experience. Thus under this model you can argue (but you don't have to) that all religions are essentially differing attempts to give expression to the same reality. This is of course the default setting for much liberal theology (e.g. Tillich).
3 The cultural-linguistic
This approach (and this is the one that Lindbeck will champion in the book) instead aims to treat religions as a culture or a language. In this model doctrines are neither propositional truth claims (model 1) or arbitrary symbols (model 2) but instead rules or regulative principles for the discourse, attitudes and actions of the religious community. This approach to religion is both an attack on liberalism (this model does not allow for all religions to be essentailly the same) but at the same time those used to more conservative patterns of thinking can'y help but feel that their presious doctrines are being shortchanged and relativised. Probably the best know exponent of a cultural-linguistic approach to Christianity is NT Wright with his "Five act play" concept of biblical authority.
As you can hopefully see, regardless of how you might feel about the validity of each model per se this is an extremely useful way of analysing what is going on behind different people's discussion of religion. You may, for example, save yourself a lot of time and effort by realising that the person you are trying to engage on a cognitive-propositional level only considers it possible to talk about religion in experiential-expressivist terms. Most of us, i believe actually use a mixture of all three in our actual religious life, even if intellectually we tend to only allow one model primacy.
let me know what you think :-)