A very satisfying first discussion in my World Religions in Historical Perspective course at the University of Guelph has just concluded. When I wrote this course about six years ago, I labelled a category of religion 'primal'. In the proposal at that time to the university Senate committee I noted that this was a far from satisfactory term, but that I could not think of one better.
In the discussion, I asked the students to supply a critique of the term, and suggest alternatives - which produced a lively and scholarly discussion.
But what I wanted to consider here is not the alternatives suggested [though I will mention one: non-literate or oral], but to ruminate on the usual practice of studying religion using the tools of scientific thought.
Newtonian science has flourished providing the world with all sorts of physical and material comforts and aids. That its fundamental precept of positing an underlying order to the universe has been challenged over the past century or so makes no practical purpose in the provision of a multitude of gadgets for all things. But, from the time I wrote this course, I began to doubt the efficacy of the methods of science in bringing a good understanding of human society in general, and of religion in particular.
Standard world religions text books [and I have a slew of these sent me free by eager publishers] have chapters on Judaism, Christianity, Islam, Buddhism, etc. etc. - each one hermetically sealed from the other. Each is written by an expert in [and often follower of] the particular religion in the chapter. The book's editor or editors try to tie them together in a general introduction which talks about the nature of religion in general - and sometimes in a concluding chapter. The overall impact, however, is of isolation - of each religion existing in isolation from the other - even where something must be said about other religions in one of the chapters, the analysis is usually perfunctory, and surprisingly to me, wrong. Even an excellent general study of one religion, such as Arthur Goldschmidt Jr.'s history of Islam 'A Concise History of the Middle East', which does present religions and other forces which had an impact on Islam accurately, still has misunderstandings, or perhaps it has a sense of preferring Islam to any other. [The link is to the 8th edition - the 7th is available in the Trellis system as an eBook, and there is an audio version 9th edition which is superbly done, also]
The only text which has come close to avoiding this problem of categorization is the now old  second edition of 'The World's Religions' by the late Ninian Smart. Perhaps because it is a book with only one author, I don't know. In this book, categories and definitions and glossaries are abundant - but he attempted to show the profound and frequent and regular interplay of religions one with the other. It does not attempt to separate religions into their own houses, but to present them within cultural context.
I think that dividing religions into discrete categories presents a false picture of the reality of religious institutional structures, of individual faith, of doctrines, of history, of theology, of philosophy - not because each of these approaches produces wrong data - but because the whole is indeed greater than, and more importantly, different than, the parts. Religion to be grasped by the inquiring mind, must first be looked at in whole, not part. The proper method would be what I call the Sound of Music approach - like that venerable movie which begins in the air, and the camera angle gradually descends from the big picture, to the image of one person on the ground, religion should be studied first in its entire context, then one should begin to look at the individual facets.
I don't see, however, how this could be effected using a printed text, or a lecture class, or even the online course as I have it structured right now - which is basically a printed book, or series of lectures, appearing on a computer or mobile device screen, with links and pretty pictures as enhancements. I would like to devise a means where students [including myself here as a permanent student] could see the whole, then delve down into any particular facet or part - to surf the site, that is. If you wanted to compare death rituals across religions, you could do that... or if you wanted to look at a particular religion by itself you could do that too - but the course would not be structured in a linear fashion, except where that made sense, such as a narrative history of change over time.
A possible model is found in a series of multimedia eBooks from a partnership between the American Historical Association and Columbia University Press.